text,label,id
"presented here Furthermore, naive improved. implementation be the can","Furthermore, the naive implementation presented here can be improved.",0
vector a in a form vector multidimensional space. Those coefficients,Those coefficients form a vector in a multidimensional vector space.,1
compatible of The model with recent is model sequential a behavioral dynamics.,The model is compatible with a recent model of sequential behavioral dynamics.,2
"but relevance outlined. hemodynamics its based is The has been electrophysiology, on model to","The model is based on electrophysiology, but its relevance to hemodynamics has been outlined.",3
of transitions lever-like involve reorientation often the Conformational structures. complexes in macromolecular,Conformational transitions in macromolecular complexes often involve the reorientation of lever-like structures.,4
"an Surprisingly, fastest of hinge. is the transition flexibility the intermediate with","Surprisingly, the transition is fastest with an intermediate flexibility of the hinge.",5
largely on the configuration framework. CLab formalization The is based,The formalization is largely based on the CLab configuration framework.,6
"these document lines. is For segmentation a tasks, text into major all step","For all these tasks, a major step is document segmentation into text lines.",7
"satisfactory proteins when is always are the zone'. not Unfortunately, in 'twilight performance their","Unfortunately, their performance is not always satisfactory when proteins are in the 'twilight zone'.",8
"constructs a HMMER-STRUCT set a pHMMs. first As step, of","As a first step, HMMER-STRUCT constructs a set of pHMMs.",9
"Properties and packing. and tertiary were primary, secondary used structures, accessibility","Properties used were primary, secondary and tertiary structures, accessibility and packing.",10
the results voting. then by prioritizes HMMER-STRUCT,HMMER-STRUCT then prioritizes the results by voting.,11
our SCOP experiments. Results: the used database to perform We,Results: We used the SCOP database to perform our experiments.,12
"we leave-one-family-out Throughout, apply protein superfamilies. over cross-validation","Throughout, we apply leave-one-family-out cross-validation over protein superfamilies.",13
"used First, the structural training proteins. aligner MAMMOTH-mult we to align the set","First, we used the MAMMOTH-mult structural aligner to align the training set proteins.",14
"two of experiments. Then, performed sets we","Then, we performed two sets of experiments.",15
"against model individual we experiment, pHMMs. compared the a In second voting","In a second experiment, we compared the voting model against individual pHMMs.",16
mixtures in evolution inhomogeneous commonly Phylogenetic molecular observed model data. the,Phylogenetic mixtures model the inhomogeneous molecular evolution commonly observed in data.,17
can be that assumption Here this we false. show,Here we show that this assumption can be false.,18
a fast also find We pathway. very non-zipper,We also find a very fast non-zipper pathway.,19
risk is The the given estimate data. to proposed of demographic method HIV tested,The proposed method is tested to estimate the risk of HIV given demographic data.,20
approach field tumor in of the the modeling application The outlined. possible of is,The possible application of the approach in the field of tumor modeling is outlined.,21
"and In large growth an are speed, terms adaptive with both state a higher balanced.","In an adaptive state with a higher growth speed, both terms are large and balanced.",22
"cellular escape time Hence, cellular state from the rate growth correlated. are negatively and a","Hence, escape time from a cellular state and the cellular growth rate are negatively correlated.",23
with Ensemble increasing thus attention. Learning receives Computation Evolutionary,Ensemble Learning with Evolutionary Computation thus receives increasing attention.,24
Evolutionary paper contributions. in two this Ensemble The (EEL) features presented Learning approach,The Evolutionary Ensemble Learning (EEL) approach presented in this paper features two contributions.,25
"the co-evolution a presented. classifier by enforcing is inspired fitness new First, function, and diversity,","First, a new fitness function, inspired by co-evolution and enforcing the classifier diversity, is presented.",26
"classification on criterion proposed. the a new Further, based margin selection is","Further, a new selection criterion based on the classification margin is proposed.",27
perceptron a model rough multi-layered and This set. on proposes paper based neuro-rough,This paper proposes a neuro-rough model based on multi-layered perceptron and rough set.,28
risk neuro-rough on is model then HIV modelling tested of the data. demographic The from,The neuro-rough model is then tested on modelling the risk of HIV from demographic data.,29
an constraints paper we scheme propose that soft this for uses abstraction semiring homomorphism. In,In this paper we propose an abstraction scheme for soft constraints that uses semiring homomorphism.,30
suffer (NM) blur. and Nuclear noise inherently medicine images amounts of large from,Nuclear medicine (NM) images inherently suffer from large amounts of noise and blur.,31
Field image Annealing used processing in Mean the technique (MFA) this research. is,Mean Field Annealing (MFA) is the image processing technique used in this research.,32
act of to confounds this day. The designers bluffing game,The act of bluffing confounds game designers to this day.,33
ANNs was this In the better study SVMs. observed perform than the it that,In this study it was observed that the ANNs perform better than the SVMs.,34
measured the generalisation This performance is two. against the of ability,This performance is measured against the generalisation ability of the two.,35
based The are are rules and Lotka-Volterra in Markovian local the model.,The local rules are Markovian and are based in the Lotka-Volterra model.,36
model This for continuous Developmental introduces Multi-cellular a paper Design.,This paper introduces a continuous model for Multi-cellular Developmental Design.,37
"the exhibit resulting almost multi-cellular perfect characteristics. self-healing Moreover, organisms","Moreover, the resulting multi-cellular organisms exhibit almost perfect self-healing characteristics.",38
Militarised the risks society. conflict is significant of that on have impact one a,Militarised conflict is one of the risks that have a significant impact on society.,39
"the to research, previous implemented have (NNs) predict a neural MID. In been networks","In a previous research, neural networks (NNs) have been implemented to predict the MID.",40
"unknown. nucleus of path the the However, folding the fiber interphase in chromatin is","However, the folding path of the chromatin fiber in the interphase nucleus is unknown.",41
"the of efficiency recursiveness are and presented, the approach. experiments showing demonstrating this Two","Two experiments are presented, demonstrating the efficiency and showing the recursiveness of this approach.",42
"to proposed respect elementary the particular, the In robustness sensitivity with w.r.t. the and behaviors,","In particular, the sensitivity with respect to the proposed elementary behaviors, and the robustness w.r.t.",43
in are controllers resulting generalization the detail. of studied,generalization of the resulting controllers are studied in detail.,44
The technique is in proposed tested and regression both classification problems.,The proposed technique is tested in both classification and regression problems.,45
diagnostic the analysis. gas in the dissolve analyzed using bushings gases The are,The diagnostic gases in the bushings are analyzed using the dissolve gas analysis.,46
terms time RBF. performance than accuracy superior gives SVM and MLP in of and training,MLP gives superior performance in terms of accuracy and training time than SVM and RBF.,47
population cylinders the the have substructures. in Some faults of in different,Some of the cylinders in the population have faults in different substructures.,48
to the used leave-one-out test performance process model. is of the cross-validation A,A leave-one-out cross-validation process is used to test the performance of the model.,49
rhythms. Eupelmidae These have two similar activity,These two Eupelmidae have similar activity rhythms.,50
Displacements place take the essentially parasitoids photophase. during of these,Displacements of these parasitoids essentially take place during the photophase.,51
accurate introduce an hitch-hiking. genetic we for Using model this approximation,Using this approximation we introduce an accurate model for genetic hitch-hiking.,52
neutral bias. without us enables to This compute the on segment SNP distributions,This enables us to compute SNP distributions on the neutral segment without bias.,53
chains one be to remains network how However coexist elucidated. synfire in,However how synfire chains coexist in one network remains to be elucidated.,54
their Fokker-Planck analyzed We the activity method. by,We analyzed their activity by the Fokker-Planck method.,55
is real a data. earthquake for trained The first single system,The system is first trained for a single real earthquake data.,56
"model transforms. the that Fourier can ""non-ideal"" cable analytically resulting We solved using be show","We show that the resulting ""non-ideal"" cable model can be solved analytically using Fourier transforms.",57
the as to This state. critical state is referred,This state is referred to as the critical state.,58
"boost protective that Furthermore, activity be should interventions against spontaneous epileptogenesis.","Furthermore, interventions that boost spontaneous activity should be protective against epileptogenesis.",59
"a Mortality, major play role in and demographic birth rates  retirement  changes.","  Mortality, birth rates and retirement play a major role in demographic changes.",60
therefore and retirement both We increase the recommend employment. the of age women,We therefore recommend the increase of both the retirement age and women employment.,61
models. result rates-across-sites of we present positive Next identifiability a showing,Next we present a positive result showing identifiability of rates-across-sites models.,62
"paper, Muller-Linow a Hutt and Marr, recent In [Phys.","In a recent paper, Marr, Muller-Linow and Hutt [Phys.",63
the reanalyze We same system kind using of modeling reaction a actual kinetics. dynamic data,We reanalyze the same kind of data using a dynamic system modeling actual reaction kinetics.,64
"from al. about our conclusions with stability, inconsistent Marr of those are analysis, et The","The conclusions about stability, from our analysis, are inconsistent with those of Marr et al.",65
for We of detailed calls issue type more that this argue a modeling.,We argue that this issue calls for a more detailed type of modeling.,66
"may recently found best hits developed other by algorithms. be more or by The blast,","The best hits may be found by blast, or by other more recently developed algorithms.",67
the The by setted is of main production its conservation problem harvests.,The main problem setted by its production is the conservation of harvests.,68
"as (bruchids). are pests destroyed the the seeds stocks, in fields the by In","In the fields as in the stocks, the seeds are destroyed by pests (bruchids).",69
species associated hymenoptera. entomophagous These are always several of with bruchids,These bruchids are always associated with several entomophagous species of hymenoptera.,70
and E ovicide Basalis. vuilleti behaviour larvicide adopts D. against,E vuilleti adopts ovicide and larvicide behaviour against D. Basalis.,71
population regression inside explains its behaviour This granaries.,This behaviour explains its population regression inside granaries.,72
"alignment These and images are segmented, PCA. before firstly feature extraction using","These images are firstly segmented, before alignment and feature extraction using PCA.",73
"calculations. modern topological give networks Conversely, facilitating standardized high-throughput experiments rise to","Conversely, modern high-throughput experiments give rise to standardized networks facilitating topological calculations.",74
based with consistent recently experimentally-obtained are structural reported that subdomains These was variability on structures.,These subdomains are consistent with structural variability that was reported recently based on experimentally-obtained structures.,75
energy of We a is occurs generically site-dependent process. and show that spontaneous localization,We show that spontaneous localization of energy occurs generically and is a site-dependent process.,76
diffusion present environment. molecularly in crowded for a a model We,We present a model for diffusion in a molecularly crowded environment.,77
network. model random in percolation consists barriers The of,The model consists of random barriers in percolation network.,78
well environment. a as observe sub- super-diffusion Thus as crowded in we,Thus we observe sub- as well as super-diffusion in a crowded environment.,79
of patients ideal radiation treatment. large virtual following simulated A an is population,A large population of virtual patients is simulated following an ideal radiation treatment.,80
"A System Tumor (ISTER), introduced. parameter, -- Immune the Efficiency characteristic is Rate","A characteristic parameter, the Immune System -- Tumor Efficiency Rate (ISTER), is introduced.",81
studied. other dependence and features ISTER is treatment of success,ISTER dependence of treatment success and other features is studied.,82
allow Statistical a us scheme. to patient classification give results,Statistical results allow us to give a patient classification scheme.,83
present task. this a We for system developed,We present a system developed for this task.,84
membrane ion transport cell of We through various models discuss channels.,We discuss various models of ion transport through cell membrane channels.,85
on equations. macroscopic not Poisson-Nernst-Planck continuum by based is models satisfied This,This is not satisfied by macroscopic continuum models based on Poisson-Nernst-Planck equations.,86
Brownian and ion are offered of dynamics. by More transport microscopic descriptions molecular realistic,More realistic descriptions of ion transport are offered by microscopic Brownian and molecular dynamics.,87
channel should a One character the take account structure. dynamical into of also,One should also take into account a dynamical character of the channel structure.,88
. yet the is not This addressed in literature,This is not yet addressed in the literature .,89
methods characterize impact We here spike to activity this and different generation. its on review,We review here different methods to characterize this activity and its impact on spike generation.,90
an variance determining for conductance points synaptic essential analysis times. spike to in role This,This analysis points to an essential role for synaptic conductance variance in determining spike times.,91
in have of proposed the for temporal models been such Lots information. litterature representing,Lots of models have been proposed in the litterature for representing such temporal information.,92
interval-based or All a derive an from point-based framework.,All derive from a point-based or an interval-based framework.,93
models rules-- inference require similarly for solving -- transitive such or tables All these tasks.,All these models require transitive tables -- or similarly inference rules-- for solving such tasks.,94
"implemented. EGFR-ERK was signal transduction level, pathway intracellular specific At molecular a the","At the molecular level, a specific EGFR-ERK intracellular signal transduction pathway was implemented.",95
of the to level. molecules at Dynamical changes were these phenotypic trigger cellular used alterations,Dynamical alterations of these molecules were used to trigger phenotypic changes at the cellular level.,96
function. for that algorithm a We finding a minimize conformations new non-bonded protein energy demonstrate,We demonstrate a new algorithm for finding protein conformations that minimize a non-bonded energy function.,97
data available are made sets These online.,These data sets are made available online.,98
whole unambiguously reasonable a HIV body indicates in eradication time. This,This unambiguously indicates a whole body HIV eradication in reasonable time.,99
layer Neurons receptive of input visual develop cortex primates in edge-like primary in the fields.,Neurons in the input layer of primary visual cortex in primates develop edge-like receptive fields.,100
"of role unknown. still is in largely the representations homeostasis sparse such learning specific However,","However, the specific role of homeostasis in learning such sparse representations is still largely unknown.",101
coding sparse is fair. optimized it when is Competition in,Competition in sparse coding is optimized when it is fair.,102
sparse excitatory neurons. of inhibitory connections populations and and separate,sparse connections and separate populations of excitatory and inhibitory neurons.,103
E. in proteins to oscillation Ongoing is Min required minicelling coli. block of sub-cellular,Ongoing sub-cellular oscillation of Min proteins is required to block minicelling in E. coli.,104
"Experimentally, newly and divided are Min minicells produced. cells no are in seen oscillations","Experimentally, Min oscillations are seen in newly divided cells and no minicells are produced.",105
this behavior been role of sedentary has determined. The in not defect,The role of sedentary behavior in this defect has not been determined.,106
connections models These have for and important consequences of dynamics protein thermodynamics.,These connections have important consequences for models of protein dynamics and thermodynamics.,107
plays role metric classification. The an important (NN) in nearest neighbor distance,The distance metric plays an important role in nearest neighbor (NN) classification.,108
"physiologically monostability. likely more That within ranges for multistability plausible becomes parameters, is, than","That is, within physiologically plausible ranges for parameters, multistability becomes more likely than monostability.",109
the often algorithms. clustering These outcome of are sets,These sets are often the outcome of clustering algorithms.,110
"analyses. provide such application, With web-based that database powerful a we OrfMapper supports but easy-to-use,","With OrfMapper we provide a powerful but easy-to-use, web-based database application, that supports such analyses.",111
last importance become has regulation post-transcriptional the The years. in clear increasingly few miRNA-mediated of,The importance of miRNA-mediated post-transcriptional regulation has become increasingly clear in the last few years.,112
experimental are proposed candidates Many verification. for new,Many new candidates are proposed for experimental verification.,113
frequently since problem been ever intermediates Darwin. The fossil has discussed the in record of,The problem of intermediates in the fossil record has been frequently discussed ever since Darwin.,114
to electrophoretic after polyacrylamide separation is on used staining detect gels. Silver proteins,Silver staining is used to detect proteins after electrophoretic separation on polyacrylamide gels.,115
"analysis digestion. after such compatible as spectrometry It is with protein processing, mass downstream","It is compatible with downstream processing, such as mass spectrometry analysis after protein digestion.",116
"for stable completed, stain weeks. the several is Once","Once completed, the stain is stable for several weeks.",117
that of We PRC. is proportional prove the to derivative STA the the,We prove that the STA is proportional to the derivative of the PRC.,118
"Conversely, ducklings agonistic hens resting, towards behaviour and increased. of preening","Conversely, resting, preening and agonistic behaviour of hens towards ducklings increased.",119
random the depth. paper this Tukey which a propose In approximates we depth,In this paper we propose a random depth which approximates the Tukey depth.,120
"Thus, high small computation in very requires time depth spaces. dimensional even random this a","Thus, this random depth requires a very small computation time even in high dimensional spaces.",121
"easily framework. is functional the to cover it Moreover, extended","Moreover, it is easily extended to cover the functional framework.",122
in We also this others compare proposed depth literature. the with some,We also compare this depth with some others proposed in the literature.,123
to by replacements and guanine cytosine lead damages thymine. These of,These damages lead to replacements of guanine and cytosine by thymine.,124
motor neuron high-fat increase Counterbalancing and with lifespan for a this prevents diet loss. extends,Counterbalancing for this increase with a high-fat diet extends lifespan and prevents motor neuron loss.,125
"Here, whether lipid investigated in model. this we animal metabolism defective is","Here, we investigated whether lipid metabolism is defective in this animal model.",126
animals. in high-fat induces defect the corrected This these typically neuroprotection by that was regimen,This defect was corrected by the high-fat regimen that typically induces neuroprotection in these animals.,127
of this dietary This protective lipids metabolic for in probably accounts effect the shift model.,This metabolic shift probably accounts for the protective effect of dietary lipids in this model.,128
in cell stochastic reactions fluctuations. Chemical subject to are intense,Chemical reactions in cell are subject to intense stochastic fluctuations.,129
withdrawn been has This paper by the author,This paper has been withdrawn by the author,130
"mathematical propose for approach programming we In the a new problem. paper, this","In this paper, we propose a new mathematical programming approach for the problem.",131
MILP optimal solution directly yields the an the joint-policy Dec-Pomdp. for of The optimal,The optimal solution of the MILP directly yields an optimal joint-policy for the Dec-Pomdp.,132
method. in problems PRF the approximation the infinite-sites inherent These from arise,These problems arise from the infinite-sites approximation inherent in the PRF method.,133
correct techniques three these introduce new problems. We that inference,We introduce three new inference techniques that correct these problems.,134
two decomposed to protein from Contributions the the VER and are contributions into water.,Contributions to the VER are decomposed into two contributions from the protein and water.,135
functions excitation are resonant angular pathways The radial modes. VER using normal and for visualized,The VER pathways are visualized using radial and angular excitation functions for resonant normal modes.,136
direct it. rewarding most of One of human promoting mechanisms is cooperation the,One of the most direct human mechanisms of promoting cooperation is rewarding it.,137
example important is model An Rvachev-Longini Math. the [{\it,An important example is the Rvachev-Longini model [{\it Math.,138
"data in ultrametric analysis are of hence spaces, nonarchimedean objects used geometry. Dendrograms","Dendrograms used in data analysis are ultrametric spaces, hence objects of nonarchimedean geometry.",139
of is that dendrograms. known It there representation exist $p$-adic,It is known that there exist $p$-adic representation of dendrograms.,140
"a of the end, At dendrogram. calculate the topology part of we the hidden","At the end, we calculate the topology of the hidden part of a dendrogram.",141
"case also unlike of in the trees They that, showed (i.e.","They also showed that, unlike in the case of trees (i.e.",142
in balance the paper three studies models. replication simplified This mutation-selection,This paper studies the mutation-selection balance in three simplified replication models.,143
of asexual model replicating spores. considers the first a production via organisms population of The,The first model considers a population of organisms replicating via the production of asexual spores.,144
sexually second gametes. model that replicating produces a identical considers The population,The second model considers a sexually replicating population that produces identical gametes.,145
that produces The sperm considers sexually third and a egg population gametes. replicating model distinct,The third model considers a sexually replicating population that produces distinct sperm and egg gametes.,146
"the develop spores the population, into asexual adult diploid asexual directly organisms. In","In the asexual population, the asexual diploid spores develop directly into adult organisms.",147
diploid organisms organisms. proceed into then develop mature The immature to resulting,The resulting immature diploid organisms then proceed to develop into mature organisms.,148
"moderate are note, their for speeds. As they valid calculations only however,","As they note, however, their calculations are valid only for moderate speeds.",149
"of Desai analysis and the Fisher's we edge. extend Here, stochastic substantially","Here, we substantially extend Desai and Fisher's analysis of the stochastic edge.",150
et compatible al. a predictions approach Our (Rouzine different made analytical with results using are,Our results are compatible with predictions made using a different analytical approach (Rouzine et al.,151
"chromosome along at called each multiple sites DNA replication is origins. In initiated eukaryotes, replication","In eukaryotes, DNA replication is initiated along each chromosome at multiple sites called replication origins.",152
"stable destroying hence Gaussian terminate cycle. transiently can the noise white the oscillations, Additive limit","Additive Gaussian white noise can transiently terminate the oscillations, hence destroying the stable limit cycle.",153
noise of spiking Further may return application the activity. system the to,Further application of the noise may return the system to spiking activity.,154
the parasite malaria by the accounts variation for exhibited falciparum). model The (P. antigenic,The model accounts for the antigenic variation exhibited by the malaria parasite (P. falciparum).,155
a exhibit show the wide dynamical We of behaviors. variety that can model,We show that the model can exhibit a wide variety of dynamical behaviors.,156
"provide for global competitive and criteria stability, exclusion, persistence. We","We provide criteria for global stability, competitive exclusion, and persistence.",157
the that We non-symmetric equilibrium cross-reactive destabilized by can disease be demonstrate responses. also,We also demonstrate that the disease equilibrium can be destabilized by non-symmetric cross-reactive responses.,158
of turbulence. the to analogous mechanism The cascade is development in the fluid,The mechanism is analogous to the development of the cascade in fluid turbulence.,159
diminished themselves. Locality bursts in the is,Locality is diminished in the bursts themselves.,160
changes Cerebral capacity. significant with in cortex oxidative in antioxidant no early impairment showed damage,Cerebral cortex showed early changes in oxidative damage with no significant impairment in antioxidant capacity.,161
"associated Recently, investigated to specific genetic regulatory process the transcription studies networks. several have","Recently, several studies have investigated the transcription process associated to specific genetic regulatory networks.",162
environments next collaborative is towards aspects. oriented for virtual training of The generation,The next generation of virtual environments for training is oriented towards collaborative aspects.,163
"new for this, We involving method a present doing simulation.","We present a new method for doing this, involving simulation.",164
useful the that Experiments provides method estimates. suggest,Experiments suggest that the method provides useful estimates.,165
role in consider different spatial the two We mechanism scales. on effects the,We consider the role spatial effects in the mechanism on two different scales.,166
quite near site. the transcription factor its dissociation lands,the transcription factor lands quite near its dissociation site.,167
"discriminate global local between two we searches: searches. of types searches Second, and","Second, we discriminate between two types of searches: global searches and local searches.",168
"to geometry the complex However, folding (i.e. more","However, folding to the more complex geometry (i.e.",169
more For is geometry folding a process. clearly this cooperative,For this geometry folding is clearly a more cooperative process.,170
"our comparing are significant. By mean-field find that stochastic with we fluctuations simulations calculations,","By comparing our simulations with mean-field calculations, we find that stochastic fluctuations are significant.",171
without proteins. cytoplasmic possible localization a motor mechanism for targeted This offers polar,This offers a possible mechanism for targeted polar localization without cytoplasmic motor proteins.,172
central coexistence one of is Species the ecology. modern in themes,Species coexistence is one of the central themes in modern ecology.,173
is diversity. a biological Coexistence prerequisite of,Coexistence is a prerequisite of biological diversity.,174
chemical presented process as equivalent Biological a to evolution reaction. is a,Biological evolution is presented as a process equivalent to a chemical reaction.,175
main interactions occur units. between The is that self-replicating point,The main point is that interactions occur between self-replicating units.,176
from mode the times relaxation were the extracted displacement coordinate. mean square transverse First in,First mode relaxation times were extracted from the mean square displacement in the transverse coordinate.,177
"usage protein-coding kingdoms Across synonmous all codons. of life, exhibit genes unequal biological of","Across all kingdoms of biological life, protein-coding genes exhibit unequal usage of synonmous codons.",178
"to one describes viz. model genetic The study scenario, present population explore specific a minimal","The present study describes a minimal population genetic model to explore one specific scenario, viz.",179
are Possible these for discussed. reasons effects,Possible reasons for these effects are discussed.,180
heterogeneous beat of The intermittent. were dynamics cells single cardiac and,The beat dynamics of single cardiac cells were heterogeneous and intermittent.,181
power-law regime. in intervals The were interbeat distributed long-time a (IBIs),The interbeat intervals (IBIs) were power-law distributed in a long-time regime.,182
nucleotide is repeat by the describe telomerase. model to processivity presented A addition and the,A model is presented to describe the nucleotide and repeat addition processivity by the telomerase.,183
processes in that vertebrates processing to found similar in with Hormonal enzymatic along occur annelids.,Hormonal processes along with enzymatic processing similar to that found in vertebrates occur in annelids.,184
"ones than exert annelids the in function physiological in found vertebrates. Furthermore, these neuropeptides similar","Furthermore, these neuropeptides exert similar physiological function in annelids than the ones found in vertebrates.",185
"or to been in specific neuropeptides also invertebrates some these animals. have annelids Nevertheless,","Nevertheless, some specific neuropeptides to annelids or invertebrates have also been in these animals.",186
on our discussion about based symmetry results A breaking is and brief also learning presented.,A brief discussion about learning and symmetry breaking based on our results is also presented.,187
signal and are pathways transduction in both ubiquitous cells. prokaryotic eukaryotic Push-pull networks in,Push-pull networks are ubiquitous in signal transduction pathways in both prokaryotic and eukaryotic cells.,188
signals allow mechanism zero-order amplify to cells the They of via strongly ultrasensitivity.,They allow cells to strongly amplify signals via the mechanism of zero-order ultrasensitivity.,189
These the cytoplasm. distributed uniformly often are in enzymes,These enzymes are often uniformly distributed in the cytoplasm.,190
how predictions could We discuss experimentally. tested be our,We discuss how our predictions could be tested experimentally.,191
networks Semantic as for systems. promising serve a general-purpose modeling complex substrate,Semantic networks serve as a promising general-purpose modeling substrate for complex systems.,192
"semantic large-scale now and models. to standardized support formats are available practical, tools network Various","Various standardized formats and tools are now available to support practical, large-scale semantic network models.",193
"the Second, triple-stores recent introduction performant of highly (i.e.","Second, the recent introduction of highly performant triple-stores (i.e.",194
"the over ages. Moreover, specific of co-evolutionary and to feature belong modules half evolutionary exhibit","Moreover, over half of the modules exhibit co-evolutionary feature and belong to specific evolutionary ages.",195
modules modules do. faster and Peripheral to cohesively evolve than core tend more,Peripheral modules tend to evolve more cohesively and faster than core modules do.,196
"denote $qe$ We them $qe$-operators, ""qualitative-enriched"" for operators. standing","We denote them $qe$-operators, $qe$ standing for ""qualitative-enriched"" operators.",197
in are this also justified in paper. details $q$-operators,$q$-operators are also justified in details in this paper.,198
correlation times. from often Monte long sampling suffer Carlo methods,Monte Carlo sampling methods often suffer from long correlation times.,199
"steps be must these an independent to for generate run methods Consequently, sample. many","Consequently, these methods must be run for many steps to generate an independent sample.",200
this proposed In to this method paper a overcome is difficulty.,In this paper a method is proposed to overcome this difficulty.,201
between chains. This through exchanges chain auxiliary coarse accomplished full and is the the,This is accomplished through exchanges between the full chain and the auxiliary coarse chains.,202
real bandwidth selection in a of heterogeneous features. multidimensional becomes case challenge Automatic,Automatic bandwidth selection becomes a real challenge in case of multidimensional heterogeneous features.,203
to solution This this problem. a paper presents,This paper presents a solution to this problem.,204
partitioning A and filtering are shift mean pseudo introduced. balloon,A pseudo balloon mean shift filtering and partitioning are introduced.,205
of histories Genetic within homologous phylogenetic heterogeneous genes. can a recombination produce set,Genetic recombination can produce heterogeneous phylogenetic histories within a set of homologous genes.,206
Super-resolution in problem difficult important processing. but is image/video an,Super-resolution is an important but difficult problem in image/video processing.,207
single a on The difficult problem hand. with image only is substantially more low-resolution,The problem is substantially more difficult with only a single low-resolution image on hand.,208
systems Evolution space. described ideas dynamical by mental metric of is in,Evolution of ideas is described by dynamical systems in metric mental space.,209
robots. of such domestic be of Creation may improvement useful psycho-robots,Creation of such psycho-robots may be useful improvement of domestic robots.,210
working are simple domestic merely devices At (e.g. moment the robots,At the moment domestic robots are merely simple working devices (e.g.,211
marker a with gametes They preferrably male zygotes than having fuse different their to own.,They fuse to zygotes preferrably with male gametes having a different marker than their own.,212
are such features two dimensional where arise. high There situations,There are two situations where such high dimensional features arise.,213
features high-order arise interactions. High when we also dimensional consider,High dimensional features also arise when we consider high-order interactions.,214
increase considered. The parameters the exponentially will of with order number,The number of parameters will increase exponentially with the order considered.,215
order. number compressed highest The the have may considering of before parameters converged possible,The number of compressed parameters may have converged before considering the highest possible order.,216
apply method We sequence logistic prediction to models. logistic models compression and this classification,We apply this compression method to logistic sequence prediction models and logistic classification models.,217
use data test simulated our real and data both We both in methods chapters. to,We use both simulated data and real data to test our methods in both chapters.,218
is The analysis. explored proteins of usually via folding these kinetics mutational,The folding kinetics of these proteins is usually explored via mutational analysis.,219
of beta-hairpins native structure a WW two form beta-sheet. comprises three-stranded domains that The these,The native structure of these WW domains comprises two beta-hairpins that form a three-stranded beta-sheet.,220
problems. and the and We open literature answers some relevant partial provide survey many,We survey the relevant literature and provide some partial answers and many open problems.,221
"sampling), significantly the test sampling). better control The performed (space-variant than (uniform group group","The test group (space-variant sampling), performed significantly better than the control group (uniform sampling).",222
"to find the enhanced, was target. as speed The pointing the was accuracy","The pointing accuracy was enhanced, as was the speed to find the target.",223
estimators algorism its Y. were The related by suggested density super-parametric and,The super-parametric density estimators and its related algorism were suggested by Y.,224
"examples, Before not can any we estimators. numerical of the applying to comment give","Before applying to numerical examples, we can not give any comment of the estimators.",225
"are splines studied covering and Bezier windows as well. B-splines,","B-splines, Bezier splines and covering windows are studied as well.",226
results The show algorism that Y. by suggested,The results show that algorism suggested by Y.,227
endemic infection. the spread Susceptible-Infected-Susceptible study an of the of model We,We study the Susceptible-Infected-Susceptible model of the spread of an endemic infection.,228
global a optimization Simulated approaching popular a solution is for method the problem. of annealing,Simulated annealing is a popular method for approaching the solution of a global optimization problem.,229
the probabilistic of recognition the retrieval enables events. similar and It,It enables the recognition and the probabilistic retrieval of similar events.,230
"model. a using inferred examples, a Bayesian probabilities and are network From Dirichlet these","From these examples, probabilities are inferred using a Bayesian network and a Dirichlet model.",231
user model a between enables links graph This interest patterns. similarity to specific to,This enables to links user interest to a specific similarity model between graph patterns.,232
a on experiments Few image proposed multi-spectral sequence recognition illustrate the spatio-temporal performed method. SPOT,Few experiments performed on a multi-spectral SPOT image sequence illustrate the proposed spatio-temporal recognition method.,233
"chemical is mechanisms. physical, by biological It and ruled","It is ruled by physical, chemical and biological mechanisms.",234
this this task. for rules developed the paper We present in,We present in this paper the rules developed for this task.,235
measure has the classical been studied This problem. MAX-SAT on,This measure has been studied on the classical MAX-SAT problem.,236
we perform In study computational Majority cellular this (CAs) the task. automata that paper,In this paper we study cellular automata (CAs) that perform the computational Majority task.,237
searching of the quantify degree We statistically this in various difficulty of ways landscape.,We statistically quantify in various ways the degree of difficulty of searching this landscape.,238
"we landscape So, go from the exploring the top.","So, we go exploring the landscape from the top.",239
subspace. of Then measure several we properties this,Then we measure several properties of this subspace.,240
"a optimal is Otherwise the bound we solution. and gap'', ``duality on a there obtain","Otherwise there is a ``duality gap'', and we obtain a bound on the optimal solution.",241
"programming techniques graphs. Thus, with applicable approach for convex dynamic optimization combines our thin","Thus, our approach combines convex optimization with dynamic programming techniques applicable for thin graphs.",242
"(e.g. induced thin subgraphs We also a small consider of set to relaxations subgraphs,","We also consider relaxations to a set of small induced subgraphs, thin subgraphs (e.g.",243
"cycles. tree by connected and a obtained ``unwinding'' loops),","loops), and a connected tree obtained by ``unwinding'' cycles.",244
"that class of introduce variables. addition, In multiscale relaxations ``summary'' new a propose we","In addition, we propose a new class of multiscale relaxations that introduce ``summary'' variables.",245
concept several the landscape the deficiencies cloud argued of overcomes is It that fitness representation.,It is argued that the fitness cloud concept overcomes several deficiencies of the landscape representation.,246
on of demonstrate scientific several from this collections journals. abstracts algorithm We,We demonstrate this algorithm on collections of scientific abstracts from several journals.,247
sense known is as of This familiarity. recognition,This sense of recognition is known as familiarity.,248
networks undergo and rearrangements during diseases. stress Cellular,Cellular networks undergo rearrangements during stress and diseases.,249
"the central position. proteasome preserve A such as basic functions, their few","A few basic functions, such as the proteasome preserve their central position.",250
"such circuit rate-equations. cannot Therefore, deterministic the as be fully repressilator methods analyzed using","Therefore, the repressilator circuit cannot be fully analyzed using deterministic methods such as rate-equations.",251
of is a complex and elements gene regulatory The characterization very process. experimental resource-intensive,The experimental characterization of gene regulatory elements is a very complex and resource-intensive process.,252
intermediate neutral are trajectories evolutionary to The be steps selectively such in or likely deleterious.,The intermediate steps in such evolutionary trajectories are likely to be selectively neutral or deleterious.,253
"and may be long times traits. such populations to required Therefore, evolve large","Therefore, large populations and long times may be required to evolve such traits.",254
this a We of test hypothesis population based model protein evolution. using,We test this hypothesis using a population based model of protein evolution.,255
a model of grained A structure-based is coarse used. protein,A structure-based coarse grained model of a protein is used.,256
large shear very Full at only occurs rates. unfolding,Full unfolding occurs only at very large shear rates.,257
along the human We behavior large-scale  chromosomes. compositional explore nucleotide of strand asymmetries ,  We explore the large-scale behavior of nucleotide compositional strand asymmetries along human chromosomes.,258
in opens for This work crystallography. exciting work and RNA possibilities further modelling,This work opens exciting possibilities for further work in RNA modelling and crystallography.,259
al. prototypical review et of model we package Niesert the Here,Here we review the prototypical package model of Niesert et al.,260
emergence forces promote nature. modularity the in selective We the that investigate of,We investigate the selective forces that promote the emergence of modularity in nature.,261
question in concerns the arising This following biology. evolutionary article computational,This article concerns the following question arising in computational evolutionary biology.,262
triplet sets. in We this all also to article the apply results note weighted that,We note that all the results in this article also apply to weighted triplet sets.,263
"its understood. are the well to and not However, relation behind complexification mechanisms the evolution","However, the mechanisms behind the complexification and its relation to evolution are not well understood.",264
mean to complexification different things different contexts. used in can Moreover be,Moreover complexification can be used to mean different things in different contexts.,265
longer Why than us? do live women,Why do women live longer than us?,266
so the Y for chromosome Why is men small?,Why is the Y chromosome for men so small?,267
the dangers marital are What of fidelity?,What are the dangers of marital fidelity?,268
of a data parameters generative the sequence. are abrupt in variations Changepoints,Changepoints are abrupt variations in the generative parameters of a data sequence.,269
on demonstrating the different data illustrate by this real-world algorithm sets. We modularity three,We illustrate this modularity by demonstrating the algorithm on three different real-world data sets.,270
"of equations monolithic their sets progress. barrier to as construction a fundamental However, current presents","However, their current construction as monolithic sets of equations presents a fundamental barrier to progress.",271
in datatypes. turn rather These programs than be as require models represented to as,These in turn require models to be represented as programs rather than as datatypes.,272
support have to a developed computational infrastructure this. We,We have developed a computational infrastructure to support this.,273
"this ""infinite""-size finite- that and equally demonstrate on effective framework networks. is Furthermore, we","Furthermore, we demonstrate that this framework is equally effective on finite- and ""infinite""-size networks.",274
modification of of is in regulation role recognized. The post-translational eukaryotic histones gene well,The role of post-translational modification of histones in eukaryotic gene regulation is well recognized.,275
be guiding in discussion in general. of the use could chromatin silencing model The on,The model could be of use in guiding the discussion on chromatin silencing in general.,276
estimate and used of collection to loop structural heterogeneity methods Both regions. ensemble are the,Both ensemble and collection methods are used to estimate the structural heterogeneity of loop regions.,277
speciation invasibility the to applied analysis is Pairwise process. illustrate,Pairwise invasibility analysis is applied to illustrate the speciation process.,278
discuss affects fluctuation diversity. also We how species,We also discuss how fluctuation affects species diversity.,279
Methods a discovery to network are problem a social node solve for presented.,Methods to solve a node discovery problem for a social network are presented.,280
to nodes nodes not which refer are directly. Covert observable the,Covert nodes refer to the nodes which are not observable directly.,281
"illustration, polarity segment is gene an As the network analyzed.","As an illustration, the segment polarity gene network is analyzed.",282
in data Generating many essential is multivariate  applications. Poisson ,  Generating multivariate Poisson data is essential in many applications.,283
computationally and for generating We multivariate appealing efficient propose a Poisson conceptually data. method,We propose a computationally efficient and conceptually appealing method for generating multivariate Poisson data.,284
we end with random asymmetric existing path-integral generalise replicator To this to approaches couplings. systems,To this end we generalise existing path-integral approaches to replicator systems with random asymmetric couplings.,285
can of onset be determined The self-consistently. instability,The onset of instability can be determined self-consistently.,286
is structural proteins local to for resolution in method similarities A presented. search atomic at,A method to search for local structural similarities in proteins at atomic resolution is presented.,287
architecture. on The robust evolve to mutants the depend to network known ability is,The ability to evolve robust mutants is known to depend on the network architecture.,288
the the Does behavior of networks? dynamical selection on global operate,Does selection operate on the global dynamical behavior of the networks?,289
kind by state-space What are structures selection? of favored,What kind of state-space structures are favored by selection?,290
we given paper this In minimum with structures stack-length. $k$-noncrossing enumerate RNA pseudoknot,In this paper we enumerate $k$-noncrossing RNA pseudoknot structures with given minimum stack-length.,291
interaction. in spots social predict relevant This to addresses empty problem a contribution,This contribution addresses a problem to predict relevant empty spots in social interaction.,292
Homogeneous underlying as the social a networks model interaction. studied and inhomogeneous are,Homogeneous and inhomogeneous networks are studied as a model underlying the social interaction.,293
problem. approach new A address to presented method predictor a the function is as heuristic,A heuristic predictor function approach is presented as a new method to address the problem.,294
homogeneous is a demonstrated over Simulation experiment network.,Simulation experiment is demonstrated over a homogeneous network.,295
the form of simulated test the from is communication. in generated baskets A data,A test data in the form of baskets is generated from the simulated communication.,296
membrane Intracellular are tool potential neuronal neurophysiology. in central of a recordings,Intracellular recordings of neuronal membrane potential are a central tool in neurophysiology.,297
are white this noise model using estimated The characteristics current of injection. first,The characteristics of this model are first estimated using white noise current injection.,298
accuracy. with demonstrate other techniques AEC yields it comparing Tests improved to that recordings,Tests comparing AEC to other techniques demonstrate that it yields recordings with improved accuracy.,299
mathematical neuro-stimulation simple models Two derived electric are discussed. of and,Two simple mathematical models of electric neuro-stimulation are derived and discussed.,300
idea for critically the that increase evidence the with examine We time. encephalization quotients,We critically examine the evidence for the idea that encephalization quotients increase with time.,301
is not feature that of evolution. convergent intelligence find human-like We a,We find that human-like intelligence is not a convergent feature of evolution.,302
extraterrestrial are discussed. intelligence search the for Implications for,Implications for the search for extraterrestrial intelligence are discussed.,303
stability on how parameters. aspect A is points of model equilibrium depends specific thereof,A specific aspect thereof is how stability of equilibrium points depends on model parameters.,304
illustrative a is to as method cascade a example. applied of The an model MAPK,The method is applied to a model of a MAPK cascade as an illustrative example.,305
this $k$-noncrossing stack-length. we minimum paper and structures enumerate arc- In pseudoknot given RNA with,In this paper we enumerate $k$-noncrossing RNA pseudoknot structures with given minimum arc- and stack-length.,306
pseudoknot prediction algorithms structures. results for importance for RNA Our of are,Our results are of importance for prediction algorithms for RNA pseudoknot structures.,307
characterize associations binary variable importance and regression pairwise (VIMP) in We study trees. variable and,We characterize and study variable importance (VIMP) and pairwise variable associations in binary regression trees.,308
common for of etc. species/goods/ideas/elements ground observation the that The is new the production,The common ground is the observation that for the production of new species/goods/ideas/elements etc.,309
pre-existence necessary of condition. elements other specific a the is,the pre-existence of specific other elements is a necessary condition.,310
we For find mixed opinions fixed of phases and topology distinct consensus.,For fixed topology we find distinct phases of mixed opinions and consensus.,311
estimate of reasonable the We barrier unfolding have obtained for ubiqutin.,We have obtained the reasonable estimate for unfolding barrier of ubiqutin.,312
"implemented and In for efficiency, called addition, techniques avoiding code reanalysis, extensionalization, we minimization.","In addition, for efficiency, we implemented techniques called extensionalization, avoiding reanalysis, and code minimization.",313
constraints invariants site-pattern vectors Phylogenetic trees. on not for are frequency only the phylogenetic,Phylogenetic invariants are not the only constraints on site-pattern frequency vectors for phylogenetic trees.,314
call these inequalities.'' We constraints additional ``edge-parameter,We call these additional constraints ``edge-parameter inequalities.'',315
self-organized explain the finding critical as phenomenon. We a,We explain the finding as a self-organized critical phenomenon.,316
interval and of patterns the implementation useful Abstractions are the operations in described.,Abstractions and patterns useful in the implementation of the interval operations are described.,317
recognition methods to blurring. object the on approach recent The based non-uniform relates for paper,The paper relates the approach to recent methods for object recognition based on non-uniform blurring.,318
"several multicellular cell states In organisms, coexist.","In multicellular organisms, several cell states coexist.",319
in systems is several of relevance mechanism The multicellular to this and differentiation development discussed.,The relevance of this mechanism to development and differentiation in several multicellular systems is discussed.,320
hexagonal three immediate with cell cellular and We neighbourhood automata consider cell-states.,We consider hexagonal cellular automata with immediate cell neighbourhood and three cell-states.,321
neighbours state. many how in each one are,how many neighbours are in each one state.,322
methods. to I motivated an Bayesian describe by approach similarity,I describe an approach to similarity motivated by Bayesian methods.,323
methods. Bayesian is using standard This a yields that similarity a learnable function,This yields a similarity function that is learnable using a standard Bayesian methods.,324
of and variable is variable approach kernel The the to metric methods relationship discussed.,The relationship of the approach to variable kernel and variable metric methods is discussed.,325
"and been physics. Many this to in chemistry biology, attempts made behaviour have model","Many attempts have been made to model this behaviour in biology, chemistry and physics.",326
"an studied. illustration, circumferential hyperelastic of a disk is growth As the free","As an illustration, the circumferential growth of a free hyperelastic disk is studied.",327
of Irregular a graphs reduced. defined successively stack pyramids are as,Irregular pyramids are defined as a stack of graphs successively reduced.,328
"is in this pyramid. the implicitly framework, often scale defined the Within as the height","Within this framework, the scale is often defined implicitly as the height in the pyramid.",329
is by by scale This Guigues. the last framework drawback addressed set designed,This last drawback is addressed by the scale set framework designed by Guigues.,330
with method and one within We two parallel propose variations sequential both. one,We propose one sequential and one parallel method with two variations within both.,331
these in energy signals. phase pattern was A of roughly detected free constant sinusoidal,A sinusoidal pattern of roughly constant phase was detected in these free energy signals.,332
are frameshift genes experimentally verified presented across eubacteria. Results using,Results are presented using experimentally verified frameshift genes across eubacteria.,333
and  perception recognition is presented. of A  theory visual neural network,  A neural network theory of visual perception and recognition is presented.,334
ordered four they a are form separators fully and semi-group. element,they are separators and form a four element fully ordered semi-group.,335
unified the for framework of Our a provides building biomolecules. work understanding blocks,Our work provides a unified framework for understanding the building blocks of biomolecules.,336
constraint implemented library. have for DDS the Gecode We programming,We have implemented DDS for the Gecode constraint programming library.,337
"problem networks study i.e. constructing the phylogenetic from We triplets, of level-k","We study the problem of constructing level-k phylogenetic networks from triplets, i.e.",338
"level-k for a defined network by We is that uniquely each its k, triplets. give,","We give, for each k, a level-k network that is uniquely defined by its triplets.",339
modeling compartmental the used can for be cardiovascular a of The system. system,The system can be used for a compartmental modeling of the cardiovascular system.,340
artificial nobody that is. really problem knows A intelligence what in fundamental intelligence is,A fundamental problem in artificial intelligence is that nobody really knows what intelligence is.,341
arbitrary These for then intelligence general mathematically a formalised of produce machines. measure to are,These are then mathematically formalised to produce a general measure of intelligence for arbitrary machines.,342
"both poorly adaptive and remains understood, distribution effects The theoretically. fitness of mutations of empirically","The distribution of fitness effects of adaptive mutations remains poorly understood, both empirically and theoretically.",343
images of This novel for paper multilevel a proposes hierarchical segmentation by method thresholding.,This paper proposes a novel method for segmentation of images by hierarchical multilevel thresholding.,344
"locations. method The global, pixel is and disregards nature in agglomerative","The method is global, agglomerative in nature and disregards pixel locations.",345
relation We process. the which variances recursive for a obtain expedites at each step the,We obtain a recursive relation at each step for the variances which expedites the process.,346
the of comparison in some shown efficacy with method methods. is well-known a The,The efficacy of the method is shown in a comparison with some well-known methods.,347
conclude also DNA search efficient to higher more a process. We density a leads that,We also conclude that a higher DNA density leads to a more efficient search process.,348
unlike empty pathways formation. These assembly capsid identified for are any,These assembly pathways are unlike any identified for empty capsid formation.,349
markup code is genetic the The assignment stop codons. of,The genetic code markup is the assignment of stop codons.,350
of this Inhibition type factor controlling as binding. of serves,Inhibition serves as controlling factor of this type of binding.,351
on depends choice The the frequency. crossover,The choice depends on the crossover frequency.,352
Close observed. is this sympatric to value speciation,Close to this value sympatric speciation is observed.,353
attracted of also debate. interest The synchronization has and role much functional,The functional role of synchronization has also attracted much interest and debate.,354
"this precision."" In enhancement synchronization: role we for study another ""collective so-called the of article,","In this article, we study another role for synchronization: the so-called ""collective enhancement of precision.""",355
biology. Similar questions may to in be concepts applicable systems,Similar concepts may be applicable to questions in systems biology.,356
consider We agent interacting environment. an an with unknown,We consider an agent interacting with an unknown environment.,357
on a has utility from environment. outputs the agent The function,The agent has a utility function on outputs from the environment.,358
world? Does construct an the efficient brain sensory representation the of,Does the brain construct an efficient representation of the sensory world?,359
decaying algebraically The density developing distributions. vasculature complex still non-hierarchical tumor is but displaying,The developing tumor vasculature is non-hierarchical but still complex displaying algebraically decaying density distributions.,360
the Hopf Turing domain. derive bifurcation on conditions and the for We spatial,We derive the conditions for Hopf and Turing bifurcation on the spatial domain.,361
for studying complexity will ecosystems. of dynamic useful It the be,It will be useful for studying the dynamic complexity of ecosystems.,362
of a be investigate can raytracing We on class that Blue performance Gene achieved supercomputers.,We investigate raytracing performance that can be achieved on a class of Blue Gene supercomputers.,363
of scaling nontrivial find at large number behavior processors. We,We find nontrivial scaling behavior at large number of processors.,364
resolution. of applications lighting to technology with this We and high advanced discuss visualization scientific,We discuss applications of this technology to scientific visualization with advanced lighting and high resolution.,365
physiological These ideal myosin's respective mechanics different functions. are for the,These different mechanics are ideal for the respective myosin's physiological functions.,366
"fitness"". following with those is statement ""Survival for non-dominated mostly The is proposed","The following statement is proposed ""Survival is mostly for those with non-dominated fitness"".",367
that immune system an We complex system. is adaptive argue,We argue that immune system is an adaptive complex system.,368
shown is that has emergent properties. It it,It is shown that it has emergent properties.,369
structure the of world Its network small is type. network,Its network structure is of the small world network type.,370
"of type, is in threshold network the autoimmunity. The avoiding which helps","The network is of the threshold type, which helps in avoiding autoimmunity.",371
tumors. a radiotherapy chemotherapy of sufficient to eradication get or/and,chemotherapy or/and radiotherapy to get a sufficient eradication of tumors.,372
example is an A attacking effectors of fractional two immune given. antigen order,A fractional order example of two immune effectors attacking an antigen is given.,373
But simple is as this? as it,But is it as simple as this?,374
"the structure, that shall In this i.e. we see lexical article","In this article we shall see that the lexical structure, i.e.",375
"linguistic by of relationships, the network match words necessarily the does not linked conceptualization. domain","the network of words linked by linguistic relationships, does not necessarily match the domain conceptualization.",376
are for non-cooperative game. Hence in equations they replicator applied,Hence they are applied in replicator equations for non-cooperative game.,377
have model It a known stable order integer is does equilibrium. that its not,It is known that its integer order model does not have a stable equilibrium.,378
to Its solution. order a have model asymptotically stable locally fractional shown is internal,Its fractional order model is shown to have a locally asymptotically stable internal solution.,379
solution. stable to game locally A a asymmetric shown asymptotically is have internal FO,A FO asymmetric game is shown to have a locally asymptotically stable internal solution.,380
the is not for its counterpart. order integer case This,This is not the case for its integer order counterpart.,381
this is matter is biological learning and examined. The the sketched basis of computing of,The biological basis of this computing is sketched and the matter of learning is examined.,382
The of of suggests advantageous. labor division is that ubiquity highly it this,The ubiquity of this division of labor suggests that it is highly advantageous.,383
"I the this evolution to specialized a present In cells. paper, study model of","In this paper, I present a model to study the evolution of specialized cells.",384
and multicellular (terminally that may contain organisms for The allows cells. somatic model unicellular differentiated),The model allows for unicellular and multicellular organisms that may contain somatic (terminally differentiated) cells.,385
Cells to quantitative a contribute additively trait.,Cells contribute additively to a quantitative trait.,386
surprisingly I cells show is the reproductive that high. always of fraction,I show that the fraction of reproductive cells is always surprisingly high.,387
model primitive of biology predictions. the the I organisms discuss with multicellular respect to,I discuss the biology of primitive multicellular organisms with respect to the model predictions.,388
the next of introduces MathPSfrag. version This article,This article introduces the next version of MathPSfrag.,389
preview Mathematica. files Moreover generated is and within shown from a these,Moreover from these files a preview is generated and shown within Mathematica.,390
for regime. plectonemic model consider rod an  twisted the We elastic DNA in ,  We consider an elastic rod model for twisted DNA in the plectonemic regime.,391
"an treated effective, tube molecule adjustable is radius. The as impenetrable with an","The molecule is treated as an impenetrable tube with an effective, adjustable radius.",392
cardinality stream in the paper of considers problem data This estimation applications.,This paper considers the problem of cardinality estimation in data stream applications.,393
in dimension analysis. for reduction primary concern are Techniques statistical of,Techniques for dimension reduction are of primary concern in statistical analysis.,394
projection algorithms that properties investigate of the use distributions. We alpha-stable exploit of the,We investigate the use of projection algorithms that exploit properties of the alpha-stable distributions.,395
Here version mean-field of a model. their we provide,Here we provide a mean-field version of their model.,396
a supervised classification method arrayCGH new We for data. propose Results: of,Results: We propose a new method for supervised classification of arrayCGH data.,397
Availability: available. publicly data algorithms All are and,Availability: All data and algorithms are publicly available.,398
function vitro. the way be this could determine in that I hepatocyte re-created microenvironments their,I this way the hepatocyte microenvironments that determine their function could be re-created in vitro.,399
load as entities the approach of hepatocyte use systems. to bioreactor The aggregates needs the,The approach needs the use of hepatocyte aggregates as entities to load the bioreactor systems.,400
produce surfaces. has patterned been printing a used Microcontact to,Microcontact printing has been used to produce a patterned surfaces.,401
processes of evolution substitution in Covarion inhomogeneities through describe character models time.,Covarion models of character evolution describe inhomogeneities in substitution processes through time.,402
"Based obtain and the formation we bifurcation numerical stability on simulation. analysis, pattern spiral via","Based on the stability and bifurcation analysis, we obtain spiral pattern formation via numerical simulation.",403
"with external we colored Additionally, the model and study periodic noise forcing.","Additionally, we study the model with colored noise and external periodic forcing.",404
Neurons nervous and variety exhibit system properties. an in physiological morphological of the outstanding,Neurons in the nervous system exhibit an outstanding variety of morphological and physiological properties.,405
traits neuron characterized. The of types been two extensively these dynamical have,The dynamical traits of these two neuron types have been extensively characterized.,406
"the interesting, of consequences It would dynamical be their understand however, to properties. information-processing","It would be interesting, however, to understand the information-processing consequences of their dynamical properties.",407
We models forms. conductance-based minimal with normal work realistic and both,We work both with realistic conductance-based models and minimal normal forms.,408
depolarizing type response stimuli. We in neurons that fire scale-free I conclude to,We conclude that type I neurons fire in response to scale-free depolarizing stimuli.,409
conditions enables under evolutionarily the be favored. will This to predict mutators us which,This enables us to predict the conditions under which mutators will be evolutionarily favored.,410
"Surprisingly, for our show effect small mutators. (i) simulations is typically strong-effect that","Surprisingly, our simulations show that effect (i) is typically small for strong-effect mutators.",411
Our and semi-quantitatively experiments directions. with evolution suggest existing agree future experimental results laboratory,Our results agree semi-quantitatively with existing laboratory evolution experiments and suggest future experimental directions.,412
areas method natural hierarchical in segmenting for images. text a present We,We present a hierarchical method for segmenting text areas in natural images.,413
"In contain simple text symbols. particular, can or the graphics","In particular, the text can contain simple graphics or symbols.",414
is to study developed theoretical dynamics framework the of protein folding. A,A theoretical framework is developed to study the dynamics of protein folding.,415
a with scale. proteins single energy model against numerical simulations We predictions test these of,We test these predictions against numerical simulations of model proteins with a single energy scale.,416
flow information of the carrier is entropy The flow.,The entropy flow is the carrier of information flow.,417
"time may scales, beget Also, ``seizures short over seizures.''","Also, over short time scales, ``seizures may beget seizures.''",418
to epilepticus. status This cluster evolve of explain may tendency the seizures into and,This may explain the tendency of seizures to cluster and evolve into status epilepticus.,419
and their thus macromolecules importance. of are is essential function of great life understanding Proteins,Proteins are essential macromolecules of life and thus understanding their function is of great importance.,420
"analyzed case in an example, sub-Poisson detail. As is the firing binomial of statistics","As an example, the case of sub-Poisson binomial firing statistics is analyzed in detail.",421
of foot were displacements a using recorded platform. force Centre pressure (CoP),Centre of foot pressure (CoP) displacements were recorded using a force platform.,422
displacements force pressure platform. foot recorded were a Centre of using,Centre of foot pressure displacements were recorded using a force platform.,423
(CoP) force foot platform. displacements were The pressure centre using of a recorded,The centre of foot pressure displacements (CoP) were recorded using a force platform.,424
have and clinical in could rehabilitative areas. implications This,This could have implications in clinical and rehabilitative areas.,425
control which a condition also in required. There no was task was concurrent,There was also a control condition in which no concurrent task was required.,426
(cGAs). cellular In genetic selection new a algorithms we in this paper introduce scheme,In this paper we introduce a new selection scheme in cellular genetic algorithms (cGAs).,427
(AS) of allows diversity selective control the promotes accurate and Selection Anisotropic pressure.,Anisotropic Selection (AS) promotes diversity and allows accurate control of the selective pressure.,428
provide a from collection trees. strategy a useful combining Consensus of for gene information methods,Consensus methods provide a useful strategy for combining information from a collection of gene trees.,429
Heterogeneity population experiencing disease. an of property a is any important,Heterogeneity is an important property of any population experiencing a disease.,430
"analyzed particular, and SIR an (susceptible-infective-removed) heterogeneity. is for sources model of different In formulated","In particular, an SIR (susceptible-infective-removed) model is formulated and analyzed for different sources of heterogeneity.",431
"very model, mechanistic which a phenomenological mimics Therefore, provided. is derivation of well, the reality","Therefore, a mechanistic derivation of the phenomenological model, which mimics reality very well, is provided.",432
may such rough theoretical graph serve The tool as a cases. models in powerful but,The graph theoretical models may serve as a rough but powerful tool in such cases.,433
"expressions are confirmation. theoretical for and computational Moreover, discussed the studies robustness are done","Moreover, the theoretical robustness expressions are discussed and computational studies are done for confirmation.",434
the properties setup. of enhanced FFL many this are find We that within,We find that many of the FFL properties are enhanced within this setup.,435
FFLs all to D. Melanogaster network. the We such identify regulatory in then proceed,We then proceed to identify all such FFLs in the D. Melanogaster regulatory network.,436
elucidate This of the may the TF type confers. help regulation,This may help elucidate the type of regulation the TF confers.,437
structure study coupled primary the dynamics and We of secondary (i.e. formation,We study the coupled dynamics of primary and secondary structure formation (i.e.,438
"simulations. numerical predicted support in physical The phenomenology finds and interpretations, satisfactory allows for natural","The predicted phenomenology allows for natural physical interpretations, and finds satisfactory support in numerical simulations.",439
a Dimensionality recent of is interest. topic reduction,Dimensionality reduction is a topic of recent interest.,440
algorithm as multiple classes The the well can semi-supervised as setting. for account,The algorithm can account for multiple classes as well as the semi-supervised setting.,441
out-of-sample an for data. We and present unlabeled labeled expressions both,We present an out-of-sample expressions for both labeled and unlabeled data.,442
using data. performance on investigate CCDR satellite hyper-spectral imagery the algorithm classification We,We investigate classification performance using the CCDR algorithm on hyper-spectral satellite imagery data.,443
"body the uniquely. and determined the composition first are In class, mass","In the first class, the body composition and mass are determined uniquely.",444
"We theory derive i.e. to the effective, apply the asymptotic","We apply the asymptotic theory to derive the effective, i.e.",445
dynamics biochemical reaction of the system. macroscopic,macroscopic dynamics of the biochemical reaction system.,446
macro-molecular of typical finally We complex more but machinery. give examples,We finally give examples of more complex but typical macro-molecular machinery.,447
argue that scientific entire We a activity. future universe simulation of from result an will,We argue that a simulation of an entire universe will result from future scientific activity.,448
real-world between distinguish also modelling. and artificial-world We,We also distinguish between real-world and artificial-world modelling.,449
population The susceptibility. in heterogeneous have individuals infectiousness the and/or,The individuals in the population have heterogeneous infectiousness and/or susceptibility.,450
bounds on probability. size upper global also conditions and find giving We,We also find conditions giving global upper bounds on size and probability.,451
network-independent. are distributions these leading bounds The to,The distributions leading to these bounds are network-independent.,452
"co-chaperones, proteins produced proteins with and associated activities, ubiquitin. Differentially included oxido-reductase","Differentially produced proteins included co-chaperones, proteins associated with oxido-reductase activities, and ubiquitin.",453
"study we replicator quenched random interaction multi-species matrices. system a Specifically, with","Specifically, we study a multi-species replicator system with quenched random interaction matrices.",454
well consider asymmetric here anti-symmetric cases. interactions and as We symmetric as,We here consider symmetric interactions as well as asymmetric and anti-symmetric cases.,455
[Phys. an is extension Tokita Our of work,Our work is an extension of Tokita [Phys.,456
"one-dimensional genes null we statistical Using constructed between and gases, chromosomes analogies a model. on","Using analogies between genes on chromosomes and one-dimensional gases, we constructed a statistical null model.",457
modular networks The and hierarchical have observed structure.,The observed networks have hierarchical and modular structure.,458
abstract [The has of manuscript information.] more the,[The abstract of the manuscript has more information.],459
in types We protein in mass-action consider two of fluctuations equilibrium networks. binding the,We consider two types of fluctuations in the mass-action equilibrium in protein binding networks.,460
used amenable As to mechanics study. statistical our of they in equilibrium methods such are,As such they are amenable to methods of equilibrium statistical mechanics used in our study.,461
Background coefficients by cameras. provided based is on modeling DCT,Background modeling is based on DCT coefficients provided by cameras.,462
the approximated for method uses model Scene updating median improved performance.,Scene model updating uses the approximated median method for improved performance.,463
"nonparametric are admit and They interpretable, priors. fast computationally simple, and scalable,","They are simple, computationally fast and scalable, interpretable, and admit nonparametric priors.",464
have evidence only weight fixed is that synapses biological discrete number of There a states.,There is evidence that biological synapses have only a fixed number of discrete weight states.,465
"storage discrete, calculate terms Shannon information. We bounded synapses of the capacity in of","We calculate the storage capacity of discrete, bounded synapses in terms of Shannon information.",466
models are macroevolution Two studied. of mathematical,Two mathematical models of macroevolution are studied.,467
"the In have models spectral different limit, noiseless power densities. the two","In the noiseless limit, the two models have different power spectral densities.",468
These fluctuations long-time indicate dynamics. population stochastic may that affect significantly results evolutionary,These results indicate that stochastic population fluctuations may significantly affect long-time evolutionary dynamics.,469
"extraction, However, or replace by (VT). training applications EM Viterbi","However, applications replace EM by Viterbi extraction, or training (VT).",470
the The then applied to One-Max optimization problem. is approach Interactive,The approach is then applied to the Interactive One-Max optimization problem.,471
the analysis provided. images of Quantitative of reconstructed accuracy is the,Quantitative analysis of the accuracy of the reconstructed images is provided.,472
motors; words: effects non-linear partial waves. molecular of collective differential shock equations; Key cargo transport;,Key words: collective effects of molecular motors; cargo transport; non-linear partial differential equations; shock waves.,473
The in first the system this presented architecture of part. global is,The global architecture of this system is presented in the first part.,474
factual is composed a is made agents. part of that multiagent system of This,This part is composed of a multiagent system that is made of factual agents.,475
model a of minimal networks. regulatory introduce the transcriptional description dynamics for We,We introduce a minimal model description for the dynamics of transcriptional regulatory networks.,476
"combines verbalization. reflection, and method The visualization, the of concepts","The method combines the concepts of reflection, visualization, and verbalization.",477
"two were experiment, learned. lessons In the","In the experiment, two lessons were learned.",478
We coding. of the neural implications the of results discuss these context in,We discuss the implications of these results in the context of neural coding.,479
the nodes Covert are which to not directly. nodes refer observable,Covert nodes refer to the nodes which are not observable directly.,480
"simulation. with Two studied, been verified Carlo Monte have coverage via examples","Two examples have been studied, with coverage verified via Monte Carlo simulation.",481
"Particularly, small are in settings. we interested sample","Particularly, we are interested in small sample settings.",482
multiagent the stages system. three inside passes situation current of by the evaluation The,The evaluation of the current situation passes by three stages inside the multiagent system.,483
"time, situation the In dynamic a represented is first a in way.","In a first time, the situation is represented in a dynamic way.",484
Information are features. the coming formatted according and environment from a taxonomy semantic to using,Information coming from the environment are formatted according to a taxonomy and using semantic features.,485
"analysis theory, set rough operation. application the hydrocyclone of paper on of describes This","This paper describes application of rough set theory, on the analysis of hydrocyclone operation.",486
"fulfilled. rules, an on these is approximate estimation decision on attribute Based","Based on these rules, an approximate estimation on decision attribute is fulfilled.",487
"with brief comparison method system a highlighted. this SOM-NFIS (briefly SONFIS) the is of Finally,","Finally, a brief comparison of this method with the SOM-NFIS system (briefly SONFIS) is highlighted.",488
"feedback simulations, haptic from comes contact forces. computation often multimodal interactive In","In multimodal interactive simulations, haptic feedback computation often comes from contact forces.",489
"on rendering the contact haptic depends of space significantly modeling. fidelity Subsequently,","Subsequently, the fidelity of haptic rendering depends significantly on contact space modeling.",490
and Contact models in often to are date methods. deformable between up simplified friction laws,Contact and friction laws between deformable models are often simplified in up to date methods.,491
keep crucial stable point to This feedback. haptic is last,This last point is crucial to keep stable haptic feedback.,492
"packaged, approach This has been and global implemented, tested.","This global approach has been packaged, implemented, and tested.",493
"such preferred. conclusive however, that frequencies is evidence of it no There why are exists,","There exists, however, no conclusive evidence of why it is that such frequencies are preferred.",494
branches functional overproduction clear. of is The significance not the of,The functional significance of the overproduction of branches is not clear.,495
efficient We and is most also branching for different dendrites. suggest that the rule axons,We also suggest that the most efficient branching rule is different for axons and dendrites.,496
test investigate statistics for We homogeneity propose to kernel testing reproducing Hilbert in spaces.,We propose to investigate test statistics for testing homogeneity in reproducing kernel Hilbert spaces.,497
uncertainty. evidence present to significantly show that underestimates some still We model BKNN,We present some evidence to show that BKNN still significantly underestimates model uncertainty.,498
is algorithms embedded on grid. are genetic the Cellular where population genetic toroidal algorithms a,Cellular genetic algorithms are genetic algorithms where the population is embedded on a toroidal grid.,499
"a switch. curve biochemical constituting signal-response appears sigmoidal, Its","Its signal-response curve appears sigmoidal, constituting a biochemical switch.",500
"in only switch i.e. However, emerges behavior the region"", the ""zero-order","However, the switch behavior only emerges in the ""zero-order region"", i.e.",501
modifies. lower the that the much molecule signal when of is it than substrate concentration,when the signal molecule concentration is much lower than that of the substrate it modifies.,502
the dynamic that at We to concentration. disorder system bulk showed is robust,We showed that the system is robust to dynamic disorder at bulk concentration.,503
biological to is relevant Such many functions. fluctuation,Such fluctuation is relevant to many biological functions.,504
with are Blueprints icosahedral provided. DNA of made cages symmetry circular of polyhedral molecules,Blueprints of polyhedral cages with icosahedral symmetry made of circular DNA molecules are provided.,505
Information (IPCA). our method refer as Preserving Component We to Analysis,We refer to our method as Information Preserving Component Analysis (IPCA).,506
of immune persistent viruses. to dynamics Population responses,Population dynamics of immune responses to persistent viruses.,507
of consider for phase. intracellular the We and times death (eclipse) age-distributed role delays the,We consider the role of age-distributed delays for death times and the intracellular (eclipse) phase.,508
of functions. These are means probability distribution processes by implemented,These processes are implemented by means of probability distribution functions.,509
convey data on their obtained about population samples information evolutionary history. Genetic,Genetic data obtained on population samples convey information about their evolutionary history.,510
"of divergences, stepwise involve size and any combination Such population scenarios changes. population admixtures","Such scenarios involve any combination of population divergences, admixtures and stepwise population size changes.",511
This program article features. in the methods key main used provides describes its and,This article describes key methods used in the program and provides its main features.,512
"Connected MAny of study, Particles we Intelligent this Systems introduce frame general (MACIPS). In","In this study, we introduce general frame of MAny Connected Intelligent Particles Systems (MACIPS).",513
"SONFIS, phase . society-government interaction particles system, intelligent SORST, Keywords: transition, connected many","Keywords: phase transition, SONFIS, SORST, many connected intelligent particles system, society-government interaction .",514
how interact. is immediately these two clear perspectives It not,It is not immediately clear how these two perspectives interact.,515
about band in Cs-DNA In spectra. this Na-DNA higher intensity than has spectra twice,In Cs-DNA spectra this band has about twice higher intensity than in Na-DNA spectra.,516
by and redundant This canceling lateral interactions. propagating information is implemented using,This is implemented by propagating and canceling redundant information using lateral interactions.,517
which a on details have numerical Success simulations these depends of number uncovered.,Success depends on a number of details which these numerical simulations have uncovered.,518
Feedback. Antitachycardia Bidomain Defibrillation; model; \emph{Keywords} drift; Resonant Re-entry; pacing; ICD;,\emph{Keywords} Re-entry; Bidomain model; Resonant drift; ICD; Defibrillation; Antitachycardia pacing; Feedback.,519
this voluntarily participated Twelve healthy young adults in experiment.,Twelve young healthy adults voluntarily participated in this experiment.,520
of foot were pressure force displacements using Centre platform. a recorded (CoP),Centre of foot pressure (CoP) displacements were recorded using a force platform.,521
showed to displacements CoP biofeedback in no-biofeedback relative Results reduced the the condition.,Results showed reduced CoP displacements in the biofeedback relative to the no-biofeedback condition.,522
are the needed to clinical strengthen potential value investigations of this device. Further,Further investigations are needed to strengthen the potential clinical value of this device.,523
The is applied invertebrate field. a rice a to families (taxa) in method network of,The method is applied to a network of invertebrate families (taxa) in a rice field.,524
This underscores versatility of the networks. the functional,This underscores the functional versatility of the networks.,525
paper problem this ways. addresses This two in,This paper addresses this problem in two ways.,526
"proof showing a formal abolished. give can be which we Second, constraints","Second, we give a formal proof showing which constraints can be abolished.",527
angular found area. that an described It was plotted by can the be data,It was found that the plotted data can be described by an angular area.,528
vector computationally tree-structured A for efficient scheme presented. is simple quantization and,A simple and computationally efficient scheme for tree-structured vector quantization is presented.,529
mines. one Out of essential seam problems of is dilution these the in,Out of seam dilution is one of the essential problems in these mines.,530
impose of cost mining dilution Therefore and the the can additional milling.,Therefore the dilution can impose the additional cost of mining and milling.,531
later methods to other unknown new cases. benefit is predict The of,The other benefit of later methods is to predict new unknown cases.,532
"sets (reducts) been by RST have obtained. So, reduced the","So, the reduced sets (reducts) by RST have been obtained.",533
"data set hydrocyclone is on of method, the the rendered. proposed Validation of the","Validation of the proposed method, on the data set of the hydrocyclone is rendered.",534
use optical of allows photocameras methods such Use commercial to measurements. for,Use of such methods allows to use commercial photocameras for optical measurements.,535
al. Moral theoretical While is et upon of works Del based the method this,While this method is based upon the theoretical works of Del Moral et al.,536
concrete This long a has history. challenging problem,This concrete challenging problem has a long history.,537
"competitive quality. segmentation output shown other faster approach than to be This has with methods,","This approach has shown to be faster than other segmentation methods, with competitive output quality.",538
"way. method the Therefore, efficient solves label problem an in the multiple of segmentation","Therefore, the method solves the problem of multiple label segmentation in an efficient way.",539
feedback regulatory The and aspects on transcriptional in global hierarchy pathways. study the of focuses,The study focuses on the global aspects of feedback and hierarchy in transcriptional regulatory pathways.,540
of film We scripts. story of style narrative using structure analyze and the the case,We analyze the style and structure of story narrative using the case of film scripts.,541
allow descriptive metadata. with users Web publish Many annotate social to content sites and,Many social Web sites allow users to publish content and annotate with descriptive metadata.,542
tags users lets Del.icio.us site related social group bookmarking bundles. The into similarly,The social bookmarking site Del.icio.us similarly lets users group related tags into bundles.,543
retained to if be relation uses determine particular a discarded. Our approach should or statistics,Our approach uses statistics to determine if a particular relation should be retained or discarded.,544
are together The hierarchies. into relations then woven larger,The relations are then woven together into larger hierarchies.,545
of marginal integrals. the in evaluation likelihood Bayesian statistics Inference involves,Inference in Bayesian statistics involves the evaluation of marginal likelihood integrals.,546
Our and priors. to apply priors Dirichlet methods uniform both,Our methods apply to both uniform priors and Dirichlet priors.,547
experiments complete Temperature-dependent of stopped-flow parametrization provide FRET thermo-kinetic near a opening.,Temperature-dependent stopped-flow FRET experiments provide a near complete thermo-kinetic parametrization of opening.,548
we decisions? unreliable reduce How the caused of can death the risk by,How can we reduce the risk of the death caused by unreliable decisions?,549
also the uncertainty in in We decisions entropy. compare of terms,We also compare the uncertainty in decisions in terms of entropy.,550
"this paper, on models contour (e.g. features statistical region-based In where focus image active we","In this paper, we focus on statistical region-based active contour models where image features (e.g.",551
intensity) whose parametric are variables belongs random family some (e.g. to distribution,intensity) are random variables whose distribution belongs to some parametric family (e.g.,552
confining to the special than rather Gaussian case. exponential) ourselves,exponential) rather than confining ourselves to the special Gaussian case.,553
of result framework multi-parameter within stated is the family. general exponential A,A general result is stated within the framework of multi-parameter exponential family.,554
on Experimental of images results applicability demonstrate real and our approach. the both synthesized,Experimental results on both synthesized and real images demonstrate the applicability of our approach.,555
"This on theory, the of application flowcharts. design granulation rock information paper of engineering describes","This paper describes application of information granulation theory, on the design of rock engineering flowcharts.",556
"flowchart, highlighted. Firstly, has been on granulation based an overall theory information","Firstly, an overall flowchart, based on information granulation theory has been highlighted.",557
"best balancing criteria on granules Using are obtained. (information different pockets),","Using different criteria on balancing best granules (information pockets), are obtained.",558
is task. time-critical of the a attack investigation terrorist The,The investigation of the terrorist attack is a time-critical task.,559
law. distribution rank-size investigated by was The analysis (Zipf's,The distribution was investigated by rank-size analysis (Zipf's law.,560
closely be distribution. an exponential to to distribution found The was more,The distribution was found to be more closely to an exponential distribution.,561
appeared languages. to to be correlation natural The similar,The correlation appeared to be similar to natural languages.,562
of memory to systems. the Segmentation be series showed analysis range short,Segmentation analysis of the series showed to be short range memory systems.,563
"deforestation. the accelerate extent increases, temperature To populations insect mean","To the extent mean temperature increases, insect populations accelerate deforestation.",564
can drivers. requiring emerge independent loop is longer self-sustaining--no feedback that positive A climate-change,A positive feedback loop can emerge that is self-sustaining--no longer requiring independent climate-change drivers.,565
novel to interventions. These open and way and detection control results nontoxic the monitoring strategies,These results open the way to novel detection and monitoring strategies and nontoxic control interventions.,566
"for the are of reliable dynamics. networks course evolution, selected the In","In the course of the evolution, networks are selected for reliable dynamics.",567
model both species-species ancestral incorporates The interactions and relationships.,The model incorporates both species-species interactions and ancestral relationships.,568
"distributions independent predicted rate of The of the speciation, i.e. are","The predicted distributions are independent of the rate of speciation, i.e.",569
gradualist of view an one or whether evolution. adopts punctuated,whether one adopts an gradualist or punctuated view of evolution.,570
was extensions maximal performing Muscular achieved trunk repetitive fatigue until by exhaustion.,Muscular fatigue was achieved by performing trunk repetitive extensions until maximal exhaustion.,571
a platform. recorded displacements pressure (CoP) foot force were using of Centre,Centre of foot pressure (CoP) displacements were recorded using a force platform.,572
findings that fatigue evidence degrades muscle ankle joint. the The sense present at force,The present findings evidence that muscle fatigue degrades force sense at the ankle joint.,573
Isochores segments  genome G+C. are long in relatively  homogeneous,  Isochores are long genome segments relatively homogeneous in G+C.,574
"end, To versions this required. are standalone","To this end, standalone versions are required.",575
"discuss present this generalized and a conflict chapter, rule. In proportional we redistribution new","In this chapter, we present and discuss a new generalized proportional conflict redistribution rule.",576
Many years. last the rules few proposed in been combination have,Many combination rules have been proposed in the last few years.,577
recognition. feature to applied during procedure learning A selection is accelerate,A feature selection procedure is applied during learning to accelerate recognition.,578
"robustness We mechanism, a improving feedback recognition simple introduce significantly attention-like in and scenes. multiple-object","We introduce a simple attention-like feedback mechanism, significantly improving recognition and robustness in multiple-object scenes.",579
plane or spatial curve. assign we each a plane To signature curve,To each plane or spatial curve we assign a plane signature curve.,580
"Curves, under equivalent signature. an same transformation, the have affine","Curves, equivalent under an affine transformation, have the same signature.",581
introduction. of invariants with in given the is The other types comparison,The comparison with other types of invariants is given in the introduction.,582
"global and signatures, signature signature. We local of types the present two the","We present two types of signatures, the global signature and the local signature.",583
of Both signatures independent sampling). (curve are parameterization,Both signatures are independent of parameterization (curve sampling).,584
establish local helps curves. It equivalence of,It helps establish local equivalence of curves.,585
"be the to the of it the experts. specificity consider can important of Moreover, responses","Moreover, it can be important to consider the specificity of the responses of the experts.",586
rules year proposed. some few unification Since are,Since few year some unification rules are proposed.,587
our of the the redistribution We previous proportional rule. have interest shown works in conflict,We have shown in our previous works the interest of the proportional conflict redistribution rule.,588
derive the transition probabilities process. We for,We derive transition probabilities for the process.,589
impressive of differences domain most the affect ligand-binding USP-RXR. The,The most impressive differences affect the ligand-binding domain of USP-RXR.,590
plantar foot system. displacements of using pressure were data pressure Centre (CoP) acquisition a recorded,Centre of foot pressure (CoP) displacements were recorded using a plantar pressure data acquisition system.,591
"first but better, approach always certainly is The not the feasible.","The first approach is certainly the better, but not always feasible.",592
"important results applications. most the produced the is in However,","However, the most important is the produced results in applications.",593
contacts as outbreaks. study We strategy to social a of effects switching the epidemic control,We study the effects of switching social contacts as a strategy to control epidemic outbreaks.,594
situations. efficient may suggests also it that an This in more be method complex,This suggests that it may also be an efficient method in more complex situations.,595
response information the in technology difficult management. Crisis crisis of many poses most,Crisis response poses many of the most difficult information technology in crisis management.,596
"we benefits response. the In in technologies this of explore crisis intelligence paper, artificial","In this paper, we explore the benefits of artificial intelligence technologies in crisis response.",597
has effects. extra-auditory both and auditory Noise,Noise has both auditory and extra-auditory effects.,598
are those extra-auditory most effects deleterious to of disturbances. leading Some the noise sleep of,Some of the most deleterious extra-auditory effects of noise are those leading to sleep disturbances.,599
on modulation of repeated sleep One is stimulation. tone by positive the focused,One is focused on the positive modulation of sleep by repeated tone stimulation.,600
noise-related disturbances that The the conclusions. have few sleep following suggest investigated studies,The few studies that have investigated noise-related sleep disturbances suggest the following conclusions.,601
"to permanently chronic noise and Third, sleep. reduces fragments exposure","Third, chronic exposure to noise permanently reduces and fragments sleep.",602
looping single-molecule use the in We to DNA techniques lac operon. dissect,We use single-molecule techniques to dissect DNA looping in the lac operon.,603
These are genome to relevant potentially the mechanism. and release modes icosahedral break symmetry,These modes break icosahedral symmetry and are potentially relevant to the genome release mechanism.,604
tiling theory dynamical viral on the such of comment considerations. in We role,We comment on the role of viral tiling theory in such dynamical considerations.,605
both coefficients? one without implement Should and linear equations with,Should one implement linear equations both with and without coefficients?,606
for introduced. model A derived is and propagators views,A model for views and derived propagators is introduced.,607
of the massive We evaluate impact derived propagators.,We evaluate the massive impact of derived propagators.,608
be dynamics Evolutionary studied or can well-mixed populations. structured in  ,  Evolutionary dynamics can be studied in well-mixed or structured populations.,609
we Here space introduce evolutionary of type phenotype game dynamics: space. to new a,Here we introduce a new type of space to evolutionary game dynamics: phenotype space.,610
might those Individuals dissimilar. or differently towards look who similar behave,Individuals might behave differently towards those who look similar or dissimilar.,611
space of study `phenotypic populations. We walk' the,We study the `phenotypic space walk' of populations.,612
present bring dynamics. We evolutionary that calculations coalescence together theory and game from ideas analytic,We present analytic calculations that bring together ideas from coalescence theory and evolutionary game dynamics.,613
"evolution we in example, phenotype space. particular the cooperation of investigate As a","As a particular example, we investigate the evolution of cooperation in phenotype space.",614
We derive evolutionary condition spaces. explore higher and dimensional fundamental phenotype any the game for,We derive the fundamental condition for any evolutionary game and explore higher dimensional phenotype spaces.,615
biochemistry theory. is system-wide statistical of approach One use studying to to graph organization the,One approach to studying the system-wide organization of biochemistry is to use statistical graph theory.,616
the encode definitions different about system. reaction Different information,Different definitions encode different information about the reaction system.,617
also choosing measure a tool as serves dimension-reduction when The natural parameters.,The measure also serves as a natural tool when choosing dimension-reduction parameters.,618
Both sample analyzed. case limit finite case the the and are,Both the finite sample case and the limit case are analyzed.,619
"we Finally, numerical results that our demonstrate claims. present","Finally, we present numerical results that demonstrate our claims.",620
project simulation The state-of-the-art multi-scale methods. uses,The project uses state-of-the-art multi-scale simulation methods.,621
basin-hopping are umbrella to minima the global reached. implemented using sampling We further confirm when,We implemented umbrella sampling using basin-hopping to further confirm when the global minima are reached.,622
optimization function transferable future scheme novel suggest results These for and a development. energy,These results suggest a novel and transferable optimization scheme for future energy function development.,623
complex are the For disjoint. more geometry actually the structurally subpopulations identified,For the more complex geometry the identified subpopulations are actually structurally disjoint.,624
broad less state heterogeneity. For the we complex with transition geometry microscopic native found a,For the less complex native geometry we found a broad transition state with microscopic heterogeneity.,625
extraction has segmentation two This and steps: splitting.,This extraction has two steps: segmentation and splitting.,626
paper framework introduces using This implement to a an SRGPA. algorithm,This paper introduces a framework to implement an algorithm using SRGPA.,627
around and framework This organization concepts: is applied of action. localization two built,This framework is built around two concepts: localization and organization of applied action.,628
unscientific the by  This due has  author paper to withdrawn extremely been errors.,  This paper has been withdrawn by the author due to extremely unscientific errors.,629
a propose models. penalty high-dimensional new generalized sparsity-smoothness for additive We,We propose a new sparsity-smoothness penalty for high-dimensional generalized additive models.,630
"large an yields version adaptive of Finally, additional sparsity-smoothness gains. our penalized performance approach","Finally, an adaptive version of our sparsity-smoothness penalized approach yields large additional performance gains.",631
scores performance single combination folding evaluated. critically is and The of predictors,The performance of single folding predictors and combination scores is critically evaluated.,632
predictors individual have than even better other These comparable or proposed methods. performance an,These predictors have an individual performance comparable or even better than other proposed methods.,633
"anisotropy. type exhibit depends on The migration, which a the of persistent cells substrate","The cells exhibit a persistent type of migration, which depends on the substrate anisotropy.",634
widespread is It suffering are that and now decline extinctions. well amphibians established,It is now well established that amphibians are suffering widespread decline and extinctions.,635
"is road massive urbanization other Among kills. habitat fragmentation reduction, causes, and for responsible habitat","Among other causes, urbanization is responsible for habitat reduction, habitat fragmentation and massive road kills.",636
"is measures. urgent appropriate assess it conservation context, this to develop and In","In this context, it is urgent to develop and assess appropriate conservation measures.",637
ultrametric. second is induced The an modelled by,The second is modelled by an induced ultrametric.,638
of data ordering account. temporal - into the,temporal - ordering of the data into account.,639
Intracellular molecular cytoskeletal on cargos based is along pull motors that transport filaments.,Intracellular transport is based on molecular motors that pull cargos along cytoskeletal filaments.,640
"one motor direction, One in always species e.g. moves","One motor species always moves in one direction, e.g.",641
exact but genetic load has for generalization. an Principle Haldane's unfamiliar,Haldane's Principle for genetic load has an exact but unfamiliar generalization.,642
of the illustrate several examples. We use results the in,We illustrate the use of the results in several examples.,643
originated the And life the force evolution of code driving in evolution. also genetic of,And the driving force of evolution of life also originated in the genetic code evolution.,644
the set described. of of examples evolution Two quasi-Mandelbrot continuous are,Two examples of continuous evolution of the quasi-Mandelbrot set are described.,645
of are number of Both any formulas information. sources presented for,Both formulas are presented for any number of sources of information.,646
with last other rules the section. given fusion comparison are application An and in,An application and comparison with other fusion rules are given in the last section.,647
oxidative of molecules. characterized by The effects biologic response a range stress various on is,The oxidative stress response is characterized by various effects on a range of biologic molecules.,648
\textit{Brassica of describes The present work novel Czern. process juncea} neoformation (L.) a,The present work describes a novel neoformation process of \textit{Brassica juncea} (L.) Czern.,649
to regenerated plants identical source plants. The were and fertile,The regenerated plants were fertile and identical to source plants.,650
[\textit{Brassica three juncea} Micropropagation of Indian (L.) of mustard genotypes Czern.],Micropropagation of three genotypes of Indian mustard [\textit{Brassica juncea} (L.) Czern.],651
significantly the in The shoot enhanced of all induction silver of nitrate addition rate genotypes.,The addition of silver nitrate significantly enhanced the rate of shoot induction in all genotypes.,652
article review specificities proteomics of membrane with This the deals the proteins. of analysis,This review article deals with the specificities of the proteomics analysis of membrane proteins.,653
"relatively the been considerable despite successes scarce. have However, effort,","However, despite considerable effort, the successes have been relatively scarce.",654
"Mitochondria organelles, requires a techniques. and proteomics analysis their combination complex are of","Mitochondria are complex organelles, and their proteomics analysis requires a combination of techniques.",655
Heavy agricultural land crop partly of contamination is productivity. responsible metal for limiting,Heavy metal contamination of agricultural land is partly responsible for limiting crop productivity.,656
type and plantlets. D.W.) the of whatever,and D.W.) whatever the type of plantlets.,657
critical thresholds than artificial cutoffs ruggedness to longer the The for ones. tolerable decreases,The artificial ruggedness decreases to tolerable thresholds for cutoffs longer than the critical ones.,658
recommended than results critical suggest cutoffs The can protein-protein that potentials. ones longer for be,The results suggest that cutoffs longer than critical ones can be recommended for protein-protein potentials.,659
"of of properties discuss native we the role Further, on crystals. the topology mechanical protein","Further, we discuss the role of native topology on the mechanical properties of protein crystals.",660
discharges intervals. and between alternate classes high-frequency neurons of silent Various,Various classes of neurons alternate between high-frequency discharges and silent intervals.,661
interpreted code. as compact be cannot representation a This firing-rate,This compact representation cannot be interpreted as a firing-rate code.,662
encode information total almost activity. burst half the investigated transmitted of neurons in The,The investigated neurons encode almost half of the total transmitted information in burst activity.,663
direct dealing The with presented. qualitative extension for transformation assignments of belief this also is,The direct extension of this transformation for dealing with qualitative belief assignments is also presented.,664
Centre a were using foot of pressure displacements platform. recorded force,Centre of foot pressure displacements were recorded using a force platform.,665
this been models dynamical have capture Many proposed phenomenon. to,Many dynamical models have been proposed to capture this phenomenon.,666
discrete a the present network We a as operon. for Boolean model lac,We present a Boolean network as a discrete model for the lac operon.,667
"traditional starting discuss multistage of models mathematical cancer theory. We from progression,","We discuss mathematical models of cancer progression, starting from traditional multistage theory.",668
cases or limit independently. follow order mutations occur arise a either linear if opposing Two,Two opposing limit cases arise if mutations either follow a linear order or occur independently.,669
"mutation we population that for stochastic accounts analyze Finally, model selection. genetics and explicitly a","Finally, we analyze a stochastic population genetics model that explicitly accounts for mutation and selection.",670
"population model, the at of In equidistant this waves intervals. sweep expansions through clonal","In this model, waves of clonal expansions sweep through the population at equidistant intervals.",671
We steps describe framework. the function general of a belief all,We describe all the steps of a general belief function framework.,672
investigate clustering cells. of We malignant glioma,We investigate clustering of malignant glioma cells.,673
mutation adhesion. strength hypothesize the affects We that the cell-cell of,We hypothesize that the mutation affects the strength of cell-cell adhesion.,674
"rate. that cluster, have cells increase form their we they evidence proliferation Second, when a","Second, when cells form a cluster, we have evidence that they increase their proliferation rate.",675
are Colour different. coarseness and skin  of  visually,  Colour and coarseness of skin are visually different.,676
proper with Grain size evaluated and diagrams. anisotropy are,Grain size and anisotropy are evaluated with proper diagrams.,677
also presence is determine the possibility pattern of defects discussed. The to,The possibility to determine the presence of pattern defects is also discussed.,678
cognitive behaviors. of to tries Each understand set cognitive science a,Each cognitive science tries to understand a set of cognitive behaviors.,679
cognitive enhancement modelling discuss and as We to computing. an modelling metacognitive,We discuss metacognitive modelling as an enhancement to cognitive modelling and computing.,680
certain a is within population Survival time a common probability measure viability. horizon of T,Survival probability within a certain time horizon T is a common measure of population viability.,681
steepest extinction method. approximate calculate using We descent times also a,We also calculate approximate extinction times using a steepest descent method.,682
"drug-therapy, drug, disease, have drug-target of introduced. networks been and number a Recently,","Recently, a number of drug-therapy, disease, drug, and drug-target networks have been introduced.",683
constructions visualisation of complex with in the distributions flux metabolic help Such networks. may,Such constructions may help with the visualisation of flux distributions in complex metabolic networks.,684
results are were for Detailed similar coli; organisms. other the found for results E. reported,Detailed results are reported for E. coli; similar results were found for the other organisms.,685
"analysis to of an propose, approach this the in sociality. article, We Darwin's","We propose, in this article, an analysis of the Darwin's approach to sociality.",686
"explains selective not process does the demonstrating on but Thus, contrary, prohibit the it. cooperation,","Thus, the selective process does not prohibit demonstrating cooperation, but on the contrary, explains it.",687
"are the noise. highly to sensitive Consequently, weight vectors","Consequently, the weight vectors are highly sensitive to noise.",688
the and single-stranded  complementary to Renaturation hybridization pairing of reactions nucleic lead  acids.,  Renaturation and hybridization reactions lead to the pairing of complementary single-stranded nucleic acids.,689
growth proceeds reaction of double then helix. fast through the a The,The reaction then proceeds through a fast growth of the double helix.,690
thermodynamic effect. from dependence a The length results excluded-volume,The length dependence results from a thermodynamic excluded-volume effect.,691
the start the proposed origin and stop is of An about explanation codons. also,An explanation is also proposed about the origin of the start and stop codons.,692
intrinsic on is We organization that processes. a focused and propose species model theoretical for,We propose a theoretical model for species that is focused on intrinsic organization and processes.,693
"only to explain is organisms. model required systems words, one In of other composed","In other words, only one model is required to explain systems composed of organisms.",694
"We are established. stability and of further this explore scheme, unique properties convergence","We explore further properties of this unique scheme, stability and convergence are established.",695
"enable consistent are interpretation These preliminary predictions and of, with, experimental data.","These predictions are consistent with, and enable interpretation of, preliminary experimental data.",696
discrepancies of suggest presence disordered the experimental These system. may states the in metastable,These discrepancies may suggest the presence of metastable disordered states in the experimental system.,697
is of for in cell maintenance eukaryotes. cytoskeleton essential the morphology The,The cytoskeleton is essential for the maintenance of cell morphology in eukaryotes.,698
"often Most based a a sites, assume on gamma rates models distribution distribution. across of","Most models assume a distribution of rates across sites, often based on a gamma distribution.",699
models. We decomposable component (PCA) Gaussian in consider analysis graphical principal,We consider principal component analysis (PCA) in decomposable Gaussian graphical models.,700
information We distribute to in the its order in exploit prior computation. models these,We exploit the prior information in these models in order to distribute its computation.,701
widely has ever result been This cited since.,This result has been widely cited ever since.,702
correspond of are the curve. to They regimes different to found species-area tri-phasic,They are found to correspond to different regimes of the tri-phasic species-area curve.,703
"In in built. neutrosophic way, logic/set/probability such new operators are","In such way, new operators in neutrosophic logic/set/probability are built.",704
promote known Such type coexistence. is interactions of species to,Such type of interactions is known to promote species coexistence.,705
"Reichenbach of al. Here, we recent generalize results et","Here, we generalize recent results of Reichenbach et al.",706
"coexistence. general diffusion can destroy species Nevertheless, in fast","Nevertheless, in general fast diffusion can destroy species coexistence.",707
to We four-state reaction also generalize rates. varying the model slightly,We also generalize the four-state model to slightly varying reaction rates.,708
"(SFA). role the spike-frequency Finally, we adaptation of investigate","Finally, we investigate the role of spike-frequency adaptation (SFA).",709
multiple Various objectives. search have problems scheduling under to applied approaches been machine recently local,Various local search approaches have recently been applied to machine scheduling problems under multiple objectives.,710
is of optimal the the Pareto set identification of Their foremost alternatives. consideration,Their foremost consideration is the identification of the set of Pareto optimal alternatives.,711
objective paper The operators presents search of flow multiple a neighbourhood for shop study scheduling.,The paper presents a study of neighbourhood search operators for multiple objective flow shop scheduling.,712
been of carried with out different Experiments have combinations twelve criteria.,Experiments have been carried out with twelve different combinations of criteria.,713
also We different phylogenetic with distinguishes reduced that networks it prove versions.,We also prove that it distinguishes phylogenetic networks with different reduced versions.,714
"considerable is interest. the interaction lupine, another also root the host-parasitoid Near of of","Near the root of the lupine, another host-parasitoid interaction is also of considerable interest.",715
this system In the of a host-parasitoid first spatial work introduced. discrete-time is model,In this work a discrete-time spatial model of the first host-parasitoid system is introduced.,716
"the that, success conservational the top metapopulation evaluates a of market. On model of","On top of that, a metapopulation model evaluates the conservational success of the market.",717
binding ligand-protein simple the binding. We here consider a kinetics four-state of in model,We consider here the binding kinetics in a simple four-state model of ligand-protein binding.,718
"ligand. has protein conformations, the model, both which this bind the can In two","In this model, the protein has two conformations, which can both bind the ligand.",719
"Nonetheless, remain essentially coalitions inequality. and societies ape weaken hierarchical, social rarely","Nonetheless, ape societies remain essentially hierarchical, and coalitions rarely weaken social inequality.",720
multiple concept tested The flow on is permutation objectives. successfully shop scheduling problems under,The concept is successfully tested on permutation flow shop scheduling problems under multiple objectives.,721
a framework The computer prototypically implemented system. in is,The framework is prototypically implemented in a computer system.,722
The optimization talk a multiple problems. for general objective describes approach genetic of algorithm a,The talk describes a general approach of a genetic algorithm for multiple objective optimization problems.,723
"that, In the support maker. to provides addition the decision powerful to program decision","In addition to that, the program provides powerful decision support to the decision maker.",724
has a approach algorithmic system. been prototypically computer implemented The in,The algorithmic approach has been prototypically implemented in a computer system.,725
"approximated single a several step. Often, by are reactions elementary","Often, several elementary reactions are approximated by a single step.",726
"situations often concentrations. relative practical use require Furthermore, the","Furthermore, practical situations often require the use relative concentrations.",727
cycle reproduces times experiments. the in clustering Our of model analytical observed,Our analytical model reproduces the clustering of cycle times observed in experiments.,728
heavily scoring on alignment depend alignments sequence Optimal parameters.,Optimal sequence alignments depend heavily on alignment scoring parameters.,729
for parameter biologically correct {\em But all suboptimal} alignments be choices. might,But biologically correct alignments might be {\em suboptimal} for all parameter choices.,730
objectives. approximation packing bin under to the multiple proposes article heuristic a problem The approach,The article proposes a heuristic approximation approach to the bin packing problem under multiple objectives.,731
algorithm An presented to solve problem. is approximation extension the the Best-Fit of,An extension of the Best-Fit approximation algorithm is presented to solve the problem.,732
recent models. many cortical to Two challenge a observations experimental pose,Two recent experimental observations pose a challenge to many cortical models.,733
Here use of these a to reconcile activity simple we model observations. cortical,Here we use a simple model of cortical activity to reconcile these observations.,734
strong it receive many neurons neurons. other that predicts than more connections some,it predicts that some neurons receive many more strong connections than other neurons.,735
local The optima. procedure overcome based search Accepting to on is heuristic Threshold,The heuristic search procedure is based on Threshold Accepting to overcome local optima.,736
index content the in libraries. search Most  of textual  digital engines documents,  Most search engines index the textual content of documents in digital libraries.,737
"figure Therefore, figures and are data extracting searching important for problems.","Therefore, searching for figures and extracting figure data are important problems.",738
shapes different overlapping that We points. to also data segregate correspond,We also segregate overlapping shapes that correspond to different data points.,739
"individual images. real-life demonstrate using algorithms, We combination of generated performance of and a","We demonstrate performance of individual algorithms, using a combination of generated and real-life images.",740
via results our We examples. illustrate theoretical several numerical,We illustrate our theoretical results via several numerical examples.,741
algebraic introduce of Here to we the exciting in biology new cophylogenetics. researchers field,Here we introduce researchers in algebraic biology to the exciting new field of cophylogenetics.,742
"various statistics propose phylogenetics, We algebraic cophylogenetics. topics in review research for for and analogs","We review various research topics in algebraic statistics for phylogenetics, and propose analogs for cophylogenetics.",743
"reconstruction, cophylogenetic we cophylogenetic propose trees, cophylogenetic invariants. of In and spaces particular","In particular we propose spaces of cophylogenetic trees, cophylogenetic reconstruction, and cophylogenetic invariants.",744
represents in a problem industrialized health overweight Significant countries. major,Significant overweight represents a major health problem in industrialized countries.,745
with obesity-associated pathology. additionally The interfering modes of novel insight present suggests,The present insight additionally suggests novel modes of interfering with obesity-associated pathology.,746
ones. of research models learning aimed at discriminative instead has reconstructive sparse been Recent purely,Recent research has been aimed at learning discriminative sparse models instead of purely reconstructive ones.,747
(ASP). extended introduce calculus for an tableau set programming We answer,We introduce an extended tableau calculus for answer set programming (ASP).,748
ASP Tableaux We empirically. of and power Extended theoretically investigate the both,We investigate the power of Extended ASP Tableaux both theoretically and empirically.,749
Theory appear To Practice of Logic Programming (TPLP). and in,To appear in Theory and Practice of Logic Programming (TPLP).,750
of are systems technical Black purely models descriptive. box,Black box models of technical systems are purely descriptive.,751
works it why the explain system does. not way do They a,They do not explain why a system works the way it does.,752
"problems. for models Thus, some black are insufficient box","Thus, black box models are insufficient for some problems.",753
essential in living and all processes Biological forces govern organisms. molecular cellular,Biological forces govern essential cellular and molecular processes in all living organisms.,754
"changes repetitive cyclic biological machines, have those of generated components. conformational in","those generated in cyclic conformational changes of biological machines, have repetitive components.",755
"stresses. mechanical repetitive process known about is how However, little proteins","However, little is known about how proteins process repetitive mechanical stresses.",756
a variation. associated particular the with Survival is probability typically 'fitness' of,Survival probability is typically associated with the 'fitness' of a particular variation.,757
Principles for also program decomposition are addressed.,Principles for program decomposition are also addressed.,758
Logic To in and Programming. Practice appear Theory of,To appear in Theory and Practice of Logic Programming.,759
DNA we this in apply to complexes kind with Here double-helical proteins. analysis of distorted,Here we apply this kind of analysis to distorted double-helical DNA in complexes with proteins.,760
with conformation are The to results respect uncertainty. robust and parameter,The results are robust with respect to parameter and conformation uncertainty.,761
and between complex of patterns reveal interaction They DNA. proteins the nano-mechanical,They reveal the complex nano-mechanical patterns of interaction between proteins and DNA.,762
in and Splines optimal smoothness higher have known to are one properties dimensions.,Splines are known to have optimal smoothness properties in one and higher dimensions.,763
The splines relationship is and to two shown dimensions. in one,The relationship to splines is shown in one and two dimensions.,764
includes against A study is benchmarking which methods. case given kriging,A case study is given which includes benchmarking against kriging methods.,765
residue. the force a have vectors each residuals vanishingly These on effect small on,These residuals have a vanishingly small effect on the force vectors on each residue.,766
proteins relevant of for biologically Implications of are study properties the discussed.,Implications for the study of biologically relevant properties of proteins are discussed.,767
related? quantities closely are so two these Why,Why are these two quantities so closely related?,768
inequality known with biological This in fits evolution. cases some well,This inequality fits well with some known cases in biological evolution.,769
engineering phase as an ISSAC in text preprocessing of implemented ontology a part system. is,ISSAC is implemented as part of a text preprocessing phase in an ontology engineering system.,770
homologous general is early for and colocalization recognition sequences of the A DNA proposed. model,A general model for the early recognition and colocalization of homologous DNA sequences is proposed.,771
"novelties of small this Consequently, sub-field extraction. term in are rare","Consequently, novelties are rare in this small sub-field of term extraction.",772
"empirically and addition, were existing mostly derived. work motivated In","In addition, existing work were mostly empirically motivated and derived.",773
"not Deleterious the mutations, establish other a hand, sector are to permanently. on able","Deleterious mutations, on the other hand, are not able to establish a sector permanently.",774
"reach front, thereby on ""surf"" frequencies. high can, and unusual the population They temporarily however,","They can, however, temporarily ""surf"" on the population front, and thereby reach unusual high frequencies.",775
"that we heavy. show be Finally, transforms to pyramid fishing biomass inverted the bottom","Finally, we show that fishing transforms the inverted biomass pyramid to be bottom heavy.",776
spp. benthic We groups two spatial of (Adelosina  analyze foraminifera the  of distributions,  We analyze the spatial distributions of two groups of benthic foraminifera (Adelosina spp.,777
study transition this velocity. the the to consequences of We properties cargo of,We study the consequences of this transition to the properties of cargo velocity.,778
we demonstrate inference for GLMMs. this the for SMC method performing use of In article,In this article we demonstrate use of the SMC method for performing inference for GLMMs.,779
domain-independent in way. present novel that a algorithm a macros We computes,We present a domain-independent algorithm that computes macros in a novel way.,780
itself magnitude responses of between be correlations of of stimulus-dependent. neurons pairs can stimulus-driven The,The magnitude of correlations between stimulus-driven responses of pairs of neurons can itself be stimulus-dependent.,781
rates. firing hold correlations relationships with Opposite when decrease,Opposite relationships hold when correlations decrease with firing rates.,782
We parametric multistability these leading bifurcation derive to in systems. trends the structure and,We derive the bifurcation structure and parametric trends leading to multistability in these systems.,783
modulated folding experiments agent is path time-resolved confirmed by apoflavodoxin's geometry. test-tube crowding that The,The test-tube experiments confirmed that apoflavodoxin's time-resolved folding path is modulated by crowding agent geometry.,784
have networks a and of impact. biological similar will likely alignment Comparison,Comparison and alignment of biological networks will likely have a similar impact.,785
"Topology-based new, provide phylogenetic have a alignments potential source independent to completely of information. the","Topology-based alignments have the potential to provide a completely new, independent source of phylogenetic information.",786
of this We key general discuss more briefly of networks mechanism consequences for neurons. biophysical,We briefly discuss key consequences of this mechanism for more general networks of biophysical neurons.,787
transfer gene from of individuals. both and between recombination a pairs population We horizontal investigate,We investigate both horizontal gene transfer from a population and recombination between pairs of individuals.,788
"these theory biology analytically solvable. nonlinear Somewhat surprisingly, to modern are generalizations of quasi-species","Somewhat surprisingly, these nonlinear generalizations of quasi-species theory to modern biology are analytically solvable.",789
"rigid. recombination, two two-parent we of selected find For is spectrally one phases, which","For two-parent recombination, we find two selected phases, one of which is spectrally rigid.",790
quasi-species holds the hypothesis These for prove models. that results deterministic mutational,These results prove that the mutational deterministic hypothesis holds for quasi-species models.,791
numerical present as phase as analytical diagrams cases. results different well We the and for,We present numerical and analytical results as well as phase diagrams for the different cases.,792
"the that Without surprisingly find we small. advantage of is exception, the filters ICA","Without exception, we find that the advantage of the ICA filters is surprisingly small.",793
scoring All on biological heuristic alignment on expertise. algorithms existing based rely schemes,All existing alignment algorithms rely on heuristic scoring schemes based on biological expertise.,794
and as build study to better. is classifiers The strong generalize critical can are that,The study is critical as to build classifiers that are strong and can generalize better.,795
The that function parameters the are varied. and nodes hidden activation the were,The hidden nodes and the activation function are the parameters that were varied.,796
shows relationship results The between structural is there observed accuracy. diversity and a that,The results observed shows that there is a relationship between structural diversity and accuracy.,797
increases as classification that observed an is ensemble the increases. of the accuracy diversity It,It is observed that the classification accuracy of an ensemble increases as the diversity increases.,798
when in rule AdaBoost's is minimizing update incremental viewed The by form. an loss derived,The update rule is derived by minimizing AdaBoost's loss when viewed in an incremental form.,799
The equations expensive. show is optimization computationally that,The equations show that optimization is computationally expensive.,800
"possible. approximation a is fast online However,","However, a fast online approximation is possible.",801
"imagery Hough lines, in will spread Properly corrected less curved and have fewer therefore space.","Properly corrected imagery will have fewer curved lines, and therefore less spread in Hough space.",802
model. the case host--parasitoid We of simulations study by a two-species,We study the two-species case by simulations of a host--parasitoid model.,803
"to case noise-sustained In leading oscillations. are the co-existence, patterns spatial there of","In the case of co-existence, there are spatial patterns leading to noise-sustained oscillations.",804
also overview A the behind brain is physiology sound of the in localization basic provided.,A basic overview of the physiology behind sound localization in the brain is also provided.,805
"a implement projection. and subspace we with process approximations dimensionality reduction simultaneous sparse Next,","Next, we implement a dimensionality reduction process with simultaneous sparse approximations and subspace projection.",806
the makeup genetic of The forest concern. artificially regenerated is of some,The genetic makeup of the artificially regenerated forest is of some concern.,807
presents the algorithm. This Optimization parser a unique article Ant using Colony a design for,This article presents a unique design for a parser using the Ant Colony Optimization algorithm.,808
the grammar. to present allocate in node a given production rule each We corresponding,We allocate a node corresponding to each production rule present in the given grammar.,809
"the path links the discovered carrying of amount concentration. optimum pheromone is maximum Eventually, by","Eventually, the optimum path is discovered by the links carrying maximum amount of pheromone concentration.",810
clustering an Data-stream of is knowledge extraction. subdomain ever-expanding,Data-stream clustering is an ever-expanding subdomain of knowledge extraction.,811
a the independence parameter. from in presentation This rigorous any results data or initialization ranking,This results in a rigorous independence from the data presentation ranking or any initialization parameter.,812
in assignment airport operations. of importance is Airport great gate,Airport gate assignment is of great importance in airport operations.,813
According et this by animal new (introduced al. Adoutte phylogeny to,According to this new animal phylogeny (introduced by Adoutte et al.,814
"traditional, deuterostome fit of details do new reconstructions. However, not many the phylogeny morphology-based","However, many details of the new deuterostome phylogeny do not fit traditional, morphology-based reconstructions.",815
"the chordate scenario, i.e. Therefore, on paper the focuses presented","Therefore, the paper presented focuses on the chordate scenario, i.e.",816
games study Here large but strategies populations of general we finite size. n in among,Here we study general games among n strategies in populations of large but finite size.,817
"stochastic evolutionary but for selection, under mutation We explore weak rate. dynamics any","We explore stochastic evolutionary dynamics under weak selection, but for any mutation rate.",818
mutation these two combination any linear A of for holds rate. conditions,A linear combination of these two conditions holds for any mutation rate.,819
of allow a of complete characterization limit weak results games Our in the selection. n*n,Our results allow a complete characterization of n*n games in the limit of weak selection.,820
simulations Model calculated from were data. to these constrained fluxes,Model simulations were constrained to fluxes calculated from these data.,821
"The Radial Multi-Layer of neural networks Basis investigated Function. types were and Perceptions,","The types of neural networks investigated were Multi-Layer Perceptions, and Radial Basis Function.",822
and these committee applied were techniques neural networks. Bayesian to,Bayesian and committee techniques were applied to these neural networks.,823
simulated in Each considered Matlab. methods of the was AI,Each of the AI methods considered was simulated in Matlab.,824
dynamics presented. is model An population agent-based of,An agent-based model of population dynamics is presented.,825
"are Individuals and selection. at potential encounter foraging through optimal natural prey, arrived strategies","Individuals encounter potential prey, and optimal foraging strategies are arrived at through natural selection.",826
DNA We for single experiments. in mechanical a present the self-contained of response theory molecule,We present a self-contained theory for the mechanical response of DNA in single molecule experiments.,827
are experimental found good in They data. with agreement,They are found in good agreement with experimental data.,828
proof-of-concept a velocity attractor form These that may integration dMEC. continuous underlie findings in dynamics,These findings form a proof-of-concept that continuous attractor dynamics may underlie velocity integration in dMEC.,829
networks. further This property of the evolvability underscores the,This property further underscores the evolvability of the networks.,830
in in problems up come Such statistics. regularized machine and estimation learning,Such problems come up in regularized estimation in machine learning and statistics.,831
"of In the requires of addition $V$. values QQMM $F$, to (sub)gradient","In addition to values of $F$, QQMM requires the (sub)gradient of $V$.",832
"good search. the iterative provides control stopping First, QQMM over","First, QQMM provides good control over stopping the iterative search.",833
"minimizers uses a for method QQMM determining of complex $F$. Secondly, trial","Secondly, QQMM uses a complex method for determining trial minimizers of $F$.",834
"in stochastic this developments evolutionary we dynamics summarize of finite populations. game review, In recent","In this review, we summarize recent developments in stochastic evolutionary game dynamics of finite populations.",835
proposed the available implements estimator. is computer shrinkage that program A,A computer program is available that implements the proposed shrinkage estimator.,836
agreement are observations. results Our clinical in with,Our results are in agreement with clinical observations.,837
stochastic model. axons spike propagation within a myelinated in noise-assisted multi-compartment Hodgkin-Huxley We consider,We consider noise-assisted spike propagation in myelinated axons within a multi-compartment stochastic Hodgkin-Huxley model.,838
each Ranvier. from channels originates in The of finite ion of number node noise a,The noise originates from a finite number of ion channels in each node of Ranvier.,839
"an exonizations effects can for Thus, negative organism. transcriptome the have of","Thus, exonizations can have negative effects for the transcriptome of an organism.",840
arbitrary The allows Metropolis-within-Gibbs implemented blocks. sampling update,The implemented Metropolis-within-Gibbs update allows arbitrary sampling blocks.,841
software and is uses C Lua written in language extension configuration. The a simple in,The software is written in C and uses a simple extension language Lua in configuration.,842
dihedral complexes. Our recent results allow us to for calculations rationalize,Our calculations allow us to rationalize recent results for dihedral complexes.,843
indicates crosstalk operates junctional mechanism between structures. that these a This,This indicates that a crosstalk mechanism operates between these junctional structures.,844
"the In this to formation. causes stimulation which stress-fiber ROCK, actin turn, leads of","In turn, this leads to the stimulation of ROCK, which causes actin stress-fiber formation.",845
"sufficient for permeability. to However, alone not account thrombin-induced is this","However, this alone is not sufficient to account for thrombin-induced permeability.",846
"Ultimately, converge both junction cell-cell leakage. pathways cause and disruption to vascular provoke","Ultimately, both pathways converge to cause cell-cell junction disruption and provoke vascular leakage.",847
the regulatory on case ubiquitous illustrate method We linear nature: a in cascades.,We illustrate the method on a ubiquitous case in nature: linear regulatory cascades.,848
"Some naturally ""gaps"" exist evolutionary among lines. evolutionary","Some evolutionary ""gaps"" naturally exist among evolutionary lines.",849
"performance increases success evolutionary dynamics, game. with an in game the evolutionary reproductive In","In evolutionary game dynamics, reproductive success increases with the performance in an evolutionary game.",850
"strategy performs will population. $B$, strategy If in spread $A$ than strategy $A$ better the","If strategy $A$ performs better than strategy $B$, strategy $A$ will spread in the population.",851
mean (or the We exit process. average with analyze times) associated this times fixation,We analyze the mean exit times (or average fixation times) associated with this process.,852
depends $v$. exit time on the only The mean term unconditional constant,The unconditional mean exit time depends only on the constant term $v$.,853
two this We for demonstrate finding evolutionary commonly processes. microscopic applied,We demonstrate this finding for two commonly applied microscopic evolutionary processes.,854
cortex. potentials relay  fire cells to retina Thalamic from information that action transmit ,  Thalamic relay cells fire action potentials that transmit information from retina to cortex.,855
"activity. factor influences however, input, neural only that is Sensory one","Sensory input, however, is only one factor that influences neural activity.",856
"networks of dynamics, neurons, For modulate intrinsic pattern. example, such oscillations also of firing as","For example, intrinsic dynamics, such as oscillations of networks of neurons, also modulate firing pattern.",857
"Here, neurons information help might oscillations if downstream. to asked to we retinal convey","Here, we asked if retinal oscillations might help to convey information to neurons downstream.",858
as multiplexed thalamic Our show spike trains results operate two that channels.,Our results show that thalamic spike trains operate as two multiplexed channels.,859
"first. the conveyed information times, channel At more second than the even","At times, the second channel conveyed even more information than the first.",860
"a we new present structure. generating method this for a work, In threshold","In this work, we present a new method for generating a threshold structure.",861
of number polyomino threshold values. a Each discrete contains fixed,Each polyomino contains a fixed number of discrete threshold values.,862
high visual structure images produced threshold quality. The with halftone have this,The halftone images produced with this threshold structure have high visual quality.,863
Autoencoder data. missing neural is implemented to network the estimate,Autoencoder neural network is implemented to estimate the missing data.,864
is implemented network Genetic data. for estimating algorithm and optimization the missing,Genetic algorithm is implemented for network optimization and estimating the missing data.,865
Missing Missing mechanism At data treated likelihood maximum implementing Random is by as algorithm.,Missing data is treated as Missing At Random mechanism by implementing maximum likelihood algorithm.,866
the calculating The determined network error performance prediction. square by mean network of is the,The network performance is determined by calculating the mean square error of the network prediction.,867
by Forest. Decision network is optimized The implementing further,The network is further optimized by implementing Decision Forest.,868
emotions involves classification expression emotions. the of standard of The categorizing,The standard classification of emotions involves categorizing the expression of emotions.,869
for disparity unreliable reliability used filter estimation of disparity map. A was eliminating increasing the,A filter eliminating unreliable disparity estimation was used for increasing reliability of the disparity map.,870
and compared. represented Obtained algorithms results by were,Obtained results by algorithms were represented and compared.,871
"difference of is dynamics. mere not This of a difference evolutionary but state,","This difference is not a mere difference of state, but of evolutionary dynamics.",872
construct are in important an Aggregates that applications. many role an ASP plays,Aggregates are an ASP construct that plays an important role in many applications.,873
"sample type, also tumor deletionsare observed. given Within amplifications specific any and","Within any given tumor type, sample specific amplifications and deletionsare also observed.",874
aberrations them. for method intuitive and prioritize to such define We an sought,We sought for an intuitive method to define such aberrations and prioritize them.,875
"generating will a finite Typically, this automata. be mechanism","Typically, this generating mechanism will be a finite automata.",876
in optimal this way. detail algorithms We to sequences generated predict,We detail optimal algorithms to predict sequences generated in this way.,877
in classification are for tools powerful regression static Neural environments. and networks,Neural networks are powerful tools for classification and regression in static environments.,878
uses Discrete Transform sparsity the (DCT) in Cosine It the images domain. of,It uses the sparsity of images in the Discrete Cosine Transform (DCT) domain.,879
can also the pixels. and corrupted The correct proposed method detect,The proposed method can also detect and correct the corrupted pixels.,880
to and Practice Guideliens practices Health physicians. are recommendations unify supposed to propose,Health Practice Guideliens are supposed to unify practices and propose recommendations to physicians.,881
for developped task. this system a We present,We present a system developped for this task.,882
tool. with precise a We of conclude evaluation the,We conclude with a precise evaluation of the tool.,883
sensitive results sudden that The the chloroplast to showed microsatellites growth. of simulations population are,The results of the simulations showed that chloroplast microsatellites are sensitive to sudden population growth.,884
learning the research conducted real-world academia unrealistic ontology applications. remains in for on Most,Most research on ontology learning conducted in the academia remains unrealistic for real-world applications.,885
construction. in dependencies to progress work reports This paper aiming address ontology such undesirable during,This paper reports work in progress aiming to address such undesirable dependencies during ontology construction.,886
"enhance anti-PIFs Conversely, protect VE-cadherin cell from and cohesion. disruption","Conversely, anti-PIFs protect from VE-cadherin disruption and enhance cell cohesion.",887
allow a completely of integration Grids in also data way. the distributed secured,Grids allow also the integration of distributed data in a completely secured way.,888
"are cost, also features. low and favored user Speed, friendliness","Speed, low cost, and user friendliness are also favored features.",889
"these Silver features, with but its staining MS many limited. is combines compatibility of","Silver staining combines many of these features, but its compatibility with MS is limited.",890
"new staining We here, variant describe a is silver of completely that formaldehyde-free.","We describe here, a new variant of silver staining that is completely formaldehyde-free.",891
buffer in as used borate sugars are Reducing alkaline developers.,Reducing sugars in alkaline borate buffer are used as developers.,892
may that cells. migration experimentally interactions the of cancer It been shown influence contact has,It has been shown experimentally that contact interactions may influence the migration of cancer cells.,893
demonstrate proteins HI and facilitate study that We folding. several small,We study several small proteins and demonstrate that HI facilitate folding.,894
controlling the E. of network metabolism. dynamics scale We coli genetic study the large also,We also study the dynamics of the large scale genetic network controlling E. coli metabolism.,895
The issues families. exponential to associated the with these broader difficulties structure discrete relate of,The issues associated with these difficulties relate to the broader structure of discrete exponential families.,896
parts. in two issues the paper This re-examines,This paper re-examines the issues in two parts.,897
"both to from theoretical its relevance and computational maximum a discuss We estimation, standpoint. likelihood","We discuss its relevance to maximum likelihood estimation, both from a theoretical and computational standpoint.",898
"results analysis ERG to we Second, the our apply of models.","Second, we apply our results to the analysis of ERG models.",899
forms two consider of We organization. genome,We consider two forms of genome organization.,900
sexual the multi-chromosomed is exception the reproduction only for pathway The genome.,The only exception is the sexual reproduction pathway for the multi-chromosomed genome.,901
"distribution, depth, habitat The to behaviour (water the coast), distance (i.e. temporal","The temporal distribution, habitat (water depth, distance to the coast), behaviour (i.e.",902
convincing. the most hypothesis was foraging advantage The,The foraging advantage hypothesis was the most convincing.,903
"benefits the associations the mixed-species depend of appeared species. to However, on","However, the benefits of mixed-species associations appeared to depend on the species.",904
"can if algorithm And we find so, construct an to one? effective","And if so, can we find an effective algorithm to construct one?",905
we paper give this case. the answer a complete to general In,In this paper we give a complete answer to the general case.,906
level-k special The main property of SN-sets a idea a to use network. is in,The main idea is to use a special property of SN-sets in a level-k network.,907
"present the Topology Fuzzy extension the Intuitionistic Topologies. as Also, of we Neutrosophic an","Also, as an extension of the Intuitionistic Fuzzy Topology we present the Neutrosophic Topologies.",908
"In all is intractable. full summation terms over correction general,","In general, full summation over all correction terms is intractable.",909
similar \textit{et Reichenbach to of recent results These results al.} are,These results are similar to recent results of Reichenbach \textit{et al.},910
below. with functional silicon neuronal is modeling presented proposed framework networks interfaced chips A for,A proposed modeling framework for functional neuronal networks interfaced with silicon chips is presented below.,911
A two-species parasitoids and system is studied. of extended hosts spatially,A two-species spatially extended system of hosts and parasitoids is studied.,912
metrics patterns thereof. characterize to introduce several the onset and We,We introduce several metrics to characterize the patterns and onset thereof.,913
experiments. can The be protocols implemented in behavioral,The protocols can be implemented in behavioral experiments.,914
"errors. pyrosequencing generated reads the However, size contain are numerous in short by and","However, the reads generated by pyrosequencing are short in size and contain numerous errors.",915
"has Furthermore, specific the a reference in position genome. each read","Furthermore, each read has a specific position in the reference genome.",916
alignment accuracy the from obtained The consensus multiple from confirmed the the alignments. is of,The accuracy of the alignment is confirmed from the consensus obtained from the multiple alignments.,917
of extended  presented. autocatalytic systems spatially Analytical analysis hypercyclic and  is,  Analytical analysis of spatially extended autocatalytic and hypercyclic systems is presented.,918
stability are stationary of studied. Existence solutions and,Existence and stability of stationary solutions are studied.,919
that stable values proved spatially some solutions it For is parameter appear. non-uniform,For some parameter values it is proved that stable spatially non-uniform solutions appear.,920
"data our to Littorina We the apply snail, marine algorithm saxatilis. from","We apply our algorithm to data from the marine snail, Littorina saxatilis.",921
set-up. results cognitive An emergent from capability this,An emergent cognitive capability results from this set-up.,922
"different into Also translating are with them documentation along examples. programs for formats, provided and","Also provided are programs for translating them into different formats, along with documentation and examples.",923
the code are all for programs included. Source binaries and,Source code and binaries for all the programs are included.,924
will for reaction of thorough studies network space. the networks This library allow of,This library of networks will allow for thorough studies of the reaction network space.,925
divided reproduction into neutral followed can two by model selection. Such be phases:,Such model can be divided into two phases: reproduction followed by neutral selection.,926
"correlations be underestimated. gene-history may situation, this significantly In the","In this situation, the gene-history correlations may be significantly underestimated.",927
the case. We why explain this is,We explain why this is the case.,928
"our remains this In sparse. neuroanatomical of genomes, knowledge circuitry of era complete surprisingly","In this era of complete genomes, our knowledge of neuroanatomical circuitry remains surprisingly sparse.",929
basic for and knowledge into however clinical function. research brain Such is both critical,Such knowledge is however critical both for basic and clinical research into brain function.,930
and knowledge rationale We review briefly and techniques. experimental detail scientific medical and existing the,We detail the scientific and medical rationale and briefly review existing knowledge and experimental techniques.,931
"being transition attractors, types called Their graph modeled. cycles, organism cell state the represent of","Their state transition graph cycles, called attractors, represent cell types of the organism being modeled.",932
finding checking. for We a present bounded an algorithm model SAT-based attractors which uses,We present an algorithm for finding attractors which uses a SAT-based bounded model checking.,933
of case consist (AEs). events pharmacovigilance The several involving drugs databases reports adverse and,The pharmacovigilance databases consist of several case reports involving drugs and adverse events (AEs).,934
"consistently all methods are to Some signals, highlight i.e. applied","Some methods are applied consistently to highlight all signals, i.e.",935
drug AE. an significant and statistically a between associations all,all statistically significant associations between a drug and an AE.,936
for have as been Boolean networks useful models control. potentially proposed genetic,Boolean networks have been proposed as potentially useful models for genetic control.,937
uncorrelated approaches network assumed stability have structure. random Previous to,Previous approaches to stability have assumed uncorrelated random network structure.,938
from gene nontrivial have network paradigm. random different topology typically significantly Real networks the,Real gene networks typically have nontrivial topology significantly different from the random network paradigm.,939
to gene work the We application experimentally of networks. our discuss also possible inferred,We also discuss the possible application of our work to experimentally inferred gene networks.,940
the algorithm both decisional estimates data. REMAPF from The new and the epsilon-machine states,The new REMAPF algorithm estimates both the epsilon-machine and the decisional states from data.,941
"for new we detection. object a paper, keypoint-based this In approach propose","In this paper, we propose a new approach for keypoint-based object detection.",942
in to Traditional pose discard classifying keypoint-based and using methods individual misclassifications. estimation points consist,Traditional keypoint-based methods consist in classifying individual points and using pose estimation to discard misclassifications.,943
video sequences experimental in Successful obtained real-time reported. detection are object for results,Successful experimental results obtained for real-time object detection in video sequences are reported.,944
a a to probe biomolecules. of new Translocation nanopore properties physical technique experimental is through,Translocation through a nanopore is a new experimental technique to probe physical properties of biomolecules.,945
"a based Moreover, incorporates consolidation mining realistic a the data architecture memory process on algorithm.","Moreover, the architecture incorporates a realistic memory consolidation process based on a data mining algorithm.",946
"However, to is domain for hard the domains, explicitly. ill-defined knowledge many define","However, for many ill-defined domains, the domain knowledge is hard to define explicitly.",947
"death. out-of-phase and both model modes, oscillation in-phase exhibits The oscillatory and","The model exhibits both in-phase and out-of-phase oscillatory modes, and oscillation death.",948
as can oscillations. such Interactions oscillatory complex behaviours generate intermittent between modes,Interactions between oscillatory modes can generate complex behaviours such as intermittent oscillations.,949
direct logic axiomatisation over We a of give first this Minspaces.,We first give a direct axiomatisation of this logic over Minspaces.,950
procedure decision a a define of tableaux form We calculus. the in next,We next define a decision procedure in the form of a tableaux calculus.,951
algorithms with huge data. Data now to amount able deal are efficiently mining of,Data mining algorithms are now able to efficiently deal with huge amount of data.,952
patterns drawback. to be an may Post that of answer processing,Post processing of patterns may be an answer to that drawback.,953
aerial of populations fly hypothetical tsetse investigated. impact on spraying The is,The hypothetical impact of aerial spraying on tsetse fly populations is investigated.,954
extra while. be worth argued is to cost The,The extra cost is argued to be worth while.,955
does and uracil hypothesis between not thymine. wobble discriminate The,The wobble hypothesis does not discriminate between uracil and thymine.,956
the of Methylation in favor could form. uracil further stabilization keto,Methylation could favor further stabilization of uracil in the keto form.,957
up pair only and keto can form adenine. is present with Thymine in but,Thymine is present in keto form only and can pair up but with adenine.,958
enol can it pair. U-G is that form; why forms the Uracil easily the construct,Uracil can easily construct the enol form; that is why it forms the U-G pair.,959
"we implications of phenomena. findings aggregation protein of discuss the our Finally, context possible in","Finally, we discuss possible implications of our findings in the context of protein aggregation phenomena.",960
XML. to new format extended networks propose represent constraint using We a,We propose a new extended format to represent constraint networks using XML.,961
or constraints This to extension either intension. defined format represent allows us in in,This format allows us to represent constraints defined either in extension or in intension.,962
constraints. allows global It also us reference to,It also allows us to reference global constraints.,963
pancreatic functional $\beta$-cells. motif in network controlling appears The of regulatory gene the activity the,The motif appears in the gene regulatory network controlling the functional activity of pancreatic $\beta$-cells.,964
appropriate model parameter exhibits hysteresis in and bistability regions. The,The model exhibits bistability and hysteresis in appropriate parameter regions.,965
the the reducing region. of has the loop of negative feedback bistable effect size The,The negative feedback loop has the effect of reducing the size of the bistable region.,966
the has withdrawn author been ali pourmohammad. by paper   This,  This paper has been withdrawn by the author ali pourmohammad.,967
A it is digraph. if with its cover regular isomorphic network is,A network is regular if it is isomorphic with its cover digraph.,968
work a consisted The game. developing present in plateau,The present work consisted in developing a plateau game.,969
"ect.) cluedo, (monopoly, traditional the ones are There","There are the traditional ones (monopoly, cluedo, ect.)",970
"uses partitioning The procedure, a trees. simple randomized variant RPtree k-d the regressor of","The regressor uses the RPtree partitioning procedure, a simple randomized variant of k-d trees.",971
"otherwise planar Mathematically, protuberances an LD as modeled were ER spherical in membrane.","Mathematically, LD were modeled as spherical protuberances in an otherwise planar ER membrane.",972
"we LD bud-off thermodynamically show cannot otherwise. addition, is demixing necessary In that as","In addition, we show that demixing is thermodynamically necessary as LD cannot bud-off otherwise.",973
typical This below of is diameter size far experimentally the LD. yeast determined,This diameter is far below the experimentally determined size of typical yeast LD.,974
the Refer to for updates. latest www.urdme.org,Refer to www.urdme.org for the latest updates.,975
"came small-world networks and into of architecture Recently, focus. scale-free","Recently, networks of small-world and scale-free architecture came into focus.",976
are quantities a quadrupole moments dipole These the distribution. of charge and,These quantities are the dipole and quadrupole moments of a charge distribution.,977
model are that to more networks of yeast. structurally similar,model networks are structurally more similar to that of yeast.,978
stochastic was recently neural of formulated. at which A statistics all networks includes orders theory,A stochastic theory of neural networks which includes statistics at all orders was recently formulated.,979
"Each yields the equations, of level i.e. closed approximation","Each level of the approximation yields closed equations, i.e.",980
use. presentations neither suitable are intended nor for Often these human,Often these presentations are neither intended nor suitable for human use.,981
patterns. these for We propagation specialised two algorithms design,We design specialised propagation algorithms for these two patterns.,982
inherit propagation a patterns these and constraints occurrence thus using specified algorithm. Counting directly,Counting and occurrence constraints specified using these patterns thus directly inherit a propagation algorithm.,983
"copula--or ""nonparanormal""--for to dimensional Gaussian show how semiparametric inference. high use a We","We show how to use a semiparametric Gaussian copula--or ""nonparanormal""--for high dimensional inference.",984
satisfaction solving factor constraint in an Symmetry many problems. is important,Symmetry is an important factor in solving many constraint satisfaction problems.,985
of common we One type have is values. symmetric when symmetry,One common type of symmetry is when we have symmetric values.,986
"we to break have of In methods papers, symmetries. value studied a recent series","In a recent series of papers, we have studied methods to break value symmetries.",987
computational symmetry. Our on eliminating value results identify limits,Our results identify computational limits on eliminating value symmetry.,988
"that general. NP-hard all For in symmetric we is instance, values prove pruning","For instance, we prove that pruning all symmetric values is NP-hard in general.",989
"that in value practice. symmetry can Nevertheless, show experiments broken much be","Nevertheless, experiments show that much value symmetry can be broken in practice.",990
requirements which analysis. for analysis exercise necessary the Our for the domain includes are reasoner,Our analysis includes requirements for the domain reasoner which are necessary for the exercise analysis.,991
This ali the paper author pourmohammad. been has by withdrawn,This paper has been withdrawn by the author ali pourmohammad.,992
cost We an online of problems. estimating the present SAT solving for method,We present an online method for estimating the cost of solving SAT problems.,993
trained linear gathered search. the uses method start Our of model at on data a,Our method uses a linear model trained on data gathered at the start of search.,994
problems. We the random of and this show effectiveness method structured using,We show the effectiveness of this method using random and structured problems.,995
be in that predictions. demonstrate made can improve to predictions later early restarts We used,We demonstrate that predictions made in early restarts can be used to improve later predictions.,996
on for scenario trees. We constraint stochastic semantics provide programs based a,We provide a semantics for stochastic constraint programs based on scenario trees.,997
"conventional constraint can programs semantics, constraint this compile Using stochastic into programs. down (non-stochastic) we","Using this semantics, we can compile stochastic constraint programs down into conventional (non-stochastic) constraint programs.",998
allows the This us exploit of to power existing constraint full solvers.,This allows us to exploit the full power of existing constraint solvers.,999
organisms. observed is the Shape for important one the characteristics living in of structures,Shape is one of the important characteristics for the structures observed in living organisms.,1000
"shapes Here, propose restricts volume possible that a we the leaves. of constraint","Here, we propose a volume constraint that restricts the possible shapes of leaves.",1001
controls leaf small the growing that development. globally volume this show folded in We,We show that growing folded in this small volume controls globally the leaf development.,1002
"activity, Experience-dependent structural to of cortical changes changes lead microcurcuts in i.e.","Experience-dependent structural changes in cortical microcurcuts lead to changes of activity, i.e.",1003
neurons. Specific of of new can patterns to lead creation external between stimulation connections synaptic,Specific patterns of external stimulation can lead to creation of new synaptic connections between neurons.,1004
development cortical the simulation neurons shows morpholess that compromise of connectivity. Numerical,Numerical simulation shows that morpholess neurons compromise the development of cortical connectivity.,1005
"combinatorial introduce model decision problems and we involving probability, To constraint programming. stochastic uncertainty","To model combinatorial decision problems involving uncertainty and probability, we introduce stochastic constraint programming.",1006
states. attractor an by stable transiently neural activity time-series of autonomous an for characterized infinite,for an autonomous neural activity characterized by an infinite time-series of transiently stable attractor states.,1007
qualitatively. stimulus the learning signal internal generated the is diffusive influences whenever unsupervised A dynamics,A diffusive learning signal is generated unsupervised whenever the stimulus influences the internal dynamics qualitatively.,1008
shows history. endstate the is that the independent also assembly on This,This also shows that the endstate is independent on the assembly history.,1009
cells without move external Eukaryotic can guided cues. being by spontaneously,Eukaryotic cells can move spontaneously without being guided by external cues.,1010
formulation homogeneous `thermodynamic' derive this used theorem functions of evolution. to Euler's for is,Euler's theorem for homogeneous functions is used to derive this `thermodynamic' formulation of evolution.,1011
model Eigen derive for the for the $h$. We principle maximum general,We derive the maximum principle for the Eigen model for general $h$.,1012
cells. cytoplasm between communication in nucleus pore and supports eukaryotic nuclear molecular The,The nuclear pore supports molecular communication between cytoplasm and nucleus in eukaryotic cells.,1013
operation experimental We a analytical and transport the of by this system approach. consider combined,We consider the operation of this transport system by a combined analytical and experimental approach.,1014
"pore does not transport se determine Clearly, the per directionality.","Clearly, the pore per se does not determine transport directionality.",1015
"a in to steady-state. reach cargo concentration of cytoplasmic Moreover, similar ratio different nuclear","Moreover, different cargo reach a similar ratio of nuclear to cytoplasmic concentration in steady-state.",1016
"a distributed representation. achieves fixed-point has attractors, information associative memory The discretely and discrete model","The associative memory model has discretely distributed fixed-point attractors, and achieves a discrete information representation.",1017
that also neural to operates close conceivable criticality. processing therefore information It is,It is therefore conceivable that also neural information processing operates close to criticality.,1018
"this is relevant we Here, networks. for real the address mechanism question whether neural","Here, we address the question whether this mechanism is relevant for real neural networks.",1019
"for are difficult result, to non-expert the systems a users. As use","As a result, the systems are difficult to use for non-expert users.",1020
control of The by precision limited of randomness the individual is this events. inevitable molecular,The precision of this control is limited by the inevitable randomness of individual molecular events.,1021
"we optimal solutions. limit, set rich of simple a this in surprisingly Even find","Even in this simple limit, we find a surprisingly rich set of optimal solutions.",1022
Subsequent discuss of will versions complete the papers problem. more,Subsequent papers will discuss more complete versions of the problem.,1023
texts. the enhancing of background readability manipulation reducing and The signals the allows digital,The digital manipulation allows reducing the background signals and enhancing the readability of texts.,1024
"are data (schemata) Sensory encountered before, similar constituents thus input expectations kindled. activates with structures","Sensory input activates data structures (schemata) with similar constituents encountered before, thus expectations are kindled.",1025
"are attempt schema, employing a triggered. new another","a new attempt employing another schema, are triggered.",1026
as is tried method examples. simulated well on The real data as,The method is tried on simulated as well as real data examples.,1027
In is formulae new this Boolean technique dynamic a proposed. paper CNF for subsumption,In this paper a new dynamic subsumption technique for Boolean CNF formulae is proposed.,1028
widely population of conifers Chloroplast recent used microsatellites years. studies in have in genetic been,Chloroplast microsatellites have been widely used in population genetic studies of conifers in recent years.,1029
data gene experiments to We expression present with real-world and the artificial evaluate method. data,We present experiments with artificial data and real-world gene expression data to evaluate the method.,1030
SDIAG algorithm this Diagonalization). AJD new In a present (Spheric work we named,In this work we present a new AJD algorithm named SDIAG (Spheric Diagonalization).,1031
"Unfortunately, poor solubility these hampers molecules. the of analysis","Unfortunately, poor solubility hampers the analysis of these molecules.",1032
"at of rapid field expanded rate. has proteomics the In the years, last a ten","In the last ten years, the field of proteomics has expanded at a rapid rate.",1033
specific proteomics organelle is for commentary provided. A,A specific commentary for organelle proteomics is provided.,1034
quantities a charge physical higher used moments Dipole are describe to and distribution.,Dipole and higher moments are physical quantities used to describe a charge distribution.,1035
color the this In dipole for color moments we images. paper define,In this paper we define the color dipole moments for color images.,1036
"have aspects, to For colors, consider. three primary in we fact, the maps color three","For color maps in fact, we have three aspects, the three primary colors, to consider.",1037
occurs computer problem in rotation vision tasks many optimal This matrix estimation. involving,This problem occurs in many computer vision tasks involving optimal rotation matrix estimation.,1038
has of in problems. been reinvented as continuously solution The fields part different specific,The solution has been continuously reinvented in different fields as part of specific problems.,1039
"standards. opposite the However, without basic is comparison not two possible","However, the comparison is not possible without two opposite basic standards.",1040
"without opposites. impossible concept of the Arguably, description of the reality is scientific","Arguably, the scientific description of reality is impossible without the concept of opposites.",1041
is a dynamics The single-domain small of folding focus of simulations proteins current and experiments.,The folding dynamics of small single-domain proteins is a current focus of simulations and experiments.,1042
"these are folders', proteins i.e. Many of 'two-state","Many of these proteins are 'two-state folders', i.e.",1043
experiments. in directly observed conformations These be short-lived and are folded partially cannot,These partially folded conformations are short-lived and cannot be observed directly in experiments.,1044
"confidence sight, of the At be computation new first infeasible. to this seems interval","At first sight, the computation of this new confidence interval seems to be infeasible.",1045
were separately characterized. components The and purified of enzyme this,The components of this enzyme were separately purified and characterized.,1046
"of A proposed. statistical conditional (CRFs), model random fields profile called is families, protein","A statistical model of protein families, called profile conditional random fields (CRFs), is proposed.",1047
suggest sleep. occur I most that replacements during,I suggest that most replacements occur during sleep.,1048
"data. to their author's is the claim, by Contrary supported not other conclusion","Contrary to the author's claim, their conclusion is not supported by other data.",1049
"fluorene the sp. genes the on in by initial involved In attack study, Sphingomonas this","In this study, the genes involved in the initial attack on fluorene by Sphingomonas sp.",1050
showed Quantification preferred substrates. oxidation resulting of that phenanthrene products were fluorene and,Quantification of resulting oxidation products showed that fluorene and phenanthrene were preferred substrates.,1051
accumulated that gas single growth. Further analysis metabolite a during by chromatography revealed,Further analysis by gas chromatography revealed that a single metabolite accumulated during growth.,1052
by chemically of was Identification lactone. synthesized the confirmed analysis,Identification was confirmed by analysis of the chemically synthesized lactone.,1053
separation proteins on polyacrylamide electrophoretic after detects gels. staining Silver,Silver staining detects proteins after electrophoretic separation on polyacrylamide gels.,1054
"is stable the weeks. for completed, several Once stain","Once completed, the stain is stable for several weeks.",1055
such applied (SNA) been networks. Classical network to from methods online social have analysis,Classical methods from social network analysis (SNA) have been applied to such online networks.,1056
Implications from data assembly experimental elongation mechanisms of determining for this discussed. constraint are,Implications of this constraint for determining elongation mechanisms from experimental assembly data are discussed.,1057
in Programming. apply answer Set proof-theoretic techniques We,We apply proof-theoretic techniques in answer Set Programming.,1058
for properties operator Gelfond-Lifschitz of characterization A program. logic of continuity,A characterization of continuity properties of Gelfond-Lifschitz operator for logic program.,1059
programs characterization to models propositional (without of formulas. logic referring of A loop stable,A propositional characterization of stable models of logic programs (without referring to loop formulas.,1060
is semimodule the over positive A fuzzy mnesor a numbers. real space,A fuzzy mnesor space is a semimodule over the positive real numbers.,1061
be for used theoretical sets. fuzzy as framework It can,It can be used as theoretical framework for fuzzy sets.,1062
stark and are deterministic used contrast approximations. findings in Our commonly to weak-noise the,Our findings are in stark contrast to the commonly used deterministic and weak-noise approximations.,1063
parameter-dependent are of certain for persistence of time-invariant provided. networks networks For classes conditions,For time-invariant networks parameter-dependent conditions for persistence of certain classes of networks are provided.,1064
"new large margins We motivated for by theory boosting algorithm, the a present boosting.","We present a new boosting algorithm, motivated by the large margins theory for boosting.",1065
the in typically use appear avoid partial models. of We which related differential equations,We avoid the use of partial differential equations which typically appear in related models.,1066
"a approximates ""memoryless"" solutions. system which true the ODE introduce We","We introduce a ""memoryless"" ODE system which approximates the true solutions.",1067
"deterministic stochastic from analyze transition the we the Finally, to the phase.","Finally, we analyze the transition from the stochastic to the deterministic phase.",1068
provided by the We equivalence Markov Zhao prove al. criterion for et that,We prove that the criterion for Markov equivalence provided by Zhao et al.,1069
tasks. in key  Point  many computer component set is vision a registration,  Point set registration is a key component in many computer vision tasks.,1070
the density probability a alignment We problem. as consider two point of sets estimation,We consider the alignment of two point sets as a probability density estimation problem.,1071
also introduce a linear. computation fast reduces that We algorithm method the complexity to,We also introduce a fast algorithm that reduces the method computation complexity to linear.,1072
framework color statistical A is learning/inference for demosaicing presented.,A statistical learning/inference framework for color demosaicing is presented.,1073
estimation An expectation-maximization of them naturally for algorithm the both. suggests itself,An expectation-maximization algorithm naturally suggests itself for the estimation of them both.,1074
"little the development about is circuits. growth At how affects neuronal known of cortical present,","At present, little is known about how cortical growth affects the development of neuronal circuits.",1075
"period. In systematically changes contrast, this their arrangement over spatial","In contrast, their spatial arrangement changes systematically over this period.",1076
"more become young are animals. mature in layouts isotropic animals more columns While in band-like,","While in young animals columns are more band-like, layouts become more isotropic in mature animals.",1077
This has significant postulates. us to three led make,This has led us to make three significant postulates.,1078
kinetics. in contacts active folding an role Also play nonnative the,Also nonnative contacts play an active role in the folding kinetics.,1079
biomass that of finds FBA optimal production the alternate the solutions one rate. maximizes,FBA finds one of the alternate optimal solutions that maximizes the biomass production rate.,1080
high-flux (HFB) backbone in networks. to This of the observation metabolic led concept,This observation led to the concept of high-flux backbone (HFB) in metabolic networks.,1081
"obtained using a In particular computed HFB optima the FBA. work, previous using been has","In previous work, the HFB has been computed using a particular optima obtained using FBA.",1082
"information. no all three to formalism kinds is of However, single there specify","However, there is no single formalism to specify all three kinds of information.",1083
a identify global (or non-binary) and important constraint. new We,We identify a new and important global (or non-binary) constraint.,1084
considerable promise. Experimental results on show problem several domains,Experimental results on several problem domains show considerable promise.,1085
a constraints multiple objectives). to new range and of to chance,to multiple chance constraints and to a range of new objectives).,1086
We new based on equivalent) (but also scenarios. a semantics provide,We also provide a new (but equivalent) semantics based on scenarios.,1087
"this (nonstochastic) Using semantics, constraint we programs. down into can compile programs constraint conventional stochastic","Using this semantics, we can compile stochastic constraint programs down into conventional (nonstochastic) constraint programs.",1088
us allows power of This to full the existing exploit constraint solvers.,This allows us to exploit the full power of existing constraint solvers.,1089
added easily new decompositions to be These solvers. can,These decompositions can be easily added to new solvers.,1090
Such constraints. sharing propagation used between improve to can be,Such sharing can be used to improve propagation between constraints.,1091
with experiments We solver. a pseudo-Boolean our report in decomposition,We report experiments with our decomposition in a pseudo-Boolean solver.,1092
the to Rewriting leads equations coplanarity solution. simpler a,Rewriting the coplanarity equations leads to a simpler solution.,1093
remaining is unknowns bases. resolution performed method an using by The Grobner algebraic,The remaining unknowns resolution is performed by an algebraic method using Grobner bases.,1094
on The the synthetic results method. this efficiency and data of real show,The results on real and synthetic data show the efficiency of this method.,1095
"through all capture pairwise We taken context, as semantics relationships.","We capture semantics through context, taken as all pairwise relationships.",1096
analysis. of of Our filmscript is application that domain,Our domain of application is that of filmscript analysis.,1097
innovative characterization. particular to develop plot an we approach In,In particular we develop an innovative approach to plot characterization.,1098
"comparisons network properties."" are Since they infeasible, ""network or heuristics, on rely computationally","Since network comparisons are computationally infeasible, they rely on heuristics, or ""network properties.""",1099
of the well-known in relationship number and the the between species. logistic-like results This area,This results in the well-known logistic-like relationship between the area and the number of species.,1100
"languages natural of generation. is services solution proposed, as ontology combination a reasoning A and","A solution is proposed, as a combination of ontology reasoning services and natural languages generation.",1101
ubiquitous is cortex. neuronal a feature of Spontaneous activity,Spontaneous neuronal activity is a ubiquitous feature of cortex.,1102
modulates output. future reflects organization past network and spatiotemporal Its input,Its spatiotemporal organization reflects past input and modulates future network output.,1103
of to ability to respond range the stimuli. the greatest,the ability to respond to the greatest range of stimuli.,1104
dynamic programming The interesting perspective suggests generalizations.,The dynamic programming perspective suggests interesting generalizations.,1105
immunogold-labeled Ras membrane-bound into segregate domains. that shown micrographs nanocluster of Ras Electron molecules have,Electron micrographs of immunogold-labeled Ras have shown that membrane-bound Ras molecules segregate into nanocluster domains.,1106
"mass assumption is this law inconsistent with of the action. However,","However, this assumption is inconsistent with the law of mass action.",1107
"large-scale analysis of membrane proteins decades of task. Despite remains a difficult extensive the research,","Despite decades of extensive research, the large-scale analysis of membrane proteins remains a difficult task.",1108
"impossible? Taking [Santoni, review and Membrane proteins Rabilloud, un this proteomics: Molloy and amour","Taking this review [Santoni, Molloy and Rabilloud, Membrane proteins and proteomics: un amour impossible?",1109
"studies. used models In theoretical particular, stochastic are to widely simplify integrate-and-fire","In particular, stochastic integrate-and-fire models are widely used to simplify theoretical studies.",1110
The the statistical current. the on of generated properties depend stimulating input spikes,The statistical properties of the generated spikes depend on the stimulating input current.,1111
"very the distribution of For good the a data. quasi-static slow provides inputs, description","For slow inputs, the quasi-static distribution provides a very good description of the data.",1112
People make within decisions a frame. time important have to,People have to make important decisions within a time frame.,1113
"is or decision imperative effective to to making. employ Hence, aid means it strategy","Hence, it is imperative to employ means or strategy to aid effective decision making.",1114
"properties cognizance temporal adequate its Thus, management with highly is its of indispensable.","Thus, its adequate management with cognizance of its temporal properties is highly indispensable.",1115
knowledge decision-problems perspective Our is respect in to projects EI. of with,Our perspective of knowledge is with respect to decision-problems projects in EI.,1116
"fly translational two types to stimuli, corresponding displacements. views of The rotational and","The fly views two types of stimuli, corresponding to rotational and translational displacements.",1117
"poor. for are stimuli expected, translational be reconstructions may As the rather","As may be expected, for translational stimuli the reconstructions are rather poor.",1118
two This tools. framework on is based formal new,This new framework is based on two formal tools.,1119
"encode a combinatorial of a First, efficiently hierarchy pyramid partitions.","First, a combinatorial pyramid encode efficiently a hierarchy of partitions.",1120
"Secondly, the important geometric precisely of estimators measure regions. parameters geometric discrete some","Secondly, discrete geometric estimators measure precisely some important geometric parameters of the regions.",1121
and are Some illustrating experiments framework discussed. shown our,Some experiments illustrating our framework are shown and discussed.,1122
In step. the this we paper on focus reduction,In this paper we focus on the reduction step.,1123
We algorithms: and two Tree algorithm. the introduce algorithm the HeightMap reduction new,We introduce two new reduction algorithms: the Tree algorithm and the HeightMap algorithm.,1124
only algorithm mentioned briefly. Tree is The,The Tree algorithm is only mentioned briefly.,1125
also discussed in detail in HeightMap The is algorithm code. pseudo given and,The HeightMap algorithm is discussed in detail and also given in pseudo code.,1126
"especially are HeightMap the our algorithms, show algorithm, that and We significantly faster.","We show that our algorithms, and especially the HeightMap algorithm, are significantly faster.",1127
complexity HeightMap O(n^d). Such multivariate the has a of algorithm time version,Such a multivariate version of the HeightMap algorithm has time complexity O(n^d).,1128
"possess dendritic Such to cells activate the T capacity na\""ive cells. activated","Such activated dendritic cells possess the capacity to activate na\""ive T cells.",1129
stimulate T the led to This cells. of activated to production dendritic unable cells,This led to the production of activated dendritic cells unable to stimulate T cells.,1130
evaluation adaptive that Our regularization significantly shows methods. outperforms approach the variational standard,Our evaluation shows that the adaptive regularization approach significantly outperforms standard variational methods.,1131
as semimodule Mnesors over min-plus of integers. defined are elements a the,Mnesors are defined as elements of a semimodule over the min-plus integers.,1132
the decision following We sequential problem. consider,We consider the following sequential decision problem.,1133
"item at to selection Measurements (possibly a prior noisy) are cost. values known allowed, of","Measurements (possibly noisy) of item values prior to selection are allowed, at a known cost.",1134
optimize the process sequential goal to measurements of and selection. is overall decision The,The goal is to optimize the overall sequential decision process of measurements and selection.,1135
contour the of prior requires a curvature. estimation metric Building,Building the metric requires a prior estimation of contour curvature.,1136
"Random convex Forests, penalty is ensembles, Garrote-style trees particular for proposed. in A","A Garrote-style convex penalty for trees ensembles, in particular Random Forests, is proposed.",1137
functional penalty of nodes in groups The the trees. selects,The penalty selects functional groups of nodes in the trees.,1138
These as individual could of predictor variables. functions be simple as monotone,These could be as simple as monotone functions of individual predictor variables.,1139
"and This function visualization easily fit, a lends yields which itself interpretation. to parsimonious","This yields a parsimonious function fit, which lends itself easily to visualization and interpretation.",1140
wide a array The demonstrated performance empirical datasets. on of is,The empirical performance is demonstrated on a wide array of datasets.,1141
"data, also Results given. disease are data, data Parkinson's and recognition microarray vowel on","Results on vowel recognition data, Parkinson's disease data, and microarray data are also given.",1142
ceramic industry. an important in control issue the Quality tile is,Quality control is an important issue in the ceramic tile industry.,1143
e.g. by are complex spaces with feature Most learning typically predictions obtained (as accurate machines,Most accurate predictions are typically obtained by learning machines with complex feature spaces (as e.g.,1144
and analytically desirable illustrated simulations. FIRM are The in of investigated properties,The desirable properties of FIRM are investigated analytically and illustrated in simulations.,1145
"(ATP), triphosphate as adenosine releasing By tonically adenosine, transmission. suppressed synaptic accumulates which astrocytes","By releasing adenosine triphosphate (ATP), which accumulates as adenosine, astrocytes tonically suppressed synaptic transmission.",1146
game reproducing of a the Evolutionary individuals. strategies describes population dynamics in of successful spreading,Evolutionary game dynamics describes the spreading of successful strategies in a population of reproducing individuals.,1147
"the recovered. from new selection, this process weak Moran under process is frequency dependent Additionally,","Additionally, under weak selection, from this new process the frequency dependent Moran process is recovered.",1148
under natural selection. extension of usual makes it the weak dynamics evolutionary a This,This makes it a natural extension of the usual evolutionary dynamics under weak selection.,1149
on optimization. segmentation predominately based are parameter-laden techniques Image,Image segmentation techniques are predominately based on parameter-laden optimization.,1150
"a painstaking, process. been weights Setting suitably has these empirical","Setting these weights suitably has been a painstaking, empirical process.",1151
death receptor the is that domain) FADD adaptor apoptotic signal. (Fas-associated molecule main transmits death,FADD (Fas-associated death domain) is the main death receptor adaptor molecule that transmits apoptotic signal.,1152
"subcellular However, unknown. still localization FADD in of vivo was","However, in vivo subcellular localization of FADD was still unknown.",1153
conditioning justified is how this in setting? But richer,But how justified is conditioning in this richer setting?,1154
several subset we this creation. In vector for methods feature and iris selection paper propose,In this paper we propose several methods for iris feature subset selection and vector creation.,1155
image. captures Contourlet intrinsic transform the iris of geometrical structures,Contourlet transform captures the intrinsic geometrical structures of iris image.,1156
In ignore method fragile this bits. we,In this method we ignore fragile bits.,1157
interval in negative plots restitution QT of displayed and Hysteresis biphasic regions vs slope. TQ,Hysteresis in QT vs TQ interval plots displayed biphasic restitution and regions of negative slope.,1158
QT/TQ beta-adrenergically Ischaemic heart Conclusion: produces induced steeper failure restitution hysteresis. during slopes,Conclusion: Ischaemic heart failure produces steeper restitution slopes during beta-adrenergically induced QT/TQ hysteresis.,1159
failing This could of underlie arrhythmias. to the propensity hearts,This could underlie the propensity of failing hearts to arrhythmias.,1160
in side chains the Ca are the the different channels. and Na model of Only,Only the side chains are different in the model of the Ca and Na channels.,1161
in are solutions. for channels all Parameters same the both,Parameters are the same for both channels in all solutions.,1162
model. the structures used in No information from crystal is,No information from crystal structures is used in the model.,1163
as are approximated spheres. grossly chains Side,Side chains are grossly approximated as spheres.,1164
properties the and are of channels Ca very predicted Na different. The,The predicted properties of the Na and Ca channels are very different.,1165
Structure forces model. in the of is the consequence computed this,Structure is the computed consequence of the forces in this model.,1166
"objective algorithm functions, optimality by two and i.e. The provides swapping between flexibility","The algorithm provides flexibility and optimality by swapping between two objective functions, i.e.",1167
defense subtractive and when strategies the required. as preferential and,the preferential and subtractive defense strategies as and when required.,1168
"Threat-Asset TE Threat is stage, In pairing done. and Ranking","In TE stage, Threat Ranking and Threat-Asset pairing is done.",1169
"is translation invariant and to approach of scale, the hand. rotation The","The approach is invariant to scale, rotation and translation of the hand.",1170
been the the undertaken proposed have experiments show to of effectiveness Some approach.,Some experiments have been undertaken to show the effectiveness of the proposed approach.,1171
graphical associative models  Collective  exploit to accurate dependence output labelings. inter-instance more,  Collective graphical models exploit inter-instance associative dependence to output more accurate labelings.,1172
of models existing kind associativity gains. very accuracy limited support However restricts which,However existing models support very limited kind of associativity which restricts accuracy gains.,1173
symmetric clique potentials. through encouraged is Agreement,Agreement is encouraged through symmetric clique potentials.,1174
solution domain for provides inference-only This an adaptation.,This provides an inference-only solution for domain adaptation.,1175
information reduction error over bibliographic extraction illustrate unseen test domains. Our on significant experiments,Our experiments on bibliographic information extraction illustrate significant test error reduction over unseen domains.,1176
"data. as is set"" subsequently training ""clean document-specific used This","This ""clean set"" is subsequently used as document-specific training data.",1177
core of Assignment the (TEWA) SCR at and Evaluation sits system. Weapon Threat,Threat Evaluation and Weapon Assignment (TEWA) sits at the core of SCR system.,1178
an and binarize We describe enhance image. a fingerprint to algorithm,We describe an algorithm to enhance and binarize a fingerprint image.,1179
"e.g. applications There many are computer cuts of vision, graph in","There are many applications of graph cuts in computer vision, e.g.",1180
Multisite omnipresent proteins is of eukaryotic in cells. modification covalent,Multisite covalent modification of proteins is omnipresent in eukaryotic cells.,1181
"mechanism In a distributive processive mechanism. essence, rebindings turn can rapid enzyme-substrate a into","In essence, rapid enzyme-substrate rebindings can turn a distributive mechanism into a processive mechanism.",1182
We support key model introduce to to the a transform. ray--based the modification,We introduce a key modification to the ray--based model to support the transform.,1183
"a valued source"", negative We emitted rays. ""virtual potentially certain insert with for light radiance","We insert a ""virtual light source"", with potentially negative valued radiance for certain emitted rays.",1184
a field elements. look-up create of table light transformers optical We of canonical,We create a look-up table of light field transformers of canonical optical elements.,1185
"wavefronts coherence analyze free--space and propagation, For different possibilities. we","For free--space propagation, we analyze different wavefronts and coherence possibilities.",1186
from optics concepts powerful formulation to wave This computer and brings graphics. vision,This formulation brings powerful concepts from wave optics to computer vision and graphics.,1187
applications displays. cubic-phase imaging plate show in We and holographic,We show applications in cubic-phase plate imaging and holographic displays.,1188
violation-based interval and task constraint new algorithms. This uses global sweep,This new global constraint uses sweep and task interval violation-based algorithms.,1189
presentations and topic illustrate on the news corpora We scientific of articles. abstracts,We illustrate the topic presentations on corpora of scientific abstracts and news articles.,1190
between interaction involves terms. This the neighboring,This involves interaction between the neighboring terms.,1191
"generalized propose a this  stratified In order first-order we for structure. paper, ontology ","  In this paper, we propose a first-order ontology for generalized stratified order structure.",1192
the then using the theory classify model-theoretic We models techniques. of,We then classify the models of the theory using model-theoretic techniques.,1193
registration difficult Medical image a problem. is,Medical image registration is a difficult problem.,1194
"""dose-response"" on curves generalizations were based proposed. Some more complicated","Some generalizations based on more complicated ""dose-response"" curves were proposed.",1195
ecosystems experimental Violations were natural law also this and in reported. of,Violations of this law in natural and experimental ecosystems were also reported.,1196
discussed. to econometrics data are interdisciplinary applications evidences are of empirical presented and Some,Some empirical data are presented and evidences of interdisciplinary applications to econometrics are discussed.,1197
analysed many facets the categories. into of distinct have We two Consciousness,We have analysed the many facets of Consciousness into two distinct categories.,1198
"in we For the numerical include paper. completeness, simulations","For completeness, we include numerical simulations in the paper.",1199
"the which wiring flow. propose to diagram, method network a reflects signal We visualize","We propose a method to visualize the wiring diagram, which reflects network signal flow.",1200
"the interaction networks. the gap Finally, analyze chemical the junction between synapse and we","Finally, we analyze the interaction between the gap junction and the chemical synapse networks.",1201
the process The pseudorandom number illustrated in is Mathematica. generator using,The process is illustrated using the pseudorandom number generator in Mathematica.,1202
in the of structure resulted decrease mutual signal in information. Introducing correlations input-output,Introducing correlations in signal structure resulted in the decrease of input-output mutual information.,1203
statistical estimators functionals. of various of consistent We investigate existence the bounded-memory,We investigate the existence of bounded-memory consistent estimators of various statistical functionals.,1204
rather negative sense. the question This in strong a is in resolved,This question is resolved in the negative in a rather strong sense.,1205
"stochastic approximations, automata We from using processes. and various propose theory bounded-memory techniques","We propose various bounded-memory approximations, using techniques from automata theory and stochastic processes.",1206
of interest for Some future work. raised are questions potential,Some questions of potential interest are raised for future work.,1207
a transcription with certain maximally within respond range. to These input motifs concentrations factors,These motifs respond maximally to input transcription factors with concentrations within a certain range.,1208
homeostatic motifs and pulse generating these From are networks derived.,From these motifs homeostatic and pulse generating networks are derived.,1209
is penalization which under give effective. variance sample conditions We,We give conditions under which sample variance penalization is effective.,1210
"by risk particular, method. bound a excess incurred we the the In on present","In particular, we present a bound on the excess risk incurred by the method.",1211
"some show We the which experimental results, theory. confirm","We show some experimental results, which confirm the theory.",1212
"results discuss Finally, compression potential of our to the schemes. application we sample","Finally, we discuss the potential application of our results to sample compression schemes.",1213
our are construction of the data for recommendations investigations. such Some derived from sets,Some recommendations for the construction of such data sets are derived from our investigations.,1214
well-known Experiments conducted using face database. FERET are the,Experiments are conducted using the well-known FERET face database.,1215
rates different in of response evolve strategies. dispersal Dispersal the relative to success,Dispersal rates evolve in response to the relative success of different dispersal strategies.,1216
"(J. a Hutson, simplified Mischaikow, et V. J. Dockery, K. al., deterministic In treatment","In a simplified deterministic treatment (J. Dockery, V. Hutson, K. Mischaikow, et al., J.",1217
activities about overview topic Web. The gives Semantic on an report the,The report gives an overview about activities on the topic Semantic Web.,1218
We of for cellular mean-field theory model the discuss a meta-learning. automata here,We discuss here the mean-field theory for a cellular automata model of meta-learning.,1219
double-layered real physical the is to The of structure. object tissue's the elastic construction closer,The construction of the double-layered elastic object is closer to the real tissue's physical structure.,1220
"localizations, connecting laterally. through and clusters network run Pathways them several typically","Pathways typically run through several network clusters and localizations, connecting them laterally.",1221
"systems change environment. can by their with Roughly descriptions interacting their these finite speaking,","Roughly speaking, these systems can change their finite descriptions by interacting with their environment.",1222
main results. contributions There four these produce used are to,There are four main contributions used to produce these results.,1223
crafted a expert. rule-based generative with by grammar a This human is specified,This is specified with a rule-based generative grammar crafted by a human expert.,1224
simulations. statistical settings encountered often are Such in,Such settings are often encountered in statistical simulations.,1225
solving two of present We different SAT the estimating  methods cost  for problems.,  We present two different methods for estimating the cost of solving SAT problems.,1226
random the these methods and We two using of compare problems. effectiveness structured,We compare the effectiveness of these two methods using random and structured problems.,1227
robust is of A subspace developed method on decomposition. classification basis sparse the,A robust classification method is developed on the basis of sparse subspace decomposition.,1228
is query the classified to the Each significantly subspace contributes whose class subspace. into decomposed,Each query is classified into the class whose subspace significantly contributes to the decomposed subspace.,1229
their from respective different classes. classes be classified queries simultaneously can Multiple into,Multiple queries from different classes can be simultaneously classified into their respective classes.,1230
subspace for the is greedy of sparse algorithm classification. practical the decomposition A designed,A practical greedy algorithm of the sparse subspace decomposition is designed for the classification.,1231
The and exploiting recognition joint rate performance achieves sparsity. high method robust present,The present method achieves high recognition rate and robust performance exploiting joint sparsity.,1232
"both satisfiability features, models. easy Based a on and compute we to classifier train runtime","Based on easy to compute features, we train both a satisfiability classifier and runtime models.",1233
these We to between use models strategies. choose restart,We use these models to choose between restart strategies.,1234
strategies. this technique restart with most present used comparing commonly We the results experimental,We present experimental results comparing this technique with the most commonly used restart strategies.,1235
effective in is that solver learning improving results demonstrate performance. Our machine,Our results demonstrate that machine learning is effective in improving solver performance.,1236
The face classification. based detection technique skin color fuzzy on information is and,The face detection technique is based on skin color information and fuzzy classification.,1237
components fiducial recognition. are points These are for by sets used wavelet described of which,These fiducial points are described by sets of wavelet components which are used for recognition.,1238
shows This powerful comparison coefficients Gabor that than distances. more geometric are,This comparison shows that Gabor coefficients are more powerful than geometric distances.,1239
agent of an We consider repeated sequence environment. interactions between an and a,We consider a sequence of repeated interactions between an agent and an environment.,1240
behavior. modeling network control disease for avoidance spread adaptive vaccine study We an disease on,We study vaccine control for disease spread on an adaptive network modeling disease avoidance behavior.,1241
distributed by vaccination susceptibles. adding is Poisson Control implemented of,Control is implemented by adding Poisson distributed vaccination of susceptibles.,1242
hydrocyclone application performance. describes methods of of approximate study analysis to some reasoning This,This study describes application of some approximate reasoning methods to analysis of hydrocyclone performance.,1243
can crisp non-crisp in be and iteration. granules implemented Balancing close-open of granules,Balancing of crisp granules and non-crisp granules can be implemented in close-open iteration.,1244
"the of Validation data set on methods, is of the hydrocyclone proposed rendered. the","Validation of the proposed methods, on the data set of the hydrocyclone is rendered.",1245
the role of of the the stability for clustering We algorithm. the investigate k-means initialization,We investigate the role of the initialization for the stability of the k-means clustering algorithm.,1246
"interested not in the clustering, costs in of the the only We solution. are actual","We are interested in the actual clustering, not only in the costs of the solution.",1247
"the view study approaches from solutions, outliers. Unlike this of other it","Unlike other solutions, this study approaches it from the view of outliers.",1248
"it rather robust a holistic runs making way, in expanding algorithm solution. a The","The expanding algorithm runs in a holistic way, making it a rather robust solution.",1249
and its power. real show Synthetic data experiments,Synthetic and real data experiments show its power.,1250
"multi-class application problems the leads Furthermore, to for of introduction algorithm. an oscillator the","Furthermore, an application for multi-class problems leads to the introduction of the oscillator algorithm.",1251
use the implies of wide corresponding the The expanding result algorithm. potential,The corresponding result implies the potential wide use of the expanding algorithm.,1252
a SIS stochastic epidemic analyse population amongst a We finite households. partitioned   into,  We analyse a stochastic SIS epidemic amongst a finite population partitioned into households.,1253
be Stochastic physiologically coherence occurs conditions relevant. that under may,Stochastic coherence occurs under conditions that may be physiologically relevant.,1254
implement and non-unit unit Should linear coefficients? both one constraints with,Should one implement linear constraints both with unit and non-unit coefficients?,1255
shows views use propagator paper derive to to how This variants. perfect,This paper shows how to use views to derive perfect propagator variants.,1256
for is A derived and propagators introduced. model views,A model for views and derived propagators is introduced.,1257
"developed. are as such systematically specialization, for and type deriving transformation, propagators Techniques generalization, conversion","Techniques for systematically deriving propagators such as transformation, generalization, specialization, and type conversion are developed.",1258
subdivision a data. image A lead to $p$-adic encoding shown hierarchical of to is interval,A hierarchical interval subdivision is shown to lead to a $p$-adic encoding of image data.,1259
of An the robustness. property inherent its is method,An inherent property of the method is its robustness.,1260
methods active exam geometric various radar segmentation. image for contour We,We exam various geometric active contour methods for radar image segmentation.,1261
images. meteorological in noises from Our is separating non-meteorological efficient method,Our method is efficient in separating non-meteorological noises from meteorological images.,1262
constructions. explore clustering these the We related between relationship also and,We also explore the relationship between these and related clustering constructions.,1263
in future. the posted be version will A rewritten,A rewritten version will be posted in the future.,1264
"the a nearest-prototype prototype introduce classifier, We (PVM). vector new machine","We introduce a new nearest-prototype classifier, the prototype vector machine (PVM).",1265
its solution. approximating propose algorithms We two for,We propose two algorithms for approximating its solution.,1266
"the of strengths This the highlights example a model. yielding low-error, interpretable PVM highly in","This example highlights the strengths of the PVM in yielding a low-error, highly interpretable model.",1267
molecular and large are machines. Proteins complex,Proteins are large and complex molecular machines.,1268
"order In their them function, need e.g. of perform most to energy,","In order to perform their function, most of them need energy, e.g.",1269
"sites stiffest Following single-site excitations, regions. within are systematically few targeted, specific the","Following single-site excitations, few specific sites are targeted, systematically within the stiffest regions.",1270
"more it models, for complex e.g. We extend Kriging","We extend it for more complex Kriging models, e.g.",1271
"a is it mind-map that emerges gets. framework, automatically the input, explorative from The dynamic","The explorative mind-map is a dynamic framework, that emerges automatically from the input, it gets.",1272
generic models classical include variation and the by model Barkley. The FitzHugh-Nagumo its,The generic models include the classical FitzHugh-Nagumo model and its variation by Barkley.,1273
are techniques. They standard singular perturbation treated with,They are treated with standard singular perturbation techniques.,1274
"introduced. of granulation several axiomatic definitions been information have Recently,","Recently, several axiomatic definitions of information granulation have been introduced.",1275
Advanced by Rgraphviz package. provided the options are plotting,Advanced plotting options are provided by the Rgraphviz package.,1276
"Recent dynamic single observe molecule in have phenomenon studies this ""age-old"" disorder directly. enzymology","Recent studies in single molecule enzymology have observe this ""age-old"" dynamic disorder phenomenon directly.",1277
"single its technique However, limitation. the molecule has","However, the single molecule technique has its limitation.",1278
"recognition. been very has date, reported for Farsi work Handwritten a Character Till little","Till date, a very little work has been reported for Handwritten Farsi Character recognition.",1279
to train The been algorithm MLP backpropagation error network. used the has,The error backpropagation algorithm has been used to train the MLP network.,1280
"MLP training system rigorous this an network. using results Finally, HCR automatic","Finally, this rigorous training results an automatic HCR system using MLP network.",1281
relations objects between knowledge. Direction important spatial are extended commonsense,Direction relations between extended spatial objects are important commonsense knowledge.,1282
"mean mixture to fit we Gaussian data. By ""covering"" a observed a model","By a ""covering"" we mean a Gaussian mixture model fit to observed data.",1283
immediate. Visual one preference for not is another over model,Visual preference for one model over another is not immediate.,1284
"are real In world multiple simultaneously. applications, involved aspects however, often spatial","In real world applications, however, multiple spatial aspects are often involved simultaneously.",1285
investigates topological arising paper combined reasoning problems This with and information. in directional,This paper investigates problems arising in reasoning with combined topological and directional information.,1286
to structure way? them a in How logical,How to structure them in a logical way?,1287
"What theory, the this is of n-opposition context? role in","What is the role of n-opposition theory, in this context?",1288
procedure problem. We to propose and efficient this simple an solve optimization,We propose an efficient and simple optimization procedure to solve this problem.,1289
data explain We graph. related how traffic then include we this to,We then explain how we include traffic related data to this graph.,1290
"selective stimulus in are or turn, upstrokes. cells, Biphasic up/down to down/up","Biphasic cells, in turn, are selective to up/down or down/up stimulus upstrokes.",1291
the setting. client-consultant study notion of a in We informedness,We study the notion of informedness in a client-consultant setting.,1292
"approach an on search. we heuristic preferred compute based propose forward-chaining To plans,","To compute preferred plans, we propose an approach based on forward-chaining heuristic search.",1293
of evaluation admissible function measuring preferences heuristic uses an partial satisfaction Our plans. over the,Our heuristic uses an admissible evaluation function measuring the satisfaction of preferences over partial plans.,1294
our evaluation effectiveness the demonstrates HTNPLAN of Our empirical heuristics.,Our empirical evaluation demonstrates the effectiveness of our HTNPLAN heuristics.,1295
fluctuation also studied. is property local The,The local fluctuation property is also studied.,1296
rendered and of crisp iteration. close-open granules rough is Balancing in granules sub,Balancing of crisp granules and sub rough granules is rendered in close-open iteration.,1297
field a The automobile where special attention. given fault a diagnosis crucial is is,The automobile is a crucial field where fault diagnosis is given a special attention.,1298
matching and techniques. done Fault using are similarity identification categorization,Fault categorization and identification are done using similarity matching techniques.,1299
automotive applications fault diagnosis In this techniques similarity discussed. in are paper for matching infotainment,In this paper similarity matching techniques for fault diagnosis in automotive infotainment applications are discussed.,1300
equations the We new dynamics sojourn this for state. each using times spent in investigate,We investigate this dynamics using new equations for the sojourn times spent in each state.,1301
calcium threshold calcium-binding oscillations. for proteins sharp set a multisite We that find,We find that multisite calcium-binding proteins set a sharp threshold for calcium oscillations.,1302
This similar to those produces strikingly in cells. real observed oscillatory behaviors,This produces oscillatory behaviors strikingly similar to those observed in real cells.,1303
an cellular in active calcium-binding may conclude and that We have proteins role important communication.,We conclude that calcium-binding proteins may have an important and active role in cellular communication.,1304
to adaptation Evolutionary peak.  hill climbing a is  likened often or,  Evolutionary adaptation is often likened to climbing a hill or peak.,1305
an important constraint Symmetry many of programs. feature is,Symmetry is an important feature of many constraint programs.,1306
class. symmetries out solutions different each pick symmetry Different in,Different symmetries pick out different solutions in each symmetry class.,1307
from problem. observations two a eliminating We in for these methods use symmetry,We use these observations in two methods for eliminating symmetry from a problem.,1308
Experimental on well two some show standard results methods perform that benchmarks. the,Experimental results show that the two methods perform well on some standard benchmarks.,1309
as the can defined Proteomics large-scale be of  proteins.  analysis,  Proteomics can be defined as the large-scale analysis of proteins.,1310
"rely with dealing techniques, chromatography electrophoresis. or These proteins peptides, on or can","These techniques, dealing with proteins or peptides, can rely on chromatography or electrophoresis.",1311
"review, techniques electrophoretic In this under scrutiny. the are","In this review, the electrophoretic techniques are under scrutiny.",1312
We causal inferring for linear method among describe relations variables. a multi-dimensional,We describe a method for inferring linear causal relations among multi-dimensional variables.,1313
Gaussian non-Gaussian as as data. well is It applicable to,It is applicable to Gaussian as well as non-Gaussian data.,1314
is characters isolated extract in and to plates recognize proposed. license A method,A method to extract and recognize isolated characters in license plates is proposed.,1315
analysis The extracted characters candidate are connected component scale doing images. DOG by different on,The candidate characters are extracted by doing connected component analysis on different scale DOG images.,1316
"strategies limited In under unbiased ""rational"" only stable are conditions. contrast,","In contrast, ""rational"" unbiased strategies are only stable under limited conditions.",1317
the upon improves best domain existing $O(\log algorithm a by of n)$. factor consistency This,This improves upon the best existing domain consistency algorithm by a factor of $O(\log n)$.,1318
are a derived these from in linear algorithms program. used The flows,The flows used in these algorithms are derived from a linear program.,1319
based points used data divergences. for Metrics on similarity are multi-dimensional between measuring symmetrical,Metrics used for measuring similarity between multi-dimensional data points are based on symmetrical divergences.,1320
"the we clustering the spectral application, method algorithms. In with also astrophysics the compare","In the astrophysics application, we also compare the method with the spectral clustering algorithms.",1321
the some here We of fuzzy be preferences where missing. consider constraint may problems,We consider here fuzzy constraint problems where some of the preferences may be missing.,1322
"possible. little ask goal However, to as is as the user our","However, our goal is to ask the user as little as possible.",1323
very effort best The algorithms. for also is the small user,The user effort is also very small for the best algorithms.,1324
interconversion the thermodynamics kinetics we the of mechanism. the From of and study characterize the,From the study of the thermodynamics and of the kinetics we characterize the interconversion mechanism.,1325
proteins Metamorphic expected. be more than may common,Metamorphic proteins may be more common than expected.,1326
unbiasedness the We of asymptotic estimate. DP-GLM mean regression conditions the for prove function,We prove conditions for the asymptotic unbiasedness of the DP-GLM regression mean function estimate.,1327
"networks transition on work Previous topology. complex, in non-random order/disorder network Boolean considered the","Previous work on the order/disorder transition in Boolean networks considered complex, non-random network topology.",1328
we to the previous behavior. paper account for current extend In work canalizing this,In the current paper we extend this previous work to account for canalizing behavior.,1329
to for flagellar bacteria physical model that a minimal the We motor enables swim. present,We present a minimal physical model for the flagellar motor that enables bacteria to swim.,1330
more food realistic conclusions numerically functional chain for response. Type-II are a with These confirmed,These conclusions are confirmed numerically for a food chain with more realistic Type-II functional response.,1331
functional of the reviews aspects paper This theory. statistical learning,This paper reviews the functional aspects of statistical learning theory.,1332
"these given. Furthermore, kernels way a to is design","Furthermore, a way to design these kernels is given.",1333
illustrate are and kernels framework To some such examples given. this sets reproducing of,To illustrate this framework some examples of such reproducing sets and kernels are given.,1334
"However, use these are and custom-made catheters. expensive","However, these are expensive and use custom-made catheters.",1335
for initial monoplane is and The feasibility algorithm on tested computer canine simulations data.,The monoplane algorithm is tested for feasibility on computer simulations and initial canine data.,1336
"especially many real-life in networks. communication Optimum Constrained on problems appear applications, (COP) Path","Constrained Optimum Path (COP) problems appear in many real-life applications, especially on communication networks.",1337
"Search. Local the modeling strengthens reuse and of modularity, the compositionality, features The Constrained-Based benefits","The modeling features the compositionality, modularity, reuse and strengthens the benefits of Constrained-Based Local Search.",1338
edge-disjoint also the paths We to apply modeling the (EDP). problem,We also apply the modeling to the edge-disjoint paths problem (EDP).,1339
can side the that in We be model. show easily added constraints,We show that side constraints can easily be added in the model.,1340
the of show significance approach. results the Computational,Computational results show the significance of the approach.,1341
"propositional (SAT for novel introduces This a hybrid HYbrid satisfiability. approach article Solver), SatHyS","This article introduces SatHyS (SAT HYbrid Solver), a novel hybrid approach for propositional satisfiability.",1342
and It local scheme. driven learning conflict clause (CDCL) combines search,It combines local search and conflict driven clause learning (CDCL) scheme.,1343
"local part search the CDCL launched. the is time minimum, local Each a reaches","Each time the local search part reaches a local minimum, the CDCL is launched.",1344
"and variables), set be of the to relaxed iteratively. re-optimized","the set of variables), to be relaxed and re-optimized iteratively.",1345
strategies the tailored The problem concerning properties. are two first,The first two strategies are tailored concerning the problem properties.,1346
upon benchmark strategies are three nurse and problem. rostering The a analyzed compared,The three strategies are analyzed and compared upon a benchmark nurse rostering problem.,1347
possibility the Promising demonstrate work results hybrid future approach. of the in,Promising results demonstrate the possibility of future work in the hybrid approach.,1348
circles). Our is detection on or based shape-detection sign (rectangles step only,Our sign detection step is based only on shape-detection (rectangles or circles).,1349
therefore backtrack instances solvers. search solved Unsatisfiable were using commonly,Unsatisfiable instances were therefore commonly solved using backtrack search solvers.,1350
"the data image volume. at capturing image device, generated in With high advancement the been","With the advancement in image capturing device, the image data been generated at high volume.",1351
"the information to are reveal users. If properly, they analyzed human useful can images","If images are analyzed properly, they can reveal useful information to the human users.",1352
categories important reveal a Grouping problem. images meaningful into useful challenging and information to is,Grouping images into meaningful categories to reveal useful information is a challenging and important problem.,1353
as color focuses framework feature. Proposed on,Proposed framework focuses on color as feature.,1354
used image Block Coding for extract are and Truncation (BTC) features dataset. Color to Moment,Color Moment and Block Truncation Coding (BTC) are used to extract features for image dataset.,1355
crowding macromolecular investigate also the assembly on the of We dynamics. effect,We also investigate the effect of macromolecular crowding on the assembly dynamics.,1356
algorithm design problems. such maximum We also clique an solving (ACF) for,We also design an algorithm (ACF) for solving such maximum clique problems.,1357
"compare with clique ACF one recently of the by finder, conceived fastest Ostergard. We furthermore","We furthermore compare ACF with one of the fastest clique finder, recently conceived by Ostergard.",1358
engine and for has many driving tasks. subtraction Background video a been computer vision analytics,Background subtraction has been a driving engine for many computer vision and video analytics tasks.,1359
"we Subsequently, a in implementation. characterize surrogates probabilistically, use practical but events low-memory, low-complexity","Subsequently, we characterize events probabilistically, but use a low-memory, low-complexity surrogates in practical implementation.",1360
"{\it some subtraction}, behavior properties. amounts Using new with to surprising algorithm a surrogates these","Using these surrogates amounts to {\it behavior subtraction}, a new algorithm with some surprising properties.",1361
"in and detection is subtraction demonstrated tool effective localization. here, behavior an anomaly As","As demonstrated here, behavior subtraction is an effective tool in anomaly detection and localization.",1362
"resulting the transcription are processes, The in of intrinsically noisy. RNA,","The processes, resulting in the transcription of RNA, are intrinsically noisy.",1363
"Here, regulated by consider the small gene model stochastic expression RNAs. we of","Here, we consider the stochastic model of gene expression regulated by small RNAs.",1364
base-pairing mRNA. Small is by RNA regulation achieved with post-transcriptional,Small RNA post-transcriptional regulation is achieved by base-pairing with mRNA.,1365
"due reveal transcriptional results description, discrepancies The the large bursting. obtained in to analytical deterministic","The obtained analytical results reveal large discrepancies in the deterministic description, due to transcriptional bursting.",1366
for The work proposed multi-species is of models. large approach a to variety expected,The proposed approach is expected to work for a large variety of multi-species models.,1367
tasks actions of learn sequence requiring a time. Animals over,Animals learn tasks requiring a sequence of actions over time.,1368
an a a action simple taking given is time before example. Waiting,Waiting a given time before taking an action is a simple example.,1369
"in heard. brief tune a humming have just humans, you","in humans, humming a brief tune you have just heard.",1370
time. a neural a activities pattern reproducing mentally of involve Re-experiencing sensory over must sequence,Re-experiencing a sensory pattern mentally must involve reproducing a sequence of neural activities over time.,1371
post-hyperpolarization to contribute Synaptic rebound adaptation cellular robustness. and,Synaptic adaptation and cellular post-hyperpolarization rebound contribute to robustness.,1372
"will Furthermore, fully the yet not appreciated processes. control I of allometric emphasize significance","Furthermore, I will emphasize the not yet fully appreciated significance of allometric control processes.",1373
used compare algorithm. objective performance PSO Two the different are functions to of,Two different objective functions are used to compare performance of the PSO algorithm.,1374
"structures of life Many and mechanisms, been unraveled. functions have","Many mechanisms, functions and structures of life have been unraveled.",1375
"we the the arising differential Here, of want calculus. possibilities use from discuss to fractional","Here, we want to discuss the possibilities arising from the use of fractional differential calculus.",1376
"was calculus Only involved image fractional processing recently, methods. in","Only recently, fractional calculus was involved in image processing methods.",1377
the al. with Li of Starting results et,Starting with the results of Li et al.,1378
present paper sequences. we two the studied In,In the present paper we studied two sequences.,1379
well-suited computing. automata concurrent not are on thus for based models are and These,These models are based on automata and thus are not well-suited for concurrent computing.,1380
the small theoretical A provided are to development. few illustrate examples,A few small examples are provided to illustrate the theoretical development.,1381
MAPK by characterized This phenomena. wave-like two is activity,This MAPK activity is characterized by two wave-like phenomena.,1382
"the develop we paper, diffusion-convection behavior. minimal this mechanistic reproduces observed that a model In","In this paper, we develop a minimal mechanistic diffusion-convection model that reproduces the observed behavior.",1383
to groups estimates trees Phylogenetic evolved. how of of display widely used are species,Phylogenetic trees are widely used to display estimates of how groups of species evolved.,1384
trees are phylogenetic When sets several for obtained (e.g. data,When phylogenetic trees are obtained for several data sets (e.g.,1385
"often contradicting. are their different for genes), then clusters","for different genes), then their clusters are often contradicting.",1386
and software. Dendroscope integrated freely it We have CASS implemented the in available,We have implemented CASS and integrated it in the freely available Dendroscope software.,1387
All the to model values. are experimentally of related parameters observable,All parameters of the model are related to experimentally observable values.,1388
open-source an published for project future SourceForge.net is extension. on and project possible development The,The project is an open-source project published on SourceForge.net for possible future development and extension.,1389
been to hypothesis has alternative the proposed. An wobble,An alternative to the wobble hypothesis has been proposed.,1390
expression gene developmental crucial temporal Precise for coordination of processes. many is,Precise temporal coordination of gene expression is crucial for many developmental processes.,1391
expression One central such robustly developmental controlled. how question is biology coordinated patterns are in,One central question in developmental biology is how such coordinated expression patterns are robustly controlled.,1392
evaluate the expression. of Drosophila stability the for sequential We network then actual,We then evaluate the stability of the actual Drosophila network for sequential expression.,1393
variables. We method both containing discrete infer and propose to causal continuous a structures,We propose a method to infer causal structures containing both discrete and continuous variables.,1394
current image algorithms is of arsenal vast. TV-based The for restoration,The current arsenal of algorithms for TV-based image restoration is vast.,1395
verified Analysis by is Mammographic mammograms Society. Image from using The algorithm,The algorithm is verified by using mammograms from Mammographic Image Analysis Society.,1396
applying these results obtained are by techniques described. The,The results obtained by applying these techniques are described.,1397
adjusted n coefficients Wavelet with in significant the bits average. least concert are (LSBs),Wavelet coefficients n least significant bits (LSBs) are adjusted in concert with the average.,1398
genomes. to individual genetic of in surplus due information the some retaining survived Populations,Populations survived due to retaining some surplus of genetic information in the individual genomes.,1399
the The size. of environment the the changes of population caused fluctuations,The changes of the environment caused the fluctuations of the population size.,1400
notion this we work of these to additive In cases. the noise extend models,In this work we extend the notion of additive noise models to these cases.,1401
are the strength resistance. proteins according The the ranked to of,The proteins are ranked according to the strength of the resistance.,1402
studied yet of predicted top-strength experimentally. have been Most not proteins the,Most of the predicted top-strength proteins have not yet been studied experimentally.,1403
are yield likely forces are Architectures to and large folds which identified.,Architectures and folds which are likely to yield large forces are identified.,1404
discovered. are New force of types clamps potent,New types of potent force clamps are discovered.,1405
"particular, disulphide They bridges slipknots. and, cysteine in involve","They involve disulphide bridges and, in particular, cysteine slipknots.",1406
class This dynamics molecular through characterized simulations. is,This class is characterized through molecular dynamics simulations.,1407
most this we paper of two. parameterize and at one matrices rank In non-negative sum,In this paper we parameterize non-negative matrices of sum one and rank at most two.,1408
"number of precisely, give family parameters. parameterizations we of a using possible the More least","More precisely, we give a family of parameterizations using the least possible number of parameters.",1409
lines EBV emerging profiles for Individual an from genomic transformed expression cell are resource investigation.,Individual expression profiles from EBV transformed cell lines are an emerging resource for genomic investigation.,1410
trans-acting additional in genetic analysis association age of discovery of variants. regulatory facilitates Incorporating,Incorporating age in analysis of association facilitates additional discovery of trans-acting regulatory genetic variants.,1411
works the the on originality segment mainly denoising of proposed The algorithm.,The originality mainly works on the denoising segment of the proposed algorithm.,1412
outs and images. real carried SAR in are Experimental both over simulated work,Experimental work outs are carried over both in simulated and real SAR images.,1413
implementation captures The individual-based realistic features. ecological,The individual-based implementation captures realistic ecological features.,1414
"extinction. can contemplate fact, spatial erratic structures emerge in In which regime oscillatory an predators'","In fact, spatial structures emerge in an erratic oscillatory regime which can contemplate predators' extinction.",1415
found coastal increasingly are habitats. structures Human-made in marine,Human-made structures are increasingly found in marine coastal habitats.,1416
null alleles. Hardy-Weinberg linked loci microsatellite disequilibrium of showed with strong likely Three presence the,Three microsatellite loci showed strong Hardy-Weinberg disequilibrium likely linked with the presence of null alleles.,1417
diversity than was in natural habitat Genetic in significantly higher habitat. artificial,Genetic diversity was significantly higher in natural habitat than in artificial habitat.,1418
to networks. extended an hypercyclic example for analysis is as catalytic Our couplings,Our analysis is extended to hypercyclic couplings as an example for catalytic networks.,1419
protein inhibition G biophysical channels characteristic Direct of recognized is N-type by modifications. calcium,Direct G protein inhibition of N-type calcium channels is recognized by characteristic biophysical modifications.,1420
emotion nowadays. Modelling challenge has a  become ,  Modelling emotion has become a challenge nowadays.,1421
"emotional have order several to express human models in been produced activity. Therefore,","Therefore, several models have been produced in order to express human emotional activity.",1422
autonomy. capabilities causing of sedentary loss can decrease behavior Aging motor or a,Aging or sedentary behavior can decrease motor capabilities causing a loss of autonomy.,1423
a emotion nowadays. challenge Modeling has become,Modeling emotion has become a challenge nowadays.,1424
"order models express human activity. Therefore, produced in emotional have several been to","Therefore, several models have been produced in order to express human emotional activity.",1425
model applied to This the ?Cascades? is,This model is applied to the ?Cascades?,1426
"pupils. arithmetical designed problem, small ten-year-old a for exercise","problem, a small arithmetical exercise designed for ten-year-old pupils.",1427
 for describes processing a an rendering. paper non-photorealistic image  The,  The paper describes an image processing for a non-photorealistic rendering.,1428
"a cover, the is An desired at canvas. applied iterative procedure level, to","An iterative procedure is applied to cover, at a desired level, the canvas.",1429
painting mimics resulting and effect The impressionist the Pointillism.,The resulting effect mimics the impressionist painting and Pointillism.,1430
"methods, using the the uniform those complex sensitivity. gave silver-diammine most long Among","Among the long methods, those using silver-diammine complex gave the most uniform sensitivity.",1431
special gels (e.g. applied however electrophoretic cannot home-made systems several require and They to be,They require however special home-made gels and cannot be applied to several electrophoretic systems (e.g.,1432
Future silver-staining will trends be also discussed. for,Future trends for silver-staining will also be discussed.,1433
a the sine of modulated wave). envelope,the envelope of a modulated sine wave).,1434
for convolution employs correlations processes. non-trivial outputs to approach One account between,One approach to account for non-trivial correlations between outputs employs convolution processes.,1435
the interpretation of convolution latent between establish dependencies a transform output Under function we variables.,Under a latent function interpretation of the convolution transform we establish dependencies between output variables.,1436
demands. this and the computational are of approach storage The associated main drawbacks,The main drawbacks of this approach are the associated computational and storage demands.,1437
In paper issues. address this these we,In this paper we address these issues.,1438
present the Gaussian through formalism. convolution constructed different output We processes for sparse dependent approximations,We present different sparse approximations for dependent output Gaussian processes constructed through the convolution formalism.,1439
naturally the in present model. exploit We the conditional independencies,We exploit the conditional independencies present naturally in the model.,1440
"a unconstrained full-rank covariance consider then models, specific including model. We four","We then consider four specific covariance models, including a full-rank unconstrained model.",1441
"in pigment imaging. (NIR) unavailable Nonetheless, Near-Infrared source rich provides VL, feature a in melanin","Nonetheless, pigment melanin provides a rich feature source in VL, unavailable in Near-Infrared (NIR) imaging.",1442
"stimulated This chemical biological not eumelanin, of in is due spectroscopy NIR. a to","This is due to biological spectroscopy of eumelanin, a chemical not stimulated in NIR.",1443
completely two consistent that give We the approaches find results.,We find that the two approaches give completely consistent results.,1444
"neighborhood such covering Recently, been have rough some as concepts sets. applied topological to","Recently, some topological concepts such as neighborhood have been applied to covering rough sets.",1445
"To the we unary analyze operations and composition themselves, introduce coverings. coverings on","To analyze the coverings themselves, we introduce unary and composition operations on coverings.",1446
relate homomorphismis approximation two provided covering notion spaces. A to of,A notion of homomorphismis provided to relate two covering approximation spaces.,1447
"respectively. approximations preserved the by the homomorphisms, properties examine of We operations also and","We also examine the properties of approximations preserved by the operations and homomorphisms, respectively.",1448
in size. population to large subject a neutral changes consider of its We population evolution,We consider neutral evolution of a large population subject to changes in its population size.,1449
"we a Within Eq. general approximation expression, coalescent have the obtained","Within the coalescent approximation we have obtained a general expression, Eq.",1450
of variations the We frequency investigate distributions. how the population-size alters,We investigate how the frequency of population-size variations alters the distributions.,1451
"close In the is equilibrium particular, distribution of to during that selection. stabilizing genotypes arising","In particular, the distribution of genotypes is close to that arising during equilibrium stabilizing selection.",1452
"the of improvements made granularity Recently, account by partitions. been some taking have into","Recently, some improvements have been made by taking into account the granularity of partitions.",1453
"measure roughness we In an this in axiomatic paper, the approach way.","In this paper, we approach the roughness measure in an axiomatic way.",1454
show unfolding pulling of this pathways protein that We the speed. depend on,We show that unfolding pathways of this protein depend on the pulling speed.,1455
process. of role Our important study non-native interactions reveals unfolding in an the,Our study reveals an important role of non-native interactions in the unfolding process.,1456
and comparison real with presented diseases Considerations are along.,Considerations and comparison with real diseases are presented along.,1457
sequence the genomes occurring to now throughout are world. cancer efforts Major,Major efforts to sequence cancer genomes are now occurring throughout the world.,1458
"we the paper, consider problem aggregation"". of ""hyper-sparse In this","In this paper, we consider the problem of ""hyper-sparse aggregation"".",1459
for this system describe acquiring We complete model. the,We describe the complete system for acquiring this model.,1460
the filters. vectors along selectivity dipole are corresponding both proteins The aligned of,The dipole vectors of both proteins are aligned along the corresponding selectivity filters.,1461
detailed a evolution results for of provide a basis the of understanding HIV. These more,These results provide a basis for a more detailed understanding of the evolution of HIV.,1462
fitness We a at below populations recombination large highest attain just critical. show that rate,We show that large populations attain highest fitness at a recombination rate just below critical.,1463
a blocks to disappear second transition allele in These selection. pure haplotype,These haplotype blocks disappear in a second transition to pure allele selection.,1464
opening closing noise channels neurons. produces stochastic and of The voltage-gated in ion,The stochastic opening and closing of voltage-gated ion channels produces noise in neurons.,1465
with the Even of persists. numbers the two models increasing difference between channels,Even with increasing numbers of channels the difference between the two models persists.,1466
"by work the ago, Since two numerous extensions nearly have Korf suggested. decades original been","Since the original work by Korf nearly two decades ago, numerous extensions have been suggested.",1467
empirically have to significant impact performance measures. various been idea shown This a has on,This idea has been empirically shown to have a significant impact on various performance measures.,1468
scheme of watermarking. invisible type proposed The and is blind,The proposed scheme is of type blind and invisible watermarking.,1469
channels of selected The into embedded watermark is pixel.,The watermark is embedded into selected channels of pixel.,1470
is by preserved using security watermark watermark the bits key. The secret of permuting,The security of watermark is preserved by permuting the watermark bits using secret key.,1471
of is convolution way processes based One on such constructing (CP). kernels,One way of constructing such kernels is based on convolution processes (CP).,1472
approach problem inference. is A for this key efficient,A key problem for this approach is efficient inference.,1473
"on our marks, prediction performance series. school demonstrate financial approaches time and of compiler We","We demonstrate our approaches on prediction of school marks, compiler performance and financial time series.",1474
consider binary surrogate question the also ``best'' of loss. the We,We also consider the question of the ``best'' surrogate binary loss.,1475
ways This suggests tuning''. characterisation ``surrogate new of,This characterisation suggests new ways of ``surrogate tuning''.,1476
principles. provides from to The first epidemics toolbox analyze framework capable a mathematical,The mathematical framework provides a capable toolbox to analyze epidemics from first principles.,1477
of recognition of (EMG) noises Electromyography signal. are major in problem Varieties,Varieties of noises are major problem in recognition of Electromyography (EMG) signal.,1478
"Hence, remove become in signal noise to analysis. methods most EMG significant","Hence, methods to remove noise become most significant in EMG signal analysis.",1479
in used noise interference represent to paper. is Gaussian this White (WGN),White Gaussian noise (WGN) is used to represent interference in this paper.,1480
WGN. tolerate research is feature novel This to that aimed present with a,This research is aimed to present a novel feature that tolerate with WGN.,1481
"noise result, As needed. algorithm a not is removal","As a result, noise removal algorithm is not needed.",1482
and median frequencies mean robust and MMDF) novel for are feature presented (MMNF Two extraction.,Two novel mean and median frequencies (MMNF and MMDF) are presented for robust feature extraction.,1483
noisy existing features in are a Sixteen two environment. and novelties evaluated,Sixteen existing features and two novelties are evaluated in a noisy environment.,1484
"with i.e. signal-to-noise various ratios WGN (SNRs),","WGN with various signal-to-noise ratios (SNRs), i.e.",1485
From feature above MMNF can be demonstrate robust new that for results the used extraction.,From the above results demonstrate that MMNF can be used for new robust feature extraction.,1486
proof merits. calibrated two provide We forecasters; yet it another has existence the of of,We provide yet another proof of the existence of calibrated forecasters; it has two merits.,1487
"valid number of for an arbitrary First, is outcomes. it finite","First, it is valid for an arbitrary finite number of outcomes.",1488
had changes. the shown to human buccal recover cells these ability Isolated,Isolated human buccal cells had shown the ability to recover these changes.,1489
the extends to domain domain graphs. the algorithms of k-means Euclidean This paper from,This paper extends k-means algorithms from the Euclidean domain to the domain of graphs.,1490
The computational a proposed approach methodology cost. at provides moderate feasible a inferential,The proposed approach provides a feasible inferential methodology at a moderate computational cost.,1491
scheme algorithm Cass Unlike only given here theoretical interest. the of is,Unlike Cass the algorithm scheme given here is only of theoretical interest.,1492
and perturbations around equilibrium by These solution an this small equilibrium. were considering solved,These were solved by considering an equilibrium solution and small perturbations around this equilibrium.,1493
quasispecies related life to dynamics. viral processes of origin describes model the The and evolutionary,The quasispecies model describes processes related to the origin of life and viral evolutionary dynamics.,1494
discuss for implications RNA highly viruses. the understanding of replication potential mutating We,We discuss potential implications for understanding the replication of highly mutating RNA viruses.,1495
multiple from trees genes evolutionary Gene representing the trees of sampled are ancestry populations.,Gene trees are evolutionary trees representing the ancestry of genes sampled from multiple populations.,1496
"methods trees Often, traditional gene phylogenetic inferred by unrooted. are however,","Often, however, gene trees inferred by traditional phylogenetic methods are unrooted.",1497
of investigate model. probabilities trees the multispecies gene We coalescent unrooted under,We investigate probabilities of unrooted gene trees under the multispecies coalescent model.,1498
not this the the species in tree root location on situation. identifiable is The of,The location of the root on the species tree is not identifiable in this situation.,1499
"fronts. the the technique, we associated optimization Pareto compute multi-objective In","In the multi-objective optimization technique, we compute the associated Pareto fronts.",1500
"information prove we systems. null Finally, that integrated is for disconnected","Finally, we prove that integrated information is null for disconnected systems.",1501
allows Dispersal that important strategy to locate an organisms exploit favorable habitats. and is,Dispersal is an important strategy that allows organisms to locate and exploit favorable habitats.,1502
"conclusion. reverse stochasticity However, this can demographic of introduction intrinsic the","However, the introduction of intrinsic demographic stochasticity can reverse this conclusion.",1503
"force. is Conversely, the to looped, once DNA-protein insensitive complex applied","Conversely, once looped, the DNA-protein complex is insensitive to applied force.",1504
findings present our identification an preliminary We such in experimental task.,We present our preliminary experimental findings in such an identification task.,1505
distributions. way probability networks powerful a belief complex are Deep to model,Deep belief networks are a powerful way to model complex probability distributions.,1506
"units, is particularly difficult. hidden with network, structure one a the belief of learning However,","However, learning the structure of a belief network, particularly one with hidden units, is difficult.",1507
reasoning received qualitative the spatial community. substantial work amongst This interest,This work received substantial interest amongst the qualitative spatial reasoning community.,1508
Our based dipole relations. condensed of semantics on the a are results,Our results are based on a condensed semantics of the dipole relations.,1509
objects. are important Both of features natural properties,Both features are important properties of natural objects.,1510
for of for spatial This straightforward reasoning allows tasks prototypical representation a agents.,This allows for a straightforward representation of prototypical reasoning tasks for spatial agents.,1511
of fast dipole illustrates constraint-based capabilities reasoning the The example the calculus.,The example illustrates the fast constraint-based reasoning capabilities of the dipole calculus.,1512
are into results integrate reasoning available. tools We two our which publicly,We integrate our results into two reasoning tools which are publicly available.,1513
"avalanches Thus, in criticality the a are brain. of neuronal robust manifestation","Thus, neuronal avalanches are a robust manifestation of criticality in the brain.",1514
of address set the in change answer problem semantics. belief We logic (nonmonotonic) programming under,We address the problem of belief change in (nonmonotonic) logic programming under answer set semantics.,1515
"merging methods revising We and for logic respectively. introduce programs,","We introduce methods for revising and merging logic programs, respectively.",1516
"IC respectively. we merging, consider merging, arbitration and operators For merging following","For merging, we consider operators following arbitration merging and IC merging, respectively.",1517
"with restricted the For problem sizes. of tree estimation, estimating graph forests consider we","For graph estimation, we consider the problem of estimating forests with restricted tree sizes.",1518
of graphs. clustering the results justify These existing domain algorithms statistically in,These results statistically justify existing clustering algorithms in the domain of graphs.,1519
quality. a by Empirical solution a significant results speedup show comparable maintaining,Empirical results show a significant speedup by maintaining a comparable solution quality.,1520
UNN a risk. by surrogate to learned Weak minimize so as classifiers are,Weak classifiers are learned by UNN so as to minimize a surrogate risk.,1521
database containing We scenes. natural experiments image on carried categorization of out eight a classes,We carried out image categorization experiments on a database containing eight classes of natural scenes.,1522
only fuzzy These quantified using can be sets.,These can only be quantified using fuzzy sets.,1523
implemented sources and using were ASP.net technique. mechanisms Rule-based the XML knowledge inference,XML Rule-based knowledge sources and the inference mechanisms were implemented using ASP.net technique.,1524
"tool ontology the the entering knowledge An and and was inferencing for built. interactive base,","An interactive tool for entering the ontology and knowledge base, and the inferencing was built.",1525
language sign need of hearing is increasing to community. The especially radically impaired,The need of sign language is increasing radically especially to hearing impaired community.,1526
(CI) Mamdani Intelligence in is an important study. Computational Model technique Fuzzy,Mamdani Fuzzy Model is an important technique in Computational Intelligence (CI) study.,1527
"technique. using of is auto Mamdani Specifically, modelled digital a function zoom camera","Specifically, auto zoom function of a digital camera is modelled using Mamdani technique.",1528
image mining have approaches this a hybrid proposed paper. been The in two manner with,The two image mining approaches with a hybrid manner have been proposed in this paper.,1529
been method the has The classify tree diagnosis. used to images for decision medical,The decision tree method has been used to classify the medical images for diagnosis.,1530
process This enhances accurate. be more the system to classification,This system enhances the classification process to be more accurate.,1531
intra-host which exist phenotypic diversity. viruses and RNA display populations great genotypic large in,RNA viruses exist in large intra-host populations which display great genotypic and phenotypic diversity.,1532
proteins function. protein of characterization to The is understand dynamical crucial,The dynamical characterization of proteins is crucial to understand protein function.,1533
"rearrangements atomic Unfortunately, between local protein relationship is global still the fluctuations elusive. and","Unfortunately, the relationship between local atomic fluctuations and global protein rearrangements is still elusive.",1534
adaptive positive guaranteeing We of Gibbs under convergence present samplers certain then results various conditions.,We then present various positive results guaranteeing convergence of adaptive Gibbs samplers under certain conditions.,1535
scheme watermarking based on paper new In this presented a log-average is luminance.,In this paper a new watermarking scheme is presented based on log-average luminance.,1536
"a more detection overcoming this As in result, becomes plagiarism issue. demanding","As a result, plagiarism detection becomes more demanding in overcoming this issue.",1537
based plagiarism developed have tools techniques. detection detection Different been various on,Different plagiarism detection tools have been developed based on various detection techniques.,1538
"in Nowadays, detection fingerprint those plays important matching technique an tools. role","Nowadays, fingerprint matching technique plays an important role in those detection tools.",1539
focused are funding biometrics. on research Governments programs,Governments are funding research programs focused on biometrics.,1540
widely gene RNAnet  used two bridge provides a databases. between Human ,  RNAnet provides a bridge between two widely used Human gene databases.,1541
Ensembl expression. transcripts sequences but describes DNA not experimental and gene,Ensembl describes DNA sequences and transcripts but not experimental gene expression.,1542
GEO levels expression from NCBI's Whilst Human contains actual samples.,Whilst NCBI's GEO contains actual expression levels from Human samples.,1543
can With seconds. access normalised cleaned anyone data quantile in RNAnet,With RNAnet anyone can access cleaned quantile normalised data in seconds.,1544
mine It data can patterns co-expression. of to be used,It can be used to data mine patterns of co-expression.,1545
genes correlated is but network The sparse. huge strongly of,The network of strongly correlated genes is huge but sparse.,1546
genes others. strongly interact thousands Thousands with of of,Thousands of genes interact strongly with thousands of others.,1547
views Biology. for Systems gives new RNAnet RNA,RNAnet gives new views for RNA Systems Biology.,1548
"more on low At contrasts, average high and contrasts. faster than at speed is variable","At low contrasts, speed is more variable and faster on average than at high contrasts.",1549
ecosystems. marine investigates often the power-law steady of This paper state the observed in stability,This paper investigates the stability of the power-law steady state often observed in marine ecosystems.,1550
steady power-law yield these a state. All of,All of these yield a power-law steady state.,1551
the stability knowledge state. properties of This of new power-law the steady provides,This provides new knowledge of the stability properties of the power-law steady state.,1552
why we proposed new breast is images. That in cancer cancer detect to algorithm mammogram,That is why we proposed new algorithm to detect cancer in mammogram breast cancer images.,1553
In using quantization vector this we technique. proposed segmentation paper,In this paper we proposed segmentation using vector quantization technique.,1554
Buzo-Gray segmentation used of images. Here Linde (LBG) we algorithm MRI for,Here we used Linde Buzo-Gray algorithm (LBG) for segmentation of MRI images.,1555
over leads does under approach not segmentation or to segmentation. This,This approach does not leads to over segmentation or under segmentation.,1556
components.  is process cell several  involving of the The a complex state regulation,  The regulation of the cell state is a complex process involving several components.,1557
"fate. contributions we the understand Consequently, in relative still not controlling cell their do","Consequently, we still do not understand their relative contributions in controlling the cell fate.",1558
gene Shwachman The for hSBDS Bodian Syndrome Diamond sapiens Homo is SDS. responsible,The Homo sapiens Shwachman Bodian Diamond Syndrome gene hSBDS is responsible for SDS.,1559
"events observed the Furthermore, the it recombination chromosomes. distributed has along been unevenly are that","Furthermore, it has been observed that the recombination events are distributed unevenly along the chromosomes.",1560
"propose current the we Fairness In book i.e. Principle, the","In the current book we propose the Fairness Principle, i.e.",1561
"examples book. inconsistent are Many consistent, in given and this weak inconsistent, strong","Many consistent, weak inconsistent, and strong inconsistent examples are given in this book.",1562
"almost on trees. primary Based available historical the on consisted the sources exclusively reports,","Based on historical reports, the available primary sources consisted almost exclusively on the trees.",1563
collapse Island community Easter structure why analysis of the about performed. is A mathematical the,A mathematical analysis about why the structure of the Easter Island community collapse is performed.,1564
print. index We threshold shoe as of global use the,We use global threshold as index of the shoe print.,1565
validate and included The method are paper this the to simulation method. results describes,The paper describes this method and simulation results are included to validate the method.,1566
results. a work of script printed review This paper Kannada their on and existing presents,This paper presents a review of existing work on printed Kannada script and their results.,1567
in Kannada Character characteristics are kcr script The discussed detail. of and Kannada Recognition System,The characteristics of Kannada script and Kannada Character Recognition System kcr are discussed in detail.,1568
increase recognition at to level proposed is fusion Finally the accuracy. the classifier,Finally fusion at the classifier level is proposed to increase the recognition accuracy.,1569
(i.e. the The of three seed co-occurring obligate-seeder dynamics bank,The seed bank dynamics of the three co-occurring obligate-seeder (i.e.,1570
remote for of from image. to base So noise removal the choose sensing method as,So as to choose the base method for removal of noise from remote sensing image.,1571
"the Wang systems, communication between To information et al. study","To study the communication between information systems, Wang et al.",1572
been there. of investigated have induced properties such Some functions relation and mappings,Some properties of such functions and induced relation mappings have been investigated there.,1573
in actual of distances taxa. additivity this divergence translates the between Mathematically of loss a,Mathematically this divergence translates in a loss of additivity of the actual distances between taxa.,1574
"et we mappings of al. some fuzzy addition, provided In improve characterizations relation Wang by","In addition, we improve some characterizations of fuzzy relation mappings provided by Wang et al.",1575
propagation input features. of design to back constitutes feed model material network prediction The forward,The input to feed forward back propagation prediction network model constitutes of material design features.,1576
furnished strategies design Parameters target are as outputs. product relevant to,Parameters relevant to product design strategies are furnished as target outputs.,1577
sign and of of between The good material ANN shows product design correlation domains. outcomes,The outcomes of ANN shows good sign of correlation between material and product design domains.,1578
(i.e. realized system by bistable is Switching often a,Switching is often realized by a bistable system (i.e.,1579
"paper, for we in this have In proposed mammogram. algorithm detecting an microcalcification","In this paper, we have proposed an algorithm for detecting microcalcification in mammogram.",1580
effectiveness experimental algorithm detection The the confirmed for is microcalcification results. of by proposed,The effectiveness of the proposed algorithm for microcalcification detection is confirmed by experimental results.,1581
"layers functions. In epithelial for separating form organisms, compartments multicellular different physiological cells responsible","In multicellular organisms, epithelial cells form layers separating compartments responsible for different physiological functions.",1582
basic features presented The MATS} described. are are and measures the briefly of {\tt implemented,The basic features of {\tt MATS} are presented and the implemented measures are briefly described.,1583
illustrated along usefulness empirical screenshots. examples {\tt of some on is The MATS} with,The usefulness of {\tt MATS} is illustrated on some empirical examples along with screenshots.,1584
Acquisition-to-acquisition inherent intensity in images. MR signal are (non-standardness) variations,Acquisition-to-acquisition signal intensity variations (non-standardness) are inherent in MR images.,1585
the performance. accuracy The that evaluations between variations degrades show images registration intensity of,The evaluations show that intensity variations between images degrades the accuracy of registration performance.,1586
choose stably often can phenotypes. Cells several heritable among,Cells can often choose among several stably heritable phenotypes.,1587
We with generalize valleys of in terms bistability associated landscape. epigenetic an this in mechanism,We generalize this mechanism in terms of bistability associated with valleys in an epigenetic landscape.,1588
The apart instances operations metamodel languages. from manipulate rewriting,The rewriting operations manipulate metamodel instances apart from languages.,1589
space metamodel using bridge languages is the A created between parsing techniques. and,A bridge is created between the metamodel space and languages using parsing techniques.,1590
from this Tools can world engineering software framework. useful the be to implement,Tools from the software engineering world can be useful to implement this framework.,1591
is important an re- constraint in task Transforming constraint systems. models programming cent,Transforming constraint models is an important task in re- cent constraint programming systems.,1592
"as rules, others, transformations an inition interesting i.e. concept-oriented Among feature of the is def-","Among others, an interesting feature is the def- inition of transformations as concept-oriented rules, i.e.",1593
method images. and this lossless compression of In offers for simple a medical paper compression,In this paper offers a simple and lossless compression method for compression of medical images.,1594
The the basis sub for analyses each prediction are equation correlation band. of,The correlation analyses are the basis of prediction equation for each sub band.,1595
is method on images. applied The and MRI CT,The method is applied on MRI and CT images.,1596
"quality, Teleophthalmology to affordability access, holds health improve a great the and in care. potential","Teleophthalmology holds a great potential to improve the quality, access, and affordability in health care.",1597
telemedicine easily image lends Ophthalmology to a largely diagnosis. itself is as it based,Ophthalmology lends itself easily to telemedicine as it is a largely image based diagnosis.,1598
Retinopathy will This of following type Drusen. Diabetic the system and diseases: diagnose,This system will diagnose the following type of diseases: Diabetic Retinopathy and Drusen.,1599
model-driven seen definitely be paradigm. can as Constraint a programming,Constraint programming can definitely be seen as a model-driven paradigm.,1600
The for programs write users problems. modeling,The users write programs for modeling problems.,1601
models calculate the are solutions. These to programs executable mapped to,These programs are mapped to executable models to calculate the solutions.,1602
and efficient on model focuses transformation). management This (definition paper,This paper focuses on efficient model management (definition and transformation).,1603
"point systems. propose design From to of revisit the constraint-programming we this of view,","From this point of view, we propose to revisit the design of constraint-programming systems.",1604
introduced solving-independent models. model-driven map to models A solving-dependent to decision is constraint architecture,A model-driven architecture is introduced to map solving-independent constraint models to solving-dependent decision models.,1605
"Image  algorithm proposed  is this for paper, In an Restoration.","  In this paper, an algorithm is proposed for Image Restoration.",1606
designed using the blurred code are The images. vectors,The code vectors are designed using the blurred images.,1607
"utilized. computationally make the efficient, restoration Nearest Quantization Neighborhood process the approaches To are","To make the restoration process computationally efficient, the Quantization Nearest Neighborhood approaches are utilized.",1608
"certain selection, the group conditions, i.e. of Under strength","Under certain conditions, the strength of group selection, i.e.",1609
(ICA) at into an statistically variables. observed Independent aims random vector independent component decomposing analysis,Independent component analysis (ICA) aims at decomposing an observed random vector into statistically independent variables.,1610
simple exact optimization This technique of kurtosis line function. search performing the consists contrast of,This simple technique consists of performing exact line search optimization of the kurtosis contrast function.,1611
"at polynomial cost, low iteration. at This rooting algebraically, can thus and performed be each","This polynomial rooting can be performed algebraically, and thus at low cost, at each iteration.",1612
The asymptotic absence performance. prewhitening improves of,The absence of prewhitening improves asymptotic performance.,1613
a demonstrated numerical are synthetic These features on by data. comparative analysis,These features are demonstrated by a comparative numerical analysis on synthetic data.,1614
or around polymers. nucleic assemble many viruses of acids Capsids other,Capsids of many viruses assemble around nucleic acids or other polymers.,1615
"eliminated first, variance. At the intensity coarse level is at on based background a","At first, the background is eliminated at a coarse level based on intensity variance.",1616
from distinct This makes other. each foreground the components,This makes the foreground components distinct from each other.,1617
are graphics. non-text Then various and using text the components removed of characteristic features,Then the non-text components are removed using various characteristic features of text and graphics.,1618
"skew are corrected further processing. for the Finally, regions text","Finally, the text regions are skew corrected for further processing.",1619
mutations our include We in back description.,We include back mutations in our description.,1620
a of derived are means results for These field-theoretic steady-state analytical by the representation.,These analytical results are derived for the steady-state by means of a field-theoretic representation.,1621
"must fluctuating, genes a environment cells which in coordinate provide function. Living cellular out-of-equilibrium","Living cells provide a fluctuating, out-of-equilibrium environment in which genes must coordinate cellular function.",1622
of this significance we codes. and functional In sparse the mechanisms analyze study olfactory,In this study we analyze the mechanisms and functional significance of sparse olfactory codes.,1623
We granule the cells. novel for propose a olfactory role,We propose a novel role for the olfactory granule cells.,1624
unevenly. of world samples parts Primate system the visual different,Primate visual system samples different parts of the world unevenly.,1625
against organism dormant poisonous environment. is The protected the,The dormant organism is protected against the poisonous environment.,1626
accurate this decisions. In theory of most the methodology assessments provides uncertainty in,In theory this methodology provides the most accurate assessments of uncertainty in decisions.,1627
leads from in posterior The sampling disproportional lack the details distribution. to,The lack in details leads to disproportional sampling from the posterior distribution.,1628
within populations is important In an many biological heterogeneity issue. processes cell,In many biological processes heterogeneity within cell populations is an important issue.,1629
accounted conditions. individual values Heterogeneity and cells by in among differences parameter is initial for,Heterogeneity among individual cells is accounted for by differences in parameter values and initial conditions.,1630
From density-based data noise-corrupted statistical model a derived. these is data,From these noise-corrupted data a density-based statistical data model is derived.,1631
"evaluate To caspase method, a considered. is the cascade activation model for the proposed","To evaluate the proposed method, a model for the caspase activation cascade is considered.",1632
sign dynamics nodes in resilience Knockout robustness. the remaining a is of the of,Knockout resilience in the dynamics of the remaining nodes is a sign of robustness.,1633
image into is card blocks. small A split,A card image is split into small blocks.,1634
on intensity are of as Some part the based classified background variance. of blocks these,Some of these blocks are classified as part of the background based on intensity variance.,1635
"is Experiment and the devices. mobile technique shows for that the efficient fast, applicable","Experiment shows that the technique is fast, efficient and applicable for the mobile devices.",1636
of the on ejection the present on measurements force ejection. dynamics We driving and self-repulsion,We present measurements on the dynamics of ejection and on the self-repulsion force driving ejection.,1637
We speculate facilitates in circularization genome cytoplasm. the this the of,We speculate this facilitates circularization of the genome in the cytoplasm.,1638
to Model metamodels. transformations conforming precisely defined models operate on,Model transformations operate on models conforming to precisely defined metamodels.,1639
"has obvious However, simple this rule some limitations.","However, this simple rule has some obvious limitations.",1640
"a instance, of use may subset For metamodel. only transformation a a","For instance, a transformation may only use a subset of a metamodel.",1641
"requires information. appropriately Therefore, transformations more chaining","Therefore, chaining transformations appropriately requires more information.",1642
case This of applied approach transformations. endogenous of library a been to has successfully the,This approach has been successfully applied to the case of a library of endogenous transformations.,1643
source target and some constraints. hidden chaining have have same the all metamodel They but,They all have the same source and target metamodel but have some hidden chaining constraints.,1644
"of first, level. At captured background image camera a coarse the is eliminated at a","At first, the background of a camera captured image is eliminated at a coarse level.",1645
are identified light-weight as technique. binarization an binarized but The CCs using then text adaptive,The CCs identified as text are then binarized using an adaptive but light-weight binarization technique.,1646
proposed. a descent specialized gradient is method,a specialized gradient descent method is proposed.,1647
horizontal contrast True localized and spatial on finally license based features. plates four are,True license plates are finally localized based on four spatial and horizontal contrast features.,1648
"We statistical supervised a documents. of introduce (sLDA), allocation labelled latent Dirichlet model","We introduce supervised latent Dirichlet allocation (sLDA), a statistical model of labelled documents.",1649
model types. variety accommodates response a of The,The model accommodates a variety of response types.,1650
(CACS). is proposed for called Solving Controller-Agents model Constraints The,The proposed model is called Controller-Agents for Constraints Solving (CACS).,1651
"Models have avoided strong networks fluctuations. learning, or like for traditionally perceptrons, neural","Models for learning, like neural networks or perceptrons, have traditionally avoided strong fluctuations.",1652
problem. insights understanding We the of here present toward provides the a novel study which,We present here a study which provides novel insights toward the understanding of the problem.,1653
features. exhibits Learning synaptic strengths and universal plastic adaptation of via occurs,Learning occurs via plastic adaptation of synaptic strengths and exhibits universal features.,1654
Even provided the that is slow. be learned sufficiently plastic complex adaptation can rules,Even complex rules can be learned provided that the plastic adaptation is sufficiently slow.,1655
"selection propose we GA-based paper, include a methodology the In to window this denoising technique.","In this paper, we propose a GA-based window selection methodology to include the denoising technique.",1656
the by procedure domain or frequency This filtering. performed spatial traditionally is in,This procedure is traditionally performed in the spatial or frequency domain by filtering.,1657
image. from aim enhancement is corrupted reconstruct image of the the to The true image,The aim of image enhancement is to reconstruct the true image from the corrupted image.,1658
technique for is enhancing a Filtering the image.,Filtering is a technique for enhancing the image.,1659
developed been are of techniques that smoothing have Thus non a linear. variety,Thus a variety of smoothing techniques have been developed that are non linear.,1660
the filter filter. is Median most of one the popular non-linear,Median filter is the one of the most popular non-linear filter.,1661
high under However condition. noise reduction the corrupted substantial and is original noise pixel,However the original pixel corrupted and noise reduction is substantial under high noise condition.,1662
the also has this image. Hence blurring affect technique on,Hence this technique has also blurring affect on the image.,1663
problem is general benchmark Handwritten Recognition of Artificial in numeral Pattern a Intelligence. and recognition,Handwritten numeral recognition is in general a benchmark problem of Pattern Recognition and Artificial Intelligence.,1664
considerable gait Analysing vision found in human has computer research. interest recent,Analysing human gait has found considerable interest in recent computer vision research.,1665
the differently. is face challenging expression expression The recognition because display individuals same different problem,The face expression recognition problem is challenging because different individuals display the same expression differently.,1666
Ultrasound is method of of detection imaging fetus. abnormality suitable early also the for,Ultrasound imaging method is also suitable for early detection of the abnormality of fetus.,1667
accurate of is detection crucial. of interest image The region in ultrasound,The accurate detection of region of interest in ultrasound image is crucial.,1668
the image more is ultrasound of analysis So challenging one.,So the analysis of ultrasound image is more challenging one.,1669
no approach noise reduction. common for Generally enhancement is there,Generally there is no common enhancement approach for noise reduction.,1670
the of framework model is DATALOG. in This worked out,This model is worked out in the framework of DATALOG.,1671
evaluation possible given. a is last At strategy,At last a possible evaluation strategy is given.,1672
work also include The alphabet. of be can extended of handwritten to Arabic OCR characters,The work can also be extended to include OCR of handwritten characters of Arabic alphabet.,1673
 for is java RefereeToolbox evidences. combination implementing operators  a fusing package,  RefereeToolbox is a java package implementing combination operators for fusing evidences.,1674
"and As is with be RefereeToolbox result, designed to evolutive. a aim the generic","As a result, RefereeToolbox is designed with the aim to be generic and evolutive.",1675
to Our attain twofold purpose. a enables us approach,Our approach enables us to attain a twofold purpose.,1676
by formalism analysis is General exemplified the experimental data. of literature,General formalism is exemplified by the analysis of literature experimental data.,1677
study of pools. to the problem We allocating stocks dark,We study the problem of allocating stocks to dark pools.,1678
"if for allowed. and approach propose allocations, continuous-valued analyze We optimal are an allocations","We propose and analyze an optimal approach for allocations, if continuous-valued allocations are allowed.",1679
are for integer-valued allocations also We when only a modification the case possible. propose,We also propose a modification for the case when only integer-valued allocations are possible.,1680
"resulting are efficient, perform adversarial inputs. in under and simulations and The well stochastic algorithms","The resulting algorithms are efficient, and perform well in simulations under stochastic and adversarial inputs.",1681
between pathways complicates which Proteins interactions often involved biological of multiple proteins. are interpretation in,Proteins are often involved in multiple biological pathways which complicates interpretation of interactions between proteins.,1682
This interpretation topology multiple and pathways greater permits biological biological easier simultaneously. refined coherence of,This refined topology permits easier interpretation and greater biological coherence of multiple biological pathways simultaneously.,1683
It was linear is image textured to invariant that EHD shift. shown,It was shown that EHD is invariant to textured image linear shift.,1684
signal Anomalous of this on point foreign objects. variations,Anomalous variations of this signal point on foreign objects.,1685
provides stationarity of model condition characteristic symmetry and polynomial periodicity. The the,The condition of characteristic polynomial symmetry provides the model stationarity and periodicity.,1686
methods from foreign Two retrieval of offered. of textured are classification objects image,Two methods of classification of retrieval from textured image foreign objects are offered.,1687
(GPUs) high-dimensional discusses optimization problems. the in This potential units of processing graphics paper,This paper discusses the potential of graphics processing units (GPUs) in high-dimensional optimization problems.,1688
criteria favor These algorithms and EM data. that MM parameters and separate,These criteria favor EM and MM algorithms that separate parameters and data.,1689
a coordinate also relaxation and To descent block lesser and extent ascent qualify.,To a lesser extent block relaxation and coordinate descent and ascent also qualify.,1690
"alter GPUs next Over the of landscape statistics. will fundamentally decade, the computational","Over the next decade, GPUs will fundamentally alter the landscape of computational statistics.",1691
statisticians to is more for It time on-board. get,It is time for more statisticians to get on-board.,1692
We using steps the two multivariate slice sampling crumb taking for methods framework. describe,We describe two slice sampling methods for taking multivariate steps using the crumb framework.,1693
target distributions favorably The highly-correlated adaptive perform on with parameters. methods low-dimensional,The adaptive methods perform favorably on low-dimensional target distributions with highly-correlated parameters.,1694
to feature due is possible calculus speed matrix for short-cuts leave-one-out addition. This based and,This speed is possible due to matrix calculus based short-cuts for leave-one-out and feature addition.,1695
"dam buildings, roads, bridges, construction of etc.","construction of buildings, roads, bridges, dam etc.",1696
and accurate satisfactory based result method. classical shows with the Classifier in ANN comparison on,Classifier based on ANN shows satisfactory and accurate result in comparison with the classical method.,1697
a applications. task in computer is vision detection various point Interest common,Interest point detection is a common task in various computer vision applications.,1698
paper to approach a proposes interest point system-theoretic Current detection.,Current paper proposes a system-theoretic approach to interest point detection.,1699
possible it Furthermore detector irredundant to equivalent is each modification. for its construct and,Furthermore for each detector it is possible to construct its irredundant and equivalent modification.,1700
lower possesses detector complexity computational and Modified preferable. is,Modified detector possesses lower computational complexity and is preferable.,1701
between compression a plates. Nanoindentation parallel by as large is tip modeled,Nanoindentation by a large tip is modeled as compression between parallel plates.,1702
process or background the individual textual pixels text classifies images. binarisation as Text in,Text binarisation process classifies individual pixels as text or background in the textual images.,1703
to between recognition necessary OCR. is by Binarization localization gap and bridge the,Binarization is necessary to bridge the gap between localization and recognition by OCR.,1704
images window method presents background. Sliding to with from binarise paper This text textual textured,This paper presents Sliding window method to binarise text from textual images with textured background.,1705
iterative are by detected thresholding. Then Edges,Then Edges are detected by iterative thresholding.,1706
population dynamics approximately the this evidence as the is ocean scale-invariant. interpret that We in,We interpret this as evidence that the population dynamics in the ocean is approximately scale-invariant.,1707
model. population of dynamical construction use the a solution and this in We invariance size-structured,We use this invariance in the construction and solution of a size-structured dynamical population model.,1708
results are that provide prone. fragments a of Our aggregation for basis wide genome search,Our results provide a basis for genome wide search of fragments that are aggregation prone.,1709
"the As we VOI selection problem. estimation study, a in measurement case examine","As a case study, we examine VOI estimation in the measurement selection problem.",1710
Then image membership the net on pixel. the value of binarized is depending each,Then the image is binarized depending on the net membership value of each pixel.,1711
mapping metadata versions We annotated content-based between for an show video. application of,We show an application for content-based metadata mapping between versions of annotated video.,1712
detail. of players been strategies has two in dynamics two studied Evolutionary great with game,Evolutionary game dynamics of two players with two strategies has been studied in great detail.,1713
"one-shot we with Here, games players. address multiple","Here, we address one-shot games with multiple players.",1714
dynamical games complexity interactions. that be pairwise a Multiplayer captured great cannot on show based,Multiplayer games show a great dynamical complexity that cannot be captured based on pairwise interactions.,1715
"hold specific any easily for game results and can applied Our for e.g. be cases,","Our results hold for any game and can easily be applied for specific cases, e.g.",1716
stag or multiplayer public hunts. goods games,public goods games or multiplayer stag hunts.,1717
were different samples free-flow data text users. and collected isolated from containing Handwritten,Handwritten data samples containing isolated and free-flow text were collected from different users.,1718
user. handwritten recognizes and isolated language-set from free-flow such collected samples test the Each designated,Each such language-set recognizes isolated and free-flow handwritten test samples collected from the designated user.,1719
Light integral is ITMS. an part of the Red Violation System Detection An (RLVDS) automated,An automated Red Light Violation Detection System (RLVDS) is an integral part of the ITMS.,1720
recognition is on studied. influence The accidental blade also of equality,The influence of accidental blade equality on recognition is also studied.,1721
"Finally, the is that compared models. model GA efficiency the developed of of previously to","Finally, the efficiency of the GA model is compared to that of previously developed models.",1722
How fitness population to does a evolve fast peak from another? one,How fast does a population evolve from one fitness peak to another?,1723
be fixed order mutations can or The of random. either,The order of mutations can either be fixed or random.,1724
weighted and using face canonical are mean vector Reduced fusion vector eigenface rule. together fused,Reduced eigenface vector and canonical face vector are fused together using weighted mean fusion rule.,1725
consisting evaluated database with The multiview of has face been faces. system UMIST,The system has been evaluated with UMIST face database consisting of multiview faces.,1726
different containing basic characters were Bangla isolated data and Handwritten users. digits samples from collected,Handwritten data samples containing isolated Bangla basic characters and digits were collected from different users.,1727
users. language-set Bangla handwritten collected the from isolated such designated Each basic test samples recognizes,Each such language-set recognizes isolated basic Bangla handwritten test samples collected from the designated users.,1728
fundamental integrals problem of Bayesian asymptotic evaluation is a likelihood in statistics. accurate The marginal,The accurate asymptotic evaluation of marginal likelihood integrals is a fundamental problem in Bayesian statistics.,1729
are on singularities. of based resolution Our results,Our results are based on resolution of singularities.,1730
Such models. models finite state include discrete,Such models include finite state discrete models.,1731
A new Gaussian for in models model determination is methodology developed. decomposable graphical,A new methodology for model determination in decomposable graphical Gaussian models is developed.,1732
"to hyper-parameters Typically, likelihood fixed maximum are the their estimations.","Typically, the hyper-parameters are fixed to their maximum likelihood estimations.",1733
scheme the datasets. simulated of real We on illustrate efficiency and this new,We illustrate the efficiency of this new scheme on simulated and real datasets.,1734
multiple users. text isolated Handwritten collected samples and free-flow from were,Handwritten isolated and free-flow text samples were collected from multiple users.,1735
samples of recognize the categories handwriting both pages. trained user-specific to Tesseract of document is,Tesseract is trained to recognize user-specific handwriting samples of both the categories of document pages.,1736
"eliminated At intensity the background on at based a first, variance. level is coarse","At first, the background is eliminated at a coarse level based on intensity variance.",1737
distinct makes This each other. foreground from components the,This makes the foreground components distinct from each other.,1738
and removed using components text characteristic graphics. Then various features of the non-text are,Then the non-text components are removed using various characteristic features of text and graphics.,1739
"further text the Finally, are skew regions and binarized corrected for processing.","Finally, the text regions are skew corrected and binarized for further processing.",1740
"the context, this maximin ones. the are and criteria two considered: dispersion usually In minimax","In this context, two dispersion criteria are usually considered: the minimax and the maximin ones.",1741
"the Moreover, thanks black-box is approximated design interpolation. typically function to kernel whatever is, the","Moreover, whatever the design is, the black-box function is typically approximated thanks to kernel interpolation.",1742
"algorithms simulated we Then, domain. maximin determine to in propose annealing bounded any designs connected","Then, we propose simulated annealing algorithms to determine maximin designs in any bounded connected domain.",1743
of schemes. convergence the We prove the different,We prove the convergence of the different schemes.,1744
paper. in Such introduced this is modification,Such modification is introduced in this paper.,1745
This and study of of are a conclusions series experiments. through this other validated numerical,This and other conclusions of this study are validated through a series of numerical experiments.,1746
be factor will results. the deciding for producing This,This will be the deciding factor for producing results.,1747
noise. against segmentation efficient Medical robust segmentation demands an and image algorithm,Medical image segmentation demands an efficient and robust segmentation algorithm against noise.,1748
spatial algorithm proposed clustering uses fuzzy calculate membership value. to information The,The proposed clustering algorithm uses fuzzy spatial information to calculate membership value.,1749
is The proposed using image ISFCM input algorithm. clustered,The input image is clustered using proposed ISFCM algorithm.,1750
and FCM the has A been comparative made between proposed study ISFCM. conventional,A comparative study has been made between the conventional FCM and proposed ISFCM.,1751
found be to conventional approach is FCM. proposed The the outperforming,The proposed approach is found to be outperforming the conventional FCM.,1752
Our current of models. understanding computational motion based number this is on of a detailed,Our current understanding of this motion is based on a number of detailed computational models.,1753
we present to Here models. computational mechanisms in distinct identify and the two analyze seek,Here we seek to identify and analyze two distinct mechanisms present in the computational models.,1754
transform; Descriptor; . recognition; words: Key Fourier signature verification Wavelet personal,Key words: signature recognition; Fourier Descriptor; Wavelet transform; personal verification .,1755
"of sequences However, (e.g. often are length as fixed","However, as sequences are often of fixed length (e.g.",1756
segmentation vital processing. Image of part image a is,Image segmentation is a vital part of image processing.,1757
images can The be medical segmented same manually.,The same medical images can be segmented manually.,1758
Medical step is image for most an analysis essential consequent tasks. segmentation image,Medical image segmentation is an essential step for most consequent image analysis tasks.,1759
Clustering algorithm. generalized approach standard This (FCM) version a CMeans of is Fuzzy,This approach is a generalized version of standard Fuzzy CMeans Clustering (FCM) algorithm.,1760
The is the standard of technique FCM eliminated technique. modifying in conventional limitation the,The limitation of the conventional FCM technique is eliminated in modifying the standard technique.,1761
is proposed The the ORL with and databases. IITK the evaluated algorithm face,The proposed algorithm is evaluated with the ORL and the IITK face databases.,1762
present results We these empirical theoretical that then corroborate bounds.,We then present empirical results that corroborate these theoretical bounds.,1763
and FIS reasoning. fuzzy logic model decision-making approximate a Each combining is,Each FIS is a decision-making model combining fuzzy logic and approximate reasoning.,1764
logic function utilizes membership fuzzy of ANFIS of rules and (MF). adjustment identification The automatic,The ANFIS utilizes automatic identification of fuzzy logic rules and adjustment of membership function (MF).,1765
evaluated FIS on comparisons. two are settings based,FIS settings are evaluated based on two comparisons.,1766
between for The ANFIS FIS comparison model classifying second events. settings structured the is and,The second comparison is the model settings between FIS and ANFIS for classifying structured events.,1767
some results spreading. us This to general about epidemic obtain allows,This allows us to obtain some general results about epidemic spreading.,1768
"cluster, an graph on each Then drawn. isomorphic is","Then on each cluster, an isomorphic graph is drawn.",1769
temporally distributions. firing Leader/follower delay times stable positively form skewed,Leader/follower firing delay times form temporally stable positively skewed distributions.,1770
with inhibitory variance. results usually Blocking in times delay reduced synapses shorter,Blocking inhibitory synapses usually results in shorter delay times with reduced variance.,1771
"robust higher cycle rates, At essential rhythms. becomes growth transcription-translation a circadian for generating","At higher growth rates, a transcription-translation cycle becomes essential for generating robust circadian rhythms.",1772
the We constraints. propagation two AllDifferent of study for algorithms conjunction,We study propagation algorithms for the conjunction of two AllDifferent constraints.,1773
"finding matchings. of simultaneous investigate bipartite problem Therefore, the we","Therefore, we investigate the problem of finding simultaneous bipartite matchings.",1774
"finding is Unfortunately, such NP-hard general. in matchings","Unfortunately, finding such matchings is NP-hard in general.",1775
that experiments significant benefits over can new existing propagator offer this Our show methods.,Our experiments show that this new propagator can offer significant benefits over existing methods.,1776
the an define We concept of symmetry. internal,We define the concept of an internal symmetry.,1777
within satisfaction This a of solution a constraint problem. a is symmety,This is a symmety within a solution of a constraint satisfaction problem.,1778
detail discuss this more this in We in paper.,We discuss this in more detail in this paper.,1779
region It and the means life from microscopic. between emerges macroscopic a that,It means that the life emerges from a region between macroscopic and microscopic.,1780
of with evolutionary finding the the An complexity sum genome. the important rule is correlated,An important finding is the sum rule correlated with the evolutionary complexity of the genome.,1781
its Applications Web-based tendering summarised in processes is also terms properties. of of,Applications of Web-based tendering processes is also summarised in terms of its properties.,1782
"implementation on in of prior details. processes approaches are in discussed tendering researches DSS Furthermore,","Furthermore, prior researches on implementation of DSS approaches in tendering processes are discussed in details.",1783
arise processes outlined. both from and Web-based of tendering Current issues paper-based are,Current issues arise from both of paper-based and Web-based tendering processes are outlined.,1784
"end of paper. Finally, this is conclusion the included at","Finally, conclusion is included at the end of this paper.",1785
"shown. have cell-to-cell transfers been Recently, P-gp","Recently, cell-to-cell P-gp transfers have been shown.",1786
of P-gp. transfer We functional confirmed cell-to-cell,We confirmed cell-to-cell transfer of functional P-gp.,1787
imposes the prior The across image constraints patches. overlapping continuity,The image prior imposes continuity constraints across the overlapping patches.,1788
"an filter. cross error back-projection We bilateral approximate an apply with technique,","We apply an error back-projection technique, with an approximate cross bilateral filter.",1789
results well both based run-time quality as for present We comparative measurements. image visual as,We present comparative results for both run-time as well as visual image quality based measurements.,1790
signature region of for interest was pre-processing new presented. approach A,A new approach for signature region of interest pre-processing was presented.,1791
this bio-secure A paper further stage based an contribution automated cropping is systems. in of,A further contribution of this paper is an automated cropping stage in bio-secure based systems.,1792
can more These be generate self combined to styles. styles,These styles can be self combined to generate more styles.,1793
the the motivates problem. for Genetic use This of Algorithms,This motivates the use of Genetic Algorithms for the problem.,1794
"order of prove to a In characters. pool of images this, we made","In order to prove this, we made a pool of images of characters.",1795
wavelet techniques packet and has discrete image as compression. wavelet Recently popular for emerged transform,Recently discrete wavelet transform and wavelet packet has emerged as popular techniques for image compression.,1796
major transform one the of compression. is The of components image wavelet processing,The wavelet transform is one of the major processing components of image compression.,1797
"which significant information Our based selects include threshold the sub-bands, algorithm entropy. on","Our algorithm selects the sub-bands, which include significant information based on threshold entropy.",1798
results better provides RLE. encoding length than suggested The is technique run enhanced,The enhanced run length encoding technique is suggested provides better results than RLE.,1799
"space and diversity, emerge in They time, high display disappear. and differentiate","They display high diversity, differentiate in space and time, emerge and disappear.",1800
"described of ignoring can other communicative using traits. methods However, be patterns mathematical organization large-scale","However, other patterns of large-scale organization can be described using mathematical methods ignoring communicative traits.",1801
We present statements. a formal set to patterns inference of statements causal from explanation,We present a set of formal inference patterns from causal statements to explanation statements.,1802
algorithms. to display improvement quality linear the reconstruction Numerical when compared in reconstruction experiments,Numerical experiments display the improvement in reconstruction quality when compared to linear reconstruction algorithms.,1803
A of is parameters. with can that these produce set needed a shapes small model,A model is needed that can produce these shapes with a small set of parameters.,1804
leaf growth of the veins. of by position The governed is lobes leaf,The growth of leaf lobes is governed by the position of leaf veins.,1805
a of This shapes. higher to an plant model leaf variety gives approximation wide,This model gives an approximation to a wide variety of higher plant leaf shapes.,1806
complex can more extended model The be types. to leaf,The model can be extended to more complex leaf types.,1807
techniques polymer for open elastic modelling We the appropriate geometry of ended the discuss molecules.,We discuss the appropriate techniques for modelling the geometry of open ended elastic polymer molecules.,1808
surface. to molecule fixed The boundary endpoints assumed a on have is,The molecule is assumed to have fixed endpoints on a boundary surface.,1809
curvature fluctuation landscape. the These domains are fitness determined of the by,These fluctuation domains are determined by the curvature of the fitness landscape.,1810
"peaks, dissipation. such curvature, Regions experience negative adaptive as with","Regions with negative curvature, such as adaptive peaks, experience dissipation.",1811
image A object logical is proposed. to approach on recognition,A logical approach to object recognition on image is proposed.,1812
biochemical and Its of chemoreceptors long been has based demethylation methylation known. mechanism on,Its biochemical mechanism based on methylation and demethylation of chemoreceptors has long been known.,1813
additions. large significant imprecision analysis adaptation Our data in for suggests,Our data analysis suggests significant imprecision in adaptation for large additions.,1814
large-scale coherence some brain produces waves states. to standing in observed The string EEG analogous,The string produces standing waves analogous to large-scale coherence EEG observed in some brain states.,1815
smaller to (mesoscopic) dynamics. springs the scale are The analogous columnar attached,The attached springs are analogous to the smaller (mesoscopic) scale columnar dynamics.,1816
a the at additionally diversity as maturation. of The model function predicts distribution full size,The full model additionally predicts the diversity distribution as a function of size at maturation.,1817
that We the accounts derive of also the results. a reliability new criterion for selection,We also derive a new selection criterion that accounts for the reliability of the results.,1818
genomic hybridisation illustrate data a and our profile. simulated We comparative methodology on both,We illustrate our methodology on both simulated data and a comparative genomic hybridisation profile.,1819
"issues of to data considered. are watermark Capacity, be quality important Robustness, & very Perceptual","Capacity, Robustness, & Perceptual quality of watermark data are very important issues to be considered.",1820
attempts Blurred of Gaussian Restored paper This the Images. study undertake to,This paper attempts to undertake the study of Restored Gaussian Blurred Images.,1821
To restored choose techniques. guesses deblurring of image this for base or the,To choose the base guesses for restored or deblurring image of this techniques.,1822
four related. these models are they review investigate how We and,We review these four models and investigate how they are related.,1823
"required. the the reticulation model minimum In of events general, influences number chosen","In general, the model chosen influences the minimum number of reticulation events required.",1824
than selective constraints of proteins rather The characteristic species. are,The selective constraints are characteristic of proteins rather than species.,1825
passes pathogen model distinct $n$ the immune a which immunologically through We surveillance of stages.,We model the immune surveillance of a pathogen which passes through $n$ immunologically distinct stages.,1826
corresponds the fixed point. unique asymptotically system's This to stable,This corresponds to the system's unique asymptotically stable fixed point.,1827
several we network resources. kinase-substrate online this mammalian For reconstructed study a literature-based from,For this study we reconstructed a literature-based mammalian kinase-substrate network from several online resources.,1828
amplitudes. of periods large effect LFP enhanced is in This,This effect is enhanced in periods of large LFP amplitudes.,1829
"computationally model inference, method parameter for and and comparison. efficient We propose a joint model","We propose a computationally efficient method for joint parameter and model inference, and model comparison.",1830
"Experimentally, or well than with equally complexity. better LiNGAM computational comparable SLIM performs","Experimentally, SLIM performs equally well or better than LiNGAM with comparable computational complexity.",1831
the basic propose extensions two to i.i.d. We,We propose two extensions to the basic i.i.d.,1832
extraction ad-hoc on has information web. like important the applications user-driven This,This has important applications like user-driven ad-hoc information extraction on the web.,1833
ranking linear  introduce vector for the support an training We efficient machine.  method,  We introduce an efficient method for training the linear ranking support vector machine.,1834
heart rate analysis. been structure of Factor detected with has periodogram factor,Factor structure of heart rate periodogram has been detected with factor analysis.,1835
there at least The are of that periodical HRV. results phenomena four showed,The results showed that there are at least four periodical phenomena of HRV.,1836
of been discovered not and explained physiologically Two them yet. have,Two of them have not been discovered and physiologically explained yet.,1837
the rages frequency of Despite differences of factors frequencies peak the the overlapped. are of,Despite of differences of the peak frequencies the frequency rages of the factors are overlapped.,1838
"separations proteomics Electrophoretic use well-established robust, protocols. for","Electrophoretic separations for proteomics use robust, well-established protocols.",1839
this mail from study objective recognize postal numerical of codes is to addresses. printed The,The objective of this study is to recognize printed numerical postal codes from mail addresses.,1840
similar understand it Thus is by important and of adhesion cells molecules. antibodies to,Thus it is important to understand adhesion of cells by antibodies and similar molecules.,1841
parameter occurs. in quantitatively We the binding also describe space which,We also quantitatively describe the parameter space in which binding occurs.,1842
dynamics study a research technique to A the task is of of suggested. solving,A technique to study the dynamics of solving of a research task is suggested.,1843
small The been to volunteer a has pilot group participants. technique applied of,The technique has been applied to a small pilot group of volunteer participants.,1844
not of successfully (non-solvers). have the and (solvers) solved them some have Some task,Some of them have successfully solved the task (solvers) and some have not (non-solvers).,1845
"It tracking object for arises, example, from motion. in and structure","It arises, for example, in object tracking and structure from motion.",1846
based decision. making has A been for Euclidean used the Distance matching technique,A Euclidean Distance based matching technique has been used for making the decision.,1847
human material snow and and river generate consequences. with damages torrents Mountain avalanches dramatic,Mountain river torrents and snow avalanches generate human and material damages with dramatic consequences.,1848
experiments. new bound the a present ChIP-seq We algorithm identification of from for regions,We present a new algorithm for the identification of bound regions from ChIP-seq experiments.,1849
a particle that new This uses MCMC method for SMC kernels dynamics. paper introduces adaptive,This paper introduces a new SMC method that uses adaptive MCMC kernels for particle dynamics.,1850
outperformed between with kernels al. a MCMC adaptive algorithm choice Haario of et the ASMC,ASMC with a choice between kernels outperformed the adaptive MCMC algorithm of Haario et al.,1851
such After projection the analysis a developed. collection alternate theoretical an find algorithm to is,After the theoretical analysis an alternate projection algorithm to find such a collection is developed.,1852
results These sampling approximate new and linear regularized rely stronger programs. on bounds for,These results rely on new and stronger sampling bounds for regularized approximate linear programs.,1853
propose homotopy also computationally efficient method. a We,We also propose a computationally efficient homotopy method.,1854
students were and hypotheses art creative writing students. tested using These,These hypotheses were tested using art students and creative writing students.,1855
findings that support the is hypothesis style creative recognizable. These,These findings support the hypothesis that creative style is recognizable.,1856
We heuristics different selection. for page analyze the of are the influence responsible that,We analyze the influence of different heuristics that are responsible for the page selection.,1857
results present a the allow we heuristics. measurement Alongside comparison to of method these our,Alongside these results we present a measurement method to allow the comparison of our heuristics.,1858
introduce Principal of Component (PCA) notion We Analysis the of gradient image orientations.,We introduce the notion of Principal Component Analysis (PCA) of image gradient orientations.,1859
of on robust favorable its subspace We demonstrate properties some estimation.,We demonstrate some of its favorable properties on robust subspace estimation.,1860
"of inequalities are also described. to reversibility corresponding realization the linear CRN the Moreover, full","Moreover, the linear inequalities corresponding to full reversibility of the CRN realization are also described.",1861
examples. The illustrated results are theoretical meaningful on,The theoretical results are illustrated on meaningful examples.,1862
An that is interesting of protocols. of set scoring elections,An interesting set of elections is that of scoring protocols.,1863
Our result main problems on complexity. is their to based families of partition the these,Our main result is to partition the problems of these families based on their complexity.,1864
are to describe Maps places far-off used .,Maps are used to describe far-off places .,1865
strategies. navigation and military is for aid an It,It is an aid for navigation and military strategies.,1866
mapping (i). important and based on lands is of the the work are Mapping,Mapping of the lands are important and the mapping work is based on (i).,1867
be assumed Pixels homogeneous independently. analyzed to within automatically are areas,Pixels within areas assumed to be automatically homogeneous are analyzed independently.,1868
"success. shown detectors tremendous offline object Recently, a have","Recently, offline object detectors have shown a tremendous success.",1869
"of arriving In can learned, not an data. use addition, once offline make detector newly","In addition, once learned, an offline detector can not make use of newly arriving data.",1870
trained Our detection benefit in manner. an online significantly results tasks object that when confirm,Our results confirm that object detection tasks benefit significantly when trained in an online manner.,1871
input A problem complex feature pattern involves face dimension of recognition vector. like recognition huge,A complex pattern recognition problem like face recognition involves huge dimension of input feature vector.,1872
performance. shown method good has very This,This method has shown very good performance.,1873
"autosomes, fully regime predicted. dominant one is only For","For fully dominant autosomes, only one regime is predicted.",1874
of affect decisions finished significantly. solver These performance the the,These decisions affect the performance of the finished solver significantly.,1875
Most forecasting methods are of accidents. series the of time fuzzy presented car road for,Most of the fuzzy time series methods are presented for forecasting of car road accidents.,1876
"methods existing of However, accuracy are good enough. the the not rates forecasting","However, the forecasting accuracy rates of the existing methods are not good enough.",1877
methods. and The belongs method proposed the order time-variant to kth,The proposed method belongs to the kth order and time-variant methods.,1878
Lateral developmental patterns system Notch-Delta contexts. the occur inhibition mediated diverse by signaling in,Lateral inhibition patterns mediated by the Notch-Delta signaling system occur in diverse developmental contexts.,1879
provide explanation for These simple development. formation results pattern plausible a lateral inhibition during and,These results provide a simple and plausible explanation for lateral inhibition pattern formation during development.,1880
physics originally of in The conserved statistical quantities. dissipation-fluctuation the seals theorem these fate,The dissipation-fluctuation theorem in statistical physics seals the fate of these originally conserved quantities.,1881
Segmentation processing. position images area holds of important the in of an image,Segmentation of images holds an important position in the area of image processing.,1882
challenges handle the recognition. face This a presents approach to of paper novel,This paper presents a novel approach to handle the challenges of face recognition.,1883
finally projected eigenspace perceptron. classified Polar are multi-layer images using into and a,Polar images are projected into eigenspace and finally classified using a multi-layer perceptron.,1884
images segmentation to choose image. satellite guesses three techniques for threshold These base on applied,These techniques applied on three satellite images to choose base guesses for threshold segmentation image.,1885
ChTX has been for time. localization The of first revealed orbitals of the frontier the,The localization of the frontier orbitals of ChTX has been revealed for the first time.,1886
the is of in key tasks Object computer vision. one detection,Object detection is one of the key tasks in computer vision.,1887
The standard. of framework the Viola become Jones de has and cascade facto,The cascade framework of Viola and Jones has become the de facto standard.,1888
algorithm a provide boosting this in work. We such,We provide such a boosting algorithm in this work.,1889
classifier et is Wu It of (LAC) linear the al. by inspired asymmetric,It is inspired by the linear asymmetric classifier (LAC) of Wu et al.,1890
function. cost in a our similar boosting that optimizes algorithm,in that our boosting algorithm optimizes a similar cost function.,1891
implemented boosting The algorithm convex in new totally-corrective is column generation the technique by optimization.,The new totally-corrective boosting algorithm is implemented by the column generation technique in convex optimization.,1892
"level We RGB(Red,Green,Blue) it because takes raw to it easy select is values use. and","We select RGB(Red,Green,Blue) because it takes raw level values and it is easy to use.",1893
"R,G,and range of this put paper we In color ratio each will for B.","In this paper we will put range ratio color for each of R,G,and B.",1894
condition. the number we pixels of Then our achieved count the that,Then we count the number of the pixels that achieved our condition.,1895
analysis some algorithm. We the of provide also convergence,We also provide some convergence analysis of the algorithm.,1896
"Finally, the is datasets. some with simulation method illustrated","Finally, the method is illustrated with some simulation datasets.",1897
offers of Turing formalization test. paper mathematical a the The,The paper offers a mathematical formalization of the Turing test.,1898
of many constraint Symmetry important feature an programs. is,Symmetry is an important feature of many constraint programs.,1899
each symmetries out pick class. different in symmetry Different solutions,Different symmetries pick out different solutions in each symmetry class.,1900
in idea used but This a simple of powerful can ways. be number different,This simple but powerful idea can be used in a number of different ways.,1901
These evidence metadata how knowledge. provide of community structured for types valuable organizes a learning,These types of structured metadata provide valuable evidence for learning how a community organizes knowledge.,1902
"ontologies data appears Structured in and many entities, taxonomies and semi-structured domains. describing","Structured and semi-structured data describing entities, taxonomies and ontologies appears in many domains.",1903
constraints. problem with structural optimization task an We as frame this,We frame this task as an optimization problem with structural constraints.,1904
the these We of samplers. consider also justification theoretical,We also consider the theoretical justification of these samplers.,1905
of the MI are image. ZM at good an shape representing and features,MI and ZM are good at representing the shape features of an image.,1906
"restrict and their application ZM non-orthogonality poor of CBIR. in MI reconstruction However, of","However, non-orthogonality of MI and poor reconstruction of ZM restrict their application in CBIR.",1907
"and needed. based an is efficient system Therefore, moment orthogonal CBIR","Therefore, an efficient and orthogonal moment based CBIR system is needed.",1908
"(LM) image Moments faster, features are computationally shape can compactly. and orthogonal, Legendre represent","Legendre Moments (LM) are orthogonal, computationally faster, and can represent image shape features compactly.",1909
"Further, improved classifier. is the Support Machine employing classification by efficiency (SVM) Vector","Further, the classification efficiency is improved by employing Support Vector Machine (SVM) classifier.",1910
"examined there needed to intrusion various remain However, (IDS). detection towards be systems issues current","However, there remain various issues needed to be examined towards current intrusion detection systems (IDS).",1911
(a) representing structure preferences. (b) Our learning and contributions are,Our contributions are (a) learning structure and (b) representing preferences.,1912
dispelled words cases this have controversy. not and of Court mountains,Court cases and mountains of words have not dispelled this controversy.,1913
develop The to employed physiological authentication trait behavioral and biometric systems. is,The physiological and behavioral trait is employed to develop biometric authentication systems.,1914
the of based connected preprocessed iris on area patterns The components. are,The iris patterns are preprocessed based on area of the connected components.,1915
into quadtree region image is split components. The,The image region is split into quadtree components.,1916
components The from training determined the variance with are minimum samples.,The components with minimum variance are determined from the training samples.,1917
the moments on are Hu components. applied,Hu moments are applied on the components.,1918
"a a signature where a expectation. case, in is is genuine The biometric, variations natural","The signature is a biometric, where variations in a genuine case, is a natural expectation.",1919
"vary of the one instance another. to In signature signature, certain from genuine parts","In the genuine signature, certain parts of signature vary from one instance to another.",1920
breakthrough as System System to achieve Information combining Game is performance. and game Information organizations',Game Information System as combining Information System and game is breakthrough to achieve organizations' performance.,1921
the problem for thus better guide into programs. heuristic a turn inference search We,We thus turn the inference problem into a heuristic search for better guide programs.,1922
global Using will access to enhance easy technology Web and use.,Using Web technology will enhance global access and easy to use.,1923
"which collective they graph, a of a triples meaning. give to set constitutes A","A set of triples constitutes a graph, to which they give a collective meaning.",1924
of stark contrast nature thoroughly with This context-sensitive the human thought. is in,This is in stark contrast with the thoroughly context-sensitive nature of human thought.,1925
states that the MDPs assume maker decision knows actions. all and (DM) Traditional,Traditional MDPs assume that the decision maker (DM) knows all states and actions.,1926
"not many this in may of interest. situations However, be true","However, this may not be true in many situations of interest.",1927
"particular, near-optimal solution In polynomial can when time. in we be found characterize a","In particular, we characterize when a near-optimal solution can be found in polynomial time.",1928
In or periodic ruled out. heteroclinic of existence is the cases some cycles solutions,In some cases the existence of periodic solutions or heteroclinic cycles is ruled out.,1929
introns. depends that of assume the length the exonization process We on,We assume that the exonization process depends on the length of introns.,1930
"Vertebrates, characterized long short unlike internal by are introns exons. and invertebrates,","Vertebrates, unlike invertebrates, are characterized by long introns and short internal exons.",1931
classifying self-organizing neural used for map The also network the is fingerprints.,The self-organizing map neural network is also used for classifying the fingerprints.,1932
period the of time to generation dominant corresponds these oscillation fish. the of The,The period of the oscillation corresponds to the dominant generation time of these fish.,1933
paper. A for this in introduced problems framework is general image solving inverse,A general framework for solving image inverse problems is introduced in this paper.,1934
"based Gaussian efficient approach a algorithm. on mixture models, via MAP-EM is computationally The estimated","The approach is based on Gaussian mixture models, estimated via a computationally efficient MAP-EM algorithm.",1935
also motivated This effective suggests for interpretation dictionary algorithm. an the MAP-EM initialization,This interpretation also suggests an effective dictionary motivated initialization for the MAP-EM algorithm.,1936
notoriously difficult for dimensional high Density modeling is data.,Density modeling is notoriously difficult for high dimensional data.,1937
as in observed density the We space. can interpreted a how model be it show,We show how it can be interpreted as a density model in the observed space.,1938
"our to are distributions, which We multimodal. highly posterior methodology apply mixture","We apply our methodology to mixture posterior distributions, which are highly multimodal.",1939
Gaussian approach We and univariate illustrate real-world with mixtures and bivariate datasets. our two,We illustrate our approach with univariate and bivariate Gaussian mixtures and two real-world datasets.,1940
between revealed transposed elements. other density SNP and differences Alu analysis,SNP density analysis revealed differences between Alu and other transposed elements.,1941
"difficult is data structure, how handle clustering problem a to A i.e. manifold with in","A difficult problem in clustering is how to handle data with a manifold structure, i.e.",1942
algorithms. can most used with The combination new be metric clustering in,The new metric can be used in combination with most clustering algorithms.,1943
schemes several sub-graphs. connecting different the for We discuss possible,We discuss several possible schemes for connecting the different sub-graphs.,1944
clustering promising shows cases In metric PKNNG all results. the,In all cases the PKNNG metric shows promising clustering results.,1945
optimal and and solving both bilinear describe algorithms We approximate programs. for analyze,We describe and analyze both optimal and approximate algorithms for solving bilinear programs.,1946
a of convergent iteration. algorithm offers The generalization analysis approximate shows this that policy,The analysis shows that this algorithm offers a convergent generalization of approximate policy iteration.,1947
bilinear algorithms programming of under analyze the briefly behavior incomplete also We samples.,We also briefly analyze the behavior of bilinear programming algorithms under incomplete samples.,1948
The researchers several and of maintenance decades. cooperation evolution for fascinated,The evolution and maintenance of cooperation fascinated researchers for several decades.,1949
evolves how can the participation. solved puzzle voluntary punishment costly under be behaviour The,The puzzle how the costly punishment behaviour evolves can be solved under voluntary participation.,1950
the is emerge compulsory? punishers Could participation if,Could the punishers emerge if participation is compulsory?,1951
punishment a behaviour selfish the altruistic inevitably an Is or behaviour?,Is the punishment inevitably a selfish behaviour or an altruistic behaviour?,1952
still The punishment enigma. motivations behind an are,The motivations behind punishment are still an enigma.,1953
is portion other contributors cost. when paying to distributed The a second,The other portion is distributed to contributors when paying a second cost.,1954
This surviving indicates more and more. gaining rule: a contributing universal,This indicates a universal surviving rule: contributing more and gaining more.,1955
"is machine-understandable. transformed words, current Web being machine-readable from In to other the","In other words, the current Web is transformed from being machine-readable to machine-understandable.",1956
"expressiveness. of but with XML on syntax, terminologies these various and based languages are Most","Most of these languages are based on XML syntax, but with various terminologies and expressiveness.",1957
"for language step. choosing is main Therefore, a an building Ontology the","Therefore, choosing a language for building an Ontology is the main step.",1958
of because knowledge-based This features. is different language different applications need kinds,This is because different kinds of knowledge-based applications need different language features.,1959
this will highlighted. of advantages neural image retrieval classification domain be and model The in,The advantages of this neural model in image classification and retrieval domain will be highlighted.,1960
three accomplishes goal its in steps. The approach main proposed,The proposed approach accomplishes its goal in three main steps.,1961
"However, methods data a image when required applied meaningful these term. segmentation to","However, when applied to image segmentation these methods required a meaningful data term.",1962
systems information. how process on emphasis The understanding is biological,The emphasis is on understanding how biological systems process information.,1963
disease high process is complicated requires of a Human level expertise. and diagnosis,Human disease diagnosis is a complicated process and requires high level of expertise.,1964
can this diagnosis. tool also to Practitioners corroborate use web-based,Practitioners can also use this web-based tool to corroborate diagnosis.,1965
to order experimented evaluate The is various proposed performance. system it's in on scenarios,The proposed system is experimented on various scenarios in order to evaluate it's performance.,1966
"system In the exhibits satisfactory all proposed cases, results.","In all the cases, proposed system exhibits satisfactory results.",1967
"two efficient we technique for measurement an similarity this ontologies. In between paper, propose","In this paper, we propose an efficient technique for similarity measurement between two ontologies.",1968
pairs of omitting all any concepts proposed candidate similar The pair. similar technique without identifies,The proposed technique identifies all candidate pairs of similar concepts without omitting any similar pair.,1969
Programming. based logical are Answer Set on formalization the and process The reasoning,The logical formalization and the reasoning process are based on Answer Set Programming.,1970
has attention model been of in a Finding received graphical much fields. the structure many,Finding the structure of a graphical model has been received much attention in many fields.,1971
"variables. exponentially of grows However, the to computational cost according its number","However, its computational cost grows exponentially according to the number of variables.",1972
illustrate We real-world artificial datasets. with the and algorithm,We illustrate the algorithm with artificial and real-world datasets.,1973
to obtain communities we in Finally the states. this apply analysis end the,Finally we apply this analysis to obtain the communities in the end states.,1974
"been set and studied. of support Notions of soft have and inversion a core, defined","Notions of core, support and inversion of a soft set have been defined and studied.",1975
soft many and has in properties presented. Membership set of a structure probed been interesting,Membership structure of a soft set has been probed in and many interesting properties presented.,1976
of to Hough have classic performance  Many techniques the proposed Transform. speedup been ,  Many techniques have been proposed to speedup the performance of classic Hough Transform.,1977
aims paper components. face extracting image at determining characteristics of the by its a This,This paper aims at determining the characteristics of a face image by extracting its components.,1978
only people with permitted preference Matchings these who appear are in lists.,Matchings are permitted only with people who appear in these preference lists.,1979
"for general demonstrated. paper, In a architectures this have two been Perceptron (MLP) Multilayer","In this paper, two general architectures for a Multilayer Perceptron (MLP) have been demonstrated.",1980
been databases. have the The and using IRIS both OTCBVS of performances methods evaluated,The performances of both the methods have been evaluated using OTCBVS and IRIS databases.,1981
"its can't main be but even. ignored advantage, the disadvantages is Its simplicity","Its simplicity is its main advantage, but the disadvantages can't be ignored even.",1982
and computation memory complexity The matter. also requirement,The memory requirement and computation complexity also matter.,1983
developed Many to limitations. these overcome are techniques,Many techniques are developed to overcome these limitations.,1984
NN techniques broadly structure less are and structure into based classified techniques.,NN techniques are broadly classified into structure less and structure based techniques.,1985
"techniques. of paper, the this present we such survey In","In this paper, we present the survey of such techniques.",1986
computational limitation structure reduce memory less and method complexity. techniques the structure based overcome The,The structure less method overcome memory limitation and structure based techniques reduce the computational complexity.,1987
"into projected are perceptron. classified images Finally, the using eigenspace a multi-layer","Finally, the images projected into eigenspace are classified using a multi-layer perceptron.",1988
database. experiments Spectrum and Beyond Visible Object we used Tracking have In Classification the (OTCBVS),In the experiments we have used Object Tracking and Classification Beyond Visible Spectrum (OTCBVS) database.,1989
efficient an been automatic technique fusion has Here presented. recognition for face,Here an efficient fusion technique for automatic face recognition has been presented.,1990
or proposed in writing system no constraints style The assumes size.,The proposed system assumes no constraints in writing style or size.,1991
both of majority weighted are using Finally combined MLP's scheme. results the,Finally the results of both MLP's are combined using weighted majority scheme.,1992
machine classifier support and fed Neural are for then to in features vector These classification.,These features are then fed to Neural classifier and in support vector machine for classification.,1993
cake propose an cutting form online of We problem. the,We propose an online form of the cake cutting problem.,1994
situations where This dividing arrive resource. the and a depart process of models during players,This models situations where players arrive and depart during the process of dividing a resource.,1995
of constraint. NVALUE global study decompositions We the,We study decompositions of the global NVALUE constraint.,1996
"inequalities. are readily encoded Finally, the decompositions linear as study we","Finally, the decompositions we study are readily encoded as linear inequalities.",1997
able integer programs. linear use to in therefore them are We,We are therefore able to use them in integer linear programs.,1998
driving evolutionary Gene duplications for primary major novelty. of are one forces,Gene duplications are one of major primary driving forces for evolutionary novelty.,1999
"in Next, transformation to face inverse wavelet obtain taken is fused image. discrete order","Next, inverse discrete wavelet transformation is taken in order to obtain fused face image.",2000
was experiments Database For IRIS used. Face Thermal/Visual,For experiments IRIS Thermal/Visual Face Database was used.,2001
solve used many be to can problems. Symmetry help,Symmetry can be used to help solve many problems.,2002
"has problem representation artificial and symmetry reasoning. In role intelligence, both an played important in","In artificial intelligence, symmetry has played an important role in both problem representation and reasoning.",2003
I constraint using work to describe solve symmetry recent satisfaction on problems. help,I describe recent work on using symmetry to help solve constraint satisfaction problems.,2004
"histogram used code features and feature The are shadow chain intersection, features.","The features used are intersection, shadow feature and chain code histogram features.",2005
"character English alphabets off-line images scheme handwritten novel, proposed. for A generic is","A novel, generic scheme for off-line handwritten English alphabets character images is proposed.",2006
a neural recognition networks. is perceptron(MLP) multilayer using scheme The,The recognition scheme is using a multilayer perceptron(MLP) neural networks.,2007
perceptron. are classified and an images into multi-layer using Log-polar finally eigenspace projected improved,Log-polar images are projected into eigenspace and finally classified using an improved multi-layer perceptron.,2008
Cs-DNA frequency spectra prominent. this is the range ion-phosphate vibrations mode at In of,In Cs-DNA spectra at this frequency range the mode of ion-phosphate vibrations is prominent.,2009
much spectra this of Its intensities Na-DNA than of intensity higher range. modes the is,Its intensity is much higher than the intensities of Na-DNA modes of this spectra range.,2010
Cs-DNA lower in have much modes of case intensities than Other of Na-DNA. the,Other modes of Cs-DNA have much lower intensities than in the case of Na-DNA.,2011
eigenspace are using images into perceptron. classified multi-layer projected a finally and Fused,Fused images are projected into eigenspace and finally classified using a multi-layer perceptron.,2012
the produces of Fusion images multiple representation more image. of efficient together a,Fusion of multiple images together produces a more efficient representation of the image.,2013
algorithms verification of considerable for has received performance multiple biometric Fusion attention. improvement,Fusion of multiple algorithms for biometric verification performance improvement has received considerable attention.,2014
is paper recognition handwritten In of characters a this method for devanagari described.,In this paper a method for recognition of handwritten devanagari characters is described.,2015
used training vectors. is feature conjugate-gradient One Perceptron with to Multi-layer classify these,One Multi-layer Perceptron with conjugate-gradient training is used to classify these feature vectors.,2016
Inverse coefficients fused of those face Daubechies images. gives transform wavelet us,Inverse Daubechies wavelet transform of those coefficients gives us fused face images.,2017
Face For was experiments used. Database Thermal/Visual IRIS,For experiments IRIS Thermal/Visual Face Database was used.,2018
with Episode applications. is popular many in Frequent a Mining framework Discovery framework Temporal Data,Frequent Episode Discovery framework is a popular framework in Temporal Data Mining with many applications.,2019
frequency this a In algorithms. of unified such counting we view present paper all,In this paper we present a unified view of all such frequency counting algorithms.,2020
all are current algorithms it. present such cases of generic special that We a algorithm,We present a generic algorithm such that all current algorithms are special cases of it.,2021
aims textual generation new VLSI from at realization This paper of face a for description.,This paper aims at VLSI realization for generation of a new face from textual description.,2022
based presents improving machine algorithm object for on tracking method a learning. paper any This,This paper presents a method for improving any object tracking algorithm based on machine learning.,2023
"triplet i.e. zones, set a zones these of we Using construct","Using these zones, we construct a triplet set of zones i.e.",2024
for recognition. on sets these techniques is Each of different applied characters of,Each of these techniques is applied on different sets of characters for recognition.,2025
MLP's to the characters. separately used recognize Two are,Two MLP's are used separately to recognize the characters.,2026
decision The combined MLP's weighted of using both majority is scheme.,The decision of both MLP's is combined using weighted majority scheme.,2027
"is differences character In stage, these divided relative two. based second into on the set","In the second stage, based on these relative differences character set is divided into two.",2028
Characters using MLP. distinct are of set first shapes of classified,Characters of distinct shapes of first set are classified using MLP.,2029
method. distance second in edit are Confused classified set characters using minimum,Confused characters in second set are classified using minimum edit distance method.,2030
found in parameter the based second on literature. is estimates The,The second is based on parameter estimates found in the literature.,2031
"the analytical First, by following series work an pionneered with power model derived is Fitzhugh.","First, an analytical model is derived with power series following the work pionneered by Fitzhugh.",2032
underlie in of the processes the considerable sex-biased interest There is evolution parasitism. that could,There is considerable interest in the processes that could underlie the evolution of sex-biased parasitism.,2033
with role. system differences may play in differences key along lifespan a Mating,Mating system differences along with differences in lifespan may play a key role.,2034
"seems test use for two-sample Therefore, it Hotelling's such reasonable to data.","Therefore, it seems reasonable to use two-sample Hotelling's test for such data.",2035
"Hotelling's increases Moreover, the well. coefficient when of power increases the correlation test as","Moreover, when the correlation coefficient increases the power of Hotelling's test increases as well.",2036
matching the mouth for used detection. Template eyes of one methods is simplest and,Template matching is one of the simplest methods used for eyes and mouth detection.,2037
"extended powerful modified become be and However, tool. to a it can","However, it can be modified and extended to become a powerful tool.",2038
the is optimum patch using method. proposed The determined shape and size,The optimum patch size and shape is determined using the proposed method.,2039
"methods with combined matching to improve Usually, other template detection order is accuracy. also in","Usually, template matching is also combined with other methods in order to improve detection accuracy.",2040
"effectiveness processing paper, two this in of image methods the Thus, i.e.","Thus, in this paper, the effectiveness of two image processing methods i.e.",2041
"Haar wavelet when analyzed. with transform, used template are matching grayscale and","grayscale and Haar wavelet transform, when used with template matching are analyzed.",2042
blur-free simplicity of and computational fine (BF) filters denoising. deliver a bilateral blend Nonlinear,Nonlinear bilateral filters (BF) deliver a fine blend of computational simplicity and blur-free denoising.,2043
"filter choices parameters. is of optimal noise-suppressing and properties, nature, little their known about However,","However, little is known about their nature, noise-suppressing properties, and optimal choices of filter parameters.",2044
to image our application is illustrate Practical CT results. to discussed denoising,Practical application to CT image denoising is discussed to illustrate our results.,2045
"SMTI the NP-hard. to problem the is SM polynomial solve, is Whilst problem","Whilst the SM problem is polynomial to solve, the SMTI problem is NP-hard.",2046
"to is quantitatively. used characterize Commonly, analysis and motility nematode movements them classify","Commonly, motility analysis is used to classify nematode movements and characterize them quantitatively.",2047
"image environment each requires parameter on tools customized often heuristic tuning. relying However, processing","However, each environment often requires customized image processing tools relying on heuristic parameter tuning.",2048
"for platform enables our of `skeletons' quantification. nematode In addition, extraction straightforward the motility","In addition, our platform enables the extraction of nematode `skeletons' for straightforward motility quantification.",2049
"overwhelming the majority for MEME examined. Overall, the approach cases threshold-based outperforms of","Overall, MEME outperforms the threshold-based approach for the overwhelming majority of cases examined.",2050
"are very However, non-experts. results to difficult interpret technical and the for","However, the results are very technical and difficult to interpret for non-experts.",2051
literature the about In stability. we high-level on a overview give paper this clustering existing,In this paper we give a high-level overview about the existing literature on clustering stability.,2052
"procedure Geographical properly conditioned, as aspects. as data, visualization illustrate well flow the","Geographical flow data, properly conditioned, illustrate the procedure as well as visualization aspects.",2053
at least boundary. birth--death dependent probabilities absorbing We examine state transition processes with and one,We examine birth--death processes with state dependent transition probabilities and at least one absorbing boundary.,2054
system random the equal fitness walk. the If a performs have types two,If the two types have equal fitness the system performs a random walk.,2055
taken it mutants have take long until does advantageous and invaded How over?,How long does it take until advantageous mutants have invaded and taken over?,2056
work Walsh using image transform. a This method registration medical for Hadamard Fast proposes,This work proposes a method for medical image registration using Fast Walsh Hadamard transform.,2057
This of images same algorithm different registers modalities. the or,This algorithm registers images of the same or different modalities.,2058
Fast basis Hadamard Walsh functions. lengthened image terms bit of in Each is,Each image bit is lengthened in terms of Fast Walsh Hadamard basis functions.,2059
consuming Also more is Walsh registration medical in reliable Hadamard less transform time. image Fast,Also Fast Walsh Hadamard transform is more reliable in medical image registration consuming less time.,2060
in a parameters of is RANSAC technique for estimating outliers. the popular presence model,RANSAC is a popular technique for estimating model parameters in the presence of outliers.,2061
estimation. speed matrix our The demonstrated increased of on method with experiments essential,The increased speed of our method demonstrated with experiments on essential matrix estimation.,2062
lowest as revised uniquely degrees [resp. lower] characterized of belief the are,lower] revised degrees of belief are uniquely characterized as the lowest [resp.,2063
certain a by that [resp. invariant highest] max-min are ones,highest] ones that are invariant by a certain max-min [resp.,2064
the by constraints. logical operation determined min-max],min-max] operation determined by the logical constraints.,2065
used Image imagery. algorithms from to cover information extract land satellite are classifiers,Image classifiers are algorithms used to extract land cover information from satellite imagery.,2066
was majority vote of final through individual the a simple determined classifiers. output The,The final output was determined through a simple majority vote of the individual classifiers.,2067
less than of the was It however RBF SVM. that,It was however less than that of the RBF SVM.,2068
over ranging permutations ranking tracking Distributions applications multi-object instances. arise to in from of,Distributions over permutations arise in applications ranging from multi-object tracking to ranking of instances.,2069
the representation hypersphere We time continuous polynomial provide projections between the permutation and $n!$-element space.,We provide polynomial time projections between the continuous hypersphere representation and the $n!$-element permutation space.,2070
article to in this work The suggests the presented a problem. sample solution two,The work presented in this article suggests a solution to the two sample problem.,2071
"phenomenon. Linnik sample similarity, Welch-Aspin the solution, problem, Two nuisance problem, parameter, Fisher-Behrens Keywords:","Keywords: Two sample problem, Welch-Aspin solution, Fisher-Behrens problem, nuisance parameter, similarity, the Linnik phenomenon.",2072
stringent the of model pathway. response dynamics develop We to a theoretical the study,We develop a theoretical model to study the dynamics of the stringent response pathway.,2073
The and foragers focuses model behavioural nectar classes; two nectar receivers. on,The model focuses on two behavioural classes; nectar foragers and nectar receivers.,2074
achieved reduced a be model. can linear Supervised rank multivariate feature extraction fitting by,Supervised linear feature extraction can be achieved by fitting a reduced rank multivariate model.,2075
rank penalized generalized constrained models. paper rank linear and vector This studies,This paper studies rank penalized and rank constrained vector generalized linear models.,2076
A setups. cross-validation for parameter in proposed is novel projective tuning such nonconvex,A novel projective cross-validation is proposed for parameter tuning in such nonconvex setups.,2077
approaches these some All limitation. of have,All of these approaches have some limitation.,2078
We and evolution. of two different discuss genomes' chromosomes' ways,We discuss two different ways of chromosomes' and genomes' evolution.,2079
multi-electrode We arrays observed high-density algorithm identify present spikes (MEAs). individual on an to neural,We present an algorithm to identify individual neural spikes observed on high-density multi-electrode arrays (MEAs).,2080
via interface. Human graphical is a and intervention minimized streamlined,Human intervention is minimized and streamlined via a graphical interface.,2081
as Attitudes (Faculty Toward Statistics). FATS This to referred be scale will,This scale will be referred to as FATS (Faculty Attitudes Toward Statistics).,2082
these Additional to properties. confirm studies psychometric required and structural are,Additional studies are required to confirm these structural and psychometric properties.,2083
"impacts, impacts negative environmental these and and environmental and between positive dependencies receptors. and","and positive and negative environmental impacts, and dependencies between these impacts and environmental receptors.",2084
"work high-level in XML-based, is designed services- oriented to architecture VERSA's environments.","VERSA's high-level architecture is designed to work in XML-based, services- oriented environments.",2085
"diversity. ecosystem high by generated A this phylogenetic rule, nested shows","A nested ecosystem generated by this rule, shows high phylogenetic diversity.",2086
"similar are neighbours. counterparts to not of This degrees species say, the phylogenetic having is","This is to say, the counterparts of species having similar degrees are not phylogenetic neighbours.",2087
and the prediction. factors part analyzes which affect First the identifies,First part analyzes and identifies the factors which affect the prediction.,2088
in detail. been by solved It an has illustrated example,It has been illustrated by an example solved in detail.,2089
simulated been and our making decision in measures Three different discussion. compared have,Three different decision making measures have been simulated and compared in our discussion.,2090
of unweighted rule. the in study elections coalitional Borda We using manipulation the problem,We study the problem of coalitional manipulation in elections using the unweighted Borda rule.,2091
approach logic set the answer semantics An of the is to revision programs presented. under,An approach to the revision of logic programs under the answer set semantics is presented.,2092
"in A that revision, stipulates A. * this AGM is K In","In AGM revision, this stipulates that A is in K * A.",2093
from respects. in update two work differs programs approach logic The main in,The approach differs from work in update logic programs in two main respects.,2094
torsion quantum is as regarded folding polypeptide between on protein a chain. The transition states,The protein folding is regarded as a quantum transition between torsion states on polypeptide chain.,2095
The previous in reviewed. deduction of the is studies rate formula folding our,The deduction of the folding rate formula in our previous studies is reviewed.,2096
generalized variation The formula the folding. in of frequency to is case rate,The rate formula is generalized to the case of frequency variation in folding.,2097
for is mass major characteristics texture. One classification the mammographic of,One of the major mammographic characteristics for mass classification is texture.,2098
factor to mass ANN important into exploits benign or the this classify malignant.,ANN exploits this important factor to classify the mass into benign or malignant.,2099
"individuals in ASD short-range lateral-frontal with electrodes. showed increased connections Conversely,","Conversely, individuals with ASD showed increased short-range connections in lateral-frontal electrodes.",2100
these special Both contain of Gaussian a the classes as case.,Both of these classes contain the Gaussian as a special case.,2101
"that paper biochemical cells can living multiplex In signals, this i.e. we show","In this paper we show that living cells can multiplex biochemical signals, i.e.",2102
these The an heuristics application of in scheme. branching results adaptive,The application of these heuristics results in an adaptive branching scheme.,2103
pathway This determining and to research powerful extinction provides for extinction. the times characterizing tools,This research provides powerful tools for determining extinction times and characterizing the pathway to extinction.,2104
been For heuristics purpose proposed. numerous have this,For this purpose numerous heuristics have been proposed.,2105
on based Results in failures more demonstrate heuristics general efficient. that are,Results demonstrate that heuristics based on failures are in general more efficient.,2106
"in role Homologous recombination genetic key functionality. generating while protein plays a diversity, maintaining","Homologous recombination plays a key role in generating genetic diversity, while maintaining protein functionality.",2107
DNA length sharply. to above sensitivity is There heterology the decreases a which characteristic,There is a characteristic DNA length above which the sensitivity to heterology decreases sharply.,2108
observable neuronal non-stationary This manifested is every in feature. behavior response of practically variability,This variability is manifested in non-stationary behavior of practically every observable neuronal response feature.,2109
authentication and The to develop biometric systems. is physiological employed behavioral trait,The physiological and behavioral trait is employed to develop biometric authentication systems.,2110
connected preprocessed on components. are The area based the patterns of iris,The iris patterns are preprocessed based on area of the connected components.,2111
is image region quadtree components. split The into,The image region is split into quadtree components.,2112
variance training minimum the from samples. determined components with The are,The components with minimum variance are determined from the training samples.,2113
components. on moments applied Hu the are,Hu moments are applied on the components.,2114
"biometric, signature natural a expectation. variations case, The in is is a where a genuine","The signature is a biometric, where variations in a genuine case, is a natural expectation.",2115
"the one In instance from signature, genuine of certain to parts signature another. vary","In the genuine signature, certain parts of signature vary from one instance to another.",2116
is area its challenging research and creativity complexity. for development a Web,Web development is a challenging research area for its creativity and complexity.,2117
"In we and how variability. multiplicity operator affect strength this show operator particular,","In particular, we show how operator strength and operator multiplicity affect this variability.",2118
all network and (i.e. functional interactions modules This contains known,This network contains all known interactions and functional modules (i.e.,2119
entry. proteins) groups to responding coordinately HCMV of,groups of proteins) coordinately responding to HCMV entry.,2120
"further novel are being Nonetheless, in literature. proposed functions","Nonetheless, further novel functions are being proposed in literature.",2121
 We kernel propose for based a  handwritten nearest classification. digit simple approach neighbor,  We propose a simple kernel based nearest neighbor approach for handwritten digit classification.,2122
"two here a The defining similarity between is ""distance"" kernel images. the actually","The ""distance"" here is actually a kernel defining the similarity between two images.",2123
"be concatenating ""normalized compression by estimated zipping the distance"". MI can sequences, and thereby yielding","MI can be estimated by concatenating and zipping sequences, yielding thereby the ""normalized compression distance"".",2124
"but results, uncontrolled this with errors. produced has far promising So","So far this has produced promising results, but with uncontrolled errors.",2125
of from get describe to robust simple pairwise global a MI approach We estimates alignments.,We describe a simple approach to get robust estimates of MI from global pairwise alignments.,2126
virtue see the in way. more of general our a we approach But main,But we see the main virtue of our approach in a more general way.,2127
are Cascade in widely used object classifiers detection. real-time,Cascade classifiers are widely used in real-time object detection.,2128
a that LAC. of cost We directly function optimizes algorithm new boosting design then the,We then design a new boosting algorithm that directly optimizes the cost function of LAC.,2129
in column totally-corrective technique implemented resulting boosting algorithm by is convex The the optimization. generation,The resulting totally-corrective boosting algorithm is implemented by the column generation technique in convex optimization.,2130
a large Motivation: variety Signaling control processes. pathways of cellular,Motivation: Signaling pathways control a large variety of cellular processes.,2131
analyses and makes cross-talk difficult. comparative This,This makes comparative and cross-talk analyses difficult.,2132
the cancer-specific cross-talks profiles. possible and expression activity We and quantified found tissue- pathway-specific of,We quantified the possible tissue- and cancer-specific activity of cross-talks and found pathway-specific expression profiles.,2133
of signaling We for novel pathways. a comparative provide and analyses Conclusions: cross-talk resource,Conclusions: We provide a novel resource for comparative and cross-talk analyses of signaling pathways.,2134
of human-computer a interaction. in focus been Object research detection has,Object detection has been a focus of research in human-computer interaction.,2135
at out. dies simulate what difference level We population the required of,We simulate at what level of required difference the population dies out.,2136
images becoming Searching and tedious the image querying the difficult. related is by,Searching the related images by the querying image is becoming tedious and difficult.,2137
demonstrate results Experimental existing superiority the techniques. with its,Experimental results demonstrate its superiority with the existing techniques.,2138
is aiming causal formalism Our from information. a at causal example capturing explanations,Our example is a formalism aiming at capturing causal explanations from causal information.,2139
efficiency of show relative and naturalness We the job. translation this,We show the naturalness and relative efficiency of this translation job.,2140
an for in interested program. the ASP writing We ease are,We are interested in the ease for writing an ASP program.,2141
translation. show facilitate ASP the We working improvements how recent in systems,We show how recent improvements in working ASP systems facilitate the translation.,2142
ovary. nongrowing the follicles is reserve (NGFs) of the in by population ovarian defined Human,Human ovarian reserve is defined by the population of nongrowing follicles (NGFs) in the ovary.,2143
the prepared tissue. identification of NGFs Direct estimation ovarian involves ovarian reserve in of,Direct estimation of ovarian reserve involves the identification of NGFs in prepared ovarian tissue.,2144
estimated al. using (C is error concordance Prediction index) index et Harrell's [Harrell,Prediction error is estimated using Harrell's concordance index (C index) [Harrell et al.,2145
also for accounts The censoring. C-index specifically,The C-index also specifically accounts for censoring.,2146
"KIRSF functions, kernel utilizing many RSF By better situations. achieves in than results","By utilizing kernel functions, KIRSF achieves better results than RSF in many situations.",2147
"functions this into we how show kernel In report, to RSF. incorporate","In this report, we show how to incorporate kernel functions into RSF.",2148
method our to KIRSF of and RSF. compare test We performance the,We test the performance of KIRSF and compare our method to RSF.,2149
better We than in performance RSF occasions. that many find is KIRSF's the,We find that the KIRSF's performance is better than RSF in many occasions.,2150
of significantly. performance the finished the These solver affect decisions,These decisions affect the performance of the finished solver significantly.,2151
a We the alldifferent use constraint case study. as,We use the alldifferent constraint as a case study.,2152
of A risk a essentially functional. wide learning variety regularized machine minimize problems,A wide variety of machine learning problems essentially minimize a regularized risk functional.,2153
emerges by A picture and studying clearer Normal model. simpler the means,A clearer and simpler picture emerges by studying the Normal means model.,2154
"hidden the statistical states of units However, resulting dependencies. exhibit","However, the resulting states of hidden units exhibit statistical dependencies.",2155
RBMs. regularizer group} We \emph{sparse trained RBMs with call the,We call RBMs trained with the regularizer \emph{sparse group} RBMs.,2156
process is global in companies. a Management Knowledge,Knowledge Management is a global process in companies.,2157
our knitting industry. We approach apply on,We apply our approach on knitting industry.,2158
of maxRPC performance this boost In propose algorithms. techniques paper the that we can,In this paper we propose techniques that can boost the performance of maxRPC algorithms.,2159
"closely algorithms. two propose Based these, related we on","Based on these, we propose two closely related algorithms.",2160
it call the We algorithm. MH random dive (RDMH),We call it the random dive MH (RDMH) algorithm.,2161
convergence (ergodicity) basic ensure kernel. of These the,These ensure basic convergence (ergodicity) of the kernel.,2162
"constraint propose problem. We of the the modifies model a which new approach,","We propose a new approach, which modifies the constraint model of the problem.",2163
"search are out imposed the that constraints done. already additional Optionally, rule","Optionally, additional constraints are imposed that rule out the search already done.",2164
new models believable method develop to for conclude a we agents. two-step To propose,To conclude we propose a two-step method to develop new models for believable agents.,2165
algorithm designed. be model the learning can Then the and,Then the model and the learning algorithm can be designed.,2166
"combinational and do better However, methods guarantee the result. not integration this","However, this integration and the combinational methods do not guarantee better result.",2167
methods these feature and extracted accuracy. set for from The obtained tested compared were,The extracted feature set obtained from these methods were compared and tested for accuracy.,2168
behaviours well work to are They new believable validate suited on models.,They are well suited to validate work on new believable behaviours models.,2169
"agents' improve However, implementation, experiments numerous can the believability. help during to","However, during the implementation, numerous experiments can help to improve agents' believability.",2170
"although, can our difficult. experiment, analysis as promising shown be method This in results' seems","This method seems promising although, as shown in our experiment, results' analysis can be difficult.",2171
and network multiple capacity. study the We patterns store,We store multiple patterns and study the network capacity.,2172
"possible. is fitness evolution moderate optimisation For rates, still change through","For moderate change rates, fitness optimisation through evolution is still possible.",2173
overcomes shock down break when level. systems resistance These some given a,These systems break down when some shock overcomes a given resistance level.,2174
"used addition, commonly and angular, evaluated. analyzed (Minkowski, and In directional-distance) measures are distance","In addition, commonly used distance measures (Minkowski, angular, and directional-distance) are analyzed and evaluated.",2175
"filter choose how a requirements. given Finally, certain suggestions on to are provided","Finally, suggestions are provided on how to choose a filter given certain requirements.",2176
"used in transformations applications. Color processing, image graphics, visualization and frequently are space","Color space transformations are frequently used in image processing, graphics, and visualization applications.",2177
activate post-synaptic such as A diffusing that glutamate neuron receptors. pre-synaptic neurotransmitters releases,A pre-synaptic neuron releases diffusing neurotransmitters such as glutamate that activate post-synaptic receptors.,2178
involves contextual classifier. decision method tree classification using pixel The a,The method involves contextual pixel classification using a decision tree classifier.,2179
plays role detecting Accurately pedestrians a computer critically applications. in vision in many images important,Accurately detecting pedestrians in images plays a critically important role in many computer vision applications.,2180
effective the Extraction of is task. this to features key,Extraction of effective features is the key to this task.,2181
"compute. and robust Promising features to be discriminative, should various easy variations to","Promising features should be discriminative, robust to various variations and easy to compute.",2182
standard The Ojala proposed et by al. LBP,The standard LBP proposed by Ojala et al.,2183
"The feature, information. CS-LBP texture proposed captures the information some gradient and in contrast,","The proposed CS-LBP feature, in contrast, captures the gradient information and some texture information.",2184
operators their have use very time-critical requirements often applications. which in high computational limits These,These operators often have very high computational requirements which limits their use in time-critical applications.,2185
minimax speed approximation vector theory. up paper This introduces the using to filters techniques,This paper introduces techniques to speed up vector filters using the minimax approximation theory.,2186
alternatives Minkowski fact strong to The metrics. demonstrate are popular the that there in results,The results demonstrate that there are in fact strong alternatives to the popular Minkowski metrics.,2187
first step the this Border in often is detection analysis.,Border detection is often the first step in this analysis.,2188
thresholding approximate of lesion ensemble then location an the algorithms. determined The of is using,The approximate location of the lesion is then determined using an ensemble of thresholding algorithms.,2189
"to speed these methods filters. we this up introduce In two paper,","In this paper, we introduce two methods to speed up these filters.",2190
with ability the feedback. computer to endows This the to interaction user's adapt certain,This interaction endows the computer with certain ability to adapt to the user's feedback.,2191
to and shown presented caluculations Also indeed. are these be numerical are invariants useful,Also numerical caluculations are presented and these invariants are shown to be useful indeed.,2192
Levesque knowledge of base. precisely introduced the only-knowing beliefs the a of capture to notion,Levesque introduced the notion of only-knowing to precisely capture the beliefs of a knowledge base.,2193
ideas their first-order also to a case. clear not generalize is It to how,It is also not clear how to generalize their ideas to a first-order case.,2194
discuss Lakemeyer. by We the Halpern and to approach comparisons earlier also,We also discuss comparisons to the earlier approach by Halpern and Lakemeyer.,2195
Two regret bounds developed. surrogate to are approaches,Two approaches to surrogate regret bounds are developed.,2196
generalization et a al. Bartlett The of first direct is,The first is a direct generalization of Bartlett et al.,2197
"exponential, hinge, sigmoid detail. losses The in squared treated then are uneven error, and","The uneven hinge, squared error, exponential, and sigmoid losses are then treated in detail.",2198
hashing to process an vector Image is of bits associating image. short a the of,Image hashing is the process of associating a short vector of bits to an image.,2199
"any relabeling Interestingly, under spectrum invariant this of nodes. is the graph","Interestingly, this spectrum is invariant under any relabeling of the graph nodes.",2200
problems is Real-time vision. of core detection the one in object computer,Real-time object detection is one of the core problems in computer vision.,2201
algorithms We develop address boosting asymmetric new learning problem. this to,We develop new boosting algorithms to address this asymmetric learning problem.,2202
objectives loss explicitly optimize that fashion. show our asymmetric a in methods corrective We totally,We show that our methods explicitly optimize asymmetric loss objectives in a totally corrective fashion.,2203
the the boosting the column At generation technique. of corrective heart totally is,At the heart of the totally corrective boosting is the column generation technique.,2204
asymmetric on outperform the methods. state-of-the-art our show boosting Experiments that detection methods face,Experiments on face detection show that our methods outperform the state-of-the-art asymmetric boosting methods.,2205
reduction of selection. diversity A describes the positive due selective strong sweep to,A selective sweep describes the reduction of diversity due to strong positive selection.,2206
which very task instability best. This of method hard the establishing objectively makes performs,This instability makes objectively very hard the task of establishing which method performs best.,2207
"In these infinitely many have solutions. general, systems","In general, these systems have infinitely many solutions.",2208
"However, shown be that may sparse uniquely. it sufficiently solutions be may identified","However, it may be shown that sufficiently sparse solutions may be identified uniquely.",2209
computational drawback of complexity. their is of current finding The methods solutions main sparse,The main drawback of current methods of finding sparse solutions is their computational complexity.,2210
"observed. texture, can be many this In drawings (geoglyphs)","In this texture, many drawings (geoglyphs) can be observed.",2211
"Layout. propose Finally, Keyboard Bangla a we","Finally, we propose a Bangla Keyboard Layout.",2212
in discuss arise from areas. also that might applications other We our work these,We also discuss other applications that might arise from our work in these areas.,2213
subserves exploration. pathway while exploitation the pathway direct the subserves Thus indirect,Thus the direct pathway subserves exploitation while the indirect pathway subserves exploration.,2214
variability with show diminishing Reaching trajectories training.,Reaching trajectories show diminishing variability with training.,2215
echoes that disease. model is PD notion the dynamical a The,The model echoes the notion that PD is a dynamical disease.,2216
mesh nonlinear with of deals for improvement based paper diffusion This vertex denoising. an,This paper deals with an improvement of vertex based nonlinear diffusion for mesh denoising.,2217
"we use of metrics. these methods To filtering, evaluate error two","To evaluate these methods of filtering, we use two error metrics.",2218
normals. the the is on the vertices on based and second first The is based,The first is based on the vertices and the second is based on the normals.,2219
Experimental with of the in our the method demonstrate effectiveness methods. proposed comparison existing results,Experimental results demonstrate the effectiveness of our proposed method in comparison with the existing methods.,2220
Many the algorithms approximate search for neighbor into data partition high-dimensional in clusters. spaces nearest,Many algorithms for approximate nearest neighbor search in high-dimensional spaces partition the data into clusters.,2221
it the are several well-known $k$-means approach often properties. desirable has since by Clusters produced,Clusters are often produced by the well-known $k$-means approach since it has several desirable properties.,2222
"it produce quite having clusters tends On to the cardinalities. different downside,","On the downside, it tends to produce clusters having quite different cardinalities.",2223
times. clusters and impact Imbalanced both the response expectation query the negatively of variance,Imbalanced clusters negatively impact both the variance and the expectation of query response times.,2224
features proposed We demonstrate compared yield performance improvement our to Haar-like features. that a,We demonstrate that our proposed features yield a performance improvement compared to Haar-like features.,2225
methods. to two and problems We on compare method the empirically it specific other illustrate,We illustrate the method on two specific problems and empirically compare it to other methods.,2226
"C-SVM, several has Compared performance on GS-SVM data better sets. with","Compared with C-SVM, GS-SVM has better performance on several data sets.",2227
the of with better effectiveness results several on show proposed the approach Experimental data performance.,Experimental results on several data show the effectiveness of the proposed approach with better performance.,2228
Probabilities the canonical to map by the the RKHS. are uniquely mean of characterized,Probabilities are uniquely characterized by the mean of the canonical map to the RKHS.,2229
an of likewise posterior The RKHS a sample. mean is weighted,The posterior is likewise an RKHS mean of a weighted sample.,2230
trees. following simulation a In the we stochastic consider for technique paper,In the following paper we consider a simulation technique for stochastic trees.,2231
typically (SMC) consists sequential techniques. Monte sampling (IS) This and importance of using Carlo,This typically consists of using importance sampling (IS) and sequential Monte Carlo (SMC) techniques.,2232
bias is of view. point a investigated from The theoretical,The bias is investigated from a theoretical point of view.,2233
proposed on effectiveness with several data performance. approach of show the better results the Experimental,Experimental results on several data show the effectiveness of the proposed approach with better performance.,2234
a two-dimensional input image wavelet. two-dimensional is analyzed wavelet stationary and by The,The input image is analyzed by two-dimensional wavelet and a two-dimensional stationary wavelet.,2235
vectors an dimension input PCA. the applied of reduce effective using We to procedure the,We applied an effective procedure to reduce the dimension of the input vectors using PCA.,2236
"regulatory components. are circuits combining circuit gene synthetic by constructed often In biology, smaller","In synthetic biology, gene regulatory circuits are often constructed by combining smaller circuit components.",2237
are achieved Connections transcription promoters. on between components acting factors by,Connections between components are achieved by transcription factors acting on promoters.,2238
"one investigate conditions: this interface In fan-out. we the paper, of","In this paper, we investigate one of the interface conditions: fan-out.",2239
the fan-out enhanced on output. self-inhibitory be We can the also regulation by show that,We also show that the fan-out can be enhanced by self-inhibitory regulation on the output.,2240
"such Non-genetic  noise, novel  induce developmental perturbations, environmental can change as or phenotypes.","  Non-genetic perturbations, such as environmental change or developmental noise, can induce novel phenotypes.",2241
"an selection induced If promote advantage, phenotype may a fitness stabilization. confers genetic its","If an induced phenotype confers a fitness advantage, selection may promote its genetic stabilization.",2242
initiate Non-genetic evolutionary innovation. perturbations thus can,Non-genetic perturbations can thus initiate evolutionary innovation.,2243
"phenotypes. perturbations, new a variation can source become non-genetic After of this","After non-genetic perturbations, this variation can become a source of new phenotypes.",2244
"model well-studied regulation To transcriptional of use we end, this a circuits.","To this end, we use a well-studied model of transcriptional regulation circuits.",2245
important innovations. evolutionary Such in circuits are many,Such circuits are important in many evolutionary innovations.,2246
= and are where N xn ai pn integers. N,xn pn = N where ai and N are integers.,2247
very segmentation analysis area. topic image research Image active a in been has,Image segmentation has been a very active research topic in image analysis area.,2248
"intuition pattern. However, this does always not human follow visual","However, human visual intuition does not always follow this pattern.",2249
in technique bioinformatics tools. various vaccine analyzed potential be The candidates silico should using,The potential vaccine candidates should be analyzed in silico technique using various bioinformatics tools.,2250
toxic invading and a system immune pathogens products. recognizes myriad The their of,The immune system recognizes a myriad of invading pathogens and their toxic products.,2251
repertoire and receptors. so It T cell a finite antibodies of does with,It does so with a finite repertoire of antibodies and T cell receptors.,2252
that here system We immune dynamics. theories the quantify describe,We here describe theories that quantify the immune system dynamics.,2253
invariant and the translation demonstrated pursuit . is Matching in K-SVD setting,Matching pursuit and K-SVD is demonstrated in the translation invariant setting .,2254
"in statistics This processing. problem machine learning, signal relevant and is","This problem is relevant in machine learning, statistics and signal processing.",2255
regulated of osteoclasts. controlling is osteoblasts by and Bone  remodeling pathways  interplay the,  Bone remodeling is regulated by pathways controlling the interplay of osteoblasts and osteoclasts.,2256
"and have interaction been recently, considered. dispersal More spatial also","More recently, spatial interaction and dispersal have also been considered.",2257
of has expectations helped our molecular patterns shape variation. formula empirical of This,This formula has helped shape our expectations of empirical patterns of molecular variation.,2258
"the null theory, with model Along statistical it rejecting neutrality. provides for coalescent of techniques","Along with coalescent theory, it provides statistical techniques for rejecting the null model of neutrality.",2259
"to well and the compare it method ""all-pairs"" We test known decomposition. our","We test our method and compare it to the well known ""all-pairs"" decomposition.",2260
new authors. tests using authorship are machine performed test learning Our a of benchmark determination,Our tests are performed using a new authorship determination benchmark test of machine learning authors.,2261
we approach modeling. a Here introduction generalized to pedagogical the present of,Here we present a pedagogical introduction to the approach of generalized modeling.,2262
generalized of places emphasis This concepts the introduction more previous modeling than underlying publications. on,This introduction places more emphasis on the underlying concepts of generalized modeling than previous publications.,2263
game theory. into it has evolutionary introduced Recently been,Recently it has been introduced into evolutionary game theory.,2264
evolutionary processes? universal for those selection microscopic How is weak,How universal is weak selection for those microscopic evolutionary processes?,2265
"find selection of the allow intensity. We which expansions, order rescaling higher a universal","We find universal higher order expansions, which allow a rescaling of the selection intensity.",2266
"But including all what human is source heuristics? of the knowledge,","But what is the source of all human knowledge, including heuristics?",2267
that results is experimental acquisition report pleasing. of showing knowledge emotionally We,We report experimental results showing that acquisition of knowledge is emotionally pleasing.,2268
knowledge satisfaction brings of through curiosity acquiring The pleasure.,The satisfaction of curiosity through acquiring knowledge brings pleasure.,2269
as immunity pathogens use Several host. against the strategy evolvability survival a acquired of,Several pathogens use evolvability as a survival strategy against acquired immunity of the host.,2270
forces paper analyzes the us driving both. to This cognitive achieve,This paper analyzes the cognitive forces driving us to achieve both.,2271
"soon able a possibly, psychological in And laboratory. measure to may we them be","And possibly, we may soon be able to measure them in a psychological laboratory.",2272
challenge: overcome long-standing analysis another has reductionism. scientific new This,This new scientific analysis has overcome another long-standing challenge: reductionism.,2273
by integrator In this perfect input. consider note the we Poisson process driven,In this note we consider the perfect integrator driven by Poisson process input.,2274
of associated from induced with We the study divergences. quasi-Newton viewpoint Bregman information methods geometry,We study quasi-Newton methods from the viewpoint of information geometry induced associated with Bregman divergences.,2275
methods. update of class quasi-Newton to The formulae self-scaling belong a proposed,The proposed update formulae belong to a class of self-scaling quasi-Newton methods.,2276
robust statistics. from the conducted the verify to of Numerical studies tools usefulness borrowed are,Numerical studies are conducted to verify the usefulness of the tools borrowed from robust statistics.,2277
of key (TFs) are gene factors expression. Transcription regulators,Transcription factors (TFs) are key regulators of gene expression.,2278
"proposed regret, and generalization method. bounds for convergence, show the We","We show regret, convergence, and generalization bounds for the proposed method.",2279
recognition Experiments on testify successfulness and dependency handwriting the approach. the parsing for of,Experiments on handwriting recognition and dependency parsing testify for the successfulness of the approach.,2280
condensed a larger in progress. of paper a This version work is,This paper is a condensed version of a larger work in progress.,2281
"they that are entries However, of the require observation the all matrix known.","However, they require that all the entries of the observation matrix are known.",2282
"view. unlikely due happen practice, is to to field of This and limited in self-occlusion","This is unlikely to happen in practice, due to self-occlusion and limited field of view.",2283
"often Also, later. long regions appear occluded that when videos, again become processing","Also, when processing long videos, regions that become occluded often appear again later.",2284
and videos illustrate artificial the with our method. real Experiments of good performance,Experiments with artificial and real videos illustrate the good performance of our method.,2285
CS. of The smaller is significantly also than conventional probability failure that,The failure probability is also significantly smaller than that of conventional CS.,2286
either sets or fuzzy S-norm T-norm. defined operators fuzzy as must serve All for fuzzy,All operators defined for fuzzy sets must serve as either fuzzy S-norm or fuzzy T-norm.,2287
"satisfy Second, each Demorgans other they so complement they law, perfectly.","Second, they satisfy Demorgans law, so they complement each other perfectly.",2288
"These and via Geometry logic. are investigated Mathematics, fuzzy viewpoints: operators three","These operators are investigated via three viewpoints: Mathematics, Geometry and fuzzy logic.",2289
"by the membrane driven in forces. we study can Here, be composition how variations fluctuating","Here, we study how variations in the membrane composition can be driven by fluctuating forces.",2290
"grey-level can this how applied we be normalize method Finally, discuss to images. further","Finally, we discuss how this method can further be applied to normalize grey-level images.",2291
"a is in Wiselib, generic implemented sensor networks. of library algorithms for protocol The","The protocol is implemented in Wiselib, a library of generic algorithms for sensor networks.",2292
Shawn. simulator conducted Simulations with are network sensor the,Simulations are conducted with the sensor network simulator Shawn.,2293
as of based known real iSense They specifications nodes. the are hardware sensor on,They are based on the specifications of real hardware known as iSense sensor nodes.,2294
"it against very robust is is the loss. Moreover, system packet shown that","Moreover, it is shown that the system is very robust against packet loss.",2295
the grounded describe brain. mathematical symbols in models We a of,We describe a mathematical models of grounded symbols in the brain.,2296
System also as Perceptual foundations It serves (PSS). computational a Symbol for,It also serves as a computational foundations for Perceptual Symbol System (PSS).,2297
"The DL, a new theory, is process-logic. mathematical","The new mathematical theory, DL, is a process-logic.",2298
property into crisp. of evolution process vague this representations A salient of is,A salient property of this process is evolution of vague representations into crisp.,2299
perceptions. applies to PSS: of it one aspect paper object The from learning situation first,The paper first applies it to one aspect of PSS: situation learning from object perceptions.,2300
"capabilities Finally, these shape-based we illustrate experiments. with image classification","Finally, we illustrate these capabilities with shape-based image classification experiments.",2301
A core  new  laughter is presented. hypothesis on,  A new core hypothesis on laughter is presented.,2302
a thesis. framework supports present this that We theoretical,We present a theoretical framework that supports this thesis.,2303
found are observed model. natural systems in several features We that mathematical the predicted by,We found that several features observed in natural systems are predicted by the mathematical model.,2304
two A by probability defined the is densities. ratio of density ratio,A density ratio is defined by the ratio of two probability densities.,2305
the is f-divergence estimator then test. for exploited homogeneity two-sample The,The f-divergence estimator is then exploited for the two-sample homogeneity test.,2306
"we the the inference. for asymptotic studies, Through adequacy illustrate theory numerical finite-sample of","Through numerical studies, we illustrate the adequacy of the asymptotic theory for finite-sample inference.",2307
eigenmaps Laplacian model and relate the We to isomap.,We relate the model to Laplacian eigenmaps and isomap.,2308
(ALLE). performs likelihood LLE of exactly: that a Acyclic We LLE variant maximum introduce,We introduce a variant of LLE that performs maximum likelihood exactly: Acyclic LLE (ALLE).,2309
interesting demonstrate that vision on our can We several build applications system. also,We also demonstrate several interesting vision applications that can build on our system.,2310
the are sources coding. at the identified selected looking present automatically The in by sub-dictionaries,The present sources are identified by looking at the sub-dictionaries automatically selected in the coding.,2311
from present We and texture applications speaker separation. and identification instrument,We present applications from speaker and instrument identification and texture separation.,2312
automatically proposed detects method independent number the The pitch of a sources on recording.,The proposed pitch independent method automatically detects the number of sources on a recording.,2313
refers a Concept to drift learning time. problem stationary over non,Concept drift refers to a non stationary learning problem over time.,2314
application and training in The the mismatch often life real problems. data,The training and the application data often mismatch in real life problems.,2315
to training formation. set issues on the We focus relevant adaptive,We focus on the issues relevant to adaptive training set formation.,2316
unique be reaction limb can force. and distinguished ground by movements Human different,Human can be distinguished by different limb movements and unique ground reaction force.,2317
the Reasoning of Qualitative is An  issue important representation direction.  in Spatial relative,  An important issue in Qualitative Spatial Reasoning is the representation of relative direction.,2318
in out is scalable these And useful granularity navigation that it tasks. turns,And it turns out that scalable granularity is useful in these navigation tasks.,2319
This intractable draws the boundary subclasses sharp a of the CDC. and between tractable,This draws a sharp boundary between the tractable and intractable subclasses of the CDC.,2320
"considered In generation problem, Theory. based Rough is reduct paper Set this on it rule","In this paper it is considered rule reduct generation problem, based on Rough Set Theory.",2321
(RG) Generation are Rule (MRG) Modified and algorithms well-known. Rule Reduct Generation,Rule Reduct Generation (RG) and Modified Rule Generation (MRG) algorithms are well-known.,2322
algorithm tree type. structured PRG data uses,PRG algorithm uses tree structured data type.,2323
PRG algorithms MRG and algorithm is with compared RG .,PRG algorithm is compared with RG and MRG algorithms .,2324
of the to appears four. process the most autoregressive accurate an method Fitting be,Fitting an autoregressive process appears to be the most accurate method of the four.,2325
comparison package more R methods extending series. and to is provided test the An for,An R package is provided for extending the comparison to more methods and test series.,2326
computers played In via is network game a article the investigated. on connected turn-based four,In the article a turn-based game played on four computers connected via network is investigated.,2327
intelligence There with intelligence. computers and are with natural one three artificial,There are three computers with natural intelligence and one with artificial intelligence.,2328
monitors. seen all in by players' table player's point Game view own each is,Game table is seen by each player's own view point in all players' monitors.,2329
distributed For systems is used. TCP/IP protocol,For distributed systems TCP/IP protocol is used.,2330
. of Number of the the multiplication is equal following to combinations: distributions,Number of the distributions is equal to the multiplication of following combinations: .,2331
make an it difficult reasons AI. to develop These,These reasons make it difficult to develop an AI.,2332
levels the article of are AI developed. In four,In the article four levels of AI are developed.,2333
with General of are and VC-dimension considered. class loss predictors functions finite,General loss functions and class of predictors with finite VC-dimension are considered.,2334
"of the focus In various we on procedures. the cross-validation proving consistency particular,","In particular, we focus on proving the consistency of the various cross-validation procedures.",2335
interest We in terms rate of of the each cross-validation point out procedure convergence. of,We point out the interest of each cross-validation procedure in terms of rate of convergence.,2336
"three are of Here, key properties uncovered. novel the RSN","Here, three key novel properties of the RSN are uncovered.",2337
"it also mutual information. measures is this observed of Second, divergence for","Second, this divergence it is observed also for measures of mutual information.",2338
The twofold. is respect with MCMC to literature the existing extension,The extension with respect to the existing MCMC literature is twofold.,2339
to ... demonstrate ability find the this . To,To demonstrate this ability to find the ... .,2340
"The show viz. study the two of classification to importance is techniques, aim this of","The aim of this study is to show the importance of two classification techniques, viz.",2341
"clustering, disabilities in of (LD) school-age decision learning children. prediction and tree of","decision tree and clustering, in prediction of learning disabilities (LD) of school-age children.",2342
extracted decision learning are Different of disabilities. from the tree rules for used prediction,Different rules extracted from the decision tree are used for prediction of learning disabilities.,2343
"identified. any LD By techniques, applying these in can classification be child","By applying these classification techniques, LD in any child can be identified.",2344
"symmetric, a to theory dispersal has for long be i.e. time assumed Metapopulation","Metapopulation theory for a long time has assumed dispersal to be symmetric, i.e.",2345
dispersing a connected migrants through bi-directionally preferred patches are without direction.,patches are connected through migrants dispersing bi-directionally without a preferred direction.,2346
"natural symmetry often for e.g. populations However, broken, is","However, for natural populations symmetry is often broken, e.g.",2347
"that suggest, results asymmetry Our negative metapopulation persistence. effects not imply on does in itself","Our results suggest, that asymmetry in itself does not imply negative effects on metapopulation persistence.",2348
disadvantages. calibration authors and have with advantages relative several Many presented their algorithms,Many authors have presented several calibration algorithms with their relative advantages and disadvantages.,2349
"stereovision system, In a task. reconstruction view is multiple a challenging","In a stereovision system, multiple view reconstruction is a challenging task.",2350
"detail total not the in been before. computational has presented However, procedure","However, the total computational procedure in detail has not been presented before.",2351
"is aspect, situation In vision computer this undesirable.","In computer vision aspect, this situation is undesirable.",2352
We it in have fashion. an done elegant,We have done it in an elegant fashion.,2353
"local camera its parameters system. are Firstly, coordinate calculated in","Firstly, camera parameters are calculated in its local coordinate system.",2354
"calibrated. is whole That is, system the","That is, the whole system is calibrated.",2355
are pose vehicle for detection Results presented.,Results for vehicle pose detection are presented.,2356
"Vermont. from St. Michael's in College Colchester,","from St. Michael's College in Colchester, Vermont.",2357
as US and train serve Air the Force Jim in to a Army meteorologist. enlisted,Jim enlisted in the US Army Air Force to train and serve as a meteorologist.,2358
the by took in army airbases war. the to him This China of close,This took him to army airbases in China by the close of the war.,2359
"and from studied the discharge graduated Following Harvard the army, M.S. at with mathematics Jim","Following discharge from the army, Jim studied mathematics at Harvard and graduated with the M.S.",2360
a the lung revealed tuberculosis. physical The routine spot admissions' and the on possibility of,The routine admissions' physical revealed a spot on the lung and the possibility of tuberculosis.,2361
in began theory research under Herbert There Robbins. compound decision he,There he began research in compound decision theory under Herbert Robbins.,2362
"told When his finish that degree. Jim felt to coming, pressure not renewal was","When told that renewal was not coming, Jim felt pressure to finish his degree.",2363
single-input-singleoutput a of some system any fuzzy expresses multi-input-single-output combination as systems. ALM,ALM expresses any multi-input-single-output system as a fuzzy combination of some single-input-singleoutput systems.,2364
an is on action Reward- goodness Penalty-Plane. The of modeled,The goodness of an action is modeled on Reward- Penalty-Plane.,2365
according updated to IDS plane. be this planes will,IDS planes will be updated according to this plane.,2366
method images with gives also. aliased result This good,This method gives good result with aliased images also.,2367
between a nodes(s). constraint and minimum put distance allowed on,and put a constraint on minimum allowed distance between nodes(s).,2368
relies models these (MCMC) techniques. Monte Markov typically inference for Bayesian chain Carlo on,Bayesian inference for these models typically relies on Markov chain Monte Carlo (MCMC) techniques.,2369
of Andrieu methods The et existing PMCMC al.,The existing PMCMC methods of Andrieu et al.,2370
Three approach. real are the demonstrate to examples used proposed,Three real examples are used to demonstrate the proposed approach.,2371
most the machine of learning. of important Classification is one tasks,Classification is one of the most important tasks of machine learning.,2372
of problem. classifying the the two-class technique reduces The with to option reject standard problem,The technique reduces the problem of classifying with reject option to the standard two-class problem.,2373
is into and then method mapped The support neural machines introduced vector networks.,The introduced method is then mapped into support vector machines and neural networks.,2374
"the is option. framework data to Finally, multiclass ordinal reject extended with","Finally, the framework is extended to multiclass ordinal data with reject option.",2375
considers may Our integrated parameters out. block approach where model a be block,Our approach considers a block model where block parameters may be integrated out.,2376
real technique. data We validate the simulated analyze to both and,We analyze both simulated and real data to validate the technique.,2377
vector co-occurrence including multi-direction descriptor and local is A a multiscale information. scattering,A scattering vector is a local descriptor including multiscale and multi-direction co-occurrence information.,2378
computed is of with a It decompositions cascade and modulus. complex wavelet,It is computed with a cascade of wavelet decompositions and complex modulus.,2379
representation locally is This deformations. invariant linearizes translation and scattering,This scattering representation is locally translation invariant and linearizes deformations.,2380
a with vectors. A classification selection algorithm computed is scattering on model PCA supervised,A supervised classification algorithm is computed with a PCA model selection on scattering vectors.,2381
for results classification. State recognition the are and texture of obtained digit art handwritten,State of the art results are obtained for handwritten digit recognition and texture classification.,2382
results. experimental the latter and proven flexibility variety of of The a functionality with is,The functionality and flexibility of the latter is proven with a variety of experimental results.,2383
"introduced. based are complementing on two thinning point processes and tests additional Furthermore,","Furthermore, two additional tests based on thinning and complementing point processes are introduced.",2384
"method structured prediction. popular probabilistic tutorial fields, describes conditional This for random a","This tutorial describes conditional random fields, a popular probabilistic method for structured prediction.",2385
"application and seen natural vision, bioinformatics. processing, wide CRFs language in computer have","CRFs have seen wide application in natural language processing, computer vision, and bioinformatics.",2386
contact We patterns. propose exploiting temporal immunization schemes two,We propose two immunization schemes exploiting temporal contact patterns.,2387
"implemented of could only local practice. these rely information and Both on obtainable, in","Both of these rely only on obtainable, local information and could implemented in practice.",2388
by primary data producers. been analyses low-resolution limited Empirical further for have,Empirical analyses have been further limited by low-resolution data for primary producers.,2389
"outliers identify used In reduce their approach addition, is Keller's and influences. to","In addition, Keller's approach is used to identify outliers and reduce their influences.",2390
Euclidean a change transformation the to distance. distance also to our defined We have,We have also defined a transformation to change our distance to the Euclidean distance.,2391
log-density of the gradient be requires computable. that It the,It requires that the gradient of the log-density be computable.,2392
"examine potential Additionally, phenomena. shifts level related of DC these the to we","Additionally, we examine shifts of the DC potential level related to these phenomena.",2393
rule computationally The extra adding cost algorithm. simple AM adaptation with the no compared is,The adaptation rule is computationally simple adding no extra cost compared with the AM algorithm.,2394
of are simulation free. All methods these,All of these methods are simulation free.,2395
the of real general. usefulness Analysis data the demonstrates in of approach,Analysis of real data demonstrates the usefulness of the approach in general.,2396
of in the quasispecies model. threshold Eigen the the error We for investigate emergence,We investigate the error threshold for the emergence of quasispecies in the Eigen model.,2397
of the agreement. the ansatz validity check methods numerical find variational and excellent by We,We check the validity of the variational ansatz by numerical methods and find excellent agreement.,2398
"the each We outperforms where other. method examples, describe","We describe examples, where each method outperforms the other.",2399
"leave-one-out interest case, this emphasized. cross-validation is of In particular special the","In this special case, the particular interest of leave-one-out cross-validation is emphasized.",2400
of General class and infinite loss functions finite predictors considered. VC-dimension with are both and,General loss functions and class of predictors with both finite and infinite VC-dimension are considered.,2401
"predictor. rule give to a how the we simple Finally, on subbag","Finally, we give a simple rule on how to subbag the predictor.",2402
Markov Carlo performance This a comparing graphical method Monte for Chain methods. paper of presents,This paper presents a graphical method for comparing performance of Markov Chain Monte Carlo methods.,2403
in detailed IEEE this the more Trans. published been has of work A version,A more detailed version of this work has been published in the IEEE Trans.,2404
the of The problem of optimization version concerns number the minimization of colors. used the,The optimization version of the problem concerns the minimization of the number of used colors.,2405
control. from prescind algorithms central Such any,Such algorithms prescind from any central control.,2406
use Male to their attract females. calls frogs,Male frogs use their calls to attract females.,2407
"calls. nearby each Interestingly, located their groups are that other of desynchronize males","Interestingly, groups of males that are located nearby each other desynchronize their calls.",2408
prove We the our formally of correctness approach.,We formally prove the correctness of our approach.,2409
a a encoded also show reinforcement problem. as We that SAT learning problem be can,We also show that a reinforcement learning problem can be encoded as a SAT problem.,2410
"analyze Yule models. Here, and and birth-death realistic we alternative more","Here, we analyze alternative and more realistic Yule and birth-death models.",2411
Unimodal disadvantages that show performance systems Experimental studies biometric accuracy. and many regarding had,Experimental studies show that Unimodal biometric systems had many disadvantages regarding performance and accuracy.,2412
biometric discuss unimodal over their advantage We systems. also,We also discuss their advantage over unimodal biometric systems.,2413
"yet to be has established. picture a complete However,","However, a complete picture has yet to be established.",2414
establish on cyclic explicitly a graph In we of dependency motifs. neural particular complexity,In particular we explicitly establish a dependency of neural complexity on cyclic graph motifs.,2415
stationary assumption relaxes proposed noise the in algorithm model coding. the usual sparse The of,The proposed algorithm relaxes the usual assumption of the stationary noise model in sparse coding.,2416
et the Woodford and al. more of algorithm recent,and the more recent algorithm of Woodford et al.,2417
"Unfortunately, on priori DTs always of a available. information favored is not structure","Unfortunately, a priori information on favored structure of DTs is not always available.",2418
"this reason, technique. for a we DTs For Bayesian MCMC new prior on the suggest","For this reason, we suggest a new prior on DTs for the Bayesian MCMC technique.",2419
"accuracy. we also is approximate, method As evaluate its only Huber-penalty based the","As the Huber-penalty based method is only approximate, we also evaluate its accuracy.",2420
modifies those al. et Karandikar of Our model,Our model modifies those of Karandikar et al.,2421
strategy. The players learning perform reactive the also against efficiently,The learning players also perform efficiently against the reactive strategy.,2422
"competitive population can simpler players but of strategies. In dynamics, invade they evolutionary adopting a","In evolutionary dynamics, they can invade a population of players adopting simpler but competitive strategies.",2423
It between social in learning explore and the dilemma relationships evolution to may situations. serve,It may serve to explore the relationships between learning and evolution in social dilemma situations.,2424
"State Iowa College, then completed a he went He to M.Sc. where","He then went to Iowa State College, where he completed a M.Sc.",2425
and various as He occupied Planning. Assistant culminating of Director Corporate Vice-President positions,He occupied various positions culminating as Assistant Vice-President and Director of Corporate Planning.,2426
analysis component suffers badly The (PCA) of of principal outliers. the presence in performance,The performance of principal component analysis (PCA) suffers badly in the presence of outliers.,2427
paper on novel robust semidefinite for PCA This approaches programming. two based proposes,This paper proposes two novel approaches for robust PCA based on semidefinite programming.,2428
also computational presents these SDPs. efficient methods for solving paper This,This paper also presents efficient computational methods for solving these SDPs.,2429
Numerical the of confirm value new these techniques. experiments,Numerical experiments confirm the value of these new techniques.,2430
choose example explanations at capturing from as an We information. causal formalism a aiming causal,We choose as an example a formalism aiming at capturing causal explanations from causal information.,2431
"We showing job. this provide implementation, of and translation efficiency an naturalness the relative","We provide an implementation, showing the naturalness and relative efficiency of this translation job.",2432
"an new, propose on approach, general We based local more a procedure. iterative optimization","We propose a new, more general approach, based on an iterative local optimization procedure.",2433
in as was a report. model systematic this then right laboratory-effects The one selected,The systematic laboratory-effects model was then selected as a right one in this report.,2434
all assumption. one on value based true Those the models were existence,Those models were all based on the one true value existence assumption.,2435
"we a web-based In this CT MySNOM, SNOMED present paper browser.","In this paper we present MySNOM, a web-based SNOMED CT browser.",2436
"was solution. Moreover, reference the water desionised as considered the","Moreover, the desionised water was considered as the reference solution.",2437
albumin. in is less constituted This alloy the by biomedical the protected and solution PBS,This biomedical alloy is the less protected in the solution constituted by PBS and albumin.,2438
"r\'ef\'erence. d'eau de ultra-pure une De solution plus, servira","De plus, une solution d'eau ultra-pure servira de r\'ef\'erence.",2439
a tests \'etudi\'ee. reproductibilit\'e La des \'et\'e,La reproductibilit\'e des tests a \'et\'e \'etudi\'ee.,2440
"by many performed specialized, applications, is SAT tools. However, for into problem-specific translation","However, for many applications, translation into SAT is performed by specialized, problem-specific tools.",2441
language paradigms. declarative a new specification imperative combines uses URSA that and programming system The,The system uses a new specification language URSA that combines imperative and declarative programming paradigms.,2442
precisely SAT specification the The language. semantics defined by is to the reduction of,The reduction to SAT is defined precisely by the semantics of the specification language.,2443
"this approach. open-source of describe an implementation paper, we also In the described","In this paper, we also describe an open-source implementation of the described approach.",2444
The systems. to performed system modelling experiments the suggest is competitive state-of-the-art that related,The performed experiments suggest that the system is competitive to state-of-the-art related modelling systems.,2445
implemented exemplified controversy. prototype MMR the is vaccine modelling The by,The implemented prototype is exemplified by modelling the MMR vaccine controversy.,2446
"Populous, an which tool ontology. populate present a We for with content to gathering","We present Populous, a tool for gathering content with which to populate an ontology.",2447
is the in ontology development. knowledge gathering stage contribution Populous's of,Populous's contribution is in the knowledge gathering stage of ontology development.,2448
"Web Semantic wikis, enhanced knowledge with community-authored Semantic technologies, systems appropriate for are models. wikis","Semantic wikis, wikis enhanced with Semantic Web technologies, are appropriate systems for community-authored knowledge models.",2449
for They collaboration. scientific are suitable particularly,They are particularly suitable for scientific collaboration.,2450
"This semantic details wiki. a the principles ofWikiBridge, design paper","This paper details the design principles ofWikiBridge, a semantic wiki.",2451
. Approach Modular Creating a Ontology: a new,Creating a new Ontology: a Modular Approach .,2452
"language it And result also, in understanding will to technology. benefits","And also, it will result in benefits to language understanding technology.",2453
been long accepted of nondeterministic have a automata. generalization finite automata Fuzzy as,Fuzzy automata have long been accepted as a generalization of nondeterministic finite automata.,2454
"measure, been bisimulation, has automata. to a introduced finite finer behavioral fuzzy Recently,","Recently, a finer behavioral measure, bisimulation, has been introduced to fuzzy finite automata.",2455
"are the results only finite-state applicable obtained to However, systems.","However, the results obtained are applicable only to finite-state systems.",2456
presented comparing for systems. The results the general here of behavior fuzzy are useful,The results presented here are useful for comparing the behavior of general fuzzy systems.,2457
process cancer proposed hypoxia-induced is the cells. in of angiogenesis study model A to,A model is proposed to study the process of hypoxia-induced angiogenesis in cancer cells.,2458
scenario. analyzed are mean-field These to ideal reference with the aspects firstly,These aspects are firstly analyzed with reference to the ideal mean-field scenario.,2459
"algorithms, previous our convergent. In most to contrast is method provably","In contrast to most previous algorithms, our method is provably convergent.",2460
brain further and the use language-like can both propose representation We that picture-like processes.,We further propose that the brain can use both picture-like and language-like representation processes.,2461
Our proposal (i.e. with representation of deals only the primary visual level,Our proposal deals only with the primary level of visual representation (i.e.,2462
of are derived. asymptotic for estimator unbiasedness zero-variance Conditions the,Conditions for asymptotic unbiasedness of the zero-variance estimator are derived.,2463
also proved limit central regularity A theorem is under conditions.,A central limit theorem is also proved under regularity conditions.,2464
actors to ensure solving overall of These the collaborate process. success the problem decision,These actors collaborate to ensure the overall success of the decision problem solving process.,2465
"for distance function introduced. and A fuzzy is sets new B dist(A,B) A","A new distance function dist(A,B) for fuzzy sets A and B is introduced.",2466
data this on prove distance. pattern on its We based properties and mathematical clustering perform,We prove its mathematical properties and perform pattern clustering on data based on this distance.,2467
between biomarkers new disease a research. 'omics' experiments potential multi-species in biomedical Translating direction is,Translating potential disease biomarkers between multi-species 'omics' experiments is a new direction in biomedical research.,2468
experimental to comparisons. healthy-diseased limited setups basic are methods simple as such existing The,The existing methods are limited to simple experimental setups such as basic healthy-diseased comparisons.,2469
is problem-solving subjective complex. behavior and is human whose Knowledge attributed to,Knowledge is attributed to human whose problem-solving behavior is subjective and complex.,2470
within reuse. of for context resolution the expressed future problem a Knowledge be must capitalized,Knowledge expressed within the context of a problem resolution must be capitalized for future reuse.,2471
. adapted system The domains can to various be,The system can be adapted to various domains .,2472
of models mind. between mathematical theory and aesthetics relationships The discusses paper,The paper discusses relationships between aesthetics theory and mathematical models of mind.,2473
The instinct foundation higher and mental aesthetic of emotions. knowledge is the abilities,The knowledge instinct is the foundation of higher mental abilities and aesthetic emotions.,2474
surrounding issue. controversy a I this discuss,I discuss a controversy surrounding this issue.,2475
perturbations. for is the true modularity same after the speed relaxation scaling The of,The same modularity scaling is true for the speed of relaxation after the perturbations.,2476
management. usually implementation chain an The requires such of appropriate supply alternatives reverse,The implementation of such alternatives usually requires an appropriate reverse supply chain management.,2477
"of is to the functioning intuition self, morality, Free of and will fundamental society. normal","Free will is fundamental to morality, intuition of self, and normal functioning of the society.",2478
"foundation not does However, idea. clear logical provide a science for this","However, science does not provide a clear logical foundation for this idea.",2479
"conclusions from Then, the in science. recent discoveries paper unexpected cognitive summarizes","Then, the paper summarizes unexpected conclusions from recent discoveries in cognitive science.",2480
fundamental of not mind. be logic turns the to out mechanism Classical,Classical logic turns out not to be the fundamental mechanism of mind.,2481
Mathematical evidence experimental conceptually. are and considered,Mathematical and experimental evidence are considered conceptually.,2482
logical counters arguments for reductionism. Dynamic logic,Dynamic logic counters logical arguments for reductionism.,2483
from literature widely structural in invariant model estimation. of considered current data an,data from an invariant model widely considered in current literature of structural estimation.,2484
"overcame recent advancement, dynamic these mathematical A logic (DL) difficulties. past","A recent mathematical advancement, dynamic logic (DL) overcame these past difficulties.",2485
"cognitive DL and related conceptually is and science, psychology, to philosophy. neuroscience, described","DL is described conceptually and related to neuroscience, psychology, cognitive science, and philosophy.",2486
"consciousness. functions: DL concepts, understanding, instincts, imagination, intuition, models emotions, cognitive higher","DL models higher cognitive functions: concepts, emotions, instincts, understanding, imagination, intuition, consciousness.",2487
are to Aesthetic related perception 'everyday' of the emotions functioning beauty of mind. and,Aesthetic emotions and perception of beauty are related to 'everyday' functioning of the mind.,2488
"methods in the geometric section, elaborated. involved following are In technique the state-of-the-art the","In the following section, the state-of-the-art methods involved in the geometric technique are elaborated.",2489
"human-accelerated HOR of acceleration, long-range find a We type regions. new","We find a new type of long-range acceleration, human-accelerated HOR regions.",2490
"after revealing significant evolution HOR construct showing rapid We human-chimpanzee alphoid difference, schemes human-chimpanzee separation.","We construct alphoid HOR schemes showing significant human-chimpanzee difference, revealing rapid evolution after human-chimpanzee separation.",2491
human and higher estimates. than between divergence This of chimpanzee estimate significantly is previous,This is significantly higher estimate of divergence between human and chimpanzee than previous estimates.,2492
Data make their interactions. about intermolecular inferences one protein and to obtained conformations allow thus,Data thus obtained allow one to make inferences about protein conformations and their intermolecular interactions.,2493
"the on and DMS proteins. residues of methylation lysine, can induce histidine glutamate","DMS can induce methylation of the lysine, histidine and glutamate residues on proteins.",2494
most constitute in road systems. elements dangerous Intersections the of one,Intersections constitute one of the most dangerous elements in road systems.,2495
real on world experiment. a The method is during evaluated data collected,The method is evaluated on data collected during a real world experiment.,2496
"that algorithm Spellman we we analyze call modified et SPCTF, al. this With","With this modified algorithm that we call SPCTF, we analyze Spellman et al.",2497
"still later by but genes several remain were studies, unclassified. whereas other identified","but were later identified by other studies, whereas several genes still remain unclassified.",2498
due motif-finders. of are is false-positives generally It assumed algorithmic to mostly weakness that,It is generally assumed that false-positives are mostly due to algorithmic weakness of motif-finders.,2499
Many conservative. of classes tractable CSPs well-known are,Many well-known tractable classes of CSPs are conservative.,2500
"accuracy in demonstrate Applications analysis. framework efficiency, and of shape proposed the invariance, its","Applications of the proposed framework demonstrate its invariance, efficiency, and accuracy in shape analysis.",2501
"is In case the a OPRAm, reference attached. direction of","In the case of OPRAm, a reference direction is attached.",2502
which with also called elevations. are works distances reference principle same local The,The same principle works also with local reference distances which are called elevations.,2503
references attachment. feature called to is features of point attaching principle a The hidden,The principle of attaching references features to a point is called hidden feature attachment.,2504
approach flow points together drifting. prevents This with from optical re-updating the the head tracker,This approach together with re-updating the optical flow points prevents the head tracker from drifting.,2505
occlusion a head. to of reduce the the We propose of partial influence also way,We also propose a way to reduce the influence of partial occlusion of the head.,2506
"has airborne played applications important to an simulation role. radars, In the related always","In the applications related to airborne radars, simulation has always played an important role.",2507
be collection example data. radar of multipolar the regarding pure simple will A,A simple example will be regarding the collection of pure multipolar radar data.,2508
OPR with We the found thermal relative stability. DPR and many proteins same,We found many DPR and OPR proteins with the same relative thermal stability.,2509
"lack carried on of the data, out data. of Because were simulated experiments","Because of the lack of data, experiments were carried out on simulated data.",2510
"the bistatic deteriorate substantially ATR not angle. per- increasing formance does Second, with","Second, the ATR per- formance does not deteriorate substantially with increasing bistatic angle.",2511
The results the in semantics loops. terms elementary useful into provide insights model stable of,The results provide useful insights into the stable model semantics in terms of elementary loops.,2512
"actually experimental one plausible is explanations. looking of interpretation the In data, for","In the interpretation of experimental data, one is actually looking for plausible explanations.",2513
We plausibility have. of should this measure define what characteristics,We define what characteristics this measure of plausibility should have.,2514
"games, the world. simplified video characters' of often virtual decision use representation systems In a","In video games, virtual characters' decision systems often use a simplified representation of the world.",2515
"and the The are detailed. modifications used of model, the the parameters we implementation","The implementation of the model, the modifications and the parameters we used are detailed.",2516
combining combining types skin evaluated Two strategies skin classifiers. combining were and of features namely,Two types of combining strategies were evaluated namely combining skin features and combining skin classifiers.,2517
A significant nucleotide vs. for content asymmetry in was found exons human introns. difference,A significant difference in nucleotide content asymmetry was found for human exons vs. introns.,2518
A T over C). and excess G of and,excess of G and T over A and C).,2519
of The extents these expression depend gene patterns. biases upon,The extents of these biases depend upon gene expression patterns.,2520
intronic The genes. of is highest introns skew housekeeping the found keto-amino in,The highest intronic keto-amino skew is found in the introns of housekeeping genes.,2521
one segmenting text regions of is of characters them and such Extraction challenges. into,Extraction of text regions and segmenting them into characters is one of such challenges.,2522
intrinsically unfoldome indicate the proteins. used The been has of recently term to disordered universe,The term unfoldome has been recently used to indicate the universe of intrinsically disordered proteins.,2523
"several some studies However, views. doubt about raise these","However, several studies raise some doubt about these views.",2524
"from Eukarya. Archaea, proteomes several scan We Bacteria and","We scan several proteomes from Archaea, Bacteria and Eukarya.",2525
estimates reported those previously studies. several lower Our in than are,Our estimates are lower than those previously reported in several studies.,2526
that unfoldome has Our often the been observations overestimated. idea support the of spread the,Our observations support the idea that the spread of the unfoldome has been often overestimated.,2527
both and magnitude of phase. to changes leads signal MR in Presence in iron,Presence of iron leads to changes in MR signal in both magnitude and phase.,2528
"Accordingly, appear (darker concentration iron than with MR hypo-intense high usual) tissues in contrasts.","Accordingly, tissues with high iron concentration appear hypo-intense (darker than usual) in MR contrasts.",2529
to inbreeding The also in species. interpretation the solution paradox the invasive provides proposed,The proposed interpretation also provides solution to the inbreeding paradox in the invasive species.,2530
Do exist copulas in and $R^d$? spherical circular there,Do there exist circular and spherical copulas in $R^d$?,2531
with processing. quantization many an graphics in applications operation important is Color and image,Color quantization is an important operation with many applications in graphics and image processing.,2532
data on Most are clustering algorithms. essentially methods quantization based,Most quantization methods are essentially based on data clustering algorithms.,2533
"color investigate the we paper, quantizer. as this In a of k-means performance","In this paper, we investigate the performance of k-means as a color quantizer.",2534
model fragmented integrated Yet data is the an missing. is and,Yet the data is fragmented and an integrated model is missing.,2535
been and used define relationships proteins widely pathways. to interactions Genetic have between functional,Genetic interactions have been widely used to define functional relationships between proteins and pathways.,2536
"multidomain that their interactions to genetic found in Moreover, domains differently. we proteins yeast contribute","Moreover, we found that domains in multidomain yeast proteins contribute to their genetic interactions differently.",2537
"lethal we pathways novel predicted also identified interactions. Furthermore, genetic compensatory synthetic the from had","Furthermore, we had also identified novel compensatory pathways from the predicted synthetic lethal genetic interactions.",2538
series. for a in this kernels new family of We time variable-length work propose,We propose in this work a new family of kernels for variable-length time series.,2539
uses describe matrix efficient techniques. low-rank generalization factorization that of We computationally a implementation this,We describe a computationally efficient implementation of this generalization that uses low-rank matrix factorization techniques.,2540
be normalizing the compute method used to This also probability of distributions. constant can,This method can also be used to compute the normalizing constant of probability distributions.,2541
been the open implemented margLikArrogance. has method R in The package source,The method has been implemented in the open source R package margLikArrogance.,2542
"are problems, also physiological Advantages, of model discussed. decoding and the tests","Advantages, problems, and physiological tests of the decoding model are also discussed.",2543
the implementation Indic methodologies. growing may power Rapidly CR of enable computational,Rapidly growing computational power may enable the implementation of Indic CR methodologies.,2544
techniques reviewed. DOCR of systems An presented is DOCR are overview available and the,An overview of DOCR systems is presented and the available DOCR techniques are reviewed.,2545
future are discussed research for The suggested. directions current and status of is DOCR,The current status of DOCR is discussed and directions for future research are suggested.,2546
also and to them was for algorithms Order way searching of described.,Order of algorithms and way to searching for them was also described.,2547
in the and questions ideas work further were mentioned for conclusion Some part.,Some questions and ideas for further work were mentioned in the conclusion part.,2548
functions important of the of One executive tasks. the is central two coordinate ability simultaneously,One of the important functions of central executive is the ability simultaneously coordinate two tasks.,2549
"is where However, are this there not followed. cases specification","However, there are cases where this specification is not followed.",2550
developed been applications to have problem. this in Several Malaysia overcome,Several applications have been developed in Malaysia to overcome this problem.,2551
shape with Altered chorionic functional associated surface efficiency. is placental lowered,Altered chorionic surface shape is associated with lowered placental functional efficiency.,2552
or true particularly integral as movies for This songs. is such items,This is particularly true for integral items such as movies or songs.,2553
"paper, common methodology check we to claim. a propose systematically In this this","In this paper, we propose a methodology to systematically check this common claim.",2554
"virulence of however, populations is, The viral evolution poorly in understood.","The evolution of virulence in viral populations is, however, poorly understood.",2555
mathematical a trade-off We introduce models infection. standard of viral into intra-host competition-colonization,We introduce a competition-colonization trade-off into standard mathematical models of intra-host viral infection.,2556
where simulations The a existence state the variants low-virulence steady few coexist. of suggest,The simulations suggest the existence of a steady state where few low-virulence variants coexist.,2557
"This ""psychological tested via projects. approach student electronics"" could be","This ""psychological electronics"" approach could be tested via student projects.",2558
that with Turing evolve population the model error threshold. towards the machines our Simulations show,Simulations with our model show that the Turing machines population evolve towards the error threshold.,2559
an this call presentation We individual angle. of viewpoints,We call this presentation of individual viewpoints an angle.,2560
in potentials attention-deficit event-related of (ADHD) (ERPs) children. on EEG hyperactivity Impact biofeedback,Impact of EEG biofeedback on event-related potentials (ERPs) in attention-deficit hyperactivity (ADHD) children.,2561
on ERPs is effectiveness approach this the Unfortunately still parameters of unknown.,Unfortunately the effectiveness of this approach on ERPs parameters is still unknown.,2562
to after changes the of therapy. Thus we study aimed neurofeedback ERPs,Thus we aimed to study the changes of ERPs after neurofeedback therapy.,2563
sequence-level analysis information. of genomic principled propose approach a to statistical first We,We propose a first principled approach to statistical analysis of sequence-level genomic information.,2564
Oriented traditional offer wavelet and multiscale dictionaries invariance. rotation may processing extend,Oriented multiscale dictionaries extend traditional wavelet processing and may offer rotation invariance.,2565
algorithms search simplify the to (sparse) for efficient representation. an Highly dictionaries require redundant specific,Highly redundant dictionaries require specific algorithms to simplify the search for an efficient (sparse) representation.,2566
"overview, on partially based overlapping The choice a of ""pictures"". of panorama an etymology suggests","The etymology of panorama suggests an overview, based on a choice of partially overlapping ""pictures"".",2567
The payoffs. with equation replicator resulting equivalent modified homogeneous has inhomogeneous an,The resulting inhomogeneous replicator equation has an homogeneous equivalent with modified payoffs.,2568
by surface BSDFs from mutually are independent. patches neighboring local Reflections represented,Reflections from neighboring surface patches represented by local BSDFs are mutually independent.,2569
on number RegenerativeMCMC of samples. a this based random use algorithms technique,RegenerativeMCMC algorithms based on this technique use a random number of samples.,2570
results intervals. These costruct to confidence are applied,These results are applied to costruct confidence intervals.,2571
MCMC simulation. are explicit compared schemes different nonasymptotic of for results Available,Available explicit nonasymptotic results are compared for different schemes of MCMC simulation.,2572
"entropy we After novel notions co-entropy, of the introducing examine properties. their and then","After introducing the novel notions of entropy and co-entropy, we then examine their properties.",2573
"particular, universes. co-entropies different discuss we In relationship the between of","In particular, we discuss the relationship of co-entropies between different universes.",2574
is The illustrative numerical theoretical by accompanied development examples.,The theoretical development is accompanied by illustrative numerical examples.,2575
"trait from initial conditions. the generic Gaussian distribution particular, emerges of In","In particular, Gaussian distribution of the trait emerges from generic initial conditions.",2576
are the exact these makes This two between derivation clear differences which models.,This derivation makes clear which are the exact differences between these two models.,2577
"with the case. compare results asexual we the Finally,","Finally, we compare the results with the asexual case.",2578
the to closer Intelligence. this pristine get In way goal we Artificial try of to,In this way we try to get closer to the pristine goal of Artificial Intelligence.,2579
The two on guidance superiority mechanism datasets. is confirmed this of,The superiority of this guidance mechanism is confirmed on two datasets.,2580
"incorporate etc.) this approach (lighting, is able particular, to In elevation, information invariance","In particular, this approach is able to incorporate invariance information (lighting, elevation, etc.)",2581
"common cerebral brain most supportive primary cells. Gliomas evolving the tumors, are from the","Gliomas are the most common primary brain tumors, evolving from the cerebral supportive cells.",2582
"clinical of follow-up, volume essential. evaluation the the preoperative is For tumor","For clinical follow-up, the evaluation of the preoperative tumor volume is essential.",2583
dynamic This of modeling video of proposes swarms paper the problem (DS). sequences,This paper proposes the problem of modeling video sequences of dynamic swarms (DS).,2584
"DS animals of Examples flocks nature, e.g., herds of and of birds. abound in","Examples of DS abound in nature, e.g., herds of animals and flocks of birds.",2585
demonstrate sequences. on method of video validity the conducting our by real experiments We,We demonstrate the validity of our method by conducting experiments on real video sequences.,2586
not involve function. fluctuations attenuation This in change subject's the of a psychometric does,This attenuation of fluctuations does not involve a change in the subject's psychometric function.,2587
databases. of amount grows The billions information of,The amount of information grows billions of databases.,2588
search tools specialize We known engine. need the information generically will to search,We need to search the information will specialize tools known generically search engine.,2589
"of many are retrieving available There information meaningful engines today, search difficult. is","There are many of search engines available today, retrieving meaningful information is difficult.",2590
in presented strongly that the results paper. validate theory Numerical is this,Numerical results strongly validate the theory that is presented in this paper.,2591
"angular and contrast, Flankers in varied sign. orientation, position,","Flankers varied in contrast, orientation, angular position, and sign.",2592
"rats likely collinear. when are more particular, are to target miss the flankers In","In particular, rats are more likely to miss the target when flankers are collinear.",2593
model We suggest the a a collinear generalization pattern-sensitive could deficit. for account of which,We suggest a pattern-sensitive generalization of the model which could account for a collinear deficit.,2594
that process. they We to a show can be learning generalized,We show that they can be generalized to a learning process.,2595
grained simple has in coarse performed a been model. survey The,The survey has been performed in a simple coarse grained model.,2596
such detail. the in clamp workings a elucidate of atomic We an,We elucidate the workings of such a clamp in an atomic detail.,2597
of in suggest oyster findings states populations. alternative stable the existence These reef,These findings suggest the existence of alternative stable states in oyster reef populations.,2598
gene the functional diversity. and complex within enable Highly varied programs macrophage expression such,Highly complex and varied gene expression programs within the macrophage enable such functional diversity.,2599
imperceivable for local an There some increasing and of use features redundant face recognition. is,There is an increasing use of some imperceivable and redundant local features for face recognition.,2600
proposed the effectiveness databases. methods verified is with The LFW face of challenging the,The effectiveness of the proposed methods is verified with the challenging LFW face databases.,2601
"Previous to as Central considered were this, American these species being separate. distinctly allopatric","Previous to this, these allopatric Central American species were considered as being distinctly separate.",2602
"however, structure to architecture. circuit not, is linked a The particular bifurcation","The bifurcation structure is not, however, linked to a particular circuit architecture.",2603
theory is quantum A system on conformation-electron presented.,A quantum theory on conformation-electron system is presented.,2604
in further terms dependences are potential analyzed torsion These temperature of parameters.,These temperature dependences are further analyzed in terms of torsion potential parameters.,2605
design principles networks interaction an is biology. Understanding important molecular of systems of goal molecular,Understanding design principles of molecular interaction networks is an important goal of molecular systems biology.,2606
dynamics. mechanisms network to the paper This patterns contributes the govern of that in identification,This paper contributes to the identification of patterns in the mechanisms that govern network dynamics.,2607
of The biology. philosophy cornerstone of Wright-Fisher and forms controversy the history a,The Wright-Fisher controversy forms a cornerstone of the history and philosophy of biology.,2608
incommensurable. theories in are These fact alternative,These alternative theories are in fact incommensurable.,2609
of evolutionary actual theory is model change Wright's detailed in populations. a dynamical,Wright's theory is a detailed dynamical model of evolutionary change in actual populations.,2610
"real concerned in arguments populations. stake issue of actual their at dynamics Instead, the the","Instead, the issue at stake in their arguments concerned the actual dynamics of real populations.",2611
combinations play fluctuations gene agreed central role (epistatic) a nonadditive that Both in evolution. of,Both agreed that fluctuations of nonadditive (epistatic) gene combinations play a central role in evolution.,2612
"emphasized stochastic combinations populations. fluctuations of small, isolated gene Wright in","Wright emphasized stochastic fluctuations of gene combinations in small, isolated populations.",2613
greedy dMotifGreedy was A named based algorithm developed. tool,A greedy algorithm based tool named dMotifGreedy was developed.,2614
also of the representable of class single SE-interpretations a by sets rule. characterise We,We also characterise the class of sets of SE-interpretations representable by a single rule.,2615
that are follows Two the inevitable. text of flaws,Two flaws of the text that follows are inevitable.,2616
"new of improve to principle, i.e. a efficiency algorithm, order the the In","In order to improve the efficiency of the algorithm, a new principle, i.e.",2617
"common addressed simplify literals principle, formulae. to the is","the common literals principle, is addressed to simplify formulae.",2618
common eliminate more us literals. to allows This,This allows us to eliminate more common literals.,2619
the frequent is formation Cooperation when fragmentation. and prosperous has network,Cooperation is prosperous when the network has frequent formation and fragmentation.,2620
"Services ontology. need compose Web semantically, To an we","To compose Web Services semantically, we need an ontology.",2621
several are in There of semantics inserting ways Services. Web,There are several ways of inserting semantics in Web Services.,2622
of languages using of OWL-S. consists One description like them,One of them consists of using description languages like OWL-S.,2623
problem is concern. widespread alignment of The economic business-IT of,The problem of business-IT alignment is of widespread economic concern.,2624
"indexed are Google search engines. other in are the rules and they by English, Since","Since the rules are in English, they are indexed by Google and other search engines.",2625
mind. This looking when rules task for useful that one a for in is has,This is useful when looking for rules for a task that one has in mind.,2626
"One run rules, again a can using the browser. then","One can then run the rules, again using a browser.",2627
cutoff frequencies biophysical The origin the understood. observed high is well not thus of,The biophysical origin of the observed high cutoff frequencies is thus not well understood.,2628
images the from skin is experiments collected internet. on method The by extensive evaluated proposed,The proposed method is evaluated by extensive experiments on the skin images collected from internet.,2629
i.e. dataset is The tested three using classifiers different,The dataset is tested using three different classifiers i.e.,2630
not special imaging and is technique cost low It any require does devices. a,It is a low cost technique and does not require any special imaging devices.,2631
"study effects noise sensing of and on the we behaviour. the motor signalling Here,","Here, we study the effects of sensing and signalling noise on the motor behaviour.",2632
analysis can be A comprehensive signalling pathways. applied to other similar two-component,A similar comprehensive analysis can be applied to other two-component signalling pathways.,2633
"a both population, games (PGG) communities. occur community-structured within public between and goods In","In a community-structured population, public goods games (PGG) occur both within and between communities.",2634
PGG games as is Such of (MPGG). referred type goods multilevel public,Such type of PGG is referred as multilevel public goods games (MPGG).,2635
noise fusion a in free Multiframe obtain performed single order to image. is,Multiframe fusion is performed in order to obtain a single noise free image.,2636
compute evaluated rules they fuzzy Index. Further to are using Criticality,Further they are evaluated using fuzzy rules to compute Criticality Index.,2637
"buses Based critical ranked. index, are this on","Based on this index, critical buses are ranked.",2638
on Bus tested Test method System. proposed is Five The,The proposed method is tested on Five Bus Test System.,2639
"non-redundant consistently interactome  Robust datasets. advances comprehensive, analysis  and annotated demand in","  Robust advances in interactome analysis demand comprehensive, non-redundant and consistently annotated datasets.",2640
"by often the the of different disagree. repositories In same result addition, annotations experimental","In addition, the annotations of the same experimental result by different repositories often disagree.",2641
"and three human yeast, we the have fruitfly. datasets: As organismal an largest illustration, processed","As an illustration, we have processed the three largest organismal datasets: yeast, human and fruitfly.",2642
optimal of NP-complete. structure problem is The finding coalition the,The problem of finding the optimal coalition structure is NP-complete.,2643
and associated techniques give several their properties. state distances to We aggregate,We give several techniques to aggregate state distances and their associated properties.,2644
two developed our serve test-beds We as for planners investigation. to,We developed two planners to serve as test-beds for our investigation.,2645
"planner regression CAltAlt, search. uses that first, is A* a conformant The","The first, CAltAlt, is a conformant regression planner that uses A* search.",2646
"uses progression AO* that is The planner conditional a POND, second, search.","The second, POND, is a conditional progression planner that uses AO* search.",2647
relative We heuristic effectiveness techniques the our within these of show planners.,We show the relative effectiveness of our heuristic techniques within these planners.,2648
"verification it So, computerized a signature system. have is important very to","So, it is very important to have a computerized signature verification system.",2649
is a system. network classifier as for this used Neural,Neural network is used as a classifier for this system.,2650
"algorithm. and provably we optimisation efficient convergent an Furthermore, derive","Furthermore, we derive an efficient and provably convergent optimisation algorithm.",2651
that incremental is training sequences our demonstrate effective. with approach to Experiments learning two,Experiments with two training sequences demonstrate that our approach to incremental learning is effective.,2652
neural (e.g. during risk tumor deficits as minimize of surgery the resection brain To,To minimize the risk of neural deficits during brain surgery as tumor resection (e.g.,2653
adaptation unclear. functional The are consequences of,The functional consequences of adaptation are unclear.,2654
net the of is and adaptation find that We subtle effect heterogeneous.,We find that the net effect of adaptation is subtle and heterogeneous.,2655
"on adaptation coding and test Depending either adaptation or mechanism can decrease stimulus, increase accuracy.","Depending on adaptation mechanism and test stimulus, adaptation can either increase or decrease coding accuracy.",2656
"et obtained levels, by answers were al. partial For higher Iersel","For higher levels, partial answers were obtained by Iersel et al.",2657
"results and are cognitive linguistics, to science, psychology. related Mathematical","Mathematical results are related to cognitive science, linguistics, and psychology.",2658
arguments theoretical and evidence are discussed. Experimental,Experimental evidence and theoretical arguments are discussed.,2659
are Approximate of minds human evolution cultures obtained. equations for and,Approximate equations for evolution of human minds and cultures are obtained.,2660
future directions Unsolved and are discussed. theoretical problems and experimental,Unsolved problems and future theoretical and experimental directions are discussed.,2661
of plays Immune role in a survival living being. vital all system the Natural the,Natural Immune system plays a vital role in the survival of the all living being.,2662
of instance. value Each the an attribute table for cell contains an,Each cell of the table contains an attribute value for an instance.,2663
"three-dimensional the segmentation these is structures closed, as of object Therefore, necessary.","Therefore, the segmentation of these structures as closed, three-dimensional object is necessary.",2664
easily. consecutive Animals learn timing the between events very,Animals learn the timing between consecutive events very easily.,2665
the for (Weber's to usually timing). proportional Their precision is law to interval time,Their precision is usually proportional to the interval to time (Weber's law for timing).,2666
'what'. a instead of learn simple We rule develop to geometric 'when',We develop a simple geometric rule to learn 'when' instead of 'what'.,2667
"for sensing from separated creativity. Representations ""internal room"" acting and gave","Representations separated sensing from acting and gave ""internal room"" for creativity.",2668
"KI compare and Kahneman-Tversky's heuristic decision creative with consider making, vs. We non-creative thinking.","We consider creative vs. non-creative decision making, and compare KI with Kahneman-Tversky's heuristic thinking.",2669
At time of grows pathways bound. number accessible without the same the alternative,At the same time the number of alternative accessible pathways grows without bound.,2670
solvers. decentralized CSP term We these algorithms,We term these algorithms decentralized CSP solvers.,2671
extract method in We these areas. propose an to paper hybrid this,We propose in this paper an hybrid method to extract these areas.,2672
The by is tools. then use extraction areas of mathematical carried the some out morphology,The areas extraction is then carried out by the use of some mathematical morphology tools.,2673
approach an computational Knowledge to tackle general problems. of compilation is the reasoning intractability,Knowledge compilation is an approach to tackle the computational intractability of general reasoning problems.,2674
decision languages. is diagram Reduced target one most the ordered (ROBDD) of binary influential,Reduced ordered binary decision diagram (ROBDD) is one of the most influential target languages.,2675
"and compilation compilation we for an ROBDD-i a algorithm algorithm. i any ROBDD-\infty Finally, propose","Finally, we propose an ROBDD-i compilation algorithm for any i and a ROBDD-\infty compilation algorithm.",2676
"maps. been the based blueprint has on map developed, The architectural cognitive","The cognitive map has been developed, based on the architectural blueprint maps.",2677
"distance Canberra phase, is In identification used. measure","In identification phase, Canberra distance measure is used.",2678
with compared The proposed discrete method (DWT). wavelet transform is,The proposed method is compared with discrete wavelet transform (DWT).,2679
Long-distance genes in between target commonplace their and higher enhancers interactions regulatory are eukaryotes.,Long-distance regulatory interactions between enhancers and their target genes are commonplace in higher eukaryotes.,2680
these long block interactions. boundaries are Interposed to able or regulatory insulators distance,Interposed boundaries or insulators are able to block these long distance regulatory interactions.,2681
unclear. relates basis activity it how and to insulator for enhancer action-at-a-distance The remains mechanistic,The mechanistic basis for insulator activity and how it relates to enhancer action-at-a-distance remains unclear.,2682
"A* runs that often researchers several Recently, into problems. with have satisficing search cost-based found","Recently, several researchers have found that cost-based satisficing search with A* often runs into problems.",2683
with We satisficing search this sized-based show immune that largely evaluation to functions problem. is,We show that satisficing search with sized-based evaluation functions is largely immune to this problem.,2684
 and rules coupled in are usually Ontologies representation  formalisms. knowledge loosely,  Ontologies and rules are usually loosely coupled in knowledge representation formalisms.,2685
To and . (TPLP) Practice Theory Programming of appear Logic in,To appear in Theory and Practice of Logic Programming (TPLP) .,2686
as are of models computationally often (GP) mathematical experiments. used Process approximations Gaussian expensive,Gaussian Process (GP) models are often used as mathematical approximations of computationally expensive experiments.,2687
"$\varepsilon_{N,m}$ random as $d\rightarrow\infty$ and In to converges variable this a $\lim_{m\to\infty}\varepsilon_{N,m}=N$. ESS case,","In this case, ESS converges to a random variable $\varepsilon_{N,m}$ as $d\rightarrow\infty$ and $\lim_{m\to\infty}\varepsilon_{N,m}=N$.",2688
Understanding age open long-lived is in a biology. evolutionary problem we why,Understanding why we age is a long-lived open problem in evolutionary biology.,2689
"to change, extinction. drive conditions competitors a senescent immortal species When can","When conditions change, a senescent species can drive immortal competitors to extinction.",2690
death the pruning by from counter-intuitive arises This caused individuals. the result elder of,This counter-intuitive result arises from the pruning caused by the death of elder individuals.,2691
can those eliminate pool. genetic Senescence from the,Senescence can eliminate those from the genetic pool.,2692
changing allows each to to It lineage adapt conditions. faster,It allows each lineage to adapt faster to changing conditions.,2693
can constraints Input tags. categorized with be,Input constraints can be categorized with tags.,2694
unexpected strategy results Our procedure. this for consequences that results theoretical show the in,Our results show that this strategy results in unexpected theoretical consequences for the procedure.,2695
least inequality squares equality problems and involve affine constraints. Many,Many least squares problems involve affine equality and inequality constraints.,2696
"the penalty recovers the constrained constant solution. In the $\infty$, limit tends to one as","In the limit as the penalty constant tends to $\infty$, one recovers the constrained solution.",2697
"path the along, exits slides In the from solution process, the constraints. various hits, and","In the process, the solution path hits, slides along, and exits from the various constraints.",2698
"of is both solution revealing. path the inspection settings, In entire","In both settings, inspection of the entire solution path is revealing.",2699
illustrate path mechanics few and A chosen well the of following. potential examples,A few well chosen examples illustrate the mechanics and potential of path following.,2700
propagation constraint. algorithm enforces that this bounds efficient global We on consistency give an,We give an efficient propagation algorithm that enforces bounds consistency on this global constraint.,2701
"on general. constraint is enforcing domain we in Finally, that consistency global prove NP-hard this","Finally, we prove that enforcing domain consistency on this global constraint is NP-hard in general.",2702
"surrounding understanding, the the discussed. to are also brain and environment of the familiarity Image","Image understanding, and the familiarity of the brain to the surrounding environment are also discussed.",2703
inspirations. We discussion a give also of,We also give a discussion of inspirations.,2704
imposed at network. Appropriate the outlet and river of nodes conditions are the upstream boundary,Appropriate boundary conditions are imposed at the outlet and upstream nodes of the river network.,2705
"(LSIR). least-squares we propose novel paper, called regression this causal In a independence algorithm inference","In this paper, we propose a novel causal inference algorithm called least-squares independence regression (LSIR).",2706
analyzed. unclear were The and marked features,The unclear features were marked and analyzed.,2707
suitable the of and these be seem to techniques approaches images. Sauvola's for Niblack's types,Niblack's and Sauvola's techniques seem to be the suitable approaches for these types of images.,2708
methods show Most these with character two for and recognition. of readability binarized images improvement,Most of binarized images with these two methods show improvement for readability and character recognition.,2709
sharp to utilize the obtain localization rate. techniques learning We,We utilize localization techniques to obtain the sharp learning rate.,2710
bound of kernels. rate the of The the characterized the associated by eigenvalues decay is,The bound is characterized by the decay rate of the eigenvalues of the associated kernels.,2711
gives decay rate faster a rate. convergence larger A,A larger decay rate gives a faster convergence rate.,2712
"overall results of regularization elastic-net MKL. the our use in suggest theoretical Thus,","Thus, our theoretical results overall suggest the use of elastic-net regularization in MKL.",2713
"useful segmentation. can For it particularly be and example, detection in edge","For example, it can be particularly useful in edge detection and segmentation.",2714
a divergence to use history. calibrate estimation The of long fossil evidence time has,The use of fossil evidence to calibrate divergence time estimation has a long history.,2715
analytical We a on results this number trees. with illustrate of small,We illustrate this with a number of analytical results on small trees.,2716
"solution practical elusive. Unfortunately, for remains calibrations multiple a","Unfortunately, a practical solution for multiple calibrations remains elusive.",2717
continuous Leslie-Gower We model. global properties time food chain of asymptotic study a,We study global asymptotic properties of a continuous time Leslie-Gower food chain model.,2718
the the this We package YADAS. in algorithm discuss software of implementation,We discuss the implementation of this algorithm in the software package YADAS.,2719
"the structure. ""manifold"" of the the to word topology properties Here, refers","Here, the word ""manifold"" refers to the topology properties of the structure.",2720
pursuit projection a built approximating is algorithm. by manifold The,The approximating manifold is built by a projection pursuit algorithm.,2721
"of of step dimension algorithm, incremented. the the the each manifold is At","At each step of the algorithm, the dimension of the manifold is incremented.",2722
"in Moreover, of the established converges is algorithm a also number it finite that steps.","Moreover, it is also established that the algorithm converges in a finite number of steps.",2723
"that, optimization numerous no We cases, is procedure required. show in","We show that, in numerous cases, no optimization procedure is required.",2724
are simulated on presented. illustrations real Some data and,Some illustrations on simulated and real data are presented.,2725
Network We Ecological Analysis these applied resource models. homogenization for calculate to,We applied Ecological Network Analysis to calculate resource homogenization for these models.,2726
results our evaluated robustness We in two the ways. further of,We further evaluated the robustness of our results in two ways.,2727
problems many Phase complex been the combinational have studied past transitions in widely decade. in,Phase transitions in many complex combinational problems have been widely studied in the past decade.,2728
"non-verbal interpersonal cues, in Facial relations. important role expressions which convey play an","Facial expressions convey non-verbal cues, which play an important role in interpersonal relations.",2729
science. may behavioural also It used be in,It may also be used in behavioural science.,2730
the measure endowed without classifiers. distance This face using the method recognition conventional,This method endowed the face recognition without using the conventional distance measure classifiers.,2731
best building from are results The category. achieved,The best results achieved are from building category.,2732
"computation results show performance experimental The detection in that both efficiency. ""DURF+ERT+SVM"" outperforms ""SIFT+ERT+SVM"" and","The experimental results show that ""DURF+ERT+SVM"" outperforms ""SIFT+ERT+SVM"" both in detection performance and computation efficiency.",2733
"results Besides, performance. DURF SIFT in better even detection and combining","Besides, combining DURF and SIFT results in even better detection performance.",2734
"on RANSAC simple tried Real-time SIFT objects, and object is e.g. detection also using","Real-time object detection using SIFT and RANSAC is also tried on simple objects, e.g.",2735
"good achieved. and result drink is can,","drink can, and good result is achieved.",2736
tools in  decreasing  effort fields varied search Heuristics of in AI. are crucial,  Heuristics are crucial tools in decreasing search effort in varied fields of AI.,2737
get is to proposed A image features' geometric information. new method,A new method is proposed to get image features' geometric information.,2738
can quantitatively of A given is performance that be definition of terms measured. in intelligence,A definition of intelligence is given in terms of performance that can be quantitatively measured.,2739
"of achieve intelligent features. exhibit To kinds have that introduced we several goal, this agents","To achieve this goal, we have introduced several kinds of agents that exhibit intelligent features.",2740
"Report agent, are agent, These and agent the Watcher agent. internal External Management Agent,","These are the Management agent, internal agent, External Agent, Watcher agent and Report agent.",2741
Metrics Vehicle Automatic of for are Checking Agent performance suggested and measurements (VCA). the evaluating,Metrics and measurements are suggested for evaluating the performance of Automatic Vehicle Checking Agent (VCA).,2742
Calibrate systems. to the facilities intelligent development suggested test and data are facilitate of,Calibrate data and test facilities are suggested to facilitate the development of intelligent systems.,2743
recognition most the Template of worldwide. methods prevalent is one matching pattern,Template matching is one of the most prevalent pattern recognition methods worldwide.,2744
most fields. has concept uses detection in visual found It,It has found uses in most visual concept detection fields.,2745
"organization In important of indexing image cases, such and retrieval efficient for becomes images.","In such cases, image indexing becomes important for efficient organization and retrieval of images.",2746
based system features proposed curvelet extraction. The uses a texture,The proposed system uses a curvelet based texture features extraction.,2747
describe common several of we models. paper in graphical this such approaches framework In the,In this paper we describe several such approaches in the common framework of graphical models.,2748
rule a based fuzzy classifier is of Design proposed.,Design of a fuzzy rule based classifier is proposed.,2749
the poses of computational from important an point This challenge view.,This poses an important challenge from the computational point of view.,2750
"and peripheral boundary texture. includes The limbs, structure protrusions,","The peripheral structure includes limbs, protrusions, and boundary texture.",2751
"our different method, images evaluate is a diffusion-weighted used. of set vendors To from","To evaluate our method, a set of diffusion-weighted images from different vendors is used.",2752
problem the context in well of a Regularization is neural networks. studied,Regularization is a well studied problem in the context of neural networks.,2753
knots. paths Polymers closure and open can polygonal as be their generates modeled,Polymers can be modeled as open polygonal paths and their closure generates knots.,2754
reduction computation. a scheme multi-component polynomial Our approach with couples the,Our approach couples a multi-component reduction scheme with the polynomial computation.,2755
"descriptions scale, viewpoint to changes and rotation, small The are in invariant articulations. resulting","The resulting descriptions are invariant to scale, rotation, small changes in viewpoint and articulations.",2756
representation skeletons. not from connected common The suffer the problems of traditional instability does the,The representation does not suffer from the common instability problems of the traditional connected skeletons.,2757
the shapes well. qualities It perceptual captures of,It captures the perceptual qualities of shapes well.,2758
among the the shapes Therefore becomes finding and easier. differences similarities,Therefore finding the similarities and the differences among shapes becomes easier.,2759
The the organism. of behind motor complexity the an metabolism is biological,The metabolism is the motor behind the biological complexity of an organism.,2760
on fundamental based standard captured randomization. These by cannot be null-models factors,These fundamental factors cannot be captured by standard null-models based on randomization.,2761
metabolism the arrangement of interpret modular as a functionality. variegated trace of evolved We more,We interpret the more variegated modular arrangement of metabolism as a trace of evolved functionality.,2762
latter This has tunneling. stochastic termed process been,This latter process has been termed stochastic tunneling.,2763
population has work focused on Moran model Previous genetics. of the,Previous work has focused on the Moran model of population genetics.,2764
also for results numerical efficiently to how populations. I finite obtain show,I also show how to efficiently obtain numerical results for finite populations.,2765
We through programming solve translation satisfaction constraint answer to (ASP). problems set,We solve constraint satisfaction problems through translation to answer set programming (ASP).,2766
computational demonstrate the Experiments approach. value of this,Experiments demonstrate the computational value of this approach.,2767
"energy variational functional. First, the classic we extend chamfer-matching to a","First, we extend the classic chamfer-matching energy to a variational functional.",2768
"can a meshless deformations. we deformation model Secondly, high-curvature significant that introduce handle","Secondly, we introduce a meshless deformation model that can handle significant high-curvature deformations.",2769
is penalizing between field fields. local regularized global deformation by The inconsistencies,The global deformation field is regularized by penalizing inconsistencies between local fields.,2770
"minimizing registration combined chamfer-energy consistency the functional is with achieved by a regularizer. Finally, variational","Finally, registration is achieved by minimizing a variational chamfer-energy functional combined with the consistency regularizer.",2771
number on of effectiveness experiments. the a We method demonstrate our of,We demonstrate the effectiveness of our method on a number of experiments.,2772
paper this auto-encoders. approach novel in present We training a deterministic for,We present in this paper a novel approach for training deterministic auto-encoders.,2773
Similar the of cycle also lengths. optimize nature noise multimodal the levels,Similar noise levels also optimize the multimodal nature of the cycle lengths.,2774
is selection acceptable A arguments. issue systems main argumentation of abstract of the for sets,A main issue for abstract argumentation systems is the selection of acceptable sets of arguments.,2775
"(J. as introduced argumentation, Bench-Capon by Value-based","Value-based argumentation, as introduced by Bench-Capon (J.",2776
"computationally accepted, are argument is subjectively an problems. or respectively, objectively whether intractable Deciding","Deciding whether an argument is subjectively or objectively accepted, respectively, are computationally intractable problems.",2777
methodology numerical The resulting with specific experiments three models. on illustrated is,The resulting methodology is illustrated with numerical experiments on three specific models.,2778
"property. is broken-triangle First, tractable of a basic class by binary using the QCSPs identified","First, a basic tractable class of binary QCSPs is identified by using the broken-triangle property.",2779
"for class, QCSPs. identify more class Finally, extendable min-of-max the a tractable generalized i.e., we","Finally, we identify a more generalized tractable class, i.e., the min-of-max extendable class for QCSPs.",2780
"models them estimates new estimating accurately more separately. the method the than simulations, In","In simulations, the new method estimates the models more accurately than estimating them separately.",2781
computational a goal macromolecules of long-sought structure biophysics. prediction Atomic-accuracy of is,Atomic-accuracy structure prediction of macromolecules is a long-sought goal of computational biophysics.,2782
"methods can resulting The are easy various efficient, implement, to accommodate types of and constraints.","The resulting methods are easy to implement, efficient, and can accommodate various types of constraints.",2783
complete focus on planners current plans. domain and Most models generating assume correct,Most current planners assume complete domain models and focus on generating correct plans.,2784
"error-prone and Unfortunately, a laborious domain modeling is task.","Unfortunately, domain modeling is a laborious and error-prone task.",2785
"We showing Probabilistic-FF, state-of-the-art planner, the results of our with approach. present a promise experimental","We present experimental results with Probabilistic-FF, a state-of-the-art planner, showing the promise of our approach.",2786
"combined By the sustained is of demographic understood. and effect contrast, fluctuations poorly recombination","By contrast, the combined effect of recombination and sustained demographic fluctuations is poorly understood.",2787
formulas. for task propositional of to propositional A compute is models logic fundamental,A fundamental task for propositional logic is to compute models of propositional formulas.,2788
for solvers. this Programs are called task developed satisfiability,Programs developed for this task are called satisfiability solvers.,2789
"In paper, this we consider a structured regularization. sparse hierarchical","In this paper, we consider a sparse hierarchical structured regularization.",2790
digital surgeon This to automatically. acetabular technique recognise implant a enables size the,This technique enables the surgeon to recognise a digital acetabular implant size automatically.,2791
presents and approach machine This cells families. paper for part novel obtaining and a new,This paper presents a new and novel approach for obtaining machine cells and part families.,2792
"the is cube"". easily-understood an Rubik's is Cube which originally ""magic puzzle, called","Rubik's Cube is an easily-understood puzzle, which is originally called the ""magic cube"".",2793
"planning problem, studied has a a is It time. well-known which for long been","It is a well-known planning problem, which has been studied for a long time.",2794
This solvers studies puzzle. to are modern whether this applicable SAT paper,This paper studies whether modern SAT solvers are applicable to this puzzle.,2795
Using efficient cannot only encoding puzzle. solve this,Using only efficient encoding cannot solve this puzzle.,2796
are new demonstrate our efficient. empirical and technique results SAT both solving The translation,The empirical results demonstrate both our SAT translation and new solving technique are efficient.,2797
"Library. with memory requirements These Philips compared (in are original implementations costs) Nexperia and speed,","These original implementations are compared (in speed, memory requirements and costs) with Philips Nexperia Library.",2798
periodic stimulated train. multiple fire on Neurons timescales when irregularly pulse a with,Neurons fire irregularly on multiple timescales when stimulated with a periodic pulse train.,2799
This intrinsic stochasticity? Does this raises significant two irregularity imply questions:,This raises two questions: Does this irregularity imply significant intrinsic stochasticity?,2800
long to at neuron existing Can be models extended readily describe behavior timescales?,Can existing neuron models be readily extended to describe behavior at long timescales?,2801
"and questions, we to Thus ""no"" above answer respectively. the ""yes""","Thus we answer ""yes"" and ""no"" to the above questions, respectively.",2802
speciation. the sympatric of problem This is,This is the problem of sympatric speciation.,2803
"However, not mating entirely need random. be","However, mating need not be entirely random.",2804
are probabilities the in rare of of in interested computer events experiments. bounding context We,We are interested in bounding probabilities of rare events in the context of computer experiments.,2805
on events variables. the output of physical with rare a input These random depend model,These rare events depend on the output of a physical model with random input variables.,2806
deal sampling then importance on to difficulty a We strategy with propose methods. this based,We then propose a strategy to deal with this difficulty based on importance sampling methods.,2807
is Kriging into the due account. properly step taken The variability to metamodeling,The variability due to the Kriging metamodeling step is properly taken into account.,2808
"study case Finally, challenging analyzed. a is real","Finally, a challenging real case study is analyzed.",2809
model. consists standard generative `spike-and-slab' of a Gaussian The underlying model and prior noise a,The underlying generative model consists of a standard `spike-and-slab' prior and a Gaussian noise model.,2810
by Closed-form solutions M-step and probabilistic generalizing are equations derived for PCA. E-,Closed-form solutions for E- and M-step equations are derived by generalizing probabilistic PCA.,2811
resulting a account. algorithm potentially posterior of can take into all multi-modal The EM modes,The resulting EM algorithm can take all modes of a potentially multi-modal posterior into account.,2812
cost of The the scales with algorithm exponentially hidden computational of the number dimensions.,The computational cost of the algorithm scales exponentially with the number of hidden dimensions.,2813
range the source model tasks. can to applied be separation of typical the Thus,Thus the model can be applied to the typical range of source separation tasks.,2814
"which To he this species random. interact large prove point, at networks analyzed in","To prove this point, he analyzed large networks in which species interact at random.",2815
"have pairs systems However, natural mutualistic predator-prey, (e.g., well-defined interactions or competitive). of in species","However, in natural systems pairs of species have well-defined interactions (e.g., predator-prey, mutualistic or competitive).",2816
all stability analytic provide criteria cases. We for,We provide analytic stability criteria for all cases.,2817
results ecology. broad in applicability These have,These results have broad applicability in ecology.,2818
"mutualistic stability is by in negatively impacted nestedness Similarly, networks. bipartite","Similarly, stability is negatively impacted by nestedness in bipartite mutualistic networks.",2819
pages. World of no longer consists just Wide HTML The Web,The World Wide Web no longer consists just of HTML pages.,2820
"the grows top On a common structured Semantic of as Web data XML, source.","On top of XML, the Semantic Web grows as a common structured data source.",2821
"we In detail. explain first of this these developments work, in each","In this work, we first explain each of these developments in detail.",2822
"domains. the also we in these the On avenues current research illustrate way,","On the way, we also illustrate the current research avenues in these domains.",2823
inferring cloud data. point methodology underlying topology the for A proposed is,A methodology is proposed for inferring the topology underlying point cloud data.,2824
"models. Gaussian an popular, extremely  flexible class  are models Latent of","  Latent Gaussian models are an extremely popular, flexible class of models.",2825
"however, tricky these is, for and consuming. inference models Bayesian time","Bayesian inference for these models is, however, tricky and time consuming.",2826
"we In the approximation this INLA and related outline its R package. paper,","In this paper, we outline the INLA approximation and its related R package.",2827
"factorization. and of finite models, It mixture conjugate calculation properties the families exploits","It exploits the calculation properties of finite mixture models, conjugate families and factorization.",2828
"certain This ability, extremely situations, is useful in most in alternative not methods. found","This ability, extremely useful in certain situations, is not found in most alternative methods.",2829
of such examples procedure Markov replica construct the a hidden for We models.,We construct the examples of such a replica procedure for hidden Markov models.,2830
and based describes parts. classifies data pattern on recognition features relationships Structural and of the,Structural pattern recognition describes and classifies data based on the relationships of features and parts.,2831
"structure number, dimension. of the characterize Topological objects like invariants, of any Euler the","Topological invariants, like the Euler number, characterize the structure of objects of any dimension.",2832
provide invariants a more to homology. refined can topological space Cohomology does algebraic than,Cohomology can provide more refined algebraic invariants to a topological space than does homology.,2833
of dimension. to homology used the chains characterize holes assigns any in to It `quantities',It assigns `quantities' to the chains used in homology to characterize holes of any dimension.,2834
and invariant scanning An is extension obtain given. to cocycles rotation,An extension to obtain scanning and rotation invariant cocycles is given.,2835
the to architecture: reasoning. perform introduce cognitive Xapagy designed a software system We narrative,We introduce the Xapagy cognitive architecture: a software system designed to perform narrative reasoning.,2836
smoother substantial a has a base bias. resulting The variance small but,The resulting base smoother has a small variance but a substantial bias.,2837
"in the their Then, package implementation. the rules available explain and stopping we","Then, we explain the stopping rules available in the package and their implementation.",2838
modeling has Hepatitis paper. dynamics A this presented (HCV) in been of C Virus mathematical,A mathematical modeling of Hepatitis C Virus (HCV) dynamics has been presented in this paper.,2839
computed on complexes. has traditionally Cohomology simplicial been ring,Cohomology ring has been traditionally computed on simplicial complexes.,2840
decisions suggest may lead and belief to A wrong actions. false harmful,A false belief may suggest wrong decisions and lead to harmful actions.,2841
as So beliefs to false as important goal possible. make unlikely is an,So an important goal is to make false beliefs as unlikely as possible.,2842
in forms cancer women Uterine of Cervical Cancer the is one most worldwide. of common,Uterine Cervical Cancer is one of the most common forms of cancer in women worldwide.,2843
"images form. in cervigrams Colposcopy, During acquired are raw colposcopic or Digital","During Digital Colposcopy, colposcopic images or cervigrams are acquired in raw form.",2844
about of half The image. raw cervix region occupies the cervigram,The cervix region occupies about half of the raw cervigram image.,2845
the of cervix. identification the tissues This information automatic irrelevant can confuse within,This irrelevant information can confuse automatic identification of the tissues within the cervix.,2846
a the solution five-point give problem. well-known pose relative to new algorithmic We,We give a new algorithmic solution to the well-known five-point relative pose problem.,2847
cubic deal an with famous the matrix. not does on approach essential constraint Our,Our approach does not deal with the famous cubic constraint on an essential matrix.,2848
induced. of can nucleation complexes or Scaffold-mediated protein either constitutive be,Scaffold-mediated nucleation of protein complexes can be either constitutive or induced.,2849
the resolved before must make This towards goals. can achieving plans be its crisis agent,This crisis must be resolved before the agent can make plans towards achieving its goals.,2850
a is forgery. of splicing Image form common image,Image splicing is a common form of image forgery.,2851
no clues alterations of tampering. may visual leave Such,Such alterations may leave no visual clues of tampering.,2852
camera are are characteristic and inherent Such unique. camera from constant manufacturing properties processes,Such constant camera characteristic properties are inherent from camera manufacturing processes and are unique.,2853
used. the from classification algorithm is independent proposal The,The proposal is independent from the classification algorithm used.,2854
"""Ariadne's planning, approach called to algorithm"". a We clew path the present new","We present a new approach to path planning, called the ""Ariadne's clew algorithm"".",2855
target. builds looks Explore the while the space accessible Search representation a for of,Explore builds a representation of the accessible space while Search looks for the target.,2856
found any about show without second a results one pre-processing. that Experimental path in is,Experimental results show that a path is found in about one second without any pre-processing.,2857
We Wheel' Optimization term general we describe approach (SWO). a optimization which to `Squeaky,We describe a general approach to optimization which we term `Squeaky Wheel' Optimization (SWO).,2858
"cycle until solution is acceptable limit reached, continues found. or some is an Construct/Analyze/Prioritize This","This Construct/Analyze/Prioritize cycle continues until some limit is reached, or an acceptable solution is found.",2859
and two prioritizations. be viewed spaces: can operating search on SWO as solutions,SWO can be viewed as operating on two search spaces: solutions and prioritizations.,2860
"prioritizations Similarly, constructing successive generated solutions. are analyzing by and","Similarly, successive prioritizations are generated by constructing and analyzing solutions.",2861
"discuss. This which `coupled search' has properties, we interesting some","This `coupled search' has some interesting properties, which we discuss.",2862
disadvantages. Top-down approaches have and specific advantages and each bottom-up theorem proving,Top-down and bottom-up theorem proving approaches each have specific advantages and disadvantages.,2863
"prover a are subgoal A they top-down generates by clauses, prover. bottom-up processed then","A top-down prover generates subgoal clauses, they are then processed by a bottom-up prover.",2864
the use second The deals in bottom-up a lemmas with top-down technique of prover. generated,The second technique deals with the use of bottom-up generated lemmas in a top-down prover.,2865
the areas We and model of apply to elimination superposition. our concept,We apply our concept to the areas of model elimination and superposition.,2866
of the article investigate default logic. latter formalizations we in Reiter's this approach In,In this article we investigate formalizations of the latter approach in Reiter's default logic.,2867
both decidable. of reasoning is is proposal in notable our that feature cases A,A notable feature of our proposal is that reasoning in both cases is decidable.,2868
concurrently explore AntNet's and exchange agents network the information. collected,AntNet's agents concurrently explore the network and exchange collected information.,2869
"asynchronous, is mediated the among the communication agents by The network indirect itself. and","The communication among the agents is indirect and asynchronous, mediated by the network itself.",2870
is typical of social stigmergy. communication form and insects of is called This,This form of communication is typical of social insects and is called stigmergy.,2871
testbeds. of The algorithms' evaluated over a performance set realistic is,The algorithms' performance is evaluated over a set of realistic testbeds.,2872
under conditions experimental performance all with AntNet to superior its respect competitors. the showed,AntNet showed superior performance under all the experimental conditions with respect to its competitors.,2873
the methods %existing previous significantly approximation that these method. known outperform Experiments show best,Experiments show that these methods significantly outperform the previous best known %existing approximation method.,2874
optimal almost randomly are to generated tested. all find in able manipulations elections the We,We are able to find optimal manipulations in almost all the randomly generated elections tested.,2875
"logic automatic extensive conducted of programs. Recently, control research generation has for on been","Recently, extensive research has been conducted on automatic generation of control for logic programs.",2876
novel this paper a automatic for algorithm of finding lowest-cost subgoal orderings. In present we,In this paper we present a novel algorithm for automatic finding of lowest-cost subgoal orderings.,2877
strategy. The divide-and-conquer works using algorithm the,The algorithm works using the divide-and-conquer strategy.,2878
"provably a optimal are yielding subsets and recursively order. The ordered merged,","The subsets are ordered recursively and merged, yielding a provably optimal order.",2879
variational a in We models. inference approximation describe probabilistic method large-scale for efficient,We describe a variational approximation method for efficient inference in large-scale probabilistic models.,2880
to probabilities of interest. that deterministic procedures Variational marginal provide and are conditional approximations methods,Variational methods are deterministic procedures that provide approximations to marginal and conditional probabilities of interest.,2881
sampling or inference provide stochastic search. on methods based to approximate They alternatives,They provide alternatives to approximate inference methods based on stochastic sampling or search.,2882
a knowledge. and network probabilistic statistical built is model large-scale graphical expert on QMR The,The QMR network is a large-scale probabilistic graphical model built on statistical and expert knowledge.,2883
The to of conditional comparative disprove Fine same result counterexample is a on used probability.,The same counterexample is used to disprove a result of Fine on comparative conditional probability.,2884
from quite simpler different those are the answers in framework. The classical,The answers are quite different from those in the simpler classical framework.,2885
"formulae. planning Boolean to we quantified conditional Instead, translate","Instead, we translate conditional planning to quantified Boolean formulae.",2886
that simple recover of proofs most previous results. provide the We,We provide simple proofs that recover most of the previous results.,2887
for game. We upper propose bounds semi-bandit new the,We propose new upper bounds for the semi-bandit game.,2888
three for derive all we Moreover feedback lower bounds assumptions.,Moreover we derive lower bounds for all three feedback assumptions.,2889
"answer et asked we al. a by Koolen Finally, question","Finally, we answer a question asked by Koolen et al.",2890
systems suggests posses that capacities. different chemical Our semantic different study,Our study suggests that different chemical systems posses different semantic capacities.,2891
Hypotheses origin evolution life. and concern of the,Hypotheses concern the origin and evolution of life.,2892
corrections without incurring terms the on Higher-order cost. much variational approach describe further computational,Higher-order terms describe corrections on the variational approach without incurring much further computational cost.,2893
to is elucidated. approaches as also other such relationship perturbational TAP The,The relationship to other perturbational approaches such as TAP is also elucidated.,2894
with of over study problem probabilistic the basic constraints events. conditional deduction We,We study the problem of probabilistic deduction with conditional constraints over basic events.,2895
constraints conditional show We complete is over basic probabilistic deduction with globally NP-hard. that events,We show that globally complete probabilistic deduction with conditional constraints over basic events is NP-hard.,2896
trees. special the on in We constraint then concentrate conditional of probabilistic case deduction,We then concentrate on the special case of probabilistic deduction in conditional constraint trees.,2897
complete globally efficient for We deduction. very techniques elaborate probabilistic,We elaborate very efficient techniques for globally complete probabilistic deduction.,2898
decidable. We the is theory first-order prove that such constraints over also,We also prove that the first-order theory over such constraints is decidable.,2899
"decomposition on techniques and on aggregation representations. focuses AI-style primarily abstraction, It based","It focuses primarily on abstraction, aggregation and decomposition techniques based on AI-style representations.",2900
"of an alignment for automatic the ontologies. present We PARIS, approach","We present PARIS, an approach for the automatic alignment of ontologies.",2901
"and instances, relations only aligns but PARIS not classes. also","PARIS aligns not only instances, but also relations and classes.",2902
schema-level. the the Alignments with at instance-level at alignments cross-fertilize,Alignments at the instance-level cross-fertilize with alignments at the schema-level.,2903
"ontology Thereby, system of provides truly our to a problem holistic alignment. the solution","Thereby, our system provides a truly holistic solution to the problem of ontology alignment.",2904
heart of the is The probabilistic. approach,The heart of the approach is probabilistic.,2905
This run any PARIS tuning. allows to parameter without,This allows PARIS to run without any parameter tuning.,2906
the precision of extensive the through efficiency its We algorithm demonstrate and experiments.,We demonstrate the efficiency of the algorithm and its precision through extensive experiments.,2907
an is subsumption is in It NP-complete language that this proven problem.,It is proven that subsumption in this language is an NP-complete problem.,2908
more and it to TLU-FU the expressive is reason TL-ALCF. languages how with Then shown,Then it is shown how to reason with the more expressive languages TLU-FU and TL-ALCF.,2909
"of computations NP-hard. these general are case, both the In","In the general case, both of these computations are NP-hard.",2910
"This a  heuristic  system paper worlds. STRIPS presents GRT, domain-independent planning for","  This paper presents GRT, a domain-independent heuristic planning system for STRIPS worlds.",2911
"local optimal domain proposed. overcoming of states, axioms, is based on a method Finally,","Finally, a method of overcoming local optimal states, based on domain axioms, is proposed.",2912
Representation formalism. efficiency a space Knowledge the of We Propositional investigate (PKR),We investigate the space efficiency of a Propositional Knowledge Representation (PKR) formalism.,2913
a This is of model reasoning and algebra its composed temporal labeled algorithms. temporal,This temporal model is composed of a labeled temporal algebra and its reasoning algorithms.,2914
"label associated of inconsistent manage constraints, mutually Reasoning disjuncts. labeled sets algorithms lists, and","Reasoning algorithms manage labeled constraints, associated label lists, and sets of mutually inconsistent disjuncts.",2915
These obtain and guarantee network. minimal algorithms a consistency,These algorithms guarantee consistency and obtain a minimal network.,2916
"constraints a be temporal of organized hierarchy Additionally, in can contexts. alternative","Additionally, constraints can be organized in a hierarchy of alternative temporal contexts.",2917
"Therefore, context-dependent on we reason metric constraints intervals on points. and can disjunctive","Therefore, we can reason on context-dependent disjunctive metric constraints on intervals and points.",2918
paper this make In we the following contributions.,In this paper we make the following contributions.,2919
of results clearly Our indicate a number conclusions.,Our results clearly indicate a number of conclusions.,2920
planning is Domain-independent a combinatorial hard problem.,Domain-independent planning is a hard combinatorial problem.,2921
account quality makes even difficult. more task Taking the plan into,Taking into account plan quality makes the task even more difficult.,2922
"introduces high-quality for efficient a This paradigm new article domain-independent planning. (PbR), by Planning Rewriting","This article introduces Planning by Rewriting (PbR), a new paradigm for efficient high-quality domain-independent planning.",2923
this have and We applied domains. to existing several planner it implemented,We have implemented this planner and applied it to several existing domains.,2924
"blush, contradictory. first these seem requirements At three","At first blush, these three requirements seem contradictory.",2925
"example, agents' risks making undesirable predictable. thereby behavior less side effects, introducing adaptation For","For example, adaptation risks introducing undesirable side effects, thereby making agents' behavior less predictable.",2926
the Our is the solution challenge following. all three of to requirements satisfying,Our solution to the challenge of satisfying all three requirements is the following.,2927
"via are (perturbation) evolutionary have online adapted automaton operators. plans, learning finite-state Agents which","Agents have finite-state automaton plans, which are adapted online via evolutionary learning (perturbation) operators.",2928
"ensure first agents' always constraints critical that satisfied, are verified. plans formally are behavioral To","To ensure that critical behavioral constraints are always satisfied, agents' plans are first formally verified.",2929
reverified adaptation. every after then are They,They are then reverified after every adaptation.,2930
"repaired. the violated, plans are are that constraints concludes If reverification","If reverification concludes that constraints are violated, the plans are repaired.",2931
the the One these ability of to emotional the of person. understand skills state is,One of these skills is the ability to understand the emotional state of the person.,2932
network-based describes emotion a approach neural classification. This for thesis,This thesis describes a neural network-based approach for emotion classification.,2933
better these can select parameters. automatically strategy we that a comparatively combination developed of,we developed a strategy that can automatically select comparatively better combination of these parameters.,2934
another propagation. back introduce We also perform way to,We also introduce another way to perform back propagation.,2935
an in We interested authentic were databases. also construction emotion,We were also interested in construction an authentic emotion databases.,2936
is important very nowadays is no available. task such there This database a because,This is a very important task because nowadays there is no such database available.,2937
genomic sequential fixations evolution implies new sequences. of Phenotypic,Phenotypic evolution implies sequential fixations of new genomic sequences.,2938
includes environments. it depends on the explicitly that a term of the Instead sequence,Instead it includes a term that explicitly depends on the sequence of the environments.,2939
the in quantum systems. classical to and geometric well-studied physical phases related is It,It is related to the well-studied geometric phases in classical and quantum physical systems.,2940
"show ordering that becomes CBJ that We dynamic a such ""perfect"" redundant. variable exists there","We show that there exists a ""perfect"" dynamic variable ordering such that CBJ becomes redundant.",2941
selection'. `sample investigates by cost methods for reducing paper annotation This,This paper investigates methods for reducing annotation cost by `sample selection'.,2942
labeling This new little redundantly examples information. that avoids contribute,This avoids redundantly labeling examples that contribute little new information.,2943
processing task to natural real-world applied of the language stochastic part-of-speech The was method tagging.,The method was applied to the real-world natural language processing task of stochastic part-of-speech tagging.,2944
powerful mean-field a The Plefka. by on are theory suggested based approximations,The approximations are based on a powerful mean-field theory suggested by Plefka.,2945
Plefka's not of is to The theory computationally application tractable. networks belief,The application of Plefka's theory to belief networks is not computationally tractable.,2946
Taylor based tackle we this new series. approximations To on problem propose,To tackle this problem we propose new approximations based on Taylor series.,2947
are attractive. the show that scale Small experiments proposed schemes,Small scale experiments show that the proposed schemes are attractive.,2948
of skill bias Typically such by the supplied hand experts. insights through and is,Typically such bias is supplied by hand through the skill and insights of experts.,2949
investigated. automatically a paper for model learning this bias is In,In this paper a model for automatically learning bias is investigated.,2950
reasoning complexity characterize and We algorithms the propositional MBNF. in provide for,We characterize the complexity and provide algorithms for reasoning in propositional MBNF.,2951
"That {\em is self-matching}. there no is,","That is, there is no {\em self-matching}.",2952
positively in of correlated number tags. the different enhancement The cooperation with is,The enhancement in cooperation is positively correlated with the number of different tags.,2953
is cooperative the a diverse more population more it becomes. The,The more diverse a population is the more cooperative it becomes.,2954
is a The technique. estimation robust multi-structure efficient highly outcome,The outcome is a highly efficient multi-structure robust estimation technique.,2955
"solved for ExpTime. that problem can we be in ALCQ, contrast, In show the","In contrast, we show that for ALCQ, the problem can be solved in ExpTime.",2956
"problem nominals, to is PSpace-complete. be known this Without","Without nominals, this problem is known to be PSpace-complete.",2957
is general. true in example not I show by that this,I show by example that this is not true in general.,2958
their The and is practical these complexity is orderings relevance of discussed. investigated,The complexity of these orderings is investigated and their practical relevance is discussed.,2959
"goal orderings are different Secondly, compute two reasonable to developed. methods","Secondly, two different methods to compute reasonable goal orderings are developed.",2960
given on data. are real results Experimental,Experimental results are given on real data.,2961
"it Interestingly, in is decide to if the time exists. polynomial possible lcs","Interestingly, it is possible to decide in polynomial time if the lcs exists.",2962
planning can for effectively provide universal OBDDs complex domains. scale can and plans,OBDDs can effectively scale and can provide universal plans for complex planning domains.,2963
"are We domains. non-deterministic, arising the in in multi-agent interested particularly addressing complexities","We are particularly interested in addressing the complexities arising in non-deterministic, multi-agent domains.",2964
"specify We multi-agent domains. language, a non-deterministic, NADL, new introduce description domain planning to","We introduce a new planning domain description language, NADL, to specify non-deterministic, multi-agent domains.",2965
of language controllable definition agents. the contributes agents The explicit and environment uncontrollable,The language contributes the explicit definition of controllable agents and uncontrollable environment agents.,2966
planning universal and OBDD-based system UMOP uses NADL algorithms. planning different The,The UMOP planning system uses NADL and different OBDD-based universal planning algorithms.,2967
includes strong It developed and planning algorithms. the strong previously cyclic,It includes the previously developed strong and strong cyclic planning algorithms.,2968
is a be system. efficient rich and shown to planning UMOP,UMOP is shown to be a rich and efficient planning system.,2969
"techniques. new of present number Second, refinements existing and novel of we a approximation methods","Second, we present a number of new approximation methods and novel refinements of existing techniques.",2970
navigation results by on from experiments theoretical the are domain. supported problem The a agent,The theoretical results are supported by experiments on a problem from the agent navigation domain.,2971
set The by are the an instance. literals backbone of which entailed is,The backbone is the set of literals which are entailed by an instance.,2972
loci mapping genomic (eQTL) that quantitative trait to transcription. gene Expression aims determine regulate regions,Expression quantitative trait loci (eQTL) mapping aims to determine genomic regions that regulate gene transcription.,2973
"method eQTL new We analysis models, via called linear a fast for eQTL. Matrix present","We present a new method for fast eQTL analysis via linear models, called Matrix eQTL.",2974
using models. ANOVA can eQTL for and and both regression test linear association Matrix model,Matrix eQTL can model and test for association using both linear regression and ANOVA models.,2975
correlated with errors. also and heteroscedastic of models supports testing It models,It also supports testing of heteroscedastic models and models with correlated errors.,2976
a finding Value for well-known POMDPs. for is algorithm optimal iteration policies,Value iteration is a well-known algorithm for finding optimal policies for POMDPs.,2977
typically to of converge. takes number a large It iterations,It typically takes a large number of iterations to converge.,2978
This of for proposes method the iteration. paper value convergence accelerating a,This paper proposes a method for accelerating the convergence of value iteration.,2979
their them Both limitations. of own and have advantages,Both of them have their own advantages and limitations.,2980
Alignments presents incorporate advantages The Topological and of snakes. the novel algorithm technique to both,The algorithm presents novel technique to incorporate the advantages of both Topological Alignments and snakes.,2981
this contributions twofold. are paper in Our,Our contributions in this paper are twofold.,2982
using studied Lasso. effectiveness the of selection work number feature recent A of,A number of recent work studied the effectiveness of feature selection using Lasso.,2983
are sparse (gLasso) the Lasso structure. consistently under which Group network procedure estimates Conditions derived,Conditions are derived under which the Group Lasso (gLasso) procedure consistently estimates sparse network structure.,2984
"involves condition a The ""false score."" key connection","The key condition involves a ""false connection score.""",2985
"been empirical no investigation yet. made However, has","However, no empirical investigation has been made yet.",2986
are planning evaluate the techniques that We FF describe algorithmic the and system. in used,We describe and evaluate the algorithmic techniques that are used in the FF planning system.,2987
"Unlike facts HSP's our be not assume independent. does heuristic, method to","Unlike HSP's heuristic, our method does not assume facts to be independent.",2988
"for knowledge. Description logics (DLs) structured well-known, Logics managing are suitable,","Description Logics (DLs) are suitable, well-known, logics for managing structured knowledge.",2989
"and, concepts supported. becomes reasoning is concepts imprecise about In particular, fuzzy thus,","In particular, concepts becomes fuzzy and, thus, reasoning about imprecise concepts is supported.",2990
main constraint of reasoning. local of Enforcing consistencies the features one is,Enforcing local consistencies is one of the main features of constraint reasoning.,2991
"relations between and of overview them, We highlight of all some them. make an","We make an overview of all of them, and highlight some relations between them.",2992
settings. in In this Pruning different present three Reduced analyses of paper we Error,In this paper we present analyses of Reduced Error Pruning in three different settings.,2993
This different analysis assumptions. is conducted under two,This analysis is conducted under two different assumptions.,2994
to are analysis that specific examples we the the distributed tree. uniformly a In assume,In a specific analysis we assume that the examples are distributed uniformly to the tree.,2995
both problem. is empirically this and purpose theoretically of the The paper study to,The purpose of this paper is to study both theoretically and empirically the problem.,2996
datasets An is classifiers described. of approach imbalanced construction to the from,An approach to the construction of classifiers from imbalanced datasets is described.,2997
imbalanced are equally not the categories is represented. A dataset classification if approximately,A dataset is imbalanced if the classification categories are not approximately equally represented.,2998
involves examples. of over-sampling the class Our method synthetic class minority creating minority,Our method of over-sampling the minority class involves creating synthetic minority class examples.,2999
"random in population that these accelerate we spread. results, show variability can temporal dispersal Using","Using these results, we show that random temporal variability in dispersal can accelerate population spread.",3000
to that this Our change. suggest results needs,Our results suggest that this needs to change.,3001
"almost particular, COIN always Braess' paradox. algorithm In avoids this","In particular, this COIN algorithm almost always avoids Braess' paradox.",3002
ones extend authors. results by already These the complete and the published,These results extend and complete the ones already published by the authors.,3003
"first knowledge, this to the our menopause. from of conception is model reserve ovarian To","To our knowledge, this is the first model of ovarian reserve from conception to menopause.",3004
empirical Recent to in commonplace. becoming trends comparison led research have planning,Recent trends in planning research have led to empirical comparison becoming commonplace.,3005
Bath are involving Database Twelve conducted single/multienrollment on presented Image tests Iris and here.,Twelve tests involving single/multienrollment and conducted on Bath Iris Image Database are presented here.,3006
by accelerates reinforcement discusses related a paper transfer learning using tasks. system that This from,This paper discusses a system that accelerates reinforcement learning by using transfer from related tasks.,3007
features early stable These are process. the easy and recognize learning to in,These features are stable and easy to recognize early in the learning process.,3008
space state of function. thus a generate partitioning and They the the,They generate a partitioning of the state space and thus the function.,3009
as is The graph. represented a partition,The partition is represented as a graph.,3010
knowledge important network they as identify engineers help are them These parameters. influential results for,These results are important for knowledge engineers as they help them identify influential network parameters.,3011
"for described mechanism packs. query is executing such Furthermore, a","Furthermore, a mechanism is described for executing such query packs.",3012
Processes in the large of planning under paper problem uncertainty (MDPs). Markov addresses Decision This,This paper addresses the problem of planning under uncertainty in large Markov Decision Processes (MDPs).,3013
"we approximate exploit that In algorithms structure this MDPs. paper, two in factored present solution","In this paper, we present two approximate solution algorithms that exploit structure in factored MDPs.",3014
"programming, the linear approximate uses One programming. second algorithm approximate dynamic and","One algorithm uses approximate linear programming, and the second approximate dynamic programming.",3015
"programs, We logic a framework for of parameterized parameter logical/mathematical statistical i.e. propose learning","We propose a logical/mathematical framework for statistical parameter learning of parameterized logic programs, i.e.",3016
parameterized facts with a probabilistic clause distribution. definite programs containing,definite clause programs containing probabilistic facts with a parameterized distribution.,3017
estimator. the scheme that this would It is observed RSS improve,It is observed that the RSS scheme would improve this estimator.,3018
are critical values desired calculated. The powers and,The desired critical values and powers are calculated.,3019
observed RSS estimators would is powers. increase also that It,It is also observed that RSS estimators would increase powers.,3020
towards steps three goal that here. We have taken describe we,We describe three steps we have taken towards that goal here.,3021
collecting for describe domain of a experts. a from definitions tool We prototype,We describe a prototype of a tool for collecting definitions from domain experts.,3022
"have detect of A meanwhile proposed outliers, definitions. lot to been lot of a algorithms","A lot of algorithms have been proposed to detect outliers, meanwhile a lot of definitions.",3023
"Unsatisfying hoc that is seem point definitions solution vague, which ad an the one. makes","Unsatisfying point is that definitions seem vague, which makes the solution an ad hoc one.",3024
"two give nature clear We definitions. of outliers, analyzed and the","We analyzed the nature of outliers, and give two clear definitions.",3025
Both study are originated from the algorithms on AI. general,Both algorithms are originated from the study on general AI.,3026
demonstrate short intelligence A discussion to Pe function of the algorithm. about added the is,A short discussion about intelligence is added to demonstrate the function of the Pe algorithm.,3027
hormone (AMH) ovarian Anti-Mullerian growing a product is follicles. Background: of,Background: Anti-Mullerian hormone (AMH) is a product of growing ovarian follicles.,3028
"The of concentration reflect AMH in follicle non-growing blood population, i.e. also may the (NGF)","The concentration of AMH in blood may also reflect the non-growing follicle (NGF) population, i.e.",3029
"and ovarian in predicting value the reserve, of lifespan. reproductive be","the ovarian reserve, and be of value in predicting reproductive lifespan.",3030
production description to of AMH the not been has reported. menopause up full A,A full description of AMH production up to the menopause has not been reported.,3031
peak there and that a shown pre-pubertal a peak. neonatal We also potential have is,We have also shown that there is a neonatal peak and a potential pre-pubertal peak.,3032
ages. model us data normative all at Our generate to allows,Our model allows us to generate normative data at all ages.,3033
key data highlight dynamics. ovarian inflection follicle These in points Conclusions/Significance:,Conclusions/Significance: These data highlight key inflection points in ovarian follicle dynamics.,3034
activity females. The gonadal of is for existence a increase neonatal in confirmed,The existence of a neonatal increase in gonadal activity is confirmed for females.,3035
function in environmental a is Mat\'ern covariance dependence data. spatial popular The for choice modeling,The Mat\'ern covariance function is a popular choice for modeling dependence in spatial environmental data.,3036
"are, however, sets. Standard for computationally data often models large Mat\'ern covariance infeasible","Standard Mat\'ern covariance models are, however, often computationally infeasible for large data sets.",3037
set how systems characterise We they in bureaucratic and show arise phylogenetics.,We characterise bureaucratic set systems and show how they arise in phylogenetics.,3038
simple of Fusions are combining logics. way a,Fusions are a simple way of combining logics.,3039
"normal For been fusions modal have detail. in logics, investigated","For normal modal logics, fusions have been investigated in detail.",3040
"description logics related normal. Though not logics, are they closely to necessarily modal are","Though description logics are closely related to modal logics, they are not necessarily normal.",3041
Recursive models (possible-world) to correspond shown counterfactual structures. of precisely a subclass to are,Recursive models are shown to correspond precisely to a subclass of (possible-world) counterfactual structures.,3042
We of modeling structural at using more the look carefully equations. causality,We look more carefully at the modeling of causality using structural equations.,3043
"choices These are, some to subjective. extent,","These choices are, to some extent, subjective.",3044
We an counts what consider choice. appropriate as,We consider what counts as an appropriate choice.,3045
spatial generated the model from The are qualitative in agreement experimental with patterns observations. the,The spatial patterns generated from the model are in qualitative agreement with the experimental observations.,3046
considers observed Our with regression article factors. model a,Our article considers a regression model with observed factors.,3047
A to chain estimate is Carlo developed Markov algorithm the Monte model.,A Markov chain Monte Carlo algorithm is developed to estimate the model.,3048
predictive our evidence better that The has suggests model performance.,The evidence suggests that our model has better predictive performance.,3049
mobile a objects scene This paper algorithm different to track new in conditions. presents,This paper presents a new algorithm to track mobile objects in different scene conditions.,3050
"similarity filtering. proposed measures main idea and multi-features estimation, includes trajectory of tracker the The","The main idea of the proposed tracker includes estimation, multi-features similarity measures and trajectory filtering.",3051
"of object trajectories usually misdetections. However, fragmented are occlusions mobile and because the","However, the mobile object trajectories are usually fragmented because of occlusions and misdetections.",3052
been has different of videos The five with scene method tested conditions.,The method has been tested with five videos of different scene conditions.,3053
the algorithms passing Both tree use structure. on message,Both algorithms use message passing on the tree structure.,3054
are cellular of reaction fundamental biochemical Motivation: interactions building underlying functions. Protein systems blocks,Motivation: Protein interactions are fundamental building blocks of biochemical reaction systems underlying cellular functions.,3055
"using is Therefore, dependencies comprehensive and about integrating required. approach for a information such","Therefore, a comprehensive approach for integrating and using information about such dependencies is required.",3056
complexes. of straightforwardly this the how demonstrate prediction framework we protein improves First,First we demonstrate how this framework straightforwardly improves the prediction of protein complexes.,3057
provided in this paper. are Those proofs,Those proofs are provided in this paper.,3058
we significantly selection number employed of feature reduce Hence to the attributes.,Hence we employed feature selection to significantly reduce the number of attributes.,3059
selection. to of we have application procedure the Additionally feature bagging the tested,Additionally we have tested application of the bagging procedure to the feature selection.,3060
The variables. using random models built forest selected were regression,The random forest regression models were built using selected variables.,3061
The confirms the cross-validation of robustness models.,The cross-validation confirms robustness of the models.,3062
the forms of These work. basis the IHIF,These IHIF forms the basis of the work.,3063
these face suitable that local Thus for images salient are most features recognition. produce,Thus these images produce salient local features that are most suitable for face recognition.,3064
useful discrimination for facial recall. associative are subsequent most and features independent These,These independent features are most useful for subsequent facial discrimination and associative recall.,3065
source audio (NMF) factorization a for matrix is tool common Nonnegative now separation.,Nonnegative matrix factorization (NMF) is now a common tool for audio source separation.,3066
abstracted sometimes games. phenomena simple be two-player biological can complex Even to,Even complex biological phenomena can sometimes be abstracted to simple two-player games.,3067
"success. But between several parties often, determines the evolutionary interaction","But often, the interaction between several parties determines evolutionary success.",3068
in expression theory. coalescence be simple obtained can by results recursions This which a using,This results in a simple expression which can be obtained by recursions using coalescence theory.,3069
"e.g. can be to variety contexts, approach of suit modified a This","This approach can be modified to suit a variety of contexts, e.g.",3070
merged together. maximum and coefficients data three details are the During approximate of other fusion,During the data fusion maximum of approximate and other three details coefficients are merged together.,3071
"used. been experiments, For Face Database thermal/visible has IRIS","For experiments, IRIS thermal/visible Face Database has been used.",3072
transformed scaling rotation. images are introduced capable complicacies and Log-polar by of handling,Log-polar transformed images are capable of handling complicacies introduced by scaling and rotation.,3073
"perceptron reduced a images are multilayer network. Finally, neural fused classified using those","Finally, those reduced fused images are classified using a multilayer perceptron neural network.",3074
the we and describe prions Also of briefly other yeast fungi.,Also we briefly describe the prions of yeast and other fungi.,3075
"behaving are i.e. infectious like genes, Prions proteins","Prions are infectious proteins behaving like genes, i.e.",3076
"its genetic only that contain i.e. tertiary information in structure, not proteins","proteins that not only contain genetic information in its tertiary structure, i.e.",3077
and to hosts. their always proteins are not detrimental ubiquitous Prion,Prion proteins are ubiquitous and not always detrimental to their hosts.,3078
algorithm. this Recognition Non In using propose we Minutiae Fingerprint (DWTFR) paper based DWT,In this paper we propose DWT based Fingerprint Recognition using Non Minutiae (DWTFR) algorithm.,3079
only. LL The local {\theta} are orientation computed band angle Dominant Coherence and on,The Dominant local orientation angle {\theta} and Coherence are computed on LL band only.,3080
DNA database of those contains like The mycoplasma sequences very molds. human genome sequence,The human genome sequence database contains DNA sequences very like those of mycoplasma molds.,3081
Nowadays growing ontologies Data in interest present Fusion a applications.,Nowadays ontologies present a growing interest in Data Fusion applications.,3082
However of none the purposes. our tools exactly meets,However none of the tools meets exactly our purposes.,3083
about ontological developed Java uncertain application that We also enables reasoning a these instances.,We also developed a Java application that enables reasoning about these uncertain ontological instances.,3084
presented. comparison with reputation A be also will,A comparison with reputation will be also presented.,3085
Individuals of an what intuitive good have a perception coincidence. makes,Individuals have an intuitive perception of what makes a good coincidence.,3086
drops. model presented complexity when suggests subjects occur The perceive that coincidences here,The model presented here suggests that coincidences occur when subjects perceive complexity drops.,3087
"than simpler separately. if together, considered are, events Co-occurring","Co-occurring events are, together, simpler than if considered separately.",3088
model redefinition This subjective leads possible of to a probability.,This model leads to a possible redefinition of subjective probability.,3089
drawbacks them. and of show We features,We show features and drawbacks of them.,3090
new is deliberately Identification of modified images face system. facial challenge a identification for,Identification of deliberately modified facial images is a new challenge for face identification system.,3091
for face scenario. in SIFT this features identification We propose efficient,We propose SIFT features for efficient face identification in this scenario.,3092
has approach. with well eigenface comparative Further known given analysis been,Further comparative analysis has been given with well known eigenface approach.,3093
real case the images Experiments performance to methods. conducted of with both evaluate been have,Experiments have been conducted with real case images to evaluate the performance of both methods.,3094
"multi-criteria any is incremental, an producing time system quality plans. LPG","LPG is an incremental, any time system producing multi-criteria quality plans.",3095
"analyzed. problems succinctly policies represented related size In to the of paper, are this some","In this paper, some problems related to the size of succinctly represented policies are analyzed.",3096
The neighbour classify k-nearest applied images. algorithms neighbour are new nearest word to and,The nearest neighbour and k-nearest neighbour algorithms are applied to classify new word images.,3097
results . The quite encouraging obtained are,The results obtained are quite encouraging .,3098
"Internet. Auctions an popular especially becoming increasingly over transacting for method are business, the","Auctions are becoming an increasingly popular method for transacting business, especially over the Internet.",3099
We present model a mathematical transfer-of-control of strategies.,We present a mathematical model of transfer-of-control strategies.,3100
on order based loosely causal UCPOP. planner (POCL) VHPOP a is partial link,VHPOP is a partial order causal link (POCL) planner loosely based on UCPOP.,3101
durative with VHPOP supports actions for temporal incorporating also techniques reasoning. constraint standard by planning,VHPOP also supports planning with durative actions by incorporating standard techniques for temporal constraint reasoning.,3102
this the extend to paper approach with we temporal In information. deal,In this paper we extend the approach to deal with temporal information.,3103
discovery. paper presents an approach subgroup expert-guided This to,This paper presents an approach to expert-guided subgroup discovery.,3104
for the describe We a specifying of system actions. effects,We describe a system for specifying the effects of actions.,3105
a deviations in model performance. present for expressive timing We generative probabilistic music,We present a probabilistic generative model for timing deviations in expressive music performance.,3106
model. a of to proposed The is switching state structure equivalent space the model,The structure of the proposed model is equivalent to a switching state space model.,3107
variables in a musical locations correspond The as switch to score. discrete note,The switch variables correspond to discrete note locations as in a musical score.,3108
variables continuous denote the The tempo. hidden,The continuous hidden variables denote the tempo.,3109
results Our with suggest results methods. simulation better sequential,Our simulation results suggest better results with sequential methods.,3110
capable planning. designed be It the temporal nature to multi-objective handling of of is metric,It is designed to be capable of handling the multi-objective nature of metric temporal planning.,3111
"(i.e., minimum over all plans). cheapest plans","minimum over all plans (i.e., cheapest plans).",3112
"on results we selected problems. report Furthermore, experimental","Furthermore, we report experimental results on selected problems.",3113
"distribution data class twenty-six and, study for learning. determine for the best sets We each,","We study twenty-six data sets and, for each, determine the best class distribution for learning.",3114
"induce domains unary domain's called causal operator the natural However, a structure, graph.","However, unary operator domains induce a natural structure, called the domain's causal graph.",3115
This and graph effect between preconditions relates each of domain operator. the,This graph relates between the preconditions and effect of each domain operator.,3116
with our to complexity serializable of results Finally we the of question relate planning subgoals.,Finally we relate our results to the question of complexity of planning with serializable subgoals.,3117
"for empirically, analyze, scheduling behavior theoretically algorithm of types. our and various distribution We the","We analyze, theoretically and empirically, the behavior of our scheduling algorithm for various distribution types.",3118
"to Latin Square our Finally, empirical applying problem. scheduling results algorithm we of present the","Finally, we present empirical results of applying our scheduling algorithm to the Latin Square problem.",3119
payoff. stochastic common-interest all games receive players identical In an,In common-interest stochastic games all players receive an identical payoff.,3120
framework populations introduced. the for structured group in is A evolution of modeling mathematical,A framework for the mathematical modeling of evolution in group structured populations is introduced.,3121
of is fixed The a number large divided groups fixed of into size. population,The population is divided into a fixed large number of groups of fixed size.,3122
the gene. a viability of leads for a to Hamilton's generalization This condition mutant of,This leads to a generalization of Hamilton's condition for the viability of a mutant gene.,3123
structural present information positive We an data efficient alone. from that algorithm about examples learns,We present an efficient algorithm that learns structural information about data from positive examples alone.,3124
on ten Web the We validated algorithm reinduction sources.,We validated the reinduction algorithm on ten Web sources.,3125
number show of these rules have that desirable a properties. We computational,We show that these rules have a number of desirable computational properties.,3126
propose number We two rules. manipulating of for also approximation these methods a,We also propose a number of approximation methods for manipulating these two rules.,3127
manipulate both Experiments in rules to difficult often demonstrate that practice. are,Experiments demonstrate that both rules are often difficult to manipulate in practice.,3128
study. further style These deserve suggest voting elimination results that rules,These results suggest that elimination style voting rules deserve further study.,3129
"competitors, so, state-of-the-art reveal Our often that experiments outperforms for problems. approach significantly our large","Our experiments reveal that our approach outperforms state-of-the-art competitors, often significantly so, for large problems.",3130
set. using data test procedure gene a the we expression well-known Finally,Finally we test the procedure using a well-known gene expression data set.,3131
decomposition systems in a task organize to knowledge. method many used is agent agent Hierarchical,Hierarchical task decomposition is a method used in many agent systems to organize agent knowledge.,3132
"the a team COM-MTDP coordination the novel provides for algorithms. Furthermore, model development basis of","Furthermore, the COM-MTDP model provides a basis for the development of novel team coordination algorithms.",3133
is  for technique a planning. major  Search,  Search is a major technique for planning.,3134
"of prohibitively the large search space sizes make expensive. search However,","However, prohibitively large sizes of the search space make search expensive.",3135
efficiency. better main Developing been the functions heuristic has search improving technique for,Developing better heuristic functions has been the main technique for improving search efficiency.,3136
searches. has POR speeding promise up in shown,POR has shown promise in speeding up searches.,3137
"for In we theory a develop this POR planning. paper, in","In this paper, we develop a theory for POR in planning.",3138
planning explained new algorithms by previous be We in POR the show can that theory.,We show that previous POR algorithms in planning can be explained by the new theory.,3139
"Based we stronger propose POR a algorithm. on new, the theory, new","Based on the new theory, we propose a new, stronger POR algorithm.",3140
domains search reduction using Experimental cost on further results algorithm. various the show new planning,Experimental results on various planning domains show further search cost reduction using the new algorithm.,3141
functions energy solved. can be for quadratic analytically dynamics Hamiltonian,Hamiltonian dynamics for quadratic energy functions can be analytically solved.,3142
online. are paper this Supplemental materials for available,Supplemental materials for this paper are available online.,3143
measure approach a is automatic based fuzziness An on histogram threshold presented.,An automatic histogram threshold approach based on a fuzziness measure is presented.,3144
(SMC) Sequential Carlo become in Monte have horses ABC. work approaches,Sequential Monte Carlo (SMC) approaches have become work horses in ABC.,3145
of if scientific evaluation likelihood any impossible. For the problematic nearly not challenging problem is,For nearly any challenging scientific problem evaluation of the likelihood is problematic if not impossible.,3146
for sets statistics are selection Such parameter estimation problems. and model both of constructed sufficient,Such sufficient sets of statistics are constructed for both parameter estimation and model selection problems.,3147
illustrative apply and to a model our range of We approach selection problems. real-world,We apply our approach to a range of illustrative and real-world model selection problems.,3148
into to lungs. fit iteratively vessels track the to cylinders them are Model the,Model cylinders are iteratively fit to the vessels to track them into the lungs.,3149
novel is (SS) metric. surface evaluation accomplished our sparse using Validation,Validation is accomplished using our novel sparse surface (SS) evaluation metric.,3150
undergone reality improvement Augmented past have years. in considerable,Augmented reality have undergone considerable improvement in past years.,3151
spread reality augmented enabled of This mass applications.,This enabled mass spread of augmented reality applications.,3152
constraint been Set represented variables subset in using have multiset bounds. programming typically and,Set and multiset variables in constraint programming have typically been represented using subset bounds.,3153
offer pruning that runtime. significantly they and demonstrate Results better,Results demonstrate that they offer significantly better pruning and runtime.,3154
for Repeated interaction individuals dilemma main in the is between maintaining cooperation situations. social mechanism,Repeated interaction between individuals is the main mechanism for maintaining cooperation in social dilemma situations.,3155
predict that evolutionary We numerically also simulations. show adaptive outcome of the dynamics appositely estimated,We also show that numerically estimated adaptive dynamics appositely predict the outcome of evolutionary simulations.,3156
achieved through measure is paradigm. stack a new filtering This,This new measure is achieved through a stack filtering paradigm.,3157
anisotropic variation entropy used this to is the measure. entropy of define directional an Then,Then the entropy variation of this directional entropy is used to define an anisotropic measure.,3158
been evaluated with real-world method and images. has degraded The artificial,The method has been evaluated with artificial and real-world degraded images.,3159
its and Kolmogorov-Smirnov test two-sample with biasedness. deals This paper,This paper deals with two-sample Kolmogorov-Smirnov test and its biasedness.,3160
This in different sizes. case in unbiased of is test general not sample,This test is not unbiased in general in case of different sample sizes.,3161
values $\alpha$. level of some distribution significance biased out for found We most,We found out most biased distribution for some values of significance level $\alpha$.,3162
"of Here, inconsistency the arises show maps. from that the connectivity the incompleteness we","Here, we show that the inconsistency arises from the incompleteness of the connectivity maps.",3163
"measurements. in transient by Similarly, inputs thalamic the patterns vivo activation recent evoked reproduce","Similarly, the activation patterns evoked by transient thalamic inputs reproduce recent in vivo measurements.",3164
Markov refer our We Carlo Monte (TMCMC). to Transformation-based new methodology as chain,We refer to our new methodology as Transformation-based Markov chain Monte Carlo (TMCMC).,3165
"properties. our convergence theoretically, establishing the proposal We validate","We validate our proposal theoretically, establishing the convergence properties.",3166
the savings Remarkable results. acceptance good computational rates are well properties convergence as as and,Remarkable computational savings as well as good convergence properties and acceptance rates are the results.,3167
performance MIPS shown competitions. two last has international distinguished planning the in,MIPS has shown distinguished performance in the last two international planning competitions.,3168
longer but Plans schedules. were actions of time-stamped sequences no,Plans were no longer sequences of actions but time-stamped schedules.,3169
detail. in is latter aspect analyzed The,The latter aspect is analyzed in detail.,3170
user making. preferences Information automated role in plays key about a decision,Information about user preferences plays a key role in automated decision making.,3171
representation circumstances. often many quite arguably a natural Such compact in and is,Such a representation is often compact and arguably quite natural in many circumstances.,3172
"In the variables. for many there choice decision is cases, considerable","In many cases, there is considerable choice for the decision variables.",3173
"can such between viewpoint. primal choose a a problems, In we and dual","In such problems, we can choose between a primal and a dual viewpoint.",3174
comparing models. methodology They constraint a for illustrate also general different,They also illustrate a general methodology for comparing different constraint models.,3175
"complex are understand, reimplement, quite these modify). or However, all systems (to","However, all these systems are quite complex (to understand, reimplement, or modify).",3176
algorithm introduce solving problem. for Physical-A* this We the (PHA*),We introduce the Physical-A* algorithm (PHA*) for solving this problem.,3177
"we an Specifically, experimental provide such implementation system. a for","Specifically, we provide an experimental implementation for such a system.",3178
recovers lost a the A surplus. significant of decommitment fraction phase subsequent,A subsequent decommitment phase recovers a significant fraction of the lost surplus.,3179
"precise determined. between has the yet and relation not been transduction Moreover, induction","Moreover, the precise relation between induction and transduction has not yet been determined.",3180
"implicitly While bound tight, routine. given is computational a this via","While tight, this bound is given implicitly via a computational routine.",3181
dynamics are simulation according to making and results. management These decision supported models with specified,These dynamics are specified according to management decision making models and supported with simulation results.,3182
"determines Value"" of ""Fitness article This by results and simulation Optimality"". of means ""Probability","This article determines simulation results by means of ""Fitness Value"" and ""Probability of Optimality"".",3183
"paper, In argue i.e. that this complexification, we","In this paper, we argue that complexification, i.e.",3184
"new goals. these through elaboration both of achieves structure, adding solutions the incremental","the incremental elaboration of solutions through adding new structure, achieves both these goals.",3185
This technique presents new on a classifier combination theory the Dempster-Shafer based evidence. of paper,This paper presents a new classifier combination technique based on the Dempster-Shafer theory of evidence.,3186
calculate Captured then are processed speed. vehicle images to,Captured images are then processed to calculate vehicle speed.,3187
for admissible functions computing for evaluation method problems. a search explore heuristic We,We explore a method for computing admissible heuristic evaluation functions for search problems.,3188
the as and we domains other extend method Here to that it show applies well.,Here we extend the method and show that it applies to other domains as well.,3189
We which call partitioned also pattern databases. dynamically for present we heuristics another method additive,We also present another method for additive heuristics which we call dynamically partitioned pattern databases.,3190
problem disjoint into of for dynamically. partition we the search Here state the each subproblems,Here we partition the problem into disjoint subproblems for each state of the search dynamically.,3191
presents model repositories. aimed at novel paper This approach a modelling building compositional,This paper presents a novel compositional modelling approach aimed at building model repositories.,3192
in field It furthers the two respects.,It furthers the field in two respects.,3193
"preferences process. the user Secondly, selection the of model incorporation enables into it","Secondly, it enables the incorporation of user preferences into the model selection process.",3194
of only plan then We low-dimensional the in belief terms features.,We then plan only in terms of the low-dimensional belief features.,3195
"been measures it Furthermore, also accuracy. that achieve has better reported constrained such can","Furthermore, it has been reported that such constrained measures can also achieve better accuracy.",3196
"result, behavior. to correct it harder is a their guarantee As","As a result, it is harder to guarantee their correct behavior.",3197
finite-state methods verification for most of the One is of checking. model successful automated systems,One of the most successful methods for automated verification of finite-state systems is model checking.,3198
"the is, The better performance the gets compact one verifier. more the from BDD","The more compact the BDD is, the better performance one gets from the verifier.",3199
"a for problem. However, optimal NP-complete is finding an order BDD an","However, finding an optimal order for a BDD is an NP-complete problem.",3200
"expert been developed have based knowledge heuristic methods several Therefore, variable for ordering. on","Therefore, several heuristic methods based on expert knowledge have been developed for variable ordering.",3201
"For evaluated. of model, sequences are a training each training number","For each training model, a number of training sequences are evaluated.",3202
"believe sub-domains caches, in (alu, models of that etc.) We","We believe that in sub-domains of models (alu, caches, etc.)",3203
will more even valuable. prove our system,our system will prove even more valuable.,3204
"known and Despite importance, ultimate its the of is the limitations strengths technique. of little","Despite its importance, little is known of the ultimate strengths and limitations of the technique.",3205
"In there years, been problems. in phase interest recent transitions has combinatorial much of","In recent years, there has been much interest in phase transitions of combinatorial problems.",3206
principle algorithm. class and lossless compression in large includes such is algorithms The of any,The class of such algorithms is large and in principle includes any lossless compression algorithm.,3207
is by quality prediction comparison the made to as respect log-loss. average The measured with,The comparison is made with respect to prediction quality as measured by the average log-loss.,3208
feasible HLC of large the computationally compute The to models. theorem dimensions effective makes it,The theorem makes it computationally feasible to compute the effective dimensions of large HLC models.,3209
compute theorem used be the can dimensions tree of to models. general effective also The,The theorem can also be used to compute the effective dimensions of general tree models.,3210
"hotel for predicting employ all some method virtually prices. Thus, participants","Thus, virtually all participants employ some method for predicting hotel prices.",3211
the paper of investigates Bayesian complexity networks. MAP in This,This paper investigates the complexity of MAP in Bayesian networks.,3212
even MAP Pr become easy. remains MPE when show hard that and We also,We also show that MAP remains hard even when MPE and Pr become easy.,3213
framework. We generic introduce approximation MAP a,We introduce a generic MAP approximation framework.,3214
some plan. facts that every solution valid point be true Landmarks at in must are,Landmarks are facts that must be true at some point in every valid solution plan.,3215
is and completely domain- Our planner-independent. methodology,Our methodology is completely domain- and planner-independent.,3216
data-sources. integration method for of abductive independent a an coherent introduce We,We introduce an abductive method for a coherent integration of independent data-sources.,3217
iteration popular a finding POMDPs. algorithm for for optimal near policies is Value,Value iteration is a popular algorithm for finding near optimal policies for POMDPs.,3218
"paper, In value this iteration belief restricted study subsets. to we","In this paper, we study value iteration restricted to belief subsets.",3219
"of 'gradual' on valuation underlying principles a discuss First, arguments we general interactions. their based","First, we discuss general principles underlying a 'gradual' valuation of arguments based on their interactions.",3220
"principles, an valuation argumentation Following we for these abstract define several system. models","Following these principles, we define several valuation models for an abstract argumentation system.",3221
"the Then, in of acceptability of 'graduality' arguments. concept introduce we","Then, we introduce 'graduality' in the concept of acceptability of arguments.",3222
NEXP-complete. been The general to has be problem shown,The general problem has been shown to be NEXP-complete.,3223
"information among also sharing which paper their studies performance. improve This decision-makers, can the","This paper also studies information sharing among the decision-makers, which can improve their performance.",3224
were conducted with key Interviews sites. in personnel affected,Interviews were conducted with key personnel in affected sites.,3225
of widely are an Metallic number more and increasing more applications. in used nanoparticles,Metallic nanoparticles are more and more widely used in an increasing number of applications.,3226
responses our the cellular This nanoparticles. requires knowledge of increase to to,This requires to increase our knowledge of the cellular responses to nanoparticles.,3227
"In appear system. this macrophages context, attractive an as","In this context, macrophages appear as an attractive system.",3228
"in e.g. foreign eliminating major play matter, They a role","They play a major role in eliminating foreign matter, e.g.",3229
We learning a traffic congestion estimation on based propose algorithm. unsupervised on-line system,We propose a traffic congestion estimation system based on unsupervised on-line learning algorithm.,3230
The does detection. or not extraction on system motion background rely,The system does not rely on background extraction or motion detection.,3231
using then clustered Mixture are The K-means and Gaussian into classes features extracted two Models(GMM).,The extracted features are then clustered into two classes using K-means and Gaussian Mixture Models(GMM).,3232
shape of the Knowledge the in sphere. built forms ILS to,Knowledge built in to the ILS forms the shape of sphere.,3233
recognition presents Kannada features. spatial vowels on and This numerals multi-font/multi-size paper based,This paper presents multi-font/multi-size Kannada numerals and vowels recognition based on spatial features.,3234
by model cargo motors for The transport Refs. introduced in molecular hopping,The hopping model for cargo transport by molecular motors introduced in Refs.,3235
essential execute movement long-range for to Jamming cargos this in is motors way. of,Jamming of motors is essential for cargos to execute long-range movement in this way.,3236
is methods. in LRR equivalent So factorization to essence,So in essence LRR is equivalent to factorization methods.,3237
when robust is LRR more the data is corruption RSI heavy. than in,RSI is more robust than LRR when the corruption in data is heavy.,3238
on real of testify synthetic improved robustness both and the to Experiments data RSI.,Experiments on both synthetic and real data testify to the improved robustness of RSI.,3239
correspondence as problem. study poses a the node  problem This  hypergraph labeling feature,  This study poses the feature correspondence problem as a hypergraph node labeling problem.,3240
group of reducing the and structure It problem is complexity relations. capable of,It is capable of reducing the complexity of problem structure and group relations.,3241
"It decision architecture, selection etc. the methods, the examines path, software the","It examines the software architecture, the selection methods, the decision path, etc.",3242
level. are Queries in abstract posed the of terms,Queries are posed in the terms of abstract level.,3243
the chaining. uses and one forward exploits hybrid one reasoning The only second first,The first one uses hybrid reasoning and the second one exploits only forward chaining.,3244
"shows it even This some algorithms good in results, algorithm outperforms cases. and other","This algorithm shows good results, and it even outperforms other algorithms in some cases.",3245
"SQL triggers Furthermore, the executable as rules SBVR-based discussed. will of be application","Furthermore, the application of SBVR-based rules as executable SQL triggers will be discussed.",3246
The highly proper performance on and the of feature selection of OCR depends set. extraction,The performance of OCR highly depends on the proper selection and extraction of feature set.,3247
images. the character handwritten printed have proposed on tested and We Hindi method Bengali and,We have tested the proposed method on printed and handwritten Bengali and Hindi character images.,3248
demonstrate Initial the results of approach. our efficacy,Initial results demonstrate the efficacy of our approach.,3249
to presented program. application iris demo segmentation is An a together with,An application to iris segmentation is presented together with a demo program.,3250
"large be of Unfortunately, data. directly cannot supervised systems due lack machine-learned scale to easily","Unfortunately, large scale systems cannot be easily machine-learned due to lack of directly supervised data.",3251
"acquisition, This semantic disambiguation. up knowledge work methods combining word-sense ends and parsing, for","This work ends up combining methods for knowledge acquisition, semantic parsing, and word-sense disambiguation.",3252
are may which arise. two cases There,There are two cases which may arise.,3253
so the virions with and system host becomes zero The HIV-free. up ends,The system ends up with zero virions and so the host becomes HIV-free.,3254
population unless In the this change. persists HIV case parameters,In this case the HIV population persists unless parameters change.,3255
methods. traditional as diagnosis are of intelligence techniques being extension fault the Computational investigated,Computational intelligence techniques are being investigated as extension of the traditional fault diagnosis methods.,3256
"selected of Algeria. clinker, The SCIMAT is in system the workshop factory cement","The system selected is the workshop of SCIMAT clinker, cement factory in Algeria.",3257
of quantitative focuses This review currents. aspects L-type on,This review focuses on quantitative aspects of L-type currents.,3258
CDI of important dynamical aspect is of cell an electrophysiology. models often,CDI is often an important aspect of dynamical models of cell electrophysiology.,3259
inactivation and related obtained. are of activation Distributions and parameters statistics several to,Distributions and statistics of several parameters related to activation and inactivation are obtained.,3260
with approximately Neurons into divided experimental half-activation groups potentials. two are,Neurons are divided approximately into two groups with experimental half-activation potentials.,3261
present generating time generate to We an O(N) algorithm for patterns Poisson-disc taking points. $N$,We present an algorithm for generating Poisson-disc patterns taking O(N) time to generate $N$ points.,3262
accuracy one Fingerprint technologies. most and is recognition of popular Biometric,Fingerprint recognition is one of most popular and accuracy Biometric technologies.,3263
"it Nowadays, applications. in real is used many","Nowadays, it is used in many real applications.",3264
"However, a recognizing is very still complex problem. in fingerprints quality images poor","However, recognizing fingerprints in poor quality images is still a very complex problem.",3265
"years, accuracy of system. given algorithms, In the to models...are many recognition recent improve","In recent years, many algorithms, models...are given to improve the accuracy of recognition system.",3266
with well that It the rigidity as tension twisting. grows as is stretching found,It is found that the stretching rigidity grows with tension as well as twisting.,3267
upon earlier light biological experimental and implications The observations. results have shed important,The results have important biological implications and shed light upon earlier experimental observations.,3268
many classification. to increase There with solve problems are aiming to efficient these approaches,There are many approaches to solve these problems with aiming to increase efficient classification.,3269
to advantages reflect fulfill this of multiform of are and image approach The classified.,The advantages of this approach are to reflect fulfill and multiform of image classified.,3270
we model Networks combines a many last which for Then applied Neural step. the develop,Then we develop a model which combines many Neural Networks applied for the last step.,3271
This each of conclusion. reliability the final gives the space classification model and evaluates,This model evaluates the reliability of each space and gives the final classification conclusion.,3272
based random (CTD) Commute on metric a Distance is graphs. Time walk,Commute Time Distance (CTD) is a random walk based metric on graphs.,3273
is metric anomaly interest inspired as the use detection. CTD by for Our a of,Our interest is inspired by the use of CTD as a metric for anomaly detection.,3274
be in backgrounded to the landscapes revealed Many could events. historical different details,Many details in the revealed landscapes could be backgrounded to different historical events.,3275
of Image Techniques. Arithmetic Filtering and Pixel-Based Fusion Methods Frequency Combination,Arithmetic Combination and Frequency Filtering Methods of Pixel-Based Image Fusion Techniques.,3276
"The (BT), (MLT). includes first Brovey Transform Normalized Transformation (CN) Multiplicative and type Color Method","The first type includes Brovey Transform (BT), Color Normalized Transformation (CN) and Multiplicative Method (MLT).",3277
is solvent cavity guanines palatinate A between and filled molecules. with formed,A cavity between palatinate guanines is formed and filled with solvent molecules.,3278
centered matrix. the has Previous to work original ability reconstruct the on,Previous work has centered on the ability to reconstruct the original matrix.,3279
is and the of We on specific that performance depends demonstrate approximation. task rank the,We demonstrate that performance is task specific and depends on the rank of the approximation.,3280
The one of most used is commonly for the techniques sharpening sharpening. technique IHS,The IHS sharpening technique is one of the most commonly used techniques for sharpening.,3281
This exact leads the the of transformation?. to confusion is IHS formula what,This leads to confusion what is the exact formula of the IHS transformation?.,3282
"with formalism introduce it demonstrate examples. several letter, this our we and In","In this letter, we introduce our formalism and demonstrate it with several examples.",3283
of neutral known two using the By variants family NK (i.e.,By using two known neutral variants of the NK family (i.e.,3284
structural differences two the some unknown found between landscapes. studied families We of neutral,We found some unknown structural differences between the two studied families of neutral landscapes.,3285
the space. enumeration of underlying methodology requires the search exhaustive Our current,Our current methodology requires the exhaustive enumeration of the underlying search space.,3286
"sampling this developed techniques before implications. be have should analysis can Therefore, practical","Therefore, sampling techniques should be developed before this analysis can have practical implications.",3287
efficient in results method. This filtering very Gaussian,This results in very efficient Gaussian filtering method.,3288
"probabilistic programming I the language offer logic alternative, As an CP-logic. of","As an alternative, I offer the probabilistic logic programming language of CP-logic.",3289
estimation compressed correlation sets of images. the This of problem paper in addresses,This paper addresses the problem of correlation estimation in sets of compressed images.,3290
correlation is estimated processing the by linear in the measurements. domain The jointly compressed,The correlation is estimated in the compressed domain by jointly processing the linear measurements.,3291
related correlated can first efficiently images the a linear using that be show We operator.,We first show that the correlated images can be efficiently related using a linear operator.,3292
then the describe domain. images Using we linear this the dependencies in between relationship compressed,Using this linear relationship we then describe the dependencies between images in the compressed domain.,3293
its such In capture users. of by ontologies vocabularies formalized terms systems shared,In such systems ontologies capture formalized vocabularies of terms shared by its users.,3294
"different many views domain, have the i.e. However users cases local of in","However in many cases users have different local views of the domain, i.e.",3295
the which a in given context of term used. is,of the context in which a given term is used.,3296
debugging to the faults. Recent use diagnosis identify approaches of methods causes ontology,Recent ontology debugging approaches use diagnosis methods to identify causes of the faults.,3297
errors where typical available. no user about is information,where no information about typical user errors is available.,3298
"this review To ODFs. we the end, of first manifold Riemannian","To this end, we first review the Riemannian manifold of ODFs.",3299
individuals these separately requires cases the are not that explicitly reasoning about of mentioned. Neither,Neither of these cases requires reasoning separately about the individuals that are not explicitly mentioned.,3300
and how a diverse techniques framework. general the propose in show the fit We framework,We propose a general framework and show how the diverse techniques fit in the framework.,3301
the matching. framework relation data consider From ontology and between interlinking we this,From this framework we consider the relation between data interlinking and ontology matching.,3302
deviations in of paper the studies the stochastic a problem. multi-armed This bandit regret,This paper studies the deviations of the regret in a stochastic multi-armed bandit problem.,3303
"trade-off the cellular In / paper, this exploitation exploration algorithms. study in genetic we","In this paper, we study the exploration / exploitation trade-off in cellular genetic algorithms.",3304
The to is used explain punctuated these results. equilibria model,The punctuated equilibria model is used to explain these results.,3305
"HyFlex, paper development software a presents of search This the methodologies. framework for cross-domain","This paper presents HyFlex, a software framework for the development of cross-domain search methodologies.",3306
Mean-shift tracking in known a and method well object is popular used tracking problems.,Mean-shift tracking is a popular and well known method used in object tracking problems.,3307
world-wide results of psychology. present scientometric in excellence We about centers,We present scientometric results about world-wide centers of excellence in psychology.,3308
"the AUC result, a this clear first Brier the As and establishes between score. connection","As a result, this establishes the first clear connection between AUC and the Brier score.",3309
increases tip the The wave. probability surfing the towards strongly of,The surfing probability strongly increases towards the tip of the wave.,3310
"front Thus, arise in at wave. the an successful intermediate most position of the mutations","Thus, most successful mutations arise at an intermediate position in the front of the wave.",3311
for association basis tools provides (and It implication rules). with reasoning,It provides tools for reasoning with implication basis (and association rules).,3312
Knowledge-Based a build to reasoning. system confidence is applied It for,It is applied to build a Knowledge-Based system for confidence reasoning.,3313
"utility recovered. ignored, expected are costs the When is maximum resource principle","When resource costs are ignored, the maximum expected utility principle is recovered.",3314
not species of selection mechanism understood. ionic is and The clearly of gating,The mechanism of gating and selection of ionic species is not clearly understood.,3315
used version with evaluation. additional extended reported in The provides that resources were the appendices,The extended version provides appendices with additional resources that were used in the reported evaluation.,3316
"algorithm, (a.k.a. OEM propose an We called","We propose an algorithm, called OEM (a.k.a.",3317
"orthogonalizing problems. for EM), ious least intended squares var-","orthogonalizing EM), intended for var- ious least squares problems.",3318
new rows. The imputes of step the the responses second,The second step imputes the responses of the new rows.,3319
"forms, until have closed convergence. third and second steps simple The iterate and","The second and third steps have simple closed forms, and iterate until convergence.",3320
Convergence of algorithm are rate and examined. the convergence,Convergence and convergence rate of the algorithm are examined.,3321
new provides theoretical This these methods. a comparison between,This provides a new theoretical comparison between these methods.,3322
to illustrate examples Numerical proposed provided the are algorithm.,Numerical examples are provided to illustrate the proposed algorithm.,3323
response such control. is One form adaptive of gain,One form of such adaptive response is gain control.,3324
variance. this most collapses proposed of system The,The system proposed collapses most of this variance.,3325
"to then in case. matched in mark face the used laboratory, The attendance our is","The matched face is then used to mark attendance in the laboratory, in our case.",3326
implies process. heterogeneity contingency the in This evolutionary,This heterogeneity implies contingency in the evolutionary process.,3327
commonplace hundreds becoming High-dimensional of many observations in data of with disciplines. thousands are,High-dimensional data with hundreds of thousands of observations are becoming commonplace in many disciplines.,3328
performance the method through simulation assessed of The studies. is,The performance of the method is assessed through simulation studies.,3329
numerous natural Large may subject to populations simultaneously contain polymorphisms selection. segregating,Large populations may contain numerous simultaneously segregating polymorphisms subject to natural selection.,3330
"when alleles, recombination is associations is disrupts strong rare. recombination draft Since between","Since recombination disrupts associations between alleles, draft is strong when recombination is rare.",3331
"Finally, through draft two accelerates the of intermediates. adaptations step rate deleterious","Finally, draft accelerates the rate of two step adaptations through deleterious intermediates.",3332
an of learns compression. We data algorithm through give a that representation,We give an algorithm that learns a representation of data through compression.,3333
can sweep single the the be from samples model through Independent pixels. drawn a by,Independent samples can be drawn from the model by a single sweep through the pixels.,3334
programming introduce new a Programming called (COP). Oriented paradigm So we Consciousness,So we introduce a new programming paradigm called Consciousness Oriented Programming (COP).,3335
the undirected directed for graphs. method of prove We and consistency,We prove consistency of the method for directed and undirected graphs.,3336
blockmodel. consistent consistent assignment stochastic for The a parameter estimation makes block possible,The consistent block assignment makes possible consistent parameter estimation for a stochastic blockmodel.,3337
graphs. is even for computationally very large also feasible Our method,Our method is also computationally feasible even for very large graphs.,3338
the of model through entities system. a flow passive they,they model the flow of passive entities through a system.,3339
commonly Self-Organizing purposes. are used unsupervised Maps for learning,Self-Organizing Maps are commonly used for unsupervised learning purposes.,3340
"to little attention relatively so knowledge their representation, specifications received Despite epistemic far. have relevance","Despite their relevance to knowledge representation, epistemic specifications have received relatively little attention so far.",3341
"we In this specification. the revisit of paper, epistemic formalism","In this paper, we revisit the formalism of epistemic specification.",3342
appearance setting. generative of Bayesian shape in a This hierarchical model a and paper investigates,This paper investigates a hierarchical Bayesian model of shape and appearance in a generative setting.,3343
"ongoing work meta-reasoning is which aims applying to MCTS, non-trivial. at This techniques","This ongoing work aims at applying meta-reasoning techniques to MCTS, which is non-trivial.",3344
"Finite-time policies empirically. is are compared the and and analysis algorithms of the asymptotic provided,","Finite-time and asymptotic analysis of the policies is provided, and the algorithms are compared empirically.",3345
the paper second Reiter The compare to of is Moore. and of our goal work,The second goal of our paper is to compare the work of Reiter and Moore.,3346
modeling information. allows one recovering imaging that Sparse lost is techniques of efficient the for,Sparse modeling is one of the efficient techniques for imaging that allows recovering lost information.,3347
solution the reconstruction. The optimal algorithm the object enables for wave developed field,The developed algorithm enables the optimal solution for the object wave field reconstruction.,3348
quality an improvement conventional to the goal Our the reconstruction algorithms. respect of is with,Our goal is an improvement of the reconstruction quality with respect to the conventional algorithms.,3349
"significant in results and accuracy, demonstrate of simulations enhancement numerical reconstruction advanced Sparse imaging. regularization","Sparse regularization results in advanced reconstruction accuracy, and numerical simulations demonstrate significant enhancement of imaging.",3350
Detecting edges processing. quality for image objects is images within of the critical,Detecting the edges of objects within images is critical for quality image processing.,3351
classic detectors. the performed than Our amoeba-based better edge edge-detection system,Our amoeba-based edge-detection system performed better than the classic edge detectors.,3352
probability variables weather disturbances. extremes of climate of consequent variance increases and the Higher catastrophic,Higher variance of climate variables increases the probability of weather extremes and consequent catastrophic disturbances.,3353
"directional variation pressure, selection the the model of are environment. and Parameters stochastic of mutation,","Parameters of the model are selection pressure, mutation, directional and stochastic variation of the environment.",3354
"directional track phenotype. the of able is the In system component the to optimum general,","In general, the system is able to track the directional component of the optimum phenotype.",3355
to first We approach. evaluations feasibility provide this experimental the show of,We provide first experimental evaluations to show the feasibility of this approach.,3356
Remembered those episodes are telling. exactly are worth which,Remembered episodes are exactly those which are worth telling.,3357
"the Using introduced allows factors swiftly. accurate here, results getting","Using the factors introduced here, allows getting accurate results swiftly.",3358
to solution compute presented. accuracy is The the mathematical generic,The generic mathematical solution to compute the accuracy is presented.,3359
described. precise approximate and computations The are,The approximate and precise computations are described.,3360
of detection. conceptions of emergence as instances the are phenomenon of reconciled two different Two,Two different conceptions of emergence are reconciled as two instances of the phenomenon of detection.,3361
"of Conversely, emerge may some as modularity trait. by-product a","Conversely, modularity may emerge as a by-product of some trait.",3362
the We modularity ask versatility metabolic networks. of whether here affects,We here ask whether versatility affects the modularity of metabolic networks.,3363
to This language. the for in encodings simpler standard compared allows ASP ones much,This allows for much simpler encodings compared to the ones in standard ASP language.,3364
"encodings novel introduced we the for semantics. Moreover, resolution-based grounded provide recently","Moreover, we provide novel encodings for the recently introduced resolution-based grounded semantics.",3365
"undertake of attempts image to So, paper fusion. study the This Feature-Level based","So, This paper attempts to undertake the study of Feature-Level based image fusion.",3366
"feature this we Hence, paper fusion. (FE) for extraction in consider","Hence, in this paper we consider feature extraction (FE) for fusion.",3367
included term in number are of A 'consciousness'. the concepts,A number of concepts are included in the term 'consciousness'.,3368
"an is epiphenomenon adaptive ? or function, it it have Does an","Does it have an adaptive function, or is it an epiphenomenon ?",3369
reject to give We arguments second alternative. shall the,We shall give arguments to reject the second alternative.,3370
"confidentiality. particular, we credulous analyze non-monotonic affects reasoning In how","In particular, we analyze how credulous non-monotonic reasoning affects confidentiality.",3371
the was by for The easy LC-MS/MS. sequence monitoring BAP status of redesigned biotinylation,The BAP sequence was redesigned for easy monitoring of the biotinylation status by LC-MS/MS.,3372
finger is possible. and shows that between discrimination This movements EEG wrist,This shows that EEG discrimination between wrist and finger movements is possible.,3373
research motor to research. introduces The tasks a BCI of new combination,The research introduces a new combination of motor tasks to BCI research.,3374
Concave provide procedures regularization sparse methods recovery. for natural,Concave regularization methods provide natural procedures for sparse recovery.,3375
"in However, dimensional to are high analyze they setting. the difficult","However, they are difficult to analyze in the high dimensional setting.",3376
"present discuss this overview Under an and of unified connections. framework, existing we their results","Under this unified framework, we present an overview of existing results and discuss their connections.",3377
relational applied of domains. learning been a range Statistical techniques successfully in wide have relational,Statistical relational learning techniques have been successfully applied in a wide range of relational domains.,3378
an established paradigm. programming (ASP) is Programming Answer-Set declarative,Answer-Set Programming (ASP) is an established declarative programming paradigm.,3379
over modularity desired results. for is increasing access place---(meta-)reasoning feature in proper and---assuming The subprogram,The feature is desired for increasing modularity and---assuming proper access in place---(meta-)reasoning over subprogram results.,3380
"sets the can be individually answer called of sub-programs Notably, accessed.","Notably, the answer sets of called sub-programs can be individually accessed.",3381
We a to product interactive present approach configuration. constraint-based,We present a constraint-based approach to interactive product configuration.,3382
"the process, optimizations. configuration show techniques underlying discuss explain the We and","We discuss the configuration process, explain the underlying techniques and show optimizations.",3383
"this tool help paper, on to CADe cancer. detect radiologist In interested to we","In this paper, we interested on CADe tool to help radiologist to detect cancer.",3384
"detection, analysis classification. three-step a based and namely, The flow; CADe on proposed is work","The proposed CADe is based on a three-step work flow; namely, detection, analysis and classification.",3385
visual approach information radiologist. gives good the This from,This approach gives good visual information from the radiologist.,3386
the present to application the mathematical hypothesis-rich We data. of genome-wide theory first association,We present the first application of the hypothesis-rich mathematical theory to genome-wide association data.,3387
late-onset sporadic Parkinson's genome-wide was analyzed. association disease study dataset,late-onset sporadic Parkinson's disease genome-wide association study dataset was analyzed.,3388
and are These in DSLs reification. used selection propagator constraint,These DSLs are used in propagator selection and constraint reification.,3389
"overheads. specifications reducing translated Prolog run-time code, interpretative compilation At to are time, these","At compilation time, these specifications are translated to Prolog code, reducing interpretative run-time overheads.",3390
NLPR database. view videos system of side proposed using The evaluated is,The proposed system is evaluated using side view videos of NLPR database.,3391
VQ compression constructed technique depends of performance The the codebook. upon image based,The performance of VQ based image compression technique depends upon the constructed codebook.,3392
design algorithm. used for A codebook VQ technique widely the is (LBG) Linde-Buzo-Gray,A widely used technique for VQ codebook design is the Linde-Buzo-Gray (LBG) algorithm.,3393
SAT important in for since is removed. performance Preprocessing a be redundancy can formula,Preprocessing a formula in SAT is important for performance since redundancy can be removed.,3394
The is Coprocessor. is called riss SAT of and part preprocessor the solver,The preprocessor is part of the SAT solver riss and is called Coprocessor.,3395
"priors intractable lead to in However, these computationally general models.","However, these priors in general lead to computationally intractable models.",3396
do how truly when know agent we an intelligent? But is,But how do we know when an agent is truly intelligent?,3397
carried be could practically such out. how then We sketch testing,We then sketch how such testing could practically be carried out.,3398
regions. tumor wavelet feature derived set texture and from is A statistical based normal,A wavelet based statistical texture feature set is derived from normal and tumor regions.,3399
"requires good high designing skills problems. and insight the usually into However, system mathematical PDE","However, designing PDE system usually requires high mathematical skills and good insight into the problems.",3400
vision results can problems PDEs well. The solve reasonably that experimental show learnt the different,The experimental results show that the learnt PDEs can solve different vision problems reasonably well.,3401
"can measures. for used be both approximate These exact and services similarity using matching,","These services can be used for both exact and approximate matching, using similarity measures.",3402
measure are adopting quality information from presented well-established of retrieval. textual borrowed a Results,Results are presented adopting a well-established measure of quality borrowed from textual information retrieval.,3403
analysis fundamental Object detection video in applications. a for automated step is many vision,Object detection is a fundamental step for automated video analysis in many vision applications.,3404
detection techniques. performed or usually in detectors by Object video a subtraction object background is,Object detection in a video is usually performed by object detectors or background subtraction techniques.,3405
"automate separate a becomes task. analysis, critical a object detection the To phase without training","To automate the analysis, object detection without a separate training phase becomes a critical task.",3406
tried People information. to have by using this tackle task motion,People have tried to tackle this task by using motion information.,3407
explain between the other We DECOLOR sparsity-based methods. relations and,We explain the relations between DECOLOR and other sparsity-based methods.,3408
long-reach passive focused the networks. of on resilient design optical an We present application,We present an application focused on the design of resilient long-reach passive optical networks.,3409
to failure. property metro important of a resilience such node is single An placement,An important property of such a placement is resilience to single metro node failure.,3410
NP-Complete. prove problem that this We is,We prove that this problem is NP-Complete.,3411
for network based We Ireland. consider case-study a detailed a on,We consider a detailed case-study based on a network for Ireland.,3412
machine recognition language. segmentation in critical of is of any phases one the Document,Document segmentation is one of the critical phases in machine recognition of any language.,3413
technique. recognition decides the segmentation symbols Correct of accuracy of character individual,Correct segmentation of individual symbols decides the accuracy of character recognition technique.,3414
most in the India. script is popular Devnagari,Devnagari is the most popular script in India.,3415
"used Sanskrit for and It Nepali languages. Marathi, writing is Hindi,","It is used for writing Hindi, Marathi, Sanskrit and Nepali languages.",3416
"most popular is the the language world. Moreover, third Hindi in","Moreover, Hindi is the third most popular language in the world.",3417
"vowels, various consonants of documents Devnagari consist and modifiers.","Devnagari documents consist of vowels, consonants and various modifiers.",3418
is proper Hence challenging. segmentation word Devnagari of,Hence proper segmentation of Devnagari word is challenging.,3419
simple paper. Devnagari documents is proposed histogram this segment to based in A approach,A simple histogram based approach to segment Devnagari documents is proposed in this paper.,3420
also Various are of challenges Devnagari script discussed. segmentation in,Various challenges in segmentation of Devnagari script are also discussed.,3421
results. paper presents this implementation series experimental performance and third discusses The ZAPs in,The third paper in this series discusses ZAPs implementation and presents experimental performance results.,3422
in Wind role an of energy increasing world-wide. energy supply plays the,Wind energy plays an increasing role in the supply of energy world-wide.,3423
of variables We energy correlation the the the reveal for output. different,We reveal the correlation of the different variables for the energy output.,3424
between network networks. regulatory Topological relationships and within motifs protein-protein functional interaction and represent,Topological network motifs represent functional relationships within and between regulatory and protein-protein interaction networks.,3425
self-contained functional Enriched modules. often units into aggregate forming motifs,Enriched motifs often aggregate into self-contained units forming functional modules.,3426
to mar/car assumptions. in-depth questions of analysis paper we provide an this several relating In,In this paper we provide an in-depth analysis of several questions relating to mar/car assumptions.,3427
fact by mar/car is question versions complicated the assumptions distinct of that This several exist.,This question is complicated by the fact that several distinct versions of mar/car assumptions exist.,3428
less show most versions and We are restrictive applications. that for sufficient distributional,We show that distributional versions are less restrictive and sufficient for most applications.,3429
at and measures local provide nominal associations evaluations levels. both of global detailed These,These measures provide detailed evaluations of nominal associations at both local and global levels.,3430
association classification. matrix rise matrix proposed also generalized to gives in the confusion The expected,The proposed association matrix also gives rise to the expected generalized confusion matrix in classification.,3431
The association equivalence vector and matrix relations also of shown. hierarchy defined by the are,The hierarchy of equivalence relations defined by the association vector and matrix are also shown.,3432
provides This drivers. driving filter also patterns insights about of,This filter also provides insights about driving patterns of drivers.,3433
also recent that make integer some significantly upon touch encodings developments competitive. more programming We,We also touch upon some recent developments that make integer programming encodings significantly more competitive.,3434
each construct first hypotheses We segmentation a of for set nested slice. neuron,We first construct a nested set of neuron segmentation hypotheses for each slice.,3435
"empirically tabu novel predictions Finally, behavior. algorithm two and regarding search discuss we substantiate","Finally, we discuss and empirically substantiate two novel predictions regarding tabu search algorithm behavior.",3436
explicitly information. and given processing for investigating,for investigating and processing explicitly given information.,3437
SNR define factor. effectiveness We by against reconstruction compression plotting,We define effectiveness by plotting reconstruction SNR against compression factor.,3438
limited to linear sensing. non-adaptive scope is The,The scope is limited to linear non-adaptive sensing.,3439
The experimentally are on of algorithms the domain resulting from bioinformatics. representation and evaluated problems,The resulting representation and algorithms are experimentally evaluated on problems from the domain of bioinformatics.,3440
relational Markov an selection (MDPs). policy Decision study large Processes We approach to for,We study an approach to policy selection for large relational Markov Decision Processes (MDPs).,3441
"suggesting to also work. limitations approach, The point our of future experiments some","The experiments also point to some limitations of our approach, suggesting future work.",3442
technique processing. a is for general Variable constraint elimination,Variable elimination is a general technique for constraint processing.,3443
space is complexity. its It discarded of high because often,It is often discarded because of its high space complexity.,3444
"combined be extremely it useful with when techniques. However, can other","However, it can be extremely useful when combined with other techniques.",3445
that the techniques these empirically. and option theoretically show both are known best We,We show that these techniques are the best known option both theoretically and empirically.,3446
"methods. include dynamic search, programming, structured and These heuristic","These include dynamic programming, heuristic search, and structured methods.",3447
"the Using that their performance. compare certain and we affect NMRDPP, problem features identify methods","Using NMRDPP, we compare the methods and identify certain problem features that affect their performance.",3448
problem framework. can Sophisticated domains be in formalised our,Sophisticated problem domains can be formalised in our framework.,3449
of paper problem learning studies policies from This diagnostic the training examples.,This paper studies the problem of learning diagnostic policies from training examples.,3450
"a costs. is there most In between kinds of tradeoff settings, diagnostic two these","In most diagnostic settings, there is a tradeoff between these two kinds of costs.",3451
diagnostic paper (MDP). as a Decision Process This Markov making decision formalizes,This paper formalizes diagnostic decision making as a Markov Decision Process (MDP).,3452
some including greedy also several over algorithms methods. previously-published paper improvements introduces The,The paper also introduces several greedy algorithms including some improvements over previously-published methods.,3453
addresses the The then of policies examples. learning question diagnostic from paper,The paper then addresses the question of learning diagnostic policies from examples.,3454
"integrated regularizers To into reduce the are overfitting, algorithms. search","To reduce overfitting, regularizers are integrated into the search algorithms.",3455
"Finally, methods benchmark sets. data diagnostic the the proposed paper compares on five","Finally, the paper compares the proposed methods on five benchmark diagnostic data sets.",3456
and commercial defense optimal solutions applications. be Provably may desirable for,Provably optimal solutions may be desirable for commercial and defense applications.,3457
work our extend Rao-Blackwellised earlier in first RDBNs. We filtering the particle described to,We first extend the Rao-Blackwellised particle filtering described in our earlier work to RDBNs.,3458
first uses predicates smooth hierarchies over estimates. the abstraction one particle the to filters The,The first one uses abstraction hierarchies over the predicates to smooth the particle filters estimates.,3459
function relational for The designed with second estimation specifically a domains. kernel kernel density employs,The second employs kernel density estimation with a kernel function specifically designed for relational domains.,3460
"integrating of require remote multiple ? these can In addition, information since tasks sources","In addition, since these tasks can require integrating multiple sources of remote information ?",3461
it efficient possible. make is to as execution desirable as,it is desirable to make execution as efficient as possible.,3462
the and THESEUS. a both have language We system executor implemented in called,We have implemented both the language and executor in a system called THESEUS.,3463
algorithm present value a randomized Perseus. called iteration point-based We,We present a randomized point-based value iteration algorithm called Perseus.,3464
same be We can action with how to extended dealing show spaces. idea continuous the,We show how the same idea can be extended to dealing with continuous action spaces.,3465
the Experimental of POMDP large in scale Perseus show problems. potential results,Experimental results show the potential of Perseus in large scale POMDP problems.,3466
"in current remain planning, AI planners. Despite progress challenging for many recent benchmarks","Despite recent progress in AI planning, many benchmarks remain challenging for current planners.",3467
strategy. Our share a common methods four-step,Our methods share a common four-step strategy.,3468
A and filtering the ranking useful procedure selects most macro-operators.,A filtering and ranking procedure selects the most useful macro-operators.,3469
"to Finally, speed used the are searches. macros up future selected","Finally, the selected macros are used to speed up future searches.",3470
international We ideas effectiveness the competitions. of from planning on benchmarks our demonstrate,We demonstrate the effectiveness of our ideas on benchmarks from international planning competitions.,3471
"sites, sites, model. finite unlinked markers or mutating a We to according biallelic consider infinite","We consider unlinked biallelic markers mutating according to a finite sites, or infinite sites, model.",3472
"text first, skew regions are extracted corrected. and At","At first, text regions are extracted and skew corrected.",3473
"into and regions characters. Then, lines segmented these binarized are and","Then, these regions are binarized and segmented into lines and characters.",3474
recognition are into module. the passed Characters,Characters are passed into the recognition module.,3475
them a easy Designing an neighborhoods fashion applying and in task. meaningful not is these,Designing these neighborhoods and applying them in a meaningful fashion is not an easy task.,3476
"order are in must appropriate which Moreover, determined. be applied they an","Moreover, an appropriate order in which they are applied must be determined.",3477
paper investigate attempt this we In to this issue.,In this paper we attempt to investigate this issue.,3478
inventory The on biobjective routing a problem. the article study presents,The article presents a study on the biobjective inventory routing problem.,3479
"safety in issue critical paramount example, settings. For is of the confidence","For example, the issue of confidence is paramount in safety critical settings.",3480
based studied Recent branch-and-bound on Boolean search. optimizers the work of validation,Recent work studied the validation of Boolean optimizers based on branch-and-bound search.,3481
oracle. This both entails unsatisfiable solutions from answers implementing and NP the for satisfiable validating,This entails implementing solutions for validating both satisfiable and unsatisfiable answers from the NP oracle.,3482
"of accompanied, birth proteins of electrophoresis Two-dimensional has and the proteomics. preceded,","Two-dimensional electrophoresis of proteins has preceded, and accompanied, the birth of proteomics.",3483
(e.g. electrophoresis Then and limitations are two-dimensional of features positive the discussed,Then the limitations and positive features of two-dimensional electrophoresis are discussed (e.g.,3484
by example taken the illustrated literature a from commented in and detailed detail. This is,This is illustrated by a detailed example taken from the literature and commented in detail.,3485
and classification Machines. performs pathway Boltzmann using identity appearance The (factored)-Restricted models deep object,The identity pathway models object appearance and performs classification using deep (factored)-Restricted Boltzmann Machines.,3486
"control location, orientation, the and models attended scale pathway of object. speed the The","The control pathway models the location, orientation, scale and speed of the attended object.",3487
of estimated particle these states posterior filtering. is The distribution with,The posterior distribution of these states is estimated with particle filtering.,3488
"chemotaxis an particular, In to chemoeffector cooperatively interact ultrasensitive bacterial to receptors concentrations. produce response","In particular, bacterial chemotaxis receptors interact cooperatively to produce an ultrasensitive response to chemoeffector concentrations.",3489
study Muller's of accumulation. effect the to provides framework ratchet a quantitative,Muller's ratchet provides a quantitative framework to study the effect of accumulation.,3490
We Methods: ratchet. maps to describe Muller's Wright-Fisher how process,Methods: We describe how Wright-Fisher process maps to Muller's ratchet.,3491
general diffusion landscape adaptive from construct analytically equation. We,We analytically construct adaptive landscape from general diffusion equation.,3492
from model adaptive generalize We landscape of application diffusion viewpoint. the,We generalize the application of diffusion model from adaptive landscape viewpoint.,3493
"More a us importantly, the model perform evaluation. to principled directed enables","More importantly, the directed model enables us to perform a principled evaluation.",3494
"cards, case it suit is rank each In of the of and playing the card.","In the case of playing cards, it is the suit and rank of each card.",3495
system passes. image processing done in two is The,The image processing system is done in two passes.,3496
"been the theory uncertainty measures of In introduced. have many functions, belief of","In the theory of belief functions, many measures of uncertainty have been introduced.",3497
"these represent. not what easy really it always to to measures try is However, understand","However, it is not always easy to understand what these measures really try to represent.",3498
"this re-interpret In belief of theory the uncertainty in some we paper, of measures functions.","In this paper, we re-interpret some measures of uncertainty in the theory of belief functions.",3499
the We and some present drawbacks measures. existing interests of,We present some interests and drawbacks of the existing measures.,3500
"of measure these we contradiction. a introduce observations, On","On these observations, we introduce a measure of contradiction.",3501
"Bayesianity non-specificity and a some Therefore, mass. present we degrees of of","Therefore, we present some degrees of non-specificity and Bayesianity of a mass.",3502
allow different choiceness. to constraints of These notions model,These constraints allow to model different notions of choiceness.,3503
"voting. approval-disapproval-preferential a new particular, give to with method deal we In","In particular, we give a new method to deal with approval-disapproval-preferential voting.",3504
produce drugs side  effects. Allosteric used fewer increasingly  are because they,  Allosteric drugs are increasingly used because they produce fewer side effects.,3505
SR is from spatially estimate to multiple given a high-resolution a technique low-resolution images. image,SR is a technique to estimate a spatially high-resolution image from given multiple low-resolution images.,3506
estimator variational is The Bayes. numerically using determined by,The estimator is numerically determined by using variational Bayes.,3507
solve We problem this Taylor simple approximations. with,We solve this problem with simple Taylor approximations.,3508
estimated within likelihood framework. the These maximum are,These are estimated within the maximum likelihood framework.,3509
fast robust approach proposed The and to noise. is,The proposed approach is fast and robust to noise.,3510
"our to of potential, approach. We the as scheme our test as well highlight limitations,","We test our scheme to highlight the potential, as well as limitations, of our approach.",3511
how fitness. variability in studies success Many have reproductive affects analyzed,Many studies have analyzed how variability in reproductive success affects fitness.,3512
causes to analyzed. framework often incomplete improperly or conceptual be fractured This applications particular,This fractured conceptual framework often causes particular applications to be incomplete or improperly analyzed.,3513
That diminishing fitness. a measure relation and relative reproductive creates success between,That relative measure creates a diminishing relation between reproductive success and fitness.,3514
reduce success. in to in fitness variability reproductive returns proportion Diminishing,Diminishing returns reduce fitness in proportion to variability in reproductive success.,3515
of relative a that The rare also measurement success frequency dependence induces favors types.,The relative measurement of success also induces a frequency dependence that favors rare types.,3516
"Second, in has populations hierarchical a structure. variability","Second, variability in populations has a hierarchical structure.",3517
individual's in in reproduction. Variable an traits that success different of individual variation affects,Variable success in different traits of an individual affects that individual's variation in reproduction.,3518
The aid algorithm peer evaluation. new released online to be will,The new algorithm will be released online to aid peer evaluation.,3519
psoriatic in presented. the patients gastrointestinal pathology Results on of are studies,Results of the studies on gastrointestinal pathology in psoriatic patients are presented.,3520
by model practical confirmed results The validity of work. the of the clinical is,The validity of the model is confirmed by the results of practical clinical work.,3521
"limited incorporated language. Ultimately, durative of actions the was notion a into","Ultimately, a limited notion of durative actions was incorporated into the language.",3522
"an we understanding. such In hand, the a to derive formal background investigation at","In the investigation at hand, we derive a formal background to such an understanding.",3523
its relating in FF. as approximation to implemented also We results present h^+,We also present results relating h^+ to its approximation as implemented in FF.,3524
regarding the same. ends dead is The provably behavior,The behavior regarding dead ends is provably the same.,3525
We direction. into preliminary that some outline we made steps,We outline some preliminary steps we made into that direction.,3526
to archaea high grow Both are able at concentrations. salt,Both archaea are able to grow at high salt concentrations.,3527
"MFCs was media resistance internal the performance. By improving increasing diminished, the the conductivity,","By increasing the media conductivity, the internal resistance was diminished, improving the MFCs performance.",3528
framework novel a discriminative integrating for into We prior knowledge classifiers. present,We present a novel framework for integrating prior knowledge into discriminative classifiers.,3529
heuristic present solving for algorithm first-order Decision Processes Markov search We (FOMDPs). a,We present a heuristic search algorithm for solving first-order Markov Decision Processes (FOMDPs).,3530
abstraction. referred kind abstraction of is This state as first-order to,This kind of abstraction is referred to as first-order state abstraction.,3531
definition Intelligence formal of Universal a Measure proposed intelligence. is recently The,The Universal Intelligence Measure is a recently proposed formal definition of intelligence.,3532
developing real-world This UIM-based studies practical a metric. performance involved issues the paper in,This paper studies the practical issues involved in developing a real-world UIM-based performance metric.,3533
in multi-agent industry. academic systems interest the are and Open community in gaining distributed,Open distributed multi-agent systems are gaining interest in the academic community and in industry.,3534
"open agents such using protocols. often settings, coordinated standardized are In conversation agent","In such open settings, agents are often coordinated using standardized agent conversation protocols.",3535
"weaknesses relative their examined. and have not been strengths However,","However, their relative strengths and weaknesses have not been examined.",3536
"not and suitability scalability been Moreover, have tasks addressed. for different their","Moreover, their scalability and suitability for different tasks have not been addressed.",3537
approach in significant This is particularly suitable improvements representation scalability for offers overhearing. and,This representation approach offers significant improvements in scalability and is particularly suitable for overhearing.,3538
full model. divergence phylogenetic Information a about using from obtained multiple outgroup genomes is,Information about divergence is obtained from multiple outgroup genomes using a full phylogenetic model.,3539
parameters expectation by The maximization. key are selection-related estimated,The key selection-related parameters are estimated by expectation maximization.,3540
"is interpretation multimodal interfaces, build effective of automated inputs user To multimodal important.","To build effective multimodal interfaces, automated interpretation of user multimodal inputs is important.",3541
"properties that steady-state yielded eliminations multi-component In systems emergent of networks. analysis has reveal biology,","In systems biology, steady-state analysis has yielded eliminations that reveal emergent properties of multi-component networks.",3542
"become previously of considered while methods Analyses analysis identical, distinct as are new revealed feasible.","Analyses previously considered distinct are revealed as identical, while new methods of analysis become feasible.",3543
illustrated the is of by The performance filter numerical new simulations.,The performance of the new filter is illustrated by numerical simulations.,3544
method as such Our comparing in protocols for different experiments useful is RNA-Seq.,Our method is useful for comparing different protocols in experiments such as RNA-Seq.,3545
how distance domains. with adapt we for probabilistic heuristics use explain based We,We explain how we adapt distance based heuristics for use with probabilistic domains.,3546
heuristics based success. of also incorporates on probability Probapop,Probapop also incorporates heuristics based on probability of success.,3547
design successes the the Probapop. implementation during explain We and encountered and difficulties of,We explain the successes and difficulties encountered during the design and implementation of Probapop.,3548
can visualize curves resulting data to structures. The then bundled cluster Bezier given be with,The resulting Bezier curves can then be bundled to visualize data with given cluster structures.,3549
"and evolutionary synergism rate. the Ultimately, variation processes genetic developmental sets between","Ultimately, the synergism between developmental processes and genetic variation sets evolutionary rate.",3550
It very be in deployment. use would easy and,It would be very easy in use and deployment.,3551
"Alternatively, into non-binary binary the translated can be problem an one. equivalent","Alternatively, the non-binary problem can be translated into an equivalent binary one.",3552
applicability the this we the paper evaluate In approach. latter of,In this paper we evaluate the applicability of the latter approach.,3553
of online We two for marginal-utility methods learning the heuristic. developed,We developed two methods for online learning of the marginal-utility heuristic.,3554
ubiquitous biology. is a alignment Multiple computational (MSA) sequence in problem,Multiple sequence alignment (MSA) is a ubiquitous problem in computational biology.,3555
"paper, previous, review this we two lines complementary of research. first In","In this paper, we first review two previous, complementary lines of research.",3556
search. heuristic Fast is Downward planning a on based classical system,Fast Downward is a classical planning system based on heuristic search.,3557
survey This in methods graphics. simulation of rendering and paper ocean presents computer a,This paper presents a survey of ocean simulation and rendering methods in computer graphics.,3558
the of the list results competition. and overview the systems We competing,We list the competing systems and overview the results of the competition.,3559
entire data large in of is too The presented far set be to full.,The entire set of data is far too large to be presented in full.,3560
summary; a appendix. detailed provide available in complete the an data is We online,We provide a detailed summary; the complete data is available in an online appendix.,3561
explain We we awarded prizes. the competition how,We explain how we awarded the competition prizes.,3562
important Distributed been in multi-agent Satisfaction has research. (DCSP) problem an considered systems long Constraint,Distributed Constraint Satisfaction (DCSP) has long been considered an important problem in multi-agent systems research.,3563
"other DCSP known, APO outperforms We present complete techniques. empirical evidence shows that that","We present empirical evidence that shows that APO outperforms other known, complete DCSP techniques.",3564
to approaches. show object avoidance in  response Many species reactions looming ,  Many species show avoidance reactions in response to looming object approaches.,3565
machine is complexity music learning. a the Modelling of world real for challenge,Modelling the real world complexity of music is a challenge for machine learning.,3566
modeling sequences the of task music the from We address genre. melodic same,We address the task of modeling melodic sequences from the same music genre.,3567
"performance. representations face been deliver Recently, shown recognizers on based linear have to the state-of-the-art","Recently, the face recognizers based on linear representations have been shown to deliver state-of-the-art performance.",3568
"In real-world suffer expressions, disguises images random applications, and from occlusions. face usually however,","In real-world applications, however, face images usually suffer from expressions, disguises and random occlusions.",3569
"in this we work, learning-inference-mixed the problem fashion. a address In","In this work, we address the problem in a learning-inference-mixed fashion.",3570
"immense to a BPRs sets, derived. Furthermore, also accommodate algorithm is boosting-like","Furthermore, to accommodate immense BPRs sets, a boosting-like algorithm is also derived.",3571
"boosted The to model, a.k.a prototype. obtains similar performance its Boosted-ORE,","The boosted model, a.k.a Boosted-ORE, obtains similar performance to its prototype.",3572
of scheduling shop characterize We the instances the problem random of job (JSP). search landscape,We characterize the search landscape of random instances of the job shop scheduling problem (JSP).,3573
"In increase vision, discriminability. computer to often filtered input machine learning and images data are","In machine learning and computer vision, input images are often filtered to increase data discriminability.",3574
"efficiency The approach, of however, on the depends this selection of points. greatly","The efficiency of this approach, however, depends greatly on the selection of points.",3575
performed POMDP standard tasks. realistic robotic domains is both using Evaluation and,Evaluation is performed using both standard POMDP domains and realistic robotic tasks.,3576
real synthetic We our performance users. of with both suggestion model-based evaluate the and techniques,We evaluate the performance of our model-based suggestion techniques with both synthetic and real users.,3577
instruments have varied approaches produce often unpredictable and results. in been Investment financial,Investment approaches in financial instruments have been varied and often produce unpredictable results.,3578
the symmetric several We results. analyze rank-one perturbation approximation for symmetric and problem derive tensors,We analyze the symmetric rank-one approximation problem for symmetric tensors and derive several perturbation results.,3579
properties We of constraints. with and monotone programs study convex,We study properties of programs with monotone and convex constraints.,3580
to from We concepts programming. these formalisms and extend logic results normal,We extend to these formalisms concepts and results from normal logic programming.,3581
"scenario. the a lags, adaptation serial coefficients arbitrary derive under We correlation for small","We derive the serial correlation coefficients for arbitrary lags, under a small adaptation scenario.",3582
a for and comprehensive describes system article applications. surveillance This monitoring,This article describes a comprehensive system for surveillance and monitoring applications.,3583
component dimensionality is space. of to further performed the analysis principal the reduce feature A,A principal component analysis is performed to further reduce the dimensionality of the feature space.,3584
is aspects Shape important in recognizing plants. an,Shape is an important aspects in recognizing plants.,3585
"plants. been Several including introduced objects, approaches to identify have","Several approaches have been introduced to identify objects, including plants.",3586
"reflect Ideally, the possible benchmarks of the should developed applications technology.","Ideally, the benchmarks should reflect possible applications of the developed technology.",3587
recognizing in plants. useful foliage method help people The is very to,The method is very useful to help people in recognizing foliage plants.,3588
that plants Foliage and various leaf. patterns have plants the in unique colors are,Foliage plants are plants that have various colors and unique patterns in the leaf.,3589
"that result better method gave performance PNN, The than Fourier shows the SVM, Transform. and","The result shows that the method gave better performance than PNN, SVM, and Fourier Transform.",3590
was also tested foliage using colors. method The plants with various,The method was also tested using foliage plants with various colors.,3591
"in representation addition, period a conversion of length. returns terms In the minimal","In addition, the conversion returns a minimal representation in terms of period length.",3592
"temporal granularities. multiple we As example, with to illustrate constraint an the application reasoning","As an example, we illustrate the application to temporal constraint reasoning with multiple granularities.",3593
"completion, one-dimensional, problems. propose a packing bin for We multicontainer branch-and-bound strategy","We propose bin completion, a branch-and-bound strategy for one-dimensional, multicontainer packing problems.",3594
"aspects. essential often real-life uncertainty are and and scenarios, preferences temporal coexisting In","In real-life temporal scenarios, uncertainty and preferences are often essential and coexisting aspects.",3595
"problems, propose presence the For the such controllability of check to we properties. algorithms","For such problems, we propose algorithms to check the presence of the controllability properties.",3596
"and tied. Thus, control information inextricably are senging,","Thus, senging, control and information are inextricably tied.",3597
A formulated this framework. can problems in variety inference of in be large AI,A large variety of inference problems in AI can be formulated in this framework.,3598
consider computer language vision and in examples We natural processing.,We consider examples in computer vision and natural language processing.,3599
We by sets that prove reduct (resp. answer,We prove that answer sets by reduct (resp.,3600
by are weakly complement) to (resp. equivalent,by complement) are equivalent to weakly (resp.,3601
fundamental ecosystem Understanding affects to event an is extinction biodiversity conservation. how,Understanding how an extinction event affects ecosystem is fundamental to biodiversity conservation.,3602
pure scenario. best This represents topological case statement the,This pure topological statement represents the best case scenario.,3603
for local sufficient proliferation diffusion. precondition is global the,sufficient local proliferation is the precondition for global diffusion.,3604
global diffusion. control controlling The strategies for chance design disease provide to results gradient a,The results provide a chance to design gradient control strategies for controlling disease global diffusion.,3605
of the pathology presented. patients Results on gastrointestinal in are studies psoriatic,Results of the studies on gastrointestinal pathology in psoriatic patients are presented.,3606
practical model confirmed The of is clinical the work. by validity of results the,The validity of the model is confirmed by the results of practical clinical work.,3607
algorithm. also overfitting greedy such of a fighting We problem the address in,We also address the problem of fighting overfitting in such a greedy algorithm.,3608
was conceived constructed as Planning originally the a International for franca lingua and Competition. PDDL,PDDL was originally conceived and constructed as a lingua franca for the International Planning Competition.,3609
Most for each activity. classical assume formulations and scheduling fixed duration known a,Most classical scheduling formulations assume a fixed and known duration for each activity.,3610
are of ones achieving have good probability best makespan. which solutions a a high The,The best solutions are ones which have a high probability of achieving a good makespan.,3611
a by linear polyhedron convex defined Let be of $K$ inequalities. closed number finite a,Let $K$ be a closed convex polyhedron defined by a finite number of linear inequalities.,3612
"and an on outer particular, is in the direction. perturbation In that we lexicographic focus","In particular, we focus on the perturbation that is lexicographic and in an outer direction.",3613
The the outperforms our deconvolution show previous method on that some problem numerical results ones.,The numerical results on the deconvolution problem show that our method outperforms some previous ones.,3614
exploiting inference convergence improves algorithms. It by memory-based,It improves convergence by exploiting memory-based inference algorithms.,3615
first algorithms step generic are a towards and unified algorithms. variable Our elimination backtrack,Our backtrack and variable elimination generic algorithms are a first step towards unified algorithms.,3616
optimization fuzzy paper system for optimization. methods inference This various compares,This paper compares various optimization methods for fuzzy inference system optimization.,3617
"annealing. algorithm, methods swarm The optimization particle optimization are simulated genetic and compared","The optimization methods compared are genetic algorithm, particle swarm optimization and simulated annealing.",3618
the We with random study transition instances. associated QXORSAT phase,We study the phase transition associated with random QXORSAT instances.,3619
"planning. actions constraints temporal to durations, an When adds such extra handling have difficulty","When actions have durations, handling such temporal constraints adds an extra difficulty to planning.",3620
performance experiments Additional planner. the show our reasoning good of temporal into the integrated techniques,Additional experiments show the good performance of the temporal reasoning techniques integrated into our planner.,3621
show a We language. each of defines generates and algorithms propositional that these,We show that each of these algorithms defines and generates a propositional language.,3622
methods these We problems evaluate inference in simulation several on a challenging study.,We evaluate these methods on several challenging inference problems in a simulation study.,3623
"is of the number solutions search. maintained A small these ""elite during","A small number of these ""elite solutions is maintained during the search.",3624
"an its used. overview provide algorithms architecture detailing of behaviour, the search and We","We provide an overview of its architecture and search behaviour, detailing the algorithms used.",3625
In problems. address paper we this both,In this paper we address both problems.,3626
"Thus, are additional features. with they unified extended and","Thus, they are unified and extended with additional features.",3627
"completeness of algorithm LRTS Second, any prove and the we framework. convergence covered by","Second, we prove completeness and convergence of any algorithm covered by the LRTS framework.",3628
"upper-bounds and the relating prove several solution parameters Third, control we quality.","Third, we prove several upper-bounds relating the control parameters and solution quality.",3629
"theory community population Applications are genetics, evolutionary to and discussed. game ecology,","Applications to community ecology, population genetics, and evolutionary game theory are discussed.",3630
approaches then previous SSA. in We with estimator this compare simulations optimization-based for,We then compare this estimator in simulations with previous optimization-based approaches for SSA.,3631
paper using the model dynamical epidemiological In susceptible-infectious (SI) study we this graphs.,In this paper we study the susceptible-infectious (SI) epidemiological model using dynamical graphs.,3632
including systems. have been applied areas complex in recently many Dynamical structures,Dynamical structures have been recently applied in many areas including complex systems.,3633
of parameter the disease. This infectivity includes also the regarding information,This parameter includes also information regarding the infectivity of the disease.,3634
"unstable. However, high classifier of in the is dimensions provably the very performance","However, the performance of the classifier in very high dimensions is provably unstable.",3635
adaptive intrinsic is We regression to k-NN that also show dimension.,We show that k-NN regression is also adaptive to intrinsic dimension.,3636
matrix Dung's theory the argumentation framework. and introduce of We to its block,We introduce matrix and its block to the Dung's theory of argumentation framework.,3637
"argumentation a provide the mathematics way ""controversial determine This to framework. arguments"" in an powerful","This provide a powerful mathematics way to determine the ""controversial arguments"" in an argumentation framework.",3638
different terms can be expansion in of interpreted paths cells. readily between This,This expansion can be readily interpreted in terms of paths between different cells.,3639
for networks. We then expressions derive correlation randomly structure in average explicit the connected,We then derive explicit expressions for the average correlation structure in randomly connected networks.,3640
"objective is images required. assessment an for fusion the spatial quality resolution of Therefore,","Therefore, an objective quality of the spatial resolution assessment for fusion images is required.",3641
"detection, CPU. both the on tracked After are hands the","After the detection, both hands are tracked on the CPU.",3642
paper has figure . This due withdrawn a to by missing author been the,This paper has been withdrawn by the author due to a missing figure .,3643
kernel multiple a We problem. the learning Gaussian of formulation consider process,We consider a Gaussian process formulation of the multiple kernel learning problem.,3644
function to process is construction imposing equivalent space. over product priors a This of heavy-tailed,This construction is equivalent to imposing a product of heavy-tailed process priors over function space.,3645
and is variational classification. regression algorithm A derived binary for inference,A variational inference algorithm is derived for regression and binary classification.,3646
algorithms thousand. larger to in dimensions trouble have Existing a scaling than,Existing algorithms have trouble in scaling to dimensions larger than a thousand.,3647
ideas. new key This requires a few,This requires a few key new ideas.,3648
centrality The concerned first the community. in distribution the of,The first concerned the distribution of centrality in the community.,3649
paper novel recognition approach for a solving present problem. The will face,The paper will present a novel approach for solving face recognition problem.,3650
reduced indirect side models. the for the is strongly This to intermediate of as ability,This ability is strongly reduced as for the intermediate to indirect side of the models.,3651
comparing approach to our a methods. We also simulation provide study existing,We also provide a simulation study comparing our approach to existing methods.,3652
is marked conventional Performance classification schemes. bench- against,Performance is bench- marked against conventional classification schemes.,3653
interactions protein the interesting post-genomic is challenges of of Predicting era. the one more,Predicting protein interactions is one of the more interesting challenges of the post-genomic era.,3654
"PPI. are predictions framework of one Importantly component however, for identifying larger a computational only","Importantly however, computational predictions are only one component of a larger framework for identifying PPI.",3655
network. density protein-protein We synaptic post interactions predicting good PPI show results for,We show good results predicting protein-protein interactions for post synaptic density PPI network.,3656
"a have de-noising multitude of methods proposed been of In hitherto. MRI, the field image","In the field of MRI, a multitude of image de-noising methods have been proposed hitherto.",3657
"the paper of present the two-fold. contribution is Accordingly,","Accordingly, the contribution of the present paper is two-fold.",3658
solvers Designing is a component-based constraint complex problem.,Designing component-based constraint solvers is a complex problem.,3659
"the there components required, are between interdependencies optional some components. Some and are are","Some components are required, some are optional and there are interdependencies between the components.",3660
"of ad-hoc this, and limited. solver approaches have to and modification previous Because design been","Because of this, previous approaches to solver design and modification have been ad-hoc and limited.",3661
valid a solver. of description problem Solving yields the this,Solving this problem yields the description of a valid solver.,3662
appearance intoxication. also noted test feeling of persons the a The alcohol of subjective,The test persons also noted the appearance of a subjective feeling of alcohol intoxication.,3663
are for Three procedures presented. and models constraint novel several shaving programming these models,Three novel constraint programming models and several shaving procedures for these models are presented.,3664
"approach heuristic was the Previously, only a algorithm. available","Previously, the only available approach was a heuristic algorithm.",3665
previous to protein linking of organisms. the aggregation summarizes This work heterozygosity paper,This paper summarizes previous work linking protein aggregation to the heterozygosity of organisms.,3666
inhibit heterozygosity aggregation. to protein can Organisms employ,Organisms can employ heterozygosity to inhibit protein aggregation.,3667
"diploid synthesize be to complex tend because Hence, aggregation-prone they proteins. organisms tend to long,","Hence, complex organisms tend to be diploid because they tend to synthesize long, aggregation-prone proteins.",3668
theory with level. also of explain may The organisms associated ecological trends ploidy different,The theory may also explain ecological trends associated with organisms of different ploidy level.,3669
with We for new observability. no present planning probabilistic a algorithm,We present a new algorithm for probabilistic planning with no observability.,3670
"the infeasible all however, This computation, but smallest problems. is for","This computation, however, is infeasible for all but the smallest problems.",3671
"approximate that analyze we allow various Therefore, for efficient Q-value computation. functions","Therefore, we analyze various approximate Q-value functions that allow for efficient computation.",3672
then gain a from We control. dynamical implements first that principles model design nonlinear,We then design from first principles a nonlinear dynamical model that implements gain control.,3673
"models equations. identification expressed to come differential Typically, the mean has this as quantitative of","Typically, this has come to mean the identification of quantitative models expressed as differential equations.",3674
"relationships by between the we machine of experimentally structural the variables. identification observed Specifically, examine","Specifically, we examine the identification by machine of the structural relationships between experimentally observed variables.",3675
form These expressed model. qualitative will of be relationship a quantitative of the in abstractions,These relationship will be expressed in the form of qualitative abstractions of a quantitative model.,3676
problem. planning decentralized as formally in can Multi-agent stochastic framed environments Markov a decision be,Multi-agent planning in stochastic environments can be framed formally as a decentralized Markov decision problem.,3677
an optimally mechanism show to is finding Dec-SMDP-Com. a equivalent optimal We solving that,We show that finding an optimal mechanism is equivalent to solving optimally a Dec-SMDP-Com.,3678
converges search provide on algorithm that the decomposition. heuristic a optimal We also,We also provide a heuristic search algorithm that converges on the optimal decomposition.,3679
results good provide show these that Empirical approximate solutions. approaches,Empirical results show that these approaches provide good approximate solutions.,3680
expressive queries an an Logics play language for Description (DLs). as Conjunctive important query role,Conjunctive queries play an important role as an expressive query language for Description Logics (DLs).,3681
"co-NP. complexity, data containment we Regarding in prove","Regarding data complexity, we prove containment in co-NP.",3682
with graphs. present three problems for complexity planning causal classes simple results We new of,We present three new complexity results for classes of planning problems with simple causal graphs.,3683
we hierarchies adding these extend and number Then role results finally restrictions. by,Then we extend these results by adding role hierarchies and finally number restrictions.,3684
intractable. combined Complete search becomes state-space soon the of,Complete search of the combined state-space soon becomes intractable.,3685
space a search in Planning the configurations. becomes smaller much of then subgraph,Planning then becomes a search in the much smaller space of subgraph configurations.,3686
blocks and are Web building initiative. the the automated of reasoning Ontologies Semantic,Ontologies and automated reasoning are the building blocks of the Semantic Web initiative.,3687
"disjunctive We terms, Q-resolution for introduce formulas normal form. used to be in on","We introduce Q-resolution on terms, to be used for formulas in disjunctive normal form.",3688
"signatures, for the algorithm. is data based on used Simulated evaluating proposed the measured spectral","Simulated data based on the measured spectral signatures, is used for evaluating the proposed algorithm.",3689
method state-of-the-art in We field. that the outperforms the method our show significantly,We show that our method significantly outperforms the state-of-the-art method in the field.,3690
birth presence processes. and of recruitment death model social in adaptive the We in networks,We model recruitment in adaptive social networks in the presence of birth and death processes.,3691
nodes recruited. subset a of can be susceptible Only,Only a susceptible subset of nodes can be recruited.,3692
system. compared direct are theoretical with simulations predictions of The the full,The theoretical predictions are compared with direct simulations of the full system.,3693
formulated Eiter al.} {\em et the logic (dl-programs) programs semantics by set Description under answer,Description logic programs (dl-programs) under the answer set semantics formulated by Eiter {\em et al.},3694
bases. been rules integrating considered formalism and as for a prominent ontology have knowledge,have been considered as a prominent formalism for integrating rules and ontology knowledge bases.,3695
"dl-programs logic. In study of this the into default paper, we embedding possibility","In this paper, we study the possibility of embedding dl-programs into default logic.",3696
"multiple current sufficient classes. slow or not can However, for be methods too distinguishing","However, current methods can be too slow or not sufficient for distinguishing multiple classes.",3697
"interoperability the importantly, semantic services. More and they also integration e-government facilitate of","More importantly, they also facilitate the semantic integration and interoperability of e-government services.",3698
using methods. a aggregation theory We mechanical statistical develop of,We develop a theory of aggregation using statistical mechanical methods.,3699
peptide/protein structures complicated of a is with An self-assembly. of several example system aggregation levels,An example of a complicated aggregation system with several levels of structures is peptide/protein self-assembly.,3700
used Euclidean measurement. is for similarity distance,Euclidean distance is used for similarity measurement.,3701
be space. calculations But the carried a in out finite-dimensional can,But the calculations can be carried out in a finite-dimensional space.,3702
linear paid penalties based operators. is attention Particular on to differential,Particular attention is paid to penalties based on linear differential operators.,3703
"In minimizer can sometimes calculate Green's functions. using the case, easily this explicitly, one","In this case, one can sometimes easily calculate the minimizer explicitly, using Green's functions.",3704
"order is in In preserve enrollment to consistency. other consistent words, system mandatory","In other words, consistent enrollment is mandatory in order to preserve system consistency.",3705
"assumption that process obtain are branching rare, bursty we innovations the speciations. of a Under","Under the assumption that innovations are rare, we obtain a bursty branching process of speciations.",3706
of rules determining and Lipophilicity are biological activities conformations of groups in functional interchangeably TCs.,Lipophilicity of functional groups and conformations interchangeably are determining rules in biological activities of TCs.,3707
affects complication encephalopathy memory. and a attention is that Hepatic neurological,Hepatic encephalopathy is a neurological complication that affects attention and memory.,3708
reversal in and were learning for tested rats water maze. Portacaval shunt Morris sham-operated the,Portacaval shunt and sham-operated rats were tested for reversal learning in the Morris water maze.,3709
for cytochrome were oxidase Brains histochemistry. processed,Brains were processed for cytochrome oxidase histochemistry.,3710
learning capacity the of measure dimension algorithms. is generalization fundamental Vapnik-Chervonenkis of a (VC),Vapnik-Chervonenkis (VC) dimension is a fundamental measure of the generalization capacity of learning algorithms.,3711
"hard analytically. from is impossible it or few to a apart calculate special However, cases,","However, apart from a few special cases, it is hard or impossible to calculate analytically.",3712
"in contexts, some voting In may the new of process. course candidates the show up","In voting contexts, some new candidates may show up in the course of the process.",3713
for reason and our They research. clue main are,They are our clue and main reason for research.,3714
their nature Here exposing investigate barriers we of process. and these mental features the,Here we investigate these barriers and their features exposing the nature of mental process.,3715
special the structures start between We objects. from to define reflect ways relations which,We start from special structures which reflect the ways to define relations between objects.,3716
be can mental a This dynamics. called,This can be called a mental dynamics.,3717
help can us be This to optimal. more,This can help us to be more optimal.,3718
use We algebra objects. and the relation invariant nature in stay abstract of to,We use abstract algebra and stay invariant in relation to the nature of objects.,3719
Facial years. recent Expression Classification in is problem research an interesting,Facial Expression Classification is an interesting research problem in recent years.,3720
are There problem. methods a solve this lot to of,There are a lot of methods to solve this problem.,3721
"preprocessing of for we region facial detection Firstly, images. in Canny local use phase,","Firstly, in preprocessing phase, we use Canny for local region detection of facial images.",3722
will (PCA). based features Component Analysis presented be Principal local Then region's on of each,Then each of local region's features will be presented based on Principal Component Analysis (PCA).,3723
"Expression Network Classification. Finally, for Artificial Neural using Facial (ANN)applies","Finally, using Artificial Neural Network (ANN)applies for Facial Expression Classification.",3724
The shows of method. the feasibility our experimental result proposal,The experimental result shows the feasibility of our proposal method.,3725
dense paper including main type Our this regularizations target in is \ellp-MKL.,Our main target in this paper is dense type regularizations including \ellp-MKL.,3726
of performances compare us to of generalization types the enables various regularizations. This,This enables us to compare the generalization performances of various types of regularizations.,3727
is as presented theory by the consistent with The known T.Y. outcome,The outcome is consistent with the known theory as presented by T.Y.,3728
higher dimensions. M. and Herman in gives et.al and insights new Khachan,Herman and M. Khachan et.al and gives new insights in higher dimensions.,3729
given proposed the the data transform distances utilizes The between points.,The proposed transform utilizes the distances between the given data points.,3730
have date studied IAMs as To of spaces. submanifolds Euclidean been embedded,To date IAMs have been studied as embedded submanifolds of Euclidean spaces.,3731
exact both mechanism can The is explain not that force-velocity known. yet curves,The exact mechanism that can explain both force-velocity curves is not yet known.,3732
simulation a agent-based such an mechanism. carried to stochastic explore out We,We carried out an agent-based stochastic simulation to explore such a mechanism.,3733
to decay possibility membranes balance recurrent with and is excitation. of synapses One modes,One possibility is to balance decay modes of membranes and synapses with recurrent excitation.,3734
"well-studied the we motion For task. dot random discrimination focus on concreteness,","For concreteness, we focus on the well-studied random dot motion discrimination task.",3735
robust leads mistuned. This to gains when performance becomes feedback integrators,This leads robust integrators to performance gains when feedback becomes mistuned.,3736
evolving this area. our perspective on paper This presents,This paper presents our perspective on this evolving area.,3737
have summarize results methods. We obtained that for been of these each the,We summarize the results that have been obtained for each of these methods.,3738
for given derived SDE ion kinetic explicitly the scheme. channel any We,We derived the SDE explicitly for any given ion channel kinetic scheme.,3739
"this the simulation demonstrated method of efficiency MC considerable over superiority Also, DA methods.","Also, the simulation efficiency of this DA method demonstrated considerable superiority over MC methods.",3740
"some do directly distances. determine to In proceed not clusters we directly cases, and the","In some cases, we proceed directly to clusters and do not directly determine the distances.",3741
We precision offer of insights on also measurement. of implications practical data,We offer insights also on practical implications of precision of data measurement.,3742
These are several well comply of with web defined processes. that composed services services,These services are composed of several web services that comply with well defined processes.,3743
method Graph based Our is on Theory.,Our method is based on Graph Theory.,3744
a the between graph. involved services through relationship the web We model semantic directed,We model the semantic relationship between the involved web services through a directed graph.,3745
"related Then, creates SWDs. it a for sub-graph most the","Then, it creates a sub-graph for the most related SWDs.",3746
"algorithm. returns by authorities document using and the It hubs the these HITS Finally, of","Finally, It returns the hubs and authorities of these document by using the HITS algorithm.",3747
classification We a consider binary standard problem.,We consider a standard binary classification problem.,3748
its dependence We also parameters classification margin the study of problem. on,We also study its dependence on margin parameters of the classification problem.,3749
behaviour. demands agent's an Solomonoff external identify We characterizing by on Induction principles,We identify principles characterizing Solomonoff Induction by demands on an agent's external behaviour.,3750
"and consistency. Key indifference computability, time are concepts rationality,","Key concepts are rationality, computability, indifference and time consistency.",3751
"case full derive we the to extensions AIXI. Furthermore, AI to discuss","Furthermore, we discuss extensions to the full AI case to derive AIXI.",3752
"value can The mean but the is variance change. is second case when constant, the","The second case is when the mean value is constant, but the variance can change.",3753
denoising. idea related variation This is to total closely,This idea is closely related to total variation denoising.,3754
"involved matrix and osteoclasts, resorbing are bone The cells producing osteoblasts mechanosensing bone osteocytes.","The cells involved are bone resorbing osteoclasts, bone matrix producing osteoblasts and mechanosensing osteocytes.",3755
other. consists each toggle mutually genes repress of that two A switch,A toggle switch consists of two genes that mutually repress each other.,3756
"protein mRNA We high binding. monomeric numbers, and focus on abundance factor transcription low","We focus on low mRNA numbers, high protein abundance and monomeric transcription factor binding.",3757
SGIM-D a on is evaluated skill learning experiment. fishing,SGIM-D is evaluated on a fishing skill learning experiment.,3758
proposed simulation of the algorithm. study the computational A shows behavior good,A simulation study shows the good computational behavior of the proposed algorithm.,3759
a of data in application An the using on real illustrates MCA. rotation set interest,An application on a real data set illustrates the interest of using rotation in MCA.,3760
"package codes ""PCAmixdata"". in All source are R the available","All source codes are available in the R package ""PCAmixdata"".",3761
"accurate. regions, For approximation rectangular integration the is highly","For rectangular integration regions, the approximation is highly accurate.",3762
case of also We the to more polyhedral integration the regions. general derivations extend,We also extend the derivations to the more general case of polyhedral integration regions.,3763
yet studied and elucidate interesting non-obvious results an detail. not in of These feature EP,These results elucidate an interesting and non-obvious feature of EP not yet studied in detail.,3764
costly. screening of consuming compound plants time be of and Pharmacology active medicinal would,Pharmacology screening of active compound of medicinal plants would be time consuming and costly.,3765
for useful approaches dimension be can These reduction and selection. variable then,These approaches can then be useful for dimension reduction and variable selection.,3766
specific of numerical for variables. have developed been Several clustering methods the,Several specific methods have been developed for the clustering of numerical variables.,3767
The package developed ClustOfVar purpose. was this R for specifically,The R package ClustOfVar was specifically developed for this purpose.,3768
with principal component obtained the method. synthetic PCAMIX is This the first variable,This synthetic variable is the first principal component obtained with the PCAMIX method.,3769
suitable to clusters. also We numbers of determine in approach order propose a bootstrap,We also propose a bootstrap approach in order to determine suitable numbers of clusters.,3770
associated and We methodologies the illustrate on small the package datasets.,We illustrate the methodologies and the associated package on small datasets.,3771
"the of the experiments, demonstrate Through we approach. usefulness proposed","Through experiments, we demonstrate the usefulness of the proposed approach.",3772
material. The supporting simulations is for code used distributed our R computer as,The R code used for our computer simulations is distributed as supporting material.,3773
the pool is feature objects. compute to A score used matching detected between two,A feature pool is used to compute the matching score between two detected objects.,3774
A noisy trajectory remove trajectories. filter is proposed to,A trajectory filter is proposed to remove noisy trajectories.,3775
component to inter-related. parts degree some are by that formed biological systems are Most,Most biological systems are formed by component parts that to some degree are inter-related.,3776
We selection degrees ratio signal-to-noise on of consequences of modularity varying the reconstruction. and explore,We explore the consequences of varying degrees of modularity and signal-to-noise ratio on selection reconstruction.,3777
"selection our controlling improves gradients. In vastly for matrices simulations, reconstruction of noise the","In our simulations, controlling matrices for noise vastly improves the reconstruction of selection gradients.",3778
counterintuitive the analysis unravels Mean-field astrocytes of mediate model that may effects.,Mean-field analysis of the model unravels that astrocytes may mediate counterintuitive effects.,3779
explicit do are they neither reactions terms. cross-inhibition The contain enantiomeric autocatalytic nor,The reactions are neither autocatalytic nor do they contain explicit enantiomeric cross-inhibition terms.,3780
a samples come into that from sites two fall number groupings. of These,These samples come from a number of sites that fall into two groupings.,3781
"types, from burials. derived samples burial The several kurgan including","The samples derived from several burial types, including kurgan burials.",3782
"dynamics We analyse terms metapopulation. in individual-based, of finite an metapopulation stochastic of model a","We analyse metapopulation dynamics in terms of an individual-based, stochastic model of a finite metapopulation.",3783
brink focus the extinction. metapopulations on We on of,We focus on metapopulations on the brink of extinction.,3784
to the estimate and extinction path describe likely extinction. the to time We most,We estimate the time to extinction and describe the most likely path to extinction.,3785
Exact in intractable. slab regression linear spike model and inference priors often with the is,Exact inference in the linear regression model with spike and slab priors is often intractable.,3786
inference. propagation be approximate can (EP) used for Expectation,Expectation propagation (EP) can be used for approximate inference.,3787
"As alternative, a convergent provably (PC-EP). algorithm EP we propose an","As an alternative, we propose a provably convergent EP algorithm (PC-EP).",3788
"when converges, similarly. By methods both contrast, perform R-EP","By contrast, when R-EP converges, both methods perform similarly.",3789
of chaos also of the algorithm. We propagation properties marginal establish,We also establish marginal propagation of chaos properties of the algorithm.,3790
high-dimensions is of SMC-based The discussed. schemes filtering some approximate accuracy also in,The accuracy in high-dimensions of some approximate SMC-based filtering schemes is also discussed.,3791
make preprocessor agents and learning properties unsupervised a for IncSFA autonomous These generally useful robots.,These properties make IncSFA a generally useful unsupervised preprocessor for autonomous learning agents and robots.,3792
fails. It where can handle SFA cases batch,It can handle cases where batch SFA fails.,3793
Diffusion for in Catalysts and that are Channels are sense Enzymes.,Channels are Catalysts for Diffusion and in that sense are Enzymes.,3794
useful idea interpretation experiments. the of This and for design is,This idea is useful for the design and interpretation of experiments.,3795
popular approaches these for Graph-based problems. are tools,Graph-based approaches are popular tools for these problems.,3796
an construction is graph-based of aspect learning. important Graph,Graph construction is an important aspect of graph-based learning.,3797
then our justify analysis. method cut through theoretically limit We,We then theoretically justify our method through limit cut analysis.,3798
"available the singularity a transformation unique of solution be However, may matrix. of because not","However, a unique solution may not be available because of the singularity of transformation matrix.",3799
"method has POCS been this overcome To used. problem,","To overcome this problem, POCS method has been used.",3800
"is their has been good the effect not of energy However, ignored. noise performance because","However, their performance is not good because the effect of noise energy has been ignored.",3801
significantly. Also the processing has time decreased,Also the processing time has decreased significantly.,3802
"recent we Letter this discuss in presented a [Phys. multiadaptive work, will a In model","In this work, we will discuss a multiadaptive model presented in a recent Letter [Phys.",3803
noise all-cooperator of state. destroy an can types Some,Some types of noise can destroy an all-cooperator state.,3804
"the the is are other so hubs hand, all-C stable, If, on then state.","If, on the other hand, hubs are stable, then so is the all-C state.",3805
for is framework a coupled The mean set-valued with stochastic asynchronous process extended approximation fields.,The framework is extended for a coupled asynchronous stochastic approximation process with set-valued mean fields.,3806
is learning process. Markov applicability of The a demonstrated approach through in decision this,The applicability of this approach is demonstrated through learning in a Markov decision process.,3807
"a to a over range conditions. of expected Thus, metric some loss may correspond operating","Thus, a metric may correspond to some expected loss over a range of operating conditions.",3808
"shaped has the biology, of life. of the increasingly history cooperative In evolution groups","In biology, the evolution of increasingly cooperative groups has shaped the history of life.",3809
on understand cooperation which Biological behavior. foundation human a to provides,Biological cooperation provides a foundation on which to understand human behavior.,3810
"biological informs evolution trace in of Historically, the two ways.","Historically, the trace of biological evolution informs in two ways.",3811
the in difficult. Matching to incorporate information new ability flexibility and recognition remains quickly animal-like,Matching animal-like flexibility in recognition and the ability to quickly incorporate new information remains difficult.,3812
yet neural addressed Limits recognition models adequately to are and in algorithms. be,Limits are yet to be adequately addressed in neural models and recognition algorithms.,3813
public benefits good. members all act that individually is costly An group a,An individually costly act that benefits all group members is a public good.,3814
"First, to goods. to selfish individual variation favors their allocation in individuals resources vary public","First, variation in individual resources favors selfish individuals to vary their allocation to public goods.",3815
"in it important. obvious is is Stated baseline of that the success way, role this","Stated in this way, it is obvious that the role of baseline success is important.",3816
"of sometimes sufficiently. this to public problems goods discussions However, point fail emphasize","However, discussions of public goods problems sometimes fail to emphasize this point sufficiently.",3817
and baseline the for simple models roles The here of resource variation success. tests suggest,The models here suggest simple tests for the roles of resource variation and baseline success.,3818
"use limited of these permit distributions. with variables continuous very random languages Consequently,","Consequently, these languages permit very limited use of random variables with continuous distributions.",3819
Practice Logic Programming). Theory of in (To and appear,(To appear in Theory and Practice of Logic Programming).,3820
unsupervised in fundamental is a learning. task Clustering,Clustering is a fundamental task in unsupervised learning.,3821
our algorithms real and synthetic both We compare on data. existing to methods,We compare our algorithms to existing methods on both synthetic and real data.,3822
in success group-structured Individual populations has two components.,Individual success in group-structured populations has two components.,3823
"individual resources. gains for neighbors its by outcompeting an First, local","First, an individual gains by outcompeting its neighbors for local resources.",3824
the of tragedy is This the commons.,This is the tragedy of the commons.,3825
"generalisation relax We exhibit a bound. of consistence, the notion the of and logarithmic","We relax the notion of consistence, and exhibit a generalisation of the logarithmic bound.",3826
"we study get results, variants Bounds Upper policies. popular To these (ucb) Confidence of","To get these results, we study variants of popular Upper Confidence Bounds (ucb) policies.",3827
task fundamental in chromosome segmentation. A human analysis chromosome is,A fundamental task in human chromosome analysis is chromosome segmentation.,3828
an important role Segmentation karyotyping. plays chromosome in,Segmentation plays an important role in chromosome karyotyping.,3829
The in performs scheme proposed segmentation two phases.,The proposed scheme performs segmentation in two phases.,3830
"other of quite on efficient separation clusters. in method, is the such hand, chromosome Our","Our method, on the other hand, is quite efficient in separation of such chromosome clusters.",3831
"find ideal agents (IFD) distribution optimal, resources. nearly We free of repeatedly a to","We repeatedly find a nearly optimal, ideal free distribution (IFD) of agents to resources.",3832
"for are solutions Finally, parameter possible. spatially which regimes we oscillating uniform find","Finally, we find parameter regimes for which spatially uniform oscillating solutions are possible.",3833
concept technique features has information a from recognition tasks. various standard image become Combining in,Combining information from various image features has become a standard technique in concept recognition tasks.,3834
promote Classical mixtures. approaches sparse to MKL,Classical approaches to MKL promote sparse mixtures.,3835
About PLoS to be ONE. submitted to,About to be submitted to PLoS ONE.,3836
from Canada. Mexico southern The annually to migrates central monarch butterfly,The monarch butterfly annually migrates from central Mexico to southern Canada.,3837
"its recent their human been reduced with decades, During population habitat. due to has interaction","During recent decades, its population has been reduced due to human interaction with their habitat.",3838
data show DNA memory. represents We assembly that structure low accurately in this graphs,We show that this data structure accurately represents DNA assembly graphs in low memory.,3839
withdrawn paper The been by authors. the has,The paper has been withdrawn by the authors.,3840
important traffic the an role in plays at A intersections. signalised control vehicle detection,A vehicle detection plays an important role in the traffic control at signalised intersections.,3841
a for This vision-based algorithm vehicles paper in presence detection introduces recognition zones.,This paper introduces a vision-based algorithm for vehicles presence recognition in detection zones.,3842
variables local to input attributes The algorithm evaluate an uses of image. linguistic,The algorithm uses linguistic variables to evaluate local attributes of an input image.,3843
"image background attributes The features. as or categorised vehicle, unknown are","The image attributes are categorised as vehicle, background or unknown features.",3844
vision-based presented. this method recognition In paper is a vehicles,In this paper a vision-based vehicles recognition method is presented.,3845
method using was implemented with fuzzy reasoning The proposed rules. system,The proposed method was implemented using reasoning system with fuzzy rules.,3846
"distorted and transform reference to new called images. applied A ""Tetrolet"" both adaptive is","A new adaptive transform called ""Tetrolet"" is applied to both reference and distorted images.",3847
coefficients Forms density of To marginal proposed. is the distribution tetrolet Bessel model (BKF) K,To model the marginal distribution of tetrolet coefficients Bessel K Forms (BKF) density is proposed.,3848
is between good presented human with judgment. A consistency comparison a showing measures these,A comparison between these measures is presented showing a good consistency with human judgment.,3849
corresponding branch length). given to a of process,corresponding to a process of given branch length).,3850
class models. the models nonhomogeneous These evolutionary of lie in,These models lie in the class of nonhomogeneous evolutionary models.,3851
based Our on algebraic tools. mainly are techniques,Our techniques are mainly based on algebraic tools.,3852
graph-theory methods These principally are or machine on learning. based,These methods are principally based on graph-theory or machine learning.,3853
The this is diffusion. of loop inner max-sum algorithm,The inner loop of this algorithm is max-sum diffusion.,3854
of inference with task probabilistic We the probabilistic models. performing consider logical,We consider the task of performing probabilistic inference with probabilistic logical models.,3855
models are for such Many inference based algorithms on with approximate sampling.,Many algorithms for approximate inference with such models are based on sampling.,3856
"these redundancy calls. the in is static there larger the repeated The more part,","The larger the static part, the more redundancy there is in these repeated calls.",3857
approximations. poor is sampling problematic since inefficient yields This,This is problematic since inefficient sampling yields poor approximations.,3858
program to sampling-based We how inference to more efficient. specialization show apply logic make,We show how to apply logic program specialization to make sampling-based inference more efficient.,3859
"and namely, building two objectives classification. of approximation the problem, examine manifold We","We examine two objectives of the manifold building problem, namely, approximation and classification.",3860
and sample identification an to Prior registering palmprint classification important a is activity.,Prior to classification and identification registering a sample palmprint is an important activity.,3861
constantly to conditions. fluctuating are exposed environmental Cells,Cells are constantly exposed to fluctuating environmental conditions.,3862
"network. human that transduction typical a a we Boolean of To end, signal model use","To that end, we use a Boolean model of a typical human signal transduction network.",3863
identification of has years. research recent Personal been field in problem a major,Personal identification problem has been a major field of research in recent years.,3864
"major and features are palm The of wrinkles lines, principal including the ridges. palm-lines,","The major features of the palm are palm-lines, including principal lines, wrinkles and ridges.",3865
the feature Another is energy robust wavelet palms. of,Another robust feature is the wavelet energy of palms.,3866
both paper combines In these this features. used feature a we which hybrid of,In this paper we used a hybrid feature which combines both of these features.,3867
"improve is to system. analysis the %Moreover, of performance applied multispectral the","%Moreover, multispectral analysis is applied to improve the performance of the system.",3868
"in hindered their practice. applications their have to further challenge However, severely due been computational","However, their further applications have been severely hindered due to their computational challenge in practice.",3869
"with paper, transformation consider setting manifolds this of parameterization. In a we known","In this paper, we consider a setting with transformation manifolds of known parameterization.",3870
an improves all performance additional on joint the classification optimization samples Performing further.,Performing an additional joint optimization on all samples improves the classification performance further.,3871
propagation  constraint. RegularGcc We the of study  global,  We study propagation of the RegularGcc global constraint.,3872
"side, two cases fixed we parameter On identify tractable. is positive where propagation the","On the positive side, we identify two cases where propagation is fixed parameter tractable.",3873
these conditions some We weighted enforce row additional automata. with,We enforce these conditions with some additional weighted row automata.,3874
these methods on demonstrate the some benchmark Experimental results of potential standard problems.,Experimental results demonstrate the potential of these methods on some standard benchmark problems.,3875
a unorganized curves fitting propose planar We points. to novel B-spline data method for,We propose a novel method for fitting planar B-spline curves to unorganized data points.,3876
"result, faster methods. a much method our than As is existing","As a result, our method is much faster than existing methods.",3877
"position. of Cells way direct embryo ""measuring"" have their in no physical developing a","Cells in a developing embryo have no direct way of ""measuring"" their physical position.",3878
"Furthermore, the the characteristics influence study model. of neuron of we","Furthermore, we study the influence of characteristics of the neuron model.",3879
ratio of is Further compounds. efficacious the most evaluation to determine two the combination recommended,Further evaluation is recommended to determine the most efficacious combination ratio of the two compounds.,3880
provides statement and proof of a paper This the complete Theorem.,This paper provides a complete statement and proof of the Theorem.,3881
heuristic operators. process of experience relies FLC majority design of knowledge on The,The majority of FLC design process relies on heuristic knowledge of experience operators.,3882
confirming Computer logic the controller. the performance presented fuzzy simulations are of constructed,Computer simulations are presented confirming the performance of the constructed fuzzy logic controller.,3883
possible this. We two for main isolate scenarios,We isolate two main possible scenarios for this.,3884
"to Alternatively, parameters be defining some need of growth a rate. function the activity promoter","Alternatively, some parameters defining promoter activity need to be a function of the growth rate.",3885
"Bayesian Roy. for computation, statistics computation: approximate J. Constructing approx- Bayesian summary Semi-automatic imate","Constructing summary statistics for approximate Bayesian computation: Semi-automatic approx- imate Bayesian computation, J. Roy.",3886
map. represented as of each regions image The salient depth are extracted a and firstly,The salient regions of each image are firstly extracted and represented as a depth map.,3887
used the is Genetic efficient here algorithm for very A optimization.,A very efficient Genetic algorithm is used here for the optimization.,3888
the The the method. results proposed superior experimental of showed performance,The experimental results showed the superior performance of the proposed method.,3889
image shape electrostatic can exact field on force perform an based recovery. model The,The image force based on electrostatic field model can perform an exact shape recovery.,3890
in term each Snakes It proves is has function. similar that GAC correspondent and and,It proves that each term in GAC and Snakes is correspondent and has similar function.,3891
"However, expressed mathematics. the models two different are using","However, the two models are expressed using different mathematics.",3892
switch and Bayes. teaching We to to  statistics undergraduates frequentist cease  should,  We should cease teaching frequentist statistics to undergraduates and switch to Bayes.,3893
over-certainty so users confusion the statistics. of rife Doing of and among reduce amount will,Doing so will reduce the amount of confusion and over-certainty rife among users of statistics.,3894
by active geometry their Cells of contractility. and sense adhesive stiffness the environment,Cells sense the geometry and stiffness of their adhesive environment by active contractility.,3895
the for we of Here compare models. different distribution shape force such network and variants,Here we compare the shape and force distribution for different variants of such network models.,3896
"resting networks, contractility the For passive mechanical of by modeled a is links. length reduced","For passive networks, contractility is modeled by a reduced resting length of the mechanical links.",3897
So the to is up adopted in today. it style rarely automatic,So it is rarely adopted in the automatic style up to today.,3898
shown experiments. is our segmentation in quality the competitive at It,It is shown competitive at the segmentation quality in our experiments.,3899
"with approach. Specifically, beta Bayesian with models deals regression a mixed effects from it","Specifically, it deals with beta regression models with mixed effects from a Bayesian approach.",3900
done amplitudes be multiresolution with is using that group. shown a analysis It transformation can,It is shown that multiresolution analysis can be done with amplitudes using a transformation group.,3901
Separation of frequency amplitude facilitated G-lets. low by and is components of high each resolution,Separation of low and high frequency components of each amplitude resolution is facilitated by G-lets.,3902
"Applications face more. extraction, this edge compression, denoising, include feature algorithm and for detection, recognition,","Applications for this algorithm include edge detection, feature extraction, denoising, face recognition, compression, and more.",3903
an We algorithm as groups using example. dihedral this analyze,We analyze this algorithm using dihedral groups as an example.,3904
We ECG results the `Lena' an and signal with image. standard the demonstrate,We demonstrate the results with an ECG signal and the standard `Lena' image.,3905
noise. salt The proposed scheme reduces and pepper efficiently identifies and,The proposed scheme efficiently identifies and reduces salt and pepper noise.,3906
"empirical clustering the an in data, Such corresponds to model ultrametric hierarchical e.g.","Such an ultrametric model corresponds to hierarchical clustering in the empirical data, e.g.",3907
epidemiological the Some methods other are discussed. of extension systems for,Some extension of the methods for other epidemiological systems are discussed.,3908
of the approach variations model its This is any and noise against parameters. robust,This approach is robust against any variations of the noise model and its parameters.,3909
the descent method an We the steepest via algorithm. use solution find iterative sparse to,We use the steepest descent method to find the sparse solution via an iterative algorithm.,3910
"additional frequency was By the obtained. these concerning networks, considering information associations several of","By considering several networks, additional information concerning the frequency of these associations was obtained.",3911
"two algorithm, that, Mau\'a Lauritzen-Nilsson we al. After consider et algorithms, and","After that, we consider two algorithms, Lauritzen-Nilsson algorithm, and Mau\'a et al.",3912
no proposed feature knowledge approach noise of of prior that the Main is required.,Main feature of the proposed approach is that no prior knowledge of noise required.,3913
developed Stefan- It law and based Boltzmann is law. on Fourier the,It is developed based on Stefan- Boltzmann law and the Fourier law.,3914
We investigate advantage computation also approach the load. less that reduction has of noise fast,We also investigate the fast noise reduction approach that has advantage of less computation load.,3915
reducing reduction comparison This of approach time. has the noise advantage by fast the,This approach has the advantage of fast noise reduction by reducing the comparison time.,3916
This medicine images applications approach will and Infrared video for conferencing. extend the,This approach will extend the Infrared images applications for medicine and video conferencing.,3917
This truth-valued concept linguistic lattice. work continuation about paper our research is the of,This paper is the continuation of our research work about linguistic truth-valued concept lattice.,3918
method Gabor spatiotemporal filters. texture presents This a dynamic based paper for on new recognition,This paper presents a new method for dynamic texture recognition based on spatiotemporal Gabor filters.,3919
"response, calculating each For energy a vector is built by feature statistic. the","For each response, a feature vector is built by calculating the energy statistic.",3920
are extract to relevant approaches information of shape. a various There,There are various approaches to extract relevant information of a shape.,3921
"comparison database both shape experimental Through a classified, previously methods are compared. using","Through experimental comparison using a shape database previously classified, both methods are compared.",3922
steps. main two by The described can be proposed method,The proposed method can be described by two main steps.,3923
"the on of based maps. images, brightness possibilities the there variations are produce To several","To produce the variations of images, there are several possibilities based on the brightness maps.",3924
of functional visibility analogous is the visibility The fringe interference. to the optical,The visibility functional is analogous to the fringe visibility of the optical interference.,3925
last relabeling by the image this at input CRF with cluster. model the associated Then,Then at last relabeling the input image by the CRF model associated with this cluster.,3926
presents a approach segment to image This correct and specific information. having label paper,This paper presents a approach to label and segment specific image having correct information.,3927
modern the thousands in Millions belong of families'. `gene Motivation: genes species only of to,Motivation: Millions of genes in the modern species belong to only thousands of `gene families'.,3928
and Genes during gained evolution. lost are,Genes are gained and lost during evolution.,3929
trees is binary specific and Reconciling simple. gene,Reconciling binary gene and specific trees is simple.,3930
"studied gene species is reconciliation non-binary in and binary refinement tree a model. Here,","Here, non-binary gene and species tree reconciliation is studied in a binary refinement model.",3931
two current context the makes In contributions. this paper,In this context the current paper makes two contributions.,3932
in and (RNNs) neuroscience networks learning neural widely applications. are used machine Recurrent computational,Recurrent neural networks (RNNs) are widely used in computational neuroscience and machine learning applications.,3933
"equations generative this we derive update that Bayesian output. Given can model, RNN decode its","Given this generative RNN model, we derive Bayesian update equations that can decode its output.",3934
"scheme implements inputs. coding dynamic Furthermore, for a predictive it","Furthermore, it implements a predictive coding scheme for dynamic inputs.",3935
the application the online an use the of (i.e. to illustrate decoding We rRNN by,We illustrate the use of the rRNN by an application to the online decoding (i.e.,3936
This considerably. enables to the compressive measurement us lower rate,This enables us to lower the compressive measurement rate considerably.,3937
"of demonstrate applications effectiveness the these the approach. Together,","Together, these applications demonstrate the effectiveness of the approach.",3938
outlined. work results shown are Initial future and,Initial results are shown and future work outlined.,3939
derived into Each a is family image sub-images. of decomposed $f$,Each image $f$ is decomposed into a family of derived sub-images.,3940
"texture After Gabor enhancing image, the features the attribute computed $v$ descriptors. of are as","After enhancing the texture attribute $v$ of the image, Gabor features are computed as descriptors.",3941
on with validate two variability. proposed We the approach datasets texture high,We validate the proposed approach on two texture datasets with high variability.,3942
evaluate our We an important leaf-texture real-world approach on application: analysis. also,We also evaluate our approach on an important real-world application: leaf-texture analysis.,3943
"two methods of suspicious Then have MCCs, been for segmentation investigated.","Then for segmentation of suspicious MCCs, two methods have been investigated.",3944
adaptive considered are: watershed and segmentation. The methods threshold,The considered methods are: adaptive threshold and watershed segmentation.,3945
cluster to results framework the new us allows concerning points. proximal prove The,The proximal framework allows us to prove new results concerning the cluster points.,3946
a first on builds geographical based information. network This road modeled simulation detailed,This simulation first builds a modeled road network based on detailed geographical information.,3947
"populations Transporters and the creates this the of simulation two On Mobiles. agents: the network,","On this network, the simulation creates two populations of agents: the Transporters and the Mobiles.",3948
circulate the agents vehicles on that embody Mobile network. the,Mobile agents embody the vehicles that circulate on the network.,3949
Trust Region} {\it update the implementation We using Mor\'{e}'s present algorithm of strategy. an,We present an implementation of the algorithm using Mor\'{e}'s {\it Trust Region} update strategy.,3950
non-quadratic to inverse applied with problem the Poisson distributed is a data. method For illustration,For illustration the method is applied to a non-quadratic inverse problem with Poisson distributed data.,3951
essential satisfaction is constraint definition problems. in distinction our The of,The distinction is essential in our definition of constraint satisfaction problems.,3952
reals. over we application satisfaction As problems treat constraint,As application we treat constraint satisfaction problems over reals.,3953
"significantly numerical of cover and These part large analysis, inequalities. nonlinear equations most a","These cover a large part of numerical analysis, most significantly nonlinear equations and inequalities.",3954
critical) modeled conditioned branching The a latter by is (supercritical or process.,The latter is modeled by a (supercritical or critical) conditioned branching process.,3955
settings. classical This paper an an to alternative as these contributes approach,This paper contributes an approach as an alternative to these classical settings.,3956
a Thresholding in major of binarization images. plays,Thresholding plays a major in binarization of images.,3957
thresholding and Thresholding can local be categorized into global thresholding.,Thresholding can be categorized into global thresholding and local thresholding.,3958
"is cases, appropriate. binarization thresholding In local with more such","In such cases, binarization with local thresholding is more appropriate.",3959
the local computational on size. mean the window time Normally depends,Normally the local mean computational time depends on the window size.,3960
to sum prior processing a technique as integral image mean. uses Our calculate local,Our technique uses integral sum image as a prior processing to calculate local mean.,3961
involve techniques. calculations standard not deviations other as local of adaptive It in does,It does not involve calculations of standard deviations as in other local adaptive techniques.,3962
"convergence situations, problems. slow some algorithm shows EM In","In some situations, EM algorithm shows slow convergence problems.",3963
reason that is update parameters possible simultaneously. One standard procedures the,One possible reason is that standard procedures update the parameters simultaneously.,3964
finite estimation. In focus on paper mixture this we,In this paper we focus on finite mixture estimation.,3965
"In EM, framework, parameters which propose a updates component-wise this the sequentially. we","In this framework, we propose a component-wise EM, which updates the parameters sequentially.",3966
calculation shows presented here [...]. that The,The calculation presented here shows that [...].,3967
"two-dimensional an algorithm. accept-reject we design In case, the","In the two-dimensional case, we design an accept-reject algorithm.",3968
is one in used the segmentation. most image The watershed tools of,The watershed is one of the most used tools in image segmentation.,3969
over and born We developed time. present concept how is its,We present how its concept is born and developed over time.,3970
"regular have more ribozymes the biggest populations. having rates, Despite smaller the growth","Despite having smaller growth rates, the more regular ribozymes have the biggest populations.",3971
self-incompatibility still is populations systems maintained debated issue. plant a How are in,How self-incompatibility systems are maintained in plant populations is still a debated issue.,3972
maintenance studies in the the self-incompatibility. of asexual explored reproduction have of function Other,Other studies have explored the function of asexual reproduction in the maintenance of self-incompatibility.,3973
by self-incompatibility system heterozygote enhanced small populations. This the in excess was,This heterozygote excess was enhanced by the self-incompatibility system in small populations.,3974
hardware trade-off detail Large-scale the neuromorphic systems chip level bear and resources. between required typically,Large-scale neuromorphic hardware systems typically bear the trade-off between detail level and required chip resources.,3975
"that be modification saves design, resources would reducing resolution. By a weight synaptic natural","By design, a natural modification that saves resources would be reducing synaptic weight resolution.",3976
represent hardware-constrained general framework setting a configuring for results synapses. Our and up,Our results represent a general framework for setting up and configuring hardware-constrained synapses.,3977
for discretization simulations. weight considered We be could to other how backends suggest large-scale dedicated,We suggest how weight discretization could be considered for other backends dedicated to large-scale simulations.,3978
are and stability. to Additional for speed improvements proposed SURF,Additional improvements to SURF are proposed for speed and stability.,3979
experimentally tightly studied process Organogenesis that regulated for has been is decades. a,Organogenesis is a tightly regulated process that has been studied experimentally for decades.,3980
"We are limbs, the for studying of development bone. models and kidneys, lungs, mechanistic currently","We are currently studying mechanistic models for the development of limbs, lungs, kidneys, and bone.",3981
"meaning, its i.e. often been While embodiment used trivial has in","While embodiment has often been used in its trivial meaning, i.e.",3982
A number case studies presented to are illustrate of the concept.,A number of case studies are presented to illustrate the concept.,3983
"concentrated visual animals grasping, perception. and locomotion, are and around and These robots involve","These involve animals and robots and are concentrated around locomotion, grasping, and visual perception.",3984
"will link low-level and between processes the establish sensory-motor we cognition. a Finally,","Finally, we will establish a link between the low-level sensory-motor processes and cognition.",3985
contexts We reinforcement of competition. several and several analyse algorithms in cooperation,We analyse several reinforcement algorithms in several contexts of cooperation and competition.,3986
distribution. is by inspired the experimental setting Darwin-Wallace Our recently developed,Our experimental setting is inspired by the recently developed Darwin-Wallace distribution.,3987
mechanisms complex includes of most developed action. all miRNA known The by model us,The most complex model developed by us includes all known mechanisms of miRNA action.,3988
from one experimental limiting another. vary can setting to The place,The limiting place can vary from one experimental setting to another.,3989
of explains the majority existing model reported. controversies This,This model explains the majority of existing controversies reported.,3990
be in mining examples can typical found Some applications. and genomics web,Some typical examples can be found in web mining and genomics applications.,3991
results on web obtained mining are datasets. State-of-art benchmark,State-of-art results are obtained on benchmark web mining datasets.,3992
experiments. models simulate of series from We community perturbation developed dynamics a derived mathematical that,We developed mathematical models that simulate community dynamics derived from a series of perturbation experiments.,3993
an Forest These in were experiments Atlantic stream. performed,These experiments were performed in an Atlantic Forest stream.,3994
was press submitted trophic level to of perturbation experiment. two community combinations The three,The three trophic level community was submitted to two combinations of press perturbation experiment.,3995
"the electrical by In level exclusion. was first, trophic the top removed","In the first, the top trophic level was removed by electrical exclusion.",3996
mechanisms. to were the repeated confidence observed Experiments in times different at increase,Experiments were repeated at different times to increase confidence in the observed mechanisms.,3997
cascade. the a existence cascading indicated interactions trophic Community of similar to classic responses,Community responses indicated the existence of cascading interactions similar to a classic trophic cascade.,3998
the successfully of community. qualitative models reproduced The the responses,The models successfully reproduced the qualitative responses of the community.,3999
are number approximate We an visible have images. of in counted cars that these also,We have also counted an approximate number of cars that are visible in these images.,4000
"to In propose problem. we this ""Model this method a paper, tackle Segmentation"" Assisted","In this paper, we propose a ""Model Assisted Segmentation"" method to tackle this problem.",4001
fully method requires user and automatic is interaction. The no,The method is fully automatic and requires no user interaction.,4002
"prior any the system require training. not Also, does","Also, the system does not require any prior training.",4003
results We a photographs of our on present car. real,We present our results on photographs of a real car.,4004
is in fact. well-established cognition process Context-dependence human a,Context-dependence in human cognition process is a well-established fact.,4005
was Result used convolution of to them. classify,Result of convolution was used to classify them.,4006
the data all image-dataset. training Moreover selected within were points images from the,Moreover the training data points were selected from all the images within image-dataset.,4007
"and main stages: hyper are there heuristic move Generally, two heuristic in selection framework, acceptance.","Generally, in hyper heuristic framework, there are two main stages: heuristic selection and move acceptance.",4008
paper Heuristic This stage on develop framework. Hyper emphasizes latter the (HH) to,This paper emphasizes on the latter stage to develop Hyper Heuristic (HH) framework.,4009
methods the have network to communication and the secure infrastructure developed been Internet. over Many,Many methods have been developed to secure the network infrastructure and communication over the Internet.,4010
addition Intrusion detection new a to relatively is such techniques.,Intrusion detection is a relatively new addition to such techniques.,4011
day big is is amount by One increasing Intrusion of problem day. which,One big problem is amount of Intrusion which is increasing day by day.,4012
"analysing IDS, network know then to effect. need about We using information attack the","We need to know about network attack information using IDS, then analysing the effect.",4013
AI of Introduction of intrusions. the necessitates importance in normalization,Introduction of AI necessitates the importance of normalization in intrusions.,4014
"similar are typically Starfish in pentameral pentagram. they typically and show shape a symmetry, to","Starfish typically show pentameral symmetry, and they are typically similar in shape to a pentagram.",4015
"In five-armed starfish autotomy, which starfish for is addition, perform crucially best on important survival.","In addition, five-armed starfish perform best on autotomy, which is crucially important for starfish survival.",4016
"conformation Nevertheless, flexibility and in can see arms. of the some we number","Nevertheless, we can see some flexibility in the number and conformation of arms.",4017
derive such quantum apply a to formula. We statistics,We apply quantum statistics to derive such a formula.,4018
"For self only simplicity, globular monomeric covered. proteins are folding","For simplicity, only monomeric self folding globular proteins are covered.",4019
immediate point out of applications formula. the We some,We point out some immediate applications of the formula.,4020
explains We very show the phenomena observed that well. the formula,We show that the formula explains the observed phenomena very well.,4021
"Intelligent Knowledge Multidimensional Agents, Discovery. Techniques, Data Mining Profile, User Visualization, KEYWORDS:","KEYWORDS: Data Mining Techniques, Intelligent Agents, User Profile, Multidimensional Visualization, Knowledge Discovery.",4022
Determining a clusters task. optimal dataset of a challenging is in number,Determining optimal number of clusters in a dataset is a challenging task.,4023
"methods there some that algorithm unique solution. available, produces no Though is are clustering","Though some methods are available, there is no algorithm that produces unique clustering solution.",4024
"infectivity in This individuals. is model, e.g., natural of susceptibility a heterogeneity or to framework","This is a natural framework to model, e.g., heterogeneity in susceptibility or infectivity of individuals.",4025
"We theory, results necessary with review, approach. along the using the obtained discussed the","We review, along with the necessary theory, the results obtained using the discussed approach.",4026
 is Spectrum cognitive a  radio. sensing in problem fundamental,  Spectrum sensing is a fundamental problem in cognitive radio.,4027
is analysis threshold for setting Theoretical algorithm of discussed. the,Theoretical analysis of threshold setting for the algorithm is discussed.,4028
schemes proposed combining probability recent and years. Many first-order been in logic have representation,Many representation schemes combining first-order logic and probability have been proposed in recent years.,4029
and slower. has logical Progress in probabilistic unifying inference been,Progress in unifying logical and probabilistic inference has been slower.,4030
variable lifted greatly show Experiments outperforms it is logical structure when elimination present. that,Experiments show that it greatly outperforms lifted variable elimination when logical structure is present.,4031
evaluations to our significant over speedup alternatives. Experimental algorithms the show provide,Experimental evaluations show our algorithms to provide significant speedup over the alternatives.,4032
symmetry and the from derived form major forms the other Pentameral is are it.,Pentameral symmetry is the major form and the other forms are derived from it.,4033
"believed However, which bilaterians. period, were the from to Cambrian originated echinoderms, be of ancestors","However, the ancestors of echinoderms, which originated from Cambrian period, were believed to be bilaterians.",4034
bilateral are during early development. larvae Echinoderm their,Echinoderm larvae are bilateral during their early development.,4035
"possess symmetrical also controls cluster, development. Starfish which Hox gene the","Starfish also possess the Hox gene cluster, which controls symmetrical development.",4036
"and a bilateral process. to Overall, developmental are mechanism echinoderms have thought","Overall, echinoderms are thought to have a bilateral developmental mechanism and process.",4037
"tendency. adult this to its corroborate bilateral focused starfish we In on article, behaviors","In this article, we focused on adult starfish behaviors to corroborate its bilateral tendency.",4038
"when EM. EDML especially interesting behaviors, exhibits to compared","EDML exhibits interesting behaviors, especially when compared to EM.",4039
The features geometric of the proposed the character system extracts contour.,The proposed system extracts the geometric features of the character contour.,4040
that forms on basic the the This based types features line skeleton. are character,This features are based on the basic line types that forms the character skeleton.,4041
feature gives output. as system its vector The a,The system gives a feature vector as its output.,4042
algorithm performing We inference models. MAP distributed anytime present in graphical for a,We present a distributed anytime algorithm for performing MAP inference in graphical models.,4043
as the linear formulated problem The over is relaxation graph. a programming of a edges,The problem is formulated as a linear programming relaxation over the edges of a graph.,4044
allows that constraint application The program Dantzig-Wolfe structure of decomposition the principle. resulting has a,The resulting program has a constraint structure that allows application of the Dantzig-Wolfe decomposition principle.,4045
be computed manner. over a individual are and can edges distributed defined in Subprograms,Subprograms are defined over individual edges and can be computed in a distributed manner.,4046
solutions not in graphs This memory. fit does to space state accommodates whose,This accommodates solutions to graphs whose state space does not fit in memory.,4047
the A to outcome formal the final more observations. directly infer approach from is,A more formal approach is to directly infer the final outcome from the observations.,4048
in potentials. consists models and of multiplying repeatedly summing graphical out Inference,Inference in graphical models consists of repeatedly multiplying and summing out potentials.,4049
an propose We potentials: their method values. quantizing simplifying for alternate,We propose an alternate method for simplifying potentials: quantizing their values.,4050
to decision use this algebraic do (ADDs) efficiently. diagrams We,We use algebraic decision diagrams (ADDs) to do this efficiently.,4051
tagged-fish and the daylight. varying strengths We the aggregation current of characterized to response,We characterized the response of the tagged-fish aggregation to varying current strengths and daylight.,4052
decision This decision focuses possibilistic trees. making in sequential paper on,This paper focuses on sequential decision making in possibilistic decision trees.,4053
"specifically to solve the optimisation framework, is this problem. SCIP used Integer programming,","Integer programming, specifically the SCIP framework, is used to solve this optimisation problem.",4054
fast particularly for this learning. is Results a exact that method BN show,Results show that this is a particularly fast method for exact BN learning.,4055
sciences. a controlling and empirical is Identifying key in problem bias,Identifying and controlling bias is a key problem in empirical sciences.,4056
making sheds under on models This result of the light uncertainty. decision,This result sheds light on the models of decision making under uncertainty.,4057
"is to achieve. difficult while often Bayes-optimal behavior, well-defined,","Bayes-optimal behavior, while well-defined, is often difficult to achieve.",4058
been face Recently proposed based recognition for the classification robust representation (SRC) sparse (FR). has,Recently the sparse representation based classification (SRC) has been proposed for robust face recognition (FR).,4059
"on computational constraint sparsity high. coefficients SRC's very coding the makes Meanwhile, cost the","Meanwhile, the sparsity constraint on the coding coefficients makes SRC's computational cost very high.",4060
joint the optimal policy as Finding optimization can problem. be an formulated,Finding the optimal joint policy can be formulated as an optimization problem.,4061
"lack too However, optimality are and/or guarantees. formulations existing verbose","However, existing formulations are too verbose and/or lack optimality guarantees.",4062
propose compact Integer We Mixed EDI-CR a instances. Program formulation Linear of,We propose a compact Mixed Integer Linear Program formulation of EDI-CR instances.,4063
variables. and us to allows fewer use similarly treat This sequences these,This allows us to treat these sequences similarly and use fewer variables.,4064
"population, studies interaction, Such etc. identification biological for social of is useful migration,","Such identification is useful for biological studies of population, social interaction, migration, etc.",4065
such preliminary that But present bounds we results possible. are show,But we present preliminary results that show such bounds are possible.,4066
"in computational abstraction, reductions exponential may applicable, when Hierarchical complexity. problem offer","Hierarchical problem abstraction, when applicable, may offer exponential reductions in computational complexity.",4067
"Viterbi abstraction the show In how apply paper, problem. to temporal this to we","In this paper, we show how to apply temporal abstraction to the Viterbi problem.",4068
processing topic an in image important field. the image is segmentation Color,Color image segmentation is an important topic in the image processing field.,4069
We study in negotiation the combinatorial of problem agent-based domains.,We study the problem of agent-based negotiation in combinatorial domains.,4070
in situations. agreements inefficient such Self-interested up often agents end negotiating,Self-interested agents often end up negotiating inefficient agreements in such situations.,4071
Markov decision widely stochastic used decision environments. are (MDPs) in making in problems processes modeling,Markov decision processes (MDPs) are widely used in modeling decision making problems in stochastic environments.,4072
"However, in precise very specification of difficult. MDPs functions reward the often is","However, precise specification of the reward functions in MDPs is often very difficult.",4073
also present method further for of the version up. an speed We approximate,We also present an approximate version of the method for further speed up.,4074
improves orders algorithm We our experimentally magnitude. that the by performance demonstrate of,We experimentally demonstrate that our algorithm improves the performance by orders of magnitude.,4075
that tasks. these is efficient contribution we develop this inference of for The paper algorithms,The contribution of this paper is that we develop efficient inference algorithms for these tasks.,4076
"methods. To such state-of-the-art employ we tasks, solve","To solve such tasks, we employ state-of-the-art methods.",4077
to smart pulse-based analysis. We pattern provide recognition stochastic-logic an design efficient blocks,We design smart pulse-based stochastic-logic blocks to provide an efficient pattern recognition analysis.,4078
specific proposed is The to rchitecture a applied navigation problem.,The proposed rchitecture is applied to a specific navigation problem.,4079
magnitude orders system resulting solutions. The than of processor-based faster is,The resulting system is orders of magnitude faster than processor-based solutions.,4080
"neuron equations. by is individual FitzHugh-Nagumo modeled each On level, the","On the individual level, each neuron is modeled by FitzHugh-Nagumo equations.",4081
"uptake. involved in roots, In various zinc ZIP are plant transporters","In plant roots, various ZIP transporters are involved in zinc uptake.",4082
"thought involved. unknown, is to bZIP largely are but be transcription Their factors regulation","Their regulation is largely unknown, but bZIP transcription factors are thought to be involved.",4083
"activator/inhibitor. up and We with activator activator three models: dimerization set only, putative","We set up three putative models: activator only, activator with dimerization and activator/inhibitor.",4084
measurements to analyzed. fitted were These and,These were fitted to measurements and analyzed.,4085
low has or in self-similarity frequencies. memory A long properties scale-invariant process,A long memory process has self-similarity or scale-invariant properties in low frequencies.,4086
many proteomic coupled techniques. separation can to MS be,MS can be coupled to many proteomic separation techniques.,4087
"However, is point. a chains analytical technological variability on these critical controlling","However, controlling technological variability on these analytical chains is a critical point.",4088
"hierarchical analytical of LC-MS a In model chain. the we Method : this introduce communication,","Method : In this communication, we introduce a hierarchical model of the LC-MS analytical chain.",4089
a mixture This model. hierarchical molecular defines graph,This molecular graph defines a hierarchical mixture model.,4090
"step. will the an consider digestion As we example,","As an example, we will consider the digestion step.",4091
this We in was that diel demonstrated amplified FAD. temporal the leading pattern,We demonstrated that this diel temporal pattern was amplified in the leading FAD.,4092
"has standard in rendering some effectiveness. However, limitations isosurface","However, standard isosurface rendering has some limitations in effectiveness.",4093
"(SDC) microdata, control disclosure of of statistical purpose a.k.a. The","The purpose of statistical disclosure control (SDC) of microdata, a.k.a.",4094
This of overview provides informal an UAI article in context.,This article provides an informal overview of UAI in context.,4095
unfavorable and consists of alternation a The periodic of habitat patches. favorable,The habitat consists of a periodic alternation of favorable and unfavorable patches.,4096
the fixed rate. In dies the at latter population,In the latter the population dies at fixed rate.,4097
"invasion may advection, the However, we intense when enough, process. activate show that","However, we show that advection, when intense enough, may activate the invasion process.",4098
doing in density. tractable density so the intractable In the transition of place emerges noise,In doing so the tractable noise density emerges in place of the intractable transition density.,4099
a of of We few the review evolution. of models the genealogies properties statistical,We review the statistical properties of the genealogies of a few models of evolution.,4100
these way of times are simple introducing how discuss We in modified selection. a,We discuss how these times are modified in a simple way of introducing selection.,4101
differentiated slow found processes. We the from emerge indeed the promoter states binding/unbinding can,We found indeed the differentiated states can emerge from the slow promoter binding/unbinding processes.,4102
"slow multiple we promoter Furthermore, found meta-stable under binding/unbinding, states. differentiated","Furthermore, under slow promoter binding/unbinding, we found multiple meta-stable differentiated states.",4103
can the explain of in experiments. origin the multiple recent observed This states,This can explain the origin of multiple states observed in the recent experiments.,4104
"In kinetic and quantify addition, the reprogramming. we differentiation pathways for","In addition, we quantify kinetic pathways for the differentiation and reprogramming.",4105
This dynamics in cells. stem multipotent progenitor the non-equilibrium captures or,This captures the non-equilibrium dynamics in multipotent stem or progenitor cells.,4106
"motif that suggest noisy we flow information a in maximizes network. Here, this","Here, we suggest that this motif maximizes information flow in a noisy network.",4107
"at In TFs, of case up-stream latter cross-regulation. the criterion hints the","In the case of TFs, the latter criterion hints at up-stream cross-regulation.",4108
nodes. consider We two with  of  types networks,  We consider networks with two types of nodes.,4109
"networks, central- in possesses fundamental widespread properties. two architecture, ized gene This","This central- ized architecture, widespread in gene networks, possesses two fundamental properties.",4110
"perturbations system. respect random the this with is organization Moreover, to of robust","Moreover, this organization is robust with respect to random perturbations of the system.",4111
be edge accurate. These assume techniques image to the,These techniques assume the edge image to be accurate.,4112
perceptually based presents techniques the This fusion important Wavelet features. Discrete combining for paper image,This paper presents the Discrete Wavelet based fusion techniques for combining perceptually important image features.,4113
modifications the SPIHT This on some presents algorithm. paper,This paper presents some modifications on the SPIHT algorithm.,4114
object ray-casting inner feature are on estimation based point. Existing projection from techniques the,Existing object projection feature estimation techniques are based on ray-casting from the inner point.,4115
"develop In this these numerical we approximations for quantities. paper,","In this paper, we develop numerical approximations for these quantities.",4116
approximation sparse for We describe coding. method of a fast,We describe a method for fast approximation of sparse coding.,4117
averaging cascades and transform operators. with convolutions It modulus non-linear wavelet,It cascades wavelet transform convolutions with non-linear modulus and averaging operators.,4118
"is heuristics inadequate. infrastructure existing However, often for search defining","However, existing infrastructure for defining search heuristics is often inadequate.",4119
"remain unexplored. a performance may As major result, in improvements","As a result, major improvements in performance may remain unexplored.",4120
from spatially a technique low-resolution observed SR is image to high-resolution images. a multiple estimate,SR is a technique to estimate a spatially high-resolution image from observed multiple low-resolution images.,4121
preserves natural compound images model edges. A provides Gaussian preferable for MRF that prior a,A compound Gaussian MRF model provides a preferable prior for natural images that preserves edges.,4122
(PSNR). for objective the estimator ratio the is signal-to-noise peak function of optimal PM,PM is the optimal estimator for the objective function of peak signal-to-noise ratio (PSNR).,4123
variational determined estimator is (VB). numerically using This by Bayes,This estimator is numerically determined by using variational Bayes (VB).,4124
"methods. method existing the experiments, In roughly overcomes proposed","In experiments, the proposed method roughly overcomes existing methods.",4125
technique. of the with showing provide the experiments performance good real We images,We provide experiments with real images showing the good performance of the technique.,4126
This been discrimination has used as tasks image framework for such classification. also,This framework has also been used for discrimination tasks such as image classification.,4127
"exploit problem. algorithm online derived an we the Then, to solve","Then, we exploit an online algorithm to solve the derived problem.",4128
"be In to set credal is particular, neither convex closed. nor assumed the","In particular, the credal set is assumed to be neither convex nor closed.",4129
"test against our it to As model, biological applications of we observations. more applied further","As further applications of our model, we applied it to test against more biological observations.",4130
partitioning describe such this approach new data. paper high-dimensional we a for In,In this paper we describe a new approach for partitioning such high-dimensional data.,4131
processing done is using simulation MATLAB complete The image model.,The complete image processing is done using MATLAB simulation model.,4132
"law. is transformation Morphological as Finally, follows Weber's organized and as paper this","Finally, this paper is organized as follows as Morphological transformation and Weber's law.",4133
quite are lower expensive. that radiation Systems risk generally have,Systems that have lower radiation risk are generally quite expensive.,4134
pedicle position orientation the screw fixation. drill and of clinically important is very The in,The position and orientation of the drill is clinically very important in pedicle screw fixation.,4135
"of fragments of a we a bacteria consider set. mixture certain this In research, genome","In this research, we consider a mixture of genome fragments of a certain bacteria set.",4136
material are the Fields by capturing real-world properties The defined Reflectance objects. the of,The material properties are defined by capturing the Reflectance Fields of the real-world objects.,4137
fusion for of Shenoy's algorithm discrete algorithm influence is The diagrams. an solution extension,The solution algorithm is an extension of Shenoy's fusion algorithm for discrete influence diagrams.,4138
and solving framework examples. two are The algorithm small illustrated by,The algorithm and framework are illustrated by solving two small examples.,4139
possibilistic knowledge is of a result often in which a formulas. set weighted base They,They often result in a possibilistic knowledge base which is a set of weighted formulas.,4140
"logic Although the inconsistency suers from possibilistic well-known is ""drowning tolerant, it effect"".","Although possibilistic logic is inconsistency tolerant, it suers from the well-known ""drowning effect"".",4141
These of change principle capture postulates minimal some the sense. in,These postulates capture the principle of minimal change in some sense.,4142
"of logic MAP introduces parameterization networks. Markov paper inference a inference, The k-bounded in MAP","The paper introduces k-bounded MAP inference, a parameterization of MAP inference in Markov logic networks.",4143
by are aids improve Rollating popular walkers to mobility used older control. balance adults,Rollating walkers are popular mobility aids used by older adults to improve balance control.,4144
and walker techniques recognize design compare activities. to We several related,We design and compare several techniques to recognize walker related activities.,4145
presented. comprehensive retirement evaluation community and walker a is users subjects from with A control,A comprehensive evaluation with control subjects and walker users from a retirement community is presented.,4146
define sets answer then generalized PASP to characterization programs. is of This naturally,This characterization is then naturally generalized to define answer sets of PASP programs.,4147
which particular world represents classes. everything real object In an is,In real world everything is an object which represents particular classes.,4148
Every described can be its object by fully attributes.,Every object can be fully described by its attributes.,4149
attributes of objects. real dataset number contains and Any world large,Any real world dataset contains large number of attributes and objects.,4150
result of shows efficiency the proposed The the method.,The result shows the efficiency of the proposed method.,4151
"BLOG, for well-known language. been a has OUPM It implemented","It has been implemented for BLOG, a well-known OUPM language.",4152
combinator that has prove several We this properties. good,We prove that this combinator has several good properties.,4153
sequential framework Decentralized expressive multi-agent an for provide decision POMDPs making.,Decentralized POMDPs provide an expressive framework for multi-agent sequential decision making.,4154
that solvers. compares on Experiments domains EM benchmark show the against favorably state-of-the-art,Experiments on benchmark domains show that EM compares favorably against the state-of-the-art solvers.,4155
"equidistantly nodes on and the the image. algorithms distribute However, uniformly graph's graph-based","However, graph-based algorithms distribute the graph's nodes uniformly and equidistantly on the image.",4156
"is a term added to cut Then, to prefer smoothness shape. a the force particular","Then, a smoothness term is added to force the cut to prefer a particular shape.",4157
"of are the when indistinguishable This can areas background. from object be the useful, strategy","This strategy can be useful, when areas of the object are indistinguishable from the background.",4158
a as consciousness by defined of propose data. relations We theory partial typical,We propose a partial theory of consciousness as relations defined by typical data.,4159
both theoretically and experimentally. proposed can applied tested theory be The and,The proposed theory can be applied and tested both theoretically and experimentally.,4160
fully is typical using established discrete how Precisely mathematics. relations determines data,Precisely how typical data determines relations is fully established using discrete mathematics.,4161
"the analytically. over filter, observations marginalize can EKF Since is the we first-order a","Since the EKF is a first-order filter, we can marginalize over the observations analytically.",4162
designed and that to an DecRSPI lack problems improve model. is explicit scalability tackle,DecRSPI is designed to improve scalability and tackle problems that lack an explicit model.,4163
Monte- a of uses Carlo to algorithm The sample states. reachable methods belief generate,The algorithm uses Monte- Carlo methods to generate a sample of reachable belief states.,4164
belief computes based it for state on a estimations. the rollout Then joint each policy,Then it computes a joint policy for each belief state based on the rollout estimations.,4165
A to policy us represent allows new representation solutions compactly.,A new policy representation allows us to represent solutions compactly.,4166
larger that existing intractable can for It problems algorithms. solve are planning,It can solve larger problems that are intractable for existing planning algorithms.,4167
effectiveness confirm and the of results scalability Experimental approach. the,Experimental results confirm the effectiveness and scalability of the approach.,4168
known as is This or block-modelling block-clustering.,This is known as block-modelling or block-clustering.,4169
integrated (SBM) The block parameters out. the blockmodel stochastic with model is,The model is the stochastic blockmodel (SBM) with block parameters integrated out.,4170
cluster The memberships. a clusters over the resulting number defines marginal distribution of posterior and,The resulting marginal distribution defines a posterior over the number of clusters and cluster memberships.,4171
allocation on The based sampler. the is algorithm,The algorithm is based on the allocation sampler.,4172
to compute them. it Our makes reformulation possible,Our reformulation makes it possible to compute them.,4173
how the procedure We done and is and this illustrate its results. show,We show how this is done and illustrate the procedure and its results.,4174
ways. unique is in Our two algorithm,Our algorithm is unique in two ways.,4175
"with lifted the First, it variables efficiency improves substantially inference domains. of continuous of","First, it substantially improves the efficiency of lifted inference with variables of continuous domains.",4176
"algorithm which way. Second, a it handles is the lifted exact first inference in RCMs","Second, it is the first exact inference algorithm which handles RCMs in a lifted way.",4177
econometrics. illustrated algorithm The example an over is from,The algorithm is illustrated over an example from econometrics.,4178
"formulas. conventional dual variable-based Unlike in our logical realm algorithms algorithms, work of the","Unlike conventional variable-based algorithms, our algorithms work in the dual realm of logical formulas.",4179
reconstruction interesting of With cameras the amount optimal large strategy. determine to it a becomes,With a large amount of cameras it becomes interesting to determine the optimal reconstruction strategy.,4180
how different changes considering error cameras. reconstruction by the,how the reconstruction error changes by considering different cameras.,4181
"work, In reconstruction derived. error is of the this approximation variance an first,","In this work, first, an approximation of the reconstruction error variance is derived.",4182
"introduce The into System Air cockpit. will sensor Generation the Next technologies advanced Transportation new,","The Next Generation Air Transportation System will introduce new, advanced sensor technologies into the cockpit.",4183
call different pilot and alerting will hazards automation plans. Different states for,Different hazards and pilot states will call for different alerting automation plans.,4184
explores alternative using compilation This algorithms techniques. paper,This paper explores alternative algorithms using compilation techniques.,4185
adaptations compilation-based first propose of We methods. possibilistic probabilistic standard,We first propose possibilistic adaptations of standard compilation-based probabilistic methods.,4186
by results. is also result experimental This confirmed,This result is also confirmed by experimental results.,4187
"quadratic present particular, a kernel. In we","In particular, we present a quadratic kernel.",4188
construction enumerative Inference based logic languages in these over is of programs. proofs on,Inference in these languages is based on enumerative construction of proofs over logic programs.,4189
"Consequently, of languages permit with variables these distributions. random use continuous very limited","Consequently, these languages permit very limited use of random variables with continuous distributions.",4190
The generalizes learning Networks. Bayesian used for PRISM and algorithm the naturally ones Hybrid,The learning algorithm naturally generalizes the ones used for PRISM and Hybrid Bayesian Networks.,4191
strategy struggle to winning rapidly. coevolutionary One is a evolve for,One strategy for winning a coevolutionary struggle is to evolve rapidly.,4192
"ecology. the Speciation of is the room"" community in ""elephant","Speciation is the ""elephant in the room"" of community ecology.",4193
"ignored speciation often its dimension. Yet, is spatial of or completely stripped","Yet, speciation is often completely ignored or stripped of its spatial dimension.",4194
allowed to theory model have based effectively Recent on complex network approaches ecologists landscapes.,Recent approaches based on network theory have allowed ecologists to effectively model complex landscapes.,4195
longer males. gave Females than estimates consistently,Females consistently gave longer estimates than males.,4196
had difference This to however mathematics. in in correlation do overall skills with nothing,This difference in correlation however had nothing to do with overall skills in mathematics.,4197
two latent We unknown blurring the a image with create by CBP kernels. same,We create a CBP by blurring the same latent image with two unknown kernels.,4198
bivariate two domain. kernels polynomials the coprime in are z when mapped The to,The two kernels are coprime when mapped to bivariate polynomials in the z domain.,4199
the organism depending and on The growth may objective conditions. cellular its vary,The cellular objective may vary depending on the organism and its growth conditions.,4200
metabolic efficiency of coli. a growth using We Eschericia model the for examined properties,We examined the properties of growth efficiency using a metabolic model for Eschericia coli.,4201
that obtained growth rate. efficiency the found maximal uptake nutrient a We is finite at,We found that the maximal growth efficiency is obtained at a finite nutrient uptake rate.,4202
content-based is Research effective place find algorithms taking image to representation for description. and,Research is taking place to find effective algorithms for content-based image representation and description.,4203
"substantial There is that shape, texture). a available of algorithms amount visual features (color, use","There is a substantial amount of algorithms available that use visual features (color, shape, texture).",4204
paper This (KDFPE). Kernel Points presents Feature Density object using representation shape an Estimator,This paper presents an object shape representation using Kernel Density Feature Points Estimator (KDFPE).,4205
then image. The is the applied to the vector of KDFPE,The KDFPE is then applied to the vector of the image.,4206
"rotation. invariant KDFPE to scale translation, is and","KDFPE is invariant to translation, scale and rotation.",4207
and to making happened wherever whenever Decision key is success. is it organizations,Decision making whenever and wherever it is happened is key to organizations success.,4208
"this paper, collaboration convergence. knowledge and In the management explain we","In this paper, we explain the knowledge management and collaboration convergence.",4209
"process the profile. UML-G MDM explicit we based on Finally,","Finally, we explicit the MDM process based on UML-G profile.",4210
the in by continued throughput. These challenged improvement sequencing algorithms are,These algorithms are challenged by the continued improvement in sequencing throughput.,4211
for available and modification. implementation is Our freely use,Our implementation is freely available for use and modification.,4212
"in datasets computational and general, method in to this We apply it discuss biology. some","We discuss this method in general, and apply it to some datasets in computational biology.",4213
is work magnification into an the of This investigation of accuracy the effects DFD. on,This work is an investigation into the effects of magnification on the accuracy of DFD.,4214
comment issues computation. on effect regarding relative blur scaling on We,We comment on issues regarding scaling effect on relative blur computation.,4215
Motivation: existing rely sequence or genotypes. accurate sequences DNA for on analysis methods Most,Motivation: Most existing methods for DNA sequence analysis rely on accurate sequences or genotypes.,4216
sequencing or discovery). multi-sample somatic low-coverage mutation,multi-sample low-coverage sequencing or somatic mutation discovery).,4217
sequence the with data new of analyzing development uncertainty. for applications These methods for press,These applications press for the development of new methods for analyzing sequence data with uncertainty.,4218
can INDELs genome. called against a and reference SNPs the by mapping unitigs be,SNPs and INDELs can be called by mapping the unitigs against a reference genome.,4219
sensitivity higher It other methods assembly based novo variant calling. has de for than,It has higher sensitivity than other de novo assembly based methods for variant calling.,4220
system This expert Troubleshooter called diagnosing paper for an problems. proposes Expert PC computer,This paper proposes an expert system called Expert PC Troubleshooter for diagnosing computer problems.,4221
phase paper response external to lagging brain stimuli. model a In of this detail we,In this paper we detail a phase lagging model of brain response to external stimuli.,4222
physics basic like is of using The law. the derived laws conservation model energy of,The model is derived using the basic laws of physics like conservation of energy law.,4223
eliminates the in the This instantaneous paradox of the action propagation potential model brain. of,This model eliminates the paradox of instantaneous propagation of the action potential in the brain.,4224
model then solution this is presented. of The,The solution of this model is then presented.,4225
"penalization terms proven have have However, strong to properties. other sparsity-inducing","However, other penalization terms have proven to have strong sparsity-inducing properties.",4226
note ends The challenges work. paper on the and the a with future,The paper ends with a note on the challenges and the future work.,4227
is since the diseases. proteins of the main Ever for cause disorder many,Ever since the disorder of proteins is the main cause for many diseases.,4228
which certain mental reported disorders fall category. in are this There to,There are certain reported mental disorders which fall in to this category.,4229
"margins reduction controlled vectors. maximizing and Generalization, by are weight the dimensionality minimizing","Generalization, dimensionality reduction and maximizing the margins are controlled by minimizing weight vectors.",4230
system biometric by multimodal good pattern paper. in proposed this Results show,Results show good pattern by multimodal biometric system proposed in this paper.,4231
an Gender of attribute people. important demographic is,Gender is an important demographic attribute of people.,4232
computer of paper This provides in human vision. survey recognition a gender,This paper provides a survey of human gender recognition in computer vision.,4233
of We survey representative the these approaches. methods challenges highlight the and faced,We highlight the challenges faced and survey the representative methods of these approaches.,4234
evolutionary change into equation components. total The partitions Price two,The Price equation partitions total evolutionary change into two components.,4235
component an selection. natural abstract expression of The first provides,The first component provides an abstract expression of natural selection.,4236
"component transmission. all The other second including evolutionary changes during subsumes processes,","The second component subsumes all other evolutionary processes, including changes during transmission.",4237
used often component natural is selection The applications. in,The natural selection component is often used in applications.,4238
expression simplicity of Those interest of their interpretation. attract applications widespread and for ease,Those applications attract widespread interest for their simplicity of expression and ease of interpretation.,4239
critiques I van and recent discuss by Veelen. the equation Price of Nowak,I discuss recent critiques of the Price equation by Nowak and van Veelen.,4240
induces to lacking. resilience over control Behavioral stimuli stressful future is control when conditions,Behavioral control over stressful stimuli induces resilience to future conditions when control is lacking.,4241
"It of (e.g. animals the field tasks started transportation deploying agriculture(bulls), in to automate with","It started with deploying animals to automate tasks in the field of agriculture(bulls), transportation (e.g.",4242
"and (pigeons). horses even and donkeys), communication","horses and donkeys), and even communication (pigeons).",4243
"suggest of a implementation real machine. we life cognitive will a And Finally,","And Finally, we will suggest a real life implementation of a cognitive machine.",4244
finite segmentation on paper This a based Bayesian algorithm introduces image mixtures.,This paper introduces a Bayesian image segmentation algorithm based on finite mixtures.,4245
developed the of is EM to estimate mixtures. Gaussian An algorithm parameters,An EM algorithm is developed to estimate parameters of the Gaussian mixtures.,4246
finite flexible mixture probabilistic The a tool. powerful modeling and is,The finite mixture is a flexible and powerful probabilistic modeling tool.,4247
provide the pattern can field in It clustering model-based of used be to a recognition.,It can be used to provide a model-based clustering in the field of pattern recognition.,4248
the bits proposed in to has been two But it method. extended,But it has been extended to two bits in the proposed method.,4249
over The of experiments images intensities. were standard carried varying,The experiments were carried over standard images of varying intensities.,4250
"are purely functions The key are computational, the benefits that i.e.","The key benefits are that the functions are purely computational, i.e.",4251
requires a and Bangla Developing algorithm OCR of methods. bunch,Developing a Bangla OCR requires bunch of algorithm and methods.,4252
developing were effort went There a on Bangla OCR. for many,There were many effort went on for developing a Bangla OCR.,4253
OCR. an But failed error to free Bangla provide of them all,But all of them failed to provide an error free Bangla OCR.,4254
existing currently problem scope about of OCR's. the Bangla We discussed,We discussed about the problem scope of currently existing Bangla OCR's.,4255
"pre-processing different and classification. into divided analysis, Our is system steps:","Our system is divided into different steps: analysis, pre-processing and classification.",4256
image. of The operation is filtering the first,The first operation is the filtering of image.,4257
"we border the remove and stamps graphics. In second, print, the","In the second, we remove the border print, stamps and graphics.",4258
based two on pre-processing steps. The approach and basic the are modeling system,The pre-processing system and the modeling approach are based on two basic steps.,4259
recognition is Graph-matching on system based The algorithm.,The recognition system is based on Graph-matching algorithm.,4260
"For and portion segmentation used. are operations image morphological face localization, processing","For face portion segmentation and localization, morphological image processing operations are used.",4261
"may Unfortunately, and a do catchment partition. not overlap basins form","Unfortunately, catchment basins may overlap and do not form a partition.",4262
on We surfaces consider first weighted graphs. topographic node defined,We first consider topographic surfaces defined on node weighted graphs.,4263
constructs operations simple iterative with basins. performs An algorithm pruning and neighborhood the catchment the,An iterative algorithm with simple neighborhood operations performs the pruning and constructs the catchment basins.,4264
algorithm is to tone The adapted gray then images.,The algorithm is then adapted to gray tone images.,4265
catchment of and extend erosions basins. graph prune pair dilations adaptative and the A the,A pair of adaptative erosions and dilations prune the graph and extend the catchment basins.,4266
detection space-based paper feature and matching present chromaticity This algorithm. new our intensity,This paper present our new intensity chromaticity space-based feature detection and matching algorithm.,4267
"searching conventional the No reducing thus is expected complexity. computational is required, it algorithm","No conventional searching algorithm is required, thus it is expected reducing the computational complexity.",4268
given recognising representation We human system present of activity content. video for a a symbolic,We present a system for recognising human activity given a symbolic representation of video content.,4269
"the some results could assumptions estimation actually be violated. if are But, distorted","But, the estimation results could be distorted if some assumptions actually are violated.",4270
using data. effectiveness of demonstrate We artificial the our method,We demonstrate the effectiveness of our method using artificial data.,4271
"value Error Propagation(EBP). Singular Keywords- back decomposition(SVD),","Keywords- Singular value decomposition(SVD), Error back Propagation(EBP).",4272
"to medium image a project, In way matching description unique is as proposed. this","In this project, a unique way to image description as matching medium is proposed.",4273
highly is the analogous which method distinctiveness concept human of DNA unique. is This to,This distinctiveness of method is analogous to the human DNA concept which is highly unique.,4274
is and datasets The return open results. evaluated promising system on,The system is evaluated on open datasets and return promising results.,4275
Support their naturally use to loss vector (SVMs) sparseness embody machines functions. of due hinge,Support vector machines (SVMs) naturally embody sparseness due to their use of hinge loss functions.,4276
"class However, probabilities. conditional directly not estimate can SVMs","However, SVMs can not directly estimate conditional class probabilities.",4277
in logit function hinge It bridges and the logistic the function regression.,It bridges the hinge function and the logit function in logistic regression.,4278
"independent random methods a these source Conventionally, variates. require of","Conventionally, these methods require a source of independent random variates.",4279
"As to required obtain results. random sometimes far side-effect, numbers fewer accurate are a","As a side-effect, sometimes far fewer random numbers are required to obtain accurate results.",4280
are widespread Food natural with in webs predation habitats. intraguild (IGP),Food webs with intraguild predation (IGP) are widespread in natural habitats.,4281
is of behaviour restructuring communities. resilience for ecological understanding and Their adaptation principal,Their adaptation and resilience behaviour is principal for understanding restructuring of ecological communities.,4282
is resource subpopulations. subdivided The distinct into,The resource is subdivided into distinct subpopulations.,4283
different rates subpopulations the are in at predators. grazed Individuals the by,Individuals in the subpopulations are grazed at different rates by the predators.,4284
parameters IGP the robust We a investigate system. in which support coexistence,We investigate which parameters support a robust coexistence in the IGP system.,4285
computation the of center surround The specific determine and strength relative achieved. the,The relative strength of the center and surround determine the specific computation achieved.,4286
the dLGN in rat. we of center-surround the architecture revisit Here neurons un-anesthetized,Here we revisit the center-surround architecture of dLGN neurons in the un-anesthetized rat.,4287
receptive (DOG) these to the curves fit model tuning of a field. spatial difference-of-Gaussians We,We fit these tuning curves to a difference-of-Gaussians (DOG) model of the spatial receptive field.,4288
can the We awake strong dLGN that the surround be rat. in of conclude antagonism,We conclude that surround antagonism can be strong in the dLGN of the awake rat.,4289
robustness Experimental effectiveness demonstrate challenging the the tracker. videos and results of proposed on,Experimental results on challenging videos demonstrate the effectiveness and robustness of the proposed tracker.,4290
susceptibility increased malaria infection. Pregnant women have to,Pregnant women have increased susceptibility to malaria infection.,4291
Mechanisms in women not are increased responsible understood. for susceptibility pregnant fully,Mechanisms responsible for increased susceptibility in pregnant women are not fully understood.,4292
model. by have malaria the been suitable Pregnancy animal of lack a studies limited,Pregnancy malaria studies have been limited by the lack of a suitable animal model.,4293
a studying research baboon to This PM. model anubis) aimed for develop (Papio,This research aimed to develop a baboon (Papio anubis) model for studying PM.,4294
placenta. P. in findings new baboon These sequesters knowlesi the show that,These new findings show that P. knowlesi sequesters in the baboon placenta.,4295
studying human been Thus a primate PM model non established. for (baboon) has,Thus a non human primate (baboon) model for studying PM has been established.,4296
understood. pregnant women for increased to responsible fully in malaria susceptibility not Mechanisms are,Mechanisms responsible for increased susceptibility to malaria in pregnant women are not fully understood.,4297
"on cells sera, done was of weekly a then mononuclear Collection basis. peripheral and","Collection of peripheral sera, and mononuclear cells was then done on a weekly basis.",4298
a basis. cell mononuclear done were proliferation Peripheral assays recall blood on weekly also,Peripheral blood mononuclear cell recall proliferation assays were also done on a weekly basis.,4299
study processes. interacting the Coupling in stochastic technique of a is theoretical widely used,Coupling is a widely used technique in the theoretical study of interacting stochastic processes.,4300
"""lifted techniques inference"" Numerous proposed. have been such for approaches","Numerous approaches for such ""lifted inference"" techniques have been proposed.",4301
We to separate amplification two decomposition the W a Schur use of mechanisms.,We use a Schur decomposition of W to separate the two amplification mechanisms.,4302
weak. must limited that find We be slowing dynamical if is amplification very kept,We find that amplification is very limited if dynamical slowing must be kept weak.,4303
samples. random consider from set learning a problem of the We,We consider the problem of learning a set from random samples.,4304
"challenging, IRP solution is of even metaheuristics. for the biobjective The rather","The solution of the biobjective IRP is rather challenging, even for metaheuristics.",4305
profound effective neighborhood structures. representations lacking are solution appropriate a of understanding We and still,We are still lacking a profound understanding of appropriate solution representations and effective neighborhood structures.,4306
understanding of work the contributes such solution Our representations. to better,Our work contributes to the better understanding of such solution representations.,4307
"margins are dimensionality reduction controlled weight Generalization, maximizing minimizing and vectors. by the","Generalization, dimensionality reduction and maximizing the margins are controlled by minimizing weight vectors.",4308
multimodal show good biometric by proposed paper. pattern this system in Results,Results show good pattern by multimodal biometric system proposed in this paper.,4309
"discussed. by large maximum inference sample are likelihood estimation for and Furthermore,","Furthermore, estimation by maximum likelihood and inference for large sample are discussed.",4310
for We an clustering. Carlo Monte based hierarchical coalescent Bayesian new sampler sequential present,We present an new sequential Monte Carlo sampler for coalescent based Bayesian hierarchical clustering.,4311
modeling Our non-i.i.d. model appropriate is for,Our model is appropriate for modeling non-i.i.d.,4312
on data closely and show Results real approaches. artificial related over significant improvements,Results on artificial and real data show significant improvements over closely related approaches.,4313
"data in proteomics, contains correlation complicated genomics, High-dimensional often common chemometrics structures. and","High-dimensional data common in genomics, proteomics, and chemometrics often contains complicated correlation structures.",4314
the point period of At oscillations bifurcation the the diverges.,At the bifurcation point the period of the oscillations diverges.,4315
"of disposes datasets geographic Today, of large of composed one's thousands objects.","Today, one's disposes of large datasets composed of thousands of geographic objects.",4316
"this In necessary. methods robust context, become sampling","In this context, robust sampling methods become necessary.",4317
"on method this we techniques. propose paper, In sampling clustering based a","In this paper, we propose a sampling method based on clustering techniques.",4318
method shows select that objects. to our allows of samples This case-study relevant,This case-study shows that our method allows to select relevant samples of objects.,4319
"the standard In supervised learning, data is the only labeled observed.","In the standard supervised learning, only the labeled data is observed.",4320
as are supervised problems The formalized learning. regression the classification and,The classification and regression problems are formalized as the supervised learning.,4321
"labeled data. unlabeled semi-supervised obtained In to addition also learning, in data is","In semi-supervised learning, unlabeled data is also obtained in addition to labeled data.",4322
"is in unlabeled to learning. improve exploiting data semi-supervised accuracy prediction Hence, the important","Hence, exploiting unlabeled data is important to improve the prediction accuracy in semi-supervised learning.",4323
problems missing estimation problem with as is This data. regarded semiparametric a,This problems is regarded as a semiparametric estimation problem with missing data.,4324
usefulness of present experiments methods. our Some the numerical,Some numerical experiments present the usefulness of our methods.,4325
In priors. a propose novel this for sparsity-inducing the of construction framework paper we,In this paper we propose a novel framework for the construction of sparsity-inducing priors.,4326
"sparse can nonconvex framework with mixtures a Furthermore, penalization. Bayesian subserve Laplace scale for learning","Furthermore, Laplace scale mixtures can subserve a Bayesian framework for sparse learning with nonconvex penalization.",4327
can be of densities EP-GIG explicitly expressed. The,The densities of EP-GIG can be explicitly expressed.,4328
"distribution posterior distribution. corresponding the inverse Moreover, generalized follows a also Gaussian","Moreover, the corresponding posterior distribution also follows a generalized inverse Gaussian distribution.",4329
learning. algorithms for These sparse EM us lead properties Bayesian to,These properties lead us to EM algorithms for Bayesian sparse learning.,4330
"extensions selection present grouped regression. variable two addition, we and In for logistic","In addition, we present two extensions for grouped variable selection and logistic regression.",4331
"mathematical age-old among ""justification"" cliche an for marriage provides This a counselors.","This provides a mathematical ""justification"" for an age-old cliche among marriage counselors.",4332
Both methods use problem error humans frequently systems artificial trial to and and solving.,Both humans and artificial systems frequently use trial and error methods to problem solving.,4333
"perfect. control knowledge Unfortunately, is rarely this","Unfortunately, this control knowledge is rarely perfect.",4334
"is artificial difficult. in often in one's of own knowledge systems-as Moreover, humans-self-evaluation","Moreover, in artificial systems-as in humans-self-evaluation of one's own knowledge is often difficult.",4335
"our of have been Thus the using approach encouraging. results far,","Thus far, the results of using our approach have been encouraging.",4336
the on repository. an Our available is package R algorithm as CRAN,Our algorithm is available as an R package on the CRAN repository.,4337
"we method In represent head a for novel paper, pose estimation. this","In this paper, we represent a novel method for head pose estimation.",4338
transform use the for we efficiency estimation feature the To enhance of extraction. contourlet,To enhance the efficiency of the estimation we use contourlet transform for feature extraction.,4339
"Then, apply as neighborhood classifiers different we distance. minimum (knn) k-nearest such and","Then, we apply different classifiers such as k-nearest neighborhood (knn) and minimum distance.",4340
available to database evaluate use of proposed the We FERET performance the public method.,We use the public available FERET database to evaluate the performance of proposed method.,4341
of results Simulation robustness method. the the superior proposed indicate,Simulation results indicate the superior robustness of the proposed method.,4342
"In problems defines clearly occurs Paper, CFA the authors the the this image. in","In this Paper, the authors clearly defines the problems occurs in the CFA image.",4343
"a rate, which vaccine evolution high Influenza challenging. design has makes","Influenza has a high evolution rate, which makes vaccine design challenging.",4344
dominant approach detection an here We new of early strains. for consider,We here consider an approach for early detection of new dominant strains.,4345
dominant define strains. We emergence of metric detect then the a to,We then define a metric to detect the emergence of dominant strains.,4346
annual appear strain detection tool vaccine influenza useful would for be to selection. This,This strain detection tool would appear to be useful for annual influenza vaccine selection.,4347
"discuss biological modularity hierarchy In and review, systems. we in this","In this review, we discuss modularity and hierarchy in biological systems.",4348
will Experimental reviewed. this support for be description,Experimental support for this description will be reviewed.,4349
"networks. will be and networks, metabolic Examples networks, from protein-protein interaction pathogen presented gene structure,","Examples will be presented from pathogen structure, metabolic networks, gene networks, and protein-protein interaction networks.",4350
"networks. will Additional pathways, social networks, and presented physiology, developmental be ecological from examples food","Additional examples will be presented from ecological food networks, developmental pathways, physiology, and social networks.",4351
under can model solved some The assumptions. be analytically simplest,The simplest model can be solved analytically under some assumptions.,4352
the examples. in two model of demonstrate the utility We,We demonstrate the utility of the model in two examples.,4353
We sequence-based to effectiveness in here introduce vaccine humans. a method predict,We here introduce a sequence-based method to predict vaccine effectiveness in humans.,4354
with cell the B probability The rate. hypermutation $p$ decreases,The probability $p$ decreases with the B cell hypermutation rate.,4355
output by The final obtained is interpolation.,The final output is obtained by interpolation.,4356
the to classroom A illustrate presented experiment in also results is practice. of our use,A classroom experiment is also presented to illustrate the use of our results in practice.,4357
then of Post-processing moving is to used the boundaries refine objects.,Post-processing is then used to refine the boundaries of moving objects.,4358
but of large fitness populations consider landscapes. finite on evolution We arbitrary the,We consider the evolution of large but finite populations on arbitrary fitness landscapes.,4359
"describe Markov, Moran process. a process evolutionary We by the","We describe the evolutionary process by a Markov, Moran process.",4360
This enantioselectivity. question puts into key-lock models for,This puts into question key-lock models for enantioselectivity.,4361
from human tracking dataset. I show Eva on We pose results widely-used Human the also,We also show results on tracking human pose from the widely-used Human Eva I dataset.,4362
to evolves to system antibodies Influenza it. escape that immune from bind virus,Influenza virus evolves to escape from immune system antibodies that bind to it.,4363
A with constant. substitution positive G$ $\Delta\Delta value decreases the binding antibody,A substitution with positive $\Delta\Delta G$ value decreases the antibody binding constant.,4364
$\Delta\Delta On charged G$ the average acid amino an uncharged highest generates to substitution values.,On average an uncharged to charged amino acid substitution generates the highest $\Delta\Delta G$ values.,4365
"Also generate on average, zero. near acids amino between $\Delta\Delta values small substitutions G$ to","Also on average, substitutions between small amino acids generate $\Delta\Delta G$ values near to zero.",4366
view. this of simulations Carlo support to virus are evolution Monte presented,Monte Carlo simulations of virus evolution are presented to support this view.,4367
"model mutation individual-based under Evolutionary analysed population is stochastic, in a branching selection. and","Evolutionary branching is analysed in a stochastic, individual-based population model under mutation and selection.",4368
"studied in the mutational dynamics evolutionary Then, of the limit step is traditionally, sizes. vanishing","Then, traditionally, the evolutionary dynamics is studied in the limit of vanishing mutational step sizes.",4369
"non-negligible approach, the small mutational present steps but are In considered.","In the present approach, small but non-negligible mutational steps are considered.",4370
it ancient in nature. and suggests Evidence is ubiquitous both,Evidence suggests it is both ancient and ubiquitous in nature.,4371
"phenomenon, consensus and unifying the importance of the principles lacking. appear Despite","Despite the importance of the phenomenon, consensus and unifying principles appear lacking.",4372
biological of integration. path neural representation is the for needed A issue fundamental space,A fundamental issue is the neural representation of space needed for biological path integration.,4373
"is type of allothetic directional also available idiothetic, The namely or cue, considered.","The type of available directional cue, namely allothetic or idiothetic, is also considered.",4374
"animals, and navigation Efficient  planning for path transportation. robotics, is and  logistics critical","  Efficient path planning and navigation is critical for animals, robotics, logistics and transportation.",4375
"connectivity. spiking wave propagates network, activity synaptic a hippocampus-like of the through altering A","A wave of spiking activity propagates through a hippocampus-like network, altering the synaptic connectivity.",4376
autocatalytic results This of from study of presents a paper new detailed sets. structure the,This paper presents new results from a detailed study of the structure of autocatalytic sets.,4377
quantitative examination This applicable analysis of the method to BSCs. is of quantitative,This quantitative method of examination is applicable to the quantitative analysis of BSCs.,4378
the with single click infinite and about finite formula We the time give potential.,We give the formula about the single click time with finite and infinite potential.,4379
provide results new understanding potential. a infinite These of,These results provide a new understanding of infinite potential.,4380
the states the whole unadaptive parameters adaptive and We regimes. analytically for demonstrate,We analytically demonstrate the adaptive and unadaptive states for the whole parameters regimes.,4381
about with parameters regions the points is issues Interesting fixed imaginary demonstrated. the,Interesting issues about the parameters regions with the imaginary fixed points is demonstrated.,4382
graphical results framework to quantitative a phenomenon. the These suggest study and absorbing,These results suggest a graphical and quantitative framework to study the absorbing phenomenon.,4383
performance method the We our with of examples. illustrate,We illustrate the performance of our method with examples.,4384
applications decade. The face its reliability in using proved last biometric has,The applications using face biometric has proved its reliability in last decade.,4385
"DWT using (DBC- DBC Recognition propose FR) Face paper, In based model. this we","In this paper, we propose DBC based Face Recognition using DWT (DBC- FR) model.",4386
algorithm. render compared better existing proposed to recognition percentage rate algorithm The the,The proposed algorithm render better percentage recognition rate compared to the existing algorithm.,4387
studied of is (RTA) A reaction-diffusion model receptor-toxin-antibody interaction numerically.,A reaction-diffusion model of receptor-toxin-antibody (RTA) interaction is studied numerically.,4388
for the prediction data. of graph problem We evolving study,We study the problem of prediction for evolving graph data.,4389
efficient and to obtain formulation oracle solvers. convex allows inequalities The,The convex formulation allows to obtain oracle inequalities and efficient solvers.,4390
"transport. Linked transport Without limited around applications data exclusively Data, to is","Without Linked Data, transport data is limited to applications exclusively around transport.",4391
these possible will because datasets. to data transport be be This linked will,This will be possible because transport data will be linked to these datasets.,4392
provide of bounds. method lower to exact guarantees approximations outer Our the,Our method guarantees to provide outer approximations of the exact lower bounds.,4393
kind a not is limited of network specific Our method structure. to,Our method is not limited to a specific kind of network structure.,4394
approaches. is computational to superior other existing The of that complexity,The computational complexity is superior to that of other existing approaches.,4395
viewed variables uncertainty as The values. their quantify to random parameters are about,The parameters are viewed as random variables to quantify uncertainty about their values.,4396
"to i.e., of responses queries; Belief to conditional are interest. used nets probabilities compute","Belief nets are used to compute responses to queries; i.e., conditional probabilities of interest.",4397
"variable. A a is the function a random parameters, hence of query","A query is a function of the parameters, hence a random variable.",4398
and variance. We accurate both approximations more mean query for develop,We develop more accurate approximations for both query mean and variance.,4399
"and demonstrate improvements, empirical studies effectiveness. their provide several to analyze We","We analyze several improvements, and provide empirical studies to demonstrate their effectiveness.",4400
for that in explanation Our show results are networks. methods MRE approaches promising Bayesian empirical,Our empirical results show that MRE methods are promising approaches for explanation in Bayesian networks.,4401
Fluorescent Hybridization) color (Multiplex provides karyotyping. In-Situ M-FISH,M-FISH (Multiplex Fluorescent In-Situ Hybridization) provides color karyotyping.,4402
accuracy improved approach The when the provided to compared pixel-by-pixel approach.,The approach provided improved accuracy when compared to the pixel-by-pixel approach.,4403
based of on feature horizontal projection is text each profile done The the words. extraction,The feature extraction is done based on the horizontal projection profile of each text words.,4404
graphical First-order probabilistic models. of representational power first-order models combine logic with,First-order probabilistic models combine representational power of first-order logic with graphical models.,4405
first-order models. for probabilistic effort lifted algorithms inference an ongoing is design There to,There is an ongoing effort to design lifted inference algorithms for first-order probabilistic models.,4406
importance lifted of processing empirical tests inference. in confirm Our the constraint,Our empirical tests confirm the importance of constraint processing in lifted inference.,4407
This study is first inference. and constraint empirical processing the in of lifted theoretical,This is the first theoretical and empirical study of constraint processing in lifted inference.,4408
use We as motivating application. forestry the planning,We use forestry planning as the motivating application.,4409
problems a reinforcement LSST using policy as solution and define present problems gradients. learning We,We define LSST problems as reinforcement learning problems and present a solution using policy gradients.,4410
different Protein function changes between conformations. involves often,Protein function often involves changes between different conformations.,4411
predominantly occurs the prior conformational unbinding. to change,the conformational change occurs predominantly prior to unbinding.,4412
"is The contrast, in i.e. conformational during change catalytic the step, an induced change,","The conformational change during the catalytic step, in contrast, is an induced change, i.e.",4413
change the after occurs chemical reaction. the,the change occurs after the chemical reaction.,4414
environments plans optimal highly dynamic Generating challenging. is in,Generating optimal plans in highly dynamic environments is challenging.,4415
"This to globally distribution, variational can efficiently approach a queried. consistent be which leads","This variational approach leads to a globally consistent distribution, which can be efficiently queried.",4416
bases. measure paper on conditional develops inconsistency an probabilistic knowledge This,This paper develops an inconsistency measure on conditional probabilistic knowledge bases.,4417
present and We on examples formal and usefulness immediate its some several application illustrate results.,We illustrate its usefulness and immediate application on several examples and present some formal results.,4418
to are graph-theoretic algorithms the concept of Our bisimulation. closely related,Our algorithms are closely related to the graph-theoretic concept of bisimulation.,4419
"cannot true keep operations such classical sets. set-theoretic for soft laws all Unfortunately,","Unfortunately, such operations cannot keep all classical set-theoretic laws true for soft sets.",4420
pattern Attribute an data as step important preprocessing recognition is reduction for and mining. viewed,Attribute reduction is viewed as an important preprocessing step for pattern recognition and data mining.,4421
Most rough are researches attribute of focused on reduction using sets. by,Most of researches are focused on attribute reduction by using rough sets.,4422
sets with reduction C.C. paper the rough in attribute [E. discussed covering,discussed attribute reduction with covering rough sets in the paper [E. C.C.,4423
difficult. specification of aMarkov can decision The process (MDP) be,The specification of aMarkov decision process (MDP) can be difficult.,4424
an determining  relationships. evolutionary in events role important Reticulate play ,  Reticulate events play an important role in determining evolutionary relationships.,4425
work publicly are three are on available. extensions theoretical of and previous based All methods,All three methods are based on extensions of previous theoretical work and are publicly available.,4426
to apply real We our data. methods also,We also apply our methods to real data.,4427
"for is model a example, clustering. representative For model mixture hierarchical a","For example, a mixture model is a representative hierarchical model for clustering.",4428
the analysis statistical matrix is applicable. based on inverse Conventional not Fisher,Conventional statistical analysis based on the inverse Fisher matrix is not applicable.,4429
estimation to paper latent-variable its performance. The applies this and determines present analysis theoretical,The present paper applies this analysis to latent-variable estimation and determines its theoretical performance.,4430
Our the distribution. convergence of results the clarify behavior posterior of,Our results clarify behavior of the convergence of the posterior distribution.,4431
Metropolis-Hastings computationally a task. describes efficient accomplishing Recent for work method this,Recent work describes a computationally efficient Metropolis-Hastings method for accomplishing this task.,4432
"constraints. incorporate how topological show to we Additionally,","Additionally, we show how to incorporate topological constraints.",4433
value analyses propagation and unify policy approximate provide We error for iteration. those that,We provide error propagation analyses that unify those for approximate policy and value iteration.,4434
"in valued. be real of may However, individuals reputations situations, multiple","However, in real situations, reputations of individuals may be multiple valued.",4435
"present by analyzing we study, gaps the fill In both model. a trinary reputation","In the present study, we fill both gaps by analyzing a trinary reputation model.",4436
to be problem known The mentioned APX-hard. is,The mentioned problem is known to be APX-hard.,4437
engineered Forward and reverse compatible. modules fully are,Forward and reverse engineered modules are fully compatible.,4438
also discussed. to gray images extension The scale are n-dimensional theoretical of GLCM,The theoretical extension of GLCM to n-dimensional gray scale images are also discussed.,4439
trace features to when Haralick indicate results CBIR. outperform The applied that features,The results indicate that trace features outperform Haralick features when applied to CBIR.,4440
known computer widely and well vision. graph used in is Spectral theory,Spectral graph theory is well known and widely used in computer vision.,4441
"most misidentified. are the features are If extracted of the cases used, all","If all the extracted features are used, most of the cases are misidentified.",4442
"with {\pi} Fuzzy-Rough In is feature paper, this proposed. function selection membership","In this paper, Fuzzy-Rough feature selection with {\pi} membership function is proposed.",4443
experimental improves method the the analysis mammograms shows proposed that accuracy. The classification,The experimental analysis shows that the proposed method improves the mammograms classification accuracy.,4444
the remove The show method mixed to the is noise. simulation results effective that,The simulation results show that the method is effective to remove the mixed noise.,4445
"abacavir remains induces pathologic by However, unclear. which T response cell the mechanism this","However, the mechanism by which abacavir induces this pathologic T cell response remains unclear.",4446
"for results provide To velocity substantiate real datasets. and we synthetic method, our field","To substantiate our method, we provide results for synthetic and real velocity field datasets.",4447
"we at response antigen. saturated levels this allow particular, become high of In to","In particular, we allow this response to become saturated at high levels of antigen.",4448
"and Sciences, of India, Delhi Hyderabad. Hills, Technology Vignan of Institute Vignan New at","of India, New Delhi at Vignan Institute of Technology and Sciences, Vignan Hills, Hyderabad.",4449
level equivalences equality capture of Rough at equalities. a than higher rough sets,Rough equivalences capture equality of sets at a higher level than rough equalities.,4450
leveled and studied. were the fact properties approximate of equalities introduced concepts In were,In fact the concepts of leveled approximate equalities were introduced and properties were studied.,4451
on alternating direction the The (ADM). is for solving based TILT existing algorithm method,The existing algorithm for solving TILT is based on the alternating direction method (ADM).,4452
times five could be It least faster method. previous the at than,It could be at least five times faster than the previous method.,4453
"of understanding we extend Secondly, the histograms. local theoretical the","Secondly, we extend the theoretical understanding of the local histograms.",4454
"preferred asymmetric is choice. our GPV is Also, an PW and measure,","Also, GPV is an asymmetric measure, and PW is our preferred choice.",4455
problem addresses This approximate in the graphical general models. MAP-MRF inference of paper,This paper addresses the problem of approximate MAP-MRF inference in general graphical models.,4456
also how show deal to efficient factors nested an with in We way.,We also show how to deal with nested factors in an efficient way.,4457
The independent based m online estimator simulation. from is Monte on an Carlo samples,The estimator is based on m samples from an independent online Monte Carlo simulation.,4458
applications terms computed. interest be correction cannot of the In many,In many applications of interest the correction terms cannot be computed.,4459
versions algorithms. of approximate We the consider,We consider approximate versions of the algorithms.,4460
learning based classification. note are representative presents for (DL) which This dictionary on some methods,This note presents some representative methods which are based on dictionary learning (DL) for classification.,4461
"roughly divide two into some them methods, representative can listing By i.e. we categories,","By listing some representative methods, we can roughly divide them into two categories, i.e.",4462
"this expect as future we extensions From can them taxonomy, of some researches.","From this taxonomy, we can expect some extensions of them as future researches.",4463
based proposes that segmentation images. can automatically method Algorithm Genetic segment gray-scale a This paper,This paper proposes a Genetic Algorithm based segmentation method that can automatically segment gray-scale images.,4464
generates image. results segmented shown quality algorithm that Experimental the good,Experimental results shown that the algorithm generates good quality segmented image.,4465
neighbor K (KNN) as classifier. nearest a used,K nearest neighbor (KNN) used as a classifier.,4466
"development models. case-cohort data, has For less done been much with AFT","For case-cohort data, much less development has been done with AFT models.",4467
is one of fundamental recognition. extraction character of problems the Feature,Feature extraction is one of the fundamental problems of character recognition.,4468
evidence that is due We integration. stimulus to this provide improvement,We provide evidence that this improvement is due to stimulus integration.,4469
their of function as latency Rats increase trial response modestly a difficulty.,Rats increase their response latency modestly as a function of trial difficulty.,4470
also ABC pharmacological aspects targeting remain transporter Several of unclear.,Several aspects of pharmacological ABC transporter targeting also remain unclear.,4471
in similarity been to data. have the approaches Several sequences proposed measure between trajectory,Several approaches have been proposed to measure the similarity between sequences in trajectory data.,4472
"wireless not of appropriate networks. in characteristics mobility are to However, spatiotemporal patterns they","However, they are not appropriate to characteristics of spatiotemporal mobility patterns in wireless networks.",4473
framework our of follows. as for is phases two composed The measure constructing,The framework for constructing our measure is composed of two phases as follows.,4474
The proved in mathematics. truth similarity and of the measures partial total are,The truth of the partial and total similarity measures are proved in mathematics.,4475
Models incorporation often these of interactions. lack to assess explicit interspecific extinction risk,Models to assess extinction risk often lack explicit incorporation of these interspecific interactions.,4476
(e.g. parameters effects We the investigate foraging of the,We investigate the effects of the foraging parameters (e.g.,4477
and time) to rate handling mean the extinction. time attack on,attack rate and handling time) on the mean time to extinction.,4478
query. link the of provided The the context is by main,The main link is provided by the context of the query.,4479
this information method With effectively. works minimal,With minimal information this method works effectively.,4480
the hospital. departments Scheduling problem can for multiple solved in be,Scheduling problem can be solved for multiple departments in the hospital.,4481
module wiki-like to minireview. extended Each contains and be documentation metadata for can a purposes,Each module contains metadata for documentation purposes and can be extended to a wiki-like minireview.,4482
manage can module. of each The database multiple versions,The database can manage multiple versions of each module.,4483
user e.g. according can defined to recombined (semi-) be automatically Modules scenarios,Modules can be (semi-) automatically recombined according to user defined scenarios e.g.,4484
"patterns expression or given disease. gene under cell certain physiological types, states of conditions, in","gene expression patterns in given cell types, under certain physiological conditions, or states of disease.",4485
one recognition. is It most face the in successful techniques of,It is one of the most successful techniques in face recognition.,4486
high database. big size drawback it has But of computational especially for,But it has drawback of high computational especially for big size database.,4487
the eigenvectors authors participated the decreases computational time. The which consequently minimize,The authors minimize the participated eigenvectors which consequently decreases the computational time.,4488
used DET illustrate Curves are the to experimental results.,DET Curves are used to illustrate the experimental results.,4489
"of muscles not. ions whether divalent calcium contract or determines concentration example, the For","For example, the concentration of divalent calcium ions determines whether muscles contract or not.",4490
The Concentrations Important Control Life Way Control Channels Important To Ion Voltages To Computers. Computers,Ion Channels Control Concentrations Important To Life The Way Computers Control Voltages Important To Computers.,4491
and atoms Chemistry is as react. configurations their electrons molecules changing about,Chemistry is about electrons changing their configurations as atoms and molecules react.,4492
ideal in studies infinitely solutions. Chemistry dilute if as reactions occurred they,Chemistry studies reactions as if they occurred in ideal infinitely dilute solutions.,4493
But solutions. nonideal in most occur reactions,But most reactions occur in nonideal solutions.,4494
this has developed interacting of deal sort. with to recently been systems Mathematics,Mathematics has recently been developed to deal with interacting systems of this sort.,4495
The spawned has liquid fluids crystals. the of complex of theory variational theory,The variational theory of complex fluids has spawned the theory of liquid crystals.,4496
"solutions be ionic In as should complex fluids. my view, viewed","In my view, ionic solutions should be viewed as complex fluids.",4497
spheres We a dielectric. present variational treatment a of in hard frictional,We present a variational treatment of hard spheres in a frictional dielectric.,4498
"this we coefficient feature In method, multiple using weight of correlations. the","In this method, we weight the feature using coefficient of multiple correlations.",4499
software A fight is to aging primitive against software method rejuvenation.,A primitive method to fight against software aging is software rejuvenation.,4500
"stopping software, It internal it. occasionally its and the cleaning state running restarting involves","It involves occasionally stopping the running software, cleaning its internal state and restarting it.",4501
was Environment) used of simulation Application implementation for Development (Java JADE and cooperation. agents,JADE (Java Application Development Environment) was used for implementation and simulation of agents cooperation.,4502
as set valuated in of are formulas. properties interpreted Data,Data in properties are interpreted as set of valuated formulas.,4503
and handwriting: offline. of on-line Two types,Two types of handwriting: on-line and offline.,4504
handwriting off-line most recognition purpose in and is protection of important authentication. The systems,The most important purpose of off-line handwriting recognition is in protection systems and authentication.,4505
scripts. scripts much to comparison in Arabic Handwriting more Latin are complicated,Arabic Handwriting scripts are much more complicated in comparison to Latin scripts.,4506
methodology introduces This to Arabic a characters. paper simple authenticate and novel handwriting,This paper introduces a simple and novel methodology to authenticate Arabic handwriting characters.,4507
"database. character our we our aim, built Reaching own","Reaching our aim, we built our own character database.",4508
show tractable. this that parameter We detecting autocatalytic sets fixed is case in,We show that in this case detecting autocatalytic sets is fixed parameter tractable.,4509
its forms. free degrades The bound and protein in,The protein degrades in its free and bound forms.,4510
balance This breaks reaction bimolecular involves step. and network detailed a single,This network breaks detailed balance and involves a single bimolecular reaction step.,4511
for proposed (PA) detail is various images tested on varying noises. algorithm The,The proposed algorithm (PA) is tested on varying detail images for various noises.,4512
histone quantum proposed. is A mechanical  on model modification ,  A quantum mechanical model on histone modification is proposed.,4513
are deduced. histone dependence The and coherence-length dependence modification the of temperature,The temperature dependence and the coherence-length dependence of histone modification are deduced.,4514
predicted. level The histone is the temperature of non-Arhenius modification dependence,The non-Arhenius temperature dependence of the histone modification level is predicted.,4515
"from however, genetic remains these a data, task. relationships Inferring challenging","Inferring these relationships from genetic data, however, remains a challenging task.",4516
sequences structures torsional non-canonical (IR) DNA repeat can Inverted stress. relieve to form cruciform in,Inverted repeat (IR) sequences in DNA can form non-canonical cruciform structures to relieve torsional stress.,4517
architecture experimental A.I. of theoretical of approach aspects This whole describes the the the,This approach describes the theoretical aspects of the architecture of the whole experimental A.I.,4518
uses to research signals. the A of measure new compression methods line similarity between,A new line of research uses compression methods to measure the similarity between signals.,4519
Our generalization anytime performance is represented algorithm composing by conditional of a approach profiles.,Our approach is a generalization of composing anytime algorithm represented by conditional performance profiles.,4520
electrophoresis proteomics. been gel has Two-dimensional the instrumental of in development,Two-dimensional gel electrophoresis has been instrumental in the development of proteomics.,4521
with without regard models Graphical doing learned inference cost are of usually to the them.,Graphical models are usually learned without regard to the cost of doing inference with them.,4522
and of (TPLP). appear Theory To Practice in Logic Programming,To appear in Theory and Practice of Logic Programming (TPLP).,4523
provide facts of from conditional a rules deriving d-separation The independence model for framework structure.,The rules of d-separation provide a framework for deriving conditional independence facts from model structure.,4524
"directed to only this models. simple However, applies graphical theory","However, this theory only applies to simple directed graphical models.",4525
"deriving theory in d-separation, independence models. relational conditional for relational We introduce a","We introduce relational d-separation, a theory for deriving conditional independence in relational models.",4526
for image The estimation. inverse primary was PSF used,The inverse PSF was used for primary image estimation.,4527
by variations of condition process on The basing was convergence. defined dynamic balance regularization iteration,The variations balance was defined by dynamic regularization basing on condition of iteration process convergence.,4528
variations. regularization previous dynamic character estimate image because on depends current has The and,The regularization has dynamic character because depends on current and previous image estimate variations.,4529
on optimal is target fluctuations invariance of functional basing to value. It argument,It is basing on target functional invariance to fluctuations of optimal argument value.,4530
in perform nearly as as hindsight. the best well to aims Online hypothesis learning,Online learning aims to perform nearly as well as the best hypothesis in hindsight.,4531
"hypothesis though, hypothesis challenging. the offline some For even best is classes, finding","For some hypothesis classes, though, even finding the best hypothesis offline is challenging.",4532
"offline guaranteed. local techniques local employed are search cases, only optimality often and In such","In such offline cases, local search techniques are often employed and only local optimality guaranteed.",4533
minimize We local with then algorithm graphs. a to present general arbitrary locality regret,We then present a general algorithm to minimize local regret with arbitrary locality graphs.,4534
drastically speed to We how learning. structure graph can show also be exploited the,We also show how the graph structure can be exploited to drastically speed learning.,4535
is variable R presented. linear estimating for latent and An models specifying package,An R package for specifying and estimating linear latent variable models is presented.,4536
methods This several compared networks. to other is approach known on,This approach is compared to several other methods on known networks.,4537
ALP proposes This approximation. a paper new,This paper proposes a new ALP approximation.,4538
LP for structure the This solving allows efficiently. new,This structure allows for solving the new LP efficiently.,4539
aspects of and study the both approach. We practical proposed theoretical,We study both practical and theoretical aspects of the proposed approach.,4540
subgraphs patterns of networks. Motifs complex are of,Motifs are patterns of subgraphs of complex networks.,4541
"prevalence chain to Specifically, diverging of the correlation. increase motifs and tends","Specifically, the prevalence of diverging and chain motifs tends to increase correlation.",4542
correctness its for proved. provided multiplicative is is and inference An models algorithm,An inference algorithm for multiplicative models is provided and its correctness is proved.,4543
and examples the models with beneficial. algorithm ends new where The using paper computationally is,The paper ends with examples where using the new models and algorithm is computationally beneficial.,4544
importance models. paper introduces The probabilistic sampling for graphical AND/OR,The paper introduces AND/OR importance sampling for probabilistic graphical models.,4545
The entities trainer is two and evaluator. different of system composed,The system is composed of two different entities trainer and evaluator.,4546
in of terms of interaction variety accuracy and We our measure classification scenarios.,We measure our accuracy in terms of variety of interaction and classification scenarios.,4547
"For BN. higher with BNs quickly datasets, BDeu finds 'true' the than most MaxWalkSat score","For most datasets, MaxWalkSat quickly finds BNs with higher BDeu score than the 'true' BN.",4548
effect information The is assessed. adding of prior,The effect of adding prior information is assessed.,4549
this the RIS result derive we theoretic Based algorithm. on,Based on this theoretic result we derive the RIS algorithm.,4550
polyacrylamide to electrophoretic proteins gels. is used on separation staining after detect Silver,Silver staining is used to detect proteins after electrophoretic separation on polyacrylamide gels.,4551
"Both features optimized are are proposed. silver in this discussed staining protocols and chapter,","Both features are discussed in this chapter, and optimized silver staining protocols are proposed.",4552
new portfolios. for and present techniques designing We evaluate algorithm,We present and evaluate new techniques for designing algorithm portfolios.,4553
"machine scheduling aspect. In a learning problem the aspect our view, a and both has","In our view, the problem has both a scheduling aspect and a machine learning aspect.",4554
in addressed largely has the Prior work two aspects of one isolation.,Prior work has largely addressed one of the two aspects in isolation.,4555
enumerated. on medium-size state fully can problems focus space be whose We,We focus on medium-size problems whose state space can be fully enumerated.,4556
"important applications, has planning and navigation under problem such numerous as This uncertainty.","This problem has numerous important applications, such as navigation and planning under uncertainty.",4557
by decomposing simpli the often be smaller task tasks arranged can Planning into ed hierarchically.,Planning can often be simpli ed by decomposing the task into smaller tasks arranged hierarchically.,4558
"line Toussaint research, al. another of In et","In another line of research, Toussaint et al.",4559
techniques approach based non-convex that this on than better scales demonstrate optimization. Experimental results previous,Experimental results demonstrate that this approach scales better than previous techniques based on non-convex optimization.,4560
models is in One cell of in of application proliferation. studies of field biology such,One field of application of such models in biology is in studies of cell proliferation.,4561
are effects-based with problem diagrams experiments an military planning discussed. issues random and and Complexity,Complexity issues and experiments with random diagrams and an effects-based military planning problem are discussed.,4562
Eberhardt builds insights in gained analysis case on from The et worst al. the algorithm,The algorithm builds on insights gained from the worst case analysis in Eberhardt et al.,4563
that conjecture correct. suggests our A is simulation,A simulation suggests that our conjecture is correct.,4564
challenge. characterization be will Systematic a these of tremendous transcripts,Systematic characterization of these transcripts will be a tremendous challenge.,4565
"of in study the model. sharing we Wright-Fisher the distribution Here, IBD","Here, we study the distribution of IBD sharing in the Wright-Fisher model.",4566
"the normally. sharing large approximately We find that for distributed cohort-averaged is cohorts,","We find that for large cohorts, the cohort-averaged sharing is distributed approximately normally.",4567
is that comparison visualization necessary. images of indicates some a three-dimensional The,The comparison of some images indicates that a three-dimensional visualization is necessary.,4568
be of terminals retinal well-segregated to origin. We find eye the by,We find the retinal terminals to be well-segregated by eye of origin.,4569
subdomains It functional sublaminae. distinct these represent whether be determined remains to,It remains to be determined whether these subdomains represent distinct functional sublaminae.,4570
method using for signal fast Bayesian introduced a sparse pursuit A is matching recovery. approach,A fast matching pursuit method using a Bayesian approach is introduced for sparse signal recovery.,4571
proposed demonstrate estimator. power results the Simulation robustness our of and,Simulation results demonstrate the power and robustness of our proposed estimator.,4572
initialization proposed also is hierarchical the random for simple algorithm. and A effective EM,A simple and effective hierarchical random initialization is also proposed for the EM algorithm.,4573
novel real in some to model-based applications is data. illustrated clustering technique The,The novel model-based clustering technique is illustrated in some applications to real data.,4574
equipped parameter scaling The a is conjugate partially with prior.,The scaling parameter is equipped with a partially conjugate prior.,4575
that the illustrate examples give Some satisfactory method results. can,Some examples illustrate that the method can give satisfactory results.,4576
another a with comparison these well. In as existing made examples is method,In these examples a comparison is made with another existing method as well.,4577
in by the the model. feasibility obtained structure Computational local GAI is exploiting,Computational feasibility is obtained by exploiting the local GAI structure in the model.,4578
in The disease of Parkinson's unknown. tremor origin exact remains,The exact origin of tremor in Parkinson's disease remains unknown.,4579
in our everyday lives. play Preferences important role an,Preferences play an important role in our everyday lives.,4580
"exist CPnets. for algorithms for but CP-nets, incomplete multi-valued Tractable these algorithms binary are","Tractable algorithms exist for binary CP-nets, but these algorithms are incomplete for multi-valued CPnets.",4581
effective POMDPs decentralized (MBDP) solving has in Memory-Bounded with Dynamic large horizons. Programming extremely proved,Memory-Bounded Dynamic Programming (MBDP) has proved extremely effective in solving decentralized POMDPs with large horizons.,4582
"larger To benchmark problem. the the improvements, new, effectiveness of evaluate we introduce a","To evaluate the effectiveness of the improvements, we introduce a new, larger benchmark problem.",4583
is biological of to large DNA wide processes. deformations subject in a range,DNA is subject to large deformations in a wide range of biological processes.,4584
persons systems disabilities for (e.g. with Assistive cognitive,Assistive systems for persons with cognitive disabilities (e.g.,4585
"for POMDPs and very labour definition a usually their procedure construction. require Unfortunately, intensive, manual","Unfortunately, POMDPs usually require a very labour intensive, manual procedure for their definition and construction.",4586
Assistance a call the resulting We POMDP SNAP Process). (SyNdetic,We call the resulting POMDP a SNAP (SyNdetic Assistance Process).,4587
"a had manually expert. To performed to be trained date, by POMDP this translation","To date, this translation had to be performed manually by a trained POMDP expert.",4588
method tasks non-experts. by the eliciting demonstrate assistance three from We,We demonstrate the method by eliciting three assistance tasks from non-experts.,4589
probabilistic search the various superiority approach We empirically networks. best-first demonstrate on of the,We demonstrate empirically the superiority of the best-first search approach on various probabilistic networks.,4590
number new of to formulation implications. The theoretical a practical leads and,The new formulation leads to a number of theoretical and practical implications.,4591
present infinite-horizon approach We solving for POMDPs. a optimization decentralized memory-bounded,We present a memory-bounded optimization approach for solving infinite-horizon decentralized POMDPs.,4592
for state controllers. stochastic Policies represented agent finite each are by,Policies for each agent are represented by stochastic finite state controllers.,4593
"are detected First, tracked. and people individually","First, people are individually detected and tracked.",4594
"Mean-Shift and clustered a Second, trajectories window analyzed are their temporal using over the algorithm.","Second, their trajectories are analyzed over a temporal window and clustered using the Mean-Shift algorithm.",4595
"formal Furthermore, language. propose a description we event","Furthermore, we propose a formal event description language.",4596
and The contributions. makes paper investigates this several potential,The paper investigates this potential and makes several contributions.,4597
despite this algorithm the that optimal We show converges to difference policy. the,We show that despite this difference the algorithm converges to the optimal policy.,4598
prescriptive descriptive utility and a generalisation It both provides expected enhancing of possibilities.,It provides a generalisation of expected utility enhancing both descriptive and prescriptive possibilities.,4599
"problem. search heuristic propose different complexity issues, algorithms we After two discussing to solve the","After discussing complexity issues, we propose two different heuristic search algorithms to solve the problem.",4600
"algorithms. reported, the experiments are the numerical practical of proposed efficiency showing Finally,","Finally, numerical experiments are reported, showing the practical efficiency of the proposed algorithms.",4601
such is inference. in difficulty models The main learning computational,The main computational difficulty in learning such models is inference.,4602
"equivalent that is we computational loopy show procedure to this propagation. sychronous Moreover, belief","Moreover, we show that this computational procedure is equivalent to sychronous loopy belief propagation.",4603
and in inference speedup learning enables a time. dramatic This,This enables a dramatic speedup in inference and learning time.,4604
error with of even is bounds known NP-hard. probability the Computing evidence,Computing the probability of evidence even with known error bounds is NP-hard.,4605
address an this settling we problem. hard this on In by paper easier problem,In this paper we address this hard problem by settling on an easier problem.,4606
that scheme inequality. proposed Our randomized uses a is approximation the Markov importance sampling,Our proposed approximation is a randomized importance sampling scheme that uses the Markov inequality.,4607
"a application bounds. to poor inequality the Markov may straight-forward lower lead of However,","However, a straight-forward application of the Markov inequality may lead to poor lower bounds.",4608
its We performance to several in practice. heuristic measures therefore improve propose,We therefore propose several heuristic measures to improve its performance in practice.,4609
of logic first-order long AI. and probability a been goal has Combining,Combining first-order logic and probability has long been a goal of AI.,4610
fall cases Many in interesting this category.,Many interesting cases fall in this category.,4611
consistent the then We measures the non-unique of set the in of structure case. examine,We then examine the structure of the set of consistent measures in the non-unique case.,4612
"important non-unique phenomena, are Many MLNs including transitions, with with measures. systems represented phase by","Many important phenomena, including systems with phase transitions, are represented by MLNs with non-unique measures.",4613
"Compiling been and recently especially processing. graphical intense modeling models probabilistic for under has investigation,","Compiling graphical models has recently been under intense investigation, especially for probabilistic modeling and processing.",4614
"work of is networks, constraint our to generalization previous on weighted a models. This","This is a generalization of our previous work on constraint networks, to weighted models.",4615
"the encouraging, of AOMDD is quite potential showing structure. the data The experimental preliminary evaluation","The preliminary experimental evaluation is quite encouraging, showing the potential of the AOMDD data structure.",4616
This paper of addresses such learning for rule the tasks. related sets problem multiple,This paper addresses the problem of learning such rule sets for multiple related tasks.,4617
"exhibit learn as to that relationships parity. They ""correlation-immunity"", are such unable","They are unable to learn relationships that exhibit ""correlation-immunity"", such as parity.",4618
"at Inference fastest significant computational at performed granularity, must be then this potentially cost.","Inference must then be performed at this fastest granularity, potentially at significant computational cost.",4619
inference (EP) al. Nodelman expectation-propagation et The of algorithm,The expectation-propagation (EP) inference algorithm of Nodelman et al.,4620
present this significant computational savings. can result experiments that in approach We demonstrating,We present experiments demonstrating that this approach can result in significant computational savings.,4621
belief exact BNs). Gaussian update BNs (CLG Conditional an architecture Linear for in We present,We present an architecture for exact belief update in Conditional Linear Gaussian BNs (CLG BNs).,4622
The examples. resulting benefits by are illustrated,The resulting benefits are illustrated by examples.,4623
Results of indicate the of a a performance architecture. proposed potential evaluation significant preliminary empirical,Results of a preliminary empirical performance evaluation indicate a significant potential of the proposed architecture.,4624
The complexity and of in creates many the control. management NAS difficulties,The complexity of the NAS creates many difficulties in management and control.,4625
is the problems pressing One most delay. of flight,One of the most pressing problems is flight delay.,4626
"operations. passengers, complaints for airlines, and Delay creates to cost high from difficulties airport","Delay creates high cost to airlines, complaints from passengers, and difficulties for airport operations.",4627
"demand delay prominent. increases, problem more the becomes As on the and system more","As demand on the system increases, the delay problem becomes more and more prominent.",4628
"spectrum was of tested. for inhibition However, isolates the the different","However, the spectrum of inhibition was different for the isolates tested.",4629
probabilistic is system gambles. Jensen's This axiom new by axiomatization of inspired,This new axiom system is inspired by Jensen's axiomatization of probabilistic gambles.,4630
Many are heuristic. the literature in approaches,Many approaches in the literature are heuristic.,4631
An metrics integral metric distributions. the probability Kantorovich such of component is between,An integral component of such metrics is the Kantorovich metric between probability distributions.,4632
We provide of empirical these an evaluation trade-offs.,We provide an empirical evaluation of these trade-offs.,4633
making means circumstances. a in involved offer more of calculations the certain efficient computation-intensive They,They offer a means of making the computation-intensive calculations involved more efficient in certain circumstances.,4634
as attention (in)dependence. are stochastic They special deserve asymmetric unlike they,They deserve special attention as they are asymmetric unlike stochastic (in)dependence.,4635
"we a graphical on their focus develop and paper suitable, representation In this i.e.","In this paper we focus on their graphical representation and develop a suitable, i.e.",4636
take disambiguating These energy. typically time expend actions and,These disambiguating actions typically take time and expend energy.,4637
"we for aid robotics. in active stereo of work, this sensing consider In vision","In this work, we consider active sensing in aid of stereo vision for robotics.",4638
"for Based selection we develop sweep. a the this estimator on coefficient an result, driving","Based on this result, we develop an estimator for the selection coefficient driving a sweep.",4639
so-called to of thanks usual They all extend 'potentials'. the use algorithms elimination variable,They all extend usual variable elimination algorithms thanks to the use of so-called 'potentials'.,4640
ratchet. restoring force delayed speeds This Muller's up,This delayed restoring force speeds up Muller's ratchet.,4641
much is probabilistic graphical challenge still very in models for domains. Inference large practical a,Inference for probabilistic graphical models is still very much a practical challenge in large domains.,4642
set reasoning metric with a spaces model measures. about up theoretic belief for We,We set up a model for reasoning about metric spaces with belief theoretic measures.,4643
these and in from metric. both The probability uncertainty stems spaces,The uncertainty in these spaces stems from both probability and metric.,4644
is formal constructed for reasoning A the system about logical distance. expected,A formal logical system is constructed for the reasoning about expected distance.,4645
logic. completeness are and Soundness this for shown,Soundness and completeness are shown for this logic.,4646
give of separability. equivalence the an sufficiency alternate We proof and on,We give an alternate proof on the equivalence of sufficiency and separability.,4647
whether a given testing computational present a is CPT for method separable. We,We present a computational method for testing whether a given CPT is separable.,4648
proposed samples in converge due fewer to requires The to scheme variance. sampling reduction,The proposed scheme requires fewer samples to converge due to reduction in sampling variance.,4649
presented complicated more efficient is and than the algorithm less techniques Our earlier.,Our algorithm is less complicated and more efficient than the techniques presented earlier.,4650
directed (CTBNs) Bayesian Continuous-time representations processes multi-component graphs. of as are graphical continuous-time networks Markov,Continuous-time Bayesian networks (CTBNs) are graphical representations of multi-component continuous-time Markov processes as directed graphs.,4651
in The direct among influences components. the edges network represent,The edges in the network represent direct influences among components.,4652
well. under examine which We also conditions UCT-Aux works,We also examine conditions under which UCT-Aux works well.,4653
linear programming set functions weights. to and techniques first-order of basis uses suitable determine a,a set of first-order basis functions and uses linear programming techniques to determine suitable weights.,4654
"represented. covering of Through concisely approximation types are the important equivalently operators matrices, three characteristic","Through the characteristic matrices, three important types of covering approximation operators are concisely equivalently represented.",4655
"decomposition. in Second, operators used boolean representations matrix matrix are approximation of covering","Second, matrix representations of covering approximation operators are used in boolean matrix decomposition.",4656
we develop algorithm decomposition. matrix And an boolean this for,And we develop an algorithm for this boolean matrix decomposition.,4657
neurophysiological a using by data. oscillator consistent model neural We do so with,We do so by using a neural oscillator model consistent with neurophysiological data.,4658
of was model the well behavioral reproduce used predictions stimulus-response shown to theory. The,The model used was shown to reproduce well the predictions of behavioral stimulus-response theory.,4659
on of This strategy the paper multi-modal functions. focuses restart on CMA-ES,This paper focuses on the restart strategy of CMA-ES on multi-modal functions.,4660
"the provided, of convergence proofs including Limited is two algorithms. analysis the","Limited analysis is provided, including the convergence proofs of the two algorithms.",4661
experiments algorithms. Computer the were validate to conducted,Computer experiments were conducted to validate the algorithms.,4662
are analysis. theoretical with consistent results the The,The results are consistent with the theoretical analysis.,4663
An to in choose an open both structure. algorithms appropriate how problem been has approximation,An open problem in both algorithms has been how to choose an appropriate approximation structure.,4664
and allow free conversion energies. between GBP Reduction these EP structures on operators,Reduction operators on these structures allow conversion between EP and GBP free energies.,4665
on of continuous network. are There nodes the no placement discrete in constraints the and,There are no constraints on the placement of continuous and discrete nodes in the network.,4666
"inference As exact shown intractable. previously, in is CTBNs","As shown previously, exact inference in CTBNs is intractable.",4667
solution processes new decision introduce Markov technique approximate for We a (FOMDPs). first-order,We introduce a new approximate solution technique for first-order Markov decision processes (FOMDPs).,4668
everyday characterize and discover of event-streams. activities to We different present from classes a framework,We present a framework to discover and characterize different classes of everyday activities from event-streams.,4669
representing as bags n-grams. of We event activities begin by,We begin by representing activities as bags of event n-grams.,4670
it system to entities the These evolves. track as state of of need the keep,These entities need to keep track of the state of the system as it evolves.,4671
"as the beliefs being of state the system are continues Furthermore, updated. to change","Furthermore, the state of the system continues to change as beliefs are being updated.",4672
developing approaches have Previous to probabilistic used static reasoning systems models. asynchronous distributed,Previous approaches to developing distributed asynchronous probabilistic reasoning systems have used static models.,4673
results our algorithm. approach to compares favorably show frontier Experimental that the factored,Experimental results show that our approach compares favorably to the factored frontier algorithm.,4674
efiectively is the area in Solving uncertainty. of planning problems such major under a challenge,Solving such problems efiectively is a major challenge in the area of planning under uncertainty.,4675
and solution classical theory. based decentralized a is synthesis on control search heuristic of Our,Our solution is based on a synthesis of classical heuristic search and decentralized control theory.,4676
MAA* significant that advantages. Experimental show results has,Experimental results show that MAA* has significant advantages.,4677
our performance algorithms We the approximate of randomly generated inference demonstrate HMNs. on,We demonstrate the performance of our approximate inference algorithms on randomly generated HMNs.,4678
maximization to show maximization to how CTBNs. (SEM) and expectation (EM) We expectation structural apply,We show how to apply expectation maximization (EM) and structural expectation maximization (SEM) to CTBNs.,4679
"of graphs, number of conditional In are only required. certain types independencies linear","In certain types of graphs, only linear number of conditional independencies are required.",4680
structural testing applications errors. linear equation The with in has correlated models result,The result has applications in testing linear structural equation models with correlated errors.,4681
solving presented were methods many for so problem. Until now this,Until now so many methods were presented for solving this problem.,4682
"proposed The technique. paper based based ""Color in image on (CBIR) this approach is retrieval""","The proposed approach in this paper is based on ""Color based image retrieval"" (CBIR) technique.",4683
"discrete for far, PSR with primarily have observations. models been domains developed Thus","Thus far, PSR models have primarily been developed for domains with discrete observations.",4684
interest In human action recognition. years in considerable been has there recent,In recent years there has been considerable interest in human action recognition.,4685
the to enhance in order have approaches been analysis. automatic Several developed video,Several approaches have been developed in order to enhance the automatic video analysis.,4686
accuracy shown different results different Our found be databases. that rates can for,Our results shown that different accuracy rates can be found for different databases.,4687
We stems simplicity LARP. show calculations of the nature the that from of the,We show that the simplicity of the calculations stems from the nature of LARP.,4688
"we for forms CPTs not explicit the the diagnostic Given networks. have evidence, do in","Given diagnostic evidence, we do not have explicit forms for the CPTs in the networks.",4689
derive the form the the exact of We optimal for function. importance first CPTs,We first derive the exact form for the CPTs of the optimal importance function.,4690
"use their approximations. we is only Since the calculation usually hard,","Since the calculation is hard, we usually only use their approximations.",4691
review popular their point and out We limitations. several strategies,We review several popular strategies and point out their limitations.,4692
known previously upon conditions. Our results improve,Our results improve upon previously known conditions.,4693
"our with seem variables binary For interactions, conditions to sharp. be (anti-)ferromagnetic","For binary variables with (anti-)ferromagnetic interactions, our conditions seem to be sharp.",4694
of tree. throughout the imposes evolution of the process evolutionary fixed rates Homogeneity,Homogeneity of the evolutionary process imposes fixed rates of evolution throughout the tree.,4695
"incidence making Euclidean used. the degree grey methods, In decision in four-dimensional space these is","In these decision making methods, the grey incidence degree in four-dimensional Euclidean space is used.",4696
final method. obtained The arrangement result Borda weighted is by,The final arrangement result is obtained by weighted Borda method.,4697
proposed the approach. An the of illustrates example applicability,An example illustrates the applicability of the proposed approach.,4698
example proposed approach. applicability of the the shows An,An example shows the applicability of the proposed approach.,4699
are DMA both the than We approach. also that easier approaches computationally show,We also show that both approaches are computationally easier than the DMA approach.,4700
"Consequently, for category. been each models have different proposed","Consequently, different models have been proposed for each category.",4701
algorithm efficient an propose utility. estimation possibility for based of We expected qualitative,We propose an efficient algorithm for estimation of possibility based qualitative expected utility.,4702
their to images is challenging owing Segmentation a complexity. task medical of,Segmentation of medical images is a challenging task owing to their complexity.,4703
is this an subset Finding setting in optimal tests in intractable of general.,Finding an optimal subset of tests in this setting is intractable in general.,4704
to iteration value our (point-based) algorithm. also We improvements recent heuristic discuss search,We also discuss recent improvements to our (point-based) heuristic search value iteration algorithm.,4705
domains. Traditional lack the to expressive languages probabilistic handle relational power,Traditional probabilistic languages lack the expressive power to handle relational domains.,4706
"but a lacks capability. plausible first-order expressive, reasoning Classical sufficiently logic coherent is","Classical first-order logic is sufficiently expressive, but lacks a coherent plausible reasoning capability.",4707
the inspired Paramount the by using present an StarTrek. example Series logic We,We present the logic using an example inspired by the Paramount Series StarTrek.,4708
approximate large We fast for matching. propose graph algorithm a,We propose a fast approximate algorithm for large graph matching.,4709
The test of the linear the representation. size in complexity the is of,The complexity of the test is linear in the size of the representation.,4710
"population handles data and important problems association in estimation. AI, including It","It handles important problems in AI, including data association and population estimation.",4711
to with concepts these matching. We demonstrate citation application,We demonstrate these concepts with application to citation matching.,4712
DMS/SHAPE and give small CMCT additional reactivities data Integrating including improvements.,Integrating DMS/SHAPE data and including CMCT reactivities give small additional improvements.,4713
"probabilistic consider Traditional sequences, symbols flat of only. sequences models however, of","Traditional probabilistic models of sequences, however, consider sequences of flat symbols only.",4714
have hidden Logical models proposed one as Markov solution. been,Logical hidden Markov models have been proposed as one solution.,4715
"over logical sequences of alphabet atoms. sequences, with logical an i.e., They deal","They deal with logical sequences, i.e., sequences over an alphabet of logical atoms.",4716
model a problem. more This the complex at selection of expense comes,This comes at the expense of a more complex model selection problem.,4717
"levels abstraction explored. different have Indeed, to be","Indeed, different abstraction levels have to be explored.",4718
effectiveness. convergence that show We provide SAGEM's and experimental results,We provide convergence and experimental results that show SAGEM's effectiveness.,4719
"such is In inference many addition, applications, inherently in uncertain.","In addition, in many applications, such inference is inherently uncertain.",4720
"the theory, probability uncertainty. Based the we on representation define of associated","Based on probability theory, we define the representation of the associated uncertainty.",4721
system enable to active results These uncertainty. handle such any,These results enable any active system to handle such uncertainty.,4722
inappropriate (e.g. use causes Possible the of include sensors,Possible causes include the use of inappropriate sensors (e.g.,4723
"looking a accumulate inaccuracies laser through (e.g. sensor glass walls),","a laser looking through glass walls), sensor inaccuracies accumulate (e.g.",4724
dynamically static grid objects). with moving used a occupancy,a static occupancy grid used with dynamically moving objects).,4725
increases find structure diagnostic best step to accuracy. that the This aims,This step aims to find the best structure that increases diagnostic accuracy.,4726
very the of the encouraging The discussed results obtained . end at paper are,The very encouraging results obtained are discussed at the end of the paper .,4727
method of can improve The the original the algorithm. mentioned weaknesses mean-shift proposed,The proposed method can improve the mentioned weaknesses of the original mean-shift algorithm.,4728
The performance case algorithm from is using test illustrated this new of the a literature.,The performance of this new algorithm is illustrated using a test case from the literature.,4729
able We are report results. promising to,We are able to report promising results.,4730
"the the addition, a of numerical properties of estimator. we provide study In statistical","In addition, we provide a numerical study of the statistical properties of the estimator.",4731
"Consequently, agent. a the protocol has formaldehyde-free devised, as carbohydrazide been developing using staining","Consequently, a formaldehyde-free staining protocol has been devised, using carbohydrazide as the developing agent.",4732
much of retained the showed and increased peptide sensitivity This coverage staining. protocol silver,This protocol showed much increased peptide coverage and retained the sensitivity of silver staining.,4733
"number Change, global This like IncreasingNValue. important subsumes of a meta-constraint Smooth constraints and","This meta-constraint subsumes a number of important global constraints like Change, Smooth and IncreasingNValue.",4734
combining Borda the results used weighted three The for methods. is method of,The weighted Borda method is used for combining the results of three methods.,4735
the of approach. shows An proposed the example applicability,An example shows the applicability of the proposed approach.,4736
"initiated, chronic of modulation the with inflammation function. leads immune Once to obesity associated cell","Once initiated, chronic inflammation associated with obesity leads to the modulation of immune cell function.",4737
planned pre-diabetic was and two into The study using model. obese mouse groups,The study was planned into two groups using obese and pre-diabetic mouse model.,4738
correction in viz. A immunological important noteworthy readouts,A noteworthy immunological correction in important readouts viz.,4739
"a indeed, especially challenging is, variations. the recognition and illumination pose task, face under Human","Human face recognition is, indeed, a challenging task, especially under the illumination and pose variations.",4740
are show pose illumination variations. features transformed shape and robust that Experiments such to,Experiments show that such transformed shape features are robust to illumination and pose variations.,4741
Comparison proposed superior. algorithms are the shows,Comparison shows the proposed algorithms are superior.,4742
"Relating certain matched. patterns, a be can correlation quickly above a noise threshold with value,","Relating noise patterns, with a correlation above a certain threshold value, can quickly be matched.",4743
"is images quality the objective Therefore, spatial an resolution of for assessment fusion required.","Therefore, an objective quality of the spatial resolution assessment for fusion images is required.",4744
effect of Arabic. for Language tools the using describes Processing Natural (NLP) learning It,It describes the effect of using Natural Language Processing (NLP) tools for learning Arabic.,4745
indexed particular The technique employment explored is the of corpora. in pedagogically case this,The technique explored in this particular case is the employment of pedagogically indexed corpora.,4746
"architecture Agent. Operator Parameter Agent three and of agents: of types User Agent, consists Our","Our architecture consists of three types of agents: User Agent, Operator Agent and Parameter Agent.",4747
process We using probit priors. (GP) probabilistic multinomial classification Gaussian consider,We consider probabilistic multinomial probit classification using Gaussian process (GP) priors.,4748
Such data. common require to to patterns methods recognize analyze,Such methods require to recognize common patterns to analyze data.,4749
a The of most often manual work consuming very process. is inpainting time,The manual work of inpainting is most often a very time consuming process.,4750
"of automatic this is faster. and to digitalization technique, Due it","Due to digitalization of this technique, it is automatic and faster.",4751
in compared longer trials Response to latency correct was trials. error,Response latency was longer in correct trials compared to error trials.,4752
dimensionality component feature is a linear reduction for analysis popular (PCA) tool extraction. and Principal,Principal component analysis (PCA) is a popular tool for linear dimensionality reduction and feature extraction.,4753
"kernel we review ideas of first paper, this PCA PCA. In the basic and","In this paper, we first review the basic ideas of PCA and kernel PCA.",4754
the for PCA. Then reconstruction kernel of on we focus pre-images,Then we focus on the reconstruction of pre-images for kernel PCA.,4755
"face it the ASMs, and construct also use kernel models. We human implement PCA-based to","We also implement the kernel PCA-based ASMs, and use it to construct human face models.",4756
a Markov our Networks Logic natural logic-based formalism. candidate are for,Markov Logic Networks are a natural candidate for our logic-based formalism.,4757
we show overcome assumptions and these under can what We how problems.,We show how and under what assumptions we can overcome these problems.,4758
Hodgkin-Huxley of giant model axon no bistability. squid The the shows,The Hodgkin-Huxley model of the squid giant axon shows no bistability.,4759
values the for two sets. between parameter to transitions analyze varied the We model parameter,We varied parameter values for the model to analyze transitions between the two parameter sets.,4760
the show bistability Na+ of characterizes that inactivation primarily We the current.,We show that bistability primarily characterizes the inactivation of the Na+ current.,4761
the is environmental chloride an that system. Cadmium affects reproductive toxic male Background:,Background: Cadmium chloride is an environmental toxic that affects the male reproductive system.,4762
"and transferrin. able and Zinc strengthen TJ was chloride to AJ, decrease","Zinc chloride was able to strengthen TJ and AJ, and decrease transferrin.",4763
"Key AJ, . prostate, Testis words: TJ, cadmium, zinc, immunohistochemistry","Key words: Testis prostate, cadmium, zinc, TJ, AJ, immunohistochemistry .",4764
"complete-data Under analyse likelihood the cluster of function we Gaussian models. assumptions, weighted","Under Gaussian assumptions, we analyse the complete-data likelihood function of cluster weighted models.",4765
"direction. that Furthermore, array elevation systems of the is larger in than lateral resolution linear","Furthermore, the elevation resolution of linear array systems is larger than that in lateral direction.",4766
degrades asymmetry arrays. This linear images using obtained C-scan,This asymmetry degrades C-scan images obtained using linear arrays.,4767
ones. comparison the yielded the This with in other best result filter,This filter yielded the best result in comparison with the other ones.,4768
reference complicated models. These values as serve statistical for results more,These results serve as reference values for more complicated statistical models.,4769
"depends their critical. However, parameters is on parameters their the determining and performance strongly","However, their performance strongly depends on their parameters and determining the parameters is critical.",4770
applied algorithm our We the to Tempering. Parallel,We applied our algorithm to the Parallel Tempering.,4771
"adaptive tunes on developed we that That the is, Tempering the Parallel fly. parameters","That is, we developed adaptive Parallel Tempering that tunes the parameters on the fly.",4772
potential usually and cons Each of choice strengths. various pros has,Each potential choice has usually pros and cons of various strengths.,4773
The VFC not cube accuracy be three-dimensionally. can in of rotated even guaranteed,The accuracy of VFC can not be guaranteed even in cube rotated three-dimensionally.,4774
time Both size of in CFD. proportional to algorithms run the the,Both algorithms run in time proportional to the size of the CFD.,4775
Case approach neurological been comparisons vs to diseases. classical control have of the the study,Case vs control comparisons have been the classical approach to the study of neurological diseases.,4776
"most However, group. fall will either patients into cleanly not","However, most patients will not fall cleanly into either group.",4777
models intermediate that the brain function describe requires To the interpolate disease between states.,To describe the intermediate brain function requires models that interpolate between the disease states.,4778
"measures Our distribution of of incorporate models e.g. activation, spatial","Our models incorporate spatial distribution of measures of activation, e.g.",4779
"behavioural fMRI Disease GPs phenotypes applied Alzheimer's We e.g. model across (AD) image to measures,","We applied GPs to model fMRI image phenotypes across Alzheimer's Disease (AD) behavioural measures, e.g.",4780
"obtained MMSE/ACE scores, and non-observed at values. predictions","scores, and obtained predictions at non-observed MMSE/ACE values.",4781
support algorithms. to efficient information and construct less require reasoning models Such,Such models require less information to construct and support efficient reasoning algorithms.,4782
"The combines with networks framework graphical representation. networks, and the belief semantics constraint defining","The framework combines belief networks with constraint networks, defining the semantics and graphical representation.",4783
filters. non-linear Stack case filters special are of a,Stack filters are a special case of non-linear filters.,4784
"a Boolean image which Each filtered function, filter. is by the binary characterizes then","Each binary image is then filtered by a Boolean function, which characterizes the filter.",4785
"propose problem, a we annealing-based algorithm, MAP To simulated this AnnealedMAP algorithm. address","To address this problem, we propose AnnealedMAP algorithm, a simulated annealing-based MAP algorithm.",4786
this tested We Bayesian algorithm real several networks. on,We tested this algorithm on several real Bayesian networks.,4787
The to can approach families of construction be other applied problems. same,The same approach can be applied to other families of construction problems.,4788
characterize terms in called expressions in algebraic of networks We Bayesian quasi-probabilities. probabilities,We characterize probabilities in Bayesian networks in terms of algebraic expressions called quasi-probabilities.,4789
improve Quasi-probabilities probabilistic easily are manipulated to the efficiency of inference.,Quasi-probabilities are easily manipulated to improve the efficiency of probabilistic inference.,4790
expected total policy minimizes optimal diagnostic cost. the An,An optimal diagnostic policy minimizes the expected total cost.,4791
as this Markov diagnosis Decision formalize Process a We (MDP). process,We formalize this diagnosis process as a Markov Decision Process (MDP).,4792
learn process. for Bayesian diagnosis We assume not a network the nor do,We do not learn nor assume a Bayesian network for the diagnosis process.,4793
Regularizers control the speed developed overfitting search. are to up and,Regularizers are developed to control overfitting and speed up the search.,4794
overfitting systematic that search. integrates research This the is into first prevention,This research is the first that integrates overfitting prevention into systematic search.,4795
describe Processes an We Markov state structure approach Decision exploiting with continuous in for variables.,We describe an approach for exploiting structure in Markov Decision Processes with continuous state variables.,4796
describe piecewise We representations. constant the for first algorithm,We first describe the algorithm for piecewise constant representations.,4797
representation generally axioms of the independence. upon of independence relations semigraphoid well-known builds The,The representation of independence relations generally builds upon the well-known semigraphoid axioms of independence.,4798
of of to capture this knowledge. We monotonicity two concepts define type,We define two concepts of monotonicity to capture this type of knowledge.,4799
can be this networks. through The paper theory achieved shows of how the credal,The paper shows how this can be achieved through the theory of credal networks.,4800
The with work to existing can algorithm this representation. be adapted,The existing algorithm can be adapted to work with this representation.,4801
to apply multiagent allows methods formulation us our also settings. to This collaborative,This formulation also allows us to apply our methods to collaborative multiagent settings.,4802
genuine a class theories. converge two trends probability to show of these We,We show these two trends converge to a class of genuine probability theories.,4803
"the model samples. only death the identified large If for unknown, probability can be is","If the death probability is unknown, the model can be identified only for large samples.",4804
We consider decision of computing the policies optimal processes. problem Markov relational for generalised,We consider the problem of computing optimal generalised policies for relational Markov decision processes.,4805
consider of the data. estimating underlying sample of distribution We the problem observed an,We consider the problem of estimating the distribution underlying an observed sample of data.,4806
symbolic probability paper generalization This theory a decision for proposes theory a (SP). of,This paper proposes a decision theory for a symbolic generalization of probability theory (SP).,4807
a interpretation elicitation. useful because subjective the uncertainty for provides It basis is,It is useful because the subjective interpretation provides a basis for uncertainty elicitation.,4808
"potentials utility used be can MTE functions. Also, approximate to","Also, MTE potentials can be used to approximate utility functions.",4809
variable MTE are by solved elimination diagrams using fusion influence algorithm. a,MTE influence diagrams are solved by variable elimination using a fusion algorithm.,4810
"evaluating and to Most for comparing approaches these problems. testbed a importantly, rigorous it provides","Most importantly, it provides a rigorous testbed for evaluating and comparing approaches to these problems.",4811
"is publicly agents, the benchmark the All of available. including software,","All of the software, including the benchmark agents, is publicly available.",4812
"Detrended of performed and (DFA) ST, was times on series analysis the fluctuation SS. SL,","Detrended fluctuation analysis (DFA) was performed on the times series of ST, SL, and SS.",4813
"contrary, parameters condition. all anti-persistent under the were On the dual-constraints three","On the contrary, all the three parameters were anti-persistent under dual-constraints condition.",4814
have soundness been convergence HSVI's and proven.,HSVI's soundness and convergence have been proven.,4815
ATL-type logical verify systems us allows This express frameworks. and BDI properties of to using,This allows us to express and verify properties of BDI systems using ATL-type logical frameworks.,4816
"graphs conventional are to context, graphs (MFGs). referred as the this factor factor In multiplicative","In this context, the conventional factor graphs are referred to as multiplicative factor graphs (MFGs).",4817
"the the an AR cases In choice impact, results. produced best the had where","In the cases where the choice had an impact, AR produced the best results.",4818
function. they However work an intractability the are due to of with to complicated likelihood,However they are complicated to work with due to an intractability of the likelihood function.,4819
tractable natural the consider function. is approximations likelihood It to to therefore,It is therefore natural to consider tractable approximations to the likelihood function.,4820
offer such to approach likelihoods constructing approximation. a principled Composite,Composite likelihoods offer a principled approach to constructing such approximation.,4821
very complexity. semantics have high by merging operators which computational are defined Most methods,Most merging operators are defined by semantics methods which have very high computational complexity.,4822
syntactical work define In merging this properties. we its operators rationality exploring some and,In this work we define some syntactical merging operators and exploring its rationality properties.,4823
The arise simple motor of from the task. learning data performance behavioural a,The data arise from the performance of a simple behavioural motor learning task.,4824
regions summary allows and contributing interpreted. centrality compact to This to brain visualized be key,This compact summary allows key brain regions contributing to centrality to be visualized and interpreted.,4825
explore illustrative these correlation We models two for of among cells. questions,We explore these questions for two illustrative models of correlation among cells.,4826
for expressions disparity of expected Yule We a derive distribution the under trait tree.,We derive expressions for the distribution of expected trait disparity under a Yule tree.,4827
fixed Our or does tree. not method simulation a require phylogenetic,Our method does not require simulation or a fixed phylogenetic tree.,4828
"of an the required. fusion assessment quality for objective images resolution is spatial So,","So, an objective quality of the spatial resolution assessment for fusion images is required.",4829
video and of behaviour recognition presents This in paper sequence. review understanding activity a human,This paper presents a review of human activity recognition and behaviour understanding in video sequence.,4830
diseases horizontally and both of A transmitted. range infectious vertically wide are,A wide range of infectious diseases are both vertically and horizontally transmitted.,4831
"disease explicit affording evolution. This adaptable infectious expression easily into insights is diseases, to specific","This explicit expression is easily adaptable to specific infectious diseases, affording insights into disease evolution.",4832
A may pertains how theoretical neuroscience resolved. deadlock this open in be to fundamental question,A fundamental open question in theoretical neuroscience pertains to how this deadlock may be resolved.,4833
"algorithms proposed are for estimation. super-resolution spectrum sparsity-based Recently,","Recently, sparsity-based algorithms are proposed for super-resolution spectrum estimation.",4834
dimensions. in screening high A spectrum computation probabilistic developed fast is for also,A probabilistic spectrum screening is also developed for fast computation in high dimensions.,4835
A Information used Criterion determine regularization to the parameters. high-dimensional of is Bayesian data-resampling version,A data-resampling version of high-dimensional Bayesian Information Criterion is used to determine the regularization parameters.,4836
connectivity. strengthen the is preserve to links aim network key to The expected,The aim is to strengthen key links to preserve the expected network connectivity.,4837
and to the result facilitate can approaches be exact used problem. metaheuristic to This,This result can be used to facilitate metaheuristic and exact approaches to the problem.,4838
future decision possible are often problems simulating by sequences. approximately Sequential action solvable,Sequential decision problems are often approximately solvable by simulating possible future action sequences.,4839
against NBUE tests and compared. alternatives discussed of These exponentiality are,These tests of exponentiality against NBUE alternatives are discussed and compared.,4840
simulations. is of empirical size the tests The by obtained,The empirical size of the tests is obtained by simulations.,4841
different simulations. Carlo popular Monte comparisons Power are using done for alternatives,Power comparisons for different popular alternatives are done using Monte Carlo simulations.,4842
both comparisons made are sizes. large sample These for and small,These comparisons are made for both small and large sample sizes.,4843
and of interface. classes a is algorithm The collection a implemented as C++ Python,The algorithm is implemented as a collection of C++ classes and a Python interface.,4844
"exceedingly gene is While precise. highly expression is erratic, embryonic development often usually","While gene expression is often highly erratic, embryonic development is usually exceedingly precise.",4845
and inter-embryonic development against variations intra- How is robust understood. is not,How development is robust against intra- and inter-embryonic variations is not understood.,4846
"in repression their enhances variations mutual robustness against morphogen Moreover, the embryo-to-embryo dramatically levels.","Moreover, mutual repression dramatically enhances their robustness against embryo-to-embryo variations in the morphogen levels.",4847
"gene include chromosome These changes fission, duplication, rearrangement. and loss, fusion, gain,","These changes include gene gain, loss, duplication, chromosome fusion, fission, and rearrangement.",4848
in that be suggest Our may hidden error pervasive alignments. breakpoint genome results,Our results suggest that hidden breakpoint error may be pervasive in genome alignments.,4849
"Europe. Judeo-Khazars of their to collapse fled empire, the Eastern Following the","Following the collapse of their empire, the Judeo-Khazars fled to Eastern Europe.",4850
European of Jewry rise The explained the therefore contribution of by is the Judeo-Khazars.,The rise of European Jewry is therefore explained by the contribution of the Judeo-Khazars.,4851
"retinal non-random of in types retinal termed patterns, positioned neurons Most mosaics. spatially are","Most types of retinal neurons are spatially positioned in non-random patterns, termed retinal mosaics.",4852
the be mosaics. to in important mechanisms formation developmental of Several thought these are,Several developmental mechanisms are thought to be important in the formation of these mosaics.,4853
studied methods Transcriptional regulatory have years. for been inference network,Transcriptional regulatory network inference methods have been studied for years.,4854
to We simulations findings. numerical theoretical our finally validate provide,We finally provide numerical simulations to validate our theoretical findings.,4855
and objects. in comprise various interactions several a involve with sub-activities sequence performed Human activities,Human activities comprise several sub-activities performed in a sequence and involve interactions with various objects.,4856
for a This reasoning affordances central task the activity about makes recognition. object,This makes reasoning about the object affordances a central task for activity recognition.,4857
distributed species Populations and widely encounter specific of to often conditions. environmental adapt,Populations of widely distributed species often encounter and adapt to specific environmental conditions.,4858
"identification bias note that loci. individual differentiated we can the populations Further, strongly adaptive of","Further, we note that strongly differentiated individual populations can bias the identification of adaptive loci.",4859
biometrics.  Fingerprint most publicized is  one of well-known and the reconstruction,  Fingerprint reconstruction is one of the most well-known and publicized biometrics.,4860
"the numerous because inherent acquisition, of popular reconstruction sources the Fingerprint of ease (e.g. is","Fingerprint reconstruction is popular because of the inherent ease of acquisition, the numerous sources (e.g.",4861
most have identification. means the positive always practical been Fingerprints of and,Fingerprints have always been the most practical and positive means of identification.,4862
performance of paper different an gives measurement overview and methods of reconstruction. fingerprint The the,The paper gives an overview of different methods and performance measurement of the fingerprint reconstruction.,4863
filter paper blocked aims image this to The boundaries. presented deblocking the in algorithm,The image deblocking algorithm presented in this paper aims to filter the blocked boundaries.,4864
common feature folding a and are bonds protein Hydrogen aggregation. in,Hydrogen bonds are a common feature in protein folding and aggregation.,4865
"a potential, designed bond hydrogen have J. Chem. see We","We have designed a hydrogen bond potential, see J. Chem.",4866
"protein carry describe tertiary we To to interactions. need studies, full out potentials additional","To carry out full protein studies, we need additional potentials to describe tertiary interactions.",4867
different We have of points discussed view. two,We have discussed two different points of view.,4868
bond one. hydrogen potential The a is with first structure-based the one combination the of,The first one is the combination of the hydrogen bond potential with a structure-based one.,4869
"peptide we aggregation have it, sequence. Thanks re-studied to including","Thanks to it, we have re-studied peptide aggregation including sequence.",4870
"Besides, folded shapes. we proteins with complete different have simulated","Besides, we have simulated complete proteins with different folded shapes.",4871
"as considerably and are usage the such computational space disk costs, runtime, memory reduced. Thereby,","Thereby, the computational costs, such as runtime, memory usage and disk space are considerably reduced.",4872
"large-scale are systems, Gas design problems and are corresponding and challenging. Transmission complex Networks control","Gas Transmission Networks are large-scale complex systems, and corresponding design and control problems are challenging.",4873
systems hybrid required networks provides by a models. framework these analysis We that present,We present these networks by a hybrid systems framework that provides required analysis models.,4874
"hybrid computational using we decision-making discuss and optimization discrete methods. Further,","Further, we discuss decision-making using computational discrete and hybrid optimization methods.",4875
are efficiency of presented the the illustrate method. to Simulations,Simulations are presented to illustrate the efficiency of the method.,4876
time successful most of system technology. and is the attendance of biometric applications Biometric one,Biometric time and attendance system is one of the most successful applications of biometric technology.,4877
biometric of the the user store minutiae a in template Most database. systems fingerprint-based,Most fingerprint-based biometric systems store the minutiae template of a user in the database.,4878
from reconstruction image proposed the The reconstructs algorithm phase minutiae.,The proposed reconstruction algorithm reconstructs the phase image from minutiae.,4879
product a has used usually as the T-norm. these merits calculating For been,For calculating these merits the product has been usually used as a T-norm.,4880
this for In different paper T-norms support confidence have calculating measures. the and used been,In this paper different T-norms have been used for calculating the confidence and support measures.,4881
to would hard find been a Such demimartingales result have whithout techniques. very,Such a result would have been very hard to find whithout demimartingales techniques.,4882
The time. spreading it disrupts in channel the transmitted signal by,The channel disrupts the transmitted signal by spreading it in time.,4883
"in is slow the it LMS algorithm robust Although, convergence. reliable, is and","Although, the LMS algorithm is robust and reliable, it is slow in convergence.",4884
"of and guaranteed. stability are reaching conditions, the The robustness proposed the with system controller","The reaching conditions, stability and robustness of the system with the proposed controller are guaranteed.",4885
"suited controller addition, proposed and for well implementation. design simple the is In","In addition, the proposed controller is well suited for simple design and implementation.",4886
of inference A data framework from lacking. populations is for diversity currently such statistical,A statistical framework for inference from diversity data of such populations is currently lacking.,4887
paper approximations optimal In considered. segmentation means constant by image is of the piecewise the,In the paper the optimal image segmentation by means of piecewise constant approximations is considered.,4888
analytically The image. experimentally on proved the the and proposed example of solution standard is,The proposed solution is proved analytically and experimentally on the example of the standard image.,4889
cases energy are Rosetta Remaining all-atom the traced to inaccuracies problem function. in,Remaining problem cases are traced to inaccuracies in the Rosetta all-atom energy function.,4890
evaluations tracks. in are this of The five divided campaign,The evaluations of this campaign are divided in five tracks.,4891
or improved OAEI Three to tracks new are of compared been campaigns. have previous these,Three of these tracks are new or have been improved compared to previous OAEI campaigns.,4892
image. chrominance of Y components are and is component;Cb Cr luminance the,Y is luminance component;Cb and Cr are chrominance components of the image.,4893
"inference Under methods. two this computational we describe model,","Under this model, we describe two computational inference methods.",4894
"for this Multiple have ""missing been heritability"" proposed. explanations","Multiple explanations for this ""missing heritability"" have been proposed.",4895
Detected contribution. interactions this two-locus only of a explain minority,Detected two-locus interactions explain only a minority of this contribution.,4896
genomic diverse The understanding fundamental Motivation: biology. comparison of genome is datasets to,Motivation: The comparison of diverse genomic datasets is fundamental to understanding genome biology.,4897
methods at We BITS existing intersections. counting demonstrate that interval outperforms,We demonstrate that BITS outperforms existing methods at counting interval intersections.,4898
for developed have Many time data. series been methods,Many methods have been developed for time series data.,4899
approach is conventional Bayesian A networks. causal,A conventional approach is causal Bayesian networks.,4900
"However, of ill-posed. networks is Bayesian estimation","However, estimation of Bayesian networks is ill-posed.",4901
underlying of uniquely algorithm propose new causal We for genes. a network the identifying discovery,We propose a new discovery algorithm for uniquely identifying the underlying causal network of genes.,4902
solution first manner. the a of algorithm in the The greedy algorithm improves second,The second algorithm improves the solution of the first algorithm in a greedy manner.,4903
two This elimination paper makes (LVE). contributions variable lifted to,This paper makes two contributions to lifted variable elimination (LVE).,4904
"operator inference novel First, introduce we group called inversion. a","First, we introduce a novel inference operator called group inversion.",4905
previously is sets. focus performance our to The evaluate the proposed feature of of analysis,The focus of our analysis is to evaluate the performance of previously proposed feature sets.,4906
this by existing casting in pipeline. common We algorithms a achieve,We achieve this by casting existing algorithms in a common pipeline.,4907
basis. basis the per-pixel analyzed and We on performance a per-image detection on a,We analyzed the detection performance on a per-image basis and on a per-pixel basis.,4908
"We and image created manipulation. for systematic real-world challenging copy-move software a dataset, a framework","We created a challenging real-world copy-move dataset, and a software framework for systematic image manipulation.",4909
Crop-wild study evolution ideal systems to represent through examine opportunities hybridization.,Crop-wild study systems represent ideal opportunities to examine evolution through hybridization.,4910
"example, ssp. wild mays conspecific For the and Zea maize teosinte","For example, maize and the conspecific wild teosinte Zea mays ssp.",4911
"of (hereafter, mexicana) highland are mexicana, Mexico. to in fields the known hybridize","mexicana, (hereafter, mexicana) are known to hybridize in the fields of highland Mexico.",4912
extent is of genomic between evolutionary the the understood. these introgression taxa nor Neither importance,Neither the genomic extent nor the evolutionary importance of introgression between these taxa is understood.,4913
"very contrast, was introgression mexicana. to found little evidence for In from maize adaptive","In contrast, very little evidence was found for adaptive introgression from maize to mexicana.",4914
one protein to discussion Our site. is binding limited,Our discussion is limited to one protein binding site.,4915
engines of specific information. domain search None provide the,None of the search engines provide domain specific information.,4916
people in can This health by used medical information care be domain. and working,This information can be used by people working in medical and health care domain.,4917
"graphical likelihood Gaussian estimation we models. this article, composite sparse discuss the of In","In this article, we discuss the composite likelihood estimation of sparse Gaussian graphical models.",4918
the and consistent offers an estimators. composite yields of objective likelihood formulation function The alternative,The composite likelihood offers an alternative formulation of the objective function and yields consistent estimators.,4919
pool represented repetitive in dealt the second mRNAs. elements The with part of,The second part dealt with repetitive elements represented in the pool of mRNAs.,4920
repetitive two a identified between difference that show biological We conditions. significant elements,We identified repetitive elements that show a significant difference between two biological conditions.,4921
"former. emphasis on was the the However,","However, the emphasis was on the former.",4922
our confirmed Laboratory with findings. qPCR validation,Laboratory validation with qPCR confirmed our findings.,4923
characters. characters to recognized The Malayalam are corresponding mapped,The recognized characters are mapped to corresponding Malayalam characters.,4924
efficiency system. the developed results show of These the,These results show the efficiency of the developed system.,4925
These offer content for highly discriminant categorization. behavior a features extracted,These extracted features offer a highly discriminant behavior for content categorization.,4926
through validated method theoretical are datasets. experimentally different analyzed properties The and of the two,The theoretical properties of the method are analyzed and validated experimentally through two different datasets.,4927
a simultaneously to We timing and the judge frequency ability sound. of study human the,We study human ability to simultaneously judge the frequency and the timing of a sound.,4928
some explain This theory recent quantitatively can findings. experimental,This theory can quantitatively explain some recent experimental findings.,4929
hypothesis multimodal region. such processes binding states depend single One upon within that a representations,One hypothesis states that such representations depend upon multimodal binding processes within a single region.,4930
cortex conceptual representation. provides is evidence that This engaged perirhinal in result,This result provides evidence that perirhinal cortex is engaged in conceptual representation.,4931
success with data The algorithm large several is tested on sets.,The algorithm is tested with success on several large data sets.,4932
objects. a shape This between new for distance measuring article presents dissimilarity,This article presents a new distance for measuring shape dissimilarity between objects.,4933
Laplace Recent descriptors. the introduced of publications as shape eigenvalues use the compact of operator,Recent publications introduced the use of eigenvalues of the Laplace operator as compact shape descriptors.,4934
through the The heat-trace. is derived definition WESD analysing of,The definition of WESD is derived through analysing the heat-trace.,4935
"its properties. and distance theoretical important prove We resulting the analyse definition, present","We analyse the resulting distance definition, present and prove its important theoretical properties.",4936
"presented. real objects experiments on Lastly, are synthetic and conducted","Lastly, experiments conducted on synthetic and real objects are presented.",4937
the They other not may fusion suitable images. of for be or may,They may or may not be suitable for the fusion of other images.,4938
exact like and involve method approximations not is time-discretization. does Our,Our method is exact and does not involve approximations like time-discretization.,4939
"in admixture locations. varied with greater among Admixture urban populations, proportions","Admixture proportions varied among populations, with greater admixture in urban locations.",4940
arms. strong in elevations diversity were in across European sample with associated the Inversions chromosome,Inversions in the European sample were associated with strong elevations in diversity across chromosome arms.,4941
conducted that selection. targets Genomic of were positive loci scans represent may to identify,Genomic scans were conducted to identify loci that may represent targets of positive selection.,4942
Distributed System the (DPS). new called The Pharaoh algorithm is,The new algorithm is called the Distributed Pharaoh System (DPS).,4943
combinatorial Traveling (GTSP) The the of Generalized is problems. Salesman optimization one NP-hard Problem,The Generalized Traveling Salesman Problem (GTSP) is one of the NP-hard combinatorial optimization problems.,4944
for solving nowadays used successfully real Agent-based problems. complex involving life are approaches,Agent-based approaches involving are successfully used nowadays for solving real life complex problems.,4945
with data For generates predictors singular categorical often subsamples. random subsampling,For data with categorical predictors random subsampling often generates singular subsamples.,4946
"estimates, coefficient discarded. be these calculate subsamples Since used be to to they cannot have","Since these subsamples cannot be used to calculate coefficient estimates, they have to be discarded.",4947
an This nonsingular improved introduces subsamples. only subsampling generates paper algorithm that,This paper introduces an improved subsampling algorithm that only generates nonsingular subsamples.,4948
of the annotation. discuss level advantages pixel We,We discuss the advantages of pixel level annotation.,4949
"scene, images. word from camera captured These view and born-digital cropped have images been street","These word images have been cropped from camera captured scene, born-digital and street view images.",4950
OCR. the version image We of Omnipage segmented Nuance recognize trial word the using,We recognize the segmented word image using the trial version of Nuance Omnipage OCR.,4951
for here analysis presented. video from framework single is group a unsupervised activity A,A framework for unsupervised group activity analysis from a single video is here presented.,4952
model. and advantages noise fast resistance The of response to over carry this,The advantages of fast response and noise resistance carry over to this model.,4953
procedure. search Our learning agent-based model local a a on approach based involving is,Our approach is based on a learning agent-based model involving a local search procedure.,4954
variance We exact rarefaction. for PD have and derived under mean of formulae the,We have derived exact formulae for the mean and variance of PD under rarefaction.,4955
models The often biomolecular Hox been collinearity approached enigma using on has based mechanisms. gene,The Hox gene collinearity enigma has often been approached using models based on biomolecular mechanisms.,4956
gene the of have Hox recently translocations model. observed biophysical been in support,Hox gene translocations have recently been observed in support of the biophysical model.,4957
"On convincing explanation. contrary, biophysical the model offers the","On the contrary, the biophysical model offers convincing explanation.",4958
"homeostasis control the unknown. are of IgM-secreting cells that the mechanisms B activated However,","However, the mechanisms that control homeostasis of the activated IgM-secreting B cells are unknown.",4959
is on These are based using stimuli experimental that defined a parametrization. model generative controlled,These stimuli are defined using a generative model that is based on controlled experimental parametrization.,4960
Python-based describe mathematically the these of and stimuli propose We implementation. an construction open-source,We describe the construction of these stimuli mathematically and propose an open-source Python-based implementation.,4961
are the of this use shown. framework of Examples,Examples of the use of this framework are shown.,4962
"as extensions touch, also such color other and propose We audition. modalities to vision,","We also propose extensions to other modalities such as color vision, touch, and audition.",4963
on The between hosts a has structure contact disease critical spread. influence,The contact structure between hosts has a critical influence on disease spread.,4964
the contact of predictions models. have a epidemiological sexual to and effect on strong networks),sexual contact networks) and to have a strong effect on the predictions of epidemiological models.,4965
the reduction. main of Unfolding for methods is dimensionality one Variance Maximum (nonlinear),Maximum Variance Unfolding is one of the main methods for (nonlinear) dimensionality reduction.,4966
"of its convergence under sample rates providing study limit, standard large specific assumptions. We","We study its large sample limit, providing specific rates of convergence under standard assumptions.",4967
"motion-based to global coding sufficient motion. is Here, develop infer predictive we that the hypothesis","Here, we develop the hypothesis that motion-based predictive coding is sufficient to infer global motion.",4968
is based representation probabilistic of implementation context-dependent a a motion. diffusion on Our of,Our implementation is based on a context-dependent diffusion of a probabilistic representation of motion.,4969
this demonstrate underlying mechanisms. two is result the solution of that We,We demonstrate that this solution is the result of two underlying mechanisms.,4970
quadratic of this (e.g. To dependence infer the arbitrary presence in,To infer this quadratic dependence in the presence of arbitrary (e.g.,4971
"with ingredient, distance network key global/local a a As procedure. bootstrap combined we employ a","As a key ingredient, we employ a global/local network distance combined with a bootstrap procedure.",4972
"are, systemic These in revealing however contrast certain low useful images diseases.","These low contrast images are, however useful in revealing certain systemic diseases.",4973
is of image noisy processing. images challenging task in Corner a detection,Corner detection of noisy images is a challenging task in image processing.,4974
transmission. Natural images corrupted during and get often by acquisition noise,Natural images often get corrupted by noise during acquisition and transmission.,4975
for same. thresholding is Adaptive applied approach the wavelet,Adaptive wavelet thresholding approach is applied for the same.,4976
transmission An application presented the an is problem education as to of illustration.,An application to the problem of education transmission is presented as an illustration.,4977
bin optimization relevant highly are packing problems Two-dimensional problems. combinatorial,Two-dimensional bin packing problems are highly relevant combinatorial optimization problems.,4978
randomized Additionally an algorithm. within of use the heuristic evolutionary this we propose one-pass,Additionally we propose the use of this randomized one-pass heuristic within an evolutionary algorithm.,4979
very compares In algorithm the evolutionary to approaches. current state-of-the-art favorably particular,In particular the evolutionary algorithm compares very favorably to current state-of-the-art approaches.,4980
plane for vital Automatic diseases. of identification cardiac features time is of important extraction,Automatic extraction of time plane features is important for identification of vital cardiac diseases.,4981
by the The wavelet bands original of are peaks achieved ECG composition signal. of Daubeheissub,The peaks are achieved by the composition of Daubeheissub bands wavelet of original ECG signal.,4982
"evolution variation genetic natural to selection, adaptive proceed. addition to In requires","In addition to natural selection, adaptive evolution requires genetic variation to proceed.",4983
are referred genetic Such limitations as often to constraints.,Such limitations are often referred to as genetic constraints.,4984
"rarely predictions clear when to expect Unfortunately, constraints are available. about","Unfortunately, clear predictions about when to expect constraints are rarely available.",4985
"an regarding Therefore, predictions that constraints. we developed experimental system specific provides","Therefore, we developed an experimental system that provides specific predictions regarding constraints.",4986
induce We mutation this would homeosis. constraint due to a genetic predicted this that,We predicted that this mutation would induce a genetic constraint due to this homeosis.,4987
of the time prove the about the event. results hybridization of We distribution,We prove results about the distribution of the time of the hybridization event.,4988
often overlooked symbionts studies invasion. Fungal are plant of in,Fungal symbionts are often overlooked in studies of plant invasion.,4989
"Nevertheless, success be their the the invader. role could essential competitive to of","Nevertheless, their role could be essential to the competitive success of the invader.",4990
in the We endophytes invasive fungal knapweed). stoebe studied Centaurea widespread (common,We studied fungal endophytes in the widespread invasive Centaurea stoebe (common knapweed).,4991
"experiment inoculants. root belonging endophytes phylotypes, as to six non-clavicipitaceous the main were employed In","In the main experiment non-clavicipitaceous endophytes belonging to six phylotypes, were employed as root inoculants.",4992
have infrastructures. systems a Network systems computer for crucial detection security become intrusion issue,Network intrusion detection systems have become a crucial issue for computer systems security infrastructures.,4993
to model mining proposed a improve this data paper on based This problem techniques.,This paper proposed a model to improve this problem based on data mining techniques.,4994
is novel to attacks rules algorithm firewall. Apriori and generate predict real-time for used,Apriori algorithm is used to predict novel attacks and generate real-time rules for firewall.,4995
data relationships among large of set extracts interesting correlation algorithm items. Apriori,Apriori algorithm extracts interesting correlation relationships among large set of data items.,4996
Apriori is best-known rules. algorithm association mine to the,Apriori is the best-known algorithm to mine association rules.,4997
to scale. on way rules association an find large innovative is This,This is an innovative way to find association rules on large scale.,4998
faster ones. and the male the theory major are theory The dominance,The dominance theory and the faster male theory are the major ones.,4999
"note, validity. I In composite scientific the challenge and this theory its","In this note, I challenge the composite theory and its scientific validity.",5000
"players way and/or same are all extent. affected the the same not However, in to","However, not all players are affected in the same way and/or to the same extent.",5001
"in worse non-diversified generics labs Small, the position. are","Small, non-diversified generics labs are in the worse position.",5002
potential for labs. smaller have a with diversification an markets across higher advantage over Labs,Labs with a higher potential for diversification across markets have an advantage over smaller labs.,5003
"contains highly variable virus envelope and (HA) hemagglutinin neuraminidase glycoproteins, two Influenza (NA).","Influenza virus contains two highly variable envelope glycoproteins, hemagglutinin (HA) and neuraminidase (NA).",5004
predefined Synchronization matched without synaptic emerges phenomenon as precise delays. transient a,Synchronization emerges as a transient phenomenon without predefined precise matched synaptic delays.,5005
"Based the and and for discussions recommendations implementation. development are review, on presented algorithm","Based on the review, recommendations and discussions are presented for algorithm development and implementation.",5006
in to pungent aldehydes present soapy largely attributed several or This cilantro. aroma is,This soapy or pungent aroma is largely attributed to several aldehydes present in cilantro.,5007
of statistical setting. novel We a offer a view AdaBoost in,We offer a novel view of AdaBoost in a statistical setting.,5008
propose label is binary noise for modeled model hierarchically. Bayesian in which classification a We,We propose a Bayesian model for binary classification in which label noise is modeled hierarchically.,5009
from experimental AdaBoost show datasets. results four and to close We connections give its,We show its close connections to AdaBoost and give experimental results from four datasets.,5010
"In this is paper, inpainting algorithm wavelets using a hierarchical proposed.","In this paper, a hierarchical inpainting algorithm using wavelets is proposed.",5011
factors. is the tested performance different proposed using of algorithm The,The performance of the proposed algorithm is tested using different factors.,5012
encodes the the We about that information image blurry rich show blur kernel. shall itself,We shall show that the blurry image itself encodes rich information about the blur kernel.,5013
"complex, understood. myopia, of be poorly to etiology known is The although","The etiology of myopia, although known to be complex, is poorly understood.",5014
This power estimation evidence. the the posterior on to focuses approach paper refining improve of,This paper focuses on refining the power posterior approach to improve estimation of the evidence.,5015
"tuning. the other tempering posterior of some involves power common with degree In algorithms,","In common with other tempering algorithms, the power posterior involves some degree of tuning.",5016
naive is mildly brute tasks. Performing for even infeasible force search a involved,Performing a naive brute force search is infeasible for even mildly involved tasks.,5017
assessment meta of The a-posteriori. is the is only possible that problem information,The problem is that assessment of the meta information is only possible a-posteriori.,5018
important at The pattern the step is this process most recognition.,The most important step at this process is the pattern recognition.,5019
the called popular structure for is Multilayer employed A Perceptron neural classification. network,A popular neural network structure called Multilayer Perceptron is employed for the classification.,5020
"the streaming video. procedures carried for are Afterwards, same out","Afterwards, the same procedures are carried out for streaming video.",5021
a are standard All PC. the of above on realized depicted steps,All of the steps depicted above are realized on a standard PC.,5022
task a of processing. noisy in detection image is images challenging Corner,Corner detection of noisy images is a challenging task in image processing.,5023
noise images and by often Natural transmission. get acquisition corrupted during,Natural images often get corrupted by noise during acquisition and transmission.,5024
is wavelet approach same. applied thresholding Adaptive for the,Adaptive wavelet thresholding approach is applied for the same.,5025
conditions The complemented information bounds. sufficient lower theoretic are with,The sufficient conditions are complemented with information theoretic lower bounds.,5026
shared proteins. constraint regions couples This different through of chromatin of the silencing reservoir ambient,This constraint couples different regions of chromatin through the shared reservoir of ambient silencing proteins.,5027
"paper, time-evolving problem of link we graphs. the in prediction In consider the","In the paper, we consider the problem of link prediction in time-evolving graphs.",5028
generalized is The using a estimate through forward-backward methods proximal efficiently computed agorithm.,The estimate is computed efficiently using proximal methods through a generalized forward-backward agorithm.,5029
this uniform verification novel in  is A for framework  paper. face and presented,  A novel and uniform framework for face verification is presented in this paper.,5030
and framework results Experimental that verification achieves verification high speed. accuracy the show,Experimental results show that the framework achieves high verification accuracy and verification speed.,5031
are for models. methods approximate used variational Mean-field many posterior in probabilistic inference widely,Mean-field variational methods are widely used for approximate posterior inference in many probabilistic models.,5032
"coordinate-ascent the methods a mean-field compute a typical application, posterior algorithm. approximately In with optimization","In a typical application, mean-field methods approximately compute the posterior with a coordinate-ascent optimization algorithm.",5033
"logistic many regression---are However, nonconjuate. models of Bayesian correlated interest---like the topic model and","However, many models of interest---like the correlated topic model and Bayesian logistic regression---are nonconjuate.",5034
extensions of introduces multidimensional and novel This paper inputs. GPs to tests structured,This paper introduces and tests novel extensions of structured GPs to multidimensional inputs.,5035
and steps include higher natural dimensional non-Gaussian grid The third second methods. and equispaced observations,The natural second and third steps include non-Gaussian observations and higher dimensional equispaced grid methods.,5036
novel introduce address these techniques We necessary both to of directions.,We introduce novel techniques to address both of these necessary directions.,5037
Homologous recombination an evolution in biological is the organisms. important operator of,Homologous recombination is an important operator in the evolution of biological organisms.,5038
applied examples. is to then POMDP algorithm The two,The algorithm is then applied to two POMDP examples.,5039
algorithmic theoretical are intelligence not models realizable. universal Existing practically,Existing theoretical universal algorithmic intelligence models are not practically realizable.,5040
not information. approach as any does samples training Our prior need,Our approach does not need any training samples as prior information.,5041
"Influenza (NA). neuraminidase contains (HA) two and hemagglutinin glycoproteins, variable envelope highly virus","Influenza virus contains two highly variable envelope glycoproteins, hemagglutinin (HA) and neuraminidase (NA).",5042
in discussed future drugs oncolytic engineered will be Those papers.,Those engineered oncolytic drugs will be discussed in future papers.,5043
phenotypes do differentiate egg? of different fertilized from how single cells Furthermore reproducibly a,Furthermore how do cells of different phenotypes differentiate reproducibly from a single fertilized egg?,5044
state reprogramming. epigenetic for The major pathways of network predicts case three this,The epigenetic state network for this case predicts three major pathways of reprogramming.,5045
observations states the pathways. Numerous experimental predicted and support,Numerous experimental observations support the predicted states and pathways.,5046
are mechanism to system Knowledge intelligent. inference make most thing and the desirable (KR) representation,Knowledge representation (KR) and inference mechanism are most desirable thing to make the system intelligent.,5047
This KR paper proposed comparison techniques presents between the hybrid the also one. with various,This paper also presents the comparison between various hybrid KR techniques with the proposed one.,5048
"proteins which the (HSP). response Hyperthermia engages mechanism component (HSR), are of main heat-shock heat-shock","Hyperthermia engages the heat-shock response mechanism (HSR), main component of which are heat-shock proteins (HSP).",5049
"i.e. thermotolerance, HSR triggers the On hand, other","On the other hand, HSR triggers thermotolerance, i.e.",5050
to a susceptibility treated their impairment show stress. hyperthermia heat-induced cells subsequent in an,hyperthermia treated cells show an impairment in their susceptibility to a subsequent heat-induced stress.,5051
"these stochastic address adapt model propose its we extension. our HSR To previous questions, and","To address these questions, we adapt our previous HSR model and propose its stochastic extension.",5052
We the formalise thermotolerance. a notion HSP-induced of,We formalise the notion of a HSP-induced thermotolerance.,5053
"Next, we the the the intensity estimate duration of and thermotolerance.","Next, we estimate the intensity and the duration of the thermotolerance.",5054
"strategy propose inhibition we and modalities. optimal an combining hyperthermia for proteasome Consequently,","Consequently, we propose an optimal strategy for combining hyperthermia and proteasome inhibition modalities.",5055
Recognition developed and techniques have in Various been have practice. theory applied in been Speech,Various Speech Recognition techniques have been developed in theory and have been applied in practice.,5056
music? to function Does fundamental point effect of Mozart cognitive a,Does Mozart effect point to a fundamental cognitive function of music?,5057
"this should in Second, result improved performance.","Second, this should result in improved performance.",5058
one image is super-resolution (SR) Image community. topics the active in of long-standing and processing,Image super-resolution (SR) is one of the long-standing and active topics in image processing community.,5059
a SR sparse Bayesian algorithm via modeling of images. natural We propose image,We propose a Bayesian image SR algorithm via sparse modeling of natural images.,5060
also are Other as available. techniques hybrid systems such neuro-fuzzy,Other hybrid techniques such as neuro-fuzzy systems are also available.,5061
data mining.  are for Rough efficient data in  pre-process sets,  Rough sets are efficient for data pre-process in data mining.,5062
of upper sets. two concepts approximations Lower and are core rough,Lower and upper approximations are two core concepts of rough sets.,5063
"rough a matroids generalized paper between close relationship this sets. and Secondly, establishes","Secondly, this paper establishes a close relationship between matroids and generalized rough sets.",5064
"matroid algebra borrows linear and terminology theory theory. the from graph of extensively Further,","Further, matroid theory borrows extensively from the terminology of linear algebra and graph theory.",5065
data for mining. Rough are efficient data pre-processing in sets,Rough sets are efficient for data pre-processing in data mining.,5066
and to sets lattices. are matroids related rough Both closely,Both rough sets and matroids are closely related to lattices.,5067
"a matroid the through the constructed function lattice. Then, of semimodular is height","Then, a matroid is constructed through the height function of the semimodular lattice.",5068
reasoning to is case-based recipes constraints. system Taaable that cooking user a adapts,Taaable is a case-based reasoning system that adapts cooking recipes to user constraints.,5069
"Within part graph. as recipes it, formalised preparation a is of the","Within it, the preparation part of recipes is formalised as a graph.",5070
"correcting required. a Therefore, and is step validating","Therefore, a validating and correcting step is required.",5071
in helps conversion prepress That department. to from RGB in of CMYK image,That helps in conversion of image from RGB to CMYK in prepress department.,5072
concept based Neighborhood rough in important is an covering sets.,Neighborhood is an important concept in covering based rough sets.,5073
and sufficient Many some necessary paid and have attention presented this conditions. issue to scholars,Many scholars have paid attention to this issue and presented some necessary and sufficient conditions.,5074
"it. from obtain and definition propose uniform a block we the sufficient Second, of condition","Second, we propose the definition of uniform block and obtain a sufficient condition from it.",5075
"repeat Third, of the degree excluded number. definitions we and propose","Third, we propose the definitions of repeat degree and excluded number.",5076
matroids several are matroids among special The relationships and studied. covering,The relationships among several special matroids and covering matroids are studied.,5077
consist regulation. oscillation often Biochemical repressive transcription of with systems negative loops feedback,Biochemical oscillation systems often consist of negative feedback loops with repressive transcription regulation.,5078
on The have been studied system fluctuations al. Gonze by et circadian free-running molecular a,The molecular fluctuations on a free-running circadian system have been studied by Gonze et al.,5079
"rapidly. are for Centromeres chromosome evolve DNA essential their yet segregation, sequences","Centromeres are essential for chromosome segregation, yet their DNA sequences evolve rapidly.",5080
with technologies. current Our compatible methods are sequencing all,Our methods are compatible with all current sequencing technologies.,5081
non-stationary neurophysiological strongly Our connectivity suggest BOLD a functional results that has origin.,Our results strongly suggest that non-stationary BOLD functional connectivity has a neurophysiological origin.,5082
algorithm been recognizing this an for In paper has proposed. speech,In this paper an algorithm for recognizing speech has been proposed.,5083
and high The experimental demonstrate algorithm. of the proposed results efficiency the gain,The experimental results demonstrate the high gain and efficiency of the proposed algorithm.,5084
the double a and a unary. is when p-algebra lattice distributive covering,a distributive lattice and a double p-algebra when the covering is unary.,5085
reproduction. time These different during color each give rendering,These give different color rendering each time during reproduction.,5086
the tissue are the classification on detection. glandular focused studies for methods of Most,Most of the studies are focused on the classification methods for glandular tissue detection.,5087
spectrogram. approaches environmental for sound presents extraction efficient novel This paper magnitude feature using,This paper presents novel approaches for efficient feature extraction using environmental sound magnitude spectrogram.,5088
on based We visual propose domain. the approach,We propose approach based on the visual domain.,5089
condition A and for sufficient this been scholars. issue has some necessary provided by,A necessary and sufficient condition for this issue has been provided by some scholars.,5090
"for we present issue. a sufficient necessary condition Second, and this","Second, we present a necessary and sufficient condition for this issue.",5091
this for We issue a as present condition sufficient and necessary well.,We present a necessary and sufficient condition for this issue as well.,5092
environments. Background: to have many to acclimate themselves evolved Microbiota,Background: Microbiota have evolved to acclimate themselves to many environments.,5093
medications of those is become and Humanity ever many antibiotics. are increasingly medicated,Humanity is become ever increasingly medicated and many of those medications are antibiotics.,5094
mining. data in efficient for pre-processing sets data Rough are,Rough sets are efficient for data pre-processing in data mining.,5095
"relations. successor predecessor we paper, by the and this more neighborhoods reveal In from relationships","In this paper, we reveal more relationships by the predecessor and successor neighborhoods from relations.",5096
"from induce Second, a circuits matroid through relation we a of matroid. the the","Second, we induce a relation from a matroid through the circuits of the matroid.",5097
an induced the prove We always is that equivalence relation. relation,We prove that the induced relation is always an equivalence relation.,5098
"have studies, of however, neglected the coding. temporal aspect Most","Most studies, however, have neglected the aspect of temporal coding.",5099
processes. address theory this Poisson of through we inhomogeneous Here a shortcoming filtering,Here we address this shortcoming through a filtering theory of inhomogeneous Poisson processes.,5100
which introduce order this we to will depend. change will AI In parameters to some,In order to change this we will introduce some parameters to which AI will depend.,5101
"value. we we calculate cannot because is say level this its Nevertheless, exact which cannot","Nevertheless, we cannot say which is this level because we cannot calculate its exact value.",5102
supports repeat interactions. in to that an may experimental expansion increase molecular This evidence lead,This supports experimental evidence that repeat expansion may lead to an increase in molecular interactions.,5103
"decrease Ka/Ks, interactions diversity, by protein-protein estimated appeared increased. the number to Genetic of as","Genetic diversity, estimated by Ka/Ks, appeared to decrease as the number of protein-protein interactions increased.",5104
"with negatively VNTRs. G+C occurrence and increasing correlated CpG Additionally, content of were","Additionally, G+C and CpG content were negatively correlated with increasing occurrence of VNTRs.",5105
pathways new interaction in suggested The inter-species possible models protein maps. also the,The models also suggested new possible pathways in the inter-species protein interaction maps.,5106
"knowledge and machine to discovery been applied data It learning, has mining.","It has been applied to machine learning, knowledge discovery and data mining.",5107
spaces. is vector of a in of theory linear The generalization matroids independence,The theory of matroids is a generalization of linear independence in vector spaces.,5108
has It and in combinatorial design. optimization used algorithm been,It has been used in combinatorial optimization and algorithm design.,5109
"two Moreover, the we constructions. between connections investigate the above","Moreover, we investigate the connections between the above two constructions.",5110
our approach of are benefits The twofold.,The benefits of our approach are twofold.,5111
"it uses prediction, i.e. for reference database First the","First it uses the reference database for prediction, i.e.",5112
setting. biomarkers provide potential clinical a in to,to provide potential biomarkers in a clinical setting.,5113
Second new statistical power increases task. on the it,Second it increases statistical power on the new task.,5114
remote tool. images sensing high classification is  a (HIS) Hyperspectral  technical,  Hyperspectral images (HIS) classification is a high technical remote sensing tool.,5115
bands pixels find of the regions. The to the to classify problematic good is how,The problematic is how to find the good bands to classify the pixels of regions.,5116
we'll accept paper In redundancy. this the useful,In this paper we'll accept the useful redundancy.,5117
"across the (N) size constant the population necessarily is However, genome. not effective","However, the effective population size (N) is not necessarily constant across the genome.",5118
"a is events. significant association between However, there {\theta}s HGT and","However, there is a significant association between {\theta}s and HGT events.",5119
learned the a implies The mapping between data. two dictionary,The learned dictionary implies a mapping between the two data.,5120
learned A thereafter. specific CFA adaptive is the to dictionary input,A specific dictionary adaptive to the input CFA is learned thereafter.,5121
PSNR the of results demonstrate in perception. performance terms visual state-of-the-art and Experimental subjective,Experimental results demonstrate the state-of-the-art performance in terms of PSNR and subjective visual perception.,5122
The is normal distribution in applications. measurement error universal with,The measurement error with normal distribution is universal in applications.,5123
"test better error measurement cost. instrument higher Generally, smaller requires and","Generally, smaller measurement error requires better instrument and higher test cost.",5124
"distribution uniform covering to error with this Recently, issue. error-range-based set proposed was rough investigate","Recently, error-range-based covering rough set with uniform distribution error was proposed to investigate this issue.",5125
The this paper contributions of are major four-fold.,The major contributions of this paper are four-fold.,5126
"data distribution on normal new a we model errors. based First, measurement build","First, we build a new data model based on normal distribution measurement errors.",5127
"data an two-dimension new is error model, range ellipse a space. in With the the","With the new data model, the error range is an ellipse in a two-dimension space.",5128
"problem is covering-based this the reduction attribute set. Third, test-cost-sensitive rough redefined on","Third, the test-cost-sensitive attribute reduction problem is redefined on this covering-based rough set.",5129
"a to heuristic with deal this algorithm proposed Fourth, problem. is","Fourth, a heuristic algorithm is proposed to deal with this problem.",5130
is - of Irvine) tested California ten datasets. UCI algorithm (University on The,The algorithm is tested on ten UCI (University of California - Irvine) datasets.,5131
is toward of applications step study realistic a This learning. cost-sensitive,This study is a step toward realistic applications of cost-sensitive learning.,5132
"we Levene's both populations. Hardy-Weinberg test, Using in reject exact","Using Levene's exact test, we reject Hardy-Weinberg in both populations.",5133
"for not each different were populations correlated. chromosome F of estimates Furthermore, in","Furthermore, estimates of F for each chromosome in different populations were not correlated.",5134
"the leaves the with interacting his/her platform, machine. every in learner traces When","When interacting with the platform, every learner leaves his/her traces in the machine.",5135
The learning can system subject. support any,The system can support any learning subject.,5136
"measure, (ILCSS). Inverse Sub-Sequence complementary Common We a Longest similarity named propose","We propose a complementary similarity measure, named Inverse Longest Common Sub-Sequence (ILCSS).",5137
"compared. Third, four lattice these geometric structures are","Third, these four geometric lattice structures are compared.",5138
yields benchmarks. algorithm competitive results and The both unimodal multimodal on,The algorithm yields competitive results on both unimodal and multimodal benchmarks.,5139
"rules by (crisp) are definitions. of instead they classical defined Moreover, Sugeno fuzzy","Moreover, they are defined by fuzzy Sugeno rules instead of classical (crisp) definitions.",5140
A at rate a geometric ergodic. that geometrically converges chain is,A chain that converges at a geometric rate is geometrically ergodic.,5141
results Our using are two examples. illustrated,Our results are illustrated using two examples.,5142
"In types covering operators. the paper, we approximation present three characterizations this of topological to","In this paper, we present the topological characterizations to three types of covering approximation operators.",5143
"these Finally, topologies are space three compared.","Finally, these three topologies space are compared.",5144
"covering-based provides word, method a In to useful sets. a the study rough topology","In a word, topology provides a useful method to study the covering-based rough sets.",5145
produce color to to main device the picture objective reproduction at output. The is identical,The main objective to color reproduction is to produce the identical picture at device output.,5146
gamut mapping that has For principles . for designed been,For that principles for gamut mapping has been designed .,5147
latency visual generally The decreases contrast of as increases. responses,The latency of visual responses generally decreases as contrast increases.,5148
Most applications in  unimodal. biometric systems real-world deployed  are,  Most biometric systems deployed in real-world applications are unimodal.,5149
The (GT). reference region the called map Ground of image is Truth,The reference image of the region is called Ground Truth map (GT).,5150
"Information to and relevant methods use bands. Some select Mutual threshold, (MI)","Some methods use Mutual Information (MI) and threshold, to select relevant bands.",5151
and paper we'll useful redundancy. between useless In this make difference,In this paper we'll make difference between useful and useless redundancy.,5152
contributes to useful redundancy it probability. if band decreasing error contains A,A band contains useful redundancy if it contributes to decreasing error probability.,5153
"complementary we To redundancy, threshold. control a introduce","To control redundancy, we introduce a complementary threshold.",5154
the from problem. objects PDF Extracting challenging is text images a,Extracting text objects from the PDF images is a challenging problem.,5155
This presents two paper classification. techniques under block-based,This paper presents two techniques under block-based classification.,5156
and consumption tested for The time are both performance for models. segmentation the metrics,The performance metrics for segmentation and time consumption are tested for both the models.,5157
"are is which AI?"" we questions consider here will ""What The","The questions which we will consider here are ""What is AI?""",5158
multi-agent definition in we Here of of present terms the will AI systems.,Here we will present the definition of AI in terms of multi-agent systems.,5159
"answer a old new in form. an but "",",""", but an old answer in a new form.",5160
that answer will help important work question. this More us the is second,More important is that this work will help us answer the second question.,5161
multi-instance single outperforms any the indicate performance approach using Results instance. than that higher verification,Results indicate that the multi-instance verification approach outperforms higher performance than using any single instance.,5162
task. resolution Recognizing degraded faces are challenging blurred yet low common images from and,Recognizing degraded faces from low resolution and blurred images are common yet challenging task.,5163
framework types adaptive proposed. are for blurred classification based on Two of dictionary image,Two types of framework for blurred image classification based on adaptive dictionary are proposed.,5164
from adaptive input The image. estimated the dictionary to Point Function is blurred Spread (PSF),The dictionary is adaptive to the Point Spread Function (PSF) estimated from input blurred image.,5165
results frameworks. effectiveness confirm experiment the the of The proposed,The experiment results confirm the effectiveness of the proposed frameworks.,5166
Optical retinal and powerful noninvasive for method a tomography is coherence (OCT) imaging.,Optical coherence tomography (OCT) is a powerful and noninvasive method for retinal imaging.,5167
not does The approach presented and on edge-based image regional texture. image information relies require,The presented approach does not require edge-based image information and relies on regional image texture.,5168
two from eliminated remaining sections The other calculations. are the,The other two sections are eliminated from the remaining calculations.,5169
information outliers. priori a does about not number require of any It the,It does not require any a priori information about the number of outliers.,5170
available Implementation the package. of is proposed extension in R-INLA the,Implementation of the proposed extension is available in the R-INLA package.,5171
expression. regulates methylation The DNA gene of,The methylation of DNA regulates gene expression.,5172
from passive heterogenity model expected The population mechanisms. the the in describes,The model describes the heterogenity in the population expected from passive mechanisms.,5173
We of the multicellular population. at on also look selection the effects,We also look at the effects of selection on the multicellular population.,5174
passive inaccessible mechanisms. that there states methylation are us tells by of model The,The model tells us that there are states of methylation inaccessible by passive mechanisms.,5175
"times, become one modern vision. has face aspects of recognition key of computer the In","In modern times, face recognition has become one of the key aspects of computer vision.",5176
There with hundreds this. associated factors not if are thousand,There are hundreds if not thousand factors associated with this.,5177
"distance, distance. e.g., pairwise determined are using The Euclidean structures generally by","The structures are generally determined by using pairwise distance, e.g., Euclidean distance.",5178
The derives a term that similar biologic the code. locality from similar observation have inputs,The locality term derives from a biologic observation that the similar inputs have similar code.,5179
relies This the on proposal sequential Monte al. sampler et Del of Carlo Moral,This proposal relies on the sequential Monte Carlo sampler of Del Moral et al.,5180
"the explosion. Really, this to is due useless combinatory algorithm","Really, this algorithm is useless due to the combinatory explosion.",5181
Intellect. our also newly to an a baby according born definition is So,So according to our definition a newly born baby also is an Intellect.,5182
automata offer will consist order final which we The of model and formulas. will first,The model which we will offer will consist of final automata and first order formulas.,5183
binarization. One important the of steps image most of is processing document,One of the most important steps of document image processing is binarization.,5184
"time shifts. since capture interactions it and However, is non-linear limited its cannot reliability","However, its reliability is limited since it cannot capture non-linear interactions and time shifts.",5185
the are regression innovations proposed where A for type problems weighted regression is heavy-tailed. procedure,A weighted regression procedure is proposed for regression type problems where the innovations are heavy-tailed.,5186
"random illustrated walk. on problems, the be first-order The including will procedure autoregressive","The procedure will be illustrated on first-order autoregressive problems, including the random walk.",5187
scheme  We processing  patches. its image based reordering on of propose an,  We propose an image processing scheme based on reordering of its patches.,5188
probability. maximization as reconstruction We problem posed outline of a super-resolution the,We outline the super-resolution reconstruction problem posed as a maximization of probability.,5189
free provided. shown the reproduces software implementation is that A results,A free software implementation that reproduces the results shown is provided.,5190
present Maximum new paper algorithm--a convexified this a graph version we clustering. of Likelihood--for In,In this paper we present a new algorithm--a convexified version of Maximum Likelihood--for graph clustering.,5191
thus challenge. Threshold important selection an remains,Threshold selection thus remains an important challenge.,5192
are and considered. Iterated the Search The NK family metaheuristic Local landscapes of,The NK family of landscapes and the Iterated Local Search metaheuristic are considered.,5193
the power landscapes. of a validates This model LONs of combinatorial as fitness study expressive,This study validates the expressive power of LONs as a model of combinatorial fitness landscapes.,5194
particular cooperation in that actors solutions in the Chain. on focus propose between We Supply,We propose in particular solutions that focus on cooperation between actors in the Supply Chain.,5195
interaction lies nuclear nuclear in strong with acids. problem An proteins with additional their,An additional problem with nuclear proteins lies in their strong interaction with nuclear acids.,5196
in of can supercoiling expression. affect level gene chromosome The the,The level of supercoiling in the chromosome can affect gene expression.,5197
our through extensive demonstrate algorithms We evaluation. the power empirical hybrid of,We demonstrate the power of our hybrid algorithms through extensive empirical evaluation.,5198
to cells that are need from noisy. information often signals Living biochemical extract,Living cells often need to extract information from biochemical signals that are noisy.,5199
can are accurately We that with study linear. signaling cells concentrations measure how networks chemical,We study how accurately cells can measure chemical concentrations with signaling networks that are linear.,5200
then effectiveness for balancing We in load demonstrate practice. its,We then demonstrate its effectiveness for load balancing in practice.,5201
complex video is in to objects track Particle in situations. Filter solution an sequences effective,Particle Filter is an effective solution to track objects in video sequences in complex situations.,5202
on is investigated numerical two Our examples method .,Our method is investigated on two numerical examples .,5203
is If this the often set ordering used can prohibitively Pareto large. be,If the Pareto ordering is used this set can often be prohibitively large.,5204
for present Online algorithm Factored Hybrid Heuristic (FHHOP) a Planning We large POMDPs.,We present a Factored Hybrid Heuristic Online Planning (FHHOP) algorithm for large POMDPs.,5205
"influence hard of the to solve. Unfortunately, described are diagrams problems by most","Unfortunately, most of the problems described by influence diagrams are hard to solve.",5206
the paper In this we influence of discuss approximately solving complexity diagrams.,In this paper we discuss the complexity of approximately solving influence diagrams.,5207
(Most Probable important query Explanation) in an plays inference. MPE probabilistic role The,The MPE (Most Probable Explanation) query plays an important role in probabilistic inference.,5208
for of group is this different maximizing The same. the elements choice,The maximizing choice for different elements of this group is the same.,5209
We ideas number a of effectiveness the on these empirically models. of demonstrate,We demonstrate the effectiveness of these ideas empirically on a number of models.,5210
"Structural Pp. Length, and Weight, Associated Indices,","Length, Weight, and Associated Structural Indices, Pp.",5211
"species. the ), mean each uncertainty the with along in for mean and","), along with the mean and uncertainty in the mean for each species.",5212
"Freshwater of Volume Press, University One, Handbook Fishery Biology, Ames.) Iowa","Handbook of Freshwater Fishery Biology, Volume One, Iowa University Press, Ames.)",5213
well our that methods in perform experiments show Extensive practical settings.,Extensive experiments show that our methods perform well in practical settings.,5214
analytical fitness. We provide and average evolutionary for population expressions rate the,We provide analytical expressions for the evolutionary rate and average population fitness.,5215
be prepress In department CMYK converted to RGB to image has image.,In prepress department RGB image has to be converted to CMYK image.,5216
biomedical and computer applications been many It informatics. has including successfully applied vision to,It has been successfully applied to many applications including computer vision and biomedical informatics.,5217
again given the and participants new repeats. process coupons The are,The new participants are given coupons again and the process repeats.,5218
"used (VH) estimator. estimator the is Volz-Heckathorn as RDS Currently, standard known in the practice","Currently, the standard RDS estimator used in practice is known as the Volz-Heckathorn (VH) estimator.",5219
strong RDS about assumptions network the the process. relies and on social It underlying,It relies on strong assumptions about the underlying social network and the RDS process.,5220
from sequentially tracking filtering hidden aims at process Bayesian an a observed one.,Bayesian filtering aims at tracking sequentially a hidden process from an observed one.,5221
validated algorithms our Finally via are simulations.,Finally our algorithms are validated via simulations.,5222
"point, interest the on points of significant the Based stability extracted. are","Based on the stability of the interest point, significant points are extracted.",5223
briefly. of Extensions distributions discussed mutations other are and models to available,Extensions to other models and distributions of available mutations are discussed briefly.,5224
State sets their functions. and represented accessing modified and by manipulating are characteristic,State sets are represented and modified by accessing and manipulating their characteristic functions.,5225
to is image the partitioning of used as BDD compute smaller the disjunction subimages.,BDD partitioning is used to compute the image as the disjunction of smaller subimages.,5226
"partitioning we a propose paper, BDD novel In this option.","In this paper, we propose a novel BDD partitioning option.",5227
binary of Let the size state the be vector. n,Let n be the size of the binary state vector.,5228
BDD the shared a efforts In are O(n). representation,In a shared BDD representation the efforts are O(n).,5229
Neural network neural in dynamics. role variability central a plays and coding neuronal,Neural variability plays a central role in neural coding and neuronal network dynamics.,5230
wrong lead to can or for applications. the These defects incomplete results,These defects can lead to wrong or incomplete results for the applications.,5231
the detection Debugging consists of and phases of repairing.,Debugging consists of the phases of detection and repairing.,5232
is-a in the missing hierarchy. relations the,the missing relations in the is-a hierarchy.,5233
of taxonomies. Previous with dealt has the work case,Previous work has dealt with the case of taxonomies.,5234
present We algorithms system. discuss and a,We present algorithms and discuss a system.,5235
"system, Before digitalized data textual by immune we approach. our n-grams experimenting the","Before experimenting our immune system, we digitalized textual data by the n-grams approach.",5236
and novelty on of hybridization n-grams The lies immune systems for clustering.,The novelty lies on hybridization of n-grams and immune systems for clustering.,5237
to Fixed-pattern analog calibration by routines. is reduced inherent circuitry noise,Fixed-pattern noise inherent to analog circuitry is reduced by calibration routines.,5238
"measure attention. AOSI particular, In to that develop algorithms we visual activities three critical assess","In particular, we develop algorithms to measure three critical AOSI activities that assess visual attention.",5239
"and This are coding scheme rate dissociable. which allows phase, in frequency a","This allows a coding scheme in which phase, rate and frequency are dissociable.",5240
on is equivalence partitions. based relations And it or,And it is based on equivalence relations or partitions.,5241
directions by of future research. identifying current closes and It,It closes by identifying directions of current and future research.,5242
classification Wavelet The using project occurrence describes and texture Matrix. Transform Co,The project describes texture classification using Wavelet Transform and Co occurrence Matrix.,5243
textures database features different is performed. of of Comparison sample texture with a of,Comparison of features of a sample texture with database of different textures is performed.,5244
"Haar, wavelets. we Symlets Daubechies use In wavelet and the transform","In wavelet transform we use the Haar, Symlets and Daubechies wavelets.",5245
of for super-resolution range report images. method a We,We report a method for super-resolution of range images.,5246
of Our approach the the HR interpretation as image grid. LR samples sparse on leverages,Our approach leverages the interpretation of LR image as sparse samples on the HR grid.,5247
"we demonstrate approach, for results the factors (e.g. proposed large Using super-resolution","Using the proposed approach, we demonstrate super-resolution results for large factors (e.g.",5248
a of been a more for It than research subject has decade.,It has been a subject of research for more than a decade.,5249
"the LASSO, Due analyses have previous to of been non-parametric. nature the","Due to the nature of LASSO, the previous analyses have been non-parametric.",5250
compare useful and makes traditional estimators. to difficult it to ignores LASSO information This,This ignores useful information and makes it difficult to compare LASSO to traditional estimators.,5251
"beamforming. case, maximum compared likelihood this is and conventional For to LASSO","For this case, LASSO is compared to maximum likelihood and conventional beamforming.",5252
"the we provide on the regularization parameter. of Thus, suggestions selection","Thus, we provide suggestions on the selection of the regularization parameter.",5253
"them China. located Mongolia, the in Inner Some are Western of","Some of them are located in the Inner Mongolia, Western China.",5254
moats. These fortresses surrounded large are by,These large fortresses are surrounded by moats.,5255
"new we framework multi-attribute for paper, this propose In graphs estimating principled a from data.","In this paper, we propose a new principled framework for estimating graphs from multi-attribute data.",5256
"efficient provide multi-attribute which an structure. Computationally, utilizes algorithm the we","Computationally, we provide an efficient algorithm which utilizes the multi-attribute structure.",5257
"which consistent guarantee provide graph we conditions Theoretically, recovery. sufficient","Theoretically, we provide sufficient conditions which guarantee consistent graph recovery.",5258
our demonstrate conditions. method under simulation of performance Extensive studies various,Extensive simulation studies demonstrate performance of our method under various conditions.,5259
anatomical and It between investigate spatial the function structure. to allows sensory relation us,It allows us to investigate the spatial relation between sensory function and anatomical structure.,5260
"have when occur, however, high to This neurons thresholds. fails","This fails to occur, however, when neurons have high thresholds.",5261
"population-genetic evolution architectures. the we develop Here genetic model of a simple, for","Here we develop a simple, population-genetic model for the evolution of genetic architectures.",5262
Published All by rights reserved. Elsevier Ltd.,Published by Elsevier Ltd. All rights reserved.,5263
under conditions. A various comparative presented is analysis,A comparative analysis is presented under various conditions.,5264
"the the results. feature affects In choice the domain, widely classification data of","In the feature classification domain, the choice of data affects widely the results.",5265
information. feature the using on we Here selection mutual focus,Here we focus on the feature selection using mutual information.,5266
"inference develop tool LD-based investigating for Here, into a admixture. versatile we comprehensively","Here, we comprehensively develop LD-based inference into a versatile tool for investigating admixture.",5267
and used tests for parametric comparison. Standard two-sample nonparametric are this often,Standard parametric and nonparametric two-sample tests are often used for this comparison.,5268
"this to we attempt an this problem. overcome In article, make","In this article, we make an attempt to overcome this problem.",5269
"applied transformed data, they on are yield Standard tests, when results. better two-sample the","Standard two-sample tests, when they are applied on the transformed data, yield better results.",5270
such field may as issues reduction. this optimization attribute in There are,There are may optimization issues in this field such as attribute reduction.,5271
widely are Matroids matrices from in optimization. used generalized,Matroids generalized from matrices are widely used in optimization.,5272
"Therefore, connect matroids is to with sets. it rough necessary","Therefore, it is necessary to connect matroids with rough sets.",5273
"a provides word, rough sets. new a matrix to In viewpoint study","In a word, matrix provides a new viewpoint to study rough sets.",5274
increases of high thyroid age. The and with very is the generally incidence nodule,The incidence of thyroid nodule is very high and generally increases with the age.,5275
cancer. thyroid Thyroid emergence the may presage of nodule,Thyroid nodule may presage the emergence of thyroid cancer.,5276
thyroid if can early. The nodule be detected completely cured,The thyroid nodule can be completely cured if detected early.,5277
is aspiration of method nodule. needle Fine cytology early a thyroid diagnosis recognized,Fine needle aspiration cytology is a recognized early diagnosis method of thyroid nodule.,5278
segmentation Image based present. at a technique graph theory on hotspot has research become,Image segmentation technique based on graph theory has become a research hotspot at present.,5279
the to leads inaccurate generates always segmentation segmentation. over segmentation in or less It which,It always generates over segmentation or less segmentation which leads to inaccurate in the segmentation.,5280
proved method by The the feasibility the experimental of results. is,The feasibility of the method is proved by the experimental results.,5281
The the related dimensionality to reduction. HSI problematic is,The problematic related to HSI is the dimensionality reduction.,5282
information the bands. to use mutual studies (MI) Many select relevant,Many studies use mutual information (MI) to select the relevant bands.,5283
"Others MI Uncertainty, like imagery normalized in the Symmetric applications. use medical forms, studies","Others studies use the MI normalized forms, like Symmetric Uncertainty, in medical imagery applications.",5284
tool This a strategy. selection and algorithm Filter feature is,This algorithm is feature selection tool and a Filter strategy.,5285
"and to control scheme effectiveness, redundancy. This is an fast","This is an effectiveness, and fast scheme to control redundancy.",5286
"factors during state. miRNAs the repress core differentiation Globally, pluripotent directly","Globally, miRNAs directly repress the pluripotent core factors during differentiation state.",5287
"state the DNA for methylation. methyltransferases, miRNAs Furthermore, pluripotent enzymes predominately at repress DNA core","Furthermore, at pluripotent state miRNAs predominately repress DNA methyltransferases, the core enzymes for DNA methylation.",5288
functions MiRNAs with stem different states. vary their cell,MiRNAs vary their functions with different stem cell states.,5289
"relation formally However, proved. this not has been still","However, this relation has still not been formally proved.",5290
"this paper, between In we both relation such analyse problems.","In this paper, we analyse such relation between both problems.",5291
"we two equivalent cases. are problems the addition, that some in In show","In addition, we show that the two problems are equivalent in some cases.",5292
"a of interactions this result allocations. In hypothesis, resource are species","In this hypothesis, species interactions are a result of resource allocations.",5293
systems have various and trade-offs. mobile operating their We the to compare tried,We have tried to compare the various mobile operating systems and their trade-offs.,5294
one-dimensional a to two-dimensional A image used method generate is back from projections. projection,A back projection method is used to generate a two-dimensional image from one-dimensional projections.,5295
are the target compared These image. to images,These images are compared to the target image.,5296
outperforms two sets. in interactions Our alternatives predicting data method on drug-protein,Our method outperforms alternatives in predicting drug-protein interactions on two data sets.,5297
by This paternity. multiple boosted effect is,This effect is boosted by multiple paternity.,5298
Our by algorithm was state-or-the-art method. the outperformed,Our algorithm was outperformed by the state-or-the-art method.,5299
"related we dataset are So, code. our and releasing","So, we are releasing our dataset and related code.",5300
random A is integer-valued for variables square-integrable generator universal introduced.,A universal generator for integer-valued square-integrable random variables is introduced.,5301
"an the an application As is such of provided. distribution example framework, Poisson-Tweedie a to","As an example of such a framework, an application to the Poisson-Tweedie distribution is provided.",5302
of \emph{a The environment-specific parameters. expressions organism- and no knowledge and/or priori} simple are require,The expressions are simple and require no \emph{a priori} knowledge of organism- and/or environment-specific parameters.,5303
recording of Motion people. capture the objects is of movement process or the,Motion capture is the process of recording the movement of objects or people.,5304
captures face includes or fingers and subtle it When .,When it includes face and fingers or captures subtle .,5305
"two text artificial scene Video text, of kinds and text. include","Video include two kinds of text, scene text and artificial text.",5306
is vigilantly artificial emergence text of consequently directed. The,The emergence of artificial text is consequently vigilantly directed.,5307
few for (KR) knowledge. representing techniques efficiently are knowledge available representation There,There are few knowledge representation (KR) techniques available for efficiently representing knowledge.,5308
"However, increase methods are with better the needed. in complexity,","However, with the increase in complexity, better methods are needed.",5309
came two researchers mechanisms methods. or up hybrid by with combining more Some,Some researchers came up with hybrid mechanisms by combining two or more methods.,5310
"and KR is and needed. of standardized demerits are There of merits method combinations,","There are merits and demerits of combinations, and standardized method of KR is needed.",5311
"KR this presented. and at details were various paper, In of schemes hybrid length explored","In this paper, various hybrid schemes of KR were explored at length and details presented.",5312
"and The are degraded filtering preprocessed denoising underwater homomorphic, anisotropic wavelet applying by sequentially. images","The degraded underwater images are preprocessed by applying homomorphic, wavelet denoising and anisotropic filtering sequentially.",5313
on right lies rectified The image and left common plane. a,The rectified left and right image lies on a common plane.,5314
"we triangulation the of Finally, images technique. estimate using depth","Finally, we estimate the depth of images using triangulation technique.",5315
is processing. Edge important an in field detection image,Edge detection is an important field in image processing.,5316
relevant result classic demonstrate the that results method proposed Experimental method. better the achieve than,Experimental results demonstrate that the proposed method achieve better result than the relevant classic method.,5317
recruits. both clonal sexually producing asexually and species genetically unique Many reproduce and,Many species reproduce both sexually and asexually producing genetically unique and clonal recruits.,5318
"we a the into a invasion for new present habitat. Here, novel model species of","Here, we present a novel model for the invasion of species into a new habitat.",5319
study mechanisms synchrony. explores potential intermittent Our this behind,Our study explores potential mechanisms behind this intermittent synchrony.,5320
of the properties affects also Dopamine neurons. cellular,Dopamine also affects the cellular properties of neurons.,5321
(H.R.) presents value neighbor algorithm high resolution This for the nearest paper (NNV),This paper presents the nearest neighbor value (NNV) algorithm for high resolution (H.R.),5322
method H.R. performances terms in proposed demonstrated of The higher,The proposed method demonstrated higher performances in terms of H.R.,5323
compared when mentioned. interpolation the conventional to algorithms,when compared to the conventional interpolation algorithms mentioned.,5324
of model. This proposed approach network overcome old limitation wavelet was to,This approach was proposed to overcome limitation of old wavelet network model.,5325
handwritten of Arabic character. interested This in is unchangeable practically feature paper the,This paper is practically interested in the unchangeable feature of Arabic handwritten character.,5326
promising the proposed approach results demonstrate Reported performance . of,Reported results demonstrate promising performance of the proposed approach .,5327
tasks in critical most Edge detection of the is analysis. one image automatic,Edge detection is one of the most critical tasks in automatic image analysis.,5328
works There no edge detection which exists all well universal method under conditions.,There exists no universal edge detection method which works well under all conditions.,5329
of the The main robustness its advantages flexibility. method are and proposed its,The main advantages of the proposed method are its robustness and its flexibility.,5330
each block PCA for efficiently done separately. is,PCA is efficiently done for each block separately.,5331
time. is in The and has algorithm execution computation high complex extremely,The algorithm is extremely complex in computation and has high execution time.,5332
this scheme explored of work. The complexity the is time BSP in,The time complexity of the BSP scheme is explored in this work.,5333
"traverse opposite classes solutions, in wheel two which color the find We directions. of","We find two classes of solutions, which traverse the color wheel in opposite directions.",5334
two near classes of cyan and hues start with red. scales The,The two classes of scales start with hues near cyan and red.,5335
"spontaneously been sets likelihood autocatalytic has long However, debated. the emerging of","However, the likelihood of autocatalytic sets emerging spontaneously has long been debated.",5336
"two progress lines. made has different Recently, been along","Recently, progress has been made along two different lines.",5337
a struggle One winning coevolutionary evolve is rapidly. strategy for to,One strategy for winning a coevolutionary struggle is to evolve rapidly.,5338
"phylogenetically in control that we the mammalian cite logic Furthermore, evidence immune lineages. conserved is","Furthermore, we cite evidence that the immune control logic is phylogenetically conserved in mammalian lineages.",5339
able were natural we to identify Furthermore isolates. pathogens fly candidate in,Furthermore we were able to identify candidate pathogens in natural fly isolates.,5340
within theory how embed the to the show we framework. Darwinian Here neutral,Here we show how to embed the neutral theory within the Darwinian framework.,5341
"the generation varying maximal study Poisson-disk radii. In this with we sets paper, of","In this paper, we study the generation of maximal Poisson-disk sets with varying radii.",5342
"present disk in gaps of we such First, analysis sets. geometric a","First, we present a geometric analysis of gaps in such disk sets.",5343
the build We and the diagram. power of on triangulation the concepts regular,We build on the concepts of the regular triangulation and the power diagram.,5344
We with `spike-and-slab' inference sparse a based model and learning study coding on prior.,We study inference and learning based on a sparse coding model with `spike-and-slab' prior.,5345
constructs data via among a random diffusion method process coordinates the walk. a The,The method constructs a diffusion process among the data coordinates via a random walk.,5346
The and algorithms. methodology produce that results successfully excellent algorithms compete state-of-the-art current proposed with,The proposed methodology and algorithms produce excellent results that successfully compete with current state-of-the-art algorithms.,5347
"both LDS the signals, and long-term antero-posterior computed. From were short-term medio-lateral the and","From the antero-posterior and the medio-lateral signals, both long-term and short-term LDS were computed.",5348
modified of dynamics of pressure the the RAC fluctuation trajectory. center the substantially,RAC substantially modified the fluctuation dynamics of the center of the pressure trajectory.,5349
that fractal observed responded to (scaling We LDS also both similarly RAC. dynamics exponents) and,We also observed that both LDS and fractal dynamics (scaling exponents) responded similarly to RAC.,5350
"observed with we graphs. reality the of many applications, errorfully are In faced","In many applications, we are faced with the reality of errorfully observed graphs.",5351
"In consider graph we $G paper, = (V,E)$. this a","In this paper, we consider a graph $G = (V,E)$.",5352
seen of and defectors mixtures cooperators Stable often nature. in are,Stable mixtures of cooperators and defectors are often seen in nature.,5353
"we the their selection, of assumption Under stability. weak analyze equilibria the and","Under the assumption of weak selection, we analyze the equilibria and their stability.",5354
rate depth control thus and has consider data. Encoding texture both to,Encoding rate control has thus to consider both texture and depth data.,5355
We a images. rate-distortion texture simplified depth based adopt and on model a framework of,We adopt a rate-distortion framework based on a simplified model of depth and texture images.,5356
and images. features main Our preserves depth of the texture model,Our model preserves the main features of depth and texture images.,5357
theoretical our results and rate efficiency strategy. of allocation Experiments our the confirm,Experiments confirm our theoretical results and the efficiency of our rate allocation strategy.,5358
"recovering article, layers of transparent the address In latent this we issue from superimposition images.","In this article, we address the issue of recovering latent transparent layers from superimposition images.",5359
"extracted assume and Here, we estimated we the layers. gradients transformations latent have of","Here, we assume we have the estimated transformations and extracted gradients of latent layers.",5360
using signals simple evolution theory of histories. Sexual preferences sexual and life models selection,Sexual selection theory models evolution of sexual signals and preferences using simple life histories.,5361
several these for pose problems evolutionary stages Early traits.,Early evolutionary stages pose several problems for these traits.,5362
"grow Few large will rarely females. males encounter will choosy they traits, and","Few males will grow large traits, and they will rarely encounter choosy females.",5363
evolutionary The unclear. of age-dependent remain therefore traits origins,The evolutionary origins of age-dependent traits therefore remain unclear.,5364
sizes weaker evolve and age-independent than at traits Age-dependent smaller traits. selection under,Age-dependent traits evolve under weaker selection and at smaller sizes than age-independent traits.,5365
possibility this this paper. investigate in We,We investigate this possibility in this paper.,5366
cause several illness diseases suffering individuals. Both in and million,Both diseases cause suffering and illness in several million individuals.,5367
"therapeutics and vaccines market. efficacy, potential are toxic Current and from far the lack are","Current therapeutics are toxic and lack efficacy, and potential vaccines are far from the market.",5368
(i-v). based papers five is on thesis This,This thesis is based on five papers (i-v).,5369
cruzi Paper repertoire iv the small of Trypanosoma non-coding epimastigotes. in RNAs describes,Paper iv describes the repertoire of small non-coding RNAs in Trypanosoma cruzi epimastigotes.,5370
a Language Swedish Sign signs. isolated of method for recognition We present,We present a method for recognition of isolated Swedish Sign Language signs.,5371
"is but language deaf children group primarily children, The with not target disorders.","The target group is not primarily deaf children, but children with language disorders.",5372
"allows be removed easily. complex Firstly, to backgrounds it","Firstly, it allows complex backgrounds to be removed easily.",5373
segment depth the We face and on hands color based skin information. and,We segment the hands and face based on skin color and depth information.,5374
"the with hand Secondly, of resolution over helps it face occlusion.","Secondly, it helps with the resolution of hand over face occlusion.",5375
can estimated motion This depth be is observable. the if,This motion can be estimated if the depth is observable.,5376
effective and data tree machine classification Decision is mining in an approach learning.,Decision tree is an effective classification approach in data mining and machine learning.,5377
"be decision should and inducing costs while In trees. test considered costs applications, misclassification","In applications, test costs and misclassification costs should be considered while inducing decision trees.",5378
symbolic only deal with algorithms data. These,These algorithms deal with only symbolic data.,5379
algorithm. our for are two issues There major,There are two major issues for our algorithm.,5380
"heuristic develop we as ratio the First, cost information gain test weighted information. the","First, we develop the test cost weighted information gain ratio as the heuristic information.",5381
"reduced. way, the cost this is In total","In this way, the total cost is reduced.",5382
consists on identifying target similar Retrieval the source are phase cases case. to that,Retrieval phase consists on identifying source cases that are similar to the target case.,5383
results results The obtained efficient typical those measures. applying more than obtained are by,The obtained results are more efficient results than those obtained by applying typical measures.,5384
"their cognitive than connectivity markedly ASD across typicals Furthermore, more alter states. subject","Furthermore, ASD subject alter more markedly than typicals their connectivity across cognitive states.",5385
"systems modular, different and environments. modularity are and in Biological this over evolves time","Biological systems are modular, and this modularity evolves over time and in different environments.",5386
show changing how steady-state be the fitness a in We randomly computed. environment can,We show how the steady-state fitness in a randomly changing environment can be computed.,5387
steady of action evolved state. find at a also least principle the We for modularity,We also find a principle of least action for the evolved modularity at steady state.,5388
the resources mechanics constrained and is energy within acquisition of the Foraging assimilation. and by,Foraging is constrained by the energy within resources and the mechanics of acquisition and assimilation.,5389
Determining manual process is human the Cobb errors. to subject angle through,Determining the Cobb angle through manual process is subject to human errors.,5390
by the shadowing autobiography. performs events with current from The reasoning system events,The system performs reasoning by shadowing current events with events from the autobiography.,5391
in image of The elliptical considered. finding be shapes an will problem,The problem of finding elliptical shapes in an image will be considered.,5392
the uses We clustering. cross-entropy discuss solution which,We discuss the solution which uses cross-entropy clustering.,5393
"for well it dimensions. works Moreover, higher search of ellipsoids in","Moreover, it works well for search of ellipsoids in higher dimensions.",5394
several for establish important T-splines and properties which analysis. fundamental design are We analysis-suitable of,We establish several fundamental properties of analysis-suitable T-splines which are important for design and analysis.,5395
"Species and, networks, predator-prey community potentially, dynamics. shape impacting interactions ecological structure","Species interactions shape predator-prey networks, impacting community structure and, potentially, ecological dynamics.",5396
autobiography. the shadowing by Reasoning performed from current events is with events,Reasoning is performed by shadowing current events with events from the autobiography.,5397
then into shadows The extrapolated are shadows (HLSs). headless,The shadows are then extrapolated into headless shadows (HLSs).,5398
The free Cu binding its valence modulates strongly energy. of,The valence of Cu strongly modulates its binding free energy.,5399
"controlled illusion a presented However, is perceived. vibration acoustical with an together if","However, if presented together with a controlled acoustical vibration an illusion is perceived.",5400
and The is basic nature illusion given. theoretical the of model is described a,The nature of the illusion is described and a basic theoretical model is given.,5401
natural-looking with Real-time in floor plans is generation dynamic vital games environments. of,Real-time generation of natural-looking floor plans is vital in games with dynamic environments.,5402
algorithm This plans presents to in floor generate paper house real-time. suburban an,This paper presents an algorithm to generate suburban house floor plans in real-time.,5403
"plans floor produce to redesigned the to However, corridor placement is real houses. similar","However, the corridor placement is redesigned to produce floor plans similar to real houses.",5404
by architect. very plans designed results an similar floor The the ones to show,The results show very similar floor plans to the ones designed by an architect.,5405
treatment to improved (ART) has recent years. combination over greatly antiretroviral Access,Access to combination antiretroviral treatment (ART) has improved greatly over recent years.,5406
ART patient generally the in well healthy. virus works and suppressed keeping the,ART generally works well in keeping the virus suppressed and the patient healthy.,5407
"law.  explanation for a present this theoretical  In paper, we possible benford's","  In this paper, we present a possible theoretical explanation for benford's law.",5408
"ideas. We a simple recursive intuitive develop between relation using probabilities, the","We develop a recursive relation between the probabilities, using simple intuitive ideas.",5409
"usually underwater lighting, diminished The images blur suffers from and contrast, low non-uniform colors.","The underwater images usually suffers from non-uniform lighting, low contrast, blur and diminished colors.",5410
images. underwater are applied sequentially on filters These degraded,These filters are applied sequentially on degraded underwater images.,5411
"the based uses anisotropic image, preprocessing image For the algorithms smoothing filter. the","For smoothing the image, the image based preprocessing algorithms uses the anisotropic filter.",5412
stages. is two The experimentation in out carried,The experimentation is carried out in two stages.,5413
"are and denoising. wavelet function Similarly, optimal shrinkage wavelet estimated filter bank optimal for","Similarly, optimal filter bank and optimal wavelet shrinkage function are estimated for wavelet denoising.",5414
"detection edge evaluated on Further, based the results. qualitatively results are","Further, the results are qualitatively evaluated based on edge detection results.",5415
in gestures or Many used applications. vision fingertips their based have track to applications manipulate,Many vision based applications have used fingertips to track or manipulate gestures in their applications.,5416
a novel for time fingertip has algorithm described Here efficient been detection.,Here a novel time efficient algorithm has been described for fingertip detection.,5417
of genome-scale research with and metabolic wide applications been models. has done range A,A wide range of applications and research has been done with genome-scale metabolic models.,5418
a novel and acquire such to spectra. paper This method reconstruct presents compressively,This paper presents a novel method to compressively acquire and reconstruct such spectra.,5419
of is efficiency demonstrated few on this test The of the experimentally objects. approach observation,The efficiency of this approach is demonstrated experimentally on the observation of few test objects.,5420
uploaded camera The user database containing motion consists realistic of videos background. and cluttered,The database consists of realistic user uploaded videos containing camera motion and cluttered background.,5421
auxiliary We derive estimate of our variable Rao-Blackwellised distribution. estimator from a a marginal,We derive our estimator from a Rao-Blackwellised estimate of a marginal auxiliary variable distribution.,5422
our applications. We method illustrate with two,We illustrate our method with two applications.,5423
SplitSampling. in package: The methods are R developed the available here,The methods developed here are available in the R package: SplitSampling.,5424
explore area the approximations Bayesian higher-order tail We simulation. of use for,We explore the use of higher-order tail area approximations for Bayesian simulation.,5425
roles play important They clinical other in psychiatric and depression disorders.,They play important roles in clinical depression and other psychiatric disorders.,5426
are sensors. hyperspectral spatial images low due presented pixels of in to Mixed hyperspectral resolution,Mixed pixels are presented in hyperspectral images due to low spatial resolution of hyperspectral sensors.,5427
mixed and decomposes endmembers spectra pixels unmixing into fractions. abundance spectra Spectral,Spectral unmixing decomposes mixed pixels spectra into endmembers spectra and abundance fractions.,5428
NMF compared on Unmixing SAD to results AAD traditional measures. method based are and,Unmixing results are compared to traditional NMF method based on SAD and AAD measures.,5429
used purposes. this for efficiently hyperspectral can that be Results method unmixing demonstrate,Results demonstrate that this method can be used efficiently for hyperspectral unmixing purposes.,5430
"respected by or vehicles, must human- be driven. either all regulation Traffic computer-","Traffic regulation must be respected by all vehicles, either human- or computer- driven.",5431
and Genomic biomedical undergoing imaging exponential data growth. are data,Genomic data and biomedical imaging data are undergoing exponential growth.,5432
"features management, web-based following: of the construction annotation, user and workbenches, community-based BioDIG rapid web-services.","BioDIG features the following: rapid construction of web-based workbenches, community-based annotation, user management, and web-services.",5433
"we In paper, development this spatial of tuning. study cortical the frequency","In this paper, we study the development of cortical spatial frequency tuning.",5434
cortical various responses. feedback analyze strength) relationship between model (including and parameters the We,We analyze the relationship between various model parameters (including cortical feedback strength) and responses.,5435
been steadily image the years. in algorithms have improving deblurring past Blind,Blind image deblurring algorithms have been improving steadily in the past years.,5436
examples. is experiments of various method of kinds on challenging The demonstrated our by effectiveness,The effectiveness of our method is demonstrated by experiments on various kinds of challenging examples.,5437
paper complexity In conjecture. we the this prove probabilistic continuous,In this paper we prove the probabilistic continuous complexity conjecture.,5438
defences The by affected population host parasite virulence is evolution structure. and of,The evolution of parasite virulence and host defences is affected by population structure.,5439
rear species of to workers that another exploit are social ants offspring. parasites Slavemaking their,Slavemaking ants are social parasites that exploit workers of another species to rear their offspring.,5440
result polydomy be This structure small-scale to population (e.g. seems a of,This small-scale population structure seems to be a result of polydomy (e.g.,5441
regulated ribosomal strategy of is Polarity in proteins. evolutionary,Polarity is regulated in evolutionary strategy of ribosomal proteins.,5442
them existing thinking synthesizes applies technological and problems. and in complex solving critical It knowledge,It synthesizes critical thinking and existing knowledge and applies them in solving complex technological problems.,5443
shedding at this light paper relationship. into aims This some,This paper aims at shedding some light into this relationship.,5444
solution variance of of mean HJE gives the fitness. the state The steady,The steady state solution of HJE gives the variance of the mean fitness.,5445
Motivation: nature. fractal in is   roughness Protein surface,  Motivation: Protein surface roughness is fractal in nature.,5446
the Max-tree compute most connected to allows simple advanced in a operators way.,Max-tree allows to compute the most advanced connected operators in a simple way.,5447
far results The sequential parallel that the to algorithm. show the superior algorithm is,The results show that the parallel algorithm is far superior to the sequential algorithm.,5448
"efficiency. methods robustness deconvolution quality, their are and restoration Candidate selected for","Candidate deconvolution methods are selected for their restoration quality, robustness and efficiency.",5449
algorithmic We optimisations scenarios. specific for discuss,We discuss algorithmic optimisations for specific scenarios.,5450
"general are times more space-invariant obtained. blur favourable For computation settings, still","For more general space-invariant blur settings, still favourable computation times are obtained.",5451
small-variance directly the applying posterior models. nonparametric Bayesian asymptotics in consider instead We to,We instead consider applying small-variance asymptotics directly to the posterior in Bayesian nonparametric models.,5452
"algorithms, all several other are scalable implement. demonstrate of which to simple We and","We demonstrate several other algorithms, all of which are scalable and simple to implement.",5453
the Empirical demonstrate framework. new benefits the results of,Empirical results demonstrate the benefits of the new framework.,5454
However is empirical the generally literature. their well not documented in performance,However their empirical performance is generally not well documented in the literature.,5455
explicit and global regulation Eigen's system is quasispecies space with considered.,Eigen's quasispecies system with explicit space and global regulation is considered.,5456
detectors. We algorithms visualize object spaces used feature to by introduce,We introduce algorithms to visualize feature spaces used by object detectors.,5457
"gain detection visualizing we of systems. spaces, feature intuitive more a can understanding our By","By visualizing feature spaces, we can gain a more intuitive understanding of our detection systems.",5458
"Moran Coalescent are derived mutation, models large admitting population from including numbers. processes, type offspring","Coalescent processes, including mutation, are derived from Moran type population models admitting large offspring numbers.",5459
and free of straightforward. the energy distance derivation are computation The also,The derivation and computation of the free energy distance are also straightforward.,5460
the of on node the applicability focuses comparison clustering and This classification. in distances graph,This comparison focuses on the applicability of the distances in graph node clustering and classification.,5461
"tasks. that the distances perform the in general, The comparison, in well parametrized shows","The comparison, in general, shows that the parametrized distances perform well in the tasks.",5462
 configuration problems their  combinatorial In problems. are nature (optimization),  In their nature configuration problems are combinatorial (optimization) problems.,5463
been in off-line used for successfully algorithms decision Symbolic representations planning processes. Markov have,Symbolic representations have been used successfully in off-line planning algorithms for Markov decision processes.,5464
on-line they show can that also improve of the We planners. performance,We show that they can also improve the performance of on-line planners.,5465
"an sRTDP), RTDP. (or introduce Programming of Dynamic We extension Real-Time Symbolic","We introduce Symbolic Real-Time Dynamic Programming (or sRTDP), an extension of RTDP.",5466
The of {sl compilation}. termed usually obtaining a is process JT,The process of obtaining a JT is usually termed {sl compilation}.,5467
"k-contingency In (OKP). algorithm planning paper, we this an present optimal for any-time","In this paper, we present an any-time algorithm for optimal k-contingency planning (OKP).",5468
present cases. applying of experimental some this to algorithm We simple test results,We present experimental results of applying this algorithm to some simple test cases.,5469
"conditional probabilities, of big representations. but compact have tables need the just cannot We","We cannot just have big tables of the conditional probabilities, but need compact representations.",5470
"standard encoded of In are weighted possibilistic means base. a information logic, prioritized knowledge by","In a standard possibilistic logic, prioritized information are encoded by means of weighted knowledge base.",5471
logic dealing This proposes for of partially information. paper an with extension ordered possibilistic,This paper proposes an extension of possibilistic logic for dealing with partially ordered information.,5472
couterparts dealing ordered have information. natural with when partially,have natural couterparts when dealing with partially ordered information.,5473
a The illustrated the example. of small algorithm behavior on is,The behavior of the algorithm is illustrated on a small example.,5474
"an of justification of axiomatic an Finally, criterion by OWA provided. robustness is refinement the","Finally, an axiomatic justification of the refinement of robustness by an OWA criterion is provided.",5475
one proceeding to the analogous at meiosis. the crossing-over of presence,the presence of crossing-over analogous to the one proceeding at meiosis.,5476
to The a is computational provide framework. quantitative basic common and idea,The basic idea is to provide a common computational and quantitative framework.,5477
"learning the algorithm. AIS-BN the avoids In addition, it stage of costly","In addition, it avoids the costly learning stage of the AIS-BN algorithm.",5478
(SD) on approach two proposed works The phases. in Deviation Standard based,The proposed approach based on Standard Deviation (SD) works in two phases.,5479
The keeping filters intact. pixels phase those others only second,The second phase filters only those pixels keeping others intact.,5480
deviation four minimum is filtering pixels. on the of based directional standard scheme The,The filtering scheme is based on minimum standard deviation of the four directional pixels.,5481
soils. assemble corn prairie two from native matched large soil metagenomes We Iowa and then,We then assemble two large soil metagenomes from matched Iowa corn and native prairie soils.,5482
models measure mean for (MSE). The performance three the the is standard error,The performance measure for the three models is the mean standard error (MSE).,5483
networks. Several proposed thus to describe languages researchers such have first-order,Several researchers have thus proposed first-order languages to describe such networks.,5484
"defines and and initial it implementation and This paper describes an introduces CLP(cal{BN}), experiments.","This paper introduces and defines CLP(cal{BN}), and it describes an implementation and initial experiments.",5485
"size, triangulation clique different of number etc. a measures (max Using quality state-space,","Using a number of different triangulation quality measures (max clique size, state-space, etc.",5486
"that triangulation improve. quality our find algorithm can we with dramatically the boundary ),","), we find that with our boundary algorithm the triangulation quality can dramatically improve.",5487
functions (pcb). decision making studies belief consonant for Walley's paper partially This,This paper studies decision making for Walley's partially consonant belief functions (pcb).,5488
"are of In partitioned. foci the set pcb, a","In a pcb, the set of foci are partitioned.",5489
"nested. Within are partition, each the foci","Within each partition, the foci are nested.",5490
includes cases. class The functions pcb extreme probability functions and possibility as,The pcb class includes probability functions and possibility functions as extreme cases.,5491
"with an a axiomatic we belief Unlike decision approach. for functions, proposals theory earlier employ","Unlike earlier proposals for a decision theory with belief functions, we employ an axiomatic approach.",5492
representation a for this theorem prove We relation.,We prove a representation theorem for this relation.,5493
"Kovac, compared was This with models. newly and proposed Saleh Swift model","This newly proposed model was compared with Kovac, Saleh and Swift models.",5494
for An results. the developed also is our Mathematical process applcation illustrating Modelling,An applcation is also developed for the Mathematical Modelling process illustrating our results.,5495
"Moreover, still methods parameterized statistical the assumption. combinatorial require independence","Moreover, parameterized combinatorial methods still require the statistical independence assumption.",5496
measures dependability We within be that can show quantitative will and classical PHA. characterized qualitative,We will show that classical qualitative and quantitative dependability measures can be characterized within PHA.,5497
used paradigm is search a can solve algorithmic that to be problems. powerful many Backtracking,Backtracking search is a powerful algorithmic paradigm that can be used to solve many problems.,5498
units. lengths in a easily to express method branch new interpretable also uses drift MixMapper,MixMapper also uses a new method to express branch lengths in easily interpretable drift units.,5499
RC We results in about this show main two paper.,We show two main results about RC in this paper.,5500
solution with non-Markovian This for number examines (NMRDPs). methods processes a rewards of paper decision,This paper examines a number of solution methods for decision processes with non-Markovian rewards (NMRDPs).,5501
"As different temporal translations. logics a result, different and adopt they","As a result, they adopt different temporal logics and different translations.",5502
step the first filling this paper This gap. is towards,This paper is the first step towards filling this gap.,5503
by developed Systems Solving the emerged from General Problem Klir. FIR G.,FIR emerged from the General Systems Problem Solving developed by G. Klir.,5504
is on a data rather driven knowledge. on structural It than methodology based systems behavior,It is a data driven methodology based on systems behavior rather than on structural knowledge.,5505
FIR available on pattern is based from reasoning rules synthesized the data.,FIR reasoning is based on pattern rules synthesized from the available data.,5506
rule rules as possible. Sugeno base pattern much The preserves as knowledge,The Sugeno rule base preserves pattern rules knowledge as much as possible.,5507
process increased. is this In information considerably is some lost robustness but,In this process some information is lost but robustness is considerably increased.,5508
present learning mean. for We framework novel counter-harmonic using operators a morphological,We present a novel framework for learning morphological operators using counter-harmonic mean.,5509
networks. and combines neural from morphology convolutional concepts It,It combines concepts from morphology and convolutional neural networks.,5510
settings. large well and to online scales datasets It,It scales well to large datasets and online settings.,5511
probabilistic relationships its probabilistic network signs. of qualitative between variables by models the means WA,WA qualitative probabilistic network models the probabilistic relationships between its variables by means of signs.,5512
Non-monotonic ambiguous an sign. have associated influences,Non-monotonic influences have associated an ambiguous sign.,5513
upon signs These results to lead uninformative inference. ambiguous typically,These ambiguous signs typically lead to uninformative results upon inference.,5514
"situational this concept we capture effect, the of sign. introduce To","To capture this effect, we introduce the concept of situational sign.",5515
"Pearl's precisely, approach More with we a independent of structural-model choice Poole's combination present logic.","More precisely, we present a combination of Pearl's structural-model approach with Poole's independent choice logic.",5516
to first-order capabilities mapping actions explicit approach. structural-model This the and modeling adds also,This mapping also adds first-order modeling capabilities and explicit actions to the structural-model approach.,5517
recent the microbial We when a re-analyzing data observe experiment. from patterns evolution same,We observe the same patterns when re-analyzing data from a recent microbial evolution experiment.,5518
"is histogram words generated. each visual of For subwindow, a","For each subwindow, a histogram of visual words is generated.",5519
Markov we hidden field. Gaussian mixture to Then it model-based random generalize,Then we generalize it to Gaussian mixture model-based hidden Markov random field.,5520
supporting and in Scotland have applications many our We and theory. UK Sweden done,We have done many applications in UK and Scotland and Sweden supporting our theory.,5521
For both only estimators and data. from provides death methodology cases both population our,For both cases our methodology provides both estimators from only death and population data.,5522
advantages The of our method straightforward. are,The advantages of our method are straightforward.,5523
the data not do make need We survey to calculations.,We do not need survey data to make the calculations.,5524
survey and test based improve methodologies. should The the estimates resulting to used be existing,The resulting estimates should be used to test and improve the existing survey based methodologies.,5525
"paper, shape current-day this of we identify In techniques. some matching of limitations the","In this paper, we identify some of the limitations of current-day shape matching techniques.",5526
any shape be that easily can method other algorithm. our augmented matching with We show,We show that our method can easily be augmented with any other shape matching algorithm.,5527
structural incorporation will Greater information of data into accuracy learning. prior on require,Greater accuracy will require incorporation of prior structural information on data into learning.,5528
of set the graph) indices. structure on,graph) structure on the set of indices.,5529
"letter Breen a In al. et recent Nature, to","In a recent letter to Nature, Breen et al.",5530
"et epistasis, ratio al. when absence calculating the dN/dS Breen However, in expected of the","However, when calculating the expected dN/dS ratio in the absence of epistasis, Breen et al.",5531
"genes and all in nuclear the for Furthermore, et al. Breen chloroplast","Furthermore, for all nuclear and chloroplast genes in the Breen et al.",5532
a time-consuming Yet procedure. this may become,Yet this may become a time-consuming procedure.,5533
real results validates paper proposed demonstrates the The data. and framework simulated on data on,The paper validates the proposed framework on simulated data and demonstrates results on real data.,5534
"Traditionally, this photographs is pattern of a aligned from contains elements. measured which distortion flat","Traditionally, this distortion is measured from photographs of a flat pattern which contains aligned elements.",5535
measurable attainable precisions. the limits This fact,This fact limits the attainable measurable precisions.,5536
"is the natural once, at thing and, process investigate. complex we This a","This is a natural process and, at once, the complex thing we investigate.",5537
of article. of Developing this destination such the is a model,Developing of such a model is the destination of this article.,5538
network. a This of corresponds anatomical AL complexity with high their,This corresponds with a high anatomical complexity of their AL network.,5539
The features. all significantly clusters two five differed in,The two clusters differed significantly in all five features.,5540
in and role We problem symmetry concept investigate its the solving. of,We investigate the concept of symmetry and its role in problem solving.,5541
of the symmetry Finally problem improving in solving. concept this attempts exploit to paper,Finally this paper attempts to exploit the concept of symmetry in improving problem solving.,5542
real synthetic on validate Simulations the data provided approach. proposed and to are,Simulations on synthetic and real data are provided to validate the proposed approach.,5543
up scale reinforcement has large to learning networks. that complex tasks require yet to Neuroevolution,Neuroevolution has yet to scale up to complex reinforcement learning tasks that require large networks.,5544
video) imply directly. very search if a high space dimensional encoded raw,raw video) imply a very high dimensional search space if encoded directly.,5545
contain Because network matrices exist (i.e. there solutions whose regularity often weight,Because there often exist network solutions whose weight matrices contain regularity (i.e.,5546
sRBM and in RBM log- models Both likelihood. outperform models Ising,Both RBM and sRBM models outperform Ising models in log- likelihood.,5547
renaissance. of a We are the heart in digital,We are in the heart of a digital renaissance.,5548
"we this computation. of of work, the In mathematical models automation, some review motivate use","In this work, we motivate the use of automation, review some mathematical models of computation.",5549
algorithm images. wide variety demonstrated using is recursive a The of this of statistical effectiveness,The effectiveness of this recursive statistical algorithm is demonstrated using a wide variety of images.,5550
always been Handwritten recognition has character a challenging task.,Handwritten character recognition has always been a challenging task.,5551
"the rate. In implemented to fusion classifier method is paper, this a improve recognition","In this paper, a classifier fusion method is implemented to improve the recognition rate.",5552
"Linear considered (LC). (KNN) the For have we fusion, classifier and neighbour classifier K-nearest","For the classifier fusion, we have considered K-nearest neighbour (KNN) and Linear classifier (LC).",5553
with streaming require An increasing applications of reasoning real-time number under uncertainty input.,An increasing number of applications require real-time reasoning under uncertainty with streaming input.,5554
framework formalism such provides a powerful applications. (dynamic) temporal The for Bayes representational net,The temporal (dynamic) Bayes net formalism provides a powerful representational framework for such applications.,5555
"existing in algorithms mind. addition, not In developed real-time were processing with","In addition, existing algorithms were not developed with real-time processing in mind.",5556
not time do each expressions of that at these The pre-computed. change parts step are,The parts of these expressions that do not change at each time step are pre-computed.,5557
"tractability are knowledge, the these approach. the for explicit structural-model results To first our","To our knowledge, these are the first explicit tractability results for the structural-model approach.",5558
the using state over the inference how be probabilistic We system show tracked can model.,We show how the system state can be tracked using probabilistic inference over the model.,5559
result well Theorem. the Hammersley-Clifford generalizes known This,This result generalizes the well known Hammersley-Clifford Theorem.,5560
Credal independence here as relations interpreted variables. encoding strong are networks among,Credal networks are here interpreted as encoding strong independence relations among variables.,5561
networks of based We sets separately first of theory present specified probabilities. credal a on,We first present a theory of credal networks based on separately specified sets of probabilities.,5562
inference also polytrees We that is setting. this with NP-hard show in,We also show that inference with polytrees is NP-hard in this setting.,5563
"world real these to fail For hold. problems, assumptions many","For many real world problems, these assumptions fail to hold.",5564
a on relies propose two-step The system Answer-Set we and Programming approach. follows,The system we propose relies on Answer-Set Programming and follows a two-step approach.,5565
future express to temporal linear representation extends Our rewards. logic (FLTL),Our representation extends future linear temporal logic (FLTL) to express rewards.,5566
translation Our the has method. the solution model-checking in of effect embedding,Our translation has the effect of embedding model-checking in the solution method.,5567
of scenarios lead theory is and scenario intended to to formal analysis. work This a,This work is intended to lead to a formal theory of scenarios and scenario analysis.,5568
decision A with is together functions step-strategy for actions. a strategy selection,A strategy is a step-strategy together with selection functions for decision actions.,5569
an for instantiation. any incorporating DAG a step-strategy introduce concept of the We GS-DAG: optimal,We introduce the concept of GS-DAG: a DAG incorporating an optimal step-strategy for any instantiation.,5570
address propagation of belief loopy in the question convergence We the algorithm. (LBP),We address the question of convergence in the loopy belief propagation (LBP) algorithm.,5571
MAP. of complexity investigates the This paper,This paper investigates the complexity of MAP.,5572
show complete for We that is NP. MAP,We show that MAP is complete for NP.,5573
for also We provide based negative algorithms. elimination complexity results,We also provide negative complexity results for elimination based algorithms.,5574
"MAP out when even turns Pr are It and hard MPE, that remains easy.","It turns out that MAP remains hard even when MPE, and Pr are easy.",5575
"investigate best results, guaranteed with there effort algorithm approximations. approximation is no we Because","Because there is no approximation algorithm with guaranteed results, we investigate best effort approximations.",5576
MAP approximation framework. a generic introduce We,We introduce a generic MAP approximation framework.,5577
an dependence. propose efficient Bayesian in for We method functional network models inference with,We propose an efficient method for Bayesian network inference in models with functional dependence.,5578
"IN TO example, FOR propagation it helps avoid large tree junction cliques.","FOR example, IN junction tree propagation it helps TO avoid large cliques.",5579
three an stratified fold iris SIFT efficient This matching for paper proposes recognition.,This paper proposes an efficient three fold stratified SIFT matching for iris recognition.,5580
conventional filter to SIFT wrongly is The objective paired matches.,The objective is to filter wrongly paired conventional SIFT matches.,5581
impairments. regions different at iris Due may to of high there some image be similarity,Due to high image similarity at different regions of iris there may be some impairments.,5582
keypoints II. detected of filtered gradient are These and Strata by finding in paired,These are detected and filtered by finding gradient of paired keypoints in Strata II.,5583
"used paired III. impairments scaling the in remove to keypoints is factor of Strata Further,","Further, the scaling factor of paired keypoints is used to remove impairments in Strata III.",5584
after be pairs retained matches iris are potential Strata likely to recognition. The for III,The pairs retained after Strata III are likely to be potential matches for iris recognition.,5585
over and This iris. the for existing SIFT of improvement matching significant marks accuracy FAR,This marks significant improvement of accuracy and FAR over the existing SIFT matching for iris.,5586
"set information, edge shadow and background, up The models of and are adaptively updated.","The models of background, edge information, and shadow are set up and adaptively updated.",5587
maximizing The field. is density solution posterior possibility the by the of obtained segmentation,The solution is obtained by maximizing the posterior possibility density of the segmentation field.,5588
fundamental uncertainty to combinatorial is the Robust with of in deal one optimization approaches optimization.,Robust optimization is one of the fundamental approaches to deal with uncertainty in combinatorial optimization.,5589
enlarge to and the . considerably problems to amenable effective solutions of class,and to enlarge considerably the class of problems amenable to effective solutions .,5590
"approximate networks are Bayesian is algorithms so necessary. monitoring in dynamic Exact intractable,","Exact monitoring in dynamic Bayesian networks is intractable, so approximate algorithms are necessary.",5591
Bayesian scoring Global (BN) metric. a the metric network new We called (GU) introduce Uniform,We introduce a new Bayesian network (BN) scoring metric called the Global Uniform (GU) metric.,5592
based metric a particular of type This on is default prior. parameter,This metric is based on a particular type of default parameter prior.,5593
is BNs for GU form derived. classes computing closed A for of special formula,A closed form formula for computing GU for special classes of BNs is derived.,5594
computing GU an Efficiently for open remains arbitrary an problem. BN,Efficiently computing GU for an arbitrary BN remains an open problem.,5595
such definition search with space. deals the analysis This of and paper one,This paper deals with the definition and analysis of one such search space.,5596
called (IASC). then Spectral clustering derive We Approximate spectral a Incremental Clustering algorithm novel,We then derive a novel spectral clustering algorithm called Incremental Approximate Spectral Clustering (IASC).,5597
model by Penna. proposed used for biological A commonly aging was,A commonly used model for biological aging was proposed by Penna.,5598
I live have delayed find corresponding that longer. that to populations senescence younger models,I find that models corresponding to delayed senescence have younger populations that live longer.,5599
We Decision first-order policies (MDPs) large select with representations. compact for Processes Markov,We select policies for large Markov Decision Processes (MDPs) with compact first-order representations.,5600
"using a taxonomic language. policies represented decision concept lists, Our as are ensembles of","Our policies are represented as ensembles of decision lists, using a taxonomic concept language.",5601
work application of the this to problem. reinforcement-learning We the relational discuss also,We also discuss the application of this work to the relational reinforcement-learning problem.,5602
"zero For collective a rescue terminates cooperative in growth rate, catastrophe.","For zero rescue rate, cooperative growth terminates in a collective catastrophe.",5603
"e.g. a signal, and spatial motion, global","global motion, and a spatial signal, e.g.",5604
modeled signal well is sparse often as some being domain. The spatial in,The spatial signal is often well modeled as being sparse in some domain.,5605
"new of First, Bayesianism a given. foundation is","First, a new foundation of Bayesianism is given.",5606
NP-SPEC a NP language for declarative problems specifying in is in way. a,NP-SPEC is a language for specifying problems in NP in a declarative way.,5607
"comprise arthropods. Pycnogonids, spiders, lineage sea or a of primitive","Pycnogonids, or sea spiders, comprise a primitive lineage of arthropods.",5608
"evolution. such, reveal hold As into arthropod insights potential they to","As such, they hold potential to reveal insights into arthropod evolution.",5609
juvenile the stage. the from molt emerge and Larvae hydroid simultaneously into,Larvae emerge from the hydroid and simultaneously molt into the juvenile stage.,5610
preceding the mature there of Over eight the are stages post-embryonic course adult. development,Over the course of post-embryonic development there are eight stages preceding the mature adult.,5611
"appendages, chelifores, some anteriormost the All structures, except transformation. degree for undergo of the","All structures, except for the anteriormost appendages, the chelifores, undergo some degree of transformation.",5612
and remain the hatching mobile are present over prior of course to Chelifores development.,Chelifores are present prior to hatching and remain mobile over the course of development.,5613
collaborative a systems. We principled and present efficient for multiagent planning dynamical algorithm,We present a principled and efficient planning algorithm for collaborative multiagent dynamical systems.,5614
connected subsystem The a hierarchy. are through subsystems,The subsystems are connected through a subsystem hierarchy.,5615
can which applied the step models intractable. be Hence to exactupdate is for it,Hence it can be applied to models for which the exactupdate step is intractable.,5616
Our for approach to successful. that results must chosen be appropriately be these this show,Our results show that these must be chosen appropriately for this approach to be successful.,5617
natural This settings. is in arguably many representation,This representation is arguably natural in many settings.,5618
"efficient. CP-semantics queries very the that strong and of Furthermore, ensures computation dominance is optimization","Furthermore, the strong CP-semantics ensures that computation of optimization and dominance queries is very efficient.",5619
decision demonstrate in value also representation this We the making. of,We also demonstrate the value of this representation in decision making.,5620
for We confidence sampling networks. in algorithms two inference present probabilistic Bayesian,We present two sampling algorithms for probabilistic confidence inference in Bayesian networks.,5621
"evidence. An of evaluation even unlikely excellent proposed for empirical performance, shows very the algorithms","An empirical evaluation of the proposed algorithms shows excellent performance, even for very unlikely evidence.",5622
We consider severalexperts. aggregating of task beliefs of the,We consider the task of aggregating beliefs of severalexperts.,5623
probabilitydistributions. beliefs We assume these as represented that are,We assume that these beliefs are represented as probabilitydistributions.,5624
"distribution aggregate Naturally, be thecombined would sets. the sample learned theideal one from","Naturally, theideal aggregate distribution would be the one learned from thecombined sample sets.",5625
Our well experiments show in practice. algorithmperforms that preliminary this,Our preliminary experiments show that this algorithmperforms well in practice.,5626
. on Mechanical to Reply Dependence Properties' 'Length Comment Scale DNA of,Reply to Comment on 'Length Scale Dependence of DNA Mechanical Properties' .,5627
"this to phenomenon."" quantify manuscript The is this purpose ""incommensurability of","The purpose of this manuscript is to quantify this ""incommensurability phenomenon.""",5628
"formed cell. from complex embryonic organism starting single a a is development, During","During embryonic development, a complex organism is formed from a single starting cell.",5629
each clustered quantitative temporal based profiles. TFs species We expression on for their further,We further clustered TFs for each species based on their quantitative temporal expression profiles.,5630
trends TF development in similar expression vertebrate showed This very and species. in insect,This showed very similar TF expression trends in development in vertebrate and insect species.,5631
fighting for provides insights Understanding new this treatments for with cancer. developing process,Understanding this process provides insights for developing new treatments for fighting with cancer.,5632
with plausible incomplete reasoning article deals spatial from large-scale This properties. knowledge about,This article deals with plausible reasoning from incomplete knowledge about large-scale spatial properties.,5633
"consisting observations,is set availableinformation, neighbour The of to points. a extrapolated of pointwise","The availableinformation, consisting of a set of pointwise observations,is extrapolated to neighbour points.",5634
"An numerical the evolutionary stable ab-initio of density-dependent, presented. dispersal strategy is study","An ab-initio numerical study of the density-dependent, evolutionary stable dispersal strategy is presented.",5635
scaling A to the laws data. are emerge few from seen general,A few general scaling laws are seen to emerge from the data.,5636
mixture (TrAp) developed algorithm problem. solving this for efficient have We an,We have developed an efficient algorithm (TrAp) for solving this mixture problem.,5637
the karyotypes using somatic the of and We tumor hypermutation datasets. method applicability demonstrate,We demonstrate the applicability of the method using tumor karyotypes and somatic hypermutation datasets.,5638
In problem in CLGs. of explore this the paper inference we,In this paper we explore the problem of inference in CLGs.,5639
in harder CLGs can that significantly than We inference in Bayes Nets. inference show be,We show that inference in CLGs can be significantly harder than inference in Bayes Nets.,5640
off work restart cut employed Previous have values. fixed strategies on,Previous work on restart strategies have employed fixed cut off values.,5641
and analyzed class Classes number size. were edges according of to,Classes were analyzed according to number of edges and class size.,5642
"Gaussian edges, By shape. the number classes of approaches distribution of a","By edges, the distribution of number of classes approaches a Gaussian shape.",5643
The generating undirected algorithm for graphs. orderly includes also program of variation a new,The program also includes a new variation of orderly algorithm for generating undirected graphs.,5644
"Although intractable, POMDPs formulation are solution. our tractable typically admits","Although POMDPs are typically intractable, our formulation admits tractable solution.",5645
inference enhanced proposed networks. has algorithm for these CLG No been exact,No exact inference algorithm has been proposed for these enhanced CLG networks.,5646
Lauritzen's algorithm to algorithm. comparable to its implement complexity in often simple Our and is,Our algorithm is simple to implement and often comparable in its complexity to Lauritzen's algorithm.,5647
We show previous that empirically higher approximate achieves algorithms. than it accuracy substantially,We show empirically that it achieves substantially higher accuracy than previous approximate algorithms.,5648
against is management accuracy. objective balance The tractability hypothesis of to,The objective of hypothesis management is to balance tractability against accuracy.,5649
Our literature. is related compared tracking work and in the approach to fusion,Our approach is compared to related work in the tracking and fusion literature.,5650
description an logics. DL-Lite family of important is,DL-Lite is an important family of description logics.,5651
belief evaluating complex querie probabilities network. a over of boolean,evaluating probabilities of complex boolean querie over a belief network.,5652
probabilistic on Preliminary evaluation effect the empirical constraint demonstrate of propagation computation.,Preliminary empirical evaluation demonstrate the effect of constraint propagation on probabilistic computation.,5653
is similarity definition issue the A of user in between key this preferences. approach,A key issue in this approach is the definition of similarity between user preferences.,5654
is {em the case of decision certainty}. This under making for,This is for the case of decision making under {em certainty}.,5655
belief approximation a to propose new value-directed for We POMDPs. approach state,We propose a new approach to value-directed belief state approximation for POMDPs.,5656
function computing concurrentoptions. This over allows for us the property value SMDPalgorithms use to,This property allows us to use SMDPalgorithms for computing the value function over concurrentoptions.,5657
on We clustering propose components a spectral local based principal (PCA). method analysis,We propose a spectral clustering method based on local principal components analysis (PCA).,5658
algorithms norm subdifferential. derived form are nonparametric a justified and using the nuclear Backfitting of,Backfitting algorithms are derived and justified using a nonparametric form of the nuclear norm subdifferential.,5659
gene methods data. are illustrated expression The on,The methods are illustrated on gene expression data.,5660
inference (BNs). networks propose approximate inBayesian recognition using for networks We,We propose using recognition networks for approximate inference inBayesian networks (BNs).,5661
input vector is to The evidential MLP thestates nodes. of of a the the,The input to the MLP is a vector of thestates of the evidential nodes.,5662
networks both and Recognition performwell observation using biases. correct incorrect,Recognition networks performwell using both correct and incorrect observation biases.,5663
systems. central in role Uncertainty spoken a plays dialogue,Uncertainty plays a central role in spoken dialogue systems.,5664
model process used to Some Markov are dialogue manager. like stochastic the decision (MDP) models,Some stochastic models like Markov decision process (MDP) are used to model the dialogue manager.,5665
uncertainty MDP-based increases. user's about when intention system degrades a fast,MDP-based system degrades fast when uncertainty about a user's intention increases.,5666
on model (POMDP). observable decision dialogue a partially propose Markov We the process based novel,We propose a novel dialogue model based on the partially observable Markov decision process (POMDP).,5667
grid methods two in point We selection algorithms. grid-based approximation also propose for,We also propose two methods for grid point selection in grid-based approximation algorithms.,5668
handles users real-time. from two-handed efficient multiple multi-threaded implementation input in The,The efficient multi-threaded implementation handles two-handed input from multiple users in real-time.,5669
forgetting variable by dilation. belief paper merging builds and This relationship between a first using,This paper first builds a relationship between belief merging and variable forgetting by using dilation.,5670
Variable applied to is forgetting operation. belief capture merging,Variable forgetting is applied to capture belief merging operation.,5671
inheritance present probabilistic We programming logic with overriding. under,We present probabilistic logic programming under inheritance with overriding.,5672
default recent to reasoning This conditional probabilistic approaches is using with constraints. done by,This is done by using recent approaches to probabilistic default reasoning with conditional constraints.,5673
properties the of entailment new We relations. analyze the semantic,We analyze the semantic properties of the new entailment relations.,5674
of is from The quantum transition. point investigated of macromolecule change biological conformational the,The conformational change of biological macromolecule is investigated from the point of quantum transition.,5675
theory A protein quantum on is folding proposed.,A quantum theory on protein folding is proposed.,5676
"the are from conformational states defined. of Simultaneously, the well local multi-minima property torsion potential","Simultaneously, from the multi-minima property of torsion potential the local conformational states are well defined.",5677
"folding protein rate the deduced. Then, formulas for general are","Then, the general formulas for protein folding rate are deduced.",5678
to We technique multi-layer clustering present the networks. feedforward applied deep neural learning,We present the clustering learning technique applied to multi-layer feedforward deep neural networks.,5679
to for technique paper this promote is systems. goal The robotic the vision of general-purpose,The goal of this paper is to promote the technique for general-purpose robotic vision systems.,5680
We in datasets datasets. report image static use and object tracking its,We report its use in static image datasets and object tracking datasets.,5681
followed structure This two strands. of disruption is in by,This is followed by disruption of structure in two strands.,5682
on results MNIST architecture The resulting state-of-the-art the classification achieves benchmark.,The resulting architecture achieves state-of-the-art results on the MNIST classification benchmark.,5683
Expectation for (EP) inference. framework provides a Propagation approximate,Expectation Propagation (EP) provides a framework for approximate inference.,5684
"EP the Through to expansion, order. see first we correct that is","Through the expansion, we see that EP is correct to first order.",5685
"be By complexity corrections higher the considering increasing of can orders, approximation. to applied polynomial","By considering higher orders, corrections of increasing polynomial complexity can be applied to the approximation.",5686
"the provide they of assessment Furthermore, error. approximation a polynomial-time","Furthermore, they provide a polynomial-time assessment of the approximation error.",5687
EP We provide both the exactness theoretical on solution. the of and also practical insights,We also provide both theoretical and practical insights on the exactness of the EP solution.,5688
"Since the not and are are provide experts, crowds necessarily rather erroneous. they labels noisy","Since crowds are not necessarily experts, the labels they provide are rather noisy and erroneous.",5689
"high-quality Although is leads mechanism not cost-effective. it labels, actually to the","Although the mechanism leads to high-quality labels, it is not actually cost-effective.",5690
"to models inference be make will surveyed. algorithms on Afterwards, these","Afterwards, algorithms to make inference on these models will be surveyed.",5691
"of the estimators proposed properties Moreover, are semiparametric asymptotic provided.","Moreover, asymptotic properties of the proposed semiparametric estimators are provided.",5692
Their practical theoretical previously relations discussed. and combination are,Their theoretical relations and practical combination are previously discussed.,5693
"descriptors. as descriptors, WSS While wavelet basis uses sub-bands treats pixel-value PSS","While PSS uses pixel-value descriptors, WSS treats wavelet sub-bands as basis descriptors.",5694
results results other compare in making. with our recent decision qualitative We,We compare our results with other recent results in qualitative decision making.,5695
work of This segmentation with indoor inputs. multi-class addresses RGB-D scenes,This work addresses multi-class segmentation of indoor scenes with RGB-D inputs.,5696
deals Bayesian paper solution the of representation asymmetric This with problems. and decision,This paper deals with the representation and solution of asymmetric Bayesian decision problems.,5697
an for asymmetric algorithm solving diagrams. influence give We,We give an algorithm for solving asymmetric influence diagrams.,5698
"and evolution. important organismic and fitness, between are links Molecular phenotypes functions, information genomic","Molecular phenotypes are important links between genomic information and organismic functions, fitness, and evolution.",5699
"called loci. also depend phenotypes, genomic often Complex quantitative traits, on multiple which are","Complex phenotypes, which are also called quantitative traits, often depend on multiple genomic loci.",5700
the and evolution We populations. derive approximate across distribution of for within equations phenotype values,We derive approximate evolution equations for the distribution of phenotype values within and across populations.,5701
evolutionary the processes. We our the of implications for predictability discuss of findings,We discuss the implications of our findings for the predictability of evolutionary processes.,5702
"we learning decision-tree consider two particular, In algorithms.","In particular, we consider two decision-tree learning algorithms.",5703
formulated as present influence solution new a to problems approach the decision We of diagrams.,We present a new approach to the solution of decision problems formulated as influence diagrams.,5704
"paper, RC. three we this key of present extensions In","In this paper, we present three key extensions of RC.",5705
We interpretation Entropy Principle terms. the game-theoretic (MaxEnt) Maximum an in of give,We give an interpretation of the Maximum Entropy (MaxEnt) Principle in game-theoretic terms.,5706
many to avoid allows problem this distinction cases. us in Our,Our distinction allows us to avoid this problem in many cases.,5707
in theorems characterization presetned are main paper. Their this,Their main characterization theorems are presetned in this paper.,5708
"scales an map is Finally, integrated information multiple by rule. final saliency the maximization across","Finally, across multiple scales is integrated the final saliency map by an information maximization rule.",5709
in The approach several appealing decomposition is ways.,The decomposition approach is appealing in several ways.,5710
"constraints propagation be challenging. because for can very a designing complex algorithm Secondly,","Secondly, because designing a propagation algorithm for complex constraints can be very challenging.",5711
can Probabilities most in is to the natural expert. the specified direction be that domain,Probabilities can be specified in the direction that is most natural to the domain expert.,5712
"networks. knowledge knowledge removes Thus, traditional the acquisition of bottleneck acquisition efficiently Bayesian the","Thus, the knowledge acquisition efficiently removes the traditional knowledge acquisition bottleneck of Bayesian networks.",5713
same shown strictly is normal multivariate for hold The also result distributions. positive to,The same result is shown to hold also for multivariate strictly positive normal distributions.,5714
"inference. speeding we this paper, evidence-specific value for Bayesian In abstraction networks use","In this paper, we use evidence-specific value abstraction for speeding Bayesian networks inference.",5715
is procedure many learning useful for complex variables. Our networks with particularly hidden,Our procedure is particularly useful for learning complex networks with many hidden variables.,5716
"are In EM parameter repeated optimization other cases, such likelihood required or techniques. for computations","In such cases, repeated likelihood computations are required for EM or other parameter optimization techniques.",5717
handwriting Arabic offline this we recognition. address word problem of paper the In,In this paper we address the problem of offline Arabic handwriting word recognition.,5718
"cut First, the several image blocks. in we","First, we cut the image in several blocks.",5719
"vector compute block, For of a each we descriptors.","For each block, we compute a vector of descriptors.",5720
"the Then, moments. we Zernik including K-means cluster to use features low-level Hu and","Then, we use K-means to cluster the low-level features including Zernik and Hu moments.",5721
recognition FAN DBN demonstrate results outperform rates and good The .,The results demonstrate FAN and DBN outperform good recognition rates .,5722
constraint reasoning a and probabilistic reasoning. document between We connection,We document a connection between constraint reasoning and probabilistic reasoning.,5723
"discuss methods. briefly some empirical We and related results,","We briefly discuss some empirical results, and related methods.",5724
a information perspective. partial Planning agents decision- theoretic state distributed considered from with is for,Planning for distributed agents with partial state information is considered from a decision- theoretic perspective.,5725
control. models of allow for decentralized We generalizations the and MDP both describe that POMDP,We describe generalizations of both the MDP and POMDP models that allow for decentralized control.,5726
Markov These and between control results a of illustrate decentralized difference fundamental centralized processes. complexity,These complexity results illustrate a fundamental difference between centralized and decentralized control of Markov processes.,5727
"models require expansion. Such and analysis, adaptation application","Such models require application analysis, adaptation and expansion.",5728
"derived from To present evaluate application our real-world data. approach, models we usage","To evaluate our approach, we present models derived from real-world application usage data.",5729
by designed peptide al. was Loose et This earlier,This peptide was earlier designed by Loose et al.,5730
linguistic of peptides. antimicrobial using model natural a,using a linguistic model of natural antimicrobial peptides.,5731
employed commonly operator used This an can strategies. as to search extension be,This operator can be used as an extension to commonly employed search strategies.,5732
artificial in and evaluated real-world It is with experiments data.,It is evaluated in experiments with artificial and real-world data.,5733
compositional and serial with parallel paper objects models multiple part This describes sharing. of,This paper describes serial and parallel compositional models of multiple objects with part sharing.,5734
executive an which yield sizes on are lattices parts represented summary These of decreasing description.,These parts are represented on lattices of decreasing sizes which yield an executive summary description.,5735
and learning algorithms for these models. inference describe We,We describe inference and learning algorithms for these models.,5736
acyclic for paper This measures associated probability independence of sets directed graphs. with analyzes concepts,This paper analyzes independence concepts for sets of probability measures associated with directed acyclic graphs.,5737
separation and Markov independence paper epistemic standard the The violate condition shows that properties. desirable,The paper shows that epistemic independence and the standard Markov condition violate desirable separation properties.,5738
"proposed, independence. To on epistemic condition strong this situation, overcome unsatisfactory is Markov based a","To overcome this unsatisfactory situation, a strong Markov condition is proposed, based on epistemic independence.",5739
discussed is sources information this The setting. in of multiple fusion,The fusion of multiple sources information is discussed in this setting.,5740
"considered of including are merging reinforcement, and classes adaptive conjunctive, operators. operators averaging Different disjunctive,","Different classes of merging operators are considered including conjunctive, disjunctive, reinforcement, adaptive and averaging operators.",5741
postulates. propose we Then terms classes of in to these analyse,Then we propose to analyse these classes in terms of postulates.,5742
under plan-generation stochastic a of Techniques for recognition model uncertainty require the process. plan,Techniques for plan recognition under uncertainty require a stochastic model of the plan-generation process.,5743
plan-generation State-Dependent an introduce (PSDGs) Probabilistic agent's process. Grammars We to represent,We introduce Probabilistic State-Dependent Grammars (PSDGs) to represent an agent's plan-generation process.,5744
"This hold anymore result, does for Bayesian however, not general networks.","This result, however, does not hold anymore for general Bayesian networks.",5745
"sequentially. thus are entropy which We models, kind new maximum computed of present a","We thus present a new kind of maximum entropy models, which are computed sequentially.",5746
"a the encodes only As basic drawback, structures. dictionary","As a drawback, the dictionary only encodes basic structures.",5747
"magneto- is This instance, or case, the (M/EEG). for multi-trial electroencephalography in","This is the case, for instance, in multi-trial magneto- or electroencephalography (M/EEG).",5748
simulated on is neuroelectric method The and data. validated real,The method is validated on simulated and real neuroelectric data.,5749
are interpretations. on Belief many-valued distributions modeled all the states possibility by set of,Belief states are modeled by possibility distributions on the set of all many-valued interpretations.,5750
lineages Phylogeographic studies multiple reveal species. morphologically-cryptic within frequently,Phylogeographic studies frequently reveal multiple morphologically-cryptic lineages within species.,5751
yet What lineages whether such evolutionary nascent or unclear is is species represent ephemera.,What is yet unclear is whether such lineages represent nascent species or evolutionary ephemera.,5752
vehicle become Monte Bayesian a has inference in major sampling Carlo approximate for networks.,Monte Carlo sampling has become a major vehicle for approximate inference in Bayesian networks.,5753
selecting for sequence. for Sobol numbers an We propose direction algorithm,We propose an algorithm for selecting direction numbers for Sobol sequence.,5754
for (nHDP) process a Dirichlet develop topic We modeling. nested hierarchical hierarchical,We develop a nested hierarchical Dirichlet process (nHDP) for hierarchical topic modeling.,5755
nature the of procedure. non-random beliefs the priors include The sampling about,The priors include beliefs about the nature of the non-random sampling procedure.,5756
uncetainties Conversations with various kinds. of abound,Conversations abound with uncetainties of various kinds.,5757
compared tree propagation. a speed-up experimental indicating to junction traditional results present We substantial,We present experimental results indicating a substantial speed-up compared to traditional junction tree propagation.,5758
this Prolog described. is implementation for A,A Prolog implementation for this is described.,5759
representations the We of discuss constructing also possibility the explicit of posterior.,We also discuss the possibility of constructing explicit representations of the posterior.,5760
public is a in health problem countries. main Malnutrition developing,Malnutrition is a main public health problem in developing countries.,5761
and the is high. still rate mortality increasing Incidence is,Incidence is increasing and the mortality rate is still high.,5762
The was lactation. a food to rat competition model applied during malnutrition provoke method to,The food competition method was applied to a rat model to provoke malnutrition during lactation.,5763
plenty delayed increase weight is available. milk is even if there The,The weight increase is delayed even if there is plenty milk available.,5764
can using a large network. dynamic MDPs be compactly represented Bayesian Many,Many large MDPs can be represented compactly using a dynamic Bayesian network.,5765
type policy improvement. of ill-suited is This weighted projection to,This type of weighted projection is ill-suited to policy improvement.,5766
process. this use a iteration then algorithm value policy in as subroutine We a determination,We then use this value determination algorithm as a subroutine in a policy iteration process.,5767
probabilistic a have in networks reasoning Qualitative been designed for probabilistic qualitative way.,Qualitative probabilistic networks have been designed for probabilistic reasoning in a qualitative way.,5768
an insightful more present unresolved algorithm We computing for results trade-offs. for,We present an algorithm for computing more insightful results for unresolved trade-offs.,5769
 compute paper topological This a rules. automatically to  SWRL presents relations using method,  This paper presents a method to compute automatically topological relations using SWRL rules.,5770
"instance, ""RailStation"". that ""Railway"" a For a a overlaps is ""Building""","For instance, a ""Building"" that overlaps a ""Railway"" is a ""RailStation"".",5771
aid of analysis Three steps selection. in the,Three steps aid in the analysis of selection.,5772
"by their component causes. First, describe phenotypes","First, describe phenotypes by their component causes.",5773
"terms models. evolutionary and Traditionally, those of the by given components regression covariance, are variance,","Traditionally, those components are given by the covariance, variance, and regression terms of evolutionary models.",5774
those I information interpret show to how with theory. to statistical expressions respect,I show how to interpret those statistical expressions with respect to information theory.,5775
applications. object challenging a a an task video multimedia is of in Segmentation from,Segmentation of an object from a video is a challenging task in multimedia applications.,5776
"video problem efficiency. segmentation address the paper, from In the we this of perspective of","In this paper, we address the problem of video segmentation from the perspective of efficiency.",5777
interactive scenarios. These automatic for both are segmentation experimented video improvements and,These improvements are experimented for both automatic and interactive video segmentation scenarios.,5778
"quality addition is tested in both segmentation the Moreover, quantitatively qualitatively. also efficiency, to and","Moreover, in addition to the efficiency, segmentation quality is also tested both quantitatively and qualitatively.",5779
"segmentation Indeed, is examples, quality. efficiency for some without observed of loss challenging time significant","Indeed, for some challenging examples, significant time efficiency is observed without loss of segmentation quality.",5780
"high-quality, assembly of data However, sequences challenging. into finished remains such genome","However, assembly of such data into high-quality, finished genome sequences remains challenging.",5781
intended to assembly. methods genome are assess Assemblathon The in current competitions state-of-the-art,The Assemblathon competitions are intended to assess current state-of-the-art methods in genome assembly.,5782
of one of daily important most communication. means Handwriting is the,Handwriting is one of the most important means of daily communication.,5783
proposed (ESR) successfully. sentence system tested and approach recognition English The is designed,The proposed approach English sentence recognition (ESR) system is designed and tested successfully.,5784
the Experimental higher were that accuracy speed show examined. results and,Experimental results show that the higher speed and accuracy were examined.,5785
ligand in binding. chain factor important an flexibility is Side,Side chain flexibility is an important factor in ligand binding.,5786
implemented This in FlexAID. the is software approach tool,This approach is implemented in the software tool FlexAID.,5787
for need system automatic lip-reading is an increasing. The ever,The need for an automatic lip-reading system is ever increasing.,5788
"In special addition, people imperative needs. visual is with information among","In addition, visual information is imperative among people with special needs.",5789
"presented. a detection is of language the approach ontology using objects Thus, knowledge-based OWL","Thus, a knowledge-based detection approach of objects using the OWL ontology language is presented.",5790
rules. to detection SWRL define This knowledge used is,This knowledge is used to define SWRL detection rules.,5791
"this the paper examines models. mathematical of using dynamics system, evolutionary This","This paper examines the evolutionary dynamics of this system, using mathematical models.",5792
are and extinction Implications avoid required to conditions discussed.,Implications and conditions required to avoid extinction are discussed.,5793
Experimental tests are model this for suggested.,Experimental tests for this model are suggested.,5794
a well-suited Poker's unique challenge in research artificial for Computer characteristics present intelligence.,Computer Poker's unique characteristics present a well-suited challenge for research in artificial intelligence.,5795
diseases. antibiotic of ability bacteria continued resistance treat The our evolution infectious to among threatens,The evolution of antibiotic resistance among bacteria threatens our continued ability to treat infectious diseases.,5796
bacterial sustainable never for need The has strategies cure infections been greater. to,The need for sustainable strategies to cure bacterial infections has never been greater.,5797
for may We results resistance. a framework anticipate conceptual these that antibiotic provide managing,We anticipate that these results may provide a conceptual framework for managing antibiotic resistance.,5798
key the the presented of is idea perception subjective of The reality. model idea,The key idea of the presented perception model is the idea of subjective reality.,5799
The the principle formulated. world is perceived of of relativity,The principle of the relativity of perceived world is formulated.,5800
which consider the We models linear mixed in observations grouped. are,We consider linear mixed models in which the observations are grouped.,5801
the a assets dependency-based game. relationships between depicts of the It,It depicts the dependency-based relationships between the assets of a game.,5802
"arsenal an weapons. he can player example, construct a must build before For","For example, a player must construct an arsenal before he can build weapons.",5803
This graph improve balancing. paper helps the to game dependency that exhibits parallel,This paper exhibits that the parallel dependency graph helps to improve game balancing.,5804
"the But, many models decisions. important include such aspects expert are to failed of","But, many such models are failed to include important aspects of the expert decisions.",5805
"decision rules. to it is Therefore, essential the minimize","Therefore, it is essential to minimize the decision rules.",5806
standard clique the tree is in networks. method doing algorithm for Bayesian inference The,The clique tree algorithm is the standard method for doing inference in Bayesian networks.,5807
in manipulating by potentials - over the clique It variables distributions a works clique.,It works by manipulating clique potentials - distributions over the variables in a clique.,5808
viewed known Many approximate this algorithms can of inference approach. instances be as,Many known approximate inference algorithms can be viewed as instances of this approach.,5809
the to improve quality gradually the Iterations are estimation. of used,Iterations are used to gradually improve the quality of the estimation.,5810
factored of We the POMDPs. are in for interested problem planning,We are interested in the problem of planning for factored POMDPs.,5811
networks. qualitative probabilistic abstractions as Qualitative of introduced networks belief Bayesian been have,Qualitative probabilistic networks have been introduced as qualitative abstractions of Bayesian belief networks.,5812
enhanced for finer present detail. qualitative of a formalism We networks level with an,We present an enhanced formalism for qualitative networks with a finer level of detail.,5813
"we the application its to examples. Finally, by method illustrate some practical","Finally, we illustrate the method by its application to some practical examples.",5814
decision theory This qualitative for of purely paper version uncertainty. under a investigates Savage's making,This paper investigates a purely qualitative version of Savage's theory for decision making under uncertainty.,5815
to presents an paper dynamic decision time-critical modeling. This approach,This paper presents an approach to time-critical dynamic decision modeling.,5816
dynamic modeling time-critical diagram knowledge A the method is called influence and proposed. representation,A knowledge representation and modeling method called the time-critical dynamic influence diagram is proposed.,5817
The abstraction model. proposed the represent has approach the within ability to space-temporal,The proposed approach has the ability to represent space-temporal abstraction within the model.,5818
algorithm An is model construction of knowledge-based also the provided. outline,An outline of the knowledge-based model construction algorithm is also provided.,5819
scheme an recently was called are The introduced. that extracted mini-bucket elimination heuristics from approximation,The heuristics are extracted from an approximation scheme called mini-bucket elimination that was recently introduced.,5820
introduced The search. paper and it Branch-and-Bound within evaluated first the idea,The first paper introduced the idea and evaluated it within Branch-and-Bound search.,5821
within current the and evaluated further In search. the idea extended Best-First paper is,In the current paper the idea is further extended and evaluated within Best-First search.,5822
algorithm probabilistic for exact present a inference Bayesian networks. parallel I in,I present a parallel algorithm for exact probabilistic inference in Bayesian networks.,5823
proof exists for based calculus. a There refutation a complete on logic procedure resolution-style Possibilistic,There exists for Possibilistic logic a proof procedure based on a refutation complete resolution-style calculus.,5824
sensors. provided any This paper validating an by information presents for time algorithm probabilistic the,This paper presents an any time probabilistic algorithm for validating the information provided by sensors.,5825
network The of consists models. Bayesian system two,The system consists of two Bayesian network models.,5826
list faulty potentially of provides sensors. It a,It provides a list of potentially faulty sensors.,5827
sensors the plant. power been The of of validation has in algorithm temperature a tested,The algorithm has been tested in the validation of temperature sensors of a power plant.,5828
described The in are paper those literature. transformations different in this the from described,The transformations described in this paper are different from those described in the literature.,5829
"particular, the In the principles latter congruence not. of former satisfies does the ordinal while","In particular, the former satisfies the principles of ordinal congruence while the latter does not.",5830
approach example. dynamic a is a contrasted with for simple network Bayesian Our medical,Our approach is contrasted with a dynamic Bayesian network for a simple medical example.,5831
the MDP and original be The solved efficiently. abstract approximates can more,The abstract MDP approximates the original and can be solved more efficiently.,5832
several which We discuss be good can generated in to ways macro-actions quality. solution ensure,We discuss several ways in which macro-actions can be generated to ensure good solution quality.,5833
the backward the to algorithms. lead Both KL-divergence and inefficient KL-divergence,Both the KL-divergence and the backward KL-divergence lead to inefficient algorithms.,5834
"mean may have For networks, large rely to one field approximations. however, on","For large networks, however, one may have to rely on mean field approximations.",5835
aggregation local Even rule probability conditions weaker out within tables. conditional,Even weaker conditions rule out local aggregation within conditional probability tables.,5836
procedure network. consensus This suggests a for straightforward a constructing Markov,This suggests a straightforward procedure for constructing a consensus Markov network.,5837
describe concepts There information not observations are that the from we extract to made. two,There are two concepts that we describe to extract information from the observations not made.,5838
"to likely others. be more if are First, before present, reported some symptoms,","First, some symptoms, if present, are more likely to be reported before others.",5839
"that there a Thus, desirable are reporting toward symptoms present. is bias","Thus, there is a desirable bias toward reporting symptoms that are present.",5840
significantly improve concepts inference. a develop simple diagnostic model can for We that these,We develop a simple model for these concepts that can significantly improve diagnostic inference.,5841
with method efficiency lazy evaluation examples. based emphasized The through is improvements obtained the,The efficiency improvements obtained with the lazy evaluation based method is emphasized through examples.,5842
framework We reasoning. presented the also probabilistic to compare other first-order with approaches,We also compare the presented framework with other approaches to first-order probabilistic reasoning.,5843
ever-increasing of There managing for uncertainty. procedures an is available variety,There is available an ever-increasing variety of procedures for managing uncertainty.,5844
these microcircuits differences unknown. of The in is synaptic structural cause,The cause of these structural differences in synaptic microcircuits is unknown.,5845
synapses. both including for depressing This heterogeneous networks and holds facilitating,This holds for heterogeneous networks including both facilitating and depressing synapses.,5846
proposed the validation suggest experiments the of for mechanism. further We,We further suggest experiments for the validation of the proposed mechanism.,5847
preference of automated form tools. component representations importnat Such decision an orderings of qualitative,Such representations of qualitative preference orderings form an importnat component of automated decision tools.,5848
arguably a is natural. compact and ofetn representation Such,Such a representation is ofetn compact and arguably natural.,5849
network with work construction. and situation-specific previous paper network This extends fragments,This paper extends previous work with network fragments and situation-specific network construction.,5850
"the formally a conditional alternative for We table. an probability network, asymmetry define representation","We formally define the asymmetry network, an alternative representation for a conditional probability table.",5851
partially present asymmetry networks. We for also representation object-oriented an specified,We also present an object-oriented representation for partially specified asymmetry networks.,5852
parsimonious. the representation We show that is,We show that the representation is parsimonious.,5853
illustration. an transmission An is used to application as education,An application to education transmission is used as an illustration.,5854
"In suggest paper, use we in contextual weak Bayesian independence (CWI) this the of networks.","In this paper, we suggest the use of contextual weak independence (CWI) in Bayesian networks.",5855
marginalizes qualitative to contribute nodes first the ambiguous approach that relationships. incrementally The,The first approach incrementally marginalizes nodes that contribute to the ambiguous qualitative relationships.,5856
"types we for of three classes, these provide HPPs. For results","For these classes, we provide three types of results for HPPs.",5857
"compute First, the we set HPP. all an develop algorithms of ground of consequences to","First, we develop algorithms to compute the set of all ground consequences of an HPP.",5858
"consistency is P input, (""Given consistent? and an HPP as P","and consistency (""Given an HPP P as input, is P consistent?",5859
process exponentially evolves. that as show a belief error the the We state in contracts,We show that the error in a belief state contracts exponentially as the process evolves.,5860
"process our bounded multiple remains error indefinitely. Thus, in with the even approximations,","Thus, even with multiple approximations, the error in our process remains bounded indefinitely.",5861
elicitation. utility We application the techniques investigate to of classification,We investigate the application of classification techniques to utility elicitation.,5862
a it user. Thus necessary for is model to new elicit utility separately each,Thus it is necessary to elicit a utility model separately for each new user.,5863
"is decomposable. not is tedious, large Elicitation and the and if particularly outcome long space","Elicitation is long and tedious, particularly if the outcome space is large and not decomposable.",5864
approaches two utility to elicitation. common function are There,There are two common approaches to utility function elicitation.,5865
this clusters minimize of distance by utility measure. do appropriate an We functions identifying that,We do this by identifying clusters of utility functions that minimize an appropriate distance measure.,5866
"are complexity, this practical in difficult of to medial skeletons applications. use Because","Because of this complexity, medial skeletons are difficult to use in practical applications.",5867
is energy in The building. first saving the,The first is energy saving in the building.,5868
this population to present dynamics a model We address question.,We present a population dynamics model to address this question.,5869
independence. definitions of Walley's addresses irrelevance these and This questions through paper,This paper addresses these questions through Walley's definitions of irrelevance and independence.,5870
decision framework a axiomatic under presents finite uncertainty setting. paper an for qualitative in This,This paper presents an axiomatic framework for qualitative decision under uncertainty in a finite setting.,5871
"(or sup-min integral. a fuzzy) called is corresponding The utility Sugeno expressed expression, by","The corresponding utility is expressed by a sup-min expression, called Sugeno (or fuzzy) integral.",5872
conflicts analysis in-depth are incurred frameworks of by an integrated without viable No uncertainties.,No integrated frameworks are viable without an in-depth analysis of conflicts incurred by uncertainties.,5873
model uncertainty. theory definite Our to extends argumentation framework,Our framework extends definite argumentation theory to model uncertainty.,5874
and uncertain three supports It conflicting knowledge. over views,It supports three views over conflicting and uncertain knowledge.,5875
"conclusions different depending Thus, on the engineers draw context knowledge can (i.e. application","Thus, knowledge engineers can draw different conclusions depending on the application context (i.e.",5876
as in are crucial awareness. real-world These aspects domains such battlefield,These aspects are crucial in real-world domains such as battlefield awareness.,5877
"SPOOK, limitations. present In this these we implemented addresses paper, that an system","In this paper, we present SPOOK, an implemented system that addresses these limitations.",5878
In paper preference we case-based the alleviating to bottleneck. approach propose a this elicitation,In this paper we propose a case-based approach to alleviating the preference elicitation bottleneck.,5879
provide for computational measures. We all techniques three,We provide computational techniques for all three measures.,5880
present of situation-specific networks. definitions completeness We query and,We present definitions of query completeness and situation-specific networks.,5881
the knowledge base describe that conditions on guarantee query We completeness.,We describe conditions on the knowledge base that guarantee query completeness.,5882
to earlier work work The our relationship KBMC also of on discussed. is,The relationship of our work to earlier work on KBMC is also discussed.,5883
?missing' characterized These information. are by essentially,These are essentially characterized by ?missing' information.,5884
complex with the phenomenon a phenotype is interacts where Development environmental genetic forming and inputs.,Development is a complex phenomenon where the forming phenotype interacts with genetic and environmental inputs.,5885
development investigated. for has developmental an their model thoroughly important studies and Teeth been are,Teeth are an important model for developmental studies and their development has been thoroughly investigated.,5886
decision tool Influence powerful serve for as problems. modelling symmetric a diagrams,Influence diagrams serve as a powerful tool for modelling symmetric decision problems.,5887
influence set diagram we for an decisions involved. solving a determine When the of strategies,When solving an influence diagram we determine a set of strategies for the decisions involved.,5888
in A over a is its a past. function decision for strategy variable principle,A strategy for a decision variable is in principle a function over its past.,5889
welldefined. is the scenario that ensures This decision constraint,This constraint ensures that the decision scenario is welldefined.,5890
diagrams paper we In this partial with i.e. influence deal,In this paper we deal with partial influence diagrams i.e.,5891
partial temporal a only with diagrams specified. influence ordering,influence diagrams with only a partial temporal ordering specified.,5892
"into decision problem these a is large, divided smaller pieces. stochastic approaches, In","In these approaches, a large, stochastic decision problem is divided into smaller pieces.",5893
examples are approach. illustrate provided A to of the number proposed,A number of examples are provided to illustrate the proposed approach.,5894
and interest logic representing for two frameworks knowledge. possibilistic Possibilistic bases graphs different of are,Possibilistic logic bases and possibilistic graphs are two different frameworks of interest for representing knowledge.,5895
graphs. possibilistic These two induce kinds two of conditioning of types,These two types of conditioning induce two kinds of possibilistic graphs.,5896
"bases into In of translation cases, a possibilistic is graphs provided. both these","In both cases, a translation of these graphs into possibilistic bases is provided.",5897
knowledge from a The described. graph base is into translation min-based possibilistic a converse also,The converse translation from a possibilistic knowledge base into a min-based graph is also described.,5898
The such with models. is here work concerned acquiring presented,The work presented here is concerned with acquiring such models.,5899
the We propagation MSBNs. Shafer-Shenoy in apply and inference to lazy,We apply the Shafer-Shenoy and lazy propagation to inference in MSBNs.,5900
solving search algorithms approaches among Stochastic hard most are combinatorial the sucessful for problems.,Stochastic search algorithms are among the most sucessful approaches for solving hard combinatorial problems.,5901
"present a recognition. abductive, theory of probabilistic We new plan","We present a new abductive, probabilistic theory of plan recognition.",5902
that There inference is really evidence matter. in numbers the don't probabilistic,There is evidence that the numbers in probabilistic inference don't really matter.,5903
"reliable. equally not are or experts equally Criteria are important, competent not","Criteria are not equally important, experts are not equally competent or reliable.",5904
decisions as planning. of becoming are (MDPs) increasing theoretic models Markov decision processes popular,Markov decisions processes (MDPs) are becoming increasing popular as models of decision theoretic planning.,5905
for belief is approximation of investigated. A foundation functions and proposed conceptual,A conceptual foundation for approximation of belief functions is proposed and investigated.,5906
consistency the requirements of and closeness. based It on is,It is based on the requirements of consistency and closeness.,5907
"of approximation the out optimal Unfortunately, be computation intractable. the to turns","Unfortunately, the computation of the optimal approximation turns out to be intractable.",5908
are the belief methods approximations to compared functions. proposed These earlier of,These methods are compared to the earlier proposed approximations of belief functions.,5909
the information the in logic. of addresses This paper problem merging framework of uncertain possibilistic,This paper addresses the problem of merging uncertain information in the framework of possibilistic logic.,5910
the logic. the These rules combination first at language meta-level possibilistic outside described are of,These combination rules are first described at the meta-level outside the language of possibilistic logic.,5911
uncertain are or decision processes domains often formulated Decision-making Markov stochastic (MDPs). in as problems,Decision-making problems in uncertain or stochastic domains are often formulated as Markov decision processes (MDPs).,5912
model promising reasoning knowledge. is a with uncertain Argumentation for,Argumentation is a promising model for reasoning with uncertain knowledge.,5913
bases. with context the is approach of knowldge illustrated in Our stratified reasoning,Our approach is illustrated in the context of reasoning with stratified knowldge bases.,5914
to optimal instance. compare the each each and to outcomes for other We their policies,We compare their outcomes to each other and to the optimal policies for each instance.,5915
technique existing POMDP algorithms. into iteration can The value any incorporated be easily,The technique can be easily incorporated into any existing POMDP value iteration algorithms.,5916
of can pruning orders We magnitude the several incremental run make that find technique faster.,We find that the technique can make incremental pruning run several orders of magnitude faster.,5917
of probabilities general for in exact updating exist algorithms two Networks. There Bayesian forms,There exist two general forms of exact algorithms for updating probabilities in Bayesian Networks.,5918
preferences. about normative a complex Classical Decision representing for and reasoning provides framework Theory,Classical Decision Theory provides a normative framework for representing and reasoning about complex preferences.,5919
set set. rough captures of in elements indiscernibility of a notion The,The notion of rough set captures indiscernibility of elements in a set.,5920
"real between the universes. many situations, relation establishes an information But, in life different system","But, in many real life situations, an information system establishes the relation between different universes.",5921
and an task recognition in (FR) Face is vision. computer important pattern recognition,Face recognition (FR) is an important task in pattern recognition and computer vision.,5922
for powerful been to framework be demonstrated (SR) has representation Sparse FR. a,Sparse representation (SR) has been demonstrated to be a powerful framework for FR.,5923
a hint. coefficients recognition provide representation then sparse The,The sparse representation coefficients then provide a recognition hint.,5924
based basic SR on a sparse Early algorithms model. are,Early SR algorithms are based on a basic sparse model.,5925
"many has based which proposed a algorithms. framework, recently BSBL block-sparse-model is advantages existing over","BSBL is a recently proposed framework, which has many advantages over existing block-sparse-model based algorithms.",5926
HORSES implement procedure. We algorithm efficient to construct the an,We construct an efficient algorithm to implement the HORSES procedure.,5927
The of human formal to second because is it avoids the reference definition being. concept,The second definition is formal because it avoids reference to the concept of human being.,5928
"unstable, is producing activity dynamics fluctuations scale-free. are Neuronal essentially intrinsically that","Neuronal dynamics is intrinsically unstable, producing activity fluctuations that are essentially scale-free.",5929
"maximal entrainment possesses itself when the natural-like, found statistics. be Response to scale-free was input","Response entrainment was found to be maximal when the input itself possesses natural-like, scale-free statistics.",5930
We methods correct this two to introduce bias.,We introduce two methods to correct this bias.,5931
techniques time co-segregating loci. selection to required in Different series detect will linked of be,Different techniques will be required to detect selection in time series of co-segregating linked loci.,5932
current discussed. of A research are number issues open,A number of current open research issues are discussed.,5933
"Moreover, do not correspondences local and the these missing works handle simultaneously deformations. large","Moreover, these works do not handle simultaneously the missing correspondences and local large deformations.",5934
"coupled technique high-throughput to Recently, have the we mutagenesis.","Recently, we have coupled the technique to high-throughput mutagenesis.",5935
graphical such is of characterization graphs given. A,A graphical characterization of such graphs is given.,5936
for networks It the criterion resembles locally. can implemented be and Bayesian d-separation well-known,It resembles the well-known d-separation criterion for Bayesian networks and can be implemented locally.,5937
decision are approaches limited resources. making Flexible under computations two and to approximate computation,Flexible and approximate computations are two approaches to decision making under limited computation resources.,5938
a an enables find computation the Approximate solution best within satisfactory to agent deadline.,Approximate computation enables an agent to find the best satisfactory solution within a deadline.,5939
"NaviCell provides addition, for community-based framework In a curation. map","In addition, NaviCell provides a framework for community-based map curation.",5940
"the an by integrated final is rule. map multiple scales information Finally, maximization across saliency","Finally, across multiple scales is integrated the final saliency map by an information maximization rule.",5941
paper concerned This with is finding policies optimal POMDPs. for,This paper is concerned with finding optimal policies for POMDPs.,5942
inference into evaluation reducing network influence paper diagram problems. about (ID) is (BN) Bayesian This,This paper is about reducing influence diagram (ID) evaluation into Bayesian network (BN) inference problems.,5943
"class functional (i.e. a class between algebra correspondence produces algebraic class's a definition Moreover,","Moreover, class algebra produces a functional correspondence between a class's class algebraic definition (i.e.",5944
"set ""intent"") (i.e. the expression of the all satisfy instances and which its","its ""intent"") and the set of all instances which satisfy the expression (i.e.",5945
in definitions. various locating thus and provides assistance counterexamples The framework examples quickly of,The framework thus provides assistance in quickly locating examples and counterexamples of various definitions.,5946
provide on regions of generated evaluation networks. empirical preliminary completeness and randomly identify We,We identify regions of completeness and provide preliminary empirical evaluation on randomly generated networks.,5947
"many the remain address environments. but latter in a problem papers Earlier problems, dropouts","Earlier papers address the latter problems, but dropouts remain a problem in many environments.",5948
correlated with not presented to actions able as effects. The deal is algorithm with,The algorithm as presented is not able to deal with actions with correlated effects.,5949
that corrects weakness. new this describe a regression operator theoretic We decision,We describe a new decision theoretic regression operator that corrects this weakness.,5950
"a straightforward, technical approach. extension complicated more somewhat While conceptually requires this","While conceptually straightforward, this extension requires a somewhat more complicated technical approach.",5951
for problems. solving among best computationally reasoning search hard are the and Stochastic algorithms,Stochastic algorithms are among the best for solving computationally hard search and reasoning problems.,5952
a The procedures of by characterized random variable. runtime is such,The runtime of such procedures is characterized by a random variable.,5953
algorithms distributions. Different to rise give different probability,Different algorithms give rise to different probability distributions.,5954
genetic is is Most inference evolution on rare. based adaptive assuming frameworks population,Most population genetic inference is based on frameworks assuming adaptive evolution is rare.,5955
presently solving most for method find is incremental efficient the POMDPs. pruning that exact We,We find that incremental pruning is presently the most efficient exact method for solving POMDPs.,5956
for of Two considered. models are local approaches combination,Two approaches for combination of local models are considered.,5957
"perturbation calculation for We variances, of global models. bounds and also discuss and utilities expected","We also discuss calculation of bounds for expected utilities and variances, and global perturbation models.",5958
of is of diversity principle FIS representation formulated. construction The,The principle of representation diversity of FIS construction is formulated.,5959
"the of paper neural i.e. suggests Besides the elements, structure","Besides the paper suggests the structure of neural elements, i.e.",5960
FIS and the construction. for are that detectors basis processors zone,zone detectors and processors that are the basis for FIS construction.,5961
this empirical The encouraging. are results of study very,The results of this empirical study are very encouraging.,5962
with the BKBs consistent requires in most Reasoning evidence. finding inferences the probable,Reasoning in BKBs requires finding the most probable inferences consistent with the evidence.,5963
to performance significantly when applied that shows cost-sharing evaluation BKBs. Empirical improves,Empirical evaluation shows that cost-sharing improves performance significantly when applied to BKBs.,5964
Image played in image has thresholding an important role segmentation.,Image thresholding has played an important role in image segmentation.,5965
images been with approach (THFCM). through Experimental have obtained results proposed test standard the,Experimental results with standard test images have been obtained through the proposed approach (THFCM).,5966
but provide We efficient a procedure for also simple nonuniform partition.,We also provide a simple but efficient procedure for nonuniform partition.,5967
evidence converges. We that empirical algorithm the provide,We provide empirical evidence that the algorithm converges.,5968
key motivate triage trauma-care We ideas in and of context the decisions. transportation,We motivate key ideas in the context of trauma-care triage and transportation decisions.,5969
that technique we decomposition. main a is relevance-based Our call contribution,Our main contribution is a technique that we call relevance-based decomposition.,5970
We our report empirical of significance results demonstrate of that tests approach. practical,We report results of empirical tests that demonstrate practical significance of our approach.,5971
shown paradox has not nonmonotonic logics do lottery that the Poole correctly. handle,Poole has shown that nonmonotonic logics do not handle the lottery paradox correctly.,5972
in Default common conceptual representing logic tasks. difficulties reasoning some sense encounters,Default logic encounters some conceptual difficulties in representing common sense reasoning tasks.,5973
in Image image one is the problems processing. fundamental denoising of,Image denoising is one of the fundamental problems in image processing.,5974
IQR kXk of filter. implemented window to A size was support,A window of size kXk was implemented to support IQR filter.,5975
of as outside treated IQR range the pixel. Each window is the pixel noisy kXk,Each pixel outside the IQR range of the kXk window is treated as noisy pixel.,5976
The obtained averaging. of by the was noisy pixels local estimation,The estimation of the noisy pixels was obtained by local averaging.,5977
customary Image by interpolation been has techniques. interpolation used spaciously,Image interpolation has been used spaciously by customary interpolation techniques.,5978
"technique prediction. implemented widely and in has been geostatistics Kriging for simulation Recently, area","Recently, Kriging technique has been widely implemented in simulation area and geostatistics for prediction.",5979
consolidate been used proposed to miscellaneous the have test technique. images A,A miscellaneous test images have been used to consolidate the proposed technique.,5980
are systems an Valuation based property studied. verifying idempotent,Valuation based systems verifying an idempotent property are studied.,5981
between structure. the lattice a order A them valuations partial giving defined is,A partial order is defined between the valuations giving them a lattice structure.,5982
efficient in computations carry an way. is It to both representations studied how out with,It is studied how to carry out computations with both representations in an efficient way.,5983
of cases convex particular are sets considered. and The finite polytopes,The particular cases of finite sets and convex polytopes are considered.,5984
independence role paper explores of This in causal influence network (ICI) of inference. the Bayesian,This paper explores the role of independence of causal influence (ICI) in Bayesian network inference.,5985
smaller factorize conditional table to allows into ICI pieces. probability a one,ICI allows one to factorize a conditional probability table into smaller pieces.,5986
imprecise overwhelming information risk operator. increases of the This proliferation the of,This proliferation of imprecise information increases the risk of overwhelming the operator.,5987
at Horizon software stage. is prototype The currently,The Horizon software is currently at prototype stage.,5988
"on this Based a is indication, causality definition subjective of forwarded.","Based on this indication, a subjective definition of causality is forwarded.",5989
"discussed. Finally, claims these of accuracy the are for further methods testing","Finally, methods for further testing the accuracy of these claims are discussed.",5990
events. on to new multiple random developed probabilistic represent A relations method is,A new method is developed to represent probabilistic relations on multiple random events.,5991
epsilon-homogeneity. a We state of introduce property call which partitions we space,We introduce a property of state space partitions which we call epsilon-homogeneity.,5992
also for The explanations. competing mechanism system will ordering a need,The system will also need a mechanism for ordering competing explanations.,5993
field Image issue is critical denoising processing. in of image a the digital,Image denoising is a critical issue in the field of digital image processing.,5994
degraded considered. with Salt images Pepper noise have been Gray-level &,Gray-level images degraded with Salt & Pepper noise have been considered.,5995
basic in construction knowledge-based of rule the unit Most model the as work knowledge. takes,Most work in knowledge-based model construction takes the rule as the basic unit of knowledge.,5996
for framework interaction. and independence asymmetric Our of representation provides intercausal canonical,Our framework provides for representation of asymmetric independence and canonical intercausal interaction.,5997
illustrated awareness. The is domain situation examples military framework the from using of,The framework is illustrated using examples from the domain of military situation awareness.,5998
inference a algorithm provide and networks domains. stochastic language associated Bayesian modeling for,Bayesian networks provide a modeling language and associated inference algorithm for stochastic domains.,5999
applications. been of applied successfully medium-scale in They variety a have,They have been successfully applied in a variety of medium-scale applications.,6000
"for hierarchies. natural These encoding can a attributes be providing objects, framework part-of themselves","These attributes can themselves be objects, providing a natural framework for encoding part-of hierarchies.",6001
functions on depends is which certain An incremental cost efficient and implementation outlined criteria. satisfying,An efficient and incremental implementation is outlined which depends on cost functions satisfying certain criteria.,6002
"paper algorithm-independent novel, network proposes to This a optimizing inference. approach belief","This paper proposes a novel, algorithm-independent approach to optimizing belief network inference.",6003
We a a to way. develop models graph unifying technique believe several describing in,We believe to develop a technique describing several graph models in a unifying way.,6004
implementation a exploits efficient present multithreading. computationally of approach We that this,We present a computationally efficient implementation of this approach that exploits multithreading.,6005
The elementary Turing-computable. the of are operations procedure,The elementary operations of the procedure are Turing-computable.,6006
which functions that The procedure used procedures not is are compute to define creative Turing-computable.,The procedure is used to define creative procedures which compute functions that are not Turing-computable.,6007
fruitful of algorithmic a These results future development. suggest direction,These results suggest a fruitful direction of future algorithmic development.,6008
Planning actions of decision problems as are modeled processes. effects be Markov where can non-deterministic,Planning problems where effects of actions are non-deterministic can be modeled as Markov decision processes.,6009
Empirical studies bring about significant that have shown speedups. the techniques can,Empirical studies have shown that the techniques can bring about significant speedups.,6010
a algorithm Convolutional to MaxPooling segment images. Networks training for We fast present,We present a fast algorithm for training MaxPooling Convolutional Networks to segment images.,6011
"hand, developing the e.g. organisms, other On","On the other hand, developing organisms, e.g.",6012
"a high is, noisy error). these (that Monte Unfortunately, Carlo be can have","Unfortunately, these can be noisy (that is, have a high Monte Carlo error).",6013
consistency of the method. We our prove,We prove the consistency of our method.,6014
lineages vertebrate evolving significantly at evidence many are different rates. that suggests Growing,Growing evidence suggests that many vertebrate lineages are evolving at significantly different rates.,6015
Physiological and water basins to were northern ecological adaptations evaluated.,Physiological and ecological adaptations to northern water basins were evaluated.,6016
water populations supposed that backwaters the water Saratov's originated sweet was sweet population. from It,It was supposed that the sweet water populations originated from Saratov's backwaters sweet water population.,6017
function of minimizing problem an the that objective orthonormal matrix. on We depends an consider,We consider the problem of minimizing an objective function that depends on an orthonormal matrix.,6018
"not higher for problems. effective However, Flury dimensional is method the","However, the Flury method is not effective for higher dimensional problems.",6019
minimization alternating The is algorithm based on on manifolds.,The algorithm is based on alternating minimization on manifolds.,6020
processing a achieves on unit realtime Implemented performance. it graphics,Implemented on a graphics processing unit it achieves realtime performance.,6021
the the gene motifs distribution in The of was repeat analyzed.,The distribution of the repeat motifs in the gene was analyzed.,6022
studies penalized that same in suggest AIC from Simulation when the regression. shortcomings used suffers,Simulation studies suggest that AIC suffers from the same shortcomings when used in penalized regression.,6023
models. only literature restricted Similar set a but for of the candidate exist in results,Similar results exist in the literature but only for a restricted set of candidate models.,6024
"networks continuous Gaussian representing of continuous that consist composite values. multivariates, Vector take nodes","Vector Gaussian continuous networks consist of composite nodes representing multivariates, that take continuous values.",6025
domain networks application estimation of monitoring and are these The problems. of,The domain of application of these networks are monitoring and estimation problems.,6026
this form deriving process illustrate of filter. the decentralized We by the Kalman,We illustrate this process by deriving the decentralized form of the Kalman filter.,6027
This theory. control artificial from work and concepts modern unifies intelligence,This work unifies concepts from artificial intelligence and modern control theory.,6028
tightness of numerical by the We experiments. illustrate the bounds,We illustrate the tightness of the bounds by numerical experiments.,6029
have brute is all possible search we as The method consider force combinations. to,The search is brute force method as we have to consider all possible combinations.,6030
are chain Markov estimated Carlo of using the Parameters Monte algorithm. a model,Parameters of the model are estimated using a Markov chain Monte Carlo algorithm.,6031
weighting likelihood These methods variants known of are algorithm. the new,These new methods are variants of the known likelihood weighting algorithm.,6032
demonstrates empirical algorithms We the evaluation improved an which of their present performance.,We present an empirical evaluation of the algorithms which demonstrates their improved performance.,6033
is lead may errors model This to the used. when inference,This may lead to inference errors when the model is used.,6034
"are Inference for algorithms for arbitrary belief impractical belief networks networks. complex large,","Inference algorithms for arbitrary belief networks are impractical for large, complex belief networks.",6035
algorithms more to networks shown efficient. been of be Inference for have specialized classes belief,Inference algorithms for specialized classes of belief networks have been shown to be more efficient.,6036
"We trees, their procedure define and binary describe a for join constructing sketch utility, them.","We define binary join trees, describe their utility, and sketch a procedure for constructing them.",6037
"Uncertainty conclusions, their characterize inferences, all their may to or be taken premises three.","Uncertainty may be taken to characterize inferences, their conclusions, their premises or all three.",6038
"characterized never itself by some is uncertainty. of Under inferences the treatments uncertainty,","Under some treatments of uncertainty, the inferences itself is never characterized by uncertainty.",6039
arbitrary using networks The to standard extends conditioning algorithm Bayesian techniques.,The algorithm extends to arbitrary Bayesian networks using standard conditioning techniques.,6040
algorithm practicality. empirically demonstrate We evaluate enumeration its the and,We empirically evaluate the enumeration algorithm and demonstrate its practicality.,6041
states. As (final) of consequences of the utility the a are of specification part,As part of the consequences are a specification of the utility of (final) states.,6042
"programming language. the similar plans, adopts The GOLOG robot to robot","The robot adopts robot plans, similar to the GOLOG programming language.",6043
problem the the expected is find The plan ?planning' to utility. with highest,The ?planning' problem is to find the plan with the highest expected utility.,6044
the inputs in discusses possibility paper revision under of This framework belief uncertain theory.,This paper discusses belief revision under uncertain inputs in the framework of possibility theory.,6045
the distribution. networks qualitatively provide language a representing Bayesian of for properties independence conditional a,Bayesian networks provide a language for qualitatively representing the conditional independence properties of a distribution.,6046
be can to representation suggest this used support We ways algorithms. effective inference which in,We suggest ways in which this representation can be used to support effective inference algorithms.,6047
problems domain. analyze diagnosis of coming the structural the from empirically circuit properties We,We analyze empirically the structural properties of problems coming from the circuit diagnosis domain.,6048
"an bounds, have Deterministic probability runtime. usually but exponential algorithms provide","Deterministic algorithms usually provide probability bounds, but have an exponential runtime.",6049
"schemes only estimates. probability provide runtime, Some but polynomial have randomized a","Some randomized schemes have a polynomial runtime, but provide only probability estimates.",6050
"resulting instantiations, bounds. randomized present algorithms We that in partial enumerate probability high-probability","We present randomized algorithms that enumerate high-probability partial instantiations, resulting in probability bounds.",6051
algorithms sampling Some of these also algorithms. are,Some of these algorithms are also sampling algorithms.,6052
through package available that A plans is ftp. constructs Quasi-Bayesian anonymous,A package that constructs Quasi-Bayesian plans is available through anonymous ftp.,6053
to m-ary We extend Bayes this with (hyperplane result observations. networks Naive separability),We extend this result (hyperplane separability) to Naive Bayes networks with m-ary observations.,6054
"the In effect decision addition, observation-observation of on dependencies we surfaces. illustrate","In addition, we illustrate the effect of observation-observation dependencies on decision surfaces.",6055
"in knowledge results research the and we of discuss implications these Finally, learning. on acquisition","Finally, we discuss the implications of these results on knowledge acquisition and research in learning.",6056
Axiomatization been used logical implications. for has testing widely,Axiomatization has been widely used for testing logical implications.,6057
of probability specify hypothesis We the the use concept structure. to,We use the concept of hypothesis to specify the probability structure.,6058
We information consider case available. probabilistic which the in is not,We consider the case in which probabilistic information is not available.,6059
"of using of the of specification we disjunctions. allow intervals, probability probabilities Instead the for","Instead of using probability intervals, we allow for the specification of the probabilities of disjunctions.",6060
goal of is default probability An algorithm a calculating the presented. for,An algorithm for calculating the default probability of a goal is presented.,6061
algorithms flexible diagrams. solving as report on work decision for influence towards problems We represented,We report on work towards flexible algorithms for solving decision problems represented as influence diagrams.,6062
and constructed incrementally. tree a is represents decision function Each,Each tree represents a decision function and is constructed incrementally.,6063
"paper present fundamental this its and components, definitions model, properties. the of its we In","In this paper we present definitions of the model, its components, and its fundamental properties.",6064
"has that noise asymmetric, we But, show logodds-normal modest effects. even","But, we show that even asymmetric, logodds-normal noise has modest effects.",6065
by algorithm the difficulty The remedies relaxing pruning (Stochastic Dominance A*) SDA* this condition.,The SDA* (Stochastic Dominance A*) algorithm remedies this difficulty by relaxing the pruning condition.,6066
employment equipment planning manages sets order that actions. of The constraints feasible agent,The planning agent manages constraints that order sets of feasible equipment employment actions.,6067
diagram. influence mapped onto abstraction an level an of intermediate is at This,This is mapped at an intermediate level of abstraction onto an influence diagram.,6068
a this optimum equivalent of simultaneous evaluation of to plans. range is Finding the,Finding this optimum is equivalent to the simultaneous evaluation of a range of plans.,6069
The area and of study been philosophy active change has in an belief AI.,The study of belief change has been an active area in philosophy and AI.,6070
inconsistencies One measured approach use in to techniques is knowledge data. to based detect,One approach is to use knowledge based techniques to detect inconsistencies in measured data.,6071
such the a presents inconsistencies. probabilistic detection for This of article model,This article presents a probabilistic model for the detection of such inconsistencies.,6072
to by applying it model. a plant is power proposed The approach exemplified,The proposed approach is exemplified by applying it to a power plant model.,6073
graphs. sparse The algorithm polynomial is on,The algorithm is polynomial on sparse graphs.,6074
a to network rapid prototyping propose approach We engineering.,We propose a rapid prototyping approach to network engineering.,6075
complex Methods discussed. network belief for are models evaluating,Methods for evaluating complex belief network models are discussed.,6076
system. A power supply faulty restoration a typical is in case distribution,A typical case is supply restoration in a faulty power distribution system.,6077
"time, decision incomplete multiple This information over and opportunities stochasticity, inaccurate involves knowledge.","This involves multiple decision opportunities over time, stochasticity, inaccurate information and incomplete knowledge.",6078
generalizations of is presented. nonmonotonic thresholded A logic,A nonmonotonic logic of thresholded generalizations is presented.,6079
elimination and way combining present a We conditioning framework. also within this of general,We also present a general way of combining conditioning and elimination within this framework.,6080
characterized for of Additional are as limitations well. lower probabilities classes other,Additional limitations are characterized for other classes of lower probabilities as well.,6081
are simply argued of is It a phenomena matter these that interpretation.,It is argued that these phenomena are simply a matter of interpretation.,6082
The and interpretation introduced is discussed. ontological,The ontological interpretation is introduced and discussed.,6083
transitions. temporally to probability -dependent assign functions state We,We assign temporally -dependent probability functions to state transitions.,6084
flight model simulation that CIRCA performance. tests We present demonstrate probabilistic our improve may how,We present flight simulation tests that demonstrate how our probabilistic model may improve CIRCA performance.,6085
framework a for actions. abstraction provides This formal practical of and sound probabilistic paper,This paper provides a formal and practical framework for sound abstraction of probabilistic actions.,6086
"types this sound Within procedures of action we for framework, three abstraction present abstraction.","Within this framework, we present sound abstraction procedures for three types of action abstraction.",6087
approximate on related abstraction. end paper work discussions and with the sound and We,We end the paper with discussions and related work on sound and approximate abstraction.,6088
knowledge shell expert SPIRIT bases. system for an is probabilistic,SPIRIT is an expert system shell for probabilistic knowledge bases.,6089
show the system. sized the of power applications Medium,Medium sized applications show the power of the system.,6090
extend develop We troubleshooting methods nonfunctioning a and existing device. for decision-theoretic,We develop and extend existing decision-theoretic methods for troubleshooting a nonfunctioning device.,6091
"this include actions. we to paradigm In this extend paper, taking","In this paper, we extend this paradigm to include taking actions.",6092
have this largest were lived on terrestrial Sauropoda to planet. animals The the ever,The Sauropoda were the largest terrestrial animals ever to have lived on this planet.,6093
"musculature the of investigated. has yet to the neck However, of the be effect","However, the effect of the musculature of the neck has yet to be investigated.",6094
experiment carried psychophysical out paradigm. We using masking,We carried out psychophysical experiment using masking paradigm.,6095
formalize function a of k-d darts. We the using notion sampling,We formalize the notion of sampling a function using k-d darts.,6096
"mutually k-d called A orthogonal, k-dimensional subspaces k-d of set independent, is flats. a dart","A k-d dart is a set of independent, mutually orthogonal, k-dimensional subspaces called k-d flats.",6097
"dart d Each for k with choose aligned axes flats, has the coordinate efficiency.","Each dart has d choose k flats, aligned with the coordinate axes for efficiency.",6098
for higher the of a accuracy also estimation volume demonstrate darts problem. dimensional We,We also demonstrate the accuracy of higher dimensional darts for a volume estimation problem.,6099
"d-separation, theory introduce relational independence deriving conditional a from We models. for facts relational","We introduce relational d-separation, a theory for deriving conditional independence facts from relational models.",6100
methods. present We with analysis for image registration performance gradient descent a,We present a performance analysis for image registration with gradient descent methods.,6101
multiclass algorithms We graph-based high-dimensional present of for data. segmentation two,We present two graph-based algorithms for multiclass segmentation of high-dimensional data.,6102
a the numerical scheme. convex minimizes algorithm functional splitting The first using,The first algorithm minimizes the functional using a convex splitting numerical scheme.,6103
"the The which Chao's theory as lower retrieve estimate estimator, species richness. of we generalizes","The theory generalizes Chao's estimator, which we retrieve as the lower estimate of species richness.",6104
is in recognition proposed report. this Iris,Iris recognition is proposed in this report.,6105
(CM) matrices FC. for LLD by were higher cross-correlation characterized Raw,Raw cross-correlation matrices (CM) for LLD were characterized by higher FC.,6106
measures network Topological between showed groups. no significant differences,Topological network measures showed no significant differences between groups.,6107
"addition, in LLD. caudate right was diffuse connectivity In more nucleus","In addition, right caudate nucleus connectivity was more diffuse in LLD.",6108
"paper studies MSBNs, presented propagation. with originally UpdateBelief, for operation inter-subnet the This","This paper studies the operation UpdateBelief, presented originally with MSBNs, for inter-subnet propagation.",6109
available statistical allow is to reliable not Often estimation. data probability enough for,Often not enough statistical data is available to allow for reliable probability estimation.,6110
the amenable encoding be Available may not network. directly for information in,Available information may not be directly amenable for encoding in the network.,6111
"probabilities. numerical reluctant provide Finally, be domain to experts may","Finally, domain experts may be reluctant to provide numerical probabilities.",6112
over then desired second-order this form probabilities. probability to derive use We distributions the canonical,We then use this canonical form to derive second-order probability distributions over the desired probabilities.,6113
exact networks. for algorithms present in two inference causal approximate We and,We present two algorithms for exact and approximate inference in causal networks.,6114
properties proposed also illustrating the some experimental results of present We algorithms. the,We also present some experimental results illustrating the properties of the proposed algorithms.,6115
been a Consider network. to evidence Bayesian has e entered the situation where some,Consider the situation where some evidence e has been entered to a Bayesian network.,6116
architecture. propagation HUGIN is into a propagation Cautious Shafer-Shenoy-like a of modification,Cautious propagation is a modification of HUGIN propagation into a Shafer-Shenoy-like architecture.,6117
of networks to the belief application diagnosis of systems. an in bottlenecks We describe computer,We describe an application of belief networks to the diagnosis of bottlenecks in computer systems.,6118
"characterized are distributions. Uncertainty and with workloads, values Gaussian predictions, counter in","Uncertainty in workloads, predictions, and counter values are characterized with Gaussian distributions.",6119
presented. in results are Initial bottlenecks diagnosing,Initial results in diagnosing bottlenecks are presented.,6120
sound but always in conservative Predict that complete. it a exhibits behavior not always is,Predict exhibits a conservative behavior in that it is always sound but not always complete.,6121
for develops paper order a reasoning. calculus This simple magnitude of,This paper develops a simple calculus for order of magnitude reasoning.,6122
completeness with A is semantics soundness and results. given,A semantics is given with soundness and completeness results.,6123
through cut justify ideas analysis. limit our We theoretically,We theoretically justify our ideas through limit cut analysis.,6124
A in benefits technique these the of key graphical idea achieving modeling. is,A key technique in achieving these benefits is the idea of graphical modeling.,6125
effective we implications algorithms. What how has we learn for it important and learn,What we learn and how we learn it has important implications for effective algorithms.,6126
analyzing reasoning. already useful proved in default Plausibility measures have,Plausibility measures have already proved useful in analyzing default reasoning.,6127
hypothesis and selection. We discuss testing model,We discuss hypothesis testing and model selection.,6128
almost method this One such is problem directly with that never are available. priors,One problem with this method is that such priors are almost never directly available.,6129
probabilistic of second is diagnosis facing model-based persistence. A modeling the problem,A second problem facing probabilistic model-based diagnosis is the modeling of persistence.,6130
"problems using from In paper, techniques these we this theory. Reliability address","In this paper, we address these problems using techniques from Reliability theory.",6131
scheme observations. a also We multiple when model handling time develop persistence tagged to,We also develop a scheme to model persistence when handling multiple time tagged observations.,6132
"general a We for metric, derive scoring Bayesian both appropriate domains.","We derive a general Bayesian scoring metric, appropriate for both domains.",6133
uncertainty. under we extend diagram decisions for this (ID) In paper the representation influence,In this paper we extend the influence diagram (ID) representation for decisions under uncertainty.,6134
constraints represent often diagrams by decision influence into nodes. want to Users of arrows,Users of influence diagrams often want to represent constraints by arrows into decision nodes.,6135
arrows decision relevance constraints nodes. We represent on by into decisions allowing,We represent constraints on decisions by allowing relevance arrows into decision nodes.,6136
(IRIDs). diagrams the representation call We information/relevance resulting influence,We call the resulting representation information/relevance influence diagrams (IRIDs).,6137
allow influence for decisions. direct diagrams representation of specification and Information/relevance constrained,Information/relevance influence diagrams allow for direct representation and specification of constrained decisions.,6138
We combination solve use IRIDs. to dynamic programming stochastic a and Gibbs of sampling,We use a combination of stochastic dynamic programming and Gibbs sampling to solve IRIDs.,6139
exact useful This is solving method for fail. IDs when methods especially,This method is especially useful when exact methods for solving IDs fail.,6140
transformations. We on present based equivalent Bayesian simple network structures local of a characterization,We present a simple characterization of equivalent Bayesian network structures based on local transformations.,6141
is the characterization The significance of twofold.,The significance of the characterization is twofold.,6142
"the of potential costs, uncertain We consider costs. among with probabilistic dependencies problem edge the","We consider the problem of uncertain edge costs, with potential probabilistic dependencies among the costs.",6143
apply debt the of problem fraud/uncollectible detection telecommunication to Bayesian services. models network We for,We apply Bayesian network models to the problem of fraud/uncollectible debt detection for telecommunication services.,6144
"C queries been true?"") ""If would true, have of counterfactual were A (e.g., Evaluation","Evaluation of counterfactual queries (e.g., ""If A were true, would C have been true?"")",6145
"planning, of and is diagnosis, determination important liability, analysis. to policy fault","is important to fault diagnosis, planning, determination of liability, and policy analysis.",6146
new taxa require taxa await description. revision and Older,Older taxa require revision and new taxa await description.,6147
mechanically their The evolution long sauropods obscure. are and necks characteristic of perplexing is,The characteristic long necks of sauropods are mechanically perplexing and their evolution is obscure.,6148
holotype characters. reveals Xenoposeidon six of preservation unique The the excellent,The excellent preservation of the Xenoposeidon holotype reveals six unique characters.,6149
"is unique particularly ilium unusual, The five exhibiting features.","The ilium is particularly unusual, exhibiting five unique features.",6150
networks. representing probability define Bayesian classes for discrete-time We a context-sensitive logic of temporal temporal,We define a context-sensitive temporal probability logic for representing classes of discrete-time temporal Bayesian networks.,6151
We for provide a language. declarative semantics our,We provide a declarative semantics for our language.,6152
approach. programming We use related concepts logic our in to justify,We use related concepts in logic programming to justify our approach.,6153
value method based on is theory. The extreme,The method is based on extreme value theory.,6154
for in Implications conclusion. are discussed the learning,Implications for learning are discussed in the conclusion.,6155
"represent networks Bayesian causal relationships. Bayesian networks acausal independence, probabilistic causal Whereas represent","Whereas acausal Bayesian networks represent probabilistic independence, causal Bayesian networks represent causal relationships.",6156
"this we examine paper, learning In types both for Bayesian networks. methods of","In this paper, we examine Bayesian methods for learning both types of networks.",6157
networks fairly developed. methods well Bayesian acausal learning are for,Bayesian methods for learning acausal networks are fairly well developed.,6158
"{em introduce sufficient assumptions, and component called independence}. mechanism {em We independence} two","We introduce two sufficient assumptions, called {em mechanism independence} and {em component independence}.",6159
"filtering. sigma-point novel to In this application methods continuous-discrete paper, a of describe we","In this paper, we describe a novel application of sigma-point methods to continuous-discrete filtering.",6160
"continuous- exactly. principle, can filtering In the problem nonlinear discrete solved be","In principle, the nonlinear continuous- discrete filtering problem can be solved exactly.",6161
"contains the terms practice, In intractible. computationally that solution are","In practice, the solution contains terms that are computationally intractible.",6162
demonstrate problems. cope with better such to is method our equipped that We,We demonstrate that our method is better equipped to cope with such problems.,6163
"obtained simultaneously. identification the Furthermore, be parameters shall","Furthermore, the identification parameters shall be obtained simultaneously.",6164
proposed be shall for an classification to The method applied example.,The proposed method shall be applied for classification to an example.,6165
of plants. be Iris {em three for Database} learnt all The shall Plant kinds,The {em Iris Plant Database} shall be learnt for all three kinds of plants.,6166
semantic in environment. research Ontology the technique of current in emerging an field is,Ontology is an emerging technique in the current field of research in semantic environment.,6167
"decision cost. strategy low repair framework, has good expected theoretic a In a","In a decision theoretic framework, a good repair strategy has low expected cost.",6168
are components failures system (b) independent. The of the,(b) The failures of the system components are independent.,6169
of functions properties acceptance General estabilished. are,General properties of acceptance functions are estabilished.,6170
planning. decision-theoretic techniques discusses paper efficient This performing for,This paper discusses techniques for performing efficient decision-theoretic planning.,6171
"method. method game efficient more than pruning the the rollback is For trees,","For game trees, the pruning method is more efficient than the rollback method.",6172
"the Also encoding examine graphical paper, in of relationships. this causal we","Also in this paper, we examine the graphical encoding of causal relationships.",6173
"causal Pearl's a establish In canonical addition, we form correspondence theory. between and","In addition, we establish a correspondence between canonical form and Pearl's causal theory.",6174
kind equation DAG structural model. one models of Recursive are,Recursive structural equation models are one kind of DAG model.,6175
belief a increasingly reasoning. networks diagnostic used bing are knowledge representation for as Bayesian,Bayesian belief networks are bing increasingly used as a knowledge representation for diagnostic reasoning.,6176
and only. observations conducting One represent system faults method to simple for reasoning diagnostic is,One simple method for conducting diagnostic reasoning is to represent system faults and observations only.,6177
"general, precomputing is policies optimal intractable. repair In","In general, precomputing optimal repair policies is intractable.",6178
is precomputation for algorithm of strategy. suitable an hierarchical repair The optimal off-line repair,The hierarchical repair algorithm is suitable for off-line precomputation of an optimal repair strategy.,6179
are vision. computer important Object problems in and detection recognition,Object detection and recognition are important problems in computer vision.,6180
"generalization properties. provides new which a good propose discriminative and We codebook, also hierarchical","We also propose a new hierarchical codebook, which provides good generalization and discriminative properties.",6181
learning. It the of capability dynamic has,It has the capability of dynamic learning.,6182
of been geometric work The has detecting images real shapes in completed. preliminary,The preliminary work of detecting geometric shapes in real images has been completed.,6183
of is the this work This focus report. preliminary,This preliminary work is the focus of this report.,6184
for detection/recognition also in is realizing object path brief. Future discussed method proposed the,Future path for realizing the proposed object detection/recognition method is also discussed in brief.,6185
complete missing solutions provides for the model. analytic component to the author's the and,and provides the missing component for the complete analytic solutions to the author's model.,6186
possible when experimentation (i.e. compared plans needed uncertainties are is resolving for Efficient,Efficient experimentation is needed for resolving uncertainties when possible plans are compared (i.e.,6187
sufficient a locally selecting with should optimal statistical significance experimentation for be (i.e. plan The,The experimentation should be sufficient for selecting with statistical significance a locally optimal plan (i.e.,6188
vastly that An incomplete models presented. is exploits to fault approach isolation,An approach to fault isolation that exploits vastly incomplete models is presented.,6189
A example illustrates of the realistic approach. benefits this,A realistic example illustrates the benefits of this approach.,6190
in systems the of so-called issue important problem. use expert An the is brittleness,An important issue in the use of expert systems is the so-called brittleness problem.,6191
model a world. of part the only limited Expert systems,Expert systems model only a limited part of the world.,6192
"is this model In represented relational as generalized the database. framework, probability a","In this framework, the probability model is represented as a generalized relational database.",6193
standard relational Subsequent processed can requests be queries. probabilistic as,Subsequent probabilistic requests can be processed as standard relational queries.,6194
systems be reasoning system. Conventional easily database approximate management for adopted can an implementing such,Conventional database management systems can be easily adopted for implementing such an approximate reasoning system.,6195
to an algorithm the give transform. We implement,We give an algorithm to implement the transform.,6196
"example, to Default represent prioritization inheritance, adequately. typically requires for specificity non-layered","Default inheritance, for example, typically requires non-layered prioritization to represent specificity adequately.",6197
logic stratified with negation-as-failure. : with programs,: with stratified logic programs with negation-as-failure.,6198
known programs previously logic equivalently Such representable circumscriptions. be predicate layered-priority to are as,Such logic programs are previously known to be representable equivalently as layered-priority predicate circumscriptions.,6199
"number the defaults the of Worst-case, exponentially. increases transform","Worst-case, the transform increases the number of defaults exponentially.",6200
is expressive special case non-?top-heaviness? such One,One such expressive special case is non-?top-heaviness?,6201
discuss other the fundamental result. aspects significance of of the We,We discuss other aspects of the significance of the fundamental result.,6202
Brewka's with them prioritization-type This an to to equipping alternative interesting provides approach precedence.,This provides an interesting alternative to Brewka's approach to equipping them with prioritization-type precedence.,6203
reasoning diagnostic potential great for networks automating Bayesian large use in tasks. scale offer,Bayesian networks offer great potential for use in automating large scale diagnostic reasoning tasks.,6204
excessive sampling take sample. can an to representative Gibbs Unfortunately a generate time,Unfortunately Gibbs sampling can take an excessive time to generate a representative sample.,6205
include techniques sampling. strategies Carlo other Gibbs chain sampling Monte Markov The than,The strategies include Monte Carlo Markov chain sampling techniques other than Gibbs sampling.,6206
distributed Emphasis be put that can is in implemented strategies on systems.,Emphasis is put on strategies that can be implemented in distributed systems.,6207
this of convex of sets probabilities. independence study concepts paper we different for In,In this paper we study different concepts of independence for convex sets of probabilities.,6208
for independence. two be will basic There ideas,There will be two basic ideas for independence.,6209
some in Theories insect accelerated and scenarios. diversification hostplant of dynamics predict,Theories of insect and hostplant dynamics predict accelerated diversification in some scenarios.,6210
affects that phylogenetic greatly diversification rate estimates. uncertainty found We,We found that phylogenetic uncertainty greatly affects diversification rate estimates.,6211
in shift for diversification not to A Danaina. the is Apocynaceae responsible of slowdown,A shift to Apocynaceae is not responsible for the slowdown of diversification in Danaina.,6212
"based paper, characterization. descriptors shape this directional new derive on we a In","In this paper, we derive new shape descriptors based on a directional characterization.",6213
"respect and translation scaling. with invariant description to a obtain rotation, We reflection,","We obtain a description invariant with respect to rotation, reflection, translation and scaling.",6214
proposed on well-defined space. associated the then is metric A feature,A well-defined metric is then proposed on the associated feature space.,6215
this of show We metric. continuity the,We show the continuity of this metric.,6216
benchmarks. here present such experiment We reliability of that the a addresses psychophysical,We here present a psychophysical experiment that addresses the reliability of such benchmarks.,6217
single generate A targets. plenty synaptic connecting can branches with axon,A single axon can generate branches connecting with plenty synaptic targets.,6218
connections Process very system. is important nervous making in of branching for central,Process of branching is very important for making connections in central nervous system.,6219
primary interstitial occurs system shaft axon branching nervous along development. The during,The interstitial branching along primary axon shaft occurs during nervous system development.,6220
and its Growing its movement leaves behind terminal. active makes points axon in pause,Growing axon makes pause in its movement and leaves active points behind its terminal.,6221
points. new these appear The branches from,The new branches appear from these points.,6222
branching network mathematical process. describe to and neural We suggest investigate model,We suggest mathematical model to describe and investigate neural network branching process.,6223
branching interstitial We from the model shaft. axon,We model the interstitial branching from axon shaft.,6224
axonal networks the are Numerical to that model simulations neural network. framework similar show in,Numerical simulations show that in the model framework axonal networks are similar to neural network.,6225
Face popular video is recognition in from biometrics or research. image a topic,Face recognition from image or video is a popular topic in biometrics research.,6226
advantages actual of other over and face based The identification uniqueness are acceptance. biometrics,The actual advantages of face based identification over other biometrics are uniqueness and acceptance.,6227
"In of accuracy identification main and this a speed field, is issue.","In this field, accuracy and speed of identification is a main issue.",6228
Monte samplers. difficulty Hamiltonian-based in this In the address Carlo widely-experienced paper tuning we,In this paper we address the widely-experienced difficulty in tuning Hamiltonian-based Monte Carlo samplers.,6229
content based retrieval the image of (CBIR). one is features Shape main in,Shape is one of the main features in content based image retrieval (CBIR).,6230
proposes signature. This a shape paper new,This paper proposes a new shape signature.,6231
"to based technique is and it is rotation. proposed scaling translation, and Fourier invariant The","The proposed technique is Fourier based and it is invariant to translation, scaling and rotation.",6232
are compared performance FSD with signatures. better those Experimental has that the shown results signature,Experimental results are shown that the FSD signature has better performance compared with those signatures.,6233
"structure Moreover, core interactome around modular mixed Sirtuin the showed sub comprising functions. the a","Moreover, the Sirtuin sub interactome showed a modular structure around the core comprising mixed functions.",6234
"clustering, function. Keywords: fuzzy algorithm numerical Cobweb, data, membership","Keywords: clustering, algorithm Cobweb, numerical data, fuzzy membership function.",6235
impact spatial the Yet heterogeneity this signaling biochemical is poorly processes understood. on of,Yet the impact of this spatial heterogeneity on biochemical signaling processes is poorly understood.,6236
that demonstrate the biochemical Here improves reliability of signaling. we partitioning,Here we demonstrate that partitioning improves the reliability of biochemical signaling.,6237
in model a signaling. ubiquitous We describing stochastic a solve membrane motif exactly,We exactly solve a stochastic model describing a ubiquitous motif in membrane signaling.,6238
quantitatively size predicted optimal The observed agrees experimentally with partition systems.,The predicted optimal partition size agrees quantitatively with experimentally observed systems.,6239
barriers. spatial These diffusion simulations explicit in results with persist,These results persist in spatial simulations with explicit diffusion barriers.,6240
"the often modelling solution this, For proposed. agent-based a is approach","For this, the agent-based approach is a modelling solution often proposed.",6241
A of pedagogical system illustration watering a smart is of then fuzzy proposed. agentification,A pedagogical illustration of fuzzy agentification of a smart watering system is then proposed.,6242
that Experiments most of have shown is schemes. one simulation likelihood the weighting promising,Experiments have shown that likelihood weighting is one of the most promising simulation schemes.,6243
"actions, networks. representation a for on states We probabilistic develop based belief and","We develop a probabilistic representation for states and actions, based on belief networks.",6244
incorporating called Maintenance are TMSs of Systems uncertainty measure Belief (BMSs).,TMSs incorporating measure of uncertainty are called Belief Maintenance Systems (BMSs).,6245
in A represent to spatial this model information is paper. presented,A model to represent spatial information is presented in this paper.,6246
"case viewed of recognition recognition. this pattern a can context, In fuzzy as template be","In this context, template recognition can be viewed as a case of fuzzy pattern recognition.",6247
"possible extensions. In we going therefore look are for alternatives and to paper, this","In this paper, we are therefore going to look for possible alternatives and extensions.",6248
"and generalize * Ranking measures. possibility which measures, Spohn-type","* Ranking measures, which generalize Spohn-type and possibility measures.",6249
large and extremely is for practical The approach models. efficient,The approach is efficient and practical for extremely large models.,6250
learned with Sparse image in have applications. representations analysis been several successful dictionaries,Sparse representations with learned dictionaries have been successful in several image analysis applications.,6251
formulated decision of as an present the We problems approach to diagrams. influence solution,We present an approach to the solution of decision problems formulated as influence diagrams.,6252
possibility propose integration We an into logics. theory of non-classical,We propose an integration of possibility theory into non-classical logics.,6253
call planning an we high-level approach conditional to introduce planning. epsilon-safe We,We introduce an approach to high-level conditional planning we call epsilon-safe planning.,6254
for describe algorithms on epsilon-safe based several planners. planning conditional We,We describe several algorithms for epsilon-safe planning based on conditional planners.,6255
is An hepatitis diagnosis paper. throughout of used example the,An example of hepatitis diagnosis is used throughout the paper.,6256
gene highly . Extensive binding with in factor expression divergence Drosophila of embryos conserved transcription,Extensive divergence of transcription factor binding in Drosophila embryos with highly conserved gene expression .,6257
"can Therefore, significantly. cascade the matching speed the improve","Therefore, the cascade can improve the matching speed significantly.",6258
learned is two cascade by steps. The,The cascade is learned by two steps.,6259
"nested features are of grouped some prepared Firstly, stages. several into kind","Firstly, some kind of prepared features are grouped into several nested stages.",6260
"with we results. a approximation Finally, present preliminary novel experimental technique","Finally, we present a novel approximation technique with preliminary experimental results.",6261
probabilistic paper the actions. discusses of This problem abstracting conditional,This paper discusses the problem of abstracting conditional probabilistic actions.,6262
intra-action identify inter-action abstraction of We distinct abstraction. and types two abstraction:,We identify two distinct types of abstraction: intra-action abstraction and inter-action abstraction.,6263
"an decision problem diagram, influence be symmetrized. To represented asymmetric an be in must","To be represented in an influence diagram, an asymmetric decision problem must be symmetrized.",6264
handling complicating plan are Techniques for features. of presented types various,Techniques are presented for handling various types of complicating plan features.,6265
propose we for method possibilistic proof Therefore alternative logic. an,Therefore we propose an alternative proof method for possibilistic logic.,6266
the a belief-network results We of showing inference. causal-independence present representation empirical benefits for,We present empirical results showing the benefits of a causal-independence representation for belief-network inference.,6267
algorithm propagation a also for networks. present We such,We also present a propagation algorithm for such networks.,6268
in the of is This method molecular biological structure. determination useful,This method is useful in the determination of biological molecular structure.,6269
for suited implementation. particularly is algorithm new parallel The,The new algorithm is particularly suited for parallel implementation.,6270
components: have a scoring The procedure. algorithms two a search metric and,The algorithms have two components: a scoring metric and a search procedure.,6271
metric. The evaluation procedure by scoring for networks search the generates,The search procedure generates networks for evaluation by the scoring metric.,6272
"a for algorithms. Bayesian-network we describe evaluating learning Third, methodology","Third, we describe a methodology for evaluating Bayesian-network learning algorithms.",6273
apply We metrics comparison approach this of a to search and procedures.,We apply this approach to a comparison of metrics and search procedures.,6274
develop We for refining network's approach structure. a the new,We develop a new approach for refining the network's structure.,6275
demonstrating approach. the present our of also experimental We effectiveness evidence,We also present experimental evidence demonstrating the effectiveness of our approach.,6276
"two an distances derived BoP nodes application, are probabilities. between families the from of As","As an application, two families of distances between nodes are derived from the BoP probabilities.",6277
"family and the interpolates second Interestingly, path between distance distance. distance the the shortest resistance","Interestingly, the second distance family interpolates between the shortest path distance and the resistance distance.",6278
algorithm which of utility. conditional present expected plans maximum computes We an,We present an algorithm which computes conditional plans of maximum expected utility.,6279
represent Non-singleton probability can ordered of beliefs. partially distributions sets,Non-singleton sets of probability distributions can represent partially ordered beliefs.,6280
It shows networks. results to simple this from applying metric counterintuitive,It shows counterintuitive results from applying this metric to simple networks.,6281
equivalent family One priors is suggested to equal of noninformative for assigning scores networks.,One family of noninformative priors is suggested for assigning equal scores to equivalent networks.,6282
is in studied independence view framework the An theory. ordinal possibility of of,An ordinal view of independence is studied in the framework of possibility theory.,6283
"definitions We possible of strength. three of dependence, increasing investigate","We investigate three possible definitions of dependence, of increasing strength.",6284
study applications. diagnosis between reasoning We and kappa calculus the connection probabilistic in,We study the connection between kappa calculus and probabilistic reasoning in diagnosis applications.,6285
their above. But done little connection been utilizing as outlined on has,But little has been done on utilizing their connection as outlined above.,6286
relation in this paper assumption when is investigate this satisfied. this experiments The not,The experiments in this paper investigate this relation when this assumption is not satisfied.,6287
couple of independence) closure is CI-statements a CI-model. stochastic semigraphoid every The of (GI=conditional,The semigraphoid closure of every couple of CI-statements (GI=conditional independence) is a stochastic CI-model.,6288
justifies as of the reasoning. CI-models use semigraphoids stochastic in of approximations This probabilistic,This justifies the use of semigraphoids as approximations of stochastic CI-models in probabilistic reasoning.,6289
corresponds an is It (each source assumption). the ATMS framework in formalized to,It is formalized in the ATMS framework (each source corresponds to an assumption).,6290
"a assume are that then with all probability. small We and sources independent ""fail"" very","We assume then that all sources are independent and ""fail"" with a very small probability.",6291
consequence compare these relations. and We study,We study and compare these consequence relations.,6292
of bases considered. briefly is The prioritized knowledge -case,The -case of prioritized knowledge bases is briefly considered.,6293
"can this represented influence causal an be dependence Using diagram. within we definition, show how","Using this definition, we show how causal dependence can be represented within an influence diagram.",6294
"Besides and the probabilistic implicitly constraints guarantee relationships, probabilistic terminological inferring consistency. existent","Besides inferring implicitly existent probabilistic relationships, the constraints guarantee terminological and probabilistic consistency.",6295
"methods commonly However, effects important many used test these cases. in to poorly predict perform","However, many commonly used methods to predict these effects perform poorly in important test cases.",6296
is information allowing base But knowledge sometimes to a limiting. contain objective only,But allowing a knowledge base to contain only objective information is sometimes limiting.,6297
"on All as such cross-entropy. bloused are techniques approaches, our well-known of","All of our techniques are bloused on well-known approaches, such as cross-entropy.",6298
"program, and influence of the performance the the We it. examine factors that","We examine the performance of the program, and the factors that influence it.",6299
networks approximate probabilistic on an anytime We evaluation this of present based idea. for procedure,We present an anytime procedure for approximate evaluation of probabilistic networks based on this idea.,6300
be the We correct. algorithm prove to,We prove the algorithm to be correct.,6301
by already proposed scheme out local that It turns our is Shenoy. computation,It turns out that our local computation scheme is already proposed by Shenoy.,6302
"are different that completely of from intuitions However, our Shenoy.","However, our intuitions are completely different from that of Shenoy.",6303
networks as belief Bayesian (IB) assignments abductive proposed Independence-based were originally to explanations.,Independence-based (IB) assignments to Bayesian belief networks were originally proposed as abductive explanations.,6304
We networks. marginal approximate use Bayesian belief to in probabilities assignments IB,We use IB assignments to approximate marginal probabilities in Bayesian belief networks.,6305
"obtained be efficiently. approximation fewer sufficient, more are assignments can a IB good and Thus,","Thus, fewer IB assignments are sufficient, and a good approximation can be obtained more efficiently.",6306
feasible is highly networks. approach this belief for that connected results Experimental show,Experimental results show that this approach is feasible for highly connected belief networks.,6307
thinking Some to and an require instances build of test hypothetical creative agent theories.,Some instances of creative thinking require an agent to build and test hypothetical theories.,6308
"have true, true?"") counterfactual A queries were C been would (e.g., Evaluation of ""If","Evaluation of counterfactual queries (e.g., ""If A were true, would C have been true?"")",6309
"planning, and important diagnosis, of fault determination liability. is to","is important to fault diagnosis, planning, and determination of liability.",6310
"some must of solve fundamental we reject the To theory. problem, postulates this","To solve this problem, we must reject some fundamental postulates of the theory.",6311
variables. two networks primitives and add networks: persistent controllable Action probabilistic variables causal to,Action networks add two primitives to probabilistic causal networks: controllable variables and persistent variables.,6312
recent paper progress. This results in both work and describes,This paper describes both recent results and work in progress.,6313
"are The these investigated, independence and non-interactivity are links between properties given. and of relations","The links between independence and non-interactivity are investigated, and properties of these relations are given.",6314
three to contributes topics. these This paper all,This paper contributes to all these three topics.,6315
with belief networks that minimallity It certain shown is properties selecting is NP-hard.,It is shown that selecting belief networks with certain minimallity properties is NP-hard.,6316
in We the efficiently be heuristic. search by it can implemented argue incorporating smoothing that,We argue that smoothing can be efficiently implemented by incorporating it in the search heuristic.,6317
belief networks that Experimental results probabilities learning is for suggest smoothing of helpful.,Experimental results suggest that for learning probabilities of belief networks smoothing is helpful.,6318
probabilities occurrence. conditions of by are and enabling The events represented,The events are represented by enabling conditions and probabilities of occurrence.,6319
issues The updating connection with networks. beliefs deals with in paper optimality in,The paper deals with optimality issues in connection with updating beliefs in networks.,6320
two processes: trees. triangulation of We junction construction and address,We address two processes: triangulation and construction of junction trees.,6321
"that integrals. a evaluates Carlo these Moreover, we Monte procedure develop new","Moreover, we develop a new Monte Carlo procedure that evaluates these integrals.",6322
It and is acceptancerejection. aspects of bootstrap sampling some on based,It is based on some aspects of bootstrap sampling and acceptancerejection.,6323
leak to enforce describe assumption of probabilities We model. in the the our use closed-world,We describe the use of leak probabilities to enforce the closed-world assumption in our model.,6324
simulation and discussion example. an by illustrative real-world Theoretical results is supplemented,Theoretical discussion is supplemented by simulation results and an illustrative real-world example.,6325
be speed-up maintaining achieved Additional can by path information. the,Additional speed-up can be achieved by maintaining the path information.,6326
also default We competing systems. comparisons make reasoning with,We also make comparisons with competing default reasoning systems.,6327
for Bayesian simulation technique is an networks. approximate inference Backward belief,Backward simulation is an approximate inference technique for Bayesian belief networks.,6328
"this expensive, impossible via indirectly is or recognition. be information must plan communication When acquired","When communication is impossible or expensive, this information must be acquired indirectly via plan recognition.",6329
license. this dis- modify of can tribution Users make terms and new under the,Users can modify and make new dis- tribution under the terms of this license.,6330
are neutral. Synonymous to affect do be mutations interplay often and this assumed not,Synonymous mutations do not affect this interplay and are often assumed to be neutral.,6331
most the Recognition today. fields one of Language Sign growing of research is,Sign Language Recognition is one of the most growing fields of research today.,6332
these have been recently techniques developed Many fields. new in,Many new techniques have been developed recently in these fields.,6333
"system and four Hand Classification. Skin Extraction of Feature Cropping, comprises The parts: Filtering,","The system comprises of four parts: Skin Filtering, Hand Cropping, Feature Extraction and Classification.",6334
decisions. time treatment a is change critical in Volumetric multiforme (GBM) in over glioblastoma factor,Volumetric change in glioblastoma multiforme (GBM) over time is a critical factor in treatment decisions.,6335
"structures. are complicated medical the an researches and Normally, exclusive more data","Normally, the medical data researches are more complicated and an exclusive structures.",6336
"expands diagnostic PET paper, in the various scope for functions. brain task In image this","In this paper, the scope diagnostic task expands for PET image in various brain functions.",6337
the sensitive Here orientation study to second-order psychophysical visual modulations. mechanisms we of show,Here we show psychophysical study of second-order visual mechanisms sensitive to the orientation modulations.,6338
"phase measured. to and Selectivity spatial is of modulation frequency orientation,","Selectivity to orientation, phase and spatial frequency of modulation is measured.",6339
field. emerged work as a has Nuclear in image research medical promising,Nuclear image has emerged as a promising research work in medical field.,6340
challenge. its Images own from different meet modality,Images from different modality meet its own challenge.,6341
world it testing It real for effectiveness its demonstrated by has sets. data patient,It has demonstrated its effectiveness by testing it for real world patient data sets.,6342
algorithm. are Experimental compared Means with K and results FCM clustering conventional,Experimental results are compared with conventional FCM and K Means clustering algorithm.,6343
compared provides algorithms SFCM performance The with PET the two of other satisfactory results .,The performance of the PET SFCM provides satisfactory results compared with other two algorithms .,6344
now Positron replaced a resolution. (PET) and Tomography caring the Emission has issues high,Positron Emission Tomography (PET) has now replaced the issues and caring a high resolution.,6345
method detection a sequences. brain based fully approach symmetry automated abnormality presents PET for This,This approach presents a fully automated symmetry based brain abnormality detection method for PET sequences.,6346
like performance accuracy considered diagnosis. the of The evolution metric by is,The performance evolution is considered by the metric like accuracy of diagnosis.,6347
obtained of result surely the surgeon seizures automated the is for focus. assists identification The,The obtained result is surely assists the surgeon for the automated identification of seizures focus.,6348
Neural by feedback often is subject currents. firing adaptation negative to,Neural firing is often subject to negative feedback by adaptation currents.,6349
between can currents correlations spikes. strong time among the induce These intervals,These currents can induce strong correlations among the time intervals between spikes.,6350
this properties. phenomenon explain paper attempts to analyze This and its,This paper attempts to explain this phenomenon and analyze its properties.,6351
 in This into ongoing  describes research paper environment. uncertain an planning,  This paper describes ongoing research into planning in an uncertain environment.,6352
strategic where in abstraction are place decisions made an hierarchy takes decisions. Planning before tactical,Planning takes place in an abstraction hierarchy where strategic decisions are made before tactical decisions.,6353
dependence networks independence variables. encoding quantifying conditional Belief among probabilistic and graphs are and,Belief networks are graphs encoding and quantifying probabilistic dependence and conditional independence among variables.,6354
presented. is the adequacy experimental of method discussed results preliminary are The and theoretical,The theoretical adequacy of the method is discussed and preliminary experimental results are presented.,6355
reasoning Evidential Artificial topic Intelligence. leading in a is now,Evidential reasoning is now a leading topic in Artificial Intelligence.,6356
is functions. Evidence represented of by variety a evidential,Evidence is represented by a variety of evidential functions.,6357
certain operation carried functions. fundamental out reasoning kinds by these Evidential of is on,Evidential reasoning is carried out by certain kinds of fundamental operation on these functions.,6358
involve decision and commonly under probabilistic variables. uncertainty random making of continuous Problems inference,Problems of probabilistic inference and decision making under uncertainty commonly involve continuous random variables.,6359
"computations. discretized a to simplify to and points, are Often assessments these few","Often these are discretized to a few points, to simplify assessments and computations.",6360
An to approximation is distributions. alternative continuous fit analytically tractable probability,An alternative approximation is to fit analytically tractable continuous probability distributions.,6361
"variables especially can advantages, accuracy and potential This has if be transformed approach simplicity first.","This approach has potential simplicity and accuracy advantages, especially if variables can be transformed first.",6362
is well-known this in algorithm. case the The procedure fitting EM,The fitting procedure in this case is the well-known EM algorithm.,6363
presented. two algorithm LED) preliminary are the on Results evaluation of networks (ALARM and of,Results of preliminary evaluation of the algorithm on two networks (ALARM and LED) are presented.,6364
algorithm open discuss some issues and also We problems. performance,We also discuss some algorithm performance issues and open problems.,6365
diagnosing uncertain logical analyses. through combining a and reasoning describe software method We problems for,We describe a method for diagnosing software problems through combining logical and uncertain reasoning analyses.,6366
"for ""compiling"" are simple decision presented of diagrams influence set Two algorithms into rules. a","Two algorithms are presented for ""compiling"" influence diagrams into a set of simple decision rules.",6367
"define rules near-optimal and complete, consistent, simple-to-execute, These decision procedures. decision","These decision rules define simple-to-execute, complete, consistent, and near-optimal decision procedures.",6368
Valuation networks graphical (VBSs). representations been as have proposed valuation-based of systems,Valuation networks have been proposed as graphical representations of valuation-based systems (VBSs).,6369
"show valuation networks how paper, independence In relations. we encode this conditional","In this paper, we show how valuation networks encode conditional independence relations.",6370
discusses several methods This from for reasoning inconsistent and presents paper bases. knowledge,This paper presents and discusses several methods for reasoning from inconsistent knowledge bases.,6371
of handled theory. priority levels the in The framework are possibility,The priority levels are handled in the framework of possibility theory.,6372
design. to Application oil the offshore platforms illustrates,Application to offshore oil platforms illustrates the design.,6373
"this system embedded in expert application, the real-time system. is normative For a","For this application, the normative system is embedded in a real-time expert system.",6374
"of the dynamic, automating construction on models. decision been little There incremental has research","There has been little research on automating the dynamic, incremental construction of decision models.",6375
proposed the diagnosis. construction for hierarchical uniform decision A of value-driven is model method complete,A uniform value-driven method of decision model construction is proposed for the hierarchical complete diagnosis.,6376
modeled and Hierarchical complete influence process is diagnostic a using stochastic as formulated diagrams. reasoning,Hierarchical complete diagnostic reasoning is formulated as a stochastic process and modeled using influence diagrams.,6377
"construction probe incrementally, models method interleaving model and evaluation. with This decision construct actions","This method construct decision models incrementally, interleaving probe actions with model construction and evaluation.",6378
tasks The method treats and uniformly. meta-level baselevel,The method treats meta-level and baselevel tasks uniformly.,6379
It calculations exactly those efficient to graphs. undirected corresponding on inference allows,It allows efficient inference calculations corresponding exactly to those on undirected graphs.,6380
weighted Watersheds been have graphs. edge defined node both for and,Watersheds have been defined both for node and edge weighted graphs.,6381
same weighted the and graph basin. minima with catchment edge),edge) weighted graph with the same minima and catchment basin.,6382
values. optimal expected The the perfect the two of information difference between value is,The value of perfect information is the difference between the two optimal expected values.,6383
the evaluating proposed. of tradeoffs the are of accuracy/efficiency Methods,Methods of evaluating the accuracy/efficiency of the tradeoffs are proposed.,6384
"still which addressed from instability, GIB-MAP be may ?approximate? a problem explanations using suffer","GIB-MAP explanations still suffer from instability, a problem which may be addressed using ?approximate?",6385
independence for as a condition irrelevance. conditional,conditional independence as a condition for irrelevance.,6386
on focuses literature the descriptors identification Most of holistic in closed-set applications. relevant,Most of the relevant literature focuses on holistic descriptors in closed-set identification applications.,6387
face to an verification approach then alternative propose SR. We via,We then propose an alternative approach to face verification via SR.,6388
utility-based to a categorization. take We approach,We take a utility-based approach to categorization.,6389
probabilities are presented. methods eliciting Parsimonious for dependent,Parsimonious methods for eliciting dependent probabilities are presented.,6390
models from we Bayesian Network In raw method learning of data. developed work previous a,In previous work we developed a method of learning Bayesian Network models from raw data.,6391
minimal on well method (MDL) the description This principle. relies known length,This method relies on the well known minimal description length (MDL) principle.,6392
work. that present we some our results In have paper arisen new this from,In this paper we present some new results that have arisen from our work.,6393
"computing of particular, In length. local description way present the a new we","In particular, we present a new local way of computing the description length.",6394
This improvements allows in search our make algorithm. to us significant,This allows us to make significant improvements in our search algorithm.,6395
involving practical networks size. of is The of our experiments feasibility a demonstrated by approach,The feasibility of our approach is demonstrated by experiments involving networks of a practical size.,6396
"while measures relative ambiguity those judgments. Probability surrounding likelihoods, vagueness measures","Probability measures relative likelihoods, while ambiguity measures vagueness surrounding those judgments.",6397
representation important of Ambiguity knowledge. is uncertain an,Ambiguity is an important representation of uncertain knowledge.,6398
"uncertainty deals or probability by subjective of belief. modeled type a with different, It","It deals with a different, type of uncertainty modeled by subjective probability or belief.",6399
"we network. a algorithm First, in for will belief present the MPE finding an","First, we will present an algorithm for finding the MPE in a belief network.",6400
are pignistic through the two subsystems transformation. connected The,The two subsystems are connected through the pignistic transformation.,6401
called by sign property of synergy. this is dependence a qualitative The product captured,The sign of this dependence is captured by a qualitative property called product synergy.,6402
"truth diagnosis. Finally, several maintenance, have argument and reasoning, Nonmonotonic networks I show applications: that","Finally, I show that argument networks have several applications: Nonmonotonic reasoning, truth maintenance, and diagnosis.",6403
belief Dynamic network for models reasoning. (DNMs) are networks temporal,Dynamic network models (DNMs) are belief networks for temporal reasoning.,6404
"belief-network forecasting, use simulation DNMs. discrete inference perform on control, event algorithms We to and","We use belief-network inference algorithms to perform forecasting, control, and discrete event simulation on DNMs.",6405
methodology an We in DNM important demonstrate medicine. the problem on forecasting,We demonstrate the DNM methodology on an important forecasting problem in medicine.,6406
information. This of allows for a wide and logical logic of range the probabilistic representation,This logic allows for the representation of a wide range of logical and probabilistic information.,6407
a as These composed network. a pieces are probability Bayesian to joint specified generate distribution,These pieces are composed to generate a joint probability distribution specified as a Bayesian network.,6408
statistical representations Bayesian models. knowledge ideal for diagrams Influence are,Influence diagrams are ideal knowledge representations for Bayesian statistical models.,6409
"manipulate. diagrams However, interpret to users and end are difficult these for to","However, these diagrams are difficult for end users to interpret and to manipulate.",6410
statistical restricted a of of subgraphs class influence as models are encoded diagram. of Elements,Elements of statistical models are encoded as subgraphs of a restricted class of influence diagram.,6411
actions construction the algorithm The called runs system that modular uses steps.,The algorithm that runs the system uses modular actions called construction steps.,6412
model for Noisy-Or describes Pearl variables. the Boolean,Pearl describes the Noisy-Or model for Boolean variables.,6413
construction This Bayesian useful networks. for modeling of a aid is generalization,This generalization is a useful modeling aid for construction of Bayesian networks.,6414
reliability examples illustrate analysis. We some circuit and diagnosis network with digital including,We illustrate with some examples including digital circuit diagnosis and network reliability analysis.,6415
must Most algorithms assumption. this at least inference make,Most inference algorithms must make at least this assumption.,6416
situation We as several belief information experts envision independently which networks. in a encode,We envision a situation in which several experts independently encode information as belief networks.,6417
"procedures, one of types for This and probabilities, for approach one graphs. requires two combination","This approach requires two types of combination procedures, one for probabilities, and one for graphs.",6418
on generalize functions. We models of it based the framework within the belief,We generalize it within the framework of the models based on belief functions.,6419
This optimal finding an paper of studies the tree. problem approximating,This paper studies the problem of finding an optimal approximating tree.,6420
time-critical making sequential a method for involving stochastic decision tasks and We describe processes.,We describe a method for time-critical decision making involving sequential tasks and stochastic processes.,6421
through probabilities. of achieved counterfactual This use is the,This is achieved through the use of counterfactual probabilities.,6422
We method to also imaging. Lewis's the demonstrate connection of,We also demonstrate the connection to Lewis's method of imaging.,6423
loudness temporal addition we to this exploit set of In a and features.,In addition to this we exploit a set of temporal and loudness features.,6424
for are different Thereafter classification. used classifiers,Thereafter different classifiers are used for classification.,6425
incorporate is propositional assumptions. extended probabilities classical the for model assumption-based to The,The classical propositional assumption-based model is extended to incorporate probabilities for the assumptions.,6426
theory. is evidence placed it the Then of into framework,Then it is placed into the framework of evidence theory.,6427
"recently techniques proposed for computing presented. degrees are support Finally, of","Finally, recently proposed techniques for computing degrees of support are presented.",6428
"limitation or the often distinctions, approach these the of ignored Bayesian Without is underestimated.","Without these distinctions, the limitation of the Bayesian approach is often ignored or underestimated.",6429
knowledge guide can to be values used elicitation. These,These values can be used to guide knowledge elicitation.,6430
"differentiation and local In adaptation maintained. case, this are","In this case, differentiation and local adaptation are maintained.",6431
migration amount linkage minimum We depends admitting invasion study pattern. how on the of this,We study how this minimum amount of linkage admitting invasion depends on the migration pattern.,6432
the This used describes PAGODA. representation mechanism paper in probabilistic inference and,This paper describes the probabilistic representation and inference mechanism used in PAGODA.,6433
probability conditional represented as distributions. are theories These,These theories are represented as conditional probability distributions.,6434
These are restricted theories. theories called predictive uniquely,These restricted theories are called uniquely predictive theories.,6435
nodes of tens with networks results thousands having of Empirical are presented.,Empirical results with networks having tens of thousands of nodes are presented.,6436
"NP-hard. Optimization generally however, energies, of such is","Optimization of such energies, however, is generally NP-hard.",6437
"and speed. it gives near minimum global deconvolution); curvature stereo, better regularization,","curvature regularization, stereo, deconvolution); it gives near global minimum and better speed.",6438
main of is curvature regularization. application Our interest,Our main application of interest is curvature regularization.,6439
"common knowledge, provides uncertainty. representing conceptual It knowledge, framework hierarchical for a and","It provides a common framework for representing conceptual knowledge, hierarchical knowledge, and uncertainty.",6440
of levels dynamic at abstraction. decision facilitates construction varying It of categorization models,It facilitates dynamic construction of categorization decision models at varying levels of abstraction.,6441
for rule examines combination the concept belief of paper This a functions.,This paper examines the concept of a combination rule for belief functions.,6442
that problem general is such exists the show a We of deciding dag whether NP-complete.,We show that the general problem of deciding whether such a dag exists is NP-complete.,6443
"to accept. Therefore, conclusions about uncertainty inconsistency in a logical causes database which","Therefore, inconsistency in a logical database causes uncertainty about which conclusions to accept.",6444
uncertainty uncertainty. This kind logical called of is,This kind of uncertainty is called logical uncertainty.,6445
"""acceptability"", define means arguments. differentiating of induces concept which for a a We","We define a concept of ""acceptability"", which induces a means for differentiating arguments.",6446
"more the we an argument, more it. acceptable The are in confident","The more acceptable an argument, the more confident we are in it.",6447
and by be aggregated variety symbolic functions. numeric flattening of Arguments can a,Arguments can be aggregated by a variety of numeric and symbolic flattening functions.,6448
with Dean Kanazawa's probabilistic approach and projection. compare our We,We compare our approach with Dean and Kanazawa's probabilistic projection.,6449
formal problem. give modelling decreasing of the a We persistence,We give a formal modelling of the decreasing persistence problem.,6450
"show persistence. Lastly, of nonmonotonic principle the how decreasing using conclusions infer to we","Lastly, we show how to infer nonmonotonic conclusions using the principle of decreasing persistence.",6451
of The of the a maximum-likelihood point. location each estimate output is,The output is a maximum-likelihood estimate of the location of each point.,6452
Gaussian all assumes that constraints It have noise.,It assumes that all constraints have Gaussian noise.,6453
"expected conceptual, of structural the and refinement decision quantitative, analyze of we models. value Specifically,","Specifically, we analyze the expected value of quantitative, conceptual, and structural refinement of decision models.",6454
examples. We with key illustrate refinement of the dimensions,We illustrate the key dimensions of refinement with examples.,6455
relevance. networks distinct notion We types based of each of similarity two a examine on,We examine two types of similarity networks each based on a distinct notion of relevance.,6456
"new Screening, design. Predictive predictor approach Correlation variable called for We introduce a to selection,","We introduce a new approach to variable selection, called Predictive Correlation Screening, for predictor design.",6457
problem. design the We Predictive predictor apply following Correlation to Screening two-stage,We apply Predictive Correlation Screening to the following two-stage predictor design problem.,6458
of is The probability scoring from concept proper for elicitation considered refinement rules.,The concept of refinement from probability elicitation is considered for proper scoring rules.,6459
The new with terms. convection system linear is hyperbolic,The new system is hyperbolic with linear convection terms.,6460
"the original new Under the recover can model. certain system assumptions,","Under certain assumptions, the new system can recover the original model.",6461
a model using consider detection natural object shapes. We generic for,We consider object detection using a generic model for natural shapes.,6462
object approach to directly models images. object matching involves recognition common A for,A common approach for object recognition involves matching object models directly to images.,6463
via grouping building intermediate processes. a generic involves Another representations approach,Another approach involves building intermediate representations via a generic grouping processes.,6464
computational two use We mechanisms. (model-based argue similar that grouping) recognition these processes may and,We argue that these two processes (model-based recognition and grouping) may use similar computational mechanisms.,6465
"easily Such by terms and expression. representations affected illumination, are of variations alignment, in pose","Such representations are easily affected by variations in terms of alignment, illumination, pose and expression.",6466
the word. of emphasizes of script the for energies of directional It significance identification,It emphasizes the significance of directional energies for identification of script of the word.,6467
sizes styles It and writing. of varied image different is robust to,It is robust to varied image sizes and different styles of writing.,6468
"of GMM. deviation is distributions of built-in and horizontal to vertical energies also Furthermore,","Furthermore, deviation of horizontal and vertical distributions of energies is also built-in to GMM.",6469
biological of often is context-dependent. genomic The significance features,The biological significance of genomic features is often context-dependent.,6470
"Second, associated cognitive transcripts function. relevant map lincRNA with functions to variants of","Second, variants associated with cognitive functions map to lincRNA transcripts of relevant function.",6471
"genes. domains highly lamina-associated are olfaction-related in Third, enriched","Third, lamina-associated domains are highly enriched in olfaction-related genes.",6472
"stratification having a optimal. other refined is hand, (too) not On the","On the other hand, having a (too) refined stratification is not optimal.",6473
points more of variations sample the or are function larger. the where noise,sample more points where the noise or variations of the function are larger.,6474
partition data is fundamental algorithms The points. clustering aim to of,The fundamental aim of clustering algorithms is to partition data points.,6475
processes. arbitrary space on Gaussian an covariate partition-valued define a We process using,We define a partition-valued process on an arbitrary covariate space using Gaussian processes.,6476
of Fiser to al. et comment A a response,A response to a comment of Fiser et al.,6477
diagnostic with Several label assess samplers tools convergence. to are switching along implemented and moves,Several samplers and label switching moves are implemented along with diagnostic tools to assess convergence.,6478
post-processing provided. are number of A for the functions also of output R,A number of R functions for post-processing of the output are also provided.,6479
as This selection. variable implemented in the is package,This is implemented in the package as variable selection.,6480
Concept experimented Concept theoretically largely grounded and are and Hierarchies well Formal Analysis methods.,Concept Hierarchies and Formal Concept Analysis are theoretically well grounded and largely experimented methods.,6481
on called analysing for Galois and They sets. rely lattices diagrams visualizing object-attribute line,They rely on line diagrams called Galois lattices for visualizing and analysing object-attribute sets.,6482
are experts. rich for Galois and visually lattices conceptually seducing,Galois lattices are visually seducing and conceptually rich for experts.,6483
reduced Semantic visual Galois which objects centred and sub-hierarchies. organize extract probes user are,Semantic probes are visual user centred objects which extract and organize reduced Galois sub-hierarchies.,6484
"text statistical this handwritten printed paper, use for texture In features classification. and we","In this paper, we use statistical texture features for handwritten and printed text classification.",6485
for south Indian level classification word We primarily aim in scripts.,We primarily aim for word level classification in south Indian scripts.,6486
document. Words are from scanned extracted the first,Words are first extracted from the scanned document.,6487
vectors classify to then classifier. used feature via k-NN These words are,These feature vectors are then used to classify words via k-NN classifier.,6488
the several validated have different over datasets. We approach,We have validated the approach over several different datasets.,6489
in manipulation elegans. we optogenetic and techniques methods describe for Here C.,Here we describe techniques and methods for optogenetic manipulation in C. elegans.,6490
Image on basis. block-by-block analysed are a sequences,Image sequences are analysed on a block-by-block basis.,6491
the network. cascade an We MAPK using give example also,We also give an example using the MAPK cascade network.,6492
memory size by or the disk The storage. available limited is buffer,The buffer size is limited by the available memory or disk storage.,6493
for geometric standard image fitting based approaches model are on features. Many pre-matched,Many standard approaches for geometric model fitting are based on pre-matched image features.,6494
"such appearances uses Typically, only feature (e.g. pre-matching","Typically, such pre-matching uses only feature appearances (e.g.",6495
"In multi-model framework. a we and contrast, problems feature joint fitting optimization matching in solve","In contrast, we solve feature matching and multi-model fitting problems in a joint optimization framework.",6496
energy generalization based a paper of proposes problem. on the several This fit-&-match assignment formulations,This paper proposes several fit-&-match energy formulations based on a generalization of the assignment problem.,6497
developed optimal We on an solver that min-cost-max-flow efficient solutions. finds near based algorithm,We developed an efficient solver based on min-cost-max-flow algorithm that finds near optimal solutions.,6498
matches. the detected Our number significantly increases approach of,Our approach significantly increases the number of detected matches.,6499
and with Our method interact can each agent that the assumes other. each environment,Our method assumes that each agent can interact with the environment and each other.,6500
of equilibrium according an set to rules. to the evolution state process converges The,The evolution process converges to an equilibrium state according to the set of rules.,6501
"autonomous agents pioneer scaffolds. of for fibroin In addition, the use silk characterizing we","In addition, we pioneer the use of autonomous agents for characterizing silk fibroin scaffolds.",6502
results analysis. can successfully that strongly employed texture for approach The our be suggest,The results strongly suggest that our approach can be successfully employed for texture analysis.,6503
in structure  many model. is of  epidemic incorporated Age types,  Age structure is incorporated in many types of epidemic model.,6504
models. should Such possibility parameterising be age-structured therefore considered when a epidemic,Such a possibility should therefore be considered when parameterising age-structured epidemic models.,6505
"manipulate. to and However, difficult notoriously they also generate are","However, they are also notoriously difficult to generate and manipulate.",6506
Extinction stochasticity. sensitivity in risks life of demographic their greatly to varied histories,Extinction risks of life histories varied greatly in their sensitivity to demographic stochasticity.,6507
age adult depended maturation. on of of Effects survival,Effects of adult survival depended on age of maturation.,6508
most variance life contributed to of Juvenile majority histories. total survival demographic in rate the,Juvenile survival rate contributed most to total demographic variance in the majority of life histories.,6509
"assign stochasticity. sensitivity to basic to used life-history-specific life be demographic information can Thus, history","Thus, basic life history information can be used to assign life-history-specific sensitivity to demographic stochasticity.",6510
value the of This is when small populations. of assessing vulnerability great,This is of great value when assessing the vulnerability of small populations.,6511
"which nature, questions to have but of This implications. may biophysical/biotechnical still allows ask biological","This allows to ask questions of biophysical/biotechnical nature, but which may still have biological implications.",6512
models rates. constant that epidemic occur Traditional consider at processes individual,Traditional epidemic models consider that individual processes occur at constant rates.,6513
"seasonality, also source the main discussed. The other interaction oscillations, with of is","The interaction with the other main source of oscillations, seasonality, is also discussed.",6514
task speckle images because is a This contaminated SAR noise. are difficult with,This is a difficult task because SAR images are contaminated with speckle noise.,6515
in under estimation multiplicative Different are point strategies literature. the discussed noise,Different point estimation strategies under multiplicative noise are discussed in the literature.,6516
be This parameter interval estimation. achieved can through,This can be achieved through interval parameter estimation.,6517
"interval. propose edge confidence that To we bootstrap-based end,","To that end, we propose bootstrap-based edge confidence interval.",6518
inference the different Monte using simulation. are Carlo strategies The compared merits relative of,The relative merits of the different inference strategies are compared using Monte Carlo simulation.,6519
"interval dataset. estimation, illustrate real analyze a edge we order In to also","In order to illustrate interval edge estimation, we also analyze a real dataset.",6520
nothing about frequently landscapes nature. in that known virtually is It is fitness stated,It is frequently stated that virtually nothing is known about fitness landscapes in nature.,6521
family TEM to We analyze the the methods $\beta$-lactamases of with resistance. antibiotic apply associated,We apply the methods to analyze the TEM family of $\beta$-lactamases associated with antibiotic resistance.,6522
fitness suggest are for empirical we suited tools The well comparisons of qualitative landscapes.,The qualitative tools we suggest are well suited for comparisons of empirical fitness landscapes.,6523
"supports alignments, chimeric end-to-end and alignment. automatically between performs paired-end reads local It chooses and","It automatically chooses between local and end-to-end alignments, supports paired-end reads and performs chimeric alignment.",6524
Enoplea. divided (indcludes Chromadorea and the The C. the is into phylum Nematoda elegans),The phylum Nematoda is divided into the Chromadorea (indcludes C. elegans) and the Enoplea.,6525
"elegans, cell Compared C. and to have of very enoplean determination. division patterns nematodes different","Compared to C. elegans, enoplean nematodes have very different patterns of cell division and determination.",6526
even lost has has toolkit of elegans. independently C. than T. more the spiralis,T. spiralis has independently lost even more of the toolkit than has C. elegans.,6527
applications. a in vision processing computer numerous is subtraction Background fundamental task low-level,Background subtraction is a fundamental low-level processing task in numerous computer vision applications.,6528
"an analysed are image Specifically, overlapping sequences block-by-block basis. on","Specifically, image sequences are analysed on an overlapping block-by-block basis.",6529
"foreground ad-hoc Unlike not masks many methods, is of post-processing pixel-based required.","Unlike many pixel-based methods, ad-hoc post-processing of foreground masks is not required.",6530
observed several discrepancies. for the propose We explanations,We propose several explanations for the observed discrepancies.,6531
"OpenMP library. To has effectiveness of algorithm using calculations, been enhance parallelized","To enhance effectiveness of calculations, algorithm has been parallelized using OpenMP library.",6532
embryonic insight provide epithelia. new results the into These healing in mechanism,These results provide new insight into the healing mechanism in embryonic epithelia.,6533
virus world. over all (HCV) the a widely is C spread Hepatitis disease,Hepatitis C virus (HCV) is a widely spread disease all over the world.,6534
makes HCV mutation that it to resistant has high antibodies. rate very,HCV has very high mutation rate that makes it resistant to antibodies.,6535
pair-wise by using was Secondly rate method calculated distance sequences. mutation between,Secondly mutation rate was calculated by using pair-wise distance method between sequences.,6536
"methods different consistent. the are their surprisingly Despite two assumptions,","Despite their different assumptions, the two methods are surprisingly consistent.",6537
mobility. a' as the identify greatest domain the showing intra-domain We,We identify the a' domain as showing the greatest intra-domain mobility.,6538
survey are of four-fold. The this contributions,The contributions of this survey are four-fold.,6539
for segments. image partitioning novel describes paper This a meaningful method into,This paper describes a novel method for partitioning image into meaningful segments.,6540
"a well-known employs technique. watershed transform, proposed image The method segmentation","The proposed method employs watershed transform, a well-known image segmentation technique.",6541
it on so operates such that adaptively. input a It organized manner in is image,It is organized in such a manner so that it operates on input image adaptively.,6542
Its all makes convenient and suitable and robustness efficiency images. of more types for it,Its robustness and efficiency makes it more convenient and suitable for all types of images.,6543
value optimization the an developed algorithm problem. for solving We general function trajectory,We developed an algorithm solving the value function for general trajectory optimization problem.,6544
to Fast the closely is related J. Our Method and algorithm Tsitsiklis's Marching,Our algorithm is closely related to the Tsitsiklis's Fast Marching Method and J.,6545
Handwriting refers to the written characters. recognition of identification,Handwriting recognition refers to the identification of written characters.,6546
recognition Arabic is word task challenging to handwriting very nature. its due a online cursive,Arabic word online handwriting recognition is a very challenging task due to its cursive nature.,6547
classifier handwriting This recognition script one and multiple Arabic on-line off-line systems. combines system,This Arabic script multiple classifier system combines one off-line and on-line handwriting recognition systems.,6548
exhibits of diversity. surprisingly chromosome genetic human The low Y levels,The human Y chromosome exhibits surprisingly low levels of genetic diversity.,6549
"observed are show we Y models Instead, of consistent selection diversity. purifying that with","Instead, we show that models of purifying selection are consistent with observed Y diversity.",6550
"substantially different stimuli, The neurons. among represented vary modalities sensory however, different and","The represented stimuli, however, vary substantially among different sensory modalities and different neurons.",6551
"only only possible, exists. a number encoded bifurcations a features of few As are restricted","As only a few bifurcations are possible, only a restricted number of encoded features exists.",6552
"and specifically on elliptic we focus square-wave Here describing bursters. parabolic,","Here we focus specifically on describing parabolic, square-wave and elliptic bursters.",6553
show bifurcations. by determined that These type of selectivity the correspondences stimulus is,These correspondences show that stimulus selectivity is determined by the type of bifurcations.,6554
genetic major defining variation. axes populations Asian of the E.,E. Asian populations defining the major axes of genetic variation.,6555
so we errors. can propagates that the between then correct context possible labelling It neighbours,It then propagates the context between neighbours so that we can correct possible labelling errors.,6556
"linearly to the it pseudo-words. number using When kd-tree, of almost proportional is a","When using a kd-tree, it is almost linearly proportional to the number of pseudo-words.",6557
of presented. gray images variety A automata of cellular is on operations,A variety of operations of cellular automata on gray images is presented.,6558
stable in of state. a a All finishing nature operations wave-front are,All operations are of a wave-front nature finishing in a stable state.,6559
"Topographic terms used: are ""dales ""lakes"", dales"". ""dales"", of","Topographic terms are used: ""lakes"", ""dales"", ""dales of dales"".",6560
asynchronous architecture. Algorithms be realized a parallel with can,Algorithms can be realized with a parallel asynchronous architecture.,6561
"graphical technique proof independence. conditional Therefore, is for a purely this","Therefore, this is a purely graphical proof technique for conditional independence.",6562
"loci understanding Furthermore, lags discovery. generally of functional associated","Furthermore, functional understanding of associated loci generally lags discovery.",6563
Conditions the sequence to small-world models larger-world which of under probabilities. given converges the are,Conditions are given under which the sequence of small-world models converges to the larger-world probabilities.,6564
as monolithic probability conditional a The dependence language the single represented distribution. original,The original language represented the dependence as a single monolithic conditional probability distribution.,6565
important to contributions classifiers Cascade are detection. the real-time most one object of,Cascade classifiers are one of the most important contributions to real-time object detection.,6566
"detectors. arising training Nonetheless, are there cascade many in problems challenging","Nonetheless, there are many challenging problems arising in training cascade detectors.",6567
classifier. with classifier issue One is symmetric is the trained common node a that,One common issue is that the node classifier is trained with a symmetric classifier.,6568
the pre-determined accelerate approach model time learning the Our reduction objective. learning helps while achieving,Our model reduction approach helps accelerate the learning time while achieving the pre-determined learning objective.,6569
on relative which the to hypotheses is of efficient Search compute. probabilities based are,Search is based on the relative probabilities of hypotheses which are efficient to compute.,6570
the important introduced. of properties Three uniqueness technique contribute to the,Three important properties contribute to the uniqueness of the technique introduced.,6571
"on the This greatly experts. the information, providers burden facility reduces of the","This facility greatly reduces the burden on the providers of information, the experts.",6572
illustrate selection. a of project problem We approach with the,We illustrate the approach with a problem of project selection.,6573
networks using dynamic reasoning computational about systems for scheme A presented. is probabilistic (causal),A computational scheme for reasoning about dynamic systems using (causal) probabilistic networks is presented.,6574
scheme The of top implemented HUGIN shell. the has been on,The scheme has been implemented on top of the HUGIN shell.,6575
for (VBS) representing drawing knowledge under general framework provides Valuation-based inferences and a uncertainty. system,Valuation-based system (VBS) provides a general framework for representing knowledge and drawing inferences under uncertainty.,6576
embedded in agents the in Deliberation the real-world. plays design role of an important rational,Deliberation plays an important role in the design of rational agents embedded in the real-world.,6577
extended which structure simple to efficiently. a permits propagate us probability have These networks,These extended networks have a simple structure which permits us to propagate probability efficiently.,6578
probabilities conditional when are given independence Improved assumptions for bounds hold.,Improved bounds for conditional probabilities are given when independence assumptions hold.,6579
is l-map. The resulting DAG a minimal,The resulting DAG is a minimal l-map.,6580
role study The into mutations. provided of insights genomic the of functional has key variation,The study of genomic variation has provided key insights into the functional role of mutations.,6581
describe motivate a this We and theory of paper. in belief,We motivate and describe a theory of belief in this paper.,6582
mind. in of developed belief human view with the following is theory This,This theory is developed with the following view of human belief in mind.,6583
Consider will occurring). that has an occur E the event is belief or occurred (or,Consider the belief that an event E will occur (or has occurred or is occurring).,6584
promising One emerging is field the psychiatry. computational particularly of approach,One particularly promising approach is the emerging field of computational psychiatry.,6585
goal. will recent this To survey towards this I exemplify efforts,To exemplify this I will survey recent efforts towards this goal.,6586
potential (SEPI). inference call evidence the We symbolic extension,We call the extension symbolic evidence potential inference (SEPI).,6587
"SEPI. we of this In procedure describe query and systematic the paper, caching","In this paper, we describe the systematic query and caching procedure of SEPI.",6588
"clique. the tree, probability various stored are in each and With distributions clique computed","With the clique tree, various probability distributions are computed and stored in each clique.",6589
"Once done, be then can step the query this computed. is","Once this step is done, the query can then be computed.",6590
"To a used. process the SPI similar process is algorithm a recursive query, to","To process a query, a recursive process similar to the SPI algorithm is used.",6591
The and computation simple. the are algorithm,The algorithm and the computation are simple.,6592
this in SEPI The be algorithm along with presented several examples. paper will,The SEPI algorithm will be presented in this paper along with several examples.,6593
"relation termed been has ""intercausal."" The generated so","The relation so generated has been termed ""intercausal.""",6594
"known is ""noisy-or"" a model The of such example distribution. an well","The well known ""noisy-or"" model is an example of such a distribution.",6595
"(SPIC). algorithm, this We of SPI SPI call Continuous the variation","We call this variation of the SPI algorithm, SPI Continuous (SPIC).",6596
"summation, operations: basic the and SPI SPIC three substitution. multiplication, modifies","SPIC modifies the three basic SPI operations: multiplication, summation, and substitution.",6597
developed. the nonlinear based filter If-then is on rules fuzzy A,A nonlinear filter based on the fuzzy If-then rules is developed.,6598
Intensive is suggest possible. justification the that results theoretical empirical of simulations,Intensive simulations suggest that theoretical justification of the empirical results is possible.,6599
expansion not maximum product is nets for entropy. probabilities conditional of belief The,The product expansion of conditional probabilities for belief nets is not maximum entropy.,6600
for a model. to of deny kind appears This assurance desirable the,This appears to deny a desirable kind of assurance for the model.,6601
"of concept to is generalization, thc of described. specialization, concept The dual","The concept of generalization, dual to thc concept of specialization, is described.",6602
Mukaidono's is described. system reason A fuzzy an maintenance which ATMS extends through logic,A reason maintenance system which extends an ATMS through Mukaidono's fuzzy logic is described.,6603
to distribution maximum entropy given The derives the subject a constraints. procedure,The procedure derives a maximum entropy distribution subject to the given constraints.,6604
raises multiple-author of interesting questions. of more variety The a settings introduction sophisticated,The introduction of more sophisticated multiple-author settings raises a variety of interesting questions.,6605
nature of consensus. involves the question and One compromise such,One such question involves the nature of compromise and consensus.,6606
in introduces prior setting. This net a Bayes paper compromises,This paper introduces prior compromises in a Bayes net setting.,6607
notions (VBS). systems valuation-based the paper and independence introduces in of This conditional independence,This paper introduces the notions of independence and conditional independence in valuation-based systems (VBS).,6608
axiomatic framework uncertainty many is capable an different representing VBS calculi. of,VBS is an axiomatic framework capable of representing many different uncertainty calculi.,6609
independence define joint valuation. of and of conditional We the independence terms in factorization,We define independence and conditional independence in terms of factorization of the joint valuation.,6610
relationship The is databases examined. relational between and networks belief,The relationship between belief networks and relational databases is examined.,6611
"diagrams. using a and is formulated influence as modeled reasoning sequential process, stochastic is Diagnostic","Diagnostic reasoning is formulated as a sequential stochastic process, and is modeled using influence diagrams.",6612
decision uses thresholds sensitivity analysis. approach class the for provide DYNASTY to equivalence an,DYNASTY uses an equivalence class approach to provide decision thresholds for the sensitivity analysis.,6613
to different. only equivalence-class actions reasoning differentiates diagnoses required This if are diagnostic the approach,This equivalence-class approach to diagnostic reasoning differentiates diagnoses only if the required actions are different.,6614
dynamically for algorithms updating of A set are network-topology necessary. the network when proposed updating,A set of network-topology updating algorithms are proposed for dynamically updating the network when necessary.,6615
to is applied layered hierarchies. distinct strategy the set Hence a control of,Hence the control strategy is applied to a layered set of distinct hierarchies.,6616
"with with a clause presents abduction, Horn framework for probabilities paper This simple hypotheses. associated","This paper presents a simple framework for Horn clause abduction, with probabilities associated with hypotheses.",6617
consultants; consultation attempt of capability emulate We human systems. this to computer in,We attempt to emulate this capability of human consultants; in computer consultation systems.,6618
for Our been in use systems. developed consultation has task-oriented mechanism,Our mechanism has been developed for use in task-oriented consultation systems.,6619
a The travel for agency. domain is chosen of we that that exploration have,The domain that we have chosen for exploration is that of a travel agency.,6620
past of intensive for the Learning study has the decade. subject Ontology been,Ontology Learning has been the subject of intensive study for the past decade.,6621
formal precise more compared Heavy-weight Ontology ontology to reasoning learning. when logic-based statistical supports Learning,Heavy-weight Ontology Learning supports more precise formal logic-based reasoning when compared to statistical ontology learning.,6622
IS-A evaluation adopted on based Standard We rich Gold Learning Ontology Wikipedia chosen documents. have,We have adopted Gold Standard based Ontology Learning evaluation on chosen IS-A rich Wikipedia documents.,6623
"relations concepts between and models semantic logic. based fuzzy of examine similarity Finally, these we","Finally, we examine relations between these concepts and similarity based semantic models of fuzzy logic.",6624
performing. are usually the The they uncertainty way on changes experts depending manage task,The way experts manage uncertainty usually changes depending on the task they are performing.,6625
"arguments example, well-known. ?consistent are behavior' betting For","For example, ?consistent betting behavior' arguments are well-known.",6626
models. paper probability another present explores for rationale The,The present paper explores another rationale for probability models.,6627
on can build of we sand? castle What kind,What kind of castle can we build on sand?,6628
"last for generations. set designed foundations our to on We bedrock, want","We want our foundations set on bedrock, designed to last for generations.",6629
directed relationships of among random graphs independence representing set variables. acyclic Bayesian networks are a,Bayesian networks are directed acyclic graphs representing independence relationships among a set of random variables.,6630
exhaustive be as can of exclusive mutually A regarded and propositions. variable set random a,A random variable can be regarded as a set of exhaustive and mutually exclusive propositions.,6631
"introspect about to must probabilities so error, subject their Such the planner are validity.","Such probabilities are subject to error, so the planner must introspect about their validity.",6632
can these be about statistics. events the probability using made of Inferences,Inferences about the probability of these events can be made using statistics.,6633
can the Inferences estimation. interval about of the using made approximations be validity,Inferences about the validity of the approximations can be made using interval estimation.,6634
inference are A and relevant set of patterns knowledge types identified.,A relevant set of inference patterns and knowledge types are identified.,6635
technique al. (Shafer The local computation et,The local computation technique (Shafer et al.,6636
expressed is Uncertain parameterized operators. means by modal of information,Uncertain information is expressed by means of parameterized modal operators.,6637
We and this and propose sound axiomatization. complete logic a give semantics for multimodal a,We propose a semantics for this multimodal logic and give a sound and complete axiomatization.,6638
beliefs analyzed. unnormalized The nature these is of,The nature of these unnormalized beliefs is analyzed.,6639
the used experts. decision structure by be to provided synthesize An the interval rules can,An interval structure can be used to synthesize the decision rules provided by the experts.,6640
expressed interval prescribed proportions. term of a is as Each linguistic,Each linguistic term is expressed as a prescribed interval of proportions.,6641
"The modelling studied rules, context. quantified this of in syllogism, the probabilistic chaining is","The quantified syllogism, modelling the chaining of probabilistic rules, is studied in this context.",6642
Irrelevance-based domain-independent for are using useful partial MAPs networks. constructs belief explanation,Irrelevance-based partial MAPs are useful constructs for domain-independent explanation using belief networks.,6643
"tool, modeling. a is proposed this In for new environment called paper, U-yraph,","In this paper, a new tool, called U-yraph, is proposed for environment modeling.",6644
report opportunities in an beliefnet parallelism inference. for investigation We into on experimental,We report on an experimental investigation into opportunities for parallelism in beliefnet inference.,6645
"clustering the parallelism find tree, at opportunity for negligible level. We or topological,","We find negligible opportunity for parallelism at the topological, or clustering tree, level.",6646
to operate network queries. answer over inference algorithms the Probabilistic,Probabilistic inference algorithms operate over the network to answer queries.,6647
paper independence paradigms Bayesian asymmetric for networks representation encoding assertions. discuses This multiple,This paper discuses multiple Bayesian networks representation paradigms for encoding asymmetric independence assertions.,6648
method proposes problems. paper a solving new decision Bayesian for This,This paper proposes a new method for solving Bayesian decision problems.,6649
that obtained identical to from homogeneous those are be Probabilities the network. obtained would,Probabilities obtained are identical to those that would be obtained from the homogeneous network.,6650
propagation shift and discuss acquired evidence. previously We to junction of different a attention tree,We discuss attention shift to a different junction tree and propagation of previously acquired evidence.,6651
"for systems. adaptive aHUGIN, creating paper a describes tool The","The paper describes aHUGIN, a tool for creating adaptive systems.",6652
reported is are a session experiments the with Finally and results discussed.,Finally a session with experiments is reported and the results are discussed.,6653
discussed. belief propagation are Several situations of,Several situations of belief propagation are discussed.,6654
discriminate information. between and priori' We 'a evidential,We discriminate between 'a priori' and evidential information.,6655
"methods them. Then, we each combination for of propose different one","Then, we propose different combination methods for each one of them.",6656
We and conditioning combination information. evidential the 'a heterogeneous of also consider as priori',We also consider conditioning as the heterogeneous combination of 'a priori' and evidential information.,6657
The evidential represented as a convex set information of is likelihood functions.,The evidential information is represented as a convex set of likelihood functions.,6658
classical These to distribution with an Possibility possibility Theory. according have associated behavior will,These will have an associated possibility distribution with behavior according to classical Possibility Theory.,6659
distribution the methods used an to derive for statistic. Asymptotic test are approximate,Asymptotic methods are used to derive an approximate distribution for the test statistic.,6660
result The VT that in the among achieved is them accuracy best both experiments. parsing,The result is that VT achieved the best parsing accuracy among them in both experiments.,6661
superior such in found not that does necessarily yield performance. case a We VT,We found that in such a case VT does not necessarily yield superior performance.,6662
Practice and To (TPLP). Theory Logic of Programming in appear,To appear in Theory and Practice of Logic Programming (TPLP).,6663
overcome preferential difficulties. of (RPMs) to work models promises on ranked these some Recent,Recent work on ranked preferential models (RPMs) promises to overcome some of these difficulties.,6664
adequate RPMs belief show change. to not iterated handle we Here are that,Here we show that RPMs are not adequate to handle iterated belief change.,6665
"always allow the for Specifically, not belief reversibility we RPMs show of that change. do","Specifically, we show that RPMs do not always allow for the reversibility of belief change.",6666
result belief. the indicates need This strengths numerical for of,This result indicates the need for numerical strengths of belief.,6667
problems decision and uncertainty discuss representation. for making Bayesian We convex,We discuss problems for convex Bayesian decision making and uncertainty representation.,6668
a as been with for Possibilistic proposed reasoning has formalism uncertainty. logic numerical,Possibilistic logic has been proposed as a numerical formalism for reasoning with uncertainty.,6669
networks of representation Influence with a uncertainty. diagram belief is graphical,Influence diagram is a graphical representation of belief networks with uncertainty.,6670
probabilistic studies model diagram. properties the influence This structural in an article of a,This article studies the structural properties of a probabilistic model in an influence diagram.,6671
"controllability observability developed structural formulated. algorithms theorems theorems are and structural In are particular, and","In particular, structural controllability theorems and structural observability theorems are developed and algorithms are formulated.",6672
system determined by be properties the Both of can ranks the matrices.,Both properties can be determined by the ranks of the system matrices.,6673
described. of is Monte-Carlo the belief calculation computationally-efficient Dempster-Shafer for algorithm very A,A very computationally-efficient Monte-Carlo algorithm for the calculation of Dempster-Shafer belief is described.,6674
decision is the complicated making knowledge involved. the often complexity Automated of by,Automated decision making is often complicated by the complexity of the knowledge involved.,6675
sensitive the Much complexity context phenomena. variations of from of arises underlying this the,Much of this complexity arises from the context sensitive variations of the underlying phenomena.,6676
"for We framework a descriptive, propose context-sensitive knowledge. representing","We propose a framework for representing descriptive, context-sensitive knowledge.",6677
uncertain knowledge a approach attempts integrate to Our and categorical network formalism. in,Our approach attempts to integrate categorical and uncertain knowledge in a network formalism.,6678
representations quantitative studied qualitative was in theory. extensively compatibility and The of probability of beliefs,The compatibility of quantitative and qualitative representations of beliefs was studied extensively in probability theory.,6679
with is of qualitative the shown compatible probability belief functions. that It structure is monotonic,It is shown that the structure of qualitative probability is compatible with monotonic belief functions.,6680
control and model dynamic construction failure of is to necessary also revision. Diagnosis model,Diagnosis of model failure is also necessary to control dynamic model construction and revision.,6681
We to a represent distribution labelings use labeling. over of each respective possibilities possibility,We use a possibility distribution over labelings to represent respective possibilities of each labeling.,6682
the a of each certainty of simple expression respective allow constraint. Necessity-valued degrees constraints,Necessity-valued constraints allow a simple expression of the respective certainty degrees of each constraint.,6683
approach is technical of The main our advantage the framework. integration CSP in its,The main advantage of our approach is its integration in the CSP technical framework.,6684
simple a demonstrated our The design on of utility is approach problem.,The utility of our approach is demonstrated on a simple design problem.,6685
to the are added evidence. specific of form in DBN Sensor basic observations the,Sensor observations are added to the basic DBN in the form of specific evidence.,6686
"sensor However, often is incorrect. or data partially totally","However, sensor data is often partially or totally incorrect.",6687
conditional with incrementally. probabilities for and rules independences Uncertain explicit maintained be bounds can,Uncertain rules with bounds for probabilities and explicit conditional independences can be maintained incrementally.,6688
"new precise particular, for entailment. bounds probabilistic In we analytical provide","In particular, we provide new precise analytical bounds for probabilistic entailment.",6689
"syntactically such Nevertheless, inferences and be characterized can both semantically.","Nevertheless, such inferences can be characterized both syntactically and semantically.",6690
by matrices. The controlled flow mass specialization is,The mass flow is controlled by specialization matrices.,6691
aspects be matrices. reasoning represented Even can certain by specialization of non-monotonic some,Even some aspects of non-monotonic reasoning can be represented by certain specialization matrices.,6692
the reasoning representing about utility knowledge agent's actions. and with an We time-dependent of discuss,We discuss representing and reasoning with knowledge about the time-dependent utility of an agent's actions.,6693
consistency is The base knowledge the analyzed. of,The consistency of the knowledge base is analyzed.,6694
"We knowledge representation needed first the and solving describe formalisms the problem, problem paradigms.","We first describe the problem, the needed knowledge representation formalisms and problem solving paradigms.",6695
can blackboard approach a clearly explained system system's be The from interactive perspective.,The system's interactive approach can be clearly explained from a blackboard system perspective.,6696
chronic techniques. marker-passing Useless paths for are problem a,Useless paths are a chronic problem for marker-passing techniques.,6697
discussed. are point to the of probabilities estimates Techniques for posterior sentences entailed calculate,Techniques to calculate point estimates for the posterior probabilities of entailed sentences are discussed.,6698
inference. on We review algorithms randomized for previous probabilistic BNRAS work algorithms- approximation,We review previous work on BNRAS algorithms- randomized approximation algorithms for probabilistic inference.,6699
"problems. the and to relate this discuss in paper We methods open we work, previous","We relate the methods in this paper to previous work, and we discuss open problems.",6700
PULCinella and comparing theories. its use uncertainty We in present,We present PULCinella and its use in comparing uncertainty theories.,6701
"Pulcinella define his allows Moreover, user specializations. easily to own the","Moreover, Pulcinella allows the user to easily define his own specializations.",6702
"using illustrate analyze four examples two To Pulcinella, above. by of theories we each the","To illustrate Pulcinella, we analyze two examples by using each of the four theories above.",6703
"In between differences focus on one, theories. intrinsic mainly first we the","In the first one, we mainly focus on intrinsic differences between theories.",6704
is current approach. art tree propagation the clique of state The,The current state of art is the clique tree propagation approach.,6705
direct in marginalization. The is each simply computation clique,The computation in each clique is simply direct marginalization.,6706
"of estimates often unavoidable mechanical design uncertainty design, performance. is in In there","In mechanical design, there is often unavoidable uncertainty in estimates of design performance.",6707
Evaluation requires alternatives of consideration of impact design of the this uncertainty.,Evaluation of design alternatives requires consideration of the impact of this uncertainty.,6708
method automobile in system for the selection for illustrate material We bumpers. a,We illustrate the method in a system for material selection for automobile bumpers.,6709
set a distributions. result of probability normalized non-necessarily is The it of convex,The result of it is a convex set of non-necessarily normalized probability distributions.,6710
the Sugeno will methods be based of and use These on Choquet integrals.,These methods will be based on the use of Sugeno and Choquet integrals.,6711
examples. compared in be basis selected to some will behaviour Their,Their behaviour will be compared in basis to some selected examples.,6712
possibility in first show and assignments. formation We uncertainty of to for how measures define,We first show how to define measures of in formation and uncertainty for possibility assignments.,6713
allows the notion the in information content among us to capture proximity It of distributions.,It allows us to capture the notion of proximity in information content among the distributions.,6714
discuss for maximum a possibility we of distributions. principle last the section uncertainty In,In the last section we discuss a principle of maximum uncertainty for possibility distributions.,6715
how as be We an such rule. inference a principle formalized show could,We show how such a principle could be formalized as an inference rule.,6716
logic theory. has control consequences and correspondence far This reaching in in,This correspondence has far reaching consequences in logic and in control theory.,6717
"Our a is presentation terse self-contained, albeit as summary. intended","Our presentation is intended as a self-contained, albeit terse summary.",6718
not we Proofs only included; are examples. illustrative offer,Proofs are not included; we only offer illustrative examples.,6719
"illustrative several forms of a example. updating, practical with of Survey","Survey of several forms of updating, with a practical illustrative example.",6720
"single is operation rule. and subtle single no is no method, a Updating there 'good'","Updating is a subtle operation and there is no single method, no single 'good' rule.",6721
The due consideration. given the always appropriate of must choice be rule,The choice of the appropriate rule must always be given due consideration.,6722
practical We the on of focus rules. meaning these,We focus on the practical meaning of these rules.,6723
"system. for presented semantics, the results consistency are and consequence Approximations,","Approximations, semantics, consistency and consequence results are presented for the system.",6724
are searching designs widely algorithms  optimal saturated  for Efficient for available.,  Efficient algorithms for searching for optimal saturated designs are widely available.,6725
(such efficiency provide a measure maximize They as given an and D-optimality) design. optimum,They maximize a given efficiency measure (such as D-optimality) and provide an optimum design.,6726
"do design. not a optimal they Nevertheless, guarantee \emph{global}","Nevertheless, they do not guarantee a \emph{global} optimal design.",6727
"design. local random a an from optimal Indeed, start initial design and find they","Indeed, they start from an initial random design and find a local optimal design.",6728
"general, the will, optimum If be in found the is different. design changed initial","If the initial design is changed the optimum found will, in general, be different.",6729
developed. tool SAS A software been has in written,A software tool written in SAS has been developed.,6730
"outline acids. hydroxylated) potential the We fatty to modified (e.g., extensions calculations also of","We also outline potential extensions of the calculations to modified (e.g., hydroxylated) fatty acids.",6731
"Edges heavy determination for detection, information classification. structural and carry","Edges carry heavy structural information for detection, determination and classification.",6732
a of center interpolation approaches focus. become Edge-adaptive,Edge-adaptive interpolation approaches become a center of focus.,6733
evaluates robustness ability. the edge-preservation accuracy measures of One and other,One evaluates accuracy and the other measures robustness of edge-preservation ability.,6734
evaluation six parameters. is based Performance on,Performance evaluation is based on six parameters.,6735
"applied including a mathematics. and applications, in variety econometrics of models wide appear statistical Such","Such statistical models appear in a wide variety of applications, including econometrics and applied mathematics.",6736
(ABC). approximate introduce We new approximation upon Bayesian computation a based,We introduce a new approximation based upon approximate Bayesian computation (ABC).,6737
of life understanding fundamental on Speciation the diversity huge of is importance Earth. to,Speciation is of fundamental importance to understanding the huge diversity of life on Earth.,6738
used this dynamic The aim. is method for programming,The dynamic programming method is used for this aim.,6739
Proposed system characteristic variance decrease estimators. method a to of allows,Proposed method allows to decrease a variance of system characteristic estimators.,6740
Methodologies Matlab entropy variations. implemented in were enhanced with and were,Methodologies were implemented in Matlab and were enhanced with entropy variations.,6741
Red cell-Hb-CSF The response functions Hb adapting sensor equilibriums. a heterotropic to as,The Red cell-Hb-CSF functions as a sensor adapting response to Hb heterotropic equilibriums.,6742
"Hence, etc., changes in pKa, sliding, concatenated vectoriality. n-H-bonds, became state, for could dipole","Hence, changes in dipole state, sliding, pKa, n-H-bonds, etc., could became concatenated for vectoriality.",6743
on has A architecture implemented the already part a humanoid robot. been of,A part of the architecture has already been implemented on a humanoid robot.,6744
"of globally modulated this method. The focuses STDP, use-case as special a on study","The study focuses on globally modulated STDP, as a special use-case of this method.",6745
processor to by wafer. embedding dedicated is Flexibility general-purpose the plasticity achieved a into,Flexibility is achieved by embedding a general-purpose processor dedicated to plasticity into the wafer.,6746
"i.e. This model its simulated is to measure performance,","This model is simulated to measure its performance, i.e.",6747
learning. the reward increase received after in,the increase in received reward after learning.,6748
"the latency reward processor. embedded This with which sent is increases to delay, the the","This latency increases the delay, with which the reward is sent to the embedded processor.",6749
We to that a accelerated find has for latency highly kept systems be to minimum.,We find that for highly accelerated systems latency has to be kept to a minimum.,6750
tested data. independently compared and with two subsystem models The are,The two subsystem models are tested independently and compared with data.,6751
sensing for widely are urban techniques land classification and analysis. Remote cover used,Remote sensing techniques are widely used for land cover classification and urban analysis.,6752
introduced. object-based paper a on hierarchical this vector is In machine based classification scheme support,In this paper object-based classification scheme based on a hierarchical support vector machine is introduced.,6753
methods has been methodology statistical compared with measures. Developed using different contemporary,Developed methodology has been compared with contemporary methods using different statistical measures.,6754
CNN with Investigations considerable over the revealed satellite achieved was images success various that approach.,Investigations over various satellite images revealed that considerable success was achieved with the CNN approach.,6755
the and study fuzzy We bipolar of soft set properties. notion fundamental introduce,We introduce the notion of bipolar fuzzy soft set and study fundamental properties.,6756
operations fuzzy set. soft study on bipolar basic We,We study basic operations on bipolar fuzzy soft set.,6757
"bipolar exdended intersection define fuzzy of soft set. We union, two","We define exdended union, intersection of two bipolar fuzzy soft set.",6758
application problem. fuzzy making give an We set soft into also decision of bipolar,We also give an application of bipolar fuzzy soft set into decision making problem.,6759
The has decade. past  connectome the been studied  over human widely,  The human connectome has been widely studied over the past decade.,6760
"methodologically. may This limited however, result, be","This result, however, may be limited methodologically.",6761
are settings presented this and and clearly results obtained experimental explained Detailed in report.,Detailed experimental settings and obtained results are clearly presented and explained in this report.,6762
competing proposed explain hypotheses Several occurrence. been have its to,Several competing hypotheses have been proposed to explain its occurrence.,6763
TSR. contrast suggests is single mechanism the for accounts that there no that This,This contrast suggests that there is no single mechanism that accounts for the TSR.,6764
varies grid scale axis entorhinal along cortex. The of the dorso-ventral the,The grid scale varies along the dorso-ventral axis of the entorhinal cortex.,6765
recognition. case called machine handwriting work same does The is in the,The same work in case the machine does is called handwriting recognition.,6766
down broken on-line. and Handwriting can recognition categories: two into be off-line,Handwriting recognition can be broken down into two categories: off-line and on-line.,6767
possible. mixtures all of Arbitrary also are three cases,Arbitrary mixtures of all three cases are also possible.,6768
"assumptions results But, the estimation violated. actually if are could distorted some be","But, the estimation results could be distorted if some assumptions actually are violated.",6769
effectiveness data and the imaging We data. brain using our artificial of method simulated demonstrate,We demonstrate the effectiveness of our method using artificial data and simulated brain imaging data.,6770
Numerical presented of examples ABC are HMMs. approximations for,Numerical examples are presented for ABC approximations of HMMs.,6771
penalized approximations then used approximate log-likelihood. the are minimize These to,These approximations are then used to minimize the approximate penalized log-likelihood.,6772
matrices. efficiency relies memory For the the implementation use sparse time of crucially and on,For time and memory efficiency the implementation relies crucially on the use of sparse matrices.,6773
R ppstat. available the package The implementation is in,The implementation is available in the R package ppstat.,6774
image A implementation non-local denoising filtering is of multiscale means new proposed. for,A new multiscale implementation of non-local means filtering for image denoising is proposed.,6775
comparison. a The modification also introduces of patch proposed for similarity measure algorithm,The proposed algorithm also introduces a modification of similarity measure for patch comparison.,6776
Euclidean comparison. for norm The patch norm Euclidean standard replaced by is weighted based,The standard Euclidean norm is replaced by weighted Euclidean norm for patch based comparison.,6777
problems. assumptions by are These real satisfied most not unfortunately,These assumptions are unfortunately not satisfied by most real problems.,6778
A for procedure distributions inference uncertainty is interval presented. marginal maximum,A maximum uncertainty inference procedure for marginal interval distributions is presented.,6779
This domain of financial the is integration in illustrated mergers acquisitions. and,This integration is illustrated in the financial domain of mergers and acquisitions.,6780
ideas system MARS. prototype implemented have in called a been These,These ideas have been implemented in a prototype system called MARS.,6781
these relatively required to operator sophisticated Both of make judgments. meteorological the systems,Both of these systems required the operator to make relatively sophisticated meteorological judgments.,6782
address encouraged participants potential the to above. at author Interested contact the are,Interested potential participants are encouraged to contact the author at the address above.,6783
are with statistical in for belief data on considerations functions detail. dealt Methodological using some,Methodological considerations for using belief functions on statistical data are dealt with in some detail.,6784
derived. aid used a conditions as an aid which Some decision are is under best,Some conditions under which a decision aid is best used as an aid are derived.,6785
diseases. for diagnosis multiple a the of We model examine probabilistic,We examine a probabilistic model for the diagnosis of multiple diseases.,6786
"the model, variables. findings In diseases as and binary are represented","In the model, diseases and findings are represented as binary variables.",6787
inference introduce a bounded called graceful to We approach probabilistic conditioning.,We introduce a graceful approach to probabilistic inference called bounded conditioning.,6788
claims wrong. both are that It is argued,It is argued that both claims are wrong.,6789
toward Non- aid. The such Monotonic Probabilist an initial represents work,The Non- Monotonic Probabilist represents initial work toward such an aid.,6790
of transformation are fastest the These Mobius (O). transformations' computing 'fast the algorithms for Mobius,These 'fast Mobius transformations' are the fastest algorithms for computing the Mobius transformation of (O).,6791
"taxonomic that an theory abductive probabilistic, We and propose integrates causal diagnosis knowledge.","We propose an abductive diagnosis theory that integrates probabilistic, causal and taxonomic knowledge.",6792
all scenario where These are observations events constitute causation a true. the,These causation events constitute a scenario where all the observations are true.,6793
is inherently reasoning is allowed. modest Abductive if power only expressive complex even,Abductive reasoning is inherently complex even if only modest expressive power is allowed.,6794
The on basic approximate sample an based representation algorithm the probabilities. of is,The algorithm is based on an approximate sample representation of the basic probabilities.,6795
specific measurement probability distributions. included be using functions may Dempster-Shafer mass,Dempster-Shafer probability mass functions may be included using specific measurement distributions.,6796
numerical The by are properties experiments. the of approach demonstrated,The properties of the approach are demonstrated by numerical experiments.,6797
is the connected. singly can be Savings not original network computational in when complexity large,Savings in computational complexity can be large when the original network is not singly connected.,6798
We subgraph computationally network. define Bayesian of a equivalent a,We define a computationally equivalent subgraph of a Bayesian network.,6799
classic about penguin is the A most Tweety. example,A most classic example is about the penguin Tweety.,6800
in on works inheritance. focused primarily probabilistic this have Previous area,Previous works in this area have primarily focused on probabilistic inheritance.,6801
Networks. and operations the state coarsening presents This two refining Bayesian paper in space for,This paper presents two operations for refining and coarsening the state space in Bayesian Networks.,6802
for also acquisition. their We knowledge implications practical discuss,We also discuss their practical implications for knowledge acquisition.,6803
"are other representing Reiter's Though this distinction, capable formalism is of original not. approaches","Though Reiter's original formalism is capable of representing this distinction, other approaches are not.",6804
of is closely to probability paradox Simpson's The theory. related problem,The problem is closely related to Simpson's paradox of probability theory.,6805
reasoning with We information problem concerned the credibility into are type introducing of systems.,We are concerned with the problem of introducing credibility type information into reasoning systems.,6806
information to provided discount The of by concept credibility allows us agents.,The concept of credibility allows us to discount information provided by agents.,6807
credibility qualification reasoning. approximate We of the suggest representational theory scheme for a in,We suggest a representational scheme for credibility qualification in the theory of approximate reasoning.,6808
We of the discuss credibility. relative concept,We discuss the concept of relative credibility.,6809
situation to structures very in This much the of nonmonotonic leads reasoning. spirit,This situation leads to structures very much in the spirit of nonmonotonic reasoning.,6810
of on approach in exploitation based is equations The max-min algebra. the,The approach is based on the exploitation of equations in max-min algebra.,6811
limit includes of case precise information. the certain and This formalism,This formalism includes the limit case of certain and precise information.,6812
adding at a time. one built The incrementally is network node,The network is built incrementally adding one node at a time.,6813
in can answer queries predicate that independencies about The a the domain. model probabilistic is,The probabilistic model is a predicate that can answer queries about independencies in the domain.,6814
implemented in In model ways. can the practice be various,In practice the model can be implemented in various ways.,6815
of research. discuss solution propose for future limitations We this this directions and,We discuss limitations of this solution and propose future directions for this research.,6816
cycles reasoning task The in most handling be difficult may probabilistic in belief directed networks.,The most difficult task in probabilistic reasoning may be handling directed cycles in belief networks.,6817
a multidisorder Such produced called for clustering. diagnosis symptom by candidate sets recent are method,Such candidate sets are produced by a recent method for multidisorder diagnosis called symptom clustering.,6818
of formalizes implementation functions Truth Assumption-based Maintenance analysis (ATMS). an Belief within the This System,This analysis formalizes the implementation of Belief functions within an Assumption-based Truth Maintenance System (ATMS).,6819
"Dempster visual VICTORS Without Shafer (i.e. all theory, possible interpretations computes","Without Dempster Shafer theory, VICTORS computes all possible visual interpretations (i.e.",6820
best the without interpretation(s). determining logical models) all,all logical models) without determining the best interpretation(s).,6821
is constraints A of from of system method calculating presented. marginal probability values a,A method of calculating probability values from a system of marginal constraints is presented.,6822
expert . This system calculating appropriate for time in method a real is probabilities,This method is appropriate for calculating probabilities in a real time expert system .,6823
normal study of multiscale presents skin mathematical This a model hybrid (vSkin).,This study presents a hybrid multiscale mathematical model of normal skin (vSkin).,6824
observations. with our as experimental are results These well as clinical consistent predictions,These predictions are consistent with our experimental results as well as clinical observations.,6825
technique error an The algorithm function. involves greedy search that using locally a minimizes,The technique involves using a greedy search algorithm that locally minimizes an error function.,6826
"reasoning, accounting the for Scenario-based the causal of generation alternative involves ""stories"" The other, evidence.","The other, Scenario-based reasoning, involves the generation of alternative causal ""stories"" accounting for the evidence.",6827
well employ as probabilistic schemes as Both causal knowledge.,Both schemes employ causal as well as probabilistic knowledge.,6828
Probabilities be may numbers. and/or as presented phrases,Probabilities may be presented as phrases and/or numbers.,6829
"control can style, completeness explanations. and abstraction Users the of","Users can control the style, abstraction and completeness of explanations.",6830
in mid show a detection peak in speeds. browsing results Our performance,Our results show a peak in detection performance in mid browsing speeds.,6831
single masses. or micro-calcifications by simulated inserting Lesion are cases,Lesion cases are simulated by inserting single micro-calcifications or masses.,6832
improve of ways them are to methods and our Limitations discussed.,Limitations of our methods and ways to improve them are discussed.,6833
Previous the linguistic probabilities and between has absolute work empirical correspondence investigated phrases.,Previous empirical work has investigated the correspondence between absolute probabilities and linguistic phrases.,6834
describe such to selected phrases probability updates. numerical best Subjects,Subjects selected such phrases to best describe numerical probability updates.,6835
the before managing of focuses This cost action. on paper deliberation,This paper focuses on managing the cost of deliberation before action.,6836
a consisting with into boolean of belief (i.e. nodes The network is only compiled network,The belief network is compiled into a network consisting only of nodes with boolean (i.e.,6837
on the is using resulting search found a best-first then MAP network. The assignment,The MAP assignment is then found using a best-first search on the resulting network.,6838
during IDEAL development. and some describes This paper lessons its learned,This paper describes IDEAL and some lessons learned during its development.,6839
representing problems. and optimization describes for solving discrete paper This systems valuation-based,This paper describes valuation-based systems for representing and solving discrete optimization problems.,6840
"called the function. objective an valuations, of factors functions, The represent","The functions, called valuations, represent the factors of an objective function.",6841
the using operations problem called and marginalization. involves optimization two Solving combination,Solving the optimization problem involves using two operations called combination and marginalization.,6842
of the factors us the objective joint function. how combine tells Combination to,Combination tells us how to combine the factors of the joint objective function.,6843
"non-serial For method systems valuation-based solution the problems, of optimization programming. to reduces dynamic","For optimization problems, the solution method of valuation-based systems reduces to non-serial dynamic programming.",6844
use dynamic that axioms programming. viewed permit as conditions the And be our can of,And our axioms can be viewed as conditions that permit the use of dynamic programming.,6845
"But, procedures to are strongly the the support Bayesian more also hypothesis. likely wrong","But, the Bayesian procedures are also more likely to strongly support the wrong hypothesis.",6846
"techniques more powerful, more Bayesian also are but error prone. are","Bayesian techniques are more powerful, but are also more error prone.",6847
"is idea this inference paper, of an policy some the detail. in explored In","In this paper, the idea of an inference policy is explored in some detail.",6848
"of characteristics are To inference this support policies standard nonstandard some exploration, the and examined.","To support this exploration, the characteristics of some standard and nonstandard inference policies are examined.",6849
introduced established. the a are expert approximations on through Bounds system into CPN-based errors,Bounds on the errors introduced into a CPN-based expert system through approximations are established.,6850
number solve algorithms belief A have developed to on problems probabilistic of inference networks. been,A number of algorithms have been developed to solve probabilistic inference problems on belief networks.,6851
sound. show and the are optimal transformations We,We show the transformations are optimal and sound.,6852
describe considerably networks. an environment generating We process simplifies of the Bayesian that belief,We describe an environment that considerably simplifies the process of generating Bayesian belief networks.,6853
Augustus of analysis proposed. is An busts,An analysis of Augustus busts is proposed.,6854
"However, of probability the aggregates, lack e.g.","However, the lack of probability aggregates, e.g.",6855
"this to In extend probability we allow paper, arbitrary DHPP aggregates.","In this paper, we extend DHPP to allow arbitrary probability aggregates.",6856
Limited of the implementation provided. empirical data on an are methodology,Limited empirical data on an implementation of the methodology are provided.,6857
a special This theories. proof of presents cases paper some between equivalence these straightforward,This paper presents a straightforward equivalence proof between some special cases of these theories.,6858
"To in an actions. the about world, operate must its intelligently reason agent","To operate intelligently in the world, an agent must reason about its actions.",6859
temporal propositional paper reasoning This representing a about and logic presents for actions. probability,This paper presents a propositional temporal probability logic for representing and reasoning about actions.,6860
that the various logic facts occur events represent The probability times. hold and can at,The logic can represent the probability that facts hold and events occur at various times.,6861
affect represent and that future. can actions It the probability events the other,It can represent the probability that actions and other events affect the future.,6862
The over time. probability model probabilities of relates,The model of probability relates probabilities over time.,6863
Several the given. logic illustrating use the of are examples,Several examples illustrating the use of the logic are given.,6864
"probabilistic belief Key words: inference, computational complexity simulation, theory, algorithms. randomized stochastic networks,","Key words: probabilistic inference, belief networks, stochastic simulation, computational complexity theory, randomized algorithms.",6865
relevant with associated time-points the of occurrence events.,time-points associated with the occurrence of relevant events.,6866
"are of analogues and Jeffrey's discussed. first of In the rule introduced conditioning cases, two","In the two first cases, analogues of Jeffrey's rule of conditioning are introduced and discussed.",6867
possibilistic Shenoy's rule that combination counterpart. well-known shown has a It is,It is shown that Shenoy's combination rule has a well-known possibilistic counterpart.,6868
None qualification has Dempster-Shafer. models the received of these of,None of these models has received the qualification of Dempster-Shafer.,6869
"down. the component conclusions is these once dynamic But break considered,","But once the dynamic component is considered, these conclusions break down.",6870
too restricted. on based is the static comparison only Any component,Any comparison based only on the static component is too restricted.,6871
object this search task. call We the,We call this the object search task.,6872
medical On-line over is time. short of relatively a data period available,On-line medical data is available over a relatively short period of time.,6873
extracted from with are opinions. Functions (Dempster-Shafer modified then theory) and Belief expert data first,Belief Functions (Dempster-Shafer theory) are first extracted from data and then modified with expert opinions.,6874
symptoms about Expert and opinions information among derived dependencies are also statistically compared.,Expert opinions and statistically derived information about dependencies among symptoms are also compared.,6875
theory for argument. presented computational of A is probabilistic framework a,A framework is presented for a computational theory of probabilistic argument.,6876
knowledge Probabilistic Reasoning levels. The encodes Environment at three,The Probabilistic Reasoning Environment encodes knowledge at three levels.,6877
of encoding a are knowledge. the deepest schemata the system's At level set domain,At the deepest level are a set of schemata encoding the system's domain knowledge.,6878
"constructed the level network Bayesian top is Finally, at from a the arguments.","Finally, at the top level is a Bayesian network constructed from the arguments.",6879
"paper, this of optimum is networks In decomposition discussed. belief","In this paper, optimum decomposition of belief networks is discussed.",6880
allow have cycles We and nommonotonic nonmonotonic inferences RUM within to rules. extended,We have extended RUM to allow nommonotonic inferences and cycles within nonmonotonic rules.,6881
a basis measures for among multiple Uncertainty defaults. provide deciding,Uncertainty measures provide a basis for deciding among multiple defaults.,6882
the defaults optimal Different for finding discussed. heuristics algorithms and are,Different algorithms and heuristics for finding the optimal defaults are discussed.,6883
problems all of involve Nearly reasoning uncertainty or another. one spatial sort,Nearly all spatial reasoning problems involve uncertainty of one sort or another.,6884
in arises used sensors inaccuracies the angles. distances due measuring of to and Uncertainty,Uncertainty arises due to the inaccuracies of sensors used in measuring distances and angles.,6885
We refer as this directional to uncertainty.,We refer to this as directional uncertainty.,6886
also Uncertainty in with spatial information location mistakenly when arises combining is one identified another.,Uncertainty also arises in combining spatial information when one location is mistakenly identified with another.,6887
as uncertainty. to recognition this refer We,We refer to this as recognition uncertainty.,6888
arise We attention the problems to due particular to uncertainty. that recognition pay,We pay particular attention to the problems that arise due to recognition uncertainty.,6889
this overly Response or has to either been view pessimistic. enthusiastic unduly,Response to this view has been either overly enthusiastic or unduly pessimistic.,6890
"expended for is available for However, reformulation time performing not inference.","However, time expended for reformulation is not available for performing inference.",6891
We uncertainty. principles for general describe of the ideal under computing partition first resources shall,We shall describe first general principles for computing the ideal partition of resources under uncertainty.,6892
problems illustrates these An example given how is arise. that,An example is given that illustrates how these problems arise.,6893
a usage. investigation that finite The be limited of probability very model shows may,The investigation shows that a finite probability model may be of very limited usage.,6894
"decision-theoretic Problem Bayesian control BPS, the applies to probabilistic inference Solver, problem-solving. resource-constrained and flexible,","BPS, the Bayesian Problem Solver, applies probabilistic inference and decision-theoretic control to flexible, resource-constrained problem-solving.",6895
"sound significantly traditional with outperform By computational effort. performing inference, less BPS can techniques","By performing sound inference, BPS can outperform traditional techniques with significantly less computational effort.",6896
"a information. means expressing however, Probability incompleteness for constitute of intervals,","Probability intervals, however, constitute a means for expressing incompleteness of information.",6897
"data a to discussed between Finally, distinguish it rare and case. how conflicting is","Finally, it is discussed how to distinguish between conflicting data and a rare case.",6898
basic search of used has control as unit. previously its Decision-theoretic,Decision-theoretic control of search has previously used as its basic unit.,6899
the of and of of computation successors. complete generation a set evaluation,of computation the generation and evaluation of a complete set of successors.,6900
"some satisficing. opportunities it Although lost results for simplifies in pruning this analysis, and","Although this simplifies analysis, it results in some lost opportunities for pruning and satisficing.",6901
of techniques may analytic more useful control The reasoning in prove settings. used for general,The analytic techniques used may prove useful for control of reasoning in more general settings.,6902
are Expectations a to PSEIKI as of presented geometrical hierarchy abstractions.,Expectations are presented to PSEIKI as a geometrical hierarchy of abstractions.,6903
dependencies. theoretic in terms to a leads of causation definition of This statistical model algorithm,This algorithm leads to a model theoretic definition of causation in terms of statistical dependencies.,6904
of major system the Knowledge bottlenecks one is design. in expert elicitation,Knowledge elicitation is one of the major bottlenecks in expert system design.,6905
probabilities). Systems based on two and structure Bayes nets (or parameters information--network types of require,Systems based on Bayes nets require two types of information--network structure and parameters (or probabilities).,6906
domain must from be elicited expert. Both the,Both must be elicited from the domain expert.,6907
Evidences to evidences). conflicting in or supports (conflicting other evidences) be each correlated may (dependent,Evidences may be correlated to each other (dependent evidences) or conflicting in supports (conflicting evidences).,6908
model. geometric on The based a is method,The method is based on a geometric model.,6909
The rules of proposed examples. performance shown different are by,The performance of proposed rules are shown by different examples.,6910
reasonably make rules results dependent that the show under evidences The . decision proposed,The results show that the proposed rules reasonably make decision under dependent evidences .,6911
probabilities database. from for arc obtained directly an Conditional are the,Conditional probabilities for an arc are obtained directly from the database.,6912
is This an entropy-based process threshold continues reached. until,This process continues until an entropy-based threshold is reached.,6913
consistently belief the fidelity. networks The system original high reproduces with,The system consistently reproduces the original belief networks with high fidelity.,6914
briefly of is approach the between idea minimizing relation and The abnormality discussed. our,The relation between our approach and the idea of minimizing abnormality is briefly discussed.,6915
paper a dynamically for framework this constructing present Bayesian we In networks.,In this paper we present a framework for dynamically constructing Bayesian networks.,6916
of evidence. Dempster-Shafer We discuss theory the,We discuss the Dempster-Shafer theory of evidence.,6917
framework property. accumulation We exhibits nonmonotonic show knowledge that of a in this the,We show that the accumulation of knowledge in this framework exhibits a nonmonotonic property.,6918
can how structure represent knowledge. commonsense belief show used or typical the be We to,We show how the belief structure can be used to represent typical or commonsense knowledge.,6919
reasonably for a systems. expert even appears to procedure be This large practical,This appears to be a practical procedure even for reasonably large expert systems.,6920
among network. restrict in a variables Functional the interactions connected potential probabilistic dependencies,Functional dependencies restrict the potential interactions among variables connected in a probabilistic network.,6921
method describe constructing for incrementally We belief a networks.,We describe a method for incrementally constructing belief networks.,6922
"classes probabilistic models. Using parameterized of language, we this define can","Using this language, we can define parameterized classes of probabilistic models.",6923
of of presented examples. artificial through field two work Applications intelligence this the to are,Applications of this work to the field of artificial intelligence are presented through two examples.,6924
used increasingly dialogue are policies to learning from Reinforcement optimise methods experience.,Reinforcement learning methods are increasingly used to optimise dialogue policies from experience.,6925
"reinforcement strategy learning. In Bayesian in alternative an grounded paper, we this model-based investigate","In this paper, we investigate an alternative strategy grounded in model-based Bayesian reinforcement learning.",6926
rules. results the prior of with The high-level capturing knowledge benefits particular domain illustrate in,The results illustrate in particular the benefits of capturing prior domain knowledge with high-level rules.,6927
"structures characterize axioms principle. consider we which Miller's Finally, satisfying","Finally, we consider which axioms characterize structures satisfying Miller's principle.",6928
"(DS) Knowledge we with suggest marrying In theory Representation Dempster-Shafer paper, (KR). this","In this paper, we suggest marrying Dempster-Shafer (DS) theory with Knowledge Representation (KR).",6929
also touched is implementation The upon. problem,The implementation problem is also touched upon.,6930
"fuzzy aggregates. arbitrary DFLP In allow paper, to this extend we","In this paper, we extend DFLP to allow arbitrary fuzzy aggregates.",6931
"logic, is non-monotonic relative additional The associated to probability conditional compatible evaluations, with evidence. conditional","The associated conditional logic, compatible with conditional probability evaluations, is non-monotonic relative to additional evidence.",6932
mentioned. are probability conditional aspects Computational of logic,Computational aspects of conditional probability logic are mentioned.,6933
possesses features. important The Lp logic some further,The logic Lp possesses some further important features.,6934
"those based of possesses sets. Second, probabilities and semantics, Lp sets transparent on","Second, Lp possesses transparent semantics, based on sets and probabilities of those sets.",6935
"the briefly, of and theory semantics, this proof syntax, present, the After we logic.","After this we present, briefly, the syntax, semantics, and proof theory of the logic.",6936
between a We planning mapping describe STRIPS duality tasks.,We describe a duality mapping between STRIPS planning tasks.,6937
be new a This as seen may challenge.,This may be seen as a new challenge.,6938
of analysed Each feature the for is independently presence type anomaly. an,Each feature type is independently analysed for the presence of an anomaly.,6939
compared. was Performance systems the of then resulting,Performance of the resulting systems was then compared.,6940
discussed. differences reasons these are for Possible,Possible reasons for these differences are discussed.,6941
concentrated paper and In associated his on issues. Nilsson computational inference,In his paper Nilsson concentrated on inference and associated computational issues.,6942
"easily such device. expressed are particular, In not a statistical by assertions","In particular, statistical assertions are not easily expressed by such a device.",6943
models numerical Many of theory--its the of examples this use do not probabilities.,Many of the examples of this theory--its models do not use numerical probabilities.,6944
"acquisition and It modification and makes test, also development, system knowledge more difficult.","It also makes knowledge acquisition and system development, test, and modification more difficult.",6945
efficiency are in expected introduced. program the the constraints shows causal The as increase,The program shows the expected increase in efficiency as the causal constraints are introduced.,6946
reasoning into enters inference two and human Uncertainty ways. at in least,Uncertainty enters into human reasoning and inference in at least two ways.,6947
"provably semantic which we sound. of derive rules inference framework, Nilsson's set are Within a","Within Nilsson's semantic framework, we derive a set of inference rules which are provably sound.",6948
in with occurrence associated There uncertainty is real life. many the of events,There is uncertainty associated with the occurrence of many events in real life.,6949
is developed temporal The simple powerful. but logic,The developed temporal logic is simple but powerful.,6950
combinations. various or can the relations their It temporal events between determine effectively uncertain,It can determine effectively the various temporal relations between uncertain events or their combinations.,6951
procedure A processing is efficient heuristic of structures. developed for these,A heuristic procedure is developed for efficient processing of these structures.,6952
inference approximation method is An continuous with for variables. presented probabilistic random,An approximation method is presented for probabilistic inference with continuous random variables.,6953
noisy knowledge in for Decision systems being used domains. tree are induction acquisition,Decision tree induction systems are being used for knowledge acquisition in noisy domains.,6954
improvements these to then are A of systems number suggested.,A number of improvements to these systems are then suggested.,6955
inference to for rank framework hypotheses. order Bayesian provides a probabilities accruing_,Bayesian inference provides a framework for accruing_ probabilities to rank order hypotheses.,6956
"sex it face that and Instead, early is is obsolescence. species widespread clonal","Instead, sex is widespread and it is clonal species that face early obsolescence.",6957
host reduce reflexively evolves The to parasitic load.,The host evolves reflexively to reduce parasitic load.,6958
outcompetes cloning by mechanisms. Sex two overlapping,Sex outcompetes cloning by two overlapping mechanisms.,6959
offset the of lesser This rate fecundity advantage sex. can,This rate advantage can offset the lesser fecundity of sex.,6960
loci interference of The under number increases markedly selection. the independent with,The interference increases markedly with the number of loci under independent selection.,6961
"is higher all parasitic model carry Since applicability. the of organisms loads, general","Since all higher organisms carry parasitic loads, the model is of general applicability.",6962
"bi-valued normally Furthermore, variables than are distributed restrictive for variables. the less conditions governing those","Furthermore, the conditions for normally distributed variables are less restrictive than those governing bi-valued variables.",6963
database been in system hospital data On-line has medical the for years. several collected,On-line medical data has been collected in the hospital database system for several years.,6964
They finding methods many significant - the of more enhance variables. classical results also statistical,They also enhance the results of classical statistical methods - finding many more significant variables.,6965
are with made presented are and other Results methods. comparisons,Results are presented and comparisons are made with other methods.,6966
that updating demonstrates prior that importance formulae This the of outweigh can assumptions. of,This demonstrates that the importance of updating formulae can outweigh that of prior assumptions.,6967
used within this to object Then a the find photographic image. is network,Then this network is used to find the object within a photographic image.,6968
"expense leads at the Greater calculations, net's to the independence expressiveness. easier of","Greater independence leads to easier calculations, at the expense of the net's expressiveness.",6969
formulation then demonstrate be solved. network a We one of can how multiply connected,We then demonstrate how one formulation of a multiply connected network can be solved.,6970
combining Modifiable synthesis are to combining functions evidence. a approaches two common of,Modifiable combining functions are a synthesis of two common approaches to combining evidence.,6971
of many They disadvantages. these some of advantages approaches the avoid offer and,They offer many of the advantages of these approaches and avoid some disadvantages.,6972
of information Proofreading/editing translation in the code. essential accurate genetic protein from synthesis for is,Proofreading/editing in protein synthesis is essential for accurate translation of information from the genetic code.,6973
steady independent of to found The concentration. in amino acid state be is editing,The editing in steady state is found to be independent of amino acid concentration.,6974
"gate"" of These include some concept. generalizations OR ""noisy the","These include some generalizations of the ""noisy OR gate"" concept.",6975
principle Classical was using Approach to This the developed probability.,This principle was developed using the Classical Approach to probability.,6976
diagnosing testbed problem We motorcycle use of as experimental the engines.,We use as experimental testbed the problem of diagnosing motorcycle engines.,6977
approach from selection The expert test heuristic employs first rules mechanics. obtained,The first approach employs heuristic test selection rules obtained from expert mechanics.,6978
cost reduce analytic (i.e. to was algorithm The the expected found decision,The decision analytic algorithm was found to reduce the expected cost (i.e.,6979
suggests difference interesting This for knowledge implications acquisition. some,This difference suggests some interesting implications for knowledge acquisition.,6980
theory weakened Arguments calculus role. is this are Bayesian relegated against decision when to the,Arguments against Bayesian decision theory are weakened when the calculus is relegated to this role.,6981
This research. current of the is focus our,This is the focus of our current research.,6982
computationally logic is We under the conditions first sampling which infeasible. review,We first review the conditions under which logic sampling is computationally infeasible.,6983
development cases stochastic Such Pearl's simulation algorithm. the motivated the of,Such cases motivated the development of the Pearl's stochastic simulation algorithm.,6984
"such distinguish path-length metric, we as will the is p-optimal. When solutions","When path-length is the metric, we will distinguish such solutions as p-optimal.",6985
is model represented angina The decision Bayesian as a network.,The angina model is represented as a Bayesian decision network.,6986
capabilities this paper approach. of of the This examples emphasizes,This paper emphasizes examples of the capabilities of this approach.,6987
use technology An its illustrates in example assessment. medical,An example illustrates its use in medical technology assessment.,6988
viewed be can others) (among partial reply as of Cheeseman's defense to probability. a It,It can be viewed as a partial reply to Cheeseman's (among others) defense of probability.,6989
and possibility We introduce qualification operation the show how. of,We introduce the operation of possibility qualification and show how.,6990
belief-network a demonstrates algorithms diagram problems. for influence to using This solve method paper,This paper demonstrates a method for using belief-network algorithms to solve influence diagram problems.,6991
"belief-network be approximation In to applied may algorithms both problems. particular, and influence-diagram solve exact","In particular, both exact and approximation belief-network algorithms may be applied to solve influence-diagram problems.",6992
Feedback explanation the decisions and can cause modified. to be,Feedback can cause the explanation and decisions to be modified.,6993
I the Inductive Theory-Based call method (T-BIL). Learning,I call the method Theory-Based Inductive Learning (T-BIL).,6994
the theoretical This this makes of the to algorithm. study following contributions paper,This paper makes the following contributions to the theoretical study of this algorithm.,6995
obtain We uniformly Gibbs as particle corollary ergodic. kernel the that direct is a,We obtain as a direct corollary that the particle Gibbs kernel is uniformly ergodic.,6996
"extend Gibbs resampling lower particle Thirdly, work we variance schemes. with to","Thirdly, we extend particle Gibbs to work with lower variance resampling schemes.",6997
discussed. method a is particular limitation has which This,This particular method has a limitation which is discussed.,6998
focuses This networks. on in multiply-connected belief probability updates paper,This paper focuses on probability updates in multiply-connected belief networks.,6999
discuss need satisfied be conditions We by selected the to that nodes.,We discuss conditions that need to be satisfied by the selected nodes.,7000
algorithm these a loop-cutset heuristic satisfies conditions. for that present finding a We,We present a heuristic algorithm for finding a loop-cutset that satisfies these conditions.,7001
belief is knowledge associated with Domain functions. rules stored as,Domain knowledge is stored as rules with associated belief functions.,7002
in uncertainty addresses of system. errors resolving This the rule-based under problem a paper,This paper addresses the problem of resolving errors under uncertainty in a rule-based system.,7003
as been this has neural-network problem A new approach reformulates developed problem. a learning that,A new approach has been developed that reformulates this problem as a neural-network learning problem.,7004
The limitations the of and explored fundamental and are this discussed. approach strength,The strength and the fundamental limitations of this approach are explored and discussed.,7005
"increasingly retrieval in and and information. become important of storage Hypertext management, the hypermedia have","Hypertext and hypermedia have become increasingly important in the storage management, and retrieval of information.",7006
"extensively packages. broad terms, hypermedia In in deliver heterogeneous bits of cross-referenced dynamic, information","In broad terms, hypermedia deliver heterogeneous bits of information in dynamic, extensively cross-referenced packages.",7007
representing Beliefs propagated the BNDL clauses of the are distributions. programs probabilistic underlying the among,Beliefs are propagated among the clauses of the BNDL programs representing the underlying probabilistic distributions.,7008
Dempster's rule controversial the has the most theory. been of (D-S) Dempster-Shafer combination of part,Dempster's rule of combination has been the most controversial part of the Dempster-Shafer (D-S) theory.,7009
give is what explanations requirement One such need of to is qualitative believed. the,One such requirement is the need to give qualitative explanations of what is believed.,7010
thereby The the in procedure practical use general made to more is case.,The procedure is thereby made more practical to use in the general case.,7011
inference. and problem frequently extension The diagnostic default multiple arises in,The multiple extension problem arises frequently in diagnostic and default inference.,7012
"reasonable, In of multiple may improbable. diagnosis, however, all however explanations the area the seem","In the area of diagnosis, however, the multiple explanations may all seem reasonable, however improbable.",7013
of them a among Choosing matter preference. is quantitative,Choosing among them is a matter of quantitative preference.,7014
in causally. preference works Quantitative diagnosis knowledge when well is modelled,Quantitative preference works well in diagnosis when knowledge is modelled causally.,7015
We construct can then probabilities compute incrementally as explanations. we,We can then compute probabilities incrementally as we construct explanations.,7016
probable partially first is explanation ordered. most are A explanations if found,A most probable explanation is found first if explanations are partially ordered.,7017
"an surroundings. analysis, often be information In drawn its can pattern regarding object from","In pattern analysis, information regarding an object can often be drawn from its surroundings.",7018
method and based theory theory. possibility on Dempster-Shafer The is,The method is based on Dempster-Shafer theory and possibility theory.,7019
a We database a that scoped temporally manages beliefs. program of present,We present a program that manages a database of temporally scoped beliefs.,7020
operates At any or of calculus. system belief of level this the measure independent belief,At this level the system operates independent of any measure of belief or belief calculus.,7021
This proposed paper an with compares other approaches the on database. image approach,This paper compares the proposed approach with other approaches on an image database.,7022
distributions. paper relationship of between This the and Shafer's convex examines functions sets probability belief,This paper examines the relationship between Shafer's belief functions and convex sets of probability distributions.,7023
implications has theoretic for decision This a extension.,This has implications for a decision theoretic extension.,7024
of rule Dempster's also rule Bayes' conditioning. compared of is combination with,Dempster's rule of combination is also compared with Bayes' rule of conditioning.,7025
I render (l) have evidence enough to S probable. the sentence,(l) I have enough evidence to render the sentence S probable.,7026
"vaguely, probabilistic and embody them shall what inference. Generally I I take call to","Generally and vaguely, I take them to embody what I shall call probabilistic inference.",7027
non-monotonic. This clearly form is of inference,This form of inference is clearly non-monotonic.,7028
this corresponds I inference. what None to of probabilistic call,None of this corresponds to what I call probabilistic inference.,7029
in from multiple Poly-trees singly are arise causal variables which causes. may networks connected,Poly-trees are singly connected causal networks in which variables may arise from multiple causes.,7030
stated conditions these are two to Sufficient methods agree. for,Sufficient conditions are stated for these two methods to agree.,7031
data and if problem involves particularly true noisy is This redundant measurements. the,This is particularly true if the problem involves noisy data and redundant measurements.,7032
of of use winner-take-all groups variables also is The discussed.,The use of winner-take-all groups of variables is also discussed.,7033
The software the to a mood the enhancing artist added helps portrait.,The software helps enhancing the mood the artist added to a portrait.,7034
estimate output be quantity an a of action. uncertain some for may recommendation The or,The output may be an estimate of some uncertain quantity or a recommendation for action.,7035
"Schemer-ll, communication computational architecture, and among problem-solving of subsystems. called for provides The interleaving various","The computational architecture, called Schemer-ll, provides for interleaving of and communication among various problem-solving subsystems.",7036
"refinement, to subsystems execution. These gathering, solution belief and approaches alternative provide construction, information solution","These subsystems provide alternative approaches to information gathering, belief refinement, solution construction, and solution execution.",7037
"knowledge This process. inferencing NAIVE, and low-level describes representation paper language a","This paper describes NAIVE, a low-level knowledge representation language and inferencing process.",7038
those about designed systems nondeterministic has like dynamic reasoning been NAIVE medicine. for in found,NAIVE has been designed for reasoning about nondeterministic dynamic systems like those found in medicine.,7039
uses functions density probability inferencing process these uncertainty. propagate The to,The inferencing process uses these probability density functions to propagate uncertainty.,7040
primitive are of framework valuations. and The the objects variables,The primitive objects of the framework are variables and valuations.,7041
framework marginalization. The primitive the are combination and of operators,The primitive operators of the framework are combination and marginalization.,7042
modus discussed. inference applying are fuzzy by ponens using for rules Different schemas,Different inference schemas for applying fuzzy rules by using modus ponens are discussed.,7043
"have, over theory. people reasons for, probability and have concerns Still, various","Still, people have, and have reasons for, various concerns over probability theory.",7044
"notion. Probability however, formal offers this to counterpart theory, no","Probability theory, however, offers no formal counterpart to this notion.",7045
"But does this provide yet not of induction, theory i.e. a","But this does not yet provide a theory of induction, i.e.",7046
no The that is there problem is completion. good crucial,The crucial problem is that there is no good completion.,7047
"this such suggest paper, want In to I solution. a","In this paper, I want to suggest such a solution.",7048
techniques. paper comparing This evidential at reasoning on reports empirical aimed work,This paper reports on empirical work aimed at comparing evidential reasoning techniques.,7049
"are are not but conclusions novel, instructive. The","The conclusions are not novel, but are instructive.",7050
"winners clear there losers. calculi, Among no the or were","Among the calculi, there were no clear winners or losers.",7051
The application been has theory recently expert systems. for to Dempster-Shafer extended its,The Dempster-Shafer theory has been extended recently for its application to expert systems.,7052
"made between is and ""statistical"" probabilities. distinction ""subjective"" A sometimes","A distinction is sometimes made between ""statistical"" and ""subjective"" probabilities.",7053
"distinction This is events. and between ""repeatable"" a events ""unique"" based on","This is based on a distinction between ""unique"" events and ""repeatable"" events.",7054
very can conception This be applied to of languages. rich probability,This conception of probability can be applied to very rich languages.,7055
more inductive is than The conclusions are logic in premises. that informative,The logic is inductive in that conclusions are more informative than premises.,7056
decision making. future of is component important an Predicting the,Predicting the future is an important component of decision making.,7057
"enough there predictions. to not make however, most In is information situations, accurate","In most situations, however, there is not enough information to make accurate predictions.",7058
"causal of uncertainty. for a In predictive inference under reasoning theory we paper, develop this","In this paper, we develop a theory of causal reasoning for predictive inference under uncertainty.",7059
"In ""conditional"" an measurefree of a was earlier new presented. objects theory paper,","In an earlier paper, a new theory of measurefree ""conditional"" objects was presented.",7060
"the is this placed In upon motivation of theory. the paper, emphasis","In this paper, emphasis is placed upon the motivation of the theory.",7061
through of system. involving an knowledge-based is The motivation part established central example this a,The central part of this motivation is established through an example involving a knowledge-based system.,7062
a published Minsky slightly Selfridge.) first (This extends & result by,(This slightly extends a result first published by Minsky & Selfridge.),7063
of probability. paper a argues modal This for view,This paper argues for a modal view of probability.,7064
concepts. for a expressing Probability logical is probabilistic both language and logic,Probability logic is a language for expressing both probabilistic and logical concepts.,7065
major A involves reasoning of use of approximations. aspect human the,A major aspect of human reasoning involves the use of approximations.,7066
reasoning work is to Our the real-time of concerned approximate with control. application,Our work is concerned with the application of approximate reasoning to real-time control.,7067
environment hardware translating programming rules fuzzy realizations. control for into a We describe,We describe a programming environment for translating fuzzy control rules into hardware realizations.,7068
hardware possible. Two are realizations of methods,Two methods of hardware realizations are possible.,7069
a inferencing. based on First purpose for The fuzzy chip is special,The First is based on a special purpose chip for fuzzy inferencing.,7070
is The second based memory on a chip. simple,The second is based on a simple memory chip.,7071
of knowledge making probabilities examined. for constraints condition linear decision Techniques on are with,Techniques for decision making with knowledge of linear constraints on condition probabilities are examined.,7072
for now inference. exist Methods propagating beliefs chains of through,Methods now exist for propagating beliefs through chains of inference.,7073
"beliefs on of same may different the Moreover, different consequent rules representations the yield hypotheses.","Moreover, different representations of the same rules may yield different beliefs on the consequent hypotheses.",7074
"different two Thus, in the in manifested is meta-levels. uncertainty system","Thus, uncertainty in the system is manifested in two different meta-levels.",7075
whole. paradigms functioning be different a into meta-levels and The integrated must uncertainty,The different uncertainty paradigms and meta-levels must be integrated into a functioning whole.,7076
"and forming natural their Bacteria complex environments infection (""bacteriophages"") coexist viruses networks. in","Bacteria and their viruses (""bacteriophages"") coexist in natural environments forming complex infection networks.",7077
"decrease increasing with bacterial that growth infection. against find should First, defense rate we","First, we find that bacterial growth rate should decrease with increasing defense against infection.",7078
"host should Second, of with infection we efficiency find viral decrease the that range.","Second, we find that the efficiency of viral infection should decrease with host range.",7079
"between curvature Next, relationship history trade-offs. establish life relative densities a of the and we","Next, we establish a relationship between relative densities and the curvature of life history trade-offs.",7080
"the to findings communities. multi-species ""Kill-the-Winner"" and compare current the model We phage-bacteria of contrast","We compare and contrast the current findings to the ""Kill-the-Winner"" model of multi-species phage-bacteria communities.",7081
seems The formalism handle the to examples nonmonotonic literature. the from,The formalism seems to handle the examples from the nonmonotonic literature.,7082
method presented. computing is probabilistic propositions for A,A method for computing probabilistic propositions is presented.,7083
"composed RUM's architecture representation, and of layers: is three control. inference,","RUM's architecture is composed of three layers: representation, inference, and control.",7084
example of situation An assessment is illustrated. naval,An example of naval situation assessment is illustrated.,7085
calculus possesses these presented The features. here,The calculus presented here possesses these features.,7086
Some scheme results with obtained this are discussed.,Some results obtained with this scheme are discussed.,7087
methods with certain I then inference maximum currently reasoning. uncertain compare entropy popular other for,I then compare maximum entropy inference with certain other currently popular methods for uncertain reasoning.,7088
"by set memory a rule (ROM). realized memory a implementation, In is only read this","In this implementation, a rule set memory is realized by a read only memory (ROM).",7089
"processor, the for path out each is inference laid data In rule. one","In the inference processor, one data path is laid out for each rule.",7090
"is processed are but parallel, All in rules rule each executed serially.","All rules are executed in parallel, but each rule is processed serially.",7091
is A two-phase nonoverlapping clocking scheme used.,A two-phase nonoverlapping clocking scheme is used.,7092
"not best indicate approach. research may recent be always the However, that results minimizing","However, recent research results indicate that minimizing may not always be the best approach.",7093
about probabilistic likelihood. Bayesian networks semantics assertions provide qualitative for a,Bayesian networks provide a probabilistic semantics for qualitative assertions about likelihood.,7094
"experts"" The of the information a Bayesian of space ""opinions combine in in fashion. experts","The experts in the space of ""opinions of experts"" combine information in a Bayesian fashion.",7095
We evidence by viewpoint. of combination the alternative for spaces this suggested present,We present alternative spaces for the combination of evidence suggested by this viewpoint.,7096
what commonsense many cases is knowledge consists knowledge usual. In of of,In many cases commonsense knowledge consists of knowledge of what is usual.,7097
paper we usual In system reasoning with develop this for information. a,In this paper we develop a system for reasoning with usual information.,7098
the We this granules. with aid system of implement possibility-probability,We implement this system with the aid of possibility-probability granules.,7099
"neither nor knowledge rule-based requires It of is tools. new frame-based, system and a engineering","It is neither rule-based nor frame-based, and requires a new system of knowledge engineering tools.",7100
with interval-valued looks at essay decision-making measures. probability This,This essay looks at decision-making with interval-valued probability measures.,7101
"information. implement we more approach, order epistemological need to In the","In order to implement the approach, we need more epistemological information.",7102
solution a the procedures to The provide problem uncertain spatial relative relationships. of general estimating,The procedures provide a general solution to the problem of estimating uncertain relative spatial relationships.,7103
"the all mode. using (These Bayes are procedures,","(These are all Bayes procedures, using the mode.",7104
in has systems reasoning often expert calculi. used ad-hoc Evidential uncertainty,Evidential reasoning in expert systems has often used ad-hoc uncertainty calculi.,7105
higher order applied deals reasoning. probabilities Metaprobability to theory with evidential,Metaprobability theory deals with higher order probabilities applied to evidential reasoning.,7106
"features has possible values, Each of three the (e.g. two","Each of the three features has two possible values, (e.g.",7107
"from sensor Beliefs and sensor are accrued A, two B. formed from sensors: evidence","Beliefs are formed from evidence accrued from two sensors: sensor A, and sensor B.",7108
Each defining corresponding sensor feature. the senses,Each sensor senses the corresponding defining feature.,7109
that third feature observed the by is any sensor. Note not,Note that the third feature is not observed by any sensor.,7110
General analyzing considered. a are database in information in probabilistic problems,General problems in analyzing information in a probabilistic database are considered.,7111
represent tree (hidden) conceptual Internal nodes inaccessible variables to observation. of the,Internal nodes of the tree represent conceptual (hidden) variables inaccessible to observation.,7112
"to remains imprecisions. method incremental, is and local, described efficient, The robust measurement","The method described is incremental, local, efficient, and remains robust to measurement imprecisions.",7113
genome-scale network of created We metabolism. have a reconstruction coli Escherichia,We have created a genome-scale network reconstruction of Escherichia coli metabolism.,7114
moves the programs procedure. choosing approach is traditional game-playing in to The minimax,The traditional approach to choosing moves in game-playing programs is the minimax procedure.,7115
is increasing search its general that depth improves use play. underlying belief The,The general belief underlying its use is that increasing search depth improves play.,7116
tree This game called pathology. phenomenon is,This phenomenon is called game tree pathology.,7117
the This strengthen depth uniform causes hypothesis pathology. property two that points that raises win,This property raises two points that strengthen the hypothesis that uniform win depth causes pathology.,7118
This the the evaluation predicts of pathological original failing function. behavior,This failing predicts the pathological behavior of the original evaluation function.,7119
by temporal setting is the compounded factory uncertainty. in Scheduling computational complexity and,Scheduling in the factory setting is compounded by computational complexity and temporal uncertainty.,7120
"hypotheses the Often, evidence hierarchical in and nature. are","Often, the evidence and hypotheses are hierarchical in nature.",7121
other logic of same Bayesian approach. the has reasoning limitations This probabilistic systems the of,This logic has the same limitations of other probabilistic reasoning systems of the Bayesian approach.,7122
"consistency very not common reasoning, sense assumption. is For natural a","For common sense reasoning, consistency is not a very natural assumption.",7123
"this propose we the In logic. of some shall probabilistic paper, extensions","In this paper, we shall propose some extensions of the probabilistic logic.",7124
"not. all the we space section, consistent the In interpretations, second or shall consider of","In the second section, we shall consider the space of all interpretations, consistent or not.",7125
"interval is its each Pls(s)]. [Spt(s), proposition belief For by function an represented s,","For each proposition s, its belief function is represented by an interval [Spt(s), Pls(s)].",7126
"to logic restricting probabilistic get Certainly, we consistent interpretations. by further Nilsson's","Certainly, we get Nilsson's probabilistic logic by further restricting to consistent interpretations.",7127
probabilistic of the theory. the framework probability brings interpretation into This logic,This interpretation brings the probabilistic logic into the framework of probability theory.,7128
have probability We distribution for proposition. each a,We have a probability distribution for each proposition.,7129
"(marginal), conditional joint absolute We probability and compute may distributions.","We may compute absolute (marginal), conditional and joint probability distributions.",7130
"of subsets probabilities, we propositions or get probabilities individual joint appropriate propositions. of summing By","By summing appropriate joint probabilities, we get probabilities of individual propositions or subsets of propositions.",7131
"In relaxation scheme logic. consider a shall for section, we probabilistic last the","In the last section, we shall consider a relaxation scheme for probabilistic logic.",7132
problem idea main consistent vision. the in from arises computer The labeling,The main idea arises from the consistent labeling problem in computer vision.,7133
is to applied This originally scene drawings. line analysis of method,This method is originally applied to scene analysis of line drawings.,7134
are Explanation expert of facilities frameworks. particularly a system feature important,Explanation facilities are a particularly important feature of expert system frameworks.,7135
in had rule-based results. It traditional area which an have mixed is frameworks expert system,It is an area in which traditional rule-based expert system frameworks have had mixed results.,7136
this effect We in that measure operating context any satisfy argue certain properties. must,We argue that any effect measure operating in this context must satisfy certain properties.,7137
"by population coalescent-based proposed rooted A tree et Nielsen model, al. originally","A coalescent-based rooted population tree model, originally proposed by Nielsen et al.",7138
"never model However, the been of proven. has identifiability this","However, the identifiability of this model has never been proven.",7139
probability using version simplified and independence assumptions. conditional a of,and a simplified version of probability using conditional independence assumptions.,7140
Importance of how ranked. the quantitatively We assessed biases illustrate may and be,We illustrate how the Importance of biases may be quantitatively assessed and ranked.,7141
given critical for a is UlS's of robustness application. selecting factor Considerations a might be,Considerations of robustness might be a critical factor is selecting UlS's for a given application.,7142
"calculus (e.g. or certainty a factors), probabilities","probabilities or certainty factors), a calculus (e.g.",7143
"(i.e. or combination and CF semantics Bayesian rules), a","Bayesian or CF combination rules), and a semantics (i.e.",7144
rational syntax studies scope and languages paper the the those This calculus grounds. on of,This paper studies the rational scope of those languages on the syntax and calculus grounds.,7145
about One the of Turing of the unorganized idea of was evolutionary computation roots machines.,One of the roots of evolutionary computation was the idea of Turing about unorganized machines.,7146
holds This observation as ants for as well for fruit flies.,This observation holds for ants as well as for fruit flies.,7147
compelling effect evidence in a also paper married provides versus unmarried The of (i.e. similar,The paper also provides compelling evidence of a similar effect in married versus unmarried (i.e.,7148
the the of regimes. A concerns natural two between dynamic question transition the,A natural question concerns the dynamic of the transition between the two regimes.,7149
"This with used factors. uncertainty, certainty by the examines to called paper MYCIN reason quantities","This paper examines the quantities used by MYCIN to reason with uncertainty, called certainty factors.",7150
shown probabilistic that accommodates is an of number unlimited It this redefinition interpretations.,It is shown that this redefinition accommodates an unlimited number of probabilistic interpretations.,7151
transformations shown be the These p(EIH)/p(El of are likelihood to ratio monotonic interpretations H).,These interpretations are shown to be monotonic transformations of the likelihood ratio p(EIH)/p(El H).,7152
assumptions in in the rarely implicit model are applications. It practical true that is emphasized,It is emphasized that assumptions implicit in the model are rarely true in practical applications.,7153
Methods the suggested. are relaxing assumptions for,Methods for relaxing the assumptions are suggested.,7154
paper a diagnosis of of model development Bayesian for causal describes This appendicitis. the the,This paper describes the development of a causal Bayesian model for the diagnosis of appendicitis.,7155
"a It axiomatic i.e. basis, well-understood yields","It yields a well-understood axiomatic basis, i.e.",7156
"conditional previous on to work confirmation interpret theory. quantitative independence,","conditional independence, to interpret previous work on quantitative confirmation theory.",7157
This flexibility requires existing uncertainty than display. more systems,This requires more flexibility than existing uncertainty systems display.,7158
when tries needs. one to satisfy these,when one tries to satisfy these needs.,7159
associated true/false These set a and their dependencies. propositions maintain systems of,These systems maintain a set of true/false propositions and their associated dependencies.,7160
Systems. of reasoning as paper This Maintenance the addresses to it probabilistic applies problem Truth,This paper addresses the problem of probabilistic reasoning as it applies to Truth Maintenance Systems.,7161
tree-like hierarchical and for formulated inference Strategies are networks Control investigated. probabilistic,Control Strategies for hierarchical tree-like probabilistic inference networks are formulated and investigated.,7162
properties Various in have use entropy information to its led theory. widespread of relative,Various properties of relative entropy have led to its widespread use in information theory.,7163
commitments to explanatory formalise Social justificatory patterns. are and used,Social commitments are used to formalise justificatory and explanatory patterns.,7164
"of The combination justification, planation, and commitments . ex-","The combination of ex- planation, justification, and commitments .",7165
directed Influence probabilities. are diagrams graph a for as representation uncertainties,Influence diagrams are a directed graph representation for uncertainties as probabilities.,7166
"the from of altering hypotheses. the updating-that preclude prevent evidence is, the probabilities","preclude updating-that is, prevent the evidence from altering the probabilities of the hypotheses.",7167
"and a specifying conjunction Each a operator. disjunction calculus negation, defined a is by","Each calculus is defined by specifying a negation, a conjunction and a disjunction operator.",7168
"The i.e. define uncertainty term the set granularity, will","The term set will define the uncertainty granularity, i.e.",7169
distinction finest uncertainty. quantifications among of level the different of,the finest level of distinction among different quantifications of uncertainty.,7170
operators. two will granularity between to the limit ability similar This differentiate,This granularity will limit the ability to differentiate between two similar operators.,7171
formalism problem tool networks as inference used representation. are the Bayesian for space and state,Bayesian inference networks and state space formalism are used as the tool for problem representation.,7172
permit general computation Pearl's localized this Trees of type sense. in,Trees of this general type permit localized computation in Pearl's sense.,7173
conditional judgments based of qualitative They are on independence.,They are based on qualitative judgments of conditional independence.,7174
describe that expert the prove in scheme useful here we will We systems. believe,We believe that the scheme we describe here will prove useful in expert systems.,7175
"demands structure. unrealistic on schemes, the for Bayesian hand, make often other","Bayesian schemes, on the other hand, often make unrealistic demands for structure.",7176
"introduction the a belief functions. section, give In next we to brief","In the next section, we give a brief introduction to belief functions.",7177
characteristics of the together discussed computational methods the from formulas are These with derived them.,These formulas are discussed together with the computational characteristics of the methods derived from them.,7178
linear between with These events of constraints the. compute necessary relations probabilities are to,These linear constraints are necessary to compute probabilities of events with relations between the.,7179
propositional formulas. calculus arbitrarily specified with boolean,arbitrarily specified with propositional calculus boolean formulas.,7180
uncertain represented be applications can inference systems Expert involve by table. that multidimensional contingency a,Expert systems applications that involve uncertain inference can be represented by a multidimensional contingency table.,7181
We a present two rigid-body approach segmentation new views. motion from to,We present a new approach to rigid-body motion segmentation from two views.,7182
Circumscription. Pointwise We a using formalization propose,We propose a formalization using Pointwise Circumscription.,7183
Projection method Rotational novel presents (RoPS). paper named a Statistics This,This paper presents a novel method named Rotational Projection Statistics (RoPS).,7184
techniques. to superior exhibited compared Our techniques existing performance proposed,Our proposed techniques exhibited superior performance compared to existing techniques.,7185
"in on expert designing This uncertain environments. making paper focuses support complex, systems decision to","This paper focuses on designing expert systems to support decision making in complex, uncertain environments.",7186
"or of in are belief belief paper, this In updates measures examined. change","In this paper, measures of change in belief or belief updates are examined.",7187
"properties of are In of in belief the spirit Cox, measure change enumerated. for a","In the spirit of Cox, properties for a measure of change in belief are enumerated.",7188
call between another supports we meta-support. facts supportive relationship fact a this When,When a fact supports another supportive relationship between facts we call this meta-support.,7189
the This reasoning both facilitates about propositional knowledge.,This facilitates reasoning about both the propositional knowledge.,7190
(CHO) genome-scale reconstruction a created cell ovary metabolism. chinese We of have hamster network,We have created a genome-scale network reconstruction of chinese hamster ovary (CHO) cell metabolism.,7191
"Preference any inefficient. of food, one kind be such for might circumstances, under","Preference for any one kind of food, under such circumstances, might be inefficient.",7192
"is the to typically which evidence is stored from Moreover, hypothesis. evidence in direction","Moreover, the direction in which evidence is stored is typically from evidence to hypothesis.",7193
with inference. and inductive dealing -models different to knowledge probabilistic are Two approaches examined,Two different approaches to dealing with probabilistic knowledge are examined -models and inductive inference.,7194
the The basic appears inductive first be sight very for approach issue different. to at,The basic issue for the inductive approach appears at first sight to be very different.,7195
"be of approach some invoked. induction requires this form Clearly, that","Clearly, this approach requires that some form of induction be invoked.",7196
"Of important manageability is additional course, an concern.","Of course, manageability is an important additional concern.",7197
inference uncertainty. the of form on strongly representation depend of rules and justification inductive The,The form and justification of inductive inference rules depend strongly on the representation of uncertainty.,7198
"information. incomplete representation, examines paper generic one This namely,","This paper examines one generic representation, namely, incomplete information.",7199
"problems A.I. Second, it to apply we in","Second, we apply it to problems in A.I.",7200
"which does different matter and these in practice, How you are it use? UISs","How different are these UISs in practice, and does it matter which you use?",7201
information. tool depicting effective Glyph-based multivariate visualization for is an,Glyph-based visualization is an effective tool for depicting multivariate information.,7202
undertaken a in conjunction with work national team. This rugby is,This work is undertaken in conjunction with a national rugby team.,7203
allowed give the interference to responses. are verbal subjects The occurs if least,The least interference occurs if the subjects are allowed to give verbal responses.,7204
in of the quantifiers Ambiguities are meanings resolved interactively.,Ambiguities in the meanings of quantifiers are resolved interactively.,7205
model examines accuracy uncertain the paper PROSPECTOR the This for of reasoning.,This paper examines the accuracy of the PROSPECTOR model for uncertain reasoning.,7206
"performance under within conditions PROSPECTOR's deteriorates. However, identified we even subset, this which","However, even within this subset, we identified conditions under which PROSPECTOR's performance deteriorates.",7207
on developed procedures diagram. the solution advances Recent influence based have,Recent advances have developed solution procedures based on the influence diagram.,7208
for automation systems. for the Mechanisms uncertainty are of expert required,Mechanisms for the automation of uncertainty are required for expert systems.,7209
to mechanisms the need obey probabilistic Sometimes properties of these reasoning.,Sometimes these mechanisms need to obey the properties of probabilistic reasoning.,7210
a does Incidence provide logic probabilistic functional with Calculus connectives. truth,Incidence Calculus does provide a probabilistic logic with truth functional connectives.,7211
types We can rules show how specific defined. be resolution of,We show how specific types of resolution rules can be defined.,7212
"particular, we general braid. chain In and a of a introduce notions the","In particular, we introduce the general notions of a chain and a braid.",7213
valid practical are rarely such We applications. that in argue restrictions,We argue that such restrictions are rarely valid in practical applications.,7214
"performance by approach, our guided both probability. are task and In learning","In our approach, both task performance and learning are guided by probability.",7215
"and normalized located revised, their are and incrementally and errors are Probabilities corrected.","Probabilities are incrementally normalized and revised, and their errors are located and corrected.",7216
"skin  a static techniquesoften model.  In Computer assume colour-based colour Vision, spatial","  In Computer Vision, colour-based spatial techniquesoften assume a static skin colour model.",7217
"lighting can skin perceived colour when changes. change camera However, a by","However, skin colour perceived by a camera can change when lighting changes.",7218
on common multiple environment In the skin. real sources light impinge,In common real environment multiple light sources impinge on the skin.,7219
the We avoiding importance of shapes. when further structured examples estimating in demonstrate misleading non-stationarity,We further demonstrate the importance of non-stationarity in avoiding misleading examples when estimating structured shapes.,7220
and the procedure. numerical We analysis theoretical assessing present performance results the of proposed,We present theoretical analysis and numerical results assessing the performance of the proposed procedure.,7221
fully are environments conform needed. that benefits computing to But to these exploit parallel algorithms,But to fully exploit these benefits algorithms that conform to parallel computing environments are needed.,7222
sequential a well presents This this simulator computing paper environment. to suited posterior,This paper presents a sequential posterior simulator well suited to this computing environment.,7223
depth the objects. facilitates by providing us information foreground of KINECT,KINECT facilitates us by providing the depth information of foreground objects.,7224
is for many step Recovering important shadows an algorithms. vision,Recovering shadows is an important step for many vision algorithms.,7225
time-lapse are that with sequences heuristics. thresholding limited work simple approaches Current to,Current approaches that work with time-lapse sequences are limited to simple thresholding heuristics.,7226
"mining of to application data techniques first propositional propose a this we In paper, satisfiability.","In this paper, we propose a first application of data mining techniques to propositional satisfiability.",7227
noise and all preserving allows while temporal removing discontinuities . spatial This,This allows removing noise while preserving all spatial and temporal discontinuities .,7228
statistical relational (MAP) RockIt maximum query is engine models. for a-posteriori a,RockIt is a maximum a-posteriori (MAP) query engine for statistical relational models.,7229
linear exploits CPA symmetries context-specific sets up bundles of local constraints. and,CPA exploits local context-specific symmetries and bundles up sets of linear constraints.,7230
referred phenomenon Comorosan as The the to effect. is investigated,The phenomenon investigated is referred to as the Comorosan effect.,7231
been by and yeast established co-workers. A mammalian enzymes between has similar Comorosan and shift,A similar shift between mammalian and yeast enzymes has been established by Comorosan and co-workers.,7232
same. is the reported not separation first incidence This the of the being,This is the first reported incidence of the separation not being the same.,7233
enable Modern genomes. DNA genetic variation identify rapidly many among technologies sequencing to human geneticists,Modern DNA sequencing technologies enable geneticists to rapidly identify genetic variation among many human genomes.,7234
functional imaging interactions magnetic via connectomes fluctuations Functional capture the in signal. resonance brain synchronized,Functional connectomes capture brain interactions via synchronized fluctuations in the functional magnetic resonance imaging signal.,7235
"measured If functional the architecture map they of brain. rest, the intrinsic during","If measured during rest, they map the intrinsic functional architecture of the brain.",7236
task-driven brain With they areas. specialized mechanisms integration represent experiments between,With task-driven experiments they represent integration mechanisms between specialized brain areas.,7237
a mitosis cancer Accurate critical of role breast plays in histopathology. detection,Accurate detection of mitosis plays a critical role in breast cancer histopathology.,7238
We discrimination to from other propose quality objects. enable mitosis clinical routine-compliant of,We propose to enable clinical routine-compliant quality of mitosis discrimination from other objects.,7239
of these means demonstrates by several The points paper examples.,The paper demonstrates these points by means of several examples.,7240
"sick bad The make low qualities, AI discomfort. cause signals which aversive us about or","The AI signals about bad low aversive qualities, which make us sick or cause discomfort.",7241
"signaling, and and causes roles dopamine mental and describe of dysfunctions. serotonin We","We describe causes and roles of dopamine and serotonin signaling, and mental dysfunctions.",7242
"describe not analysis current However, methods can texture adequately patterns. of heterogeneous such","However, current methods of texture analysis can not adequately describe such heterogeneous patterns.",7243
"still scales, at but multi-scale verifies preserving patterns the Furthermore, analysis. the different homogeneous analysis","Furthermore, multi-scale analysis verifies the patterns at different scales, but still preserving the homogeneous analysis.",7244
four Tests performed databases dynamic and one static in texture were one.,Tests were performed in four static texture databases and one dynamic one.,7245
systems composite for approaches describes (as combinatorial The development paper (modular) reengineering). (improvement/extension),The paper describes development (improvement/extension) approaches for composite (modular) systems (as combinatorial reengineering).,7246
strategies illustrative including are numerical Various system examples. improvement/extension presented,Various system improvement/extension strategies are presented including illustrative numerical examples.,7247
"tool semi-automatic paper the OntoRich support evaluation. a ontology presents framework, This and for enrichment","This paper presents the OntoRich framework, a support tool for semi-automatic ontology enrichment and evaluation.",7248
streams. is The extract to RSS used WordNet ontology from candidates for enrichment dynamic,The WordNet is used to extract candidates for dynamic ontology enrichment from RSS streams.,7249
metrics. are evaluated against qualitative The several ontologies enriched,The enriched ontologies are evaluated against several qualitative metrics.,7250
Worlds action propose for approximation (PWS) of Possible an planning. the Semantics We,We propose an approximation of the Possible Worlds Semantics (PWS) for action planning.,7251
wrt. A novelty support postdiction for is,A novelty is support for postdiction wrt.,7252
"may the species the However, involve interactions two individuals. multiple between","However, the interactions between the two species may involve multiple individuals.",7253
"multiplayer games. to resort evolutionary To we situation, a analyse such","To analyse such a situation, we resort to evolutionary multiplayer games.",7254
for of multiplayer can features distribution The the be of games crucial benefits. underlying,The underlying features of multiplayer games can be crucial for the distribution of benefits.,7255
"in algorithm decline our ways: a and complete We greedy, one. incomplete, two approach an","We decline our approach in two ways: an incomplete, greedy, algorithm and a complete one.",7256
and a new is and Modeling nascent emotional-cognition wide-open stage in discussions. for ideas therefore,Modeling emotional-cognition is in a nascent stage and therefore wide-open for new ideas and discussions.,7257
power The tests is the of also assessed.,The power of the tests is also assessed.,7258
of and data capabilities to the discrimination actual distances. these Applications identification illustrate homogeneity,Applications to actual data illustrate the discrimination and homogeneity identification capabilities of these distances.,7259
Mobile applications. an has computer the object important vision tracking in role,Mobile object tracking has an important role in the computer vision applications.,7260
"taxonomy target-based a use present tracked In object the algorithms. we to tracking paper, this","In this paper, we use a tracked target-based taxonomy to present the object tracking algorithms.",7261
"extracellular complex in theory cable embedded arbitrarily this generalize In we media. membranes paper, for","In this paper, we generalize cable theory for membranes embedded in arbitrarily complex extracellular media.",7262
"We outline the generalized equations, cable cases. specific then consider","We outline the generalized cable equations, then consider specific cases.",7263
the traditional generalized to this We comparing cable. always the by cable illustrate numerically,We illustrate this numerically always by comparing the generalized cable to the traditional cable.,7264
and proposed living modeling and systems. of study is for A description life-like the formalism,A modeling formalism is proposed for the description and study of living and life-like systems.,7265
real for abstract provides model framework evolution of biological It organisms. life and an conceptual,It provides an abstract conceptual model framework for real life and evolution of biological organisms.,7266
themselves sequences the unexplored. also remains of However indexing utility relatively query the,However the utility of also indexing the query sequences themselves remains relatively unexplored.,7267
of allows the proposed structured wikipedia. mass-collaborative editing arguments style of framework in semantic The,The proposed framework allows mass-collaborative editing of structured arguments in the style of semantic wikipedia.,7268
"setting, and decidability verification. In establish we complexity of this","In this setting, we establish decidability and complexity of verification.",7269
"properties of study here Using we basic combinatorial Omega-trees. functions, some generating","Using generating functions, we study here some basic combinatorial properties of Omega-trees.",7270
of focus We of on number distribution leaves. with the subtrees two the,We focus on the distribution of the number of subtrees with two leaves.,7271
"our method. of we Through demonstrate usefulness experiments, the","Through experiments, we demonstrate the usefulness of our method.",7272
connectivity These an linkages area-specific focus studies. to impose,These linkages impose an area-specific focus to connectivity studies.,7273
"this a paper, algorithm to propose new transformation. spatial we novel make a In image","In this paper, we propose a new algorithm to make a novel spatial image transformation.",7274
The depth for used approach reduce to image bit aims storage. the proposed,The proposed approach aims to reduce the bit depth used for image storage.,7275
the The operator. for based transformation modulus basic of technique proposed the is,The basic technique for the proposed transformation is based of the modulus operator.,7276
predefined goal of is image The whole integer. to multiples transform the into,The goal is to transform the whole image into multiples of predefined integer.,7277
fields image in used scheme a could be compression. as It processing embedded especially other,It could be used as a scheme embedded in other image processing fields especially compression.,7278
coin coins a at to of the ancient aims issuer. given its Classification assigning,Classification of the ancient coins aims at assigning a given coin to its issuer.,7279
for coins. issuers their Various various used symbols,Various issuers used various symbols for their coins.,7280
information incorporating Improvements spatial made this by have been to the technique.,Improvements have been made by incorporating the spatial information to this technique.,7281
the to problem We technique BoVWs for apply classification. and use three coarse-grained our symbols,We apply the BoVWs technique to our problem and use three symbols for coarse-grained classification.,7282
"log-polar to use We BoVWs. tiling circular tiling, tiling information and rectangular to spatial incorporate","We use rectangular tiling, log-polar tiling and circular tiling to incorporate spatial information to BoVWs.",7283
features the analysis ground truth gathered versus measurements. accuracy present a detailed we From,From the gathered features we present a detailed accuracy analysis versus ground truth measurements.,7284
"else nothing is However, rate constants. these known about","However, nothing else is known about these rate constants.",7285
The constraint notion the expresses are solutions FOCUS concentrated. that,The FOCUS constraint expresses the notion that solutions are concentrated.,7286
"suffers rigidity this from constraint semantics. In practice, of its the","In practice, this constraint suffers from the rigidity of its semantics.",7287
"we propose To of FOCUS issue, three the generalizations constraint. tackle this","To tackle this issue, we propose three generalizations of the FOCUS constraint.",7288
each decompositions. as a for well algorithm one We provide complete as filtering discussing,We provide for each one a complete filtering algorithm as well as discussing decompositions.,7289
information. geometric Each symbol is individually by leveraged categorizing learned visual by features,Each symbol is individually learned by categorizing visual features leveraged by geometric information.,7290
"in geometric granularity. naturally Then, define parts of body symbols these fine a contexts","Then, these symbols naturally define geometric contexts of body parts in a fine granularity.",7291
datasets methods. approach art the suggest on large Experiments our state outperforms two of,Experiments on two large datasets suggest our approach outperforms state of the art methods.,7292
methods RNA-Seq study overview the from data. of to an splicing short We available provide,We provide an overview of the methods available to study splicing from short RNA-Seq data.,7293
assembly is transcript de addressed by methods. This reconstruction novo and,This is addressed by transcript reconstruction and de novo assembly methods.,7294
RNA-Seq the data visualization of alternative facilitate splicing. of the tools Various in the context,Various tools facilitate the visualization of the RNA-Seq data in the context of alternative splicing.,7295
"we do describe behind In specific mathematical review, each models not method. this the","In this review, we do not describe the specific mathematical models behind each method.",7296
is Learning Fuzzy This Rule. achieved Backpropagation using,This is achieved using Fuzzy Backpropagation Learning Rule.,7297
analysis demonstrated This the Computational fact Complexity the Algorithm. of by is Neuro-Fuzzy,This fact is demonstrated by the Computational Complexity analysis of the Neuro-Fuzzy Algorithm.,7298
simulated Algorithm software. proposed MATLAB effectively is using The,The proposed Algorithm is simulated effectively using MATLAB software.,7299
comparison a of sources Presented error. is of,Presented is a comparison of sources of error.,7300
the shown of error in converges limiting the that GCM (i). the It is case,It is shown that the error of the GCM converges in the limiting case (i).,7301
"the the The of r-K equation results the logistic equation, model. solution in differential","The solution of the differential equation results in the logistic equation, the r-K model.",7302
experimental Cs profiles Cp- The well. the fitted model,The model fitted the experimental Cp- Cs profiles well.,7303
up possibilities to deconvolution. iterative investigate image for We speed algorithms non-blind,We investigate possibilities to speed up iterative algorithms for non-blind image deconvolution.,7304
faster than computation constellations Fourier the domain. some via allows In this,In some constellations this allows faster computation than via the Fourier domain.,7305
"iterations can Second, be subsets progress, of restricted as computation to of convolutions pixels.","Second, as iterations progress, computation of convolutions can be restricted to subsets of pixels.",7306
"unconstrained recognition However, the drastically. images, rates applied when to drop","However, when applied to unconstrained images, the recognition rates drop drastically.",7307
In image approach we a convex global for this hallucination. propose paper,In this paper we propose a global convex approach for image hallucination.,7308
al. on Tappen is based et Our work the paper of,Our work is based on the paper of Tappen et al.,7309
they image for use non-convex where model hallucination. a,where they use a non-convex model for image hallucination.,7310
with details high-frequency incorporate database to use high-resolution the We a output. images face to,We use a database with face images to incorporate high-frequency details to the high-resolution output.,7311
that convex using show results approach. by We state-of-the-art we can a achieve,We show that we can achieve state-of-the-art results by using a convex approach.,7312
"this in In Making applications Decision of significance. Problems Modeling particular, have technique remarkable","In particular, applications of this Modeling technique in Decision Making Problems have remarkable significance.",7313
other Relation advantages are compared Soft paradigms of Fuzzy discussed. to The,The advantages of Fuzzy Soft Relation compared to other paradigms are discussed.,7314
Our important to several analysis conclusions. leaded,Our analysis leaded to several important conclusions.,7315
"The curves and include clothoid, involute log-aesthetic (equiangular) the curves. logarithmic spiral,","The log-aesthetic curves include the logarithmic (equiangular) spiral, clothoid, and involute curves.",7316
surface filters formulation new on propose the discrete log-aesthetic Furthermore we surface . based,Furthermore we propose new discrete surface filters based on the log-aesthetic surface formulation .,7317
result conclude the We diagram a shape in of parameters.,We conclude the result in a shape diagram of parameters.,7318
to compute. intuitive more simplicity form this and makes The of efficient characterization,The simplicity of this form makes characterization more intuitive and efficient to compute.,7319
"circles. and discussed curves The C-shape of separated are between two contact, transition S-shape","The transition curves discussed are S-shape and C-shape of contact, between two separated circles.",7320
curve one monotone of a curvature A decreasing spiral is of sign. increasing monotone or,A spiral is a curve of monotone increasing or monotone decreasing curvature of one sign.,7321
"inflection spiral point or extreme. Thus, cannot curvature an a have","Thus, a spiral cannot have an inflection point or curvature extreme.",7322
Namely in is modulation. biochemical property remote ATP recycled hydrolyzation,Namely ATP hydrolyzation is recycled in remote biochemical property modulation.,7323
et proposed al. Generalized first by Cornu Ali Spiral The (GCS) was,The Generalized Cornu Spiral (GCS) was first proposed by Ali et al.,7324
a LCG The analysis GCS is the aesthetic generalized that claim supports of curve. indeed,The analysis of LCG supports the claim that GCS is indeed a generalized aesthetic curve.,7325
oscillatory Effective sharply pattern. times to conform irradiation a,Effective irradiation times conform to a sharply oscillatory pattern.,7326
and this extend aspect of latter to The presented confirm the here serve experiments phenomenon.,The experiments presented here serve to confirm and extend this latter aspect of the phenomenon.,7327
perform is chains the automatically algorithm and An hierarchical initialization presented clustering. to using Markov,An algorithm is presented to automatically perform the initialization using Markov chains and hierarchical clustering.,7328
values paper handwritten in records. temperature automatic recognition addresses weather the of This,This paper addresses the automatic recognition of handwritten temperature values in weather records.,7329
detection profiles. based localization table on cells using The is line projection of,The localization of table cells is based on line detection using projection profiles.,7330
"line based stroke-preserving method removal a images gradient proposed. is on which is Further,","Further, a stroke-preserving line removal method which is based on gradient images is proposed.",7331
"the resistance. of the introduction by been has drugs new Repeatedly, of evolution followed","Repeatedly, the introduction of new drugs has been followed by the evolution of resistance.",7332
distributions al. is et in Nakayama for applied general This originally idea Fisher-Bingham,This idea is originally applied for general Fisher-Bingham distributions in Nakayama et al.,7333
time. OR-gate turns XOR-gate into a over that an,an OR-gate that turns into a XOR-gate over time.,7334
local problems. could the and both This strategy solve maxima outlier the,This strategy could solve both the outlier and the local maxima problems.,7335
and change diversity? why does How mortality generate this,How and why does mortality change generate this diversity?,7336
deal to solvers propose extension this We an of with SAT CDCL-based fundamental problem.,We propose an extension of CDCL-based SAT solvers to deal with this fundamental problem.,7337
the edges ones detects This than conventional method. accurately the Sobel the obtained by technique,This technique detects the edges accurately than the ones obtained by the conventional Sobel method.,7338
proposed The merit of the the proposed of the substantiate technique. results method,The results of the proposed method substantiate the merit of the proposed technique.,7339
"the consider model, binary integrate-and-fire We the process. the neuron leaky Hawkes model, and","We consider the binary neuron model, the leaky integrate-and-fire model, and the Hawkes process.",7340
closed covariance. form for Both the classes allow solutions,Both classes allow closed form solutions for the covariance.,7341
The framework results transfer unified between to enables models. us,The unified framework enables us to transfer results between models.,7342
applicable suitable averages. network approach and Our general is to population for structures,Our approach is applicable to general network structures and suitable for population averages.,7343
in-degree. exact and out-degree The fixed are for derived approximate architectures fixed network averages for,The derived averages are exact for fixed out-degree network architectures and approximate for fixed in-degree.,7344
analysis. role The plays critical $D$ number of our a condition in,The condition number of $D$ plays a critical role in our analysis.,7345
our consistent with results. theoretical are Numerical simulations,Numerical simulations are consistent with our theoretical results.,7346
"measures, are distinguishing in on community types. other effective hand, OTU-based the less","OTU-based measures, on the other hand, are less effective in distinguishing community types.",7347
developed. model framework counterions in vibrations studied of are of with phenomenological conformational Z-DNA The,The conformational vibrations of Z-DNA with counterions are studied in framework of phenomenological model developed.,7348
The Z-DNA. results spectra the experimental Raman obtained describe well of,The obtained results describe well the experimental Raman spectra of Z-DNA.,7349
"to generalize we this high-dimensional In of Bernoulli templates. learning their theory mixture paper,","In this paper, we generalize their theory to learning mixture of high-dimensional Bernoulli templates.",7350
We the theoretical by illustrate synthetic results real and examples.,We illustrate the theoretical results by synthetic and real examples.,7351
experiments Adaptation and Flower set. data We set Domain conduct data on Oxford,We conduct experiments on Domain Adaptation data set and Oxford Flower data set.,7352
The improve all proposed method from standard that cases. demonstrate descriptors results in our,The results demonstrate that our proposed method improve from standard descriptors in all cases.,7353
data simulations illustrate results our real a theoretical with We experiment. and,We illustrate our theoretical results with simulations and a real data experiment.,7354
is the proposed algorithm improvised An work Yoshimoto Harada. of on and based,An improvised algorithm is proposed based on the work of Yoshimoto and Harada.,7355
remains intense efforts Its decades. despite etiology research over enigmatic many,Its etiology remains enigmatic despite intense research efforts over many decades.,7356
competitive with a rim. to core Hypercellularity leads a contact-inhibited proliferating,Hypercellularity leads to a contact-inhibited core with a competitive proliferating rim.,7357
"changing coexist, the whole. Several of combinations can the phenotypes of fitness","Several combinations of phenotypes can coexist, changing the fitness of the whole.",7358
"exact  applications, distance  very the important. many of In geodesic computation graphics is","  In many graphics applications, the computation of exact geodesic distance is very important.",7359
the number simple to windows. also of control a effective adopt yet total We strategy,We also adopt a simple yet effective strategy to control the total number of windows.,7360
the domain treasure. and pathologist's is a therapy personalized targeted Identifying for targets,Identifying targets for personalized targeted therapy is the pathologist's domain and a treasure.,7361
tests suggested content described. somatic report for is The pathology mutation,The suggested pathology report content for somatic mutation tests is described.,7362
is necessary Visuo-spatial analysis multi-level step analysis. for a,Visuo-spatial analysis is a necessary step for multi-level analysis.,7363
"position seems such represent to to codes' cortex `grid accurately. multi-scale use stochastic periodic Indeed,","Indeed, cortex seems to use such stochastic multi-scale periodic `grid codes' to represent position accurately.",7364
also this are that for statistically hold Does not finding neurons independent?,Does this finding also hold for neurons that are not statistically independent?,7365
"correlations the of mismatches grids neurons, increasing the noise For the fall between rapidly. two","For increasing mismatches between the grids of the two neurons, the noise correlations fall rapidly.",7366
on dictionary new enhance proposes based method paper image This a learning.,This paper proposes a new image enhance method based on dictionary learning.,7367
"atoms. image adjusts by of the rarity proposed method the the Particularly, manipulating dictionary","Particularly, the proposed method adjusts the image by manipulating the rarity of dictionary atoms.",7368
"Firstly, on coding sparse blocks. learn algorithms dictionary the through divided sub-image","Firstly, learn the dictionary through sparse coding algorithms on divided sub-image blocks.",7369
"the rarity the sparse statistics coefficients. corresponding of atoms Secondly, of on dictionary compute","Secondly, compute the rarity of dictionary atoms on statistics of the corresponding sparse coefficients.",7370
"the Thirdly, form according and specific application a new to rarity adjust dictionary.","Thirdly, adjust the rarity according to specific application and form a new dictionary.",7371
"image sparse and dictionary the the coefficients. updated reconstruct Finally, using","Finally, reconstruct the image using the updated dictionary and sparse coefficients.",7372
The results experimental the proposed demonstrate of image enhance the method. effectiveness,The experimental results demonstrate the effectiveness of the proposed image enhance method.,7373
been article developed. this segmentation a color for In algorithm novel image has,In this article a novel algorithm for color image segmentation has been developed.,7374
"formulate channels integrated again that, to color the After are new modified new image. a","After that, the new modified channels are integrated again to formulate a new color image.",7375
distortion. of suffers The kind resulted image from some,The resulted image suffers from some kind of distortion.,7376
This by very significant process eye. the ocular looks,This process looks very significant by the ocular eye.,7377
results were proposed variety test presented a support the Experimental images algorithm. on to of,Experimental results were presented on a variety of test images to support the proposed algorithm.,7378
"into it EA. all easily turned a (EA), be Like Algorithms Pareto-based Evolutionary Multi-Objective can","Like all Evolutionary Algorithms (EA), it can easily be turned into a Pareto-based Multi-Objective EA.",7379
of also notion to extend the closure We component. rational Abox the,We also extend the notion of rational closure to the Abox component.,7380
computational evolution predict design networks to features of use processing ligand We categorization.,We use computational evolution to predict design features of networks processing ligand categorization.,7381
of a immune response The as is considered important case-study. early problem,The important problem of early immune response is considered as a case-study.,7382
be models substructures of recognition. immune Corresponding in identified current network can,Corresponding network substructures can be identified in current models of immune recognition.,7383
and biochemical between work deep immune Our a adaptation. recognition analogy draws,Our work draws a deep analogy between immune recognition and biochemical adaptation.,7384
"profiling. than beyond and also This sentiment much classical customer analysis, goes further","This goes beyond classical sentiment analysis, and also much further than customer profiling.",7385
used statistical emulating models expensive process simulators. (GP) metamodels are computer commonly Gaussian for,Gaussian process (GP) models are commonly used statistical metamodels for emulating expensive computer simulators.,7386
a model by of This implements version modified the paper slightly et Ranjan proposed al.,This paper implements a slightly modified version of the model proposed by Ranjan et al.,7387
for used test a comparison popular with are functions R Several mlegp. package performance,Several test functions are used for performance comparison with a popular R package mlegp.,7388
a of segmentation based present images. a novel representation on hierarchical We algorithm,We present a novel segmentation algorithm based on a hierarchical representation of images.,7389
has three exploratory approach main This goals.,This exploratory approach has three main goals.,7390
the in Simple objects tree articulated decade. for models last prevails,Simple tree models for articulated objects prevails in the last decade.,7391
VOC challenge further images the findings. on Experiments support from our animal,Experiments on animal images from the VOC challenge further support our findings.,7392
"impaired. structure above sequence However, recombination) hence difference threshold, is (and a","However, above a sequence difference threshold, structure (and hence recombination) is impaired.",7393
"algorithm bandits. con-textual Exponentiated LINUCB, Gradient multi-armed for We an present","We present Exponentiated Gradient LINUCB, an algorithm for con-textual multi-armed bandits.",7394
of uses optimal to Exponentiated find LINUCB. algorithm Gradient the the exploration This,This algorithm uses Exponentiated Gradient to find the optimal exploration of the LINUCB.,7395
results demonstrate algorithms. experimental The that algorithm surveyed outperforms our,The experimental results demonstrate that our algorithm outperforms surveyed algorithms.,7396
surface example with use car splines aesthetics body of is of given. the The modelling,The example of car body surface modelling with the use of aesthetics splines is given.,7397
the for representations. of of three dots compare identification needed number average among We types,We compare the average number of dots needed for identification among three types of representations.,7398
will predicts the results similar. Our three from that hypothesis be representations the,Our hypothesis predicts that the results from the three representations will be similar.,7399
"affect our these we shape does From triangulation recognition. that conclude experiments,","From these experiments, we conclude that triangulation does affect our shape recognition.",7400
"can be However, tests how universal far taken?","However, how far can universal tests be taken?",7401
"paper space-time terms subjects, resolution, in of environments, This question and this analyses interfaces. rewards","This paper analyses this question in terms of subjects, environments, space-time resolution, rewards and interfaces.",7402
means be adaptive. must that universal necessarily This tests,This means that universal tests must be necessarily adaptive.,7403
"in to water are biology, organs. in from important molecules Ions","Ions in water are important in biology, from molecules to organs.",7404
"gas. a noninteracting are in in Classically, as perfect treated ions ideal particles water","Classically, ions in water are treated as ideal noninteracting particles in a perfect gas.",7405
was energy ion of free zero. Excess,Excess free energy of ion was zero.,7406
"interactions ions deal available to with was with not or consistently Mathematics flows, or boundaries.","Mathematics was not available to deal consistently with flows, or interactions with ions or boundaries.",7407
in ions because interact. and flow needed conditions are Non-classical approaches biological,Non-classical approaches are needed because ions in biological conditions flow and interact.,7408
multiscale deal interactions is with approach to variational and A needed flow.,A variational multiscale approach is needed to deal with interactions and flow.,7409
ascribed conformation behavior by Non-ideal changes. been to has classically mediated protein allosteric interactions,Non-ideal behavior has classically been ascribed to allosteric interactions mediated by protein conformation changes.,7410
the of Computation multiscale approach. alternative popular a to atoms is all,Computation of all atoms is a popular alternative to the multiscale approach.,7411
from systems very different exist atomic on scales motion. Biological,Biological systems exist on very different scales from atomic motion.,7412
called set. generalized a In new soft this paper we present concept neutrosophic,In this paper we present a new concept called generalized neutrosophic soft set.,7413
study this also properties We some concept. of,We also study some properties of this concept.,7414
Some set. been and neutrosophic on soft definitions operations generalized have introduced,Some definitions and operations have been introduced on generalized neutrosophic soft set.,7415
application of in soft problem. we generalized making present set an Finally neuutrosophic decision,Finally we present an application of generalized neuutrosophic soft set in decision making problem.,7416
when ingredient today Parameter is tuning tackling as an a crucial problem. recognized optimization,Parameter tuning is recognized today as a crucial ingredient when tackling an optimization problem.,7417
the also proposed issues Other are discussed. tuning parameter pertaining in context to,Other issues pertaining to parameter tuning in the proposed context are also discussed.,7418
Belief combination Expectation Loopy of Maximization a inference perform Propagation. We using and,We perform inference using a combination of Expectation Maximization and Loopy Belief Propagation.,7419
quality depends context on video tracking usually Object (e.g.,Object tracking quality usually depends on video context (e.g.,7420
"for learned phase, offline parameters video are satisfactory an context In tracking clusters.","In an offline phase, satisfactory tracking parameters are learned for video context clusters.",7421
about native Much been outcompeting written introduced rainbow and interbreeding cutthroat trout trout. has,Much has been written about introduced rainbow trout interbreeding and outcompeting native cutthroat trout.,7422
"polyps colon major cancer, to a Colorectal are health precursors important problem.","Colorectal polyps are important precursors to colon cancer, a major health problem.",7423
then the is analyzed video for presence The polyps. of sequence,The video sequence is then analyzed for the presence of polyps.,7424
two logistic populations model to study two-dimensional with paper genders. This a considers,This paper considers a two-dimensional logistic model to study populations with two genders.,7425
that are learning. We order-of-magnitude speedups parallelizing by also possible show,We also show that order-of-magnitude speedups are possible by parallelizing learning.,7426
facial propose recognition. this approach for easiest  expression  In we an paper,  In this paper we propose an easiest approach for facial expression recognition.,7427
SVM are Here Classification. Expression for concept we using of,Here we are using concept of SVM for Expression Classification.,7428
problem in is divided sub three main Main modules.,Main problem is sub divided in three main modules.,7429
using in which one and segmentation. skin are filter First Face Face detection we is,First one is Face detection in which we are using skin filter and Face segmentation.,7430
We feature more on are stress Extraction. given,We are given more stress on feature Extraction.,7431
enough required. is effective application is method where fast This for execution,This method is effective enough for application where fast execution is required.,7432
"for essential Extraction is Second, expression recognition. part Feature Facial which","Second, Facial Feature Extraction which is essential part for expression recognition.",7433
Edge Analysis. Projection used we this module In,In this module we used Edge Projection Analysis.,7434
towards features Recognition. passed for SVM Expression classifier vector extracted Finally is,Finally extracted features vector is passed towards SVM classifier for Expression Recognition.,7435
"are and six Disgust, Expressions Fear, We basic Joy, Surprise) . considering Sadness, (Anger,","We are considering six basic Expressions (Anger, Fear, Disgust, Joy, Sadness, and Surprise) .",7436
of classical Uniform Latin generation problem. squares is random a,Uniform random generation of Latin squares is a classical problem.,7437
of analysis. It recent makes for graph tools use,It makes use of recent tools for graph analysis.,7438
doubts raise serious results However the their about interpretation conclusion. and this,However the results and their interpretation raise serious doubts about this conclusion.,7439
Influenza human of a challenging problem. virus presents dynamics theoretical A evolutionary The,The evolutionary dynamics of human Influenza A virus presents a challenging theoretical problem.,7440
"United States. a is recorded Bay, method evaluated The dataset in on Massachusetts","The method is evaluated on a dataset recorded in Massachusetts Bay, United States.",7441
"extent due time-consuming is iteration. However, it some still the inner to to","However, it is still time-consuming to some extent due to the inner iteration.",7442
composed The neurons. of binding network is,The network is composed of binding neurons.,7443
"in mechanisms delays. are implemented propagation the Two namely, strength and/or adjusting synaptic network, learning","Two learning mechanisms are implemented in the network, namely, adjusting synaptic strength and/or propagation delays.",7444
the in is This inhibition-controlled network. as interpreted binding,This is interpreted as inhibition-controlled binding in the network.,7445
We extensional constraints constraints). n-ary consider (table,We consider n-ary extensional constraints (table constraints).,7446
"an play role Prediction Image Processing, Science. Image and Optimization key a of Computer in","Image Processing, Optimization and Prediction of an Image play a key role in Computer Science.",7447
captured segmentation by for camera. outdoor surveillance a scenes This semantic method a proposes paper,This paper proposes a semantic segmentation method for outdoor scenes captured by a surveillance camera.,7448
types combines proposed The approach of information. two different,The proposed approach combines two different types of information.,7449
"is regions. First, color performed the into perceptually segmentation to similar scene divide","First, color segmentation is performed to divide the scene into perceptually similar regions.",7450
"generative done Bayesian is classifier. as prediction Na\""ive a a Network The using","The prediction is done using a Na\""ive Bayesian Network as a generative classifier.",7451
provide of for useful systems. and analysis methods complex Mathematical design  the framework ,  Mathematical methods provide useful framework for the analysis and design of complex systems.,7452
"the we analytical estimates. develop Further, bounds computational upper for error","Further, we develop analytical upper bounds for the computational error estimates.",7453
in communities. invoked coexistence species are trade-offs Ecological explain often species between to ecological,Ecological trade-offs between species are often invoked to explain species coexistence in ecological communities.,7454
a Here of a present model community characterization. allows a which we such plant,Here we present a model of a plant community which allows such a characterization.,7455
and Species differ ability. in stress tolerance competitive both,Species differ both in stress tolerance and competitive ability.,7456
to ability win Competitive the the propensity sites. empty for quantified competition is as,Competitive ability is quantified as the propensity to win the competition for empty sites.,7457
"abundances. discrete-time for species system We derive deterministic, the dynamical the","We derive the deterministic, discrete-time dynamical system for the species abundances.",7458
equilibrium. in stable We plant coexist the prove species can a under which conditions,We prove the conditions under which plant species can coexist in a stable equilibrium.,7459
II the HapMap Phase has protocol The been to data. applied,The protocol has been applied to the HapMap Phase II data.,7460
"Moreover, previous reported those four least studies. than smaller the in are times panels at","Moreover, the panels are at least four times smaller than those reported in previous studies.",7461
challenge of basis understanding of these The the associations. next consists molecular,The next challenge consists of understanding the molecular basis of these associations.,7462
"gene of this association integration including can contribute expression The datasets, goal. multiple datasets, to","The integration of multiple association datasets, including gene expression datasets, can contribute to this goal.",7463
"adaptive into entry hypotheses of zones elusive. remain However, tests Simpson's new for the direct","However, direct tests of Simpson's hypotheses for the entry into new adaptive zones remain elusive.",7464
through problem. Experimental is a analysis face detection provided,Experimental analysis is provided through a face detection problem.,7465
obtained has results that and are method value. the encouraging demonstrate practical The,The obtained results are encouraging and demonstrate that the method has practical value.,7466
"its However, a task. complicated design is","However, its design is a complicated task.",7467
constrained approximate We a corresponds that to factored this show inference. expectation,We show that this corresponds to a factored expectation constrained approximate inference.,7468
dynamic very (HDR). world of the have high real scenes range a Most,Most of the real world scenes have a very high dynamic range (HDR).,7469
model the Bayes Gaussian structured the We using a parameters latent variational of approach. infer,We infer the parameters of the latent Gaussian model using a structured variational Bayes approach.,7470
is theory of Kin a causal kind analysis. selection,Kin selection theory is a kind of causal analysis.,7471
"selection ascribed initial kin to relatedness. The benefits, costs, of genetic form and cause","The initial form of kin selection ascribed cause to costs, benefits, and genetic relatedness.",7472
causal analysis because Controversy followed inevitably views. opposing attracts,Controversy followed because causal analysis inevitably attracts opposing views.,7473
It possible into different causes. is separate always to component effects total,It is always possible to separate total effects into different component causes.,7474
"significant example, group a particular causal For is selection limitations. advantages and with certain scheme","For example, group selection is a particular causal scheme with certain advantages and significant limitations.",7475
on This analysis. Curvelet paper model on focuses coefficients improved edge based,This paper focuses on improved edge model based on Curvelet coefficients analysis.,7476
representation of edge. multiresolution object powerful for a with is Curvelet tool transform anisotropic,Curvelet transform is a powerful tool for multiresolution representation of object with anisotropic edge.,7477
"letter, the (NLM) shrinkage we denoising. problem investigate this image In the non-local for means","In this letter, we investigate the shrinkage problem for the non-local means (NLM) image denoising.",7478
propose We blockwise fast constant algorithm also complexity shrinkage. a allowing,We also propose a constant complexity algorithm allowing fast blockwise shrinkage.,7479
image be patchwise Similar other applicable denoising can to techniques. ideas,Similar ideas can be applicable to other patchwise image denoising techniques.,7480
for successful inference demonstrate that is vertices We out-of-sample possible. the then,We then demonstrate that successful inference for the out-of-sample vertices is possible.,7481
images. using processing objects in Image detecting for for radar is,Image processing is using for detecting for objects in radar images.,7482
detectors time. have edge developed over Many classical been,Many classical edge detectors have been developed over time.,7483
the distribution first and to masks been second for Gaussian has derivative. build used Also,Also Gaussian distribution has been used to build masks for the first and second derivative.,7484
"this shape. limit distribution However, to only symmetric has","However, this distribution has limit to only symmetric shape.",7485
images masks constructed to applied we obtained The and are good results.,The constructed masks are applied to images and we obtained good results.,7486
proposed. novel image is background grayscale A from segmenting method for bright dark for objects,A novel method for segmenting bright objects from dark background for grayscale image is proposed.,7487
used of Relationships Recruitment predict population are often dynamics. to (SRRs) Stock fish Models,Models of Stock Recruitment Relationships (SRRs) are often used to predict fish population dynamics.,7488
unit was drug prescription the of line. The analysis,The unit of analysis was the drug prescription line.,7489
"negatives. were most errors the false diagnostic by system, the Unlike pharmacists made alert","Unlike the alert system, most diagnostic errors made by the pharmacists were false negatives.",7490
analysis with from along details analysis. implementation Complexity drawn and experimental are conclusions provided,Complexity analysis and implementation details are provided along with conclusions drawn from experimental analysis.,7491
is pedagogical paper no with This results. a purely new,This is a purely pedagogical paper with no new results.,7492
obvious A* several The use their to take in is maximum. to admissible heuristics way,The obvious way to use several admissible heuristics in A* is to take their maximum.,7493
In the on reduce aim paper computing to this spent heuristics. we time,In this paper we aim to reduce the time spent on computing heuristics.,7494
causes malaria. the falciparum parasite severe the of that is Plasmodium form most,Plasmodium falciparum is the parasite that causes the most severe form of malaria.,7495
"be thus presents analysis are that challenges their SSNs addressed. should quasi-complete networks, different","SSNs are quasi-complete networks, thus their analysis presents different challenges that should be addressed.",7496
Statistical spectrum study. with are site of of Lambda-coalescents frequency the objects associated properties our,Statistical properties of the site frequency spectrum associated with Lambda-coalescents are our objects of study.,7497
"and this about the summarize the event. editorial, background facts organizers In","In this editorial, the organizers summarize facts and background about the event.",7498
"to function this sine the Kendall's transformation. tau are context, In connected a through correlations","In this context, the correlations are connected to Kendall's tau through a sine function transformation.",7499
"$\hat{\Sigma}$ Kendall's with natural is tau plug-in for statistic. a estimate Hence, estimator the $\Sigma$","Hence, a natural estimate for $\Sigma$ is the plug-in estimator $\hat{\Sigma}$ with Kendall's tau statistic.",7500
$\hat{\Sigma}-\Sigma$. bound obtain the a of We operator first sharp on norm,We first obtain a sharp bound on the operator norm of $\hat{\Sigma}-\Sigma$.,7501
entirely estimation are All data-driven. procedures our of,All of our estimation procedures are entirely data-driven.,7502
Transcriptional delay networks. can the gene dynamics significantly impact of,Transcriptional delay can significantly impact the dynamics of gene networks.,7503
such bistable we examine delay systems. Here how affects,Here we examine how such delay affects bistable systems.,7504
"introduce analytically tractable explain a reduced we model. non-Markovian, this, To","To explain this, we introduce a non-Markovian, analytically tractable reduced model.",7505
we that manner. behaves systems simulate Each of bistable in this the,Each of the bistable systems that we simulate behaves in this manner.,7506
relevant find is The also advised on how software to for R proteomics. reader,The reader is also advised on how to find relevant R software for proteomics.,7507
physical closings a of of flooding all properties surface. Reconstruction have a topographic,Reconstruction closings have all properties of a physical flooding of a topographic surface.,7508
a On and edges each basin a catchment represents connect node RAG neighboring nodes.,On a RAG each node represents a catchment basin and edges connect neighboring nodes.,7509
weighted The regions. by are of both the adjacent between altitude pass the point edges,The edges are weighted by the altitude of the pass point between both adjacent regions.,7510
represented on flood is the the type each nodes on The of graphs. of level,The level of the flood is represented on the nodes on each type of graphs.,7511
may a TG on be a or flooding The modeled thus same RAG. on,The same flooding may thus be modeled on a TG or on a RAG.,7512
or path weighted by This node flood using shortest permits to edge algorithms. remark graphs,This remark permits to flood edge or node weighted graphs by using shortest path algorithms.,7513
inference. a way such for higher-order computational account take exhibit to We systematically terms into,We exhibit a way to systematically take into account such higher-order terms for computational inference.,7514
"advanced for post-processing. relaxations schemes along Polynomially with are as rounding studied solvable well,","Polynomially solvable relaxations are studied as well, along with advanced rounding schemes for post-processing.",7515
damaged Image of inpainting art an predicting image. is of regions the,Image inpainting is the art of predicting damaged regions of an image.,7516
time consuming. inpainting a way of is manual image The,The manual way of image inpainting is a time consuming.,7517
"machine multimodal difficult. distributions probabilistic inference learning involving quite is statistics, and In","In machine learning and statistics, probabilistic inference involving multimodal distributions is quite difficult.",7518
based so-called reasoning qualitative calculi. on Qualitative temporal spatial and is,Qualitative spatial and temporal reasoning is based on so-called qualitative calculi.,7519
implications properties reasoning calculi on algorithms. Algebraic of have these several,Algebraic properties of these calculi have several implications on reasoning algorithms.,7520
calculus? a exactly qualitative what is But,But what exactly is a qualitative calculus?,7521
these meet do qualitative And demands? which to the extent proposed calculi,And to which extent do the qualitative calculi proposed meet these demands?,7522
many applications. one steps crucial is of unmixing hyperspectral Hyperspectral for the,Hyperspectral unmixing is one of the crucial steps for many hyperspectral applications.,7523
hyperspectral for model a robust paper presents unmixing. This unsupervised,This paper presents a robust model for unsupervised hyperspectral unmixing.,7524
"naturally model our addition, constraints, can generate sparse abundances. sparsity In with","In addition, with sparsity constraints, our model can naturally generate sparse abundances.",7525
manner. knowledge efficient a representation They for tool structured an in become,They become an efficient tool for knowledge representation in a structured manner.,7526
than more ontology for is the There always one domain. same,There is always more than one ontology for the same domain.,7527
"Because between low. reasons, is ontologies interoperability of very these the","Because of these reasons, interoperability between the ontologies is very low.",7528
"use tools mostly Current and inference edit ontology. ontology build, to functions the","Current ontology tools mostly use functions to build, edit and inference the ontology.",7529
domain tools. for included in heterogeneous are ontologies not merging Methods most,Methods for merging heterogeneous domain ontologies are not included in most tools.,7530
proposed. mechanism of A of the is in oligomer replication world proto-RNAs,A mechanism of the replication of proto-RNAs in oligomer world is proposed.,7531
simulation the the By acquires computer products diversity of and complexity. replication,By computer simulation the products of the replication acquires diversity and complexity.,7532
the complexity Such bases of and are evolution. the diversity,Such diversity and complexity are the bases of the evolution.,7533
"SOAPdenovo-Trans, present novo a specifically assembler designed Results: de transcriptome RNA-Seq. for We","Results: We present SOAPdenovo-Trans, a de novo transcriptome assembler designed specifically for RNA-Seq.",7534
datasets performance and transcriptome on was mouse. evaluated from rice Its,Its performance was evaluated on transcriptome datasets from rice and mouse.,7535
"contiguity, faster Our that SOAPdenovo-Trans and redundancy, is execution. higher lower conclusion provides","Our conclusion is that SOAPdenovo-Trans provides higher contiguity, lower redundancy, and faster execution.",7536
in animals. Migration trait is a fundamental humans and,Migration is a fundamental trait in humans and animals.,7537
findings the of the for migration These significance long-range evolution adaptive of demonstrate cooperation.,These findings demonstrate the significance of adaptive long-range migration for the evolution of cooperation.,7538
"in finite. population are However, sizes reality,","However, in reality, population sizes are finite.",7539
"playing only on most However, of focus populations games. single studies these","However, most of these studies focus on populations playing only single games.",7540
"was the invitation wrote I the revoked. After review,","After I wrote the review, the invitation was revoked.",7541
"which cytometry, often for used In are patient samples, critical. issues multiparametric is analyzing such","In multiparametric cytometry, which is often used for analyzing patient samples, such issues are critical.",7542
probability JCM distribution. population multivariate a models every with robust,JCM models every population with a robust multivariate probability distribution.,7543
"applications systems-level variation, supports cohorts. practical biomedical JCM By involving tackling large","By tackling systems-level variation, JCM supports practical biomedical applications involving large cohorts.",7544
permanent We a of by computation of binary n the n the consider matrix.,We consider the computation of the permanent of a binary n by n matrix.,7545
computation is exact complete It well- that known problem. is a #P the,It is well- known that the exact computation is a #P complete problem.,7546
in issue AGI. Universal crucial induction is a,Universal induction is a crucial issue in AGI.,7547
efficiently. subsequent This more machine solving updatable should tasks for be,This machine should be updatable for solving subsequent tasks more efficiently.,7548
extinction dynamics the close is metapopulations of Understanding management. for vital to of importance,Understanding the dynamics of metapopulations close to extinction is of vital importance for management.,7549
that interactions. features tRNA-protein for enriched are CIFs tRNA underlie,tRNA CIFs are enriched for features that underlie tRNA-protein interactions.,7550
"propose imaging compressive paper, this In lensless a we architecture.","In this paper, we propose a lensless compressive imaging architecture.",7551
"components, of two a The an sensor. aperture architecture assembly consists and","The architecture consists of two components, an aperture assembly and a sensor.",7552
dimensional aperture elements. The array aperture assembly of a two consists of,The aperture assembly consists of a two dimensional array of aperture elements.,7553
element independently The of aperture is controllable. transmittance each,The transmittance of each aperture element is independently controllable.,7554
single The element. detection is a sensor,The sensor is a single detection element.,7555
architecture is and reliable proposed used. is simple because The lens no,The proposed architecture is simple and reliable because no lens is used.,7556
images used with may sensors single to aperture a assembly capture be simultaneously. multi-view Multiple,Multiple sensors may be used with a single aperture assembly to capture multi-view images simultaneously.,7557
earlier. are here obtained developed ITC results those comparable based to,ITC based results obtained here are comparable to those developed earlier.,7558
coordinated intrinsic a complex Cell by signals. environmental is and process size in growth,Cell growth in size is a complex process coordinated by intrinsic and environmental signals.,7559
"alternative model studied. are growth the space, models To further rules and division explore","To further explore the model space, alternative growth models and division rules are studied.",7560
"a alternative, Liu robust As [J. and Han","As a robust alternative, Han and Liu [J.",7561
"both not any cases, moment need condition. In do we","In both cases, we do not need any moment condition.",7562
decisions based their outcomes. make Animals  humans and expected  on,  Animals and humans make decisions based on their expected outcomes.,7563
"for derive subjective mathematical theory, representation the we Using expression of a time. this","Using this theory, we derive a mathematical expression for the subjective representation of time.",7564
"objects number the monitored of framework, In region, i.e. the in our presenting","In our framework, the number of objects presenting in the monitored region, i.e.",7565
"space specified sampling variables, does be need not labeling the of beforehand. to","the sampling space of labeling variables, does not need to be specified beforehand.",7566
"Instead, fly. determined can it the be on automatically","Instead, it can be determined automatically on the fly.",7567
distances distances transportation are histograms. family parameterized Optimal of fundamental a for,Optimal transportation distances are a fundamental family of parameterized distances for histograms.,7568
distances We the classical over problem. improved performance benchmark transportation MNIST also report on optimal,We also report improved performance over classical optimal transportation distances on the MNIST benchmark problem.,7569
"inference a (DAG). graph Causal on structure relies a of the often directed acyclic graph,","Causal inference relies on the structure of a graph, often a directed acyclic graph (DAG).",7570
in and different graphs distributions. statements different result may inference causal intervention Different,Different graphs may result in different causal inference statements and different intervention distributions.,7571
well-suited computing are interventions. graphs used for It therefore is for evaluating that,It is therefore well-suited for evaluating graphs that are used for computing interventions.,7572
and have approaches solvers. achieved Algorithm improvements single portfolio remarkable selection over,Algorithm portfolio and selection approaches have achieved remarkable improvements over single solvers.,7573
for to makes researchers problems. for techniques it different This specific difficult explore their,This makes it difficult for researchers to explore different techniques for their specific problems.,7574
three method ones. previously We our compare proposed with well-known,We compare our method with three well-known previously proposed ones.,7575
Our potential the the demonstrate approach. new experimental good of results,Our experimental results demonstrate the good potential of the new approach.,7576
model. probabilistic attempts inference the to symmetries in by up Lifting speed exploiting,Lifting attempts to speed up probabilistic inference by exploiting symmetries in the model.,7577
"there no currently equivalent for is results nor inference. complexity structure analogous lifted However,","However, there is currently no equivalent structure nor analogous complexity results for lifted inference.",7578
"FO-dtrees, propositional upgrade first-order to In introduce dtrees level. which paper, the we this","In this paper, we introduce FO-dtrees, which upgrade propositional dtrees to the first-order level.",7579
"been not before For the have estimators in these all know, we given literature.","For all we know, these estimators have not been given before in the literature.",7580
"further effects. Sometimes, priming instance, only for analyzed, for unconscious these trials are","Sometimes, only these trials are further analyzed, for instance, for unconscious priming effects.",7581
"of bias. effects response it the Most potential importantly, ignores","Most importantly, it ignores the potential effects of response bias.",7582
"discuss admixture spatial ancestral using graph. structure, and and models generative We recombination the drift,","We discuss generative models using drift, admixture and spatial structure, and the ancestral recombination graph.",7583
"principle and statistical methods. analysis for contrasted with other are component `non-parametric' inference, models These","These are contrasted with statistical models for inference, principle component analysis and other `non-parametric' methods.",7584
examples. are and approaches these simulated explored real-data between both with relationships The,The relationships between these approaches are explored with both simulated and real-data examples.,7585
contrasted. and discussed are practical The tools software state-of-the-art,The state-of-the-art practical software tools are discussed and contrasted.,7586
spp.) species  indigenous oak The (Quercus ,  The indigenous oak species (Quercus spp.),7587
"human recolonization, Since ecological and have subjected changes to been they post-glacial impact.","Since post-glacial recolonization, they have been subjected to ecological changes and human impact.",7588
clustering method. The a is algorithm partitional k-means,The k-means algorithm is a partitional clustering method.,7589
consequence its of part simplicity a and of large efficiency. in popularity k-means is The,The popularity of k-means is in large part a consequence of its simplicity and efficiency.,7590
speedups. indicate easily and adaptive strategy Our implemented empirical significant be can results,Our adaptive strategy can be easily implemented and empirical results indicate significant speedups.,7591
on a timely presents literature this of subject. review the comprehensive and paper This,This paper presents a comprehensive and timely review of the literature on this subject.,7592
research Consequences of fact in work. this investigated are this,Consequences of this fact are investigated in this research work.,7593
"with DE-CCA. is greater achieved discriminative improvement variant, An the even","An even greater improvement is achieved with the discriminative variant, DE-CCA.",7594
research area an since last decades. has been OCR active few,OCR has been an active research area since last few decades.,7595
"several segmentation, pre-processing, can processing. The like stages OCR recognition have process and post","The OCR process can have several stages like pre-processing, segmentation, recognition and post processing.",7596
technique filtering technique. comparison to The results improvised in median shows proposed,The proposed technique shows improvised results in comparison to median filtering technique.,7597
A multispectral generalization case multidimensional images foreseen. color is of for and,A generalization for multidimensional case of color and multispectral images is foreseen.,7598
characterized loci and on FSTs. were discriminant provenance Species based,Species and provenance discriminant loci were characterized based on FSTs.,7599
then structure. in treated analysis genetic were Data separately a of Bayesian,Data were then treated separately in a Bayesian analysis of genetic structure.,7600
"to three probability species By achieved. species inferred membership high groups using loci, discriminant was","By using three species discriminant loci, high membership probability to inferred species groups was achieved.",7601
"identification species of an molecular provided is Thus, and effective tool provenances. for both","Thus, an effective tool is provided for molecular identification of both species and provenances.",7602
"genome selection, species Species discriminant identity. directional represent might affected which by maintains loci regions","Species discriminant loci might represent genome regions affected by directional selection, which maintains species identity.",7603
are Stack of special non-linear a filters case filters.,Stack filters are a special case of non-linear filters.,7604
"binary function, then filter. which a characterizes image the is Each Boolean by filtered","Each binary image is then filtered by a Boolean function, which characterizes the filter.",7605
"Classification. system stage, of three comprises stages: Preprocessing The Feature and proposed Extraction","The proposed system comprises of three stages: Preprocessing stage, Feature Extraction and Classification.",7606
"filtering, skin stage matching. Preprocessing includes histogram","Preprocessing stage includes skin filtering, histogram matching.",7607
walks problems in source of the are very (SAW) Self-avoiding enumerative and combinatorics. difficult probabilities,Self-avoiding walks (SAW) are the source of very difficult problems in probabilities and enumerative combinatorics.,7608
(un)folded presented article. of is self-avoiding first this study A in walks,A first study of (un)folded self-avoiding walks is presented in this article.,7609
of survey about contribution The known is majorly sets. currently is these what a,The contribution is majorly a survey of what is currently known about these sets.,7610
"we obtained, first and have results sets. the theoretically these computationally, on or","and the first results we have obtained, theoretically or computationally, on these sets.",7611
"and temperature, emotional ambient intake. state, alcohol","emotional state, ambient temperature, and alcohol intake.",7612
addresses which novel a describe challenges. these method We,We describe a novel method which addresses these challenges.,7613
which This (AAM) follows appearance warping is by achieved affine active fitting. model piecewise,This is achieved by piecewise affine warping which follows active appearance model (AAM) fitting.,7614
representation both pose changes. scale robust much and This makes proposed more the to,This makes the proposed representation much more robust both to pose and scale changes.,7615
shape method propose of detect curves. We a and new in to visualize outliers samples,We propose a new method to visualize and detect shape outliers in samples of curves.,7616
reinforcement computationally Optimal learning approach infeasible. probabilistic in is,Optimal probabilistic approach in reinforcement learning is computationally infeasible.,7617
"and temperature, intake. ambient alcohol emotional state,","emotional state, ambient temperature, and alcohol intake.",7618
method this these challenges. we novel which addresses a major paper In describe,In this paper we describe a novel method which addresses these major challenges.,7619
achieved by affine AAM follows warping which fitting. This is piecewise,This is achieved by piecewise affine warping which follows AAM fitting.,7620
distance camera). subject varying due a the of from to the,due to the varying distance of a subject from the camera).,7621
lengths. evolution We with of different consider birth-and-death stochastic genotypes,We consider birth-and-death stochastic evolution of genotypes with different lengths.,7622
effect. to death and The are length selection which rates corresponds birth dependent a,The birth and death rates are length dependent which corresponds to a selection effect.,7623
study genotypes. an of We an for of a asymptotic collection infinite behavior density,We study an asymptotic behavior of a density for an infinite collection of genotypes.,7624
and considered. are of homogeneous heterogeneous space cases space The densities,The cases of space homogeneous and space heterogeneous densities are considered.,7625
two-dimensional extend quantitatively expansions. Our to results and qualitatively,Our results extend qualitatively and quantitatively to two-dimensional expansions.,7626
"to universal, induction known Solomonoff but be is incomputable.","Solomonoff induction is known to be universal, but incomputable.",7627
bounds which rationality Therefore flexibly-bounded within the is exercised. expands rationality,Therefore flexibly-bounded rationality expands the bounds within which rationality is exercised.,7628
the properties dynamic Here we formulate histone patterns. model mathematical a of investigate to modification,Here we formulate a mathematical model to investigate the dynamic properties of histone modification patterns.,7629
"should for considered be parts data of clustering? which First,","First, which parts of data should be considered for clustering?",7630
"clustering for Second, suitable method what is de-noising? data","Second, what data clustering method is suitable for de-noising?",7631
to dictionary. is learn an over exploited complete Clustering,Clustering is exploited to learn an over complete dictionary.,7632
a heritage is quantified evolutionary diversity of by The phylogenetic biodiversity frequently (PD). measure called,The evolutionary heritage of biodiversity is frequently quantified by a measure called phylogenetic diversity (PD).,7633
in for flexibility these three-dimensional the motions is How intrinsic structures? their encoded required,How is the intrinsic flexibility required for these motions encoded in their three-dimensional structures?,7634
"be determinant accessibility solvent of local flexibility. is primary residue a local Finally, shown to","Finally, local solvent accessibility is shown to be a primary determinant of local residue flexibility.",7635
"objects, actors,and simple does compositions the How human of brain represent actions?","How does the human brain represent simple compositions of objects, actors,and actions?",7636
"viewed again scanning the while ), during sequence (fMRI). action subjects","), again while subjects viewed the action sequence during scanning (fMRI).",7637
"the and systems, TPJ, i.e. 'where' and 'what'","the 'what' and 'where' systems, and TPJ, i.e.",7638
of evidence interspecific crossings progeny are that provides Analysis rare. arrays,Analysis of progeny arrays provides evidence that interspecific crossings are rare.,7639
chloroplast introgression finding is limited This of supported further genomes. by,This finding is further supported by limited introgression of chloroplast genomes.,7640
"might this account phenology, post-zygotic incompatibilities) asynchronous for result.","asynchronous phenology, post-zygotic incompatibilities) might account for this result.",7641
morphologically. species observed studied has Hybridization already been between the and two,Hybridization between the two species has already been observed and studied morphologically.,7642
"of genetic evidence little exists about extent However, the introgression.","However, little evidence exists about the extent of genetic introgression.",7643
were of used microsatellites a analysis DNA genetic barriers. for Chloroplast spatial,Chloroplast DNA microsatellites were used for a spatial analysis of genetic barriers.,7644
main explained the variation genetic groups. species partition was part by The into of nuclear,The main part of the nuclear genetic variation was explained by partition into species groups.,7645
"funnels. and landscapes repeats with multiple embedded related fundamentally structural Similarly, mosaics to be may","Similarly, repeats and structural mosaics may be fundamentally related to landscapes with multiple embedded funnels.",7646
structural protein molecules. and repetitions to detect Here in tools compare present we analytical,Here we present analytical tools to detect and compare structural repetitions in protein molecules.,7647
recognize different propose to gestures. Hence hand approach an we easy,Hence we propose an easy approach to recognize different hand gestures.,7648
"to algorithm mean a adapt the field. Therefore, the we fluctuations Gillespie also around reproduce","Therefore, we also adapt a Gillespie algorithm to reproduce the fluctuations around the mean field.",7649
characterize The stochastic scheme exploited is model. DE the to,The DE scheme is exploited to characterize the stochastic model.,7650
"a any much more allows detailed our the of case, description model fights. In","In any case, our model allows a much more detailed description of the fights.",7651
is celebrated known Auto-Regression. as problem This a,This is a celebrated problem known as Auto-Regression.,7652
speech simple done estimation simulations and on Two have spectrum coding. been,Two simple simulations have been done on spectrum estimation and speech coding.,7653
for up ground rapid data set corresponding truth. acquisition of allows annotation and Our with,Our set up allows for rapid acquisition and annotation of data with corresponding ground truth.,7654
algorithm proposed The are detail. properties and evaluated the discrimination in of robustness,The robustness and discrimination properties of the proposed algorithm are evaluated in detail.,7655
"this as embodied matrix In Decomposition. by Value invariants spectral we use paper, Singular","In this paper, we use spectral matrix invariants as embodied by Singular Value Decomposition.",7656
well-suited hardware. are All current computer to,All are well-suited to current computer hardware.,7657
"are a hardware with performance including software comparison different the Empirical configurations. of presented, results","Empirical results are presented, including a performance comparison of the software with different hardware configurations.",7658
features SPM exceptional task. on car distance achieve for the performance MDS The recognition based,The MDS features based on SPM distance achieve exceptional performance for the car recognition task.,7659
"operator. estimation blur Consequently, non-uniform the requires a of successful spatially-varying blind or deconvolution","Consequently, successful blind deconvolution requires the estimation of a spatially-varying or non-uniform blur operator.",7660
a free. parameter virtually is can that algorithm using implemented The be strategy majorization-minimization,The algorithm can be implemented using a majorization-minimization strategy that is virtually parameter free.,7661
real theoretical proposed and images the serve validation to method. Detailed empirical analysis validate on,Detailed theoretical analysis and empirical validation on real images serve to validate the proposed method.,7662
"only association the in edges pairwise However, statistics depend fields of on natural scenes.","However, association fields only depend on the pairwise statistics of edges in natural scenes.",7663
embedding reveals analysis spectral a An using probabilistic components. curvature-dependent,An analysis using a probabilistic spectral embedding reveals curvature-dependent components.,7664
"that the of) show we divergence and (half equivalent are beta statistical measures. deviance Further,","Further, we show that beta divergence and (half of) the statistical deviance are equivalent measures.",7665
"open of bases still However, largely a rule is evolution handling problem. the","However, handling the evolution of rule bases is still a largely open problem.",7666
symmetry with robust We consider covariance constraints. estimation group,We consider robust covariance estimation with group symmetry constraints.,7667
"sets of examples structures. such proper include: and complex/quaternion Practical circulant, persymmetric","Practical examples of such sets include: circulant, persymmetric and complex/quaternion proper structures.",7668
"Bayesian a only estimation. However, few papers considered","However, only a few papers considered Bayesian estimation.",7669
"article the intelligence"". This an of is theory overview of ""SP","This article is an overview of the ""SP theory of intelligence"".",7670
"""multiple of a alignment"" central is The idea. concept powerful","The concept of ""multiple alignment"" is a powerful central idea.",7671
"probabilities calculated. multiple be alignment, each may For","For each multiple alignment, probabilities may be calculated.",7672
the of These for calculating provide probabilities the basis inferences.,These provide the basis for calculating the probabilities of inferences.,7673
partial from patterns. matches new learns between The structures system,The system learns new structures from partial matches between patterns.,7674
viewpoint on focussing Wide-baseline change extreme is considered. matching with problems,Wide-baseline matching focussing on problems with extreme viewpoint change is considered.,7675
actions. of with different a classify dataset couple dozens Currently only most systems,Currently most systems only classify dataset with a couple of dozens different actions.,7676
"extraction computational from often is complex. Moreover, data the feature","Moreover, feature extraction from the data is often computational complex.",7677
Often questions based these reasoning. require simulation,Often these questions require simulation based reasoning.,7678
"higher colored we constructs a tokens. As result, level need like","As a result, we need higher level constructs like colored tokens.",7679
to Our modeling approach can domains with other be needs. similar adapted,Our approach can be adapted to other domains with similar modeling needs.,7680
covariance events define tiger An brown prawns. used at of to analysis was fishing targeted,An analysis of covariance was used to define fishing events targeted at brown tiger prawns.,7681
variations fishing general A linear estimated of model power. inter-annual,A general linear model estimated inter-annual variations of fishing power.,7682
effort recent in were this fishery The fishing causes of in discussed. decline,The causes of recent decline in fishing effort in this fishery were discussed.,7683
It and the itself structures inferential lends knowledge processes. visualisation to of,It lends itself to the visualisation of knowledge structures and inferential processes.,7684
to this experiment declarative the problem. encourage readers approaches also We different with to,We also encourage the readers to experiment with different declarative approaches to this problem.,7685
computational is the most Protein one of in problems structure biology. prediction important,Protein structure prediction is one of the most important problems in computational biology.,7686
composite non-directional model We sensory on search a developed of cues. based,We developed a model of composite search based on non-directional sensory cues.,7687
method efficient the about be At will last concluding we i.e.,At last we will be concluding about the efficient method i.e.,7688
methods. performance compared gives that better the the to other method,the method that gives better performance compared to the other methods.,7689
within SMC algorithms use Such dynamics. the algorithms transition afore-mentioned the,Such algorithms use the afore-mentioned SMC algorithms within the transition dynamics.,7690
part the can In reasoning. through knowledge many of recover missing cases one,In many cases one can recover part of the missing knowledge through reasoning.,7691
then the discuss answering and how this We respect with implication to of why questions.,We then discuss the implication of this with respect to answering why and how questions.,7692
They hunt not rarely protein. do and of rich sources encounter consequently,They rarely hunt and consequently do not encounter rich sources of protein.,7693
they Instead carbohydrates. to adapted primarily have of diet a,Instead they have adapted to a diet of primarily carbohydrates.,7694
till advancements present paper presents those art of This the day. state areas in,This paper presents state of the art advancements in those areas till present day.,7695
this Our comparison superior that dataset. produced of systems method to on results,Our method produced results superior to that of comparison systems on this dataset.,7696
threading allows An individual for sequences. of extension subtrees of instead,An extension allows for threading of subtrees instead of individual sequences.,7697
these called methods program in a have We computer ARGweaver. implemented,We have implemented these methods in a computer program called ARGweaver.,7698
to reducing Constraint powerful is (CP) Nogood approach a in learning Programming search solvers.,Nogood learning is a powerful approach to reducing search in Constraint Programming (CP) solvers.,7699
learning resolution Nogood be as solvers seen proof systems. can,Nogood learning solvers can be seen as resolution proof systems.,7700
"CP system, problem. the solve faster can it stronger The the proof a","The stronger the proof system, the faster it can solve a CP problem.",7701
"systems \emph{extended as exist. stronger such resolution} proof However,","However, stronger proof systems such as \emph{extended resolution} exist.",7702
noncentral a Student's $t$-distribution of The is the generalization $t$-distribution.,The noncentral $t$-distribution is a generalization of the Student's $t$-distribution.,7703
"solving SAT become portfolio approaches In common. CSPs years, problems to have increasingly and recent","In recent years, portfolio approaches to solving SAT problems and CSPs have become increasingly common.",7704
encodings a are There representing number instances. also CSPs as different of SAT for,There are also a number of different encodings for representing CSPs as SAT instances.,7705
dramatically the and requirement. burden memory This computational increases,This dramatically increases the computational burden and memory requirement.,7706
"proposed the enhancement. luminance-independent method this way, In a contrast can provide","In this way, the proposed method can provide a luminance-independent contrast enhancement.",7707
"a established. the is amount From varying of data, population","From the data, a varying amount of population is established.",7708
of nature questions these correlations. surround Open basic the,Open questions surround the basic nature of these correlations.,7709
"sexual The of reproduction of prevalence enigma an biology. (""sex"") is evolutionary eukaryotes in","The prevalence of sexual reproduction (""sex"") in eukaryotes is an enigma of evolutionary biology.",7710
long-term increases its superiority essence. in variation tells genetic only Sex,Sex increases genetic variation only tells its long-term superiority in essence.,7711
ubiquitous causes for accumulation of an and mutations organisms. harmful The immediate pressure,The accumulation of harmful mutations causes an immediate and ubiquitous pressure for organisms.,7712
irreversibly reduces A reproductive than lower population become the extinct to rate will once R*.,A population will become irreversibly extinct once the reproductive rate reduces to lower than R*.,7713
explains asexual and as That persistence facultative of widespread organisms. the as sex the well,That explains the widespread sex as well as the persistence of facultative and asexual organisms.,7714
the for in real and illustrated adaptive are datasets. simulated The improvements in efficiency samplers,The improvements in efficiency for the adaptive samplers are illustrated in real and simulated datasets.,7715
Decreases well production fisheries decreased loads are in also documented. are when nutrient,Decreases in fisheries production when nutrient loads are decreased are also well documented.,7716
very of many image crucial is processing quality image applications. Measurement to,Measurement of image quality is very crucial to many image processing applications.,7717
codecs. prevalent most technique the compression compression is for JPEG lossy which image is,JPEG compression is the lossy compression which is most prevalent technique for image codecs.,7718
artifacts. Various reduce to deblocking used blocking filters are,Various deblocking filters are used to reduce blocking artifacts.,7719
introduce approach PSNR-B deblocked for of of quality analyzing We images. a new,We introduce a new approach of PSNR-B for analyzing quality of deblocked images.,7720
approach new results of called modified show Simulation PSNR-B that PSNR-B.,Simulation results show that new approach of PSNR-B called modified PSNR-B.,7721
known existing . gives better specific it compared well even blockiness results to indices,it gives even better results compared to existing well known blockiness specific indices .,7722
genetic recent human evidence found have population growth Population for history. in dramatic studies,Population genetic studies have found evidence for dramatic population growth in recent human history.,7723
"association significant number detected. demography of the affect not Moreover, does here","Moreover, here demography does not affect the number of significant association detected.",7724
We propose fair max-min timetabling the course curriculum-based decomposition (MMF-CB-CTT) problem. of a,We propose a decomposition of the max-min fair curriculum-based course timetabling (MMF-CB-CTT) problem.,7725
assignment the The optimization subproblem generalized bottleneck (LBOP). room models as decomposition a problem lexicographic,The decomposition models the room assignment subproblem as a generalized lexicographic bottleneck optimization problem (LBOP).,7726
"the a efficiently. As can consequence, subproblem room problem of the be assignment MMF-CB-CTT solved","As a consequence, the room assignment subproblem of the MMF-CB-CTT problem can be solved efficiently.",7727
We for proposed a use the problem. insight improve MMF-CB-CTT algorithm previously to this heuristic,We use this insight to improve a previously proposed heuristic algorithm for the MMF-CB-CTT problem.,7728
"scholarships. the more becoming fees more expensive, are As for apply   education students","  As the education fees are becoming more expensive, more students apply for scholarships.",7729
"hundreds even by sponsor. Consequently, and thousands be of handled to the applications need","Consequently, hundreds and even thousands of applications need to be handled by the sponsor.",7730
"problems, need solve on some several the (criteria) to attributes selected. alternatives be based To","To solve the problems, some alternatives based on several attributes (criteria) need to be selected.",7731
were data. Data fuzzy used a crisp and,Data used were a crisp and fuzzy data.,7732
"response. tended spatial to purely retinal in correlations increase Meanwhile, correlations the","Meanwhile, purely spatial correlations tended to increase correlations in the retinal response.",7733
model The traditional such contour on images. work active not does,The traditional active contour model does not work on such images.,7734
"contour Manifold to In active model this on generalize paper, the we images. work valued","In this paper, we generalize the active contour model to work on Manifold valued images.",7735
"regions our values algorithm expected, detects with image. As similar in Manifold the","As expected, our algorithm detects regions with similar Manifold values in the image.",7736
and results images. texture demonstrate segmentation manifold images for We valued,We demonstrate segmentation results for manifold valued images and texture images.,7737
sensory requires predictions inputs. Guiding brain future behavior about make to the,Guiding behavior requires the brain to make predictions about future sensory inputs.,7738
Proper subsequent and measures similarity two applied. are statistical of introduced analysis is curves a,Proper similarity measures of two curves are introduced and a subsequent statistical analysis is applied.,7739
mathematical is of though. no texture a definition There,There is no mathematical definition of a texture though.,7740
Automating far trivial. this a for computer from is task,Automating this task for a computer is far from trivial.,7741
"algorithm an paper, this for texture based unsupervised propose In contour novel active segmentation. we","In this paper, we propose an active contour based novel unsupervised algorithm for texture segmentation.",7742
on and of real both our demonstrate We texture algorithm images. performance artificial,We demonstrate performance of our algorithm on both artificial and real texture images.,7743
defined these weighted-sum a as Vulnerability is of components.,Vulnerability is defined as a weighted-sum of these components.,7744
was A reduced penalty for vulnerability. imposed,A penalty was imposed for reduced vulnerability.,7745
"also region, but focus not context only can Our shape. system and global region preserve","Our system can preserve not only focus region, but also context region and global shape.",7746
backed thorough Our by studies. theoretical results are numerical up,Our theoretical results are backed up by thorough numerical studies.,7747
uncertainty partial equations. propagating ordinary for a and Polynomial through chaos technique is powerful differential,Polynomial chaos is a powerful technique for propagating uncertainty through ordinary and partial differential equations.,7748
spatial resolution to of images hyperspectral Hyperspectral low pixels sensors. mixed contain due,Hyperspectral images contain mixed pixels due to low spatial resolution of hyperspectral sensors.,7749
called pixels pixels than containing distinct material more Mixed one endmembers. are,Mixed pixels are pixels containing more than one distinct material called endmembers.,7750
of percentages presence endmembers pixels The are called in mixed fractions. abundance,The presence percentages of endmembers in mixed pixels are called abundance fractions.,7751
"characterizing five at present aimed paper, axioms this we inconsistency In indices.","In this paper, we present five axioms aimed at characterizing inconsistency indices.",7752
"assumptions, The used theoretical developed framework algorithms. here SMC several commonly cover, will under","The theoretical framework developed here will cover, under assumptions, several commonly used SMC algorithms.",7753
genes. have gene to Such generally as marker families referred been,Such gene families have generally been referred to as marker genes.,7754
phylogenetic to (for We refer and these markers. phylogenetic ecology) as therefore PhyEco,We therefore refer to these as PhyEco (for phylogenetic and phylogenetic ecology) markers.,7755
"In is paper, an deterministic this scheme this proposed. for mean approximating iterative","In this paper, an iterative deterministic scheme for approximating this mean is proposed.",7756
Hence transformations are geometric called all these elementary.,Hence all these geometric transformations are called elementary.,7757
is of the important component current An methods assembly scaffolding process.,An important component of current assembly methods is the scaffolding process.,7758
"theory, for signal suitable Based representation dictionary sparse CS a the of is on required.","Based on CS theory, a suitable dictionary for sparse representation of the signal is required.",7759
optimal. adaptive to proved dictionary be is The,The adaptive dictionary is proved to be optimal.,7760
CR-DL start-of-the-art and DR-DL with al. the proposed Compared by et Merlet methods,Compared with the start-of-the-art CR-DL and DR-DL methods proposed by Merlet et al.,7761
"al., Bilgic offers espectively, our work and et advantages. the following","and Bilgic et al., espectively, our work offers the following advantages.",7762
"learned dictionary diffusion signals. optimal proved to be First, is the Gaussian for","First, the learned dictionary is proved to be optimal for Gaussian diffusion signals.",7763
"the first a is Second, this learn to to our voxel-adaptive work dictionary. knowledge,","Second, to our knowledge, this is the first work to learn a voxel-adaptive dictionary.",7764
will of theoretically in the dictionary reconstruction importance demonstrated adaptive The and EAP be empirically.,The importance of the adaptive dictionary in EAP reconstruction will be demonstrated theoretically and empirically.,7765
"we alternative model. an propose VAR estimating in the In manuscript, this way","In this manuscript, we propose an alternative way in estimating the VAR model.",7766
"theoretical into new brings insights the model In addition, our analysis. VAR method","In addition, our method brings new theoretical insights into the VAR model analysis.",7767
Such indirect the and matrices transparent. conditions not about transition are,Such conditions are indirect about the transition matrices and not transparent.,7768
both prediction. for provide of convergence explicit estimation rates and We,We provide explicit rates of convergence for both estimation and prediction.,7769
"such assumption how it affected if However, is are is last dropped. the unclear tests","However, it is unclear how such tests are affected if the last assumption is dropped.",7770
A frequency the moments to the spectrum. is key of ingredient first specify two,A key ingredient is to specify the first two moments of the frequency spectrum.,7771
"populations, may mutations simultaneously spreading. In be large multiple beneficial","In large populations, multiple beneficial mutations may be simultaneously spreading.",7772
"or management not. huge a whether has role hospitalized of Imaging patients, the occupied in","Imaging has occupied a huge role in the management of patients, whether hospitalized or not.",7773
"variety a patients Depending imaging on were use. modalities the available clinical problem, for of","Depending on the patients clinical problem, a variety of imaging modalities were available for use.",7774
of birth process. image medical This of the gave annotation,This gave birth of the annotation of medical image process.,7775
semantic gap. is problem to annotation the of solve image intended The and analysis,The annotation is intended to image analysis and solve the problem of semantic gap.,7776
increase is due to in acquisition images. of for annotation The reason image,The reason for image annotation is due to increase in acquisition of images.,7777
the the activity design most The successful phase in of important is process. building annotation,The design phase is the most important activity in the successful building of annotation process.,7778
consistent obtained Values seem observed. those with,Values obtained seem consistent with those observed.,7779
image is Segmentation Image partitioning into classes. technique of some distinct original a the,Image Segmentation is a technique of partitioning the original image into some distinct classes.,7780
"multilevel segmentation. thresholding technique been In method, used has proposed for our image","In our proposed method, multilevel thresholding technique has been used for image segmentation.",7781
is (CS) of optimal of value. threshold for Cuckoo used approach A selection Search new,A new approach of Cuckoo Search (CS) is used for selection of optimal threshold value.,7782
"quality. segmentation MSE measured understand are to PSNR Finally, and the","Finally, MSE and PSNR are measured to understand the segmentation quality.",7783
complexity of as of serves a the measure text. structural It,It serves as a measure of structural complexity of the text.,7784
It in is data detection anomaly applied to transmission.,It is applied to anomaly detection in data transmission.,7785
demonstrate the that integration leads sources to of robust multiple data more We performance.,We demonstrate that the integration of multiple data sources leads to more robust performance.,7786
A copula is of to dependencies these approach a express in terms function. common,A common approach is to express these dependencies in terms of a copula function.,7787
"To estimation the account of framework for proposed. for is a Bayesian copulas this, conditional","To account for this, a Bayesian framework for the estimation of conditional copulas is proposed.",7788
our We and by cancer advocate important view relevant combining concepts. reviewing,We advocate our view by reviewing and combining important cancer relevant concepts.,7789
many existed years. markets India for In have financial,In India financial markets have existed for many years.,7790
"exist maturity markets varying Today debt, and of commodities foreign in exchange. equity,","Today markets of varying maturity exist in equity, debt, commodities and foreign exchange.",7791
"Fuzzy Computational use of Sets, such Networks, etc. as Intelligence Genetic Algorithms, The Systems Neural","The use of Computational Intelligence Systems such as Neural Networks, Fuzzy Sets, Genetic Algorithms, etc.",7792
has Predictions established. Stock widely Market been for,for Stock Market Predictions has been widely established.,7793
the knowledge from extract to is stock daily in of The form movements. process rules,The process is to extract knowledge in the form of rules from daily stock movements.,7794
can These used rules guide be then to investors.,These rules can then be used to guide investors.,7795
paradigm is develop used rule extraction a The thus algorithm. to,The paradigm is thus used to develop a rule extraction algorithm.,7796
problem Problem widely has Transportation Research domain. studied is in which been Operations an important,Transportation Problem is an important problem which has been widely studied in Operations Research domain.,7797
simulate life It often real been to has used problems. different,It has been often used to simulate different real life problems.,7798
"In application remarkable Hard has Problems in of this significance. particular, Problem NP a","In particular, application of this Problem in NP Hard Problems has a remarkable significance.",7799
"the complexity computational in we Finally, involved problem. discuss the","Finally, we discuss the computational complexity involved in the problem.",7800
ETP is Hard optimization combinatorial problem. NP,ETP is NP Hard combinatorial optimization problem.,7801
"NSOU, we develop technique. this Kolkata, using models India Paper, mathematical for FILP In three","In this Paper, we develop three mathematical models for NSOU, Kolkata, India using FILP technique.",7802
deal impreciseness with through numbers. we variables vagueness To and allocation various model fuzzy,To deal with impreciseness and vagueness we model various allocation variables through fuzzy numbers.,7803
obtained using solution The Fuzzy the ranking number problem to method. is,The solution to the problem is obtained using Fuzzy number ranking method.,7804
function. objective feasible by Fuzzy has solution obtained fuzzy number Each,Each feasible solution has fuzzy number obtained by Fuzzy objective function.,7805
FILP with compared The are used proposed heuristic viz. models commonly,The proposed FILP models are compared with commonly used heuristic viz.,7806
which experimental gives heuristic. quality of an on approach about ILP data idea,ILP approach on experimental data which gives an idea about quality of heuristic.,7807
good serves heuristics. for The benchmark other formulation as thus,The formulation thus serves as good benchmark for other heuristics.,7808
"restrictions global with guarantee tractability for constraints. such However, many to CSPs fail","However, many such restrictions fail to guarantee tractability for CSPs with global constraints.",7809
regression model fitness of a NK mean has explained in as landscapes. been This result,This has been explained as a result of mean regression in NK model fitness landscapes.,7810
we quite fitness this Here phenomenon landscapes. show generally occurs in that,Here we show that this phenomenon occurs quite generally in fitness landscapes.,7811
relationships of knowledge concepts a description their in a domain. and provide formal Ontologies,Ontologies provide a formal description of concepts and their relationships in a knowledge domain.,7812
between a reveals correlation properties. Our result these analytical two,Our analytical result reveals a correlation between these two properties.,7813
The such pure principle discussed. by of games Fuzzy is strategies minimax-maximin with solution,The solution of such Fuzzy games with pure strategies by minimax-maximin principle is discussed.,7814
"pay-off Method. to by Here, Dominance matrix is pay-off reduced matrix","Here, pay-off matrix is reduced to pay-off matrix by Dominance Method.",7815
by This means is Numerical fact illustrated of Example.,This fact is illustrated by means of Numerical Example.,7816
"should used. be We such that call algorithms and \texttt{AdapFail}, conclude they not","We call such algorithms \texttt{AdapFail}, and conclude that they should not be used.",7817
object challenge central recognition. a and detection remains variation Illumination in,Illumination variation remains a central challenge in object detection and recognition.,7818
determine Molecular signaling to are robustness. networks cancer believed,Molecular signaling networks are believed to determine cancer robustness.,7819
"the with curable. In modular less more are signaling networks cancers particular,","In particular, the cancers with less modular signaling networks are more curable.",7820
consistent modularity--robustness is a prediction hypothesis. result the This from with,This result is consistent with a prediction from the modularity--robustness hypothesis.,7821
"connectivity. do heterogeneous the of not the results these However, contradict importance","However, these results do not contradict the importance of the heterogeneous connectivity.",7822
Transportation an which important in has Operations widely been aspect is Problem domain. studied Research,Transportation Problem is an important aspect which has been widely studied in Operations Research domain.,7823
studied real to simulate life It different has been problems.,It has been studied to simulate different real life problems.,7824
"this Hard Problems significance. remarkable In application has of particular, in a NP- Problem","In particular, application of this Problem in NP- Hard Problems has a remarkable significance.",7825
Logic a classical uncertainty. Fuzzy is reasoning computational that paradigm for two-valued logic generalizes under,Fuzzy Logic is a computational paradigm that generalizes classical two-valued logic for reasoning under uncertainty.,7826
Sets uncertain handling vague information. and The nature allows of Fuzzy multi-valued,The multi-valued nature of Fuzzy Sets allows handling uncertain and vague information.,7827
a Theory. and is clever disguise of model-less approach It a Probability,It is a model-less approach and a clever disguise of Probability Theory.,7828
discuss the of approaches simulation We and Computational results comparative Complexity. both give,We give comparative simulation results of both approaches and discuss the Computational Complexity.,7829
"and been models regression forecasting Research to Fuzzy several applied prediction. viz., have applications Operations","Fuzzy regression models have been applied to several Operations Research applications viz., forecasting and prediction.",7830
they But non-uniform with of cannot spreads. the problem deal,But they cannot deal with the problem of non-uniform spreads.,7831
of The improving inference used for Tagaki-Sugeno is forecasts. system the accuracy,The Tagaki-Sugeno inference system is used for improving the accuracy of forecasts.,7832
applied model to processes. The new also is object recognition,The new model is also applied to object recognition processes.,7833
The experimental our show is with less requirement. that results much memory method efficient primary,The primary experimental results show that our method is efficient with much less memory requirement.,7834
"often are sometimes which and real noisy ineffective. extinction-prone, such methods render can populations Moreover,","Moreover, real populations are often noisy and extinction-prone, which can sometimes render such methods ineffective.",7835
"a this significance. remarkable of particular, theory in Problems In application has NP-Hard","In particular, application of this theory in NP-Hard Problems has a remarkable significance.",7836
Combinatorial to It a Optimization. new Stochastic approach viable is,It is a viable new approach to Stochastic Combinatorial Optimization.,7837
We results. the Longest the simulation give methodology apply proposed to and Common Subsequence Problem,We apply the proposed methodology to Longest Common Subsequence Problem and give the simulation results.,7838
is by of effectiveness Computational efficient this The demonstrated Complexity. approach,The effectiveness of this approach is demonstrated by efficient Computational Complexity.,7839
of Community  has graphs detection  many in been subject algorithms. the,  Community detection in graphs has been the subject of many algorithms.,7840
"are directed graphs. these and unipartite, to algorithms multipartite applied","these algorithms are applied to unipartite, multipartite and directed graphs.",7841
our show of Several approach. interest experiments the,Several experiments show the interest of our approach.,7842
numerical result illustrated with the algorithm and theoretical experiments. The are,The theoretical result and the algorithm are illustrated with numerical experiments.,7843
"computing is distributed sets. scalable MapReduce for proven, huge a on data solution","MapReduce is a proven, scalable solution for distributed computing on huge data sets.",7844
autocatalytic theory provides (reflexively and study such to processes. a formal RAF food-generated) way,RAF (reflexively autocatalytic and food-generated) theory provides a formal way to study such processes.,7845
has (HE) to an addition essential Enhancement Equalization world. Image been Histogram the,Histogram Equalization (HE) has been an essential addition to the Image Enhancement world.,7846
"various etc. SNR, MSE, calculated the have RMSE, parameters We like PSNR,","We have calculated the various parameters like PSNR, SNR, RMSE, MSE, etc.",7847
"controversial. the general Meanwhile, remains of existence landscape","Meanwhile, the general existence of landscape remains controversial.",7848
neighborhoods. neighborhoods of form covering All called a new covering the a,All the neighborhoods form a new covering called a covering of neighborhoods.,7849
on of studies of the degree coverings This application repeat further paper neighborhoods.,This paper studies further the application of repeat degree on coverings of neighborhoods.,7850
"give according method of the we repeat calculating the covering Finally, to degree.","Finally, we give the method of calculating the covering according to repeat degree.",7851
We conclude giving by future for impulses research.,We conclude by giving impulses for future research.,7852
sensing. of satellite briefly reviews paper the remote This limitations,This paper briefly reviews the limitations of satellite remote sensing.,7853
"fusion image reviews techniques. of on the Also, problems","Also, reviews on the problems of image fusion techniques.",7854
classification. introduces (ALwassaiProcess) for program and also image developed It fusion,It also introduces program (ALwassaiProcess) developed for image fusion and classification.,7855
"than faster the retain blocks cannot degree. expected the grow To transitivity,","To retain transitivity, the blocks cannot grow faster than the expected degree.",7856
"Thus, blocks must small. remain asymptotically models, the sparse in","Thus, in sparse models, the blocks must remain asymptotically small.",7857
"variable neurons sensory the repeat presentations same responses. Over of show stimulus,","Over repeat presentations of the same stimulus, sensory neurons show variable responses.",7858
unifying yields work the into insights role correlations. noise of This,This work yields unifying insights into the role of noise correlations.,7859
as This is generality other well. our results for true,This generality is true for our other results as well.,7860
a correlation set improve can show structures of coding. that diverse We,We show that a diverse set of correlation structures can improve coding.,7861
"these of correlations. the experimentally do violate SR, as observed Many","Many of these violate the SR, as do experimentally observed correlations.",7862
set can data. fixed with prior be power-priors expressed imaginary set-ups These of,These prior set-ups can be expressed power-priors with fixed set of imaginary data.,7863
and reading the used impaired Braille has The been visually system for writing. by,The Braille system has been used by the visually impaired for reading and writing.,7864
enhance to documents are the processed Braille The reduce pre the noise. and dots,The Braille documents are pre processed to enhance the dots and reduce the noise.,7865
of to language. are appropriate alphabets mapped the These the,These are mapped to the appropriate alphabets of the language.,7866
spoken The through converted a is out text synthesizer. speech,The converted text is spoken out through a speech synthesizer.,7867
and to typed alphabet The mapped is the Braille character spoken out.,The typed Braille character is mapped to the alphabet and spoken out.,7868
a language. The representation Braille mapping for cell each standard differs but has the,The Braille cell has a standard representation but the mapping differs for each language.,7869
"Hindi and this are In mapping of English, considered. paper Tamil","In this paper mapping of English, Hindi and Tamil are considered.",7870
The input audio. be the will output system as scene and of,The input of the system will be scene and output as audio.,7871
minimum dynamic scaling. have extraction spaces using Features to are,Features are extraction to have minimum spaces using dynamic scaling.,7872
"scene, and scale on From the objects formed shape. of are clusters based a","From a scene, clusters of objects are formed based on the scale and shape.",7873
of detect The object minimum is the operation performed. shape possible to the,The minimum operation to detect the possible shape of the object is performed.,7874
"speedy In such way, of this finds way study detecting a objects.","In such way, this study finds a speedy way of detecting objects.",7875
(e.g. are cellular Protein that machinery species conserved processes complexes indicate core to across,Protein complexes conserved across species indicate processes that are core to cellular machinery (e.g.,7876
complexes damage-repair yeast). or cell-cycle human conserved DNA across and,cell-cycle or DNA damage-repair complexes conserved across human and yeast).,7877
with in concerned discrete dynamical This determining is article the solutions time. system of finite,This article is concerned with determining solutions of the discrete dynamical system in finite time.,7878
of kneading matrices appealing address and the by kneading this We problem concept to determinants.,We address this problem by appealing to the concept of kneading matrices and kneading determinants.,7879
of in can terms theses also dynamics The be infinite systems described of matrices.uation. Leslie,The dynamics of theses systems can also be described in terms of infinite Leslie matrices.uation.,7880
only. the on correspondences is based image Calibration,Calibration is based on the image correspondences only.,7881
on validated data. image implemented and algorithm is synthetic The,The algorithm is implemented and validated on synthetic image data.,7882
from is most the among in common outcrossing to transitions self-fertilization shift The plants.,The shift from outcrossing to self-fertilization is among the most common transitions in plants.,7883
haplotypes) founder(s). founding rubella's C. in present,founding haplotypes) present in C. rubella's founder(s).,7884
experiments. vary the on of These the wildly conditions strategies depending can,These strategies can vary wildly depending on the conditions of the experiments.,7885
doses with balanced explained more of are reinforcement Other inference. experiments and,Other experiments are explained with more balanced doses of reinforcement and inference.,7886
consistent These with results experimental findings. are,These results are consistent with experimental findings.,7887
"true That explored makes or usually it not background is which is, false. NC information","That is, it is usually not explored which background information makes NC true or false.",7888
to this gap. fill paper The present aims,The present paper aims to fill this gap.,7889
We in specific study NC two settings.,We study NC in two specific settings.,7890
"of In complete some individuals is a first description known, case, the e.g.","In the first case, a complete description of some individuals is known, e.g.",7891
"a e.g. individuals given, property the is In particular having second case, number of the","In the second case, the number of individuals having a particular property is given, e.g.",7892
"are to interaction often networks their new used topology. ""learn"" from Protein (PINs) biological function","Protein interaction networks (PINs) are often used to ""learn"" new biological function from their topology.",7893
PIN to LP existing predict uses links. the and topology missing spurious,LP uses the existing PIN topology to predict missing and spurious links.,7894
"After de-noise to evaluating the methods, them use different PINs. we","After evaluating the different methods, we use them to de-noise PINs.",7895
executables are available upon freely Software request.,Software executables are freely available upon request.,7896
on identification transition the are constraints. based of Many of so-called them successive,Many of them are based on the successive identification of so-called transition constraints.,7897
improve assemblies. goal produce trimmed reads Our genome and error-corrected is to to,Our goal is to produce trimmed and error-corrected reads to improve genome assemblies.,7898
are widely in the in of brain. activity neurons correlations the Positive observed,Positive correlations in the activity of neurons are widely observed in the brain.,7899
"Here, by positive that factors. enhance show can astronomical coding performance we correlations","Here, we show that positive correlations can enhance coding performance by astronomical factors.",7900
"can error be magnitude. discrimination of many by the suppressed probability of Specifically, orders","Specifically, the probability of discrimination error can be suppressed by many orders of magnitude.",7901
"Likewise, of similarly be enhanced by large factors. number stimuli the capacity--can encoded--the","Likewise, the number of stimuli encoded--the capacity--can be enhanced by similarly large factors.",7902
commonly in show from heterogeneity activity. We further that both benefit seen population effects,We further show that both effects benefit from heterogeneity commonly seen in population activity.,7903
a capacity of Error suppression enhancement pattern and upon rest correlation.,Error suppression and capacity enhancement rest upon a pattern of correlation.,7904
experimental discuss We it. of and pattern identify this nature the suggest tests to,We discuss the nature of this pattern and suggest experimental tests to identify it.,7905
likelihood balancing ratio composite Here for detecting tests selection. two we develop,Here we develop two composite likelihood ratio tests for detecting balancing selection.,7906
services are and Increased thought prevention care to costs. decrease improved primary,Increased prevention services and improved primary care are thought to decrease costs.,7907
for construct controlling mathematical We regulatory a the core network model EMT. TGF-\beta-induced,We construct a mathematical model for the core regulatory network controlling TGF-\beta-induced EMT.,7908
coupled irreversible governed switches. the system Mechanistically and reversible by is bistable,Mechanistically the system is governed by coupled reversible and irreversible bistable switches.,7909
bistable variation coupled are switches parameter to and molecular robust Such noise.,Such coupled bistable switches are robust to parameter variation and molecular noise.,7910
mechanistic multiple a provide observations. on explanation experimental We,We provide a mechanistic explanation on multiple experimental observations.,7911
a range is image segmentation Image method step wide of a processing crucial in systems.,Image segmentation is a crucial step in a wide range of method image processing systems.,7912
image. objects the the is in useful visualization It different of present in,It is useful in visualization of the different objects present in the image.,7913
an towards edges very of Detection is step features. understanding image important in a image,Detection of edges in an image is a very important step towards understanding image features.,7914
several from objectively. be detection measured of criteria edge can Quality The,The Quality of edge detection can be measured from several criteria objectively.,7915
can tumor specifically enhancing be mammographic in for the This images. area used,This can be used specifically for enhancing the tumor area in mammographic images.,7916
"Different like Sobel, for edge Prewitt, detection Canny, are methods operators. Roberts, edge available Log","Different methods are available for edge detection like Roberts, Sobel, Prewitt, Canny, Log edge operators.",7917
In detection edge paper mammographic for algorithms proposed a been novel images. this has for,In this paper a novel algorithms for edge detection has been proposed for mammographic images.,7918
"Breast method. be location region and pectoral clearly this using can boundary, seen by tumor","Breast boundary, pectoral region and tumor location can be seen clearly by using this method.",7919
of demonstrate the Experimental effectiveness the approach. results proposed,Experimental results demonstrate the effectiveness of the proposed approach.,7920
located randomly genomes. along are Genes not,Genes are not located randomly along genomes.,7921
theory to be grown area an has graph intense study. of Evolutionary,Evolutionary graph theory has grown to be an area of intense study.,7922
differences value. These accounted reproductive are by for,These differences are accounted for by reproductive value.,7923
"mating, random gene sampled a Under parental progeny's pools. the alleles are from independently","Under random mating, a progeny's alleles are independently sampled from the parental gene pools.",7924
such of number are existence theoretical The and established. selection coefficients,The theoretical existence and number of such selection coefficients are established.,7925
AMP. methods available classes use different approach prediction antimicrobial Most (AMP) of common peptides for,Most available antimicrobial peptides (AMP) prediction methods use common approach for different classes of AMP.,7926
"image high. processing, low, three in are They middle into is types that classified and","They are classified into three types in image processing, that is low, middle and high.",7927
enhancing method of KNN. an is MKNN,MKNN is an enhancing method of KNN.,7928
classification proposed KNN is The called MKNN.,The proposed KNN classification is called MKNN.,7929
to its according of validity point is neighbors. each The computed,The validity of each point is computed according to its neighbors.,7930
"logically alignments for incoherent. are produced large the ontologies of However, most","However, most of the alignments produced for large ontologies are logically incoherent.",7931
study and functions. sparsity-inducing paper a we penalty propose family In of this,In this paper we propose and study a family of sparsity-inducing penalty functions.,7932
"Moreover, coordinate algorithm the we especially for a appropriate show is that descent function. KEP","Moreover, we show that a coordinate descent algorithm is especially appropriate for the KEP function.",7933
"gain. I-O reducing response current thus the spike-dependent curve the divisively, adaptation A alters","A spike-dependent adaptation current alters the I-O curve divisively, thus reducing the response gain.",7934
"always inhibition, variability. ISI the Synaptic increases however,","Synaptic inhibition, however, always increases the ISI variability.",7935
produce coherently regions colored Segmentation algorithms of pixels.,Segmentation algorithms produce regions of coherently colored pixels.,7936
the processes serve proto-objects determine These of attentional which on image relevance. portions as,These serve as proto-objects on which the attentional processes determine image portions of relevance.,7937
necessarily the not of full object---constitutes A attention. focus region---which single a represents,A single region---which not necessarily represents a full object---constitutes the focus of attention.,7938
against challenge. antimicrobial antibiotic of new agents nowadays Development resistance pathogens the the is,Development of the new antimicrobial agents against antibiotic resistance pathogens is the nowadays challenge.,7939
antibiotic current templates agents. have increasingly Therefore they prospective research in for become as recognized,Therefore they have become increasingly recognized in current research as templates for prospective antibiotic agents.,7940
"action, physically AMP disrupt bacterial to mechanisms supposed with According of the and membranes. interact","According to supposed mechanisms of action, AMP interact with and physically disrupt the bacterial membranes.",7941
of them its permeability membrane damage. give ability and,give them ability of membrane permeability and its damage.,7942
coexistence is transition At these separating there phase. discontinuous line the non-ergodic phases a,At the discontinuous transition line separating these phases there is a non-ergodic coexistence phase.,7943
object recognition. role play a vital in Shapes,Shapes play a vital role in object recognition.,7944
"and The in scale, orientation. position shapes image can an varying occur basic in","The basic shapes in an image can occur in varying scale, position and orientation.",7945
satisfactory In demonstrated experiments method proposed the the all results.,In all the experiments the proposed method demonstrated satisfactory results.,7946
algorithms. is common tuning a for issue Parameter many tracking,Parameter tuning is a common issue for many tracking algorithms.,7947
has Efficient management security todays an important parameter world. in become,Efficient security management has become an important parameter in todays world.,7948
for subtraction propose real using we this model background In a paper time AGMM.,In this paper we propose a model for real time background subtraction using AGMM.,7949
"changes, adaptable The motion. repetitive and illumination background, to is fast dynamic proposed model robust","The proposed model is robust and adaptable to dynamic background, fast illumination changes, repetitive motion.",7950
color Also using method a have detecting model. incorporated Horpresert we shadows for the,Also we have incorporated a method for detecting shadows using the Horpresert color model.,7951
theoretical to and treatment. neuroscience ongoing disease medical challenge to an Parkinson's is,Parkinson's disease is an ongoing challenge to theoretical neuroscience and to medical treatment.,7952
organisms. on gradient Models of based are for sensing individual chemotaxis,Models for chemotaxis are based on gradient sensing of individual organisms.,7953
of Keller-Segel derivation the to model model. original surprisingly similar is This,This model is surprisingly similar to the original derivation of Keller-Segel model.,7954
and traveling numerical of A band solutions comprehensive provided front is picture with simulations.,A comprehensive picture of traveling band and front solutions is provided with numerical simulations.,7955
"conferences. new the main published vision models Each year, more in and are more computer","Each year, more and more new models are published in the main computer vision conferences.",7956
"assess proposed of models to saliency. In paper, is new framework visual a this","In this paper, a new framework is proposed to assess models of visual saliency.",7957
to answer three Statistical questions. is used analysis each these here of,Statistical analysis is used here to answer each of these three questions.,7958
without significantly high Pruning savings accuracy. sacrificing classification achieve can a computational of degree,Pruning can achieve a high degree of computational savings without significantly sacrificing classification accuracy.,7959
"found does that however, classification we performance not improve. weighting, saliency In","In saliency weighting, however, we found that classification performance does not improve.",7960
using salient regions combine them the Multiple non-salient treat and We separately Learning. and Kernel,We treat the salient and non-salient regions separately and combine them using Multiple Kernel Learning.,7961
"often contours, is initial active segmentation the by contour required. In object","In object segmentation by active contours, the initial contour is often required.",7962
"initial the by contour is Conventionally, user. the provided","Conventionally, the initial contour is provided by the user.",7963
optimization. Projected-gradient the solve constrained to are derived descent equations,Projected-gradient descent equations are derived to solve the constrained optimization.,7964
has malignant. The texture from mammogram benign of different,The benign of mammogram has different texture from malignant.,7965
enormous task of analysis Automatic is sets of sciences. the a in critical images life,Automatic analysis of the enormous sets of images is a critical task in life sciences.,7966
to the A framework. presented study investigate case is proposed,A case study is presented to investigate the proposed framework.,7967
Many influence sequence gene by DNA phenotypes expression. variants altering,Many DNA sequence variants influence phenotypes by altering gene expression.,7968
single-cell of tags. through abundance protein The the method green-fluorescent-protein use measures,The method measures single-cell protein abundance through the use of green-fluorescent-protein tags.,7969
phenotypic diversity Evolutionary populations shape genetic to forces patterns variation. contribute within of and,Evolutionary forces shape patterns of genetic diversity within populations and contribute to phenotypic variation.,7970
"empirical In significant selection has particular, interest attracted recurrent positive in both theoretical studies. and","In particular, recurrent positive selection has attracted significant interest in both theoretical and empirical studies.",7971
the parameter a We theoretical analysis of detailed space. across robustness rescaling the perform of,We perform a detailed theoretical analysis of the robustness of rescaling across the parameter space.,7972
"data have component important In transcriptome past many decade, the studies. an become phylogenetic of","In the past decade, transcriptome data have become an important component of many phylogenetic studies.",7973
have or semiautomated. Phylogenomic manual therefore studies been,Phylogenomic studies have therefore been manual or semiautomated.,7974
"that an automated We present Agalma, phylogenomic tool conducts analyses.","We present Agalma, an automated tool that conducts phylogenomic analyses.",7975
built-in Agalma test includes data and data. these a a analysis of set test,Agalma includes a test data set and a built-in test analysis of these data.,7976
"to phylogenomic modified, reproduced, This and will be studies enable readily extended.","This will enable phylogenomic studies to be readily reproduced, modified, and extended.",7977
neighborhood is in common Similar models. mathematical local for structure disregard,Similar disregard for local neighborhood structure is common in mathematical models.,7978
game evolutionary progression between dynamics an cellular phenotypes. the of as disease view Methods: We,Methods: We view the dynamics of disease progression as an evolutionary game between cellular phenotypes.,7979
(go) strategy. the show spatial can promote structure Results: invasive We that,Results: We show that spatial structure can promote the invasive (go) strategy.,7980
"using available SBML packages, models and and Qual BioModels are encoded through All Core Database.","All models are encoded using SBML Core and Qual packages, and available through BioModels Database.",7981
Most also understand maps. models SBGN available graphical are to easy as,Most models are also available as easy to understand graphical SBGN maps.,7982
these of experiments variety A of support models. for each have claimed transport,A variety of transport experiments have claimed support for each of these models.,7983
coalescent perform sequences generate to and at host subsequently simulations loci. parasite We the,We subsequently perform coalescent simulations to generate sequences at the host and parasite loci.,7984
 is with process Wright-Fisher The an population in tool  genetics selection important theory.,  The Wright-Fisher process with selection is an important tool in population genetics theory.,7985
relies process diffusion the of analysis approximation. on Traditional this,Traditional analysis of this process relies on the diffusion approximation.,7986
is The differential usually framework. approximation partial equations in diffusion studied a,The diffusion approximation is usually studied in a partial differential equations framework.,7987
"be system. target required the applied When adaptable are systems the to to robotics, to","When applied to robotics, the systems are required to be adaptable to the target system.",7988
identify such population to are posed Current signals. techniques not well genomic,Current population genomic techniques are not well posed to identify such signals.,7989
"timepoint applicable Here, inference develop escape datasets. rates for single we methods to CTL","Here, we develop inference methods for CTL escape rates applicable to single timepoint datasets.",7990
A signal Eurasian is eastern present Africa. throughout west of similar ancestry,A similar signal of west Eurasian ancestry is present throughout eastern Africa.,7991
"this paper, In we images managing of problem tagged address with a hybrid summarization.","In this paper, we address a problem of managing tagged images with hybrid summarization.",7992
experimental of demonstrate effectiveness and approach. the The efficiency roposed the results,The experimental results demonstrate the effectiveness and efficiency of the roposed approach.,7993
to in genetic observed diversity the of genetic attributes reduction Simulation most drift.,Simulation attributes most of the observed reduction in genetic diversity to genetic drift.,7994
method accuracy The values mean by proposed using the and evaluated error. is of,The proposed method is evaluated by using the mean values of accuracy and error.,7995
breast method more accurately. comparative the The identifies that the shows analysis proposed region,The comparative analysis shows that the proposed method identifies the breast region more accurately.,7996
not is they formed understood. are How,How they are formed is not understood.,7997
"Potentiation time. the weights reduction gradual, decay activity-independent over synaptic is of","Potentiation decay is the gradual, activity-independent reduction of synaptic weights over time.",7998
length depend on the distributions potentiation demonstrate decay. chain We that,We demonstrate that chain length distributions depend on the potentiation decay.,7999
nervous waves ultrasonic excite neurons. can central and Low-intensity (CNS) nondestructively remotely system,Low-intensity ultrasonic waves can remotely and nondestructively excite central nervous system (CNS) neurons.,8000
source software open available under and is Public the described The License. freely GNU,The software described is freely available and open source under the GNU Public License.,8001
between genes. of relationship duplicated the and evolutionary traced We timing replication age the,We traced the relationship between replication timing and the evolutionary age of duplicated genes.,8002
"system the has ground truth. the of many certain to In problem scenarios, statistics access","In many problem scenarios, the system has access to certain statistics of the ground truth.",8003
are usually maximize Rational to built rewards. agents,Rational agents are usually built to maximize rewards.,8004
"maximizing ways agents find can of AGI However, prior undesirable function. any reward","However, AGI agents can find undesirable ways of maximizing any prior reward function.",8005
learning crucial Therefore is for AGI. value safe,Therefore value learning is crucial for safe AGI.,8006
We via action for attributes present of dictionary learning maximization. information approach an,We present an approach for dictionary learning of action attributes via information maximization.,8007
the and demonstrate results our Experimental approach effectiveness recognition action summarization. in of,Experimental results demonstrate the effectiveness of our approach in action recognition and summarization.,8008
on geometric simulate novel We methodology a effects. lens to magnification approach based present,We present a novel methodology based on geometric approach to simulate magnification lens effects.,8009
promote aim new to geometric powerful computing. in Our is of techniques applications modeling visual,Our aim is to promote new applications of powerful geometric modeling techniques in visual computing.,8010
user spectrum can of a different effects. specify for also easily The visual metrics,The user can also easily specify a spectrum of metrics for different visual effects.,8011
"robustness, framework. our the effectiveness, and demonstrate of efficiency experimental results Various","Various experimental results demonstrate the effectiveness, robustness, and efficiency of our framework.",8012
effect Most are to thought protein be because non-synonymous their sequence. of on deleterious mutations,Most non-synonymous mutations are thought to be deleterious because of their effect on protein sequence.,8013
We stratified labelings argumentation as to approach introduce a frameworks. novel abstract semantical,We introduce stratified labelings as a novel semantical approach to abstract argumentation frameworks.,8014
a subspace transformation-learning to clustering. propose We low-rank framework robustify,We propose a low-rank transformation-learning framework to robustify subspace clustering.,8015
"further method. prospects the In we discuss conclusion, of","In conclusion, we discuss further prospects of the method.",8016
adaptively The in are to circumstance. transmitted the pulses observed carrier designed frequencies response of,The carrier frequencies of transmitted pulses are adaptively designed in response to the observed circumstance.,8017
"carrier frequencies, investigate are devised. efficient the criterion design then methods and We to","We investigate the criterion to design carrier frequencies, and efficient methods are then devised.",8018
benchmarks. the files for Supplementary replicating provided are,Supplementary files are provided for replicating the benchmarks.,8019
for implications on We the comment evolution. networks their and real,We comment on the implications for real networks and their evolution.,8020
a dataset. experiments a conduct and We collected Flickr in dataset manually,We conduct experiments in a manually collected dataset and a Flickr dataset.,8021
clustering that shows correlation performance. result such the the The improves,The result shows that the such correlation improves the clustering performance.,8022
"best phage ""Well-temperate"" population our long-term title growth is from by rate. characterized the","""Well-temperate"" phage from our title is characterized by the best long-term population growth rate.",8023
compensatory The be mutations. that complex extended to require method escapes can,The method can be extended to complex escapes that require compensatory mutations.,8024
the of lasso group to We an of design introduce experiments. application,We introduce an application of the group lasso to design of experiments.,8025
the explain we to trying experimental group design that NOT lasso. for are Note,Note that we are NOT trying to explain experimental design for the group lasso.,8026
"be traded Energy can trade-off, receptors. while against a time allows power-speed","Energy allows a power-speed trade-off, while time can be traded against receptors.",8027
Data Big scientists. challenges opportunities to modern society to bring new data and,Big Data bring new opportunities to modern society and challenges to data scientists.,8028
statistical and computational paradigm. require These and are distinguished challenges new,These challenges are distinguished and require new computational and statistical paradigm.,8029
the We various Data on analysis computation. provide new Big perspectives also and,We also provide various new perspectives on the Big Data analysis and computation.,8030
to conclusions. can inferences statistical scientific consequently and They lead wrong wrong,They can lead to wrong statistical inferences and consequently wrong scientific conclusions.,8031
of and simulation likelihoods of to fast tree trees. lead reconstructed extremely computation CPP,CPP lead to extremely fast computation of tree likelihoods and simulation of reconstructed trees.,8032
distribution the uniform topology shapes on ranked follows (URT). always Their tree,Their topology always follows the uniform distribution on ranked tree shapes (URT).,8033
"Here, to two fully laws are scaling explain these relations. derived cerebral","Here, two laws are derived to fully explain these cerebral scaling relations.",8034
"would decrease of Moreover, paradoxically, a further increase information flow. the efficiency connections long-range in","Moreover, paradoxically, a further increase in long-range connections would decrease the efficiency of information flow.",8035
representation This for paper algorithm images. proposes sparse an image interpolation exploiting natural,This paper proposes an image interpolation algorithm exploiting sparse representation for natural images.,8036
"a to yield multiscale dictionary modeling, is sparse the a representation. directional shearlet chosen For","For the sparse modeling, a shearlet dictionary is chosen to yield a multiscale directional representation.",8037
the How information? combine these does of two brain sources,How does the brain combine these two sources of information?,8038
"the fine of these algorithms remove them blur. edges and However, make many","However, many of these algorithms remove the fine edges and make them blur.",8039
"for paper, ensemble we this image explore method of denoising. learning In Bayesian","In this paper, we explore Bayesian method of ensemble learning for image denoising.",8040
"algorithms. are ensemble and Bayesian successful of models very Non-local the recent Fields Experts, Means","Bayesian ensemble models are Non-local Means and Fields of Experts, the very successful recent algorithms.",8041
an of The contains that amount self-similarity. extensive image the Non-local presumes Means,The Non-local Means presumes that the image contains an extensive amount of self-similarity.,8042
on image natural models two images. performed is denoising and The are implemented,The two models are implemented and image denoising is performed on natural images.,8043
real recognition technique describes a head gesture of system. This time paper,This paper describes a technique of real time head gesture recognition system.,8044
various proposed implemented be The can system. in control model,The proposed model can be implemented in various control system.,8045
result are implementation presenting We mentioned method. both of and also the,We are also presenting the result and implementation of both mentioned method.,8046
grasping kept theoretical eye on problem. the At we an same of time the,At the same time we kept an eye on theoretical grasping of the problem.,8047
on limit Based simulations computer the our opinion distribution is Gaussian. our is that,Based on our computer simulations our opinion is that the limit distribution is Gaussian.,8048
used frequently coordinates graphics Barycentric as shade computer to images. are interpolants,Barycentric coordinates are frequently used as interpolants to shade computer graphics images.,8049
computation is using a representation. amenable block-normalized to technique integer This,This technique is amenable to computation using a block-normalized integer representation.,8050
drawback applications. serious potential and overshadows modeling a Such many analysis,Such a serious drawback overshadows many potential modeling and analysis applications.,8051
volumetric to this techniques well-known solution the modeling become problem. Consequently,Consequently volumetric modeling techniques become the well-known solution to this problem.,8052
"models, probabilistic mixture Gaussian learning tasks. widely models, for unsupervised used as such are Hierarchical","Hierarchical probabilistic models, such as Gaussian mixture models, are widely used for unsupervised learning tasks.",8053
"errors, square using was mean accuracy The evaluated correlation. mutual information and registration","The registration accuracy was evaluated using mean square errors, mutual information and correlation.",8054
We with investigate games evolutionary zealots two-strategy finite dynamics in populations. of matrix,We investigate evolutionary dynamics of two-strategy matrix games with zealots in finite populations.,8055
take strategy the fitness. regardless either of are assumed Zealots to,Zealots are assumed to take either strategy regardless of the fitness.,8056
study in time scenario. We fixation this,We study fixation time in this scenario.,8057
We examples games. social various our dilemma results illustrate with of,We illustrate our results with examples of various social dilemma games.,8058
these regulating also in miRNAs genes contribution was of demonstrated. differential The,The contribution of miRNAs in regulating these differential genes was also demonstrated.,8059
account. noise few taking analysis and recent A papers have performed multiplexing into characteristics,A few recent papers have performed analysis taking multiplexing and noise characteristics into account.,8060
"we present paper, framework In three components. a analysis this comprehensive incorporating all","In this paper, we present a comprehensive analysis framework incorporating all three components.",8061
A confers unique two GMM characteristics. prior,A GMM prior confers two unique characteristics.,8062
population curve how shows paper responds in size. LD changes to This the,This paper shows how the LD curve responds to changes in population size.,8063
size generates an flat. but curve that A LD relatively high in reduction is,A reduction in size generates an LD curve that is high but relatively flat.,8064
"population suggesting the of a curve is steep, data, In European expansion. history","In European data, the curve is steep, suggesting a history of population expansion.",8065
the in which Stochastic an edges weights. probability graphs unknown are graphs have distribution,Stochastic graphs are graphs in which the edges have an unknown distribution probability weights.,8066
"be which determines At the eDLA stage algorithm, to edges sampled. proposed of each","At each stage of the proposed algorithm, eDLA determines which edges to be sampled.",8067
the endemic reproduction equilibrium. basic analysis conducted forward number sensitivity Static the is and on,Static forward sensitivity analysis is conducted on the basic reproduction number and the endemic equilibrium.,8068
"In pathway, deleterious fixed to becomes occurrence a prior the one mutation of mutation. compensatory","In one pathway, a deleterious mutation becomes fixed prior to occurrence of the compensatory mutation.",8069
"two are In loci the simultaneously other, the polymorphic.","In the other, the two loci are simultaneously polymorphic.",8070
on effect simulation. of of The process the the are explored using dynamics substitution recombination,The effect of recombination on the dynamics of the substitution process are explored using simulation.,8071
for an MaLeS theorem tuning automatic framework automated is provers.,MaLeS is an automatic tuning framework for automated theorem provers.,8072
finding as both It the strategy for strategy provides well scheduling as the solutions problem.,It provides solutions for both the strategy finding as well as the strategy scheduling problem.,8073
algorithms  real-time video-game heuristic in have outstanding search pathfinding. demonstrated  Recent performance,  Recent real-time heuristic search algorithms have demonstrated outstanding performance in video-game pathfinding.,8074
"applications to limited thus have domain. been far However, their that","However, their applications have been thus far limited to that domain.",8075
"a We IQA and deviation magnitude model, effective called gradient similarity present (GMSD). new efficient","We present a new effective and efficient IQA model, called gradient magnitude similarity deviation (GMSD).",8076
examine model-based we from MID Here a perspective.,Here we examine MID from a model-based perspective.,8077
primate illustrate methods simulations cortex. with visual We these data from and,We illustrate these methods with simulations and data from primate visual cortex.,8078
stable very nonlinearity. Only coexistence by lead specific relative trade-offs to evolutionarily,Only very specific trade-offs lead to evolutionarily stable coexistence by relative nonlinearity.,8079
"competing simulation to benchmarking However, exist enable few correction of methods. tools advanced","However, few advanced simulation tools exist to enable benchmarking of competing correction methods.",8080
relapses in early Most are breast the cancer category. in,Most relapses in breast cancer are in the early category.,8081
Forget suggests mechanism. unexpected an Recent from et al data,Recent data from Forget et al suggests an unexpected mechanism.,8082
with therapy. were conventional treated Patients adjuvant and mastectomy,Patients were treated with mastectomy and conventional adjuvant therapy.,8083
heuristic. algorithm has as been The previously EqM motivated a,The EqM algorithm has previously been motivated as a heuristic.,8084
interactions two beneficial. when networks between species are mutually are of classes the Mutualistic formed,Mutualistic networks are formed when the interactions between two classes of species are mutually beneficial.,8085
examples evolution. of shaped are They important cooperation by,They are important examples of cooperation shaped by evolution.,8086
organization plays in a communities. animals ecological and between Mutualism key of plants the role,Mutualism between animals and plants plays a key role in the organization of ecological communities.,8087
data. modeling of a general We for reduced-rank framework propose matrix-valued,We propose a general framework for reduced-rank modeling of matrix-valued data.,8088
"estimated using Moreover, optimization. these class scalably of large models a convex can be","Moreover, a large class of these models can be estimated scalably using convex optimization.",8089
"signals. of In communication combinations system, signals a some the combinatorial of consist other","In a combinatorial communication system, some signals consist of the combinations of other signals.",8090
basic the principles derive three We of about combinatorial communication. emergence,We derive three basic principles about the emergence of combinatorial communication.,8091
"other species, ability. probably but have this no Humans,","Humans, but probably no other species, have this ability.",8092
We tested that threatening the hypothesis creativity. enhance situations,We tested the hypothesis that threatening situations enhance creativity.,8093
were of level for The creativity. rated stories,The stories were rated for level of creativity.,8094
Theoretical with consistent findings frameworks are these discussed.,Theoretical frameworks consistent with these findings are discussed.,8095
alternating of the we study Prisoner's Here direct reciprocity in context the Dilemma.,Here we study direct reciprocity in the context of the alternating Prisoner's Dilemma.,8096
implemented consider strategies can that be We automata. and by two-state all one,We consider all strategies that can be implemented by one and two-state automata.,8097
payoff all of calculate the presence noise. encounters pairwise in the matrix We of,We calculate the payoff matrix of all pairwise encounters in the presence of noise.,8098
mutation. explore without with deterministic and selection dynamics We,We explore deterministic selection dynamics with and without mutation.,8099
strict are always-defect uncooperative Two of representing equilibria them and Grim. Nash (ALLD),Two of them are uncooperative strict Nash equilibria representing always-defect (ALLD) and Grim.,8100
and facilitates noise. cooperation of the presence exploitation in Forgiveness stable,Forgiveness facilitates stable cooperation in the presence of exploitation and noise.,8101
epigenetic control implications of for the results have The mechanics. DNA,The results have implications for the epigenetic control of DNA mechanics.,8102
significantly viewers positively were Authenticity correlated. ratings amongst,Authenticity ratings amongst viewers were significantly positively correlated.,8103
performers were all and not between viewers Ratings significant positive. but,Ratings between viewers and performers were not significant but all positive.,8104
relationship The between authenticity and creativity discussed. is,The relationship between authenticity and creativity is discussed.,8105
"genotypes. similarity that show Here, this we to extends","Here, we show that this similarity extends to genotypes.",8106
"strength theory assumes most population genetic of constant. the remains that However, selection","However, most population genetic theory assumes that the strength of selection remains constant.",8107
"applications it found has in non-monotonic Since in in reasoning, further reasoning. then particular defeasible","Since then it has found further applications in non-monotonic reasoning, in particular in defeasible reasoning.",8108
How stimuli time-dependent populations encode do rate? firing population in their neuronal,How do neuronal populations encode time-dependent stimuli in their population firing rate?,8109
study The systems. problem within autonomous present for solving framework gives mathematical self-evolution a,The present study gives a mathematical framework for self-evolution within autonomous problem solving systems.,8110
come How be so intelligent? we did to,How did we come to be so intelligent?,8111
fuzzy one the provide about tools core measures sets. Similarity that enable reasoning of,Similarity measures provide one of the core tools that enable reasoning about fuzzy sets.,8112
present for Gibbs state a We modified spaces. sampler general,We present a modified Gibbs sampler for general state spaces.,8113
extend The models and pathway biology. network reproduce resulting known,The resulting network models reproduce and extend known pathway biology.,8114
diverse molecular is biology. The areas larger technology to in for suitable application of systems,The technology is suitable for application to larger systems in diverse areas of molecular biology.,8115
maintains GMIC MIC. asymptotic key We properties define it prove of and several,We define GMIC and prove it maintains several key asymptotic properties of MIC.,8116
"correlation, and are to Pearson distance results The the MIC. compared correlation,","The results are compared to the Pearson correlation, distance correlation, and MIC.",8117
"statistics some surpasses GMIC For functional calculated. all relationships, other","For some functional relationships, GMIC surpasses all other statistics calculated.",8118
"of underlying in powerful relationships. our for many forms Nonetheless, more simulations correlation was distance","Nonetheless, distance correlation was in our simulations more powerful for many forms of underlying relationships.",8119
and concern influenza Seasonal public health the is a in significant globally. United States,Seasonal influenza is a significant public health concern in the United States and globally.,8120
are often used persistent that such Networks support to activity. bump model attractors,Networks that support bump attractors are often used to model such persistent activity.,8121
"bumps not away. in position and stable, do decay Thus are neutrally activity perturbations","Thus activity bumps are neutrally stable, and perturbations in position do not decay away.",8122
attractor strength balancing connections. requires the Such of recurrent an precisely synaptic,Such an attractor requires precisely balancing the strength of recurrent synaptic connections.,8123
with strategy environment. The often coping an its behaviour organism a for reflects of,The behaviour of an organism often reflects a strategy for coping with its environment.,8124
Protein function structure both and on depends protein sequence. amino (aa) acid,Protein function depends on both protein structure and amino acid (aa) sequence.,8125
Spinodal crossings by in marked of are algebraic sequences. of functions slope breaks aa,Spinodal crossings are marked by breaks in slope of algebraic functions of aa sequences.,8126
is associated new effective temperature with than The standard the a lower scale. scale,The new scale is associated with a lower effective temperature than the standard scale.,8127
we review work a the this present on four popular paper HCR scripts. on In,In this paper we present a review on HCR work on the four popular scripts.,8128
We direction HCR. of some research future also added on,We also added some future direction of research on HCR.,8129
solving industrial SAT progress on instances. have remarkable a experienced solvers Modern,Modern SAT solvers have experienced a remarkable progress on solving industrial instances.,8130
developed the after techniques experimental process. intensive of have been an Most testing,Most of the techniques have been developed after an intensive experimental testing process.,8131
not the that We clauses. this affected is of also learnt addition dimension show by,We also show that this dimension is not affected by the addition of learnt clauses.,8132
"properties state-of-the-art we empirical Finally, can in evidence be portfolios. used these graph that give","Finally, we give empirical evidence that these graph properties can be used in state-of-the-art portfolios.",8133
is return in Inhibit out. saliency of the sequence (IOR) applied object pop of judging,Inhibit of return (IOR) is applied in judging the sequence of saliency object pop out.,8134
such applied many method can as into fields The be surveillance and security. video,The method can be applied into many fields such as video surveillance and security.,8135
"reference of the human the distribution genome. question, address evaluate To k-mer we this","To address this question, we evaluate the k-mer distribution of the human reference genome.",8136
noisy problem to is up now. a documents challenging ancient or Characterizing,Characterizing noisy or ancient documents is a challenging problem up to now.,8137
approaches. local less Global and approaches in exact general robust are than,Global approaches are in general less robust and exact than local approaches.,8138
"Next, match the applied request. the is which show SIFT well to images","Next, the SIFT is applied to show which images match well the request.",8139
the complementary of nodes (nearly) strings. links connect The with network,The links of the network connect nodes with (nearly) complementary strings.,8140
an marrow. influx There the idiotype lymphocytes random bone of with is from,There is an influx of lymphocytes with random idiotype from the bone marrow.,8141
"A software is scalable reality architecture system virtual proposed. and flexible, affordable","A flexible, scalable and affordable virtual reality software system architecture is proposed.",8142
object properties hidden Implicit (i.e. features represent,Implicit properties represent hidden object features (i.e.,8143
"the we this, To mathematically carcinogenesis. of modeled effect on dedifferentiation address","To address this, we mathematically modeled the effect of dedifferentiation on carcinogenesis.",8144
next done A features. phase global of and is by combining local wordspotting,A phase of wordspotting is done next by combining local and global features.,8145
ones. We cartographic applied fontspotting both ancient documents and on,We applied fontspotting on both ancient documents and cartographic ones.,8146
a requires it. of the Recognition compose of recognition then all word that characters (OCR),Recognition of a word then requires the recognition of all characters (OCR) that compose it.,8147
classes mainly the of images old of in work focuses characterization documents. on Our,Our work focuses mainly on the characterization of classes in images of old documents.,8148
toolbox documents. for classes We Som use finding in,We use Som toolbox for finding classes in documents.,8149
ancient fractal applied and and also dimensions documents. match points to interest categorize of We,We applied also fractal dimensions and points of interest to categorize and match ancient documents.,8150
specially method The a approximating partitions transformed designed for on is image.,The method is specially designed for approximating partitions on a transformed image.,8151
proposed of BIQA learning The one is method to rank.,The proposed BIQA method is one of learning to rank.,8152
strategy to but A scores effective is estimate simple quality then image presented. perceptual,A simple but effective strategy to estimate perceptual image quality scores is then presented.,8153
"new can be to the proposed distortion easily method Moreover, extended categories.","Moreover, the proposed method can be easily extended to new distortion categories.",8154
"possible instance, probabilistic dynamic ensemble behaviour. model for predictions, The allows about resulting for","The resulting model ensemble allows for probabilistic predictions, for instance, about possible dynamic behaviour.",8155
"adding can be or data the more By tighter predictions more precise. made constraints,","By adding more data or tighter constraints, the predictions can be made more precise.",8156
is elasticity MATLAB available. freely code sampling for,MATLAB code for elasticity sampling is freely available.,8157
bone long on growing How spatial such unresolved. is domain the organization emerges still,How such spatial organization emerges on the growing long bone domain is still unresolved.,8158
patterns Turing interactions a general may to nature. mechanism constitute generate thus Receptor-ligand in,Receptor-ligand interactions may thus constitute a general mechanism to generate Turing patterns in nature.,8159
to different of use approaches. avian modeling under culling understand mathematical dynamics the influenza We,We use mathematical modeling to understand the dynamics of avian influenza under different culling approaches.,8160
"box proposed performance. the Moreover, model deblurring on to is constraint imposed a guarantee high","Moreover, a box constraint is imposed on the proposed model to guarantee high deblurring performance.",8161
"the and reconstructing distribution, on changes (i.e. these branch length that, depending show We tree","We show that, depending on the tree and branch length distribution, reconstructing these changes (i.e.",8162
"trees holds. computer for Moreover, that this more demonstrate scenarios, simulations and complex result still","Moreover, computer simulations demonstrate that for more complex trees and scenarios, this result still holds.",8163
"experiment of thought Leibniz is thinking and perception, sensing, reconsidered.","Leibniz thought experiment of perception, sensing, and thinking is reconsidered.",8164
We Leibniz to try our in picture neuroscience. view understand basic of knowledge of,We try to understand Leibniz picture in view of our knowledge of basic neuroscience.,8165
principle understood. can in we see emergence In the of could be how consciousness particular,In particular we can see how the emergence of consciousness could in principle be understood.,8166
changes images. be as These of texture can features regarded temperature,These temperature changes can be regarded as texture features of images.,8167
"transform multi-direction analyze tool is to of changes multi-scale, good Wavelet a texture.","Wavelet transform is a good tool to analyze multi-scale, multi-direction changes of texture.",8168
Local patterns type binary for vision. used (LBP) feature are a classification of computer in,Local binary patterns (LBP) are a type of feature used for classification in computer vision.,8169
"approaches to the from two region. cropped are different extract the features Secondly, used face","Secondly, two different approaches are used to extract the features from the cropped face region.",8170
performed feature is dimensionality reeducation. PCA set separately individual on for the,PCA is performed separately on the individual feature set for dimensionality reeducation.,8171
classify classifiers different are face to two images. used Finally,Finally two different classifiers are used to classify face images.,8172
network neural classifier minimum is feed classifier forward classifier. One another and distance such multi-layer,One such classifier multi-layer feed forward neural network and another classifier is minimum distance classifier.,8173
and These components over to rendering static allow and implement animated developers techniques algorithms meshes.,These components allow developers to implement rendering techniques and algorithms over static and animated meshes.,8174
first Pearl causation definition Judea the to of actual propose was models. a using causal,Judea Pearl was the first to propose a definition of actual causation using causal models.,8175
very causal potentially complex. models are Extended,Extended causal models are potentially very complex.,8176
given. regret-based A this complete of likelihood axiomatic notion characterization is of,A complete axiomatic characterization of this notion of regret-based likelihood is given.,8177
in Both important dynamics. of population cell evolution the roles the play,Both play important roles in the evolution of the cell population dynamics.,8178
variations the natural continuous the growth The selection. sustains thus among switching subpopulation,The switching sustains the variations among the subpopulation growth thus continuous natural selection.,8179
versa. switching obtained of limiting are cases results growth and vice dominating the Analytical for,Analytical results are obtained for the limiting cases of growth dominating switching and vice versa.,8180
capable constant under hence maintaining and surrounding Humans are of temperature. different homoeothermic temperature,Humans are homoeothermic and hence capable of maintaining constant temperature under different surrounding temperature.,8181
of normally sizes large. the facial are images The,The sizes of the facial images are normally large.,8182
"image transform similarity used before is wavelet So, measured. is the","So, the wavelet transform is used before image similarity is measured.",8183
"used Haar to low the frequency from is band cropped face region. extract wavelet Secondly,","Secondly, Haar wavelet is used to extract low frequency band from the cropped face region.",8184
the they microscopic One one. two advantage is scales: macroscopic and that the bridge different,One advantage is that they bridge two different scales: the microscopic and the macroscopic one.,8185
biological often case. the this systems not In is real,In real biological systems this is often not the case.,8186
"depends longer explicitly Moreover, its no the solution universal: on deterministic it is system.","Moreover, its solution is universal: it no longer depends explicitly on the deterministic system.",8187
We demonstrate of populations. models technique the simple neuronal on,We demonstrate the technique on simple models of neuronal populations.,8188
"In excellent the limit, Fokker-Planck are diffusive achieved. equations approximations of","In the diffusive limit, excellent approximations of Fokker-Planck equations are achieved.",8189
wide-band desirable. spectral An and algorithm HFOs for detecting analysis is automated for,An automated algorithm for detecting HFOs and for wide-band spectral analysis is desirable.,8190
spectra Wide-band time-frequency HFO-associated frequency presented. also are and,Wide-band time-frequency and HFO-associated frequency spectra are also presented.,8191
low-frequency structure. wide-band intricate HFO-associated data show,HFO-associated wide-band data show intricate low-frequency structure.,8192
HFO intensity of ease the may Significance: labor detection. DOOD,Significance: DOOD may ease the labor intensity of HFO detection.,8193
normal HFOs. DOOD may wide-band future from analysis pathological in help distinguish,DOOD wide-band analysis may in future help distinguish normal from pathological HFOs.,8194
the algorithms of straightforward. thinning tables acyclicity render Using implementation,Using acyclicity tables render the implementation of thinning algorithms straightforward.,8195
image. an The goal in detection of is object to objects find,The goal of object detection is to find objects in an image.,8196
"locations object of pairs. accepts $(x,y)$ detector list an as and a An produces image","An object detector accepts an image and produces a list of locations as $(x,y)$ pairs.",8197
we new introduce Here {\bf a concept: boosting}. location-based,Here we introduce a new concept: {\bf location-based boosting}.,8198
was found this small performs samples. procedure that It good in,It was found that this procedure performs good in small samples.,8199
"was parameter However, set. optimal the a on analysis based single","However, the analysis was based on a single optimal parameter set.",8200
ranges. To a within parameter parameter global that conducted screen end physiological the we,To that end we conducted a global parameter screen within the physiological parameter ranges.,8201
"Eventually, reproduced parameter all correctly. situation all sets data that physiological the predicted available","Eventually, all parameter sets that reproduced all available data predicted the physiological situation correctly.",8202
higher reveal . Experimental results degree the performance,Experimental results reveal the higher degree performance .,8203
In HGD for with constraints develop MLE paper . this we,In this paper we develop HGD with constraints for MLE .,8204
Many as represented coordination DCOPs. can multi-agent problems be,Many multi-agent coordination problems can be represented as DCOPs.,8205
under results geometric qualitative We find same hold Fisher's the that model.,We find that the same qualitative results hold under Fisher's geometric model.,8206
the subspace of considers noise. under problem clustering paper This,This paper considers the problem of subspace clustering under noise.,8207
massive datasets any fit single can approach data on too Our handle to also repository.,Our approach can also handle datasets too massive to fit on any single data repository.,8208
images tasks in recognition are (SAR) used nowadays. target widely radar aperture Synthetic,Synthetic aperture radar (SAR) images are widely used in target recognition tasks nowadays.,8209
camera. thermal The infra-red is captured face by of human thermogram,The thermogram of human face is captured by thermal infra-red camera.,8210
of perfusion skin. blood under vessels distribution to related data the face Blood are,Blood perfusion data are related to distribution of blood vessels under the face skin.,8211
Two extracting perfusion methods for minutiae points discussed blood data. from are different,Two different methods are discussed for extracting minutiae points from blood perfusion data.,8212
the A network is propagation five feed-forward layer tool. classification used back as neural,A five layer feed-forward back propagation neural network is used as the classification tool.,8213
problem is Low-rank practical of importance. completion a immense matrix,Low-rank matrix completion is a problem of immense practical importance.,8214
actual takes be graded resulting and The comparative. causation to account both,The resulting account takes actual causation to be both graded and comparative.,8215
show would standard how a handle cases. We of number then our account,We then show how our account would handle a number of standard cases.,8216
is protons the itself may Nearby also make protonated. unless possible reaction C,Nearby protons may also make reaction possible unless the C itself is protonated.,8217
interpretation Therefore the requires of reconsideration. substitution results C,Therefore interpretation of the C substitution results requires reconsideration.,8218
distribution under blood are face of perfusion skin. vessels the data related to Blood,Blood perfusion data are related to distribution of blood vessels under the face skin.,8219
"network backpropagation has For used. neural classification, feed-forward five been a layer","For classification, a five layer feed-forward backpropagation neural network has been used.",8220
the performed Experiments at created laboratory. database been have our own on,Experiments have been performed on the database created at our own laboratory.,8221
"above-chance For at they both the works of guessed sets, levels. creators the","For both sets, they guessed the creators of the works at above-chance levels.",8222
"model (REN) biometric re-weighted we In elastic net propose a for this paper, recognition.","In this paper, we propose a re-weighted elastic net (REN) model for biometric recognition.",8223
is to and model geometric components. new into color spatial The data separated applied,The new model is applied to data separated into geometric and color spatial components.,8224
"according for classification are sparse obtained. to Next, purposes representations fusion-scheme, a Bayesian","Next, according to a Bayesian fusion-scheme, sparse representations for classification purposes are obtained.",8225
using a projection The gradient is solved scheme algorithm. (GP) numerically,The scheme is numerically solved using a gradient projection (GP) algorithm.,8226
new Arabic In recognition cursive system. presented approach for This text we paper,In This paper we presented new approach for cursive Arabic text recognition system.,8227
rapid handwritten to offline propose objective methodology The analytical recognition is implementation. Arabic for of,The objective is to propose methodology analytical offline recognition of handwritten Arabic for rapid implementation.,8228
of with periodic stable studied. evolutionary are Two cases strategy payoffs,Two cases of evolutionary stable strategy with periodic payoffs are studied.,8229
generalization a et of al. The first is Uyttendaele,The first is a generalization of Uyttendaele et al.,8230
prisoner's second with dilemma periodic The is payoff.,The second is prisoner's dilemma with periodic payoff.,8231
appearance culture of Many first Paleolithic. in elements the their Upper made,Many elements of culture made their first appearance in the Upper Paleolithic.,8232
forth explain put unprecedented of hypotheses to this wanting. burst are Previous creativity found,Previous hypotheses put forth to explain this unprecedented burst of creativity are found wanting.,8233
dendrites. extended spatially Neurons that receive are on and process structures their inputs,Neurons are spatially extended structures that receive and process inputs on their dendrites.,8234
"Inevitably, omit applications these a fundamental property neuronal of computation. thus","Inevitably, these applications thus omit a fundamental property of neuronal computation.",8235
to derive closed version is cable used the the formalism Green's of equation. function,Green's function formalism is used to derive the closed version of the cable equation.,8236
correctness the in any genome an of step assemblies project. is important genome Assessing,Assessing the correctness of genome assemblies is an important step in any genome project.,8237
"most cases, computationally exist, are and, in but inappropriate. intensive some methods Several","Several methods exist, but most are computationally intensive and, in some cases, inappropriate.",8238
"popular time-uniform the we convergence adaptive filter. a byproduct, As of particle prove resampling","As a byproduct, we prove time-uniform convergence of the popular adaptive resampling particle filter.",8239
preexisting from that many evolved natural systems signaling Current evolutionary cues. theory suggests,Current evolutionary theory suggests that many natural signaling systems evolved from preexisting cues.,8240
"and signals both benefit systems, warning prey. aposematic predator In prey","In aposematic systems, prey warning signals benefit both predator and prey.",8241
"In observed (i.e. evolution predators, rapid cue of recognition we","In predators, we observed rapid evolution of cue recognition (i.e.",8242
active responses) presented with poisonous behavioral prey. sufficiently when,active behavioral responses) when presented with sufficiently poisonous prey.,8243
conditions utilization. cue all prey evolved that under to mimicry) in led,mimicry) evolved in prey under all conditions that led to cue utilization.,8244
our problem obtained to solution is over optimizing parameters. these The by then,The solution to our problem is then obtained by optimizing over these parameters.,8245
analysis rigorous to present justify cut approach. limit results We our,We present rigorous limit cut analysis results to justify our approach.,8246
resolution mainly single on and Algorithms of wavelet domain. domain are spatial based image,Algorithms of single image resolution are mainly based on wavelet domain and spatial domain.,8247
back and reconstruction is projection Algorithm minimize to error. iterative use,Algorithm is iterative and use back projection to minimize reconstruction error.,8248
denoising is based method noise. Wavelet to introduced also remove,Wavelet based denoising method is also introduced to remove noise.,8249
based optimization there of escape domain crisp on Usually no algorithms. it-erative is,Usually there is no escape of it-erative optimization based on crisp domain algorithms.,8250
network memristor and implement promising algorithms. structures Recently fuzzy appeared neural structures to,Recently memristor structures appeared promising to implement neural network structures and fuzzy algorithms.,8251
two processes. is the between no correlation There,There is no correlation between the two processes.,8252
"closely to predictability evolution linked universality of lineages. trait Furthermore, is across","Furthermore, universality is closely linked to predictability of trait evolution across lineages.",8253
"can be in of congestion any for sector Then, probability airspace the the computed.","Then, the probability of congestion for any sector in the airspace can be computed.",8254
consensus of develop hierarchical novel We clusterings. a,We develop a novel consensus of hierarchical clusterings.,8255
conventions by chess are in composing composers human Composition guidelines used problems.,Composition conventions are guidelines used by human composers in composing chess problems.,8256
particularly tournaments. significant They are in composition,They are particularly significant in composition tournaments.,8257
"but statistically significant found to a does, that but only point. We it slight evidence","We found slight but statistically significant evidence that it does, but only to a point.",8258
These two conflicting suggest main findings things. seemingly,These seemingly conflicting findings suggest two main things.,8259
this defined quadrature to Monte-Carlo problem. specifically algorithms is Clenshaw-Curtis for compared The,The Clenshaw-Curtis quadrature is compared to Monte-Carlo algorithms defined specifically for this problem.,8260
pathways within wide-ranging environmental operate Eukaryotic adaptation conditions stimulus saturation. without,Eukaryotic adaptation pathways operate within wide-ranging environmental conditions without stimulus saturation.,8261
thus mechanisms in on Our new the complex of sheds systems. work light adaptation evolution,Our work thus sheds new light on the evolution of adaptation mechanisms in complex systems.,8262
MH choosing sampling between PCG guidance within Theoretical provide schemes. when arguments different,Theoretical arguments provide guidance when choosing between different MH within PCG sampling schemes.,8263
of most model. each events the of probabilities these parameters important The the of provide,The probabilities of each of these events provide the most important parameters of the model.,8264
"not differentiated death divide, each proportion a programmed do in but cells undergoes generation. Fully","Fully differentiated cells do not divide, but a proportion undergoes programmed death in each generation.",8265
"growth has started, In is process most irreversible. cases, this once exponential","In most cases, once exponential growth has started, this process is irreversible.",8266
paper timetables way This to presents for schools. generate a high novel,This paper presents a novel way to generate timetables for high schools.,8267
"search. through initial and optimization phase tabu Pre-scheduling,","Pre-scheduling, initial phase and optimization through tabu search.",8268
"Finally, algorithm solution based diversification. on tabu optimized frequency is using based the search","Finally, the solution is optimized using tabu search algorithm based on frequency based diversification.",8269
acceptable that effectively show build the Experiments proposed can timetables. algorithm,Experiments show that the proposed algorithm can effectively build acceptable timetables.,8270
mogeneity. to minimizing functions algorithms superpixel color on rely enforce ho- state-of-the-art objective Many,Many state-of-the-art superpixel algorithms rely on minimizing objective functions to enforce color ho- mogeneity.,8271
"real-time they applications. too computationally result, are for expensive As a","As a result, they are computationally too expensive for real-time applications.",8272
based optimization. a new a introduce simple on hill-climbing approach We,We introduce a new approach based on a simple hill-climbing optimization.,8273
"by continuously boundaries. an from Starting the it superpixel modifying superpixels partitioning, the initial refines","Starting from an initial superpixel partitioning, it continuously refines the superpixels by modifying the boundaries.",8274
method spherical a paper numerical both analysis and proposes This with Monte simulation. theoretical Carlo,This paper proposes a spherical Monte Carlo method with both theoretical analysis and numerical simulation.,8275
on some experiments Monte claims. calculation these probabilities Carlo Extensive confirm,Extensive Monte Carlo experiments on some probabilities calculation confirm these claims.,8276
image generated The the results than supervised MS better did. classification fusion of,The supervised classification results of fusion image generated better than the MS did.,8277
results QuickBird best and the classifier than did. other with ED the,QuickBird and the best results with ED classifier than the other did.,8278
common using experimental reproduced biologically observations on justified model parameters. heterosis Our only,Our model reproduced common experimental observations on heterosis using only biologically justified parameters.,8279
"it are that often Moreover, now established signaling well pathways. is between components shared","Moreover, it is now well established that components are often shared between signaling pathways.",8280
These observations oscillatory that ideal are signals. for suggest signals multiplexing,These observations suggest that oscillatory signals are ideal for multiplexing signals.,8281
show that social behaviour change oscillations. trigger We alone sustained may,We show that social behaviour change alone may trigger sustained oscillations.,8282
"show seasonal how briefly of imply chaos. inclusion the may in we variations contacts Finally,","Finally, we briefly show how the inclusion of seasonal variations in contacts may imply chaos.",8283
"ability a discrimination never was across prime However, individuals there relationship and between priming.","However, across individuals there was never a relationship between prime discrimination ability and priming.",8284
"A. and M. S. Perelson. Taylor A. Makalkin, A.","A. Makalkin, M. A. Taylor and A. S. Perelson.",8285
Nonlinear tumors: Parameter and estimation of immunogenic bifurcation analysis. global dynamics,Nonlinear dynamics of immunogenic tumors: Parameter estimation and global bifurcation analysis.,8286
Paternal of number age conception found genetic mutations. been new at predict to the has,Paternal age at conception has been found to predict the number of new genetic mutations.,8287
association a the parental had order IQ for negative. positive-zero of effect turning Controlling,Controlling for parental IQ had the effect of turning a positive-zero order association negative.,8288
"part discuss. which program general also of The a is more we project,","The program is part of a more general project, which we also discuss.",8289
a median mesh technique weighted uses smoothing. filtering method present for The,The present method uses a weighted mesh median filtering technique for smoothing.,8290
Computations of are binations all these com- desired variables. for,Computations are desired for all com- binations of these variables.,8291
an simulation of The a study array. typically object result final is,The final result object of a simulation study is typically an array.,8292
achieve above to The the tasks. provides several package R tools simsalapar,The R package simsalapar provides several tools to achieve the above tasks.,8293
"the analyzing pro- results are vided. graphics Furthermore, or tables tools via for","Furthermore, tools for analyzing the results via tables or graphics are pro- vided.",8294
Temporal networks. approach coding is one in neural information representing spiking to,Temporal coding is one approach to representing information in spiking neural networks.,8295
"Hence, noise inputs caused reduced. is across delays these by different synaptic","Hence, noise caused by different synaptic delays across these inputs is reduced.",8296
distributions. Markov We (MCMC) for new Monte Carlo target a method propose Chain constrained,We propose a new Markov Chain Monte Carlo (MCMC) method for constrained target distributions.,8297
"resulting sphere, parameter the space it $D$-dimensional S}^D$. augments the to ${\bf Then,","Then, it augments the resulting parameter space to the $D$-dimensional sphere, ${\bf S}^D$.",8298
provide biological can processes. of into images microscopy insight many Analysis,Analysis of microscopy images can provide insight into many biological processes.,8299
flexible non-parametric Bayesian Tree is  density estimation. P\'{o}lya  a model (OPT) Optional for,  Optional P\'{o}lya Tree (OPT) is a flexible non-parametric Bayesian model for density estimation.,8300
"inference its OPT merits, challenging. for computation the is Despite","Despite its merits, the computation for OPT inference is challenging.",8301
two improvements simulations. the these of demonstrate using We performance,We demonstrate the performance of these two improvements using simulations.,8302
"are large In clusters. systems, many spatially sensory transmembrane organized in receptors","In many sensory systems, transmembrane receptors are spatially organized in large clusters.",8303
multiple of may and facilitate stimuli. arrangement Such signal amplification integration the,Such arrangement may facilitate signal amplification and the integration of multiple stimuli.,8304
in receptor activity fluctuations quantified been experimentally. have Large,Large fluctuations in receptor activity have been quantified experimentally.,8305
"That important. choice means, may learning methods of dictionary not be the","That means, the choice of dictionary learning methods may not be important.",8306
to different closure moment different assumptions lead mean-field systems. Three three,Three different moment closure assumptions lead to three different mean-field systems.,8307
full the for numerical This with biologically simulations parameter reasonable confirmed network of values. is,This is confirmed with numerical simulations of the full network for biologically reasonable parameter values.,8308
"regression. This data advanced such requires functional methods, of analysis the use as","This requires the use of advanced data analysis methods, such as functional regression.",8309
an experimental methodology vineyards. a is to design case in applied study involving French The,The methodology is applied to a case study involving an experimental design in French vineyards.,8310
"well are genericity robustness. discussed, and the as as method results The","The results are discussed, as well as the method genericity and robustness.",8311
"primary the thinking Like faculties of today, were individuals. reasoning and","Like today, reasoning and thinking were the primary faculties of individuals.",8312
cortex Its smaller the brain convoluted. less was was and apparently,Its brain was smaller and the cortex was apparently less convoluted.,8313
"rapid a it that, Despite proliferated at pace.","Despite that, it proliferated at a rapid pace.",8314
running Endurance sapiens. Homo the to led of emergence,Endurance running led to the emergence of Homo sapiens.,8315
"such searched, minimally that an its filter gaussian. inverse Hence, is output is","Hence, an inverse filter is searched, such that its output is minimally gaussian.",8316
of signal. the use We the non-gaussianity measure the of a as kurtosis,We use the kurtosis as a measure of the non-gaussianity of the signal.,8317
kurtosis maximum deconvolving A the is of filter function coefficients the as of searched. a,A maximum of the kurtosis as a function of the deconvolving filter coefficients is searched.,8318
may signal This applied the original obtain be the to undistorted distorted filter signal. to,This filter may be applied to the distorted signal to obtain the original undistorted signal.,8319
"a framework, its Dirichlet solution for this process. on variational Based provides full the","Based on this framework, its provides a full variational solution for the Dirichlet process.",8320
that computationally to numerical very efficient when results compared the method is show The MCMC.,The numerical results show that the method is very computationally efficient when compared to MCMC.,8321
information matrix The the from ratio. arises test,The test arises from the information matrix ratio.,8322
is and The corresponding derived normality is test proven. asymptotic its statistic,The corresponding test statistic is derived and its asymptotic normality is proven.,8323
process is of circulating mechanism the (CTCs). physical this of tumour biological dissemination cells Preceeding,Preceeding this biological mechanism is the physical process of dissemination of circulating tumour cells (CTCs).,8324
growth The than expected. normally was magnitude of around two-orders larger assumed rate,The assumed growth rate was around two-orders of magnitude larger than normally expected.,8325
"is In lack of script, official there such Devnagari benchmark.","In Devnagari script, there is lack of such official benchmark.",8326
offline format The sample it memory. in occupies TIFF as stored images image are less,The offline sample images are stored in TIFF image format as it occupies less memory.,8327
"data reduced. memory presented level further is binary Also, in so the requirement that is","Also, the data is presented in binary level so that memory requirement is further reduced.",8328
and Background: role vexing Understanding distinction the and function difficult. is between,Background: Understanding the distinction between function and role is vexing and difficult.,8329
them the test against Ontology Investigations practice (OBI). Biomedical I for in,I test them in practice against the Ontology for Biomedical Investigations (OBI).,8330
"discuss in these and definitions Finally, I methods for applying an give practice. axiomatisation","Finally, I give an axiomatisation and discuss methods for applying these definitions in practice.",8331
"are practice. paper current applicable, definitions formalizing Conclusions: this The in","Conclusions: The definitions in this paper are applicable, formalizing current practice.",8332
matrix nonnegative noisy algorithm for factorization present under a separability. We problems numerical (NMF),We present a numerical algorithm for nonnegative matrix factorization (NMF) problems under noisy separability.,8333
"to report experimental results. the Finally, apply document algorithm clustering, and the we","Finally, we apply the algorithm to document clustering, and report the experimental results.",8334
determine A understand challenge major to is molecular how biology in processes phenotypic features.,A major challenge in biology is to understand how molecular processes determine phenotypic features.,8335
We history  . the mathematical induction review automation of the of ,  We review the history of the automation of mathematical induction .,8336
tasks Many inference large over exponentially probabilistic summations sets. involve,Many probabilistic inference tasks involve summations over exponentially large sets.,8337
is Compressed (CS-THz) imaging Sensing imaging a based technique. Terahertz computational,Compressed Sensing based Terahertz imaging (CS-THz) is a computational imaging technique.,8338
under CS state-of-the-art conducted systematic CS-THz scan algorithms the architecture. evaluation based We of,We conducted systematic evaluation of state-of-the-art CS algorithms under the scan based CS-THz architecture.,8339
We this a approach tackling problem. principled therefore propose for,We therefore propose a principled approach for tackling this problem.,8340
"this for a propose approximately optimization greedy problem. dynamic, intractable solving We algorithm","We propose a dynamic, greedy algorithm for approximately solving this intractable optimization problem.",8341
Bayesian estimating the model white from matter propose atlas HYDIs. for a then We,We then propose a Bayesian model for estimating the white matter atlas from HYDIs.,8342
We work al. in Hosseinbor the adopt et given,We adopt the work given in Hosseinbor et al.,8343
stochastic volatility a and apply specifications. We range the method univariate of for bivariate,We apply the method for a range of univariate and bivariate stochastic volatility specifications.,8344
find networks. for optimal Bayesian to provably guarantee networks learning algorithms Exact,Exact algorithms for learning Bayesian networks guarantee to find provably optimal networks.,8345
"may in limited due fail difficult or they to time However, tasks memory. learning","However, they may fail in difficult learning tasks due to limited time or memory.",8346
to heuristic research algorithms we networks. Bayesian anytime learn search-based adapt this several In,In this research we adapt several anytime heuristic search-based algorithms to learn Bayesian networks.,8347
fair optimization Multiobjective Decision Markov This (MOMDPs). in paper devoted Processes to is,This paper is devoted to fair optimization in Multiobjective Markov Decision Processes (MOMDPs).,8348
"policies In setting, this study to of tradeoffs. we determination Lorenz-non-dominated the leading","In this setting, we study the determination of policies leading to Lorenz-non-dominated tradeoffs.",8349
of subsets polynomial-sized are approximations those The solutions.,The approximations are polynomial-sized subsets of those solutions.,8350
"learning. approach Our information-driven intrinsic task-dependent support to task-independent, use motivation(s) is to","Our approach is to use task-independent, information-driven intrinsic motivation(s) to support task-dependent learning.",8351
"examine protein in Drosophila pattern the embryo, Here Bicoid reproducible we the gradient. the earliest","Here we examine the earliest reproducible pattern in the Drosophila embryo, the Bicoid protein gradient.",8352
We consider the from causal distribution. directed graphs of learning joint an observational problem acyclic,We consider the problem of learning causal directed acyclic graphs from an observational joint distribution.,8353
correct an is prove evaluation. empirical the We RESIT and population provide setting in that,We prove that RESIT is correct in the population setting and provide an empirical evaluation.,8354
goalkeeping and job. a an (GK) is soccer is expert professional Goalkeeper complete in,Goalkeeper (GK) is an expert in soccer and goalkeeping is a complete professional job.,8355
"seems without reliable In a success impossible achieving fact, GK.","In fact, achieving success seems impossible without a reliable GK.",8356
than dominant is effect successes other failures more and in players. His,His effect in successes and failures is more dominant than other players.,8357
game are mistakes those goalkeeper's. of a visible in The most,The most visible mistakes in a game are those of goalkeeper's.,8358
goalkeepers' done find soccer. the researches indexes Previously in used to are,Previously done researches are used to find the goalkeepers' indexes in soccer.,8359
some qualifications. have Soccer have experts a GK should found successful that,Soccer experts have found that a successful GK should have some qualifications.,8360
"Here consider such class stratified models. we termed an models, of as additional graphical","Here we consider an additional class of such models, termed as stratified graphical models.",8361
Models Random discusses Utility {General (GRUMs)}. paper This,This paper discusses {General Random Utility Models (GRUMs)}.,8362
for algorithm general MAP a couple with GRUMs. inference this (MC-EM) Monte-Carlo-Expectation-Maximization We under based,We couple this with a general Monte-Carlo-Expectation-Maximization (MC-EM) based algorithm for MAP inference under GRUMs.,8363
for of also functions of the prove GRUMs. likelihood a We uni-modality class,We also prove uni-modality of the likelihood functions for a class of GRUMs.,8364
statistical emulating computationally are Gaussian expensive surrogates popular models for (GP) computer simulators. Process used,Gaussian Process (GP) models are popular statistical surrogates used for emulating computationally expensive computer simulators.,8365
the methods optimizing Previous function likelihood (e.g. for,Previous methods for optimizing the likelihood function (e.g.,8366
"package, method GPMfit. proposed Matlab implemented a is The in","The proposed method is implemented in a Matlab package, GPMfit.",8367
based on characterized in We regulatory their studies kinetic relationships hormones. different nine,We characterized their regulatory relationships in nine kinetic studies based on different hormones.,8368
in main two that network the in identified We guard hubs cells. transport nitrate,We identified two main hubs in the network that transport nitrate in guard cells.,8369
transport a that is This sunflower nitrate physiological drought. of suggests critical response aspect to,This suggests that nitrate transport is a critical aspect of sunflower physiological response to drought.,8370
on inferences concept computational complexity irrelevance/independence adopted. the such The on of models depends,The computational complexity of inferences on such models depends on the irrelevance/independence concept adopted.,8371
independence inferences are trees ternary that show under We even strong in NP-hard variables. with,We show that inferences under strong independence are NP-hard even in trees with ternary variables.,8372
"a text its problem. challenging value, images research general remains in Despite identifying however,","Despite its value, however, identifying general text in images remains a challenging research problem.",8373
on benchmark proposed The datasets. a detection detector state-of-the-art text methods outperforms scene few text,The proposed text detector outperforms state-of-the-art methods on a few benchmark scene text detection datasets.,8374
probability distributions provide for language over compact preference orderings. a representing PCP-nets,PCP-nets provide a compact language for representing probability distributions over preference orderings.,8375
are We useful that modelling argue noisy they aggregating preferences preferences. for or,We argue that they are useful for aggregating preferences or modelling noisy preferences.,8376
Graphical Potentials received Order considerable have interest with High recent (HOPs) years. models in,Graphical models with High Order Potentials (HOPs) have received considerable interest in recent years.,8377
of learning discrete the networks complete problem We Bayesian (BNs) consider from data.,We consider the problem of learning Bayesian networks (BNs) from complete discrete data.,8378
of as discrete (IP). an is integer problem formulated This program optimisation,This problem of discrete optimisation is formulated as an integer program (IP).,8379
the we of this allow have describe to IP. We steps efficient solving various taken,We describe the various steps we have taken to allow efficient solving of this IP.,8380
"These sometimes earlier over results. show dramatic, improvements,","These show improvements, sometimes dramatic, over earlier results.",8381
bias-variance interesting introduce algorithms our that show several new We trade-offs.,We show that our new algorithms introduce several interesting bias-variance trade-offs.,8382
both the for (feedback directed latent procedure allows of and cycles variables. presence loops) The,The procedure allows for both directed cycles (feedback loops) and the presence of latent variables.,8383
the effect how and algorithm of assumptions the Simulations scales. illustrate discovery present these on,Simulations illustrate the effect of these assumptions on discovery and how the present algorithm scales.,8384
Bayesian networks. learns PC oriented maximally The algorithm causal,The PC algorithm learns maximally oriented causal Bayesian networks.,8385
(RCD) that causal algorithm discovery relational causal learns We relational present models. the,We present the relational causal discovery (RCD) algorithm that learns causal relational models.,8386
is applications. of pivotal Reliable importance tracking surveillance in,Reliable tracking is of pivotal importance in surveillance applications.,8387
is allow cases not z-identifiability to where first from generalize necessarily Z disjoint X. We,We first generalize z-identifiability to allow cases where Z is not necessarily disjoint from X.,8388
for complete also z-transportability. that show do-calculus is Our results,Our results also show that do-calculus is complete for z-transportability.,8389
"we from CRAN available R implement package an have freely MPBART, To mpbart repositories. developed","To implement MPBART, we have developed an R package mpbart available freely from CRAN repositories.",8390
Browser enhancer as Genome developmental UCSC freely predictions be a Our genome-wide will available track.,Our genome-wide developmental enhancer predictions will be freely available as a UCSC Genome Browser track.,8391
effect is the change immediate rivers. The of lakes size and of,The immediate effect is the change of size of lakes and rivers.,8392
happen ignored. may their are Misclassification contents since color,Misclassification may happen since their color contents are ignored.,8393
"implemented. colored achieve kinds of different this To and purpose, are descriptors SIFT introduced","To achieve this purpose, different kinds of colored SIFT descriptors are introduced and implemented.",8394
several The are real benchmarks. on out experiments carried,The real experiments are carried out on several benchmarks.,8395
in patterns embedded stochastic capture The proposes data. modeling second part to the the better,The second part proposes stochastic modeling to better capture the patterns embedded in the data.,8396
"However, laws for systems structure. fail complex these network with","However, these laws fail for systems with complex network structure.",8397
in Here show cases that superellipses. the these are we eigenvalues by described,Here we show that in these cases the eigenvalues are described by superellipses.,8398
complex develop also a We dominant of method eigenvalue the estimate new networks. to analytically,We also develop a new method to analytically estimate the dominant eigenvalue of complex networks.,8399
Each solutions presentations. oral group complement presented comprehensive to,Each group presented comprehensive solutions to complement oral presentations.,8400
forms techniques. modern amenable mathematical are to programming Both,Both forms are amenable to modern mathematical programming techniques.,8401
Differences by were motif (HsHop) identified PfHop Hop human between analysis. and,Differences between human Hop (HsHop) and PfHop were identified by motif analysis.,8402
in problem not relatively and is literature. new well-explored This the,This problem is relatively new and not well-explored in the literature.,8403
"In features, i.e. to extract propose we contrast, cross-image","In contrast, we propose to extract cross-image features, i.e.",8404
"weight the discriminative we Consequently, most to select use AdaBoost features. and","Consequently, we use AdaBoost to select and weight the most discriminative features.",8405
a mathematical groups. formulate of resources different of for model We between age competition representatives,We formulate a mathematical model of competition for resources between representatives of different age groups.,8406
done. numerical simulation initial the of has for evolution of distributions resources different A,A numerical simulation of the evolution of resources for different initial distributions has done.,8407
around interested always world. Human identification has a researchers the that been topic,Human identification has always been a topic that interested researchers around the world.,8408
"performed sensors. cameras, scanners can Data or using acquisition be","Data acquisition can be performed using cameras, scanners or sensors.",8409
compared for segmenting test set ear Level few efficiency their A in algorithms images. are,A few Level set algorithms are compared for their efficiency in segmenting test ear images.,8410
"unrepresented much comparative ecdysozoan remains in diversity analyses. However, genomic of","However, much of ecdysozoan diversity remains unrepresented in comparative genomic analyses.",8411
is The detectors clustering. alleviated among redundancy k-medoids using,The redundancy among detectors is alleviated using k-medoids clustering.,8412
an result show the We for random general also dot model. analogous graph product more,We also show an analogous result for the more general random dot product graph model.,8413
introduced motion internal sites. interactions to to the and hide its target binding We,We introduced motion and internal interactions to the target to hide its binding sites.,8414
binding access the a site. the IDP intermediate The provides for state scaffold hidden to,The intermediate state provides a scaffold for the IDP to access the hidden binding site.,8415
network data. the from a learning investigate problem We of the structure of Markov,We investigate the problem of learning the structure of a Markov network from data.,8416
recognition. This action optical for (STA) appearance flow descriptors proposes with combining paper human spatio-temporal,This paper proposes combining spatio-temporal appearance (STA) descriptors with optical flow for human action recognition.,8417
of The recognition. in our encouraging suggest potential action human the experimental results online approach,The encouraging experimental results suggest the potential of our approach in online human action recognition.,8418
for It data. is a complex proposed multidimensional channel encoding valued,It is proposed a complex valued channel encoding for multidimensional data.,8419
contains basic complex nonlinear of The mappings. overlapping approach,The basic approach contains overlapping of complex nonlinear mappings.,8420
work. be examined possibility will in This future,This possibility will be examined in future work.,8421
pattern A end-user transparent simplifies flowchart. the proposed the is to and,A proposed pattern is transparent to the end-user and simplifies the flowchart.,8422
is describe of practical pattern. the We solution implementation a proposed a developed that software,We describe a developed software solution that is a practical implementation of the proposed pattern.,8423
real-life components. example an We of electric give use industrial in production its of,We give an example of its real-life use in industrial production of electric components.,8424
These amoeba are based et al. introduced so-called Lerallut by concept on a morphological filters,These so-called morphological amoeba filters are based on a concept introduced by Lerallut et al.,8425
results the as similar and PDEs. self-snakes geodesic achieve active contour They well-known,They achieve similar results as the well-known geodesic active contour and self-snakes PDEs.,8426
geodesic contour identical PDE similar but is the structurally not active This to equation.,This PDE is structurally similar but not identical to the geodesic active contour equation.,8427
special reproduces previous for the cases. filters results PDE as amoeba approximation It median,It reproduces the previous PDE approximation results for amoeba median filters as special cases.,8428
observers synchronization on develop contraction Nonlinear to these perform and theory used systems. analysis is,Nonlinear contraction theory is used to develop observers and perform synchronization analysis on these systems.,8429
by other or state are estimated partial observers. state using All full variables,All other variables are estimated by using partial state or full state observers.,8430
effects. explaining provide We these intuitions some,We provide some intuitions explaining these effects.,8431
is constrained multi-project notoriously combinatorial problem scheduling The difficult optimization a problem. multi-mode (RCMMMPSP) resource,The resource constrained multi-mode multi-project scheduling problem (RCMMMPSP) is a notoriously difficult combinatorial optimization problem.,8432
goal given a set the solution under It's restrictions. of was time instances running of,It's goal was the solution of a given set of instances under running time restrictions.,8433
with here to have the this contributed presented We challenge approach.,We have contributed to this challenge with the here presented approach.,8434
is The groundtruth a by GPS consumer IMU. without device grade supplied,The groundtruth is supplied by a consumer grade GPS device without IMU.,8435
Synchronization readings between frames are after acquisition. and recovered GPS stereo alignment the and,Synchronization and alignment between GPS readings and stereo frames are recovered after the acquisition.,8436
the allows practice. conclusions draw that groundtruth in accuracy useful to We show attained,We show that the attained groundtruth accuracy allows to draw useful conclusions in practice.,8437
a local on Therefore results. method RSR illumination we estimation estimating global based for propose,Therefore we propose a method for estimating global illumination estimation based on local RSR results.,8438
we database a publicly our constancy available tests color testing. use image for For,For our tests we use a publicly available color constancy image database for testing.,8439
"replacing on framework i.e. keypoints This by is based keygraphs,","This framework is based on replacing keypoints by keygraphs, i.e.",8440
"isomorph explore order whose structural to are relative information. graphs and vertices in directed keypoints,","isomorph directed graphs whose vertices are keypoints, in order to explore relative and structural information.",8441
"issue. time to we keytuples solve novel of introduce particular, the a In running concept","In particular, we introduce the novel concept of keytuples to solve a running time issue.",8442
speed is tracking real-time with The two cameras different illustrated by in hardware. ordinary,The speed is illustrated by real-time tracking with two different cameras in ordinary hardware.,8443
Knowing important in density the both diagnostic breast is applications. and detection computer-aided,Knowing the breast density is important both in diagnostic and computer-aided detection applications.,8444
"the dynamical this with In on we filtering focus models paper, in jumps. problem statistical","In this paper, we focus on the statistical filtering problem in dynamical models with jumps.",8445
increased of The recent vehicles road decades. significantly number in,The number of road vehicles significantly increased in recent decades.,8446
of road the infrastructure used Cameras or as part are vehicles. in,Cameras are used as part of the road infrastructure or in vehicles.,8447
"Generally, with (CSD) cropping period increased system local decreasing disturbances latitude. the","Generally, the local cropping system disturbances (CSD) period increased with decreasing latitude.",8448
the of functionality to These and the improve power contribute results simulated computational networks.,These simulated results contribute to improve the computational power and functionality of the networks.,8449
energy Potts function applications. computer vision problem frequently in the minimizing of occurs The,The problem of minimizing the Potts energy function frequently occurs in computer vision applications.,8450
"e.g. significant in The some applications, pixels can be number ""labeled"" of","The number of ""labeled"" pixels can be significant in some applications, e.g.",8451
"technique, al. of et generalize the derive we To algorithm our Felzenszwalb","To derive our technique, we generalize the algorithm of Felzenszwalb et al.",8452
"a compare numerical particular, In related results by with we simulations. of our means model","In particular, we compare our results with a related model by means of numerical simulations.",8453
literature. discussed in dipole the assumption satisfies Our the neuroscientific widespread work consistently,Our work consistently satisfies the widespread dipole assumption discussed in the neuroscientific literature.,8454
"(MSE, three MSDM). error using Hausdorff, metrics is Compression measured","Compression error is measured using three metrics (MSE, Hausdorff, MSDM).",8455
use visual learning propose pipeline for feature Here we tracking. to the,Here we propose to use the feature learning pipeline for visual tracking.,8456
is the resulted framework using Tracking implemented and effective. yet simple is very tracking-by-detection,Tracking is implemented using tracking-by-detection and the resulted framework is very simple yet effective.,8457
"on classifier a Finally, features. linear trained these simple is","Finally, a simple linear classifier is trained on these features.",8458
al. et into it plugging the by feature flexibility Hare also We of demonstrate learning,We also demonstrate the flexibility of feature learning by plugging it into Hare et al.,8459
"estimation accuracy error the Thus, critical. of is","Thus, the accuracy of error estimation is critical.",8460
their finite-sample these approximations analytic accuracy we From and via examples. produce demonstrate numerical,From these we produce analytic finite-sample approximations and demonstrate their accuracy via numerical examples.,8461
figures. and Material derivations equations contains added for some The Supplementary,The Supplementary Material contains derivations for some equations and added figures.,8462
is an biology. structure between currently in sequence Mapping problem open and structural,Mapping between sequence and structure is currently an open problem in structural biology.,8463
its sequence. this fold given a protein predicting the pave Answering for may question way,Answering this question may pave the way for predicting a protein fold given its sequence.,8464
doctoral focused My particular to relationship. relevant studies the protein sequence-structure phenomenon on a have,My doctoral studies have focused on a particular phenomenon relevant to the protein sequence-structure relationship.,8465
"to to robust easy outliers, is high-dimensional and to variables. appropriate It implement, powerful","It is robust to outliers, easy to implement, powerful and appropriate to high-dimensional variables.",8466
These important are properties in applications. many,These properties are important in many applications.,8467
"de well-known theorems, and of theorem theorem. such A number Cox's Finetti's as","A number of well-known theorems, such as Cox's theorem and de Finetti's theorem.",8468
"change directional such However, is inevitable. not","However, such directional change is not inevitable.",8469
been neural recently for success deep has networks reported Significant using classification.,Significant success has been reported recently using deep neural networks for classification.,8470
"computationally networks large be Such even over. after training can intensive, is","Such large networks can be computationally intensive, even after training is over.",8471
Humor and. proceedings of We summarize Workshop on the the,We summarize the proceedings of the Workshop on Humor and.,8472
"The slippage is of humor. aggression-based principal and defined humor humor, contrasted with considered, type","The principal type of humor considered, slippage humor, is defined and contrasted with aggression-based humor.",8473
"frame of presented. of humor, Hofstadter's blend, on slippage A variety based a is notion","A variety of slippage humor, based on Hofstadter's notion of a frame blend, is presented.",8474
produce humor. of two varying can of frames blending Diverse ways types degrees and,Diverse ways of blending two frames can produce varying degrees and types of humor.,8475
"blends. This amongst the and defective highlights relationships frame analogies, good jokes,","This highlights the relationships amongst good jokes, defective analogies, and frame blends.",8476
"is presented. value humor of the Finally, a adaptive about theory speculative","Finally, a speculative theory about the adaptive value of humor is presented.",8477
algebras languages. knowledge used spatial Qualitative temporal as representation and are,Qualitative algebras are used as spatial and temporal knowledge representation languages.,8478
given. the domain cooking A from temporal example is,A temporal example from the cooking domain is given.,8479
"iterative this with robust for method discussed. deconvolution an constraints paper, In positivity is","In this paper, an iterative method for robust deconvolution with positivity constraints is discussed.",8480
demonstrate synthetic of on proposed method. data and the image Experiments performance the real-world,Experiments on synthetic and real-world image data demonstrate the performance of the proposed method.,8481
the shape studied an representation of essential is skeleton compact shape. providing The a characteristic,The skeleton is an essential shape characteristic providing a compact representation of the studied shape.,8482
grid many on the raises computation Its issues. image,Its computation on the image grid raises many issues.,8483
make to shape to matching a is goal Our recognition. for skeleton dedicated,Our goal is to make a skeleton dedicated to shape matching for recognition.,8484
"a obtain the After thinning skeleton. pruning and we process,","After a thinning and pruning process, we obtain the skeleton.",8485
to finally compared The methods. method proposed recent is fairly,The proposed method is finally compared to fairly recent methods.,8486
"fossils. FBD for Moreover, model inclusion all of available allows the","Moreover, the FBD model allows for inclusion of all available fossils.",8487
IQRray . Availability: is at: ftp://lausanne.isb-sib.ch/pub/databases/Bgee/general/IQRray.R The R of available implementation,Availability: The R implementation of IQRray is available at: ftp://lausanne.isb-sib.ch/pub/databases/Bgee/general/IQRray.R .,8488
We of synchronization adaptation investigated how tendency coupled neurons. affect the currents model these,We investigated how these adaptation currents affect the synchronization tendency of coupled model neurons.,8489
Ecology variety complexity. its in and biodiversity studies,Ecology studies biodiversity in its variety and complexity.,8490
describes perform distribute how in species changes. to and It response environmental,It describes how species distribute and perform in response to environmental changes.,8491
adaptive. and Ecological are and complex structures highly processes,Ecological processes and structures are highly complex and adaptive.,8492
apparent ecology. patterns becomes dealing with especially scaling This when in,This becomes especially apparent when dealing with scaling patterns in ecology.,8493
"ecological all scale of are Indeed, dependent. nearly patterns","Indeed, nearly all of ecological patterns are scale dependent.",8494
power predictive problems and inference. hampers Such in creates our scale our dependence,Such scale dependence hampers our predictive power and creates problems in our inference.,8495
case the These aspects of two further learning-related in are activity. illustrated brain,These two aspects are further illustrated in the case of learning-related brain activity.,8496
"Finally, experimental methods are data a reviewed. to from extract temperature","Finally, methods to extract a temperature from experimental data are reviewed.",8497
and (aa) Protein on depends amino function protein structure acid both sequence.,Protein function depends on both protein structure and amino acid (aa) sequence.,8498
predicted. center edge Some of fragments features and are the,Some edge and center features of the fragments are predicted.,8499
been research focusing this topic. There much on work has,There has been much research work focusing on this topic.,8500
"peers knowledge a is and beliefs jor ma- challenge. settings, autonomous Combining in distributed of","Combining knowledge and beliefs of autonomous peers in distributed settings, is a ma- jor challenge.",8501
of within the Ontologies SHIQ are Description Logics. fragment,Ontologies are within the SHIQ fragment of Description Logics.,8502
"satisfiability-preserving are clause often elimination procedures used. Additionally,","Additionally, satisfiability-preserving clause elimination procedures are often used.",8503
"is to this i.e. setting preprocessing in has requirement be key sound, The that the","The key requirement in this setting is that the preprocessing has to be sound, i.e.",8504
to this MaxSAT. paper show these we how preprocessing adapt to In techniques,In this paper we show how to adapt these preprocessing techniques to MaxSAT.,8505
the data. eigenspace elliptical covariance the the estimates of of ECA matrix,ECA estimates the eigenspace of the covariance matrix of the elliptical data.,8506
"is exploited. statistic To rank heavy-tailed multivariate a cope distributions, elliptical with","To cope with heavy-tailed elliptical distributions, a multivariate rank statistic is exploited.",8507
"Methodologically, non-sparse ECA procedures settings. we sparse both and for propose","Methodologically, we propose ECA procedures for both non-sparse and sparse settings.",8508
"theoretical non-asymptotic Theoretically, analyses quantifying of provide both the asymptotic we and ECA. performances","Theoretically, we provide both non-asymptotic and asymptotic analyses quantifying the theoretical performances of ECA.",8509
vibration pattern particular to creates Misfiring unique attributed cylinder. a a,Misfiring creates a unique vibration pattern attributed to a particular cylinder.,8510
patterns to and extracted be analyzed these can misfire. detect be from Useful can features,Useful features can be extracted from these patterns and can be analyzed to detect misfire.,8511
vibration were features Statistical these from extracted. signals,Statistical features from these vibration signals were extracted.,8512
algorithm presented. In analysis is of Kstar paper performance this,In this paper performance analysis of Kstar algorithm is presented.,8513
prostate for organ of gland. delineation) central the limits,delineation) of organ limits for the prostate central gland.,8514
on segmentation spherical The a algorithm used template. automated graph-driven is based,The automated segmentation algorithm used is graph-driven based on a spherical template.,8515
classrooms. can incorporate creativity It in challenging to be,It can be challenging to incorporate creativity in classrooms.,8516
this leaves general. the whether of question holds open in interesting This,This leaves open the interesting question of whether this holds in general.,8517
diallelic precisely analyze combine modes loci. into at how We IBD distributions identical different genotype,We analyze precisely how different IBD modes combine into identical genotype distributions at diallelic loci.,8518
an relatedness are characterization Our statistics always exhaustive identifiable. that analysis of yields,Our analysis yields an exhaustive characterization of relatedness statistics that are always identifiable.,8519
alike. ubiquitous is DNA eukaryotes and regulation Large-scale deformation in prokaryotes transcriptional in,Large-scale DNA deformation is ubiquitous in transcriptional regulation in prokaryotes and eukaryotes alike.,8520
"DWSVM solution developed. theoretical asymptotic normality of properties, Several Fisher the consistency including are and","Several theoretical properties, including Fisher consistency and asymptotic normality of the DWSVM solution are developed.",8521
establishes approach. data real further usefulness application our A the of,A real data application further establishes the usefulness of our approach.,8522
constrain halo introduce without transmission introducing regularization structure-aware We non-local properly artifacts. estimation the to,We introduce non-local structure-aware regularization to properly constrain transmission estimation without introducing the halo artifacts.,8523
combination system on three the based of is approaches. The,The system is based on the combination of three approaches.,8524
the this semiparametric Cross frequently of version We method efficient yields show that Entropy estimators.,We show that this semiparametric version of the Cross Entropy method frequently yields efficient estimators.,8525
The similarities identify differences DWD. the to FLAME helps between and and family SVM,The FLAME family helps to identify the similarities and differences between SVM and DWD.,8526
classifiers asymptotic are of FLAME the studied. Several properties,Several asymptotic properties of the FLAME classifiers are studied.,8527
of illustrate usefulness applications Simulations to real the the FLAME are classifiers. and data investigated,Simulations and real data applications are investigated to illustrate the usefulness of the FLAME classifiers.,8528
"recombination, quantitative forqs traits, of and forward-in-time is simulation selection. a","forqs is a forward-in-time simulation of recombination, quantitative traits, and selection.",8529
command- program. C++ line is implemented as forqs a,forqs is implemented as a command- line C++ program.,8530
manifold. the is Riemannian based ODF constructed distribution of on then The ODFs probabilistic,The probabilistic distribution of ODFs is then constructed based on the ODF Riemannian manifold.,8531
study. hotspots quantitative for translocation to identification A this crucial was of the method,A quantitative method for the identification of translocation hotspots was crucial to this study.,8532
determine experiments performed Single-factor were food for factor to preferred each the location.,Single-factor experiments for each factor were performed to determine the preferred food location.,8533
a is such date established possible To approaches. not distinction with,To date such a distinction is not possible with established approaches.,8534
"principal one-shot a recognition. paper motion new introduces components method (PMC), gesture This for","This paper introduces principal motion components (PMC), a new method for one-shot gesture recognition.",8535
PMC with set. same with competitive proposed the data methods alternative are obtained Results for,Results obtained with PMC are competitive with alternative methods proposed for the same data set.,8536
"shapes into has given Lapalce-Beltrami Recently, embedding to been of attention operator. the the eigenspace","Recently, attention has been given to embedding shapes into the eigenspace of the Lapalce-Beltrami operator.",8537
"under the transformations. isometric invariant preserves Laplace-Beltrami diffusion eigenspace and is distance, The","The Laplace-Beltrami eigenspace preserves the diffusion distance, and is invariant under isometric transformations.",8538
"However, incompatible eigenfunctions independently other. Laplace-Beltrami different each for computed are with often shapes","However, Laplace-Beltrami eigenfunctions computed independently for different shapes are often incompatible with each other.",8539
"approach we introduce Here, for eigenfunctions. statistical matching a","Here, we introduce a statistical approach for matching eigenfunctions.",8540
"statistics. order using indirectly, Comparing high done is distributions","Comparing distributions is done indirectly, using high order statistics.",8541
proposed feature the We the present experiments of applied point method correspondence. success demonstrating to,We present experiments demonstrating the success of the proposed method applied to feature point correspondence.,8542
"pixels. full of locations from images, learned be from a random Sensor may subsample or","Sensor locations may be learned from full images, or from a random subsample of pixels.",8543
"and origin, claims. its discuss We narrative","We discuss its origin, narrative and claims.",8544
unacceptable. scientifically We why this explain is concept,We explain why this concept is scientifically unacceptable.,8545
work the extend by al. Bolic original et We,We extend the original work by Bolic et al.,8546
during to processes oscillate Many the cell continue intracellular cycle.,Many intracellular processes continue to oscillate during the cell cycle.,8547
barycenters propose build compute two that subgradient method. to original Wasserstein the upon We algorithms,We propose two original algorithms to compute Wasserstein barycenters that build upon the subgradient method.,8548
is that in enzyme an the DNA participates process Polymerase DNA of replication.,DNA Polymerase is an enzyme that participates in the process of DNA replication.,8549
in  many of all areas the across ancestor  China. Muslims Ajjal Sayyid is,  Sayyid Ajjal is the ancestor of many Muslims in areas all across China.,8550
the evidence concepts dynamics. fundamental Empirical questions of population human,Empirical evidence questions fundamental concepts of the human population dynamics.,8551
implementation between allows logic of for functions such gene interdependency the An in interactions regulation.,An interdependency between such interactions allows for the implementation of logic functions in gene regulation.,8552
behavior emergence of of a long-range demonstrate We correlation power-law this the linear function.,We demonstrate the emergence of a long-range power-law behavior of this linear correlation function.,8553
elucidate can model these conditions invasion two-locus processes. Understanding a the in for,Understanding the conditions for invasion in a two-locus model can elucidate these processes.,8554
recognition method a based functions. statistical paper Numerical presents Devnagari This discriminant on,This paper presents a Devnagari Numerical recognition method based on statistical discriminant functions.,8555
"Linear, used Quadratic, Diagquadratic for are Mahalanobis Diaglinear, and distance classification.","Linear, Quadratic, Diaglinear, Diagquadratic and Mahalanobis distance are used for classification.",8556
"Mahalanobis discriminant Quadratic results. better and that Experimental Linear, provide functions show results","Experimental results show that Linear, Quadratic and Mahalanobis discriminant functions provide better results.",8557
these classifier. are Combination fed three voting to Results majority type a of Discriminants,Results of these three Discriminants are fed to a majority voting type Combination classifier.,8558
over Combination It that classifier classifiers. individual offers is results found better,It is found that Combination classifier offers better results over individual classifiers.,8559
kernels. then and translation-invariant on weighted We polynomial focus,We then focus on translation-invariant and weighted polynomial kernels.,8560
objects that human capture attention aims object locate images. to detection within Salient,Salient object detection aims to locate objects that capture human attention within images.,8561
Previous approaches problem pose often image of as a contrast analysis. this,Previous approaches often pose this as a problem of image contrast analysis.,8562
knockdowns. identified whose was then by expression genes We affected the,We then identified genes whose expression was affected by the knockdowns.,8563
TF infer us This allowed combination to functional data binding. of,This combination of data allowed us to infer functional TF binding.,8564
resolution. technological limited restrictions sensors has a hyperspectral of Because spatial,Because of technological restrictions hyperspectral sensors has a limited spatial resolution.,8565
spatial panchromatic the better other image On resolution. has hand a,On the other hand panchromatic image has a better spatial resolution.,8566
understanding target provide of better can together the information this scene. Combining a,Combining this information together can provide a better understanding of the target scene.,8567
method AVRIS Pines simulated been The has on Indian applied using data datasets. proposed,The proposed method has been applied on simulated data using AVRIS Indian Pines datasets.,8568
that in information hyperspectral method this can and panchromatic combine Results images. effectively show,Results show that this method can effectively combine information in hyperspectral and panchromatic images.,8569
color show natural grayscale similar Hyperspectral photographic images to properties or statistical images.,Hyperspectral images show similar statistical properties to natural grayscale or color photographic images.,8570
"processing lead manifolds. particular These peculiarities indetermination signal by problems, mainly characterized and complex to","These peculiarities lead to particular signal processing problems, mainly characterized by indetermination and complex manifolds.",8571
of gained statistical The framework learning last decade. popularity has the in,The framework of statistical learning has gained popularity in the last decade.,8572
tutuorial main illustrative through This hyperspectral classification image sensing advances the remote examples. for reviews,This tutuorial reviews the main advances for hyperspectral remote sensing image classification through illustrative examples.,8573
differ the $v$. size methods the The by of of neighborhood two,The two methods differ by the size of the neighborhood of $v$.,8574
conducted experiments the approach. been the Various effectiveness have evaluate to of proposed,Various experiments have been conducted to evaluate the effectiveness of the proposed approach.,8575
the to work common a PCBs. This detect proposes most defects novel in approach,This work proposes a novel approach to detect most common defects in the PCBs.,8576
analysis. quantitative world has A board used for been real circuit printed dataset,A real world printed circuit board dataset has been used for quantitative analysis.,8577
the demonstrated method. proposed results of the efficacy Experimental,Experimental results demonstrated the efficacy of the proposed method.,8578
"can to bundles DTI fiber determined, gain about structures. eloquent be information Using data, brain","Using DTI data, fiber bundles can be determined, to gain information about eloquent brain structures.",8579
"the fiber bundle has to determined. boundary Therefore, be","Therefore, the fiber bundle boundary has to be determined.",8580
"of approach is estimation tubular a ray-based novel this In paper, boundary presented. for structures","In this paper, a novel ray-based approach for boundary estimation of tubular structures is presented.",8581
develop organism integration of elegans. the We neuro-sensory Caenorhabditis biophysical model model a in,We develop a biophysical model of neuro-sensory integration in the model organism Caenorhabditis elegans.,8582
"responses critical as crawling. and an swimming are bifurcation, in such induced explaining behavioral Hopf","an induced Hopf bifurcation, are critical in explaining behavioral responses such as swimming and crawling.",8583
a Scientific large of There Historical and importance. of documents is Handwritten English paper collection,There is a large collection of Handwritten English paper documents of Historical and Scientific importance.,8584
by directly computer. not documents But recognized are paper,But paper documents are not recognized directly by computer.,8585
digital document storing Hence closest the their documents of is by these image. indexing way,Hence the closest way of indexing these documents is by storing their document digital image.,8586
a replace images large can document database of documents. Hence the paper,Hence a large database of document images can replace the paper documents.,8587
of shape on code then size word query. based and based character on First,First based on word size and then based on character shape code of query.,8588
makes efficient and the It multiple pre-processing. and faster more reduces need of the process,It makes the process faster and more efficient and reduces the need of multiple pre-processing.,8589
novel paper. this methodology is presented gender in classification for A,A novel methodology for gender classification is presented in this paper.,8590
difference. face local It color intensity feature extracts from using of region a gray,It extracts feature from local region of a face using gray color intensity difference.,8591
obtained FERET very It is high collected from accuracy. on database evaluated the and images,It is evaluated on the images collected from FERET database and obtained very high accuracy.,8592
of Acquiring genomes study many microbiota. in at applications has single-cell resolution the such as,Acquiring genomes at single-cell resolution has many applications such as in the study of microbiota.,8593
sparse in sense. Biologically important are often samples that,Biologically important samples are often sparse in that sense.,8594
of in vision. research salient is Automated computer regions area visually an active of detection,Automated detection of visually salient regions is an active area of research in computer vision.,8595
We different achieving this. two propose for approaches,We propose two different approaches for achieving this.,8596
progress computational the networks. in spiking capabilities has of uncovering neural made Much been,Much progress has been made in uncovering the computational capabilities of spiking neural networks.,8597
"problem. analysis, a this classic is stiff numerical In","In numerical analysis, this is a classic stiff problem.",8598
neurons study much difficult to analytically. Spiking more also are,Spiking neurons are also much more difficult to study analytically.,8599
theory of system a gives The field mean the rate-like description.,The mean field theory of the system gives a rate-like description.,8600
expansion keep first The in the order of perturbation terms track covariances.,The first order terms in the perturbation expansion keep track of covariances.,8601
this Kernel apply work this to (MKL) Learning we problem. Multiple In,In this work we apply Multiple Kernel Learning (MKL) to this problem.,8602
"is be estimation results actually distorted the could if However, assumption either violated.","However, the estimation results could be distorted if either assumption is actually violated.",8603
using real-world artificial demonstrate our data. the effectiveness and We of method,We demonstrate the effectiveness of our method using artificial and real-world data.,8604
"is commonly from In impostor scores. similarity uniqueness facial inferred Biometrics,","In Biometrics, facial uniqueness is commonly inferred from impostor similarity scores.",8605
the is retina the retinopathy. of symptom in The earliest hemorrhages diabetic of presence,The presence of hemorrhages in the retina is the earliest symptom of diabetic retinopathy.,8606
of number of and shape disease. indicate severity is the the The to used hemorrhages,The number and shape of hemorrhages is used to indicate the severity of the disease.,8607
hemorrhage reduce blindness. automated of can help Early detection incidence the,Early automated hemorrhage detection can help reduce the incidence of blindness.,8608
separates method and The the recognizes proposed all DH.,The proposed method recognizes and separates all the DH.,8609
"a of suggest notion model. we minimal a Secondly, complexity","Secondly, we suggest a notion of a minimal complexity model.",8610
"we the propose strategy a for composing models Thirdly, together. reduced","Thirdly, we propose a strategy for composing the reduced models together.",8611
"advantageous the can Connection detailed applications. preserved, with be some is model which in","Connection with the detailed model is preserved, which can be advantageous in some applications.",8612
involve phenomena cell physiological migration.  directional  Many,  Many physiological phenomena involve directional cell migration.,8613
chemical It gradients vivo. to attributed in is usually,It is usually attributed to chemical gradients in vivo.,8614
We concept associated protrusion and direction of an efficient index. identified the,We identified the concept of efficient protrusion and an associated direction index.,8615
patterns. external the the changing varied by cues adhesive We,We varied the external cues by changing the adhesive patterns.,8616
"using We drug the activity. also internal modified modified the protrusion treatments, cues which","We also modified the internal cues using drug treatments, which modified the protrusion activity.",8617
Stochasticity long-term steps. short- and affects the,Stochasticity affects the short- and long-term steps.,8618
"we document ongoing, biological differentiation, clear introgression. bidirectional Despite","Despite clear biological differentiation, we document ongoing, bidirectional introgression.",8619
"a is issue. and imaging, quite application problem of practical limited-angle In important tomography","In application of tomography imaging, limited-angle problem is a quite practical and important issue.",8620
"two control respectively. algorithm And scaling factors introduces the the processes, to","And the algorithm introduces scaling factors to control the two processes, respectively.",8621
"no model data introgression negative control sets. two detects Further, our in","Further, our model detects no introgression in two negative control data sets.",8622
is and an . a both The challenge... point paper entry present,The present paper is both an entry point and a challenge... .,8623
"of ascertainment also an this phenomenon sampling bias might However, due to SNPs. be nonrandom","However, this phenomenon might be also an ascertainment bias due to nonrandom sampling of SNPs.",8624
microphonic. the electrical be to These cochlear reduced cannot potentials,These electrical potentials cannot be reduced to the cochlear microphonic.,8625
"both important findings and our implications Thus, for have practice. theory","Thus, our findings have important implications for both theory and practice.",8626
"using we approach examples econometrics. Finally, our from illustrate","Finally, we illustrate our approach using examples from econometrics.",8627
Dempster-Shafer under is (D-S theory) uncertain the used widely making theory environment. decision in,Dempster-Shafer theory (D-S theory) is widely used in decision making under the uncertain environment.,8628
open issue. Ranking belief assignments basic an is (BBAs) now,Ranking basic belief assignments (BBAs) now is an open issue.,8629
"evidence new measure issue, address a (RED) To is proposed. ranking distance this","To address this issue, a new ranking evidence distance (RED) measure is proposed.",8630
that claim is unwarranted. show We this,We show that this claim is unwarranted.,8631
Michaelson demonstrate et of model We the that al. also,We also demonstrate that the model of Michaelson et al.,8632
"and studies during e.g. perceptual tasks, of processing interaction social (mu), in and somatomotor","during perceptual and social interaction tasks, and in studies of somatomotor processing (mu), e.g.",8633
coordination computer and brain bimanual in interfaces.,in bimanual coordination and brain computer interfaces.,8634
"theories our and needed. further advance the new understanding concepts are To of brain,","To further advance our understanding of the brain, new concepts and theories are needed.",8635
"and adoption, fate process Cardiac complex, cell differentiation development morphogenesis. multiscale encompassing is","Cardiac development is complex, multiscale process encompassing cell fate adoption, differentiation and morphogenesis.",8636
interactions gene predicted profile The gene prioritized are uniqueness semantic using and metrics. similarity,The predicted gene interactions are prioritized using semantic similarity and gene profile uniqueness metrics.,8637
specific dynamic networks in and context are Biological nature.,Biological networks are context specific and dynamic in nature.,8638
We synthetic using principle demonstrated performance validated the our and and datasets. of method,We demonstrated and validated the principle and performance of our method using synthetic datasets.,8639
"cells. \delta\ controlling islets, \alpha, \beta, homeostasis, and glucose Pancreatic of consist","Pancreatic islets, controlling glucose homeostasis, consist of \alpha, \beta, and \delta\ cells.",8640
"discrete, the We problem consider low lattices. clustering on of data that dimensional reside","We consider the problem of clustering data that reside on discrete, low dimensional lattices.",8641
setting in point extraction. examples image key Canonical segmentation this found are for and,Canonical examples for this setting are found in image segmentation and key point extraction.,8642
We the replace means processing by convolutions. of in costly algorithm steps original,We replace costly processing steps in the original algorithm by means of convolutions.,8643
for and These efficient significantly allow reduce thus highly implementations runtime.,These allow for highly efficient implementations and thus significantly reduce runtime.,8644
signal gap machine therefore learning paper bridges processing. between a and This,This paper therefore bridges a gap between machine learning and signal processing.,8645
"synchrony, of out epochs allows It also sudden forming synchrony. short leaps of hence","It also allows sudden leaps out of synchrony, hence forming short epochs of synchrony.",8646
of Transcription the gene genes most of forms of is regulation expression. of focus,Transcription of genes is the focus of most forms of regulation of gene expression.,8647
"available of a propose Instead, is mechanism which we data. initiation, consistent the with two-step","Instead, we propose a two-step mechanism of initiation, which is consistent with the available data.",8648
model. are Current mainly on focused mixing works the spatial research information that linear consider,Current research works that consider spatial information are mainly focused on the linear mixing model.,8649
"with the real effectiveness of results, synthetic the illustrate Experimental and data, both proposed scheme.","Experimental results, with both synthetic and real data, illustrate the effectiveness of the proposed scheme.",8650
over synthetic experimental data. are run Experiments and,Experiments are run over synthetic and experimental data.,8651
"impact a has dictionary and careful significant consideration. performance, requires on This","This dictionary has a significant impact on performance, and requires careful consideration.",8652
for on analysis the future dictionary This theoretical paves design. KLMS way investigations,This theoretical analysis paves the way for future investigations on KLMS dictionary design.,8653
"compression"". paradox apparent theory ""decompression SP the The resolves by of","The SP theory resolves the apparent paradox of ""decompression by compression"".",8654
for method propose models. jointly based a We kernel all graphical estimating,We propose a kernel based method for jointly estimating all graphical models.,8655
factors. is different many by driven Speciation,Speciation is driven by many different factors.,8656
"and no contains The sympatric. strictly parameters, is model therefore spatial","The model contains no spatial parameters, and is therefore strictly sympatric.",8657
negative selection. The persist organisms frequency-dependent through specialized,The specialized organisms persist through negative frequency-dependent selection.,8658
"a Such on patterns different have for scales, can prominent calling hierarchically structured model. networks","Such networks can have prominent patterns on different scales, calling for a hierarchically structured model.",8659
"effect examined. the estimators covariation is on Monte sample of the Carlo the Additionally,","Additionally, the effect of the sample covariation on the Monte Carlo estimators is examined.",8660
"the phylogeneticists to species. reveal of history most seek Yet,","Yet, most phylogeneticists seek to reveal the history of species.",8661
have jointly species few attempted trees to and infer gene a trees. Only studies,Only a few studies have attempted to jointly infer gene trees and species trees.,8662
"models loss, transfer account incomplete for and duplication These gene or sorting. lineage","These models account for gene duplication and loss, transfer or incomplete lineage sorting.",8663
Current are mean methods based. nonparametric Fr\'echet,Current nonparametric methods are Fr\'echet mean based.,8664
illustrated a inference objects. near-Earth real-world dataset of on is Posterior,Posterior inference is illustrated on a real-world dataset of near-Earth objects.,8665
translucent a We novel materials. BSSRDF for present rendering,We present a novel BSSRDF for rendering translucent materials.,8666
incorporated by Angular previous lacking effects formulation. models using a in BSSRDF dual-beam are,Angular effects lacking in previous BSSRDF models are incorporated by using a dual-beam formulation.,8667
method diffusion employ of We the Placzek's a images theory. of Lemma and discard interpretation,We employ a Placzek's Lemma interpretation of the method of images and discard diffusion theory.,8668
modifications the improve accuracy previous method-of-images of can Our also BSSRDFs.,Our method-of-images modifications can also improve the accuracy of previous BSSRDFs.,8669
"ions. a copper of be with However, only obtained these fraction effects could","However, only a fraction of these effects could be obtained with copper ions.",8670
"showed study nanoparticles. functions macrophage are this altered conclusion, that copper-based significantly by In","In conclusion, this study showed that macrophage functions are significantly altered by copper-based nanoparticles.",8671
An matroids topic excellent to the of and rough to introduction due Zhu is Wang.,An excellent introduction to the topic of rough matroids is due to Zhu and Wang.,8672
"a some sets properties with First, respect of we to the covering. definable investigate","First, we investigate some properties of the definable sets with respect to a covering.",8673
"coverings rough are properties explored. on some matroids of based Finally,","Finally, some properties of rough matroids based on coverings are explored.",8674
"an presented. rough based is equivalent formulation of Moreover, coverings matroids on","Moreover, an equivalent formulation of rough matroids based on coverings is presented.",8675
and many These and results between connections matroids. exhibit rough sets important potential interesting,These interesting and important results exhibit many potential connections between rough sets and matroids.,8676
"naturally our of convergence. mini-batch update afford it and speed can Moreover, method up gives","Moreover, our method can naturally afford mini-batch update and it gives speed up of convergence.",8677
"We that, our show method mild under exponentially. converges assumptions,","We show that, under mild assumptions, our method converges exponentially.",8678
our experiments The numerical actually performs efficiently. that method show,The numerical experiments show that our method actually performs efficiently.,8679
central is genotypes the phenotypes evolution. biological study to The biological and of between mapping,The mapping between biological genotypes and phenotypes is central to the study of biological evolution.,8680
"this method in is reveal However, its ability to sequential limited dependencies.","However, this method is limited in its ability to reveal sequential dependencies.",8681
single-subject raw also We the provide behavioral data.,We also provide the raw single-subject behavioral data.,8682
is to network flow related problems. It closely theory of the,It is closely related to the theory of network flow problems.,8683
as a is The network. measure of an of a connectivity robustness its matroid important,The connectivity of a matroid is an important measure of its robustness as a network.,8684
"under necessary very which Therefore, conditions is the it to investigate connected. matroid is a","Therefore, it is very necessary to investigate the conditions under which a matroid is connected.",8685
"studied through this matroids sets. for is rough paper, relation-based In connectivity the","In this paper, the connectivity for matroids is studied through relation-based rough sets.",8686
"matroid, introduced graph an Moreover, a by general generalized. undirected is the relation through","Moreover, through the relation introduced by a general matroid, an undirected graph is generalized.",8687
"graph investigated the the can be of sets. by connection rough Specifically, relation-based the","Specifically, the connection of the graph can be investigated by the relation-based rough sets.",8688
of relation-based important application sets results an matroids. to show rough These,These results show an important application of relation-based rough sets to matroids.,8689
We this based particular the exploit of conductance investigate to property. models interconnection remarkable structure,We exploit the particular interconnection structure of conductance based models to investigate this remarkable property.,8690
"compare spatial signals, for a a images, must scale. temporal To choose scale, one","To compare images, one must choose a spatial scale, for signals, a temporal scale.",8691
cases some a shown is simpler It solution available. that is in,It is shown that in some cases a simpler solution is available.,8692
checked reduction. strongly how be a can distance hyperparameter dimensionality using depend relationships on It,It can be checked how strongly distance relationships depend on a hyperparameter using dimensionality reduction.,8693
"variant embeds multi-dimensional formulated, of (MDS) as is scaling datapoints which A curves. dynamical","A variant of dynamical multi-dimensional scaling (MDS) is formulated, which embeds datapoints as curves.",8694
on variant hyperparameters is dependence also analyze the multiple A to presented.,A variant to analyze the dependence on multiple hyperparameters is also presented.,8695
"use is implement, is cMDS that to and extend provided. algorithm A straightforward","A cMDS algorithm that is straightforward to implement, use and extend is provided.",8696
"real-world possibilities cMDS several of illustrate applied to sets. cMDS, is data the To","To illustrate the possibilities of cMDS, cMDS is applied to several real-world data sets.",8697
an to the adapting rearrangements have potential chromosomal organism. be advantageous to Gross evolutionarily,Gross chromosomal rearrangements have the potential to be evolutionarily advantageous to an adapting organism.,8698
denoising. for simple are image powerful and filters techniques Nonlocal,Nonlocal filters are simple and powerful techniques for image denoising.,8699
The allows this perform simplicity to its detailed properties. of of a formulation analysis,The simplicity of this formulation allows to perform a detailed analysis of its properties.,8700
problems of non-submodular require optimization Many computer binary energies. vision,Many computer vision problems require optimization of binary non-submodular energies.,8701
(LSA). framework approximations on based general We a optimization local submodular propose,We propose a general optimization framework based on local submodular approximations (LSA).,8702
"optimization other the local hand, unlike On methods (e.g. standard","On the other hand, unlike standard local optimization methods (e.g.",8703
"higher-order While energies, pairwise to problems. focused paper on extend our ideas is our","While our paper is focused on pairwise energies, our ideas extend to higher-order problems.",8704
"computer in and vision (non-linear) High-order become stereo problems. other functionals segmentation, have very popular","High-order (non-linear) functionals have become very popular in segmentation, stereo and other computer vision problems.",8705
"experimentally activity, observed neuron irregular activity. firing highly self-sustained During similar each patterns, to exhibits","During self-sustained activity, each neuron exhibits highly irregular firing patterns, similar to experimentally observed activity.",8706
"implications for self-sustained the activity memory learning, propagation. signal and we Finally, discuss of","Finally, we discuss the implications of self-sustained activity for learning, memory and signal propagation.",8707
in work. and for is Transform The learning Fourier fast adopted detection Fast this,The Fast Fourier Transform is adopted for fast learning and detection in this work.,8708
models. similarities there are these fundamental we Here show two between,Here we show there are fundamental similarities between these two models.,8709
sleep-wake The important implications are two with phenomena. illustrated,The implications are illustrated with two important sleep-wake phenomena.,8710
"preserves structures to fine In details. it contrast length, elongated and","In contrast to length, it preserves elongated structures and fine details.",8711
derive geometry. model We curvature integral computing a based squared for new on,We derive a new model for computing squared curvature based on integral geometry.,8712
straight cliques. responses line triple counts model The of,The model counts responses of straight line triple cliques.,8713
pairwise submodular supermodular and potentials. energy corresponding The into decomposes,The corresponding energy decomposes into submodular and supermodular pairwise potentials.,8714
time Signature for verification been has authentication for and used long purpose.,Signature has been used for long time for verification and authentication purpose.,8715
but digitized. they nowadays getting were are methods Earlier manual,Earlier methods were manual but nowadays they are getting digitized.,8716
using Basis an paper provides recognition Radial Function Network. efficient This method signature to,This paper provides an efficient method to signature recognition using Radial Basis Function Network.,8717
database. with sample The trained network is in images,The network is trained with sample images in database.,8718
is before Feature them extraction for training. performed using,Feature extraction is performed before using them for training.,8719
environments pose tremendous to populations. challenges Fluctuating bacterial,Fluctuating environments pose tremendous challenges to bacterial populations.,8720
"this discuss phenomena. temporal to problems, In various associated we paper","In this paper we discuss various problems, associated to temporal phenomena.",8721
way. a are in and integrated Objects natural processes,Objects and processes are integrated in a natural way.,8722
applied algorithms each Three to in parallel are image. then,Three algorithms are then applied in parallel to each image.,8723
a most purchasing functions the selection one department. is Suuplier important of of,Suuplier selection is one of the most important functions of a purchasing department.,8724
"added. data visualization exploration been have addition, new In and tools options","In addition, new data exploration tools and visualization options have been added.",8725
"of into the was denoising the In algorithms, account. hardly particularity taken image this image","In this image denoising algorithms, the particularity of the image was hardly taken into account.",8726
"underlying by graph: occupy connected which structure vertices, individuals is The are edges. a","The underlying structure is a graph: individuals occupy vertices, which are connected by edges.",8727
"vertex, in an their for individual an they receive suited increase is If fecundity.","If an individual is suited for their vertex, they receive an increase in fecundity.",8728
suitable restricted framework habitat. to allows attention of be to arrangement This spatial the,This framework allows attention to be restricted to the spatial arrangement of suitable habitat.,8729
this some model We prove find of properties results. and counter-intuitive some basic,We prove some basic properties of this model and find some counter-intuitive results.,8730
"maintain also out-cross, hybridization. While capacity selfing for individuals some they primarily small and","While individuals primarily out-cross, they also maintain some small capacity for selfing and hybridization.",8731
background I on the and graph Laplacians. necessary graphs include,I include the necessary background on graphs and graph Laplacians.,8732
Laplacians. of is graph application an This attractive,This is an attractive application of graph Laplacians.,8733
this main normalized The method thrust paper cuts. the of is of,The main thrust of this paper is the method of normalized cuts.,8734
"problem, is used. To this technique correlation overcome","To overcome this problem, correlation technique is used.",8735
rearrangements. by GC short previously detection has genomic been through reads confounded,GC detection through short reads has previously been confounded by genomic rearrangements.,8736
"material paper, with scale information computer information - The quantity. - of discs usually bases","The material bases of information - paper, computer discs - usually scale with information quantity.",8737
bases. Large quantities large of information usually material require,Large quantities of information usually require large material bases.,8738
"savants, in as always memories, Large heads. should large require","Large memories, as in savants, should always require large heads.",8739
small memories. with Small heads scale always should,Small heads should always scale with small memories.,8740
"have Yet some intelligence, these of or of long-term indicating little advanced normal impairment memory.","Yet some of these have normal or advanced intelligence, indicating little impairment of long-term memory.",8741
should afield? looking further Perhaps we be,Perhaps we should be looking further afield?,8742
"sets develop modulo for method surfaces a in spaces, We shape optimization re-parametrization. of i.e.,","We develop a method for optimization in shape spaces, i.e., sets of surfaces modulo re-parametrization.",8743
of meshes. triangular on Latter of an the sparks development method direction regularizer alternating multipliers,Latter regularizer sparks the development of an alternating direction method of multipliers on triangular meshes.,8744
a anything naive the Imagine who world. organism know does about not,Imagine a naive organism who does not know anything about the world.,8745
can signals through sensors it actions. its make It capture and can,It can capture signals through its sensors and it can make actions.,8746
about is knowledge world accessible of organism? the to kind What the,What kind of knowledge about the world is accessible to the organism?,8747
"text, main this with present I examples. of In illustrated the concrete concepts physics, subjective","In this text, I present the main concepts of subjective physics, illustrated with concrete examples.",8748
provide to our CIG with high method the We probability. for correctly analytical conditions identify,We provide analytical conditions for our method to correctly identify the CIG with high probability.,8749
The survival adaptive essential accumulation for in mutations is environments. of novel,The accumulation of adaptive mutations is essential for survival in novel environments.,8750
independently genetic parallelism The mutations in of basis a adaptive populations. the evolving striking revealed,The genetic basis of the adaptive mutations revealed a striking parallelism in independently evolving populations.,8751
"populations, Interestingly phenotypic without observed variation. complete genetic we a most of loss in sweep","Interestingly in most populations, we observed a complete phenotypic sweep without loss of genetic variation.",8752
model by reaction-diffusion-delay and modifies al. This et model the extends Graham,This model extends and modifies the reaction-diffusion-delay model by Graham et al.,8753
classification benchmark. recently Convolutional have demonstrated Large performance the models impressive on Network ImageNet,Large Convolutional Network models have recently demonstrated impressive classification performance on the ImageNet benchmark.,8754
paper issues. this both we In address,In this paper we address both issues.,8755
the performance model We also to layers. perform contribution different study discover from an ablation,We also perform an ablation study to discover the performance contribution from different model layers.,8756
"variants this be can In posttranslational favorable some by cases, separated procedure.","In some favorable cases, posttranslational variants can be separated by this procedure.",8757
enables of Compressed from data. the signals high-resolution reconstruction sensing under-sampled,Compressed sensing enables the reconstruction of high-resolution signals from under-sampled data.,8758
video constructing demonstrate real-time this a We by compressive camera.,We demonstrate this by constructing a real-time compressive video camera.,8759
produce and We of Internet images. that also descriptions they materials intuitive demonstrate,We also demonstrate that they produce intuitive descriptions of materials and Internet images.,8760
on shows excellent performance datasets. classification Our both approach,Our approach shows excellent classification performance on both datasets.,8761
considerable disciplines. The of many aggregation in is importance problem,The problem of aggregation is considerable importance in many disciplines.,8762
The according weights importance of graph. are the visibility in obtained the data to the,The weights are obtained according to the importance of the data in the visibility graph.,8763
These sources. the of instrumental neuronal been signal in also have microelectrodes localization,These microelectrodes have also been instrumental in the localization of neuronal signal sources.,8764
"be not Therefore, experimentally these could validated. methods","Therefore, these methods could not be experimentally validated.",8765
tetrode respect to the location The with of stimulator. the varied was,The location of the tetrode was varied with respect to the stimulator.,8766
new appropriate for by assigning then classifying will model plan. the cases This operate,This model will then operate for classifying new cases by assigning the appropriate plan.,8767
bursts. single-neuron complete of reproduce A spikes correctly and must the model firing,A complete single-neuron model must correctly reproduce the firing of spikes and bursts.,8768
has deconvolution the past progress made decade. Blind in significant,Blind deconvolution has made significant progress in the past decade.,8769
a-Posteriori either Maximum or are algorithms classified successful as Most ($MAP$). Variational,Most successful algorithms are classified either as Variational or Maximum a-Posteriori ($MAP$).,8770
for deconvolution. the required of a unified blind Our principles observations successful lead understanding to,Our observations lead to a unified understanding of the principles required for successful blind deconvolution.,8771
of neural conditions the These design layers. sufficient give invertible network for,These give sufficient conditions for the design of invertible neural network layers.,8772
"and via are reference performed assembly completed, encoding. Once lossless positional is based classification compression","Once assembly and classification are completed, lossless reference based compression is performed via positional encoding.",8773
Haar to to decomposition single-level detailed coefficients. Overall obtain subjected and wavelet are approximate DMFs,Overall DMFs are subjected to single-level Haar wavelet decomposition to obtain approximate and detailed coefficients.,8774
and periodicities row help coefficients column in Extracted in determination of directions.,Extracted coefficients help in determination of periodicities in row and column directions.,8775
of Growth signs population no stagnation. human of shows,Growth of human population shows no signs of stagnation.,8776
of convincingly the The contradicted concept of is by Malthusian empirical Stagnation Epoch evidence.,The concept of the Epoch of Malthusian Stagnation is convincingly contradicted by empirical evidence.,8777
"other the computationally On efficient. the is formulation second always hand,","On the other hand, the second formulation is always computationally efficient.",8778
an this that form for is edges The reason object. the outline of,The reason for this is that edges form the outline of an object.,8779
paper been This author(s) has by withdrawn . the,This paper has been withdrawn by the author(s) .,8780
and association structures studies genome-wide of elements can (GWAS). gene inform regulatory Annotations,Annotations of gene structures and regulatory elements can inform genome-wide association studies (GWAS).,8781
of multiple annotations. model incorporates naturally The types,The model naturally incorporates multiple types of annotations.,8782
collected commonly Symmetric entities in relations representing matrices among are many binary areas.,Symmetric binary matrices representing relations among entities are commonly collected in many areas.,8783
to in markets. financial an consider world We also application co-movements,We also consider an application to co-movements in world financial markets.,8784
rigid bodies animation or numerical soft often for applications of instabilities. Physics-based suffers from real-time,Physics-based animation of soft or rigid bodies for real-time applications often suffers from numerical instabilities.,8785
analyse integration strategy. of common the of numerical sources the most We one unwanted behaviour:,We analyse one of the most common sources of unwanted behaviour: the numerical integration strategy.,8786
"requires and the information. imprecise treatment complex vague, of systems manipulation uncertain The of often","The treatment of complex systems often requires the manipulation of vague, imprecise and uncertain information.",8787
"Indeed, natural such handling being the a in systems is of human competent in way.","Indeed, the human being is competent in handling of such systems in a natural way.",8788
system a In work developed this was tested. skin and recognition,In this work a skin recognition system was developed and tested.,8789
very gave during system results The face. neural generalization network the encouraging,The system gave very encouraging results during the neural network generalization face.,8790
"carry African spretus Further, mice sympatric more European M. than the alleles ones.","Further, European mice carry more M. spretus alleles than the sympatric African ones.",8791
"tracts evolutionary introgressed contain Third, genes fate determine might driver those that the tracts. of","Third, introgressed tracts might contain driver genes that determine the evolutionary fate of those tracts.",8792
"are to exposed genetic they relatively Therefore, drift. little","Therefore, they are exposed to relatively little genetic drift.",8793
information sometimes more communicating way words. and it Emotions are best carry information; than of,Emotions are best way of communicating information; and sometimes it carry more information than words.,8794
an face uses proposed Haar-like features The in method image. detect to,The proposed method uses Haar-like features to detect face in an image.,8795
dimensionality the of (PCA). using is reduced Principal eigenspace Component Analysis The,The dimensionality of the eigenspace is reduced using Principal Component Analysis (PCA).,8796
An method. test implementation given and to is this validate,An implementation is given to test and validate this method.,8797
settings. We attribute tasks substantial challenging vs. show state-of-the-art unconstrained in on improvement classification methods,We show substantial improvement vs. state-of-the-art methods on challenging attribute classification tasks in unconstrained settings.,8798
"brain operation, most a degree Reasoning, is important the human by charactrized fuzziness.","Reasoning, the most important human brain operation, is charactrized by a degree fuzziness.",8799
illustrating classroom number a in results our practice. We present of experiments also,We also present a number of classroom experiments illustrating our results in practice.,8800
propagation the transfer. by Light media in is radiative driven equation of turbid,Light propagation in turbid media is driven by the equation of radiative transfer.,8801
method. are The to Wang) the so-called Wang-Prahl estimating resulting then compared methods (or,The resulting estimating methods are then compared to the so-called Wang-Prahl (or Wang) method.,8802
estimator The estimates the is first one simultaneously. which joint parameters a all,The first one is a joint estimator which estimates all the parameters simultaneously.,8803
We state-of-the-art datasets. our advantages and on real over the model methods synthetic demonstrate of,We demonstrate the advantages of our model over state-of-the-art methods on synthetic and real datasets.,8804
are Results photographs pose for presented. in detection real vehicle,Results for vehicle pose detection in real photographs are presented.,8805
points Effective high. vantage panoramic taken photographs that are are from,Effective panoramic photographs are taken from vantage points that are high.,8806
"vastly reduces to the size aircraft the Also, small of these risks flight. inherent","Also, the small size of these aircraft vastly reduces the risks inherent to flight.",8807
This the of uncertainty the of effects tone-mapping. inherent undoing considers paper,This paper considers the inherent uncertainty of undoing the effects of tone-mapping.,8808
is An influenza example data discussed. with,An example with influenza data is discussed.,8809
tests. rely independence current for methods this Most nonparametric achieving on,Most current methods for achieving this rely on nonparametric independence tests.,8810
world contemporary the a identification of Character in Image processing. plays role vital,Character identification plays a vital role in the contemporary world of Image processing.,8811
It composite can makes solve work easier. many problems humans and,It can solve many composite problems and makes humans work easier.,8812
Character the Networks Tamil distinguish recognition to Neural uses system them. Handwritten,Tamil Handwritten Character recognition system uses the Neural Networks to distinguish them.,8813
used to written characters. characteristics structural and instruct Neural Network and recognize are,Neural Network and structural characteristics are used to instruct and recognize written characters.,8814
every monitoring can the Unobtrusive day for important of use elderly population. be health,Unobtrusive every day health monitoring can be of important use for the elderly population.,8815
and by method area. first the Our detecting works face eyes the,Our method works by first detecting the face and the eyes area.,8816
show our and both approach. efficiency effectiveness result the of the Experimental,Experimental result show both the efficiency and the effectiveness of our approach.,8817
statistics. model Markov modulated (switching) important is applied state The paradigm in an space,The Markov modulated (switching) state space is an important model paradigm in applied statistics.,8818
summary The subsequently presented. algorithmic an detailed descriptions including are,The detailed descriptions including an algorithmic summary are subsequently presented.,8819
inference. symmetries probabilistic speed up algorithms in Lifted exploit to models inference,Lifted inference algorithms exploit symmetries in probabilistic models to speed up inference.,8820
"show with binary particular, we conditioning In rank efficient. on is evidence bounded that Boolean","In particular, we show that conditioning on binary evidence with bounded Boolean rank is efficient.",8821
"to present series. route coherences multivariate another partial evaluate time we the Here, directed in","Here, we present another route to evaluate the partial directed coherences in multivariate time series.",8822
comparison to usual term complexity in detailed. of the A is approach,A comparison to the usual approach in term of complexity is detailed.,8823
"For long is proposed models, AR interest. of approach the","For long AR models, the proposed approach is of interest.",8824
Overall be appear results to extremely encouraging.,Overall results appear to be extremely encouraging.,8825
approach. accelerate to two use the different We ways,We use two different ways to accelerate the approach.,8826
"version acceleration Moreover, TBB. intensively an without paralleled implement GPU we via","Moreover, we implement an intensively paralleled version via GPU acceleration without TBB.",8827
"is unclear. regulate how enzymes on these to function cell-type-specific expression Furthermore, enhancers gene","Furthermore, how these enzymes function on enhancers to regulate cell-type-specific gene expression is unclear.",8828
Internet its are very usage increasing services and available on web Number of fast.,Number of web services available on Internet and its usage are increasing very fast.,8829
with composition. and be can service Development for applications effort reduced new time,Development time and effort for new applications can be reduced with service composition.,8830
out of approaches automated to carry Various are web discussed in composition literature. services,Various approaches to carry out automated composition of web services are discussed in literature.,8831
approaches. one of Web is effective using the composition ontologies service,Web service composition using ontologies is one of the effective approaches.,8832
provide a framework services We fulfil propose to user requirements. web precomposed to,We propose a framework to provide precomposed web services to fulfil user requirements.,8833
which process. for We the be whole used can merging ontology composition detail how expedites,We detail how ontology merging can be used for composition which expedites the whole process.,8834
discuss how We customer repository. and merging provides ontology specific framework,We discuss how framework provides customer specific ontology merging and repository.,8835
how of carried We on out. merging ontologies elaborate also is,We also elaborate on how merging of ontologies is carried out.,8836
evolutionary of  biology. is goal evolutionary Predicting a  adaptive primary trajectories,  Predicting adaptive evolutionary trajectories is a primary goal of evolutionary biology.,8837
have conducted. not been diploid experiments in organisms Similar,Similar experiments in diploid organisms have not been conducted.,8838
neural show perception. to results their and relation importance the codes These exciting of,These exciting results show the importance of neural codes and their relation to perception.,8839
The dynamics} neural the codes. to next of challenge is \emph{model,The next challenge is to \emph{model the dynamics} of neural codes.,8840
electrophysiological then data. We a neural from network the derive,We then derive a neural network from the electrophysiological data.,8841
"unsatisfactory traditional approaches produce matching results. often input, such For shape","For such input, traditional shape matching approaches often produce unsatisfactory results.",8842
coarse correspondences optimization given method ensemble correspondences. obtain to that improves an dense We propose,We propose an ensemble optimization method that improves given coarse correspondences to obtain dense correspondences.,8843
end part-based we introduce this shape To novel generative statistical a model.,To this end we introduce a novel part-based generative statistical shape model.,8844
"In methods. a substantial in obtain practical quality experiments, improvement over we correspondence state-of-the-art","In practical experiments, we obtain a substantial improvement in correspondence quality over state-of-the-art methods.",8845
or object modeled composite piecewise quadric an linear A is model quadric (CQM) by patches.,A composite quadric model (CQM) is an object modeled by piecewise linear or quadric patches.,8846
in to topography characteristics. of is nanomaterials metal nanosensor crucial nanograins compound The,The topography of metal nanograins in compound nanomaterials is crucial to nanosensor characteristics.,8847
three The templates completing of such of method consists steps.,The method of completing such templates consists of three steps.,8848
knowing logics make non-trivial. axiomatizing features whether These,These features make axiomatizing knowing whether logics non-trivial.,8849
whether over classes. We knowing frame axiomatize various logic,We axiomatize knowing whether logic over various frame classes.,8850
compare We proposals. recent work detail our two in similar to,We compare our work in detail to two recent similar proposals.,8851
all Existing methods systematic testing be the test possible cannot for cases. used of,Existing methods cannot be used for the systematic testing of all possible test cases.,8852
here We exhaustive describe new dynamic a verification method.,We describe here a new exhaustive dynamic verification method.,8853
"universality network. evinces structural After this in deviation range, from underlying the features","After this range, deviation from the universality evinces underlying structural features in network.",8854
"the Gabor extracted. Firstly, images low-energized from blocks are wavelet transformed","Firstly, the low-energized blocks from Gabor wavelet transformed images are extracted.",8855
function is The kernel method. KDCV cosine in include method extended to the discriminating,The KDCV method is extended to include cosine kernel function in the discriminating method.,8856
proteomics. Two-dimensional nurtured has of birth electrophoresis the,Two-dimensional electrophoresis has nurtured the birth of proteomics.,8857
"e.g., for Previous presented guided works improvements, Horner schemes. heuristically","Previous works presented heuristically guided improvements, e.g., for Horner schemes.",8858
The common remaining then is subexpression by further elimination. reduced expression,The remaining expression is then further reduced by common subexpression elimination.,8859
"the we intuitive provide of explanation an of in algorithm. terms behavior First, exploration-exploitation the","First, we provide an intuitive explanation in terms of the exploration-exploitation behavior of the algorithm.",8860
"our test expressions of we three origins. Then, different algorithm large on","Then, we test our algorithm on three large expressions of different origins.",8861
advantage We methods to the integrate three mentioned full take the approaches. above of,We integrate the above mentioned methods to take full advantage of the three approaches.,8862
for human been lip used recognizing have identification by Methods prints. of,Methods have been used for identification of human by recognizing lip prints.,8863
a Gaussian using is lip print In Filter. smoothened first this paper,In this paper lip print is first smoothened using a Gaussian Filter.,8864
identification personal identification. This both will method useful and be in forensics of criminal,This method of identification will be useful both in criminal forensics and personal identification.,8865
necessary discussion for steps predicting the distributions. and expertise The helps clarify,The discussion helps clarify the necessary steps and expertise for predicting distributions.,8866
method An paper. proposed on salient is features in based this image fusion,An image fusion method based on salient features is proposed in this paper.,8867
feature individual for calculated and are images. those are stored as vectors,are calculated and those are stored as feature vectors for individual images.,8868
"used Here, K-NN recognize probe classifier is to face-sketch.","Here, K-NN classifier is used to recognize probe face-sketch.",8869
applications has It for law digital enforcement and both useful entertainment.,It has useful applications for both law enforcement and digital entertainment.,8870
data These in sequencing platforms. multiple findings reproducible from are,These findings are reproducible in data from multiple sequencing platforms.,8871
features dimensional face informative Thus can space. extracted the low in of a be a,Thus the informative features of a face can be extracted in a low dimensional space.,8872
estimation of error results reduces the This efficient implementation in algorithm variance. and an the,This results in an efficient implementation of the algorithm and reduces the estimation error variance.,8873
"model expectation identification algorithm, with enables unknown the of Together online recursive this parameters. maximization","Together with the online expectation maximization algorithm, this enables recursive identification of unknown model parameters.",8874
for contour generalized shape based develop on a templates. formalism segmentation image We active,We develop a generalized active contour formalism for image segmentation based on shape templates.,8875
We the show functional certain under optimal energy conditions. is contrast that,We show that the contrast energy functional is optimal under certain conditions.,8876
show theorem. use that calculations of through are We Green's made the efficient,We show that the calculations are made efficient through use of Green's theorem.,8877
for feature a technique has scanning selection. been better proposed spiral,a spiral scanning technique has been proposed for better feature selection.,8878
"drastically dependency long-range problem HMMs inherent the been has traditional to Thus, reduced.","Thus, the long-range dependency problem inherent to traditional HMMs has been drastically reduced.",8879
"mixing of the is However, sampler Carlo the behavior Markov poor. sometimes Monte (MCMC) Chain","However, the mixing behavior of the Markov Chain Monte Carlo (MCMC) sampler is sometimes poor.",8880
"poor to can problems, as mixing under-representing inferential uncertainty. lead This such","This poor mixing can lead to inferential problems, such as under-representing uncertainty.",8881
"develop paper, we mechanisms sampling. this In for efficient novel proposal","In this paper, we develop novel proposal mechanisms for efficient sampling.",8882
first while second perturbation call the proposal a rule is tree we The rotation.,The first is a rule perturbation proposal while the second we call tree rotation.,8883
"as are classification which these feature are used peaks Then concatenated process. Hough together, in","Then these Hough peaks are concatenated together, which are used as feature in classification process.",8884
models core The break to is conditionally into distributions. complex independent idea,The core idea is to break complex models into conditionally independent distributions.,8885
of Theory Practice (TPLP). and Programming in Logic appear To,To appear in Theory and Practice of Logic Programming (TPLP).,8886
"large However, OCT high-speed speckle to noise. the of is imaging the quality limited due","However, the imaging quality of high-speed OCT is limited due to the large speckle noise.",8887
"To volume. problem, multi-frame paper to proposes a denoise this address method this algorithmic OCT","To address this problem, this paper proposes a multi-frame algorithmic method to denoise OCT volume.",8888
truncated propose integrate probability method a to efficiently densities. We,We propose a method to efficiently integrate truncated probability densities.,8889
with is case demonstrated The studies. numerical method,The method is demonstrated with numerical case studies.,8890
and enhancements limitations are Possible the method to discussed.,Possible enhancements to the method and limitations are discussed.,8891
"to relating is to KECA, After extended the kernel entropy Renyi function. that include cosine","After that KECA, relating to the Renyi entropy is extended to include cosine kernel function.",8892
"Cognitive limitations. Artificial example, a intelligent number of Traditional robots) Systems share (for","Traditional Artificial Cognitive Systems (for example, intelligent robots) share a number of limitations.",8893
"made time-shared that Second, and that are of components up are re-usable.","Second, that are made up of components that are time-shared and re-usable.",8894
"through that Third, provide increased self-repair. robustness","Third, that provide increased robustness through self-repair.",8895
process is observed birth-death reconstructed evolutionary The homogeneous without lineages. a extinct process,The homogeneous reconstructed evolutionary process is a birth-death process without observed extinct lineages.,8896
suspicious paper which The email model incorporates presents enhanced feature detection a selection.,The paper presents a suspicious email detection model which incorporates enhanced feature selection.,8897
"good the accuracy various the literature, In task. desired achieved algorithms for","In the literature, various algorithms achieved good accuracy for the desired task.",8898
phenotypic evolutionary Understanding causes biology. and of is a in goal patterns central the divergence,Understanding the patterns and causes of phenotypic divergence is a central goal in evolutionary biology.,8899
mRNA related Much closely shown work that has between highly abundance variable species. is,Much work has shown that mRNA abundance is highly variable between closely related species.,8900
"are extent mechanisms evolution of unknown. gene However, largely regulatory the and post-transcriptional","However, the extent and mechanisms of post-transcriptional gene regulatory evolution are largely unknown.",8901
abundance Compensatory differences translation in may transcript gene of regulation. efficiency and increase the robustness,Compensatory differences in transcript abundance and translation efficiency may increase the robustness of gene regulation.,8902
a photo through sketch face for todays task database researchers. face recognize To is challenging,To recognize face sketch through face photo database is a challenging task for todays researchers.,8903
"(PCA). from Principal features images Analysis extract Component After transformed that, using","After that, extract features from transformed images using Principal Component Analysis (PCA).",8904
"classifier we SVM Thereafter, classification. and better for classifier use K-NN","Thereafter, we use SVM classifier and K-NN classifier for better classification.",8905
Our on relies results matrix from random procedure theory. adapting,Our procedure relies on adapting results from random matrix theory.,8906
"method. number to nonparametric a our hLDA, topic superior is procedure Performance recovery of","Performance of our topic number recovery procedure is superior to hLDA, a nonparametric method.",8907
This motivates practical a algorithm. description also,This description also motivates a practical algorithm.,8908
"new high-dimensional with up of to algorithms Ultimately, our problems scale millions variables.","Ultimately, our new algorithms scale up to high-dimensional problems with millions of variables.",8909
"improvement attempt through analysis. Here, theoretical at quantifying provide an we this","Here, we provide an attempt at quantifying this improvement through theoretical analysis.",8910
to set. is and well a data real work This technique simulations shown through on,This technique is shown to work well through simulations and on a real data set.,8911
implementation We Multi-Bernoulli the filter. $\delta$-Generalized an multi-target present efficient numerical tracking Labeled of,We present an efficient numerical implementation of the $\delta$-Generalized Labeled Multi-Bernoulli multi-target tracking filter.,8912
adaptive question the studies males. female for   value preferences older Several of,  Several studies question the adaptive value of female preferences for older males.,8913
mates. visible not are females choosing to These mutations,These mutations are not visible to females choosing mates.,8914
"genes."" mutations oppose Germ-line could preferences for ""good","Germ-line mutations could oppose preferences for ""good genes.""",8915
depends includes on mutation the time generation bias My population. that dynamical the equation of,My dynamical equation includes mutation bias that depends on the generation time of the population.,8916
findings My genes with agree theory. good,My findings agree with good genes theory.,8917
Females males. above-average at always select equilibrium,Females at equilibrium always select above-average males.,8918
trait preferred with direct by of females correlates size directly preference. the costs The the,The trait size preferred by females directly correlates with the direct costs of the preference.,8919
rate accentuate mutational can at preference a higher equilibrium than parameters. Direct costs the,Direct costs can accentuate the equilibrium preference at a higher rate than mutational parameters.,8920
"can Females always offset ornamented mating more costs older, direct by males. with","Females can always offset direct costs by mating with older, more ornamented males.",8921
"an mutations essential selection. than Rather eliminating germ-line preferences, female ingredient in provide sexual","Rather than eliminating female preferences, germ-line mutations provide an essential ingredient in sexual selection.",8922
knowledge classes in method propose related for a semantically ImageNet. We transfer between,We propose a method for knowledge transfer between semantically related classes in ImageNet.,8923
the space low-dimensional of among appearance recovers AE windows. latent variations image,AE recovers the latent low-dimensional space of appearance variations among image windows.,8924
to correspond (e.g. AE The of appearance dimensions to space aspects of window tend,The dimensions of AE space tend to correspond to aspects of window appearance (e.g.,8925
"self-assessment, perform probabilistic i.e. allows The nature GP of to method our","The probabilistic nature of GP allows our method to perform self-assessment, i.e.",8926
estimate output. to own quality a its assigning,assigning a quality estimate to its own output.,8927
their returned off It quality. amount annotations for trading the enables of,It enables trading off the amount of returned annotations for their quality.,8928
"epsilon-random the specific we graphs, models, and graphs. neighbor consider Finally, graph torus, k-nearest","Finally, we consider specific graph models, the torus, k-nearest neighbor graphs, and epsilon-random graphs.",8929
"addition, occur can rate evolutionary outcome. the interactions alter at In game the which","In addition, the rate at which game interactions occur can alter the evolutionary outcome.",8930
measures derive We populations account appropriate that for for heterogeneous bias. this the,We derive the appropriate measures for heterogeneous populations that account for this bias.,8931
task the expertise Ontology development requiring a in ontological language. is non-trivial chosen,Ontology development is a non-trivial task requiring expertise in the chosen ontological language.,8932
for referral and Conclusions: substantially reduce Quicker gonorrhoea. and treatment re-infection can partner rates chlamydia,Conclusions: Quicker partner referral and treatment can substantially reduce re-infection rates for chlamydia and gonorrhoea.,8933
"cells, and The the of elements model immune host's response, treatments. tumor includes","The model includes tumor cells, elements of the host's immune response, and treatments.",8934
match Our for results closely results protocols. these trial clinical,Our results closely match clinical trial results for these protocols.,8935
"we and values. examined equilibria system's parameter to the sensitivity Additionally,","Additionally, we examined the system's equilibria and sensitivity to parameter values.",8936
"is affected parameters. treatment most When PK/PD drug-specific is introduced, by tumor growth","When treatment is introduced, tumor growth is most affected by drug-specific PK/PD parameters.",8937
we statistical the test first recessive wide genome for Here present selection.,Here we present the first genome wide statistical test for recessive selection.,8938
test to explicitly populations demographic of the non-equilibrium differences infer selection. This between mode uses,This test uses explicitly non-equilibrium demographic differences between populations to infer the mode of selection.,8939
"a paper, we for data search. nearest this approximate propose neighbor In structure new","In this paper, we propose a new data structure for approximate nearest neighbor search.",8940
graph. bridge with a neighborhood structure This graph the augments,This structure augments the neighborhood graph with a bridge graph.,8941
the eyes. of in thickness distribution three-dimensional demonstrated layers retinal normal SD-OCT,SD-OCT demonstrated the three-dimensional thickness distribution of retinal layers in normal eyes.,8942
and sectors. with and of varied sex in Thickness age layers different,Thickness of layers varied with sex and age and in different sectors.,8943
thickness. considered evaluating should macular while These variables be,These variables should be considered while evaluating macular thickness.,8944
Phishing. malicious activities such has also to rise given This as,This has also given rise to malicious activities such as Phishing.,8945
Using proposes Rough Phishing paper Set towards approach This Theory. Detection an,This paper proposes an approach towards Phishing Detection Using Rough Set Theory.,8946
"factors, Strata. four directly Thirteen into are basic towards Phishing, responsible The grouped","The Thirteen basic factors, directly responsible towards Phishing, are grouped into four Strata.",8947
determines of Reliability or suspected Valid Factor possibility to Fake. site be the a,Reliability Factor determines the possibility of a suspected site to be Valid or Fake.,8948
Using towards least the also Theory Rough set determined. factors most and are Phishing influential,Using Rough set Theory most and the least influential factors towards Phishing are also determined.,8949
"geometric Therefore, lattices to combine to coverings with it solve meaningful is problems. the optimization","Therefore, it is meaningful to combine coverings with geometric lattices to solve the optimization problems.",8950
"a transversal matroids. through lattice is geometric covering constructed of a First, structure","First, a geometric lattice structure of a covering is constructed through transversal matroids.",8951
are used atoms Then to lattice. the its studied and describe,Then its atoms are studied and used to describe the lattice.,8952
least analyzed using At genes twenty the were proposed study.,At least twenty genes were analyzed using the proposed study.,8953
types of number were repeats and observed. A mirror,A number and types of mirror repeats were observed.,8954
also repeats. We have tried nomenclature give to these to,We have also tried to give nomenclature to these repeats.,8955
data by ideas utilized mining. Graph theoretical science especially fields computer are highly,Graph theoretical ideas are highly utilized by computer science fields especially data mining.,8956
"In tree. in data can this of form a designed field, structure the be","In this field, a data structure can be designed in the form of tree.",8957
"to matroid theory. theory studying graph paper approach summary, a and new In provides this","In summary, this paper provides a new approach to studying graph theory and matroid theory.",8958
some All under proposed have certain methods conditions. drawbacks,All methods proposed have some drawbacks under certain conditions.,8959
these propose multi-model drawbacks we a . overcome approach To,To overcome these drawbacks we propose a multi-model approach .,8960
metabolically member aeruginosa a the flexible is of Gammaproteobacteria. Pseudomonas,Pseudomonas aeruginosa is a metabolically flexible member of the Gammaproteobacteria.,8961
"the for this mathematical this knowledge, of is our bacterium. To first denitrification model","To our knowledge, this is the first mathematical model of denitrification for this bacterium.",8962
transitions. applied thermodynamics to A simple conformational generated protein using concepts elementary is in expression,A simple expression is generated using concepts in elementary thermodynamics applied to protein conformational transitions.,8963
basic is data representation and issue reduction a knowledge in mining. Attribute,Attribute reduction is a basic issue in knowledge representation and data mining.,8964
the Rough provide for sets a issue. foundation theoretical,Rough sets provide a theoretical foundation for the issue.,8965
"rough with the is combine sets to to solve optimization problems. Therefore, it matroids meaningful","Therefore, it is meaningful to combine matroids with rough sets to solve the optimization problems.",8966
"of a space is constructed. First, dependence matroids","First, a dependence space of matroids is constructed.",8967
"reduction. approaches new attribute a word, this to paper In provides study","In a word, this paper provides new approaches to study attribute reduction.",8968
"In the C-Means paper, this algorithm. an extension we Fuzzy of propose classical clustering","In this paper, we propose an extension of the classical Fuzzy C-Means clustering algorithm.",8969
"a Further, propose we scheme using achieve algorithm. the crisp VFC clustering to","Further, we propose a scheme to achieve crisp clustering using the VFC algorithm.",8970
the networks. complex studies gene of behavior structure biology regulatory and Systems,Systems biology studies the structure and behavior of complex gene regulatory networks.,8971
(P.E.M. By synthetic of the methods biology using,By using the methods of synthetic biology (P.E.M.,8972
rebalance nowadays them. why measures system This to providers bike is sharing most take,This is why nowadays most bike sharing system providers take measures to rebalance them.,8973
"iris recently biometric system, attracted stable attention among As has great the researchers. a","As a stable biometric system, iris has recently attracted great attention among the researchers.",8974
"faster localization So, pupil and be robust. will","So, pupil localization will be faster and robust.",8975
"has Finally, been matching using criterion. iris the a semi-correlation performed","Finally, the iris matching has been performed using a semi-correlation criterion.",8976
The threatens expansion human ocean-wide activities of rapid loss. biodiversity,The rapid expansion of human activities threatens ocean-wide biodiversity loss.,8977
measure distance to histograms relationships. is used to between differences Quadratic-Chi capture family The cross-bin,The Quadratic-Chi distance family is used to measure differences between histograms to capture cross-bin relationships.,8978
available the code has in repository. publicly been made The MLOSS,The code has been made publicly available in the MLOSS repository.,8979
"food or dispersal) nutrients. seed of pollination, instead","pollination, seed dispersal) instead of food or nutrients.",8980
". and services, Abstract numerical functional resources, Keywords: response steady-state, mutualism,","Abstract Keywords: mutualism, resources, services, steady-state, functional and numerical response .",8981
"latest idea, advertise paper an our Ogryzko. also described I in (Bordonaro,","I also advertise an idea, described in our latest paper (Bordonaro, Ogryzko.",8982
"highlighted (i.e. are In particular, symptoms motor four related PD","In particular, four PD related motor symptoms are highlighted (i.e.",8983
"and dyskinesia) tremor, their and freezing gait bradykinesia, summarized. details of","tremor, bradykinesia, freezing of gait and dyskinesia) and their details summarized.",8984
$n$-site phosphorylation distributive frequently. Mathematical studied sequential are models of therefore,Mathematical models of $n$-site sequential distributive phosphorylation are therefore studied frequently.,8985
conjecture are counterexamples to Our Wang therefore and of the Sontag. results,Our results therefore are counterexamples to the conjecture of Wang and Sontag.,8986
to first-order extends representations. procedure Our Skolemization of the counters model applicability these,Our Skolemization procedure extends the applicability of first-order model counters to these representations.,8987
"counting Moreover, it of algorithms. model the design simplifies lifted","Moreover, it simplifies the design of lifted model counting algorithms.",8988
Marine microbial viruses community. structure the the shape of,Marine viruses shape the structure of the microbial community.,8989
"are, the thus, biogeochemical the most in cycles key determinant of They planet. a important","They are, thus, a key determinant of the most important biogeochemical cycles in the planet.",8990
"The in ways. example, two different cycle, modeled indistinctly for infection is very","The infection cycle, for example, is indistinctly modeled in two very different ways.",8991
"computational analysis. eco-evolutionary end, use To and mathematical we this","To this end, we use mathematical and eco-evolutionary computational analysis.",8992
"descriptions interactions. long-term can fail rate-based describe microbe-virus properly to However,","However, rate-based descriptions can fail to describe properly long-term microbe-virus interactions.",8993
method Deep estimation on based pose We human (DNNs). Networks a Neural propose for,We propose a method for human pose estimation based on Deep Neural Networks (DNNs).,8994
towards problem pose a joints. body The DNN-based is regression estimation as formulated,The pose estimation is formulated as a DNN-based regression problem towards body joints.,8995
results precision DNN high in of pose such estimates. We a regressors cascade which present,We present a cascade of such DNN regressors which results in high precision pose estimates.,8996
evolution GRNs of the Explosion. explain alone Cambrian life in cannot multicellular the,GRNs alone cannot explain the evolution of multicellular life in the Cambrian Explosion.,8997
Networks links. are network are on addressing construct used to which systems based,Networks are based on addressing systems which are used to construct network links.,8998
genes). CENES These networks (for new called Control control are developmental,These new developmental control networks are called CENES (for Control genes).,8999
evolution Explosion was networks explain these of how the would possible. The Cambrian higher new,The evolution of these new higher networks would explain how the Cambrian Explosion was possible.,9000
"we of the detection model symmetries. automatic common address work, problematic In this","In this work, we address the automatic detection of common problematic model symmetries.",9001
several symmetries. broad detect derive to of to classes We how algorithms show local,We show how to derive algorithms to detect several broad classes of local symmetries.,9002
"contextual use notably temporal discounting nor neither classical certain cases, Nevertheless, cannot discounting. model uniform","Nevertheless, neither classical uniform discounting nor contextual cannot model certain use cases, notably temporal discounting.",9003
"discounting optimistic, proposed. and new contextual proportional schemes, are article, In this conservative,","In this article, new contextual discounting schemes, conservative, proportional and optimistic, are proposed.",9004
operations examined. are discounting properties Some these of,Some properties of these discounting operations are examined.,9005
be of to is shown these schemes. a discounting Classical case special,Classical discounting is shown to be a special case of these schemes.,9006
are temporal motivating cases reliability discounting. of Two and to discussed: application source modelling,Two motivating cases are discussed: modelling of source reliability and application to temporal discounting.,9007
in the developed The were MATLAB. implementations described software,The implementations described were developed in the software MATLAB.,9008
There features proposed of to set two initial an correspondences. approaches are establish,There are proposed two approaches to establish an initial set of features correspondences.,9009
We propose location. Bayesian a information and textural with to similarity color about approach merge,We propose a Bayesian approach to merge textural similarity with information about color and location.,9010
"of our the real through demonstrate Lastly, experiments on images. utility we algorithm","Lastly, we demonstrate the utility of our algorithm through experiments on real images.",9011
the volume one the changes lineage. spectacular brain is in in of Growth most hominid,Growth in brain volume is one of the most spectacular changes in the hominid lineage.,9012
agrees anthropological that on community The point.,The anthropological community agrees on that point.,9013
"contributing pressures growth. been reached No however, selection on to that consensus, has","No consensus, however, has been reached on selection pressures contributing to that growth.",9014
accepted were this peculiar. dynamics growth the broadly It is that of,It is broadly accepted that the dynamics of this growth were peculiar.,9015
"Growth very in or was scale of even the time. rapid evolutionary fast,","Growth was very fast, or even rapid in the evolutionary scale of time.",9016
paper Bernstein penalization nonconvex using functions. this In we study,In this paper we study nonconvex penalization using Bernstein functions.,9017
this losses. by public toxin The economic caused raised caused outbreaks health and serious crisis,The outbreaks caused by this toxin raised serious public health crisis and caused economic losses.,9018
detection relevant food-borne food Rapid of pathogens. contamination is of the containment therefore for,Rapid detection of food contamination is therefore relevant for the containment of food-borne pathogens.,9019
"ensemble-of-forests introduce the ensemble-of-trees We of a model. model, generalization the","We introduce the ensemble-of-forests model, a generalization of the ensemble-of-trees model.",9020
addresses with the state-space filtering model. This problem of a paper,This paper addresses the problem of filtering with a state-space model.,9021
probabilistic model assume that a approaches (i.e. for Standard filtering observations for,Standard approaches for filtering assume that a probabilistic model for observations (i.e.,9022
observation given model) or least is at explicitly parametrically. the,the observation model) is given explicitly or at least parametrically.,9023
demonstrate by experiments. first findings theoretical We these synthetic,We first demonstrate these theoretical findings by synthetic experiments.,9024
DNN Training. on data are by clickthrough The learned multi-task weights Ring of,The weights of multi-task DNN are learned on clickthrough data by Ring Training.,9025
show real effectiveness simulated results of proposed on dataset the method. the both and Experimental,Experimental results on both simulated and real dataset show the effectiveness of the proposed method.,9026
an be excellent test used games Video techniques. as Artificial for Intelligence can bed (AI),Video games can be used as an excellent test bed for Artificial Intelligence (AI) techniques.,9027
"They and challenging difficult very makes AI to non-deterministic, it write players. are strong this","They are challenging and non-deterministic, this makes it very difficult to write strong AI players.",9028
game Pac-Man. such is a Ms. example An of video,An example of such a video game is Ms. Pac-Man.,9029
its adjacent cells pass update. each value at onto) state,pass its state value onto) adjacent cells at each update.,9030
local complex global interactions Simple are with results achieved. truly,Simple local interactions with complex global results are truly achieved.,9031
Video computer representation and in an task the community. challenging vision important is,Video representation is an important and challenging task in the computer vision community.,9032
call approximation We that we an ridge {matrix approach discuss approximation}.,We discuss an approximation approach that we call {matrix ridge approximation}.,9033
"data analysis. in approach the of multivariate illustrate the Finally, applications approximation we","Finally, we illustrate the applications of the approximation approach in multivariate data analysis.",9034
Microsoft built is Basic. system The Visual using,The system is built using Microsoft Visual Basic.,9035
structure Ganglia central loops. and multiple a The subcortical cortical is in Basal involved,The Basal Ganglia is a central structure involved in multiple cortical and subcortical loops.,9036
saccade these target believed for loops responsible of selection. are Some to be,Some of these loops are believed to be responsible for saccade target selection.,9037
two and loop. interactions loops It one the parallel is structured cortico-basal around of tecto-basal,It is structured around the interactions of two parallel cortico-basal loops and one tecto-basal loop.,9038
or These concert reward may in regarding disturb work different each maximization. loops other,These different loops may work in concert or disturb each other regarding reward maximization.,9039
are these tasks. capabilities saccade and learning loops tested Interactions between on their different,Interactions between these loops and their learning capabilities are tested on different saccade tasks.,9040
"model the predicts of dominate. the that loop absence in control, should prefrontal spatial Finally,","Finally, the model predicts that in absence of prefrontal control, the spatial loop should dominate.",9041
onto these that information assume neighbours We can within pass graph. the others,We assume that these neighbours can pass information onto others within the graph.,9042
ecologically temperature. humidity different and of in Both terms regions are also,Both regions are also ecologically different in terms of humidity and temperature.,9043
"and been some composition showing identified, in abundance. differences regional orders have Twenty-eight","Twenty-eight orders have been identified, showing some regional differences in composition and abundance.",9044
difference in the diversity an This non-protected unexpected area. shows greater,This difference shows an unexpected greater diversity in the non-protected area.,9045
agriculture) genus important becomes level. more at,agriculture) becomes more important at genus level.,9046
"However, tools of these architecture features development. simultaneously require to further brain probe","However, tools to simultaneously probe these features of brain architecture require further development.",9047
"contrast with null For we those obtained statistical our comparison, several models. results benchmark for","For statistical comparison, we contrast our results with those obtained for several benchmark null models.",9048
that weighted organizational profiles graphs. diagnostic capture complex multi-resolution work in curves Our demonstrates,Our work demonstrates that multi-resolution diagnostic curves capture complex organizational profiles in weighted graphs.,9049
a probabilistic extension of is logic CP-logic the FO(ID).,CP-logic is a probabilistic extension of the logic FO(ID).,9050
levels gene of sustain expression different under conditions. identical systems gene Multistable external regulatory,Multistable gene regulatory systems sustain different levels of gene expression under identical external conditions.,9051
"known. not of a a which pathway transition is generally such However, triggers component","However, which component of a pathway triggers such a transition is generally not known.",9052
atoms. In we constructs as any these to of refer paper generalized this,In this paper we refer to any of these constructs as generalized atoms.,9053
effective authentication is method Biometrics automatically recognizing individuals. an for,Biometrics authentication is an effective method for automatically recognizing individuals.,9054
enrollment phase identification of verification consists phase. The an authentication or and an,The authentication consists of an enrollment phase and an identification or verification phase.,9055
presents interest for paper a technology. print (ROI) use palm region of of This,This paper presents use of a region of interest (ROI) for palm print technology.,9056
of existing identification First been print have the for palm methods some introduced.,First some of the existing methods for palm print identification have been introduced.,9057
work the Several extraction of region examined. topic have in been existing,Several existing work in the topic of region extraction have been examined.,9058
The considerably algorithms and are related than fixed-point faster algorithms. new currently used,The new algorithms are considerably faster than currently used fixed-point and related algorithms.,9059
accelerate $M$-estimators with incomplete initially. connection our with to symmetrized procedures $U$-statistics work In we,In connection with symmetrized $M$-estimators we work with incomplete $U$-statistics to accelerate our procedures initially.,9060
"EXMOVES, actions for introduces This of learned videos. features recognition exemplar-based in paper efficient","This paper introduces EXMOVES, learned exemplar-based features for efficient recognition of actions in videos.",9061
approach demonstrated are the recognition and benchmarks. The accuracy efficiency of several action on large-scale,The accuracy and efficiency of the approach are demonstrated on several large-scale action recognition benchmarks.,9062
"observable by described instances modeling called In traditional attributes. are predictive the variables,","In the traditional predictive modeling instances are described by observable variables, called attributes.",9063
model for a the for to target variable instances. goal predicting is unseen The learn,The goal is to learn a model for predicting the target variable for unseen instances.,9064
emphasize instances sensitive actions. all are changes equally in that to We not,We emphasize that not all instances are equally sensitive to changes in actions.,9065
way other). opinion not have who do one the a or users strong,users who do not have a strong opinion one way or the other).,9066
between has answer negation set been useful negation default The programming. strong in distinction and,The distinction between strong negation and default negation has been useful in answer set programming.,9067
model similar by with semantics note Cabalar. stable that also hold the results We functional,We also note that similar results hold with the functional stable model semantics by Cabalar.,9068
"problem However, to sequence be shortest is known a NP-hard. of the finding synchronizing","However, the problem of finding a shortest synchronizing sequence is known to be NP-hard.",9069
"ASP synchronizing finite sequence, . shortest automata, Keywords:","Keywords: finite automata, shortest synchronizing sequence, ASP .",9070
"This classification (ConvNets). paper using Networks of the learnt Convolutional deep addresses image models, visualisation","This paper addresses the visualisation of image classification models, learnt using deep Convolutional Networks (ConvNets).",9071
"class to map, second saliency a a computes given technique specific image class. and The","The second technique computes a class saliency map, specific to a given image and class.",9072
These a new in research (CASP). programming resulted set answer constraint area: efforts,These efforts resulted in a new research area: constraint answer set programming (CASP).,9073
"systems non-trivial is CASP requiring difficult, Yet, multiple in of expertise development the areas.","Yet, the development of CASP systems is difficult, requiring non-trivial expertise in multiple areas.",9074
a for of This study need identifying principles development systems. suggests general hybrid a,This suggests a need for a study identifying general development principles of hybrid systems.,9075
employed Computer in is eye-movement Study (HCI) of Human Interaction being research.,Study of eye-movement is being employed in Human Computer Interaction (HCI) research.,9076
generalisation powerful dialectical Dung's frameworks argumentation a abstract Abstract of are (ADFs) frameworks.,Abstract dialectical frameworks (ADFs) are a powerful generalisation of Dung's abstract argumentation frameworks.,9077
We why ASP significant queries current solvers. for the challenges pose practical explain these,We explain why these queries pose significant practical challenges for the current ASP solvers.,9078
proof a as concatenation constructor. or acting,acting as a concatenation or proof constructor.,9079
detected once the rules output. These four produced are directly edge images to and applied,These four rules are directly applied once to the images and produced edge detected output.,9080
"with problem, multiple images challenging planes Detecting one many is a in applications. but","Detecting multiple planes in images is a challenging problem, but one with many applications.",9081
"the homographies the us in participating planes. Features scene multiple multiple detected, provide","Features participating in the multiple homographies detected, provide us the multiple scene planes.",9082
classification to and merging perform better .,to perform better classification and merging .,9083
further and RAD-Seq-like between datasets. the explored We methods using simulated differences Seq-cap-,We further explored the differences between methods using simulated Seq-cap- and RAD-Seq-like datasets.,9084
"concentrate interest. resources on ignoring background can By object generative models of clutter, the their","By ignoring background clutter, generative models can concentrate their resources on the object of interest.",9085
"As to of chief physical is subject molecule DNA extensive life, the manipulations. informational","As the chief informational molecule of life, DNA is subject to extensive physical manipulations.",9086
"vivo not Surprisingly, affect regulation. sequence-dependent flexibility in does gene","Surprisingly, sequence-dependent flexibility does not affect in vivo gene regulation.",9087
arbitrary problem. a is hard in Recognizing unconstrained text photographs natural multi-character,Recognizing arbitrary multi-character text in unconstrained natural photographs is a hard problem.,9088
"this this In we domain viz. an hard address equally in sub-problem paper,","In this paper, we address an equally hard sub-problem in this domain viz.",9089
arbitrary Street recognizing numbers multi-digit imagery. from View,recognizing arbitrary multi-digit numbers from Street View imagery.,9090
"steps. typically and this the localization, solve to out recognition approaches separate segmentation, Traditional problem","Traditional approaches to solve this problem typically separate out the localization, segmentation, and recognition steps.",9091
"Manual costly, inefficient, ineffective. be and roadside potentially ultimately to enforcement known is dangerous,","Manual roadside enforcement is known to be inefficient, costly, potentially dangerous, and ultimately ineffective.",9092
"automated need occupancy Therefore, HOV/HOT there for detection enforcement. vehicle to lane a support is","Therefore, there is a need for automated vehicle occupancy detection to support HOV/HOT lane enforcement.",9093
"classification, detection. integrated for Convolutional and an for using framework Networks localization present We","We present an integrated framework for using Convolutional Networks for classification, localization and detection.",9094
are confidence. then than increase Bounding suppressed rather detection order boxes in to accumulated,Bounding boxes are then accumulated rather than suppressed in order to increase detection confidence.,9095
tasks can using show that single learned a We simultaneously be shared different network.,We show that different tasks can be learned simultaneously using a single shared network.,9096
"task. state the In of a post-competition art work, establish the we new for detection","In post-competition work, we establish a new state of the art for the detection task.",9097
"best OverFeat. Finally, from model release extractor a we feature called our","Finally, we release a feature extractor from our best model called OverFeat.",9098
"kernel-based paper, propose method for an outlier-robust this identification. linear regularized In system we","In this paper, we propose an outlier-robust regularized kernel-based method for linear system identification.",9099
program. we make step to Now want next the create to this and,Now we want to make the next step and to create this program.,9100
"wavelet network, for two-layer a object We classification. introduce scattering","We introduce a two-layer wavelet scattering network, for object classification.",9101
"reinforcement avoid In to seek and rewards punishments. learning, goal is the","In reinforcement learning, the goal is to seek rewards and avoid punishments.",9102
different coinfection Two modes based been experimental observation. posited of on have,Two different modes of coinfection have been posited based on experimental observation.,9103
"the one In the independently. virus mode, virophage and host enter","In one mode, the virophage and virus enter the host independently.",9104
ecological the effects of coinfection? modes of Here ask: we are what different these,Here we ask: what are the ecological effects of these different modes of coinfection?,9105
"differences we qualitative models. find do between However,","However, we do find qualitative differences between models.",9106
is soft-tissue an medical imaging information providing about advanced MRI technique the rich human anatomy.,MRI is an advanced medical imaging technique providing rich information about the human soft-tissue anatomy.,9107
segmentation different in MRI of described. are also and The detection procedures use image,The use of MRI image detection and segmentation in different procedures are also described.,9108
alternative based We on ASP. an approach here develop,We here develop an alternative approach based on ASP.,9109
aspartame re-uses sugar parts and system The parsing for of CSPs. resulting normalizing,The resulting system aspartame re-uses parts of sugar for parsing and normalizing CSPs.,9110
our approach empirically establish sugar. by and aspartame the competitiveness We contrasting of,We establish the competitiveness of our approach by empirically contrasting aspartame and sugar.,9111
"If manually be it time-consuming. their the own medical experts with knowledge, do will segmentation","If experts do the segmentation manually with their own medical knowledge, it will be time-consuming.",9112
the also A Zanjan Center. MRI by provided database is,A database is also provided by the Zanjan MRI Center.,9113
Experimental algorithm. dependability the of show precision and results proposed the high,Experimental results show the high precision and dependability of the proposed algorithm.,9114
"convex represents or as a set observations combinations patterns, Archetypal pure archetypes. of analysis of","Archetypal analysis represents a set of observations as convex combinations of pure patterns, or archetypes.",9115
"practical is with not unfortunately, situations. compatible This, many","This, unfortunately, is not compatible with many practical situations.",9116
tool analysis better. archetypal solution summarize We visualization an present to also appropriate,We also present an appropriate visualization tool to summarize archetypal analysis solution better.,9117
rational costs decision-making problem specifically off and information-processing this by utility. addresses trading expected Bounded,Bounded rational decision-making addresses this problem by specifically trading off information-processing costs and expected utility.,9118
"entropy describing trade-off changes Interestingly, a arises and between systems. energy in when thermodynamic similar","Interestingly, a similar trade-off between energy and entropy arises when describing changes in thermodynamic systems.",9119
This has to agents. rational recently been describe used similarity bounded,This similarity has been recently used to describe bounded rational agents.,9120
"we aiming gap. contributions make three this Here, bridge to","Here, we make three contributions aiming to bridge this gap.",9121
"be can effects by models, by captured standard but thermodynamics. These not diffusion-based facilitated","These effects can be captured by facilitated diffusion-based models, but not by standard thermodynamics.",9122
"exposed Veterinarians, and are uniquely processors to producers emerging zoonoses.","Veterinarians, producers and processors are uniquely exposed to emerging zoonoses.",9123
models. approach combines This discrete continuous the and of dynamic features,This approach combines the features of continuous and discrete dynamic models.,9124
to We molecular of model investigation actomyosin motor. this apply,We apply this model to investigation of actomyosin molecular motor.,9125
using analytically system equations introduced wavelet theory. p-adic The is solved of,The introduced system of equations is solved analytically using p-adic wavelet theory.,9126
stationary regime. and explicit the relaxation We behavior in find solutions,We find explicit stationary solutions and behavior in the relaxation regime.,9127
adapt required methods flexibility such lack the changes. to Geometric to,Geometric methods lack the required flexibility to adapt to such changes.,9128
We training by address set. this re-sampling the,We address this by re-sampling the training set.,9129
model a term the form selection penalty a for We propose procedure. of,We propose a form of a penalty term for the model selection procedure.,9130
approach numerical our data. experiments real simulated and through We also illustrate using,We also illustrate our approach through numerical experiments using real and simulated data.,9131
"DLPP power In the exploited from of algorithm, discriminating further are our aspects. two","In our algorithm, the discriminating power of DLPP are further exploited from two aspects.",9132
We to our apply algorithm face recognition.,We apply our algorithm to face recognition.,9133
often is detection the in this Border analysis. first step,Border detection is often the first step in this analysis.,9134
Carlo algorithm. chain is Markov Monte inherently an serial,Markov chain Monte Carlo is an inherently serial algorithm.,9135
"approach is and The implement, directions research. and suggests simple additional easy for to further","The approach is simple and easy to implement, and suggests additional directions for further research.",9136
from retrieve objects challenge task. effectively with is To corpus accuracy large high a,To effectively retrieve objects from large corpus with high accuracy is a challenge task.,9137
segmentation. This automatic proposes method paper a retinal for vessel,This paper proposes a method for automatic retinal vessel segmentation.,9138
"For Morlet based vectors are this classification, organized feature on special to responses wavelet.","For this classification, special feature vectors are organized based on responses to Morlet wavelet.",9139
is classifier its Gaussian mixture model used likelihood is Bayesian function. and (GMM),Bayesian classifier is used and Gaussian mixture model (GMM) is its likelihood function.,9140
of the features rely Pattern in inherent images. upon the problem pattern recognition,Pattern recognition problem rely upon the features inherent in the pattern of images.,9141
regions from pixel this identified are Face region. skin,Face regions are identified from this skin pixel region.,9142
"located. such nose Facial and mouth features are then eyes, as","Facial features such as eyes, nose and mouth are then located.",9143
using recognized from Faces images an based color network. neural are RBF,Faces are recognized from color images using an RBF based neural network.,9144
means to algorithm with K different locate clustering used Automata is Cellular elements. facial Unsupervised,Unsupervised Cellular Automata with K means clustering algorithm is used to locate different facial elements.,9145
"inter mouth position, etc. Cosine nose coefficients like (DCT) length, Transform Discrete distance, Parameters eye","Parameters like inter eye distance, nose length, mouth position, Discrete Cosine Transform (DCT) coefficients etc.",9146
a and for computed Radial used based (RBF) are neural network. Basis Function,are computed and used for a Radial Basis Function (RBF) based neural network.,9147
"expressions head, sequence etc. in with This face orientation works for reliably approach","This approach reliably works for face sequence with orientation in head, expressions etc.",9148
state-of-the-arts. experimental approach Extensive demonstrate results outperform the that the proposed,Extensive experimental results demonstrate that the proposed approach outperform the state-of-the-arts.,9149
rate for Kramers theory analogous is chemical reactions. This to formula,This rate formula is analogous to Kramers theory for chemical reactions.,9150
limiting studied are the cases behaviors The two significantly resulting from different previously.,The resulting behaviors are significantly different from the two limiting cases studied previously.,9151
solving framework used for a problems. modeling Problem A constrained is and Constraint (CSP) Satisfaction,A Constraint Satisfaction Problem (CSP) is a framework used for modeling and solving constrained problems.,9152
proposed choosing for heuristics ordering. Various have been variable good,Various heuristics have been proposed for choosing good variable ordering.,9153
are global Constraint in employed weighting a CSP. to basically identify bottlenecks,Constraint weighting are basically employed to identify global bottlenecks in a CSP.,9154
is of network organization. on based principles The cortical,The network is based on principles of cortical organization.,9155
obtaining the about information for biological method atomic-scale is macromolecules. predominant crystallography X-ray,X-ray crystallography is the predominant method for obtaining atomic-scale information about biological macromolecules.,9156
"process crystallization the proceeds knowledge-informed through practice, empiricism. In","In practice, the crystallization process proceeds through knowledge-informed empiricism.",9157
The electrostatic other crystallization interactions the in identifies specific not previously context. described,The other identifies specific electrostatic interactions not previously described in the crystallization context.,9158
as are important of events indicators being undesired Unusual possible consequences.,Unusual events are important as being possible indicators of undesired consequences.,9159
"second utilizes cross-validation low-resolution the extra method and first measurements. measurements, one extra Our exploits","Our first method utilizes extra cross-validation measurements, and the second one exploits extra low-resolution measurements.",9160
age on of the Properties and synchrony status. depended disease,Properties of synchrony depended on the age and disease status.,9161
"of However, desynchronization episodes multiple regardless age prevailed disease and short status.","However, multiple short desynchronization episodes prevailed regardless of age and disease status.",9162
importance functional short implies of dynamics. desynchronizations This,This implies functional importance of short desynchronizations dynamics.,9163
"logistic fused learning for a We regression, sparse the present binary classification. multi-task approach","We present the fused logistic regression, a sparse multi-task learning approach for binary classification.",9164
et this In employ model paper Compte based of the we al. conductance,In this paper we employ the conductance based model of Compte et al.,9165
In we a NN/HMM hybrid handwriting propose work for recognition. Arabic online this model,In this work we propose a hybrid NN/HMM model for online Arabic handwriting recognition.,9166
decoded this The character to provide by HMMs recognition. is output network of level,The output of this network is decoded by HMMs to provide character level recognition.,9167
"echo-processing our mechanisms were of conducted investigated bottlenose in on The experiments, dolphins.","The mechanisms of echo-processing were investigated in our experiments, conducted on bottlenose dolphins.",9168
"revealed. perception, animals was describing organized Hierarchically of dimensions, system echoes independent in","Hierarchically organized system of independent dimensions, describing echoes in animals perception, was revealed.",9169
discrimination in established. of echoes dolphins recognition and were of The rules,The rules of discrimination and recognition of echoes in dolphins were established.,9170
"paper, both properties algorithms. systematically theoretical the we greedy analyze forward-backward of this In","In this paper, we systematically analyze the theoretical properties of both forward-backward greedy algorithms.",9171
"satisfy view algorithm and discriminability this of security, biometric template protection should In cancelability.","In view of this biometric template protection algorithm should satisfy security, discriminability and cancelability.",9172
face used evaluation. available benchmark databases are for publicly Three,Three publicly available benchmark face databases are used for evaluation.,9173
decision. for ensembled final the perspectives are Different,Different perspectives are ensembled for the final decision.,9174
paper been has Automatic targeted. authentication of money,Automatic authentication of paper money has been targeted.,9175
and recognition techniques used to the design overall processing are Image approach. pattern,Image processing and pattern recognition techniques are used to design the overall approach.,9176
The fake the is ability thoroughly embedded analysed currencies. aspects of detecting security for,The ability of the embedded security aspects is thoroughly analysed for detecting fake currencies.,9177
of reported system processing in The both accuracy and is performance speed. terms,The system performance is reported in terms of both accuracy and processing speed.,9178
are together. and Modeling fields that consciousness-related are of phenomena growing rapidly neuroengineering,Modeling of consciousness-related phenomena and neuroengineering are fields that are rapidly growing together.,9179
"significantly has progressed using decades. been Thus, recent the computer systems predicting in","Thus, using the predicting computer systems has been significantly progressed in recent decades.",9180
"system recognition proposed training the and In approach, in two modes: classification. operates","In proposed approach, the recognition system operates in two modes: training and classification.",9181
"the stage. In information concept are centre of system, compression alignment"" SP and ""multiple a","In the SP system, information compression and a concept of ""multiple alignment"" are centre stage.",9182
of use Jacobian achieved pseudo-inverse Matrix. The has been through matching of seamless the,The seamless matching has been achieved through the use of pseudo-inverse of Jacobian Matrix.,9183
stochastic high suffer variance implementations Naive this setting. in of from approximation,Naive implementations of stochastic approximation suffer from high variance in this setting.,9184
such how we show description bear Here insight. fruitful a can higher-level,Here we show how such a higher-level description can bear fruitful insight.,9185
intriguing We this and numerically phenomenon analyze analytically.,We analyze this intriguing phenomenon numerically and analytically.,9186
"a two-step has proposed. Here, approach been","Here, a two-step approach has been proposed.",9187
"text into contours. line is on line words First, information segmented based the using","First, the text line is segmented into words using information based on line contours.",9188
gap. First of word value are to used text blocks the order gradient the find,First order gradient value of the text blocks are used to find the word gap.,9189
"sent for binarized is to purpose. recognition Secondly, text OCR block this","Secondly, this binarized text block is sent to OCR for recognition purpose.",9190
in different Epigenetic important histone modifications an role phenotypes. of the maintenance play cell,Epigenetic histone modifications play an important role in the maintenance of different cell phenotypes.,9191
generations modification remains The mechanism the patterns exact inheritance of for cell over molecular elusive.,The exact molecular mechanism for inheritance of the modification patterns over cell generations remains elusive.,9192
central a notion theory. Exchangeability in statistics is probability and,Exchangeability is a central notion in statistics and probability theory.,9193
"renders probabilistic tractable finite a property less inference well-understood. exchangeability as statistical is that However,","However, finite exchangeability as a statistical property that renders probabilistic inference tractable is less well-understood.",9194
and inference. develop its to relation tractable probabilistic exchangeability finite We a theory of,We develop a theory of finite exchangeability and its relation to tractable probabilistic inference.,9195
is conditional independence. theory complementary The that and independence to of,The theory is complementary to that of independence and conditional independence.,9196
These straightforwardly concrete more abstract move to operations allow representations. from to,These operations allow to move straightforwardly from concrete to more abstract representations.,9197
dominant a assume cognition. role associative in rules that propose I,I propose that associative rules assume a dominant role in cognition.,9198
problem. concept the definition of for also a I propose,I also propose a definition for the concept of problem.,9199
for possible illustrate the briefly I At will systems. model end a cognitive general,At the end I will briefly illustrate a possible general model for cognitive systems.,9200
interest. of regions segmentation locate to we image an split mean By,By segmentation we mean split an image to locate regions of interest.,9201
"training a and an classes Here, we partition image basis. according define discriminate to","Here, we discriminate and define an image partition classes according to a training basis.",9202
"of training consists This urban basis classes: aquatic, vegetation and three pattern regions.","This training basis consists of three pattern classes: aquatic, urban and vegetation regions.",9203
been gestures mouse have performed. Thus various recognizing events the,Thus recognizing the gestures various mouse events have been performed.,9204
We multiscale then to optimal method distance tangent of the the solution. convergence the study,We then study the convergence of the multiscale tangent distance method to the optimal solution.,9205
distance of in performance method We the examine applications. tangent classification finally image the,We finally examine the performance of the tangent distance method in image classification applications.,9206
expectation-maximization also We solution. (EM) (point) for a MAP consider,We also consider expectation-maximization (EM) for a MAP (point) solution.,9207
The service. of process the by businessprocesses web web pages customizing composite by using,The process of customizing businessprocesses by composite the web pages by using web service.,9208
to DNA is transfer genome the host of Lateral Wolbachia common.,Lateral transfer of Wolbachia DNA to the host genome is common.,9209
sequence. with definitively were co-assembled fragments nematode-derived Many,Many fragments were co-assembled with definitively nematode-derived sequence.,9210
expression the found genes. evidence We of limited Wolbachia-derived of,We found limited evidence of expression of the Wolbachia-derived genes.,9211
recommendation a describe completely scale fashion. large automated visual for We system,We describe a completely automated large scale visual recommendation system for fashion.,9212
two methods importance on new We Monte based sampling. sequential Carlo propose,We propose two new Monte Carlo methods based on sequential importance sampling.,9213
computation. Parallel (PVM) Virtual use to Machine We up speed the,We use Parallel Virtual Machine (PVM) to speed up the computation.,9214
hallucinations speech absence perceptions Purpose: stimulation. in (AVHs) of verbal a external Auditory are the,Purpose: Auditory verbal hallucinations (AVHs) are speech perceptions in the absence of a external stimulation.,9215
"is a on predictive control The which account based self-monitoring in model, is implemented. verbal","The account is based on a predictive control model, in which verbal self-monitoring is implemented.",9216
during activity articulation) (without Lip muscle covert was and rest. AVHs Methods: recorded,Methods: Lip muscle activity was recorded during covert AVHs (without articulation) and rest.,9217
electromyography patients. eleven schizophrenia used (EMG) was on Surface,Surface electromyography (EMG) was used on eleven schizophrenia patients.,9218
This how AVHs understanding better occur. help will,This will help better understanding how AVHs occur.,9219
and sequence. from exhibit Cells phenotypes different often the DNA same stable,Cells often exhibit different and stable phenotypes from the same DNA sequence.,9220
epigenetic of for regulation. our implications deep These have understanding results,These results have deep implications for our understanding of epigenetic regulation.,9221
"an this paper, we modes in a present In histogram. algorithm automatically detect to meaningful","In this paper, we present an algorithm to automatically detect meaningful modes in a histogram.",9222
is based in scale-space proposed local representation. behavior of on method minima the The a,The proposed method is based on the behavior of local minima in a scale-space representation.,9223
"parameters. fast, require and does easy not to is The algorithm any implement,","The algorithm is easy to implement, fast, and does not require any parameters.",9224
details offer note we on to practice. this this implement algorithm how in efficiently In,In this note we offer details on how to efficiently implement this algorithm in practice.,9225
" an  for from We interactive algorithm learning present policy demonstration. Confidence-Based Autonomy (CBA),","  We present Confidence-Based Autonomy (CBA), an interactive algorithm for policy learning from demonstration.",9226
its in simulated compared and complex are individual and domain. CBA a components evaluated driving,CBA and its individual components are compared and evaluated in a complex simulated driving domain.,9227
extensions number the refinements also of are to introduced. and A basic algorithm,A number of refinements and extensions to the basic algorithm are also introduced.,9228
"including It scheduling, resource allocation, applied bioinformatics, been has etc. various problems, to","It has been applied to various problems, including resource allocation, bioinformatics, scheduling, etc.",9229
"domains. very these to solve problems However, techniques not with suited are well large","However, these techniques are not well suited to solve problems with very large domains.",9230
"complexity. improved arc soft asymptotic to BAC consistency, and Compared space significantly time provides","Compared to soft arc consistency, BAC provides significantly improved time and space asymptotic complexity.",9231
some well of results. as empirical evaluation the as We an complexity approach worst-case provide,We provide an empirical evaluation of the approach as well as some worst-case complexity results.,9232
results macros. This several presents for on based tractability planning new paper,This paper presents several new tractability results for planning based on macros.,9233
work of consider spatio-temporal In this videos. we behavior detecting the problem anomalous in,In this work we consider the problem of detecting anomalous spatio-temporal behavior in videos.,9234
reduction approach and is sparse estimate models. Our the parameter to using covariance,Our approach is to estimate the covariance using parameter reduction and sparse models.,9235
algorithms problem. We our to learning relevant propose,We propose learning algorithms relevant to our problem.,9236
that SP of shown solutions. estimates over covers clusters marginals that was It represent,It was shown that SP estimates marginals over covers that represent clusters of solutions.,9237
generalizes SP. naturally interpretation This cover the of,This naturally generalizes the cover interpretation of SP.,9238
on weighted RSP Max-SAT random solvers also Max-SAT instances. state-of-the-art weighted outperforms,RSP also outperforms state-of-the-art weighted Max-SAT solvers on random weighted Max-SAT instances.,9239
of our its demonstrate variations. and method all accuracy Experimental new results the,Experimental results demonstrate the accuracy of our new method and all its variations.,9240
consistency CP-nets. dominance We in computational investigate testing and the complexity of,We investigate the computational complexity of testing dominance and consistency in CP-nets.,9241
negative often and arguments. positive is matter a a and listing decision of Making comparing,Making a decision is often a matter of listing and comparing positive and negative arguments.,9242
"Prospect is Cumulative Theory. is in That done, example, for what","That is what is done, for example, in Cumulative Prospect Theory.",9243
"terms, decision well other qualitative is bipolar. In as the as process","In other terms, the decision process is qualitative as well as bipolar.",9244
proposed. rules former More that decisive refine the are also,More decisive rules that refine the former are also proposed.,9245
"modelling call includes which responsible the incompleteness process. process for we incompleteness, the This the","This includes modelling the process responsible for the incompleteness, which we call the incompleteness process.",9246
allow behaviour process We be partly to the unknown.,We allow the process behaviour to be partly unknown.,9247
known common hard. Multiagent be and to and planning are problems computationally coordination,Multiagent planning and coordination problems are common and known to be computationally hard.,9248
be wide programs. two-agent bilinear range formulated that as problems We a of show can,We show that a wide range of two-agent problems can be formulated as bilinear programs.,9249
the A evolutionary among species. tree shows relationships phylogenetic,A phylogenetic tree shows the evolutionary relationships among species.,9250
Internal represent correspond and the leaf nodes events tree to speciation of species. nodes,Internal nodes of the tree represent speciation events and leaf nodes correspond to species.,9251
rooted This inspires programming a for encoding trees. constraint,This inspires a constraint programming encoding for rooted trees.,9252
allows constraint-based an construction this We efficient that show supertree to the problem. solution,We show that this allows an efficient constraint-based solution to the supertree construction problem.,9253
of the simple of model model with a Soft goals classical extend preferences. planning,Soft goals extend the classical model of planning with a simple model of preferences.,9254
in has complexity the situations. operators resulting than approach other some much characteristics better Hence,Hence the resulting approach has much better complexity characteristics than other operators in some situations.,9255
"obtain yields operator. revision, compositional the most natural that the under We Satoh definition, revision","We obtain that compositional revision, under the most natural definition, yields the Satoh revision operator.",9256
magnitude variation was as donors. donors with unrelated in in greater matched This related compared,This variation was greater in magnitude in unrelated donors as compared with matched related donors.,9257
"piece-wise we prior. general First, using a constant methodology a use Bayesian Dirichlet","First, we use a piece-wise constant Bayesian methodology using a general Dirichlet prior.",9258
analysis. correlation with We linear these also compare,We also compare these with linear correlation analysis.,9259
this framework algorithm for We configuration problem. describe automatic an,We describe an automatic framework for this algorithm configuration problem.,9260
"procedures, algorithm achieved and improvements. using consistent substantial our performance Nevertheless, automated configuration we","Nevertheless, using our automated algorithm configuration procedures, we achieved substantial and consistent performance improvements.",9261
the analysis paper model. new for introduced co-sparse recently algorithm learning addresses a This,This paper addresses a new learning algorithm for the recently introduced co-sparse analysis model.,9262
"the to training, we technique called operators. a bi-level analysis For introduce learn optimization","For training, we introduce a technique called bi-level optimization to learn the analysis operators.",9263
to framework implement. a and intuitive easy develops is approach to understand Our that,Our approach develops a framework that is intuitive to understand and easy to implement.,9264
in reasoning. problem Inference in with important probabilistic applications numerous Nets Bayes is (BAYES) an,Inference in Bayes Nets (BAYES) is an important problem with numerous applications in probabilistic reasoning.,9265
"sum-of-products members these class (SUMPROD) problems, of and problems. the Both others, are of","Both these problems, and others, are members of the class of sum-of-products (SUMPROD) problems.",9266
is for optimization A search algorithm problems solving constraint distributed new presented. (DisCOPs),A new search algorithm for solving distributed constraint optimization problems (DisCOPs) is presented.,9267
variables assign compute sequentially assignments partial on Agents asynchronously. bounds and,Agents assign variables sequentially and compute bounds on partial assignments asynchronously.,9268
based bounds on the is partial asynchronous assignments. of computation The propagation,The asynchronous bounds computation is based on the propagation of partial assignments.,9269
described is in its The and detail algorithm proven. correctness,The algorithm is described in detail and its correctness proven.,9270
presented fully a resulting form. batch variational iterative automated first are The in schemes,The resulting fully automated variational schemes are first presented in a batch iterative form.,9271
"AI several technique has the become for recent advances, Thanks applications. underlying to Planning","Thanks to recent advances, AI Planning has become the underlying technique for several applications.",9272
current planning The tools. belief support update limited for in severely is,The support for belief update is severely limited in current planning tools.,9273
DEC-POMDPs. algorithm paper an contribution solving iteration main optimal of The policy this for is,The main contribution of this paper is an optimal policy iteration algorithm for solving DEC-POMDPs.,9274
to algorithm represent The policies. stochastic uses controllers finite-state,The algorithm uses stochastic finite-state controllers to represent policies.,9275
domains. We observable for identifying algorithms in deterministic-actions dynamic present partially exact and effects preconditions,We present exact algorithms for identifying deterministic-actions effects and preconditions in dynamic partially observable domains.,9276
Such real common in scenarios applications. are world,Such scenarios are common in real world applications.,9277
and assumptions Our from models. action observations traditional partial work about departs,Our work departs from traditional assumptions about partial observations and action models.,9278
modified such domains. algorithms for We for tractable yield the problem,We yield tractable algorithms for the modified problem for such domains.,9279
"exactly. identify tractability experiments verify guarantees, Our models theoretical action that we the and show","Our experiments verify the theoretical tractability guarantees, and show that we identify action models exactly.",9280
"use already Several adventure-game playing planning, and applications these exploration, in autonomous results.","Several applications in planning, autonomous exploration, and adventure-game playing already use these results.",9281
"diagnosis. also probabilistic They are for and reinforcement observable learning, partially promising settings,","They are also promising for probabilistic settings, partially observable reinforcement learning, and diagnosis.",9282
several to decision tasks real-world expensive choose among require us Many observations. making,Many real-world decision making tasks require us to choose among several expensive observations.,9283
It heuristic-guided observations. selecting been practice general has to for procedures use,It has been general practice to use heuristic-guided procedures for selecting observations.,9284
"label optimally example, Models hidden (HMMs). Hidden allow our Markov For variables to algorithms in","For example, our algorithms allow to optimally label hidden variables in Hidden Markov Models (HMMs).",9285
is for even polytrees. value optimizing of that prove $NP^{PP}$-hard We information the,We prove that the optimizing value of information is $NP^{PP}$-hard even for polytrees.,9286
of two the compiling model. We algorithms for a graphical provide AOMDD,We provide two algorithms for compiling the AOMDD of a graphical model.,9287
potential AOMDDs. We demonstrates the that provide evaluation an experimental of,We provide an experimental evaluation that demonstrates the potential of AOMDDs.,9288
"paper, distance implement plant a researched were this system. In to several foliage retrieval measures","In this paper, several distance measures were researched to implement a foliage plant retrieval system.",9289
and chart future We experimental results work. provide,We provide experimental results and chart future work.,9290
"dimensionality techniques these problems. scale impractical large the on Nevertheless, curse makes of","Nevertheless, the curse of dimensionality makes these techniques impractical on large scale problems.",9291
focus of consisting cycle. on of one single the Here permutations we variables,Here we focus on permutations of the variables consisting of one single cycle.,9292
them as a analyze We function classes and these of cube the quantify dimensionality.,We analyze these classes and quantify them as a function of the cube dimensionality.,9293
Distributed agent-coordination constraint problems optimization problems. a popular and of are way solving (DCOP) formulating,Distributed constraint optimization (DCOP) problems are a popular way of formulating and solving agent-coordination problems.,9294
It often algorithms. desirable DCOP with to solve and asynchronous problems memory-bounded is,It is often desirable to solve DCOP problems with memory-bounded and asynchronous algorithms.,9295
of the DisCSP. subproblems search algorithm partitions into the different The,The algorithm partitions the search into different subproblems of the DisCSP.,9296
"algorithm, version Asynchronous resulting completeness. of the its Complete The ensures Overlay Partial (CompAPO),","The resulting version of the algorithm, Complete Asynchronous Partial Overlay (CompAPO), ensures its completeness.",9297
Formal for of proofs soundness and CompAPO are the given. completeness,Formal proofs for the soundness and completeness of CompAPO are given.,9298
"of algorithm, the and discussed, CompOptAPO, an is optimization Additionally, evaluated. presented, version","Additionally, an optimization version of the algorithm, CompOptAPO, is presented, discussed, and evaluated.",9299
as this The main thesis follows. of contributions are,The main contributions of this thesis are as follows.,9300
useful information the to provide uses technique SAR blanket image method for classification. This,This method uses the blanket technique to provide useful information for SAR image classification.,9301
"a classification for technique different five terrain As result, a types, i.e.","As a result, a classification technique for five different terrain types, i.e.",9302
"and is suburban, urban, in mountain presented sea, this paper. rural,","urban, suburban, rural, mountain and sea, is presented in this paper.",9303
"is solving except to often problems due However, a for intractable small their complexity. POMDP","However, solving a POMDP is often intractable except for small problems due to their complexity.",9304
"storytelling, particular an and of the in important part human Narrative, is experience.","Narrative, and in particular storytelling, is an important part of the human experience.",9305
success to its is Central this understandability.,Central to this success is its understandability.,9306
"intentional the be agents. perceived must to by audience characters be Specifically,","Specifically, characters must be perceived by the audience to be intentional agents.",9307
to a with is satisfiability eminent Planning principal planning many approach advantages. as,Planning as satisfiability is a principal approach to planning with many eminent advantages.,9308
encodings STRIPS. The as usually use satisfiability from existing compiled techniques planning,The existing planning as satisfiability techniques usually use encodings compiled from STRIPS.,9309
the (SASE) SAT scheme encoding formalism. on a We based SAS+ introduce novel,We introduce a novel SAT encoding scheme (SASE) based on the SAS+ formalism.,9310
statistical to results We analysis. empirical support give our,We give empirical statistical results to support our analysis.,9311
paper develops make algorithm that the FODD improvements approach This to practical. the several,This paper develops several improvements to the FODD algorithm that make the approach practical.,9312
"planning system, relational planning paper a presents the Incorporating for solving these, FODD-Planner, stochastic problems.","Incorporating these, the paper presents a planning system, FODD-Planner, for solving relational stochastic planning problems.",9313
its in The comparison. problem has genome application,The problem has its application in genome comparison.,9314
"to solutions a MAX-MIN quality ant the employ we system achieve problem. for high Subsequently,","Subsequently, we employ a MAX-MIN ant system to achieve high quality solutions for the problem.",9315
by is achieved test. statistical improvement The justified also standard,The improvement achieved is also justified by standard statistical test.,9316
cortex in debated. highly orientation emergence primary the visual underlying of selectivity the Mechanisms are,Mechanisms underlying the emergence of orientation selectivity in the primary visual cortex are highly debated.,9317
"accounts perform a which for mean-field analysis Here, this process. we","Here, we perform a mean-field analysis which accounts for this process.",9318
The problem cycle an (HCP) important problem many in combinatorial Hamiltonian with is areas. applications,The Hamiltonian cycle problem (HCP) is an important combinatorial problem with applications in many areas.,9319
main this work contribution algorithm the DHCP. an is of effective The for,The main contribution of this work is an effective algorithm for the DHCP.,9320
graphs key task network mapping is of data. complex the of in Visual a rendering,Visual rendering of graphs is a key task in the mapping of complex network data.,9321
centrality-constrained determined the embeddings. optimization yielding to solutions meaningful are problems Closed-form resulting,Closed-form solutions to the resulting centrality-constrained optimization problems are determined yielding meaningful embeddings.,9322
hypotheses. leads to diagnostic number Model-based reasoning a large often diagnostic of,Model-based diagnostic reasoning often leads to a large number of diagnostic hypotheses.,9323
descriptions multimodal action in work logic. about The domain is present changing,The present work is about changing action domain descriptions in multimodal logic.,9324
"method the be intensive. can Consequently, computationally extremely","Consequently, the method can be extremely computationally intensive.",9325
for We expensive that to are proposals avoid rejected. therefore be calculations likely,We therefore avoid expensive calculations for proposals that are likely to be rejected.,9326
"additional comes model Our efficient. with is that advantage, the trained inference extremely along","Our trained model comes along with the additional advantage, that inference is extremely efficient.",9327
candidates: This family concentrates paper search. of counting-based on one,This paper concentrates on one family of candidates: counting-based search.,9328
other search generally heuristics. results that show counting-based Experimental outperforms generic,Experimental results show that counting-based search generally outperforms other generic heuristics.,9329
"this problem In in an paper, system. address the we argumentation abstract change of","In this paper, we address the problem of change in an abstract argumentation system.",9330
more according may applications. specific these less properties desirable All to or be,All these properties may be more or less desirable according to specific applications.,9331
preferred are semantics: the and semantics. They studied under grounded two particular,They are studied under two particular semantics: the grounded and preferred semantics.,9332
image produces output. fused Applying to the algorithm final deconvolution the blind a,Applying a blind deconvolution algorithm to the fused image produces the final output.,9333
the presence in discuss We then cost functions. of configuration multiple interactive,We then discuss interactive configuration in the presence of multiple cost functions.,9334
instances. of even fraction a are very times second generally within Response for large a,Response times are generally within a fraction of a second even for very large instances.,9335
methods usefulness by of The the numerical are illustrated results. thorough proposed,The usefulness of the proposed methods are illustrated by thorough numerical results.,9336
research. active act area observable partially remains to how an of Deciding in environments,Deciding how to act in partially observable environments remains an active area of research.,9337
"problem, been this to have overcome In order approaches proposed. many","In order to overcome this problem, many approaches have been proposed.",9338
"Among ones. are variants and them, ILAO* of RTDP state-of-the-art","Among them, ILAO* and variants of RTDP are state-of-the-art ones.",9339
analysis reachability heuristic some search methods backups. These avoid unnecessary and use to,These methods use reachability analysis and heuristic search to avoid some unnecessary backups.,9340
"MDP we algorithms. this paper, optimal present two In","In this paper, we present two optimal MDP algorithms.",9341
"other close-to-equal-sized when an algorithms TVI multiple, outperforms has VI state-of-the-art MDP SCCs. and vastly","TVI outperforms VI and other state-of-the-art algorithms vastly when an MDP has multiple, close-to-equal-sized SCCs.",9342
"topological value extension of second TVI. algorithm, an iteration The is focused (FTVI),","The second algorithm, focused topological value iteration (FTVI), is an extension of TVI.",9343
its for MDP. FTVI components connected that to relevant attention restricts are solving the,FTVI restricts its attention to connected components that are relevant for solving the MDP.,9344
"magnitude, of domains. averaged order demonstrate by FTVI outperforms that an across We several TVI","We demonstrate that FTVI outperforms TVI by an order of magnitude, averaged across several domains.",9345
or advances activities. to in lead of completion delays,lead to delays or advances in completion of activities.,9346
"to that between provide we dependencies enhancements search Furthermore, local exploit also activities. temporal","Furthermore, we also provide enhancements to local search that exploit temporal dependencies between activities.",9347
"prune variables a able its is explanation. result, to MRE less As relevant from automatically","As a result, MRE is able to automatically prune less relevant variables from its explanation.",9348
heuristic to planning for expensive functions Current are evaluation compute.,Current evaluation functions for heuristic planning are expensive to compute.,9349
policy. as The first action classifier resulting using of consists the one an,The first one consists of using the resulting classifier as an action policy.,9350
"Drake, with work executive This for dynamic temporal plans a choice. presents","This work presents Drake, a dynamic executive for temporal plans with choice.",9351
"on of representation non-dominated maintaining maintenance the minimal a focuses system particular, In constraints.","In particular, the maintenance system focuses on maintaining a minimal representation of non-dominated constraints.",9352
"American-style program Dr.Fill, We solves puzzles. that crossword describe a","We describe Dr.Fill, a program that solves American-style crossword puzzles.",9353
formally then are and studied induced The empirically. admissible heuristics,The induced admissible heuristics are then studied formally and empirically.,9354
We on networks. large-scale experimentally Bayesian real techniques randomly generated evaluated our and,We experimentally evaluated our techniques on large-scale real and randomly generated Bayesian networks.,9355
"up as As they problems become larger. well scale result, a","As a result, they scale up well as problems become larger.",9356
the parameter Estimation methodology. crucial is transformation this in of,Estimation of the transformation parameter is crucial in this methodology.,9357
An is for covariate comparative purposes. method also artificial included,An artificial covariate method is also included for comparative purposes.,9358
to the are Simulation studies proposed the of implemented methods. compare effectiveness,Simulation studies are implemented to compare the effectiveness of the proposed methods.,9359
methods are data illustrated on real life also The two different applications.,The methods are also illustrated on two different real life data applications.,9360
"for an R Moreover, package proposed implementation. AID is","Moreover, an R package AID is proposed for implementation.",9361
several probabilistic rules model a relational world promising for are representation Noisy reasons.,Noisy probabilistic relational rules are a promising world model representation for several reasons.,9362
over world instantiations. compact and generalize are They,They are compact and generalize over world instantiations.,9363
grounded rules We relational in domains. such investigate reasoning with,We investigate reasoning with such rules in grounded relational domains.,9364
efficient the for and decision-theoretic compactness of flexible rules algorithms Our planning. exploit,Our algorithms exploit the compactness of rules for efficient and flexible decision-theoretic planning.,9365
Empirical our methods methods that results fail. can where solve existing problems show,Empirical results show that our methods can solve problems where existing methods fail.,9366
execution system Our and preferences. handles failures multi-objective,Our system handles execution failures and multi-objective preferences.,9367
problem throughout important and machine features structure in are planning. learning representing decision-theoretic Domain-specific,Domain-specific features are important in representing problem structure throughout machine learning and decision-theoretic planning.,9368
features often crafted in domains applications experts. real-world human Successful by require,Successful applications in real-world domains often require features crafted by human experts.,9369
first Our successfully domain-independent human-engineered provides (without that plays features). the Tetris approach method,Our method provides the first domain-independent approach that plays Tetris successfully (without human-engineered features).,9370
some relation. We establish dominance key of properties the,We establish some key properties of the dominance relation.,9371
negative different give properties results results. also complexity few and underlie novel which a some,different properties and also give a few novel complexity results which underlie some negative results.,9372
optimization differentiation the We implicit the bi-level solve using technique. problem,We solve the bi-level optimization problem using the implicit differentiation technique.,9373
"of priors. Under of framework, compare we the unified the performance two can this types","Under this unified framework, we can compare the performance of the two types of priors.",9374
"computationally I-DIDs expect, solving As we is hard. may","As we may expect, solving I-DIDs is computationally hard.",9375
methods these into aggregating equivalence models involve Both individual classes.,Both these methods involve aggregating individual models into equivalence classes.,9376
"we equivalent identical time actions single that cluster at prescribe step. actionally a models Specifically,","Specifically, we cluster actionally equivalent models that prescribe identical actions at a single time step.",9377
to us identifying the require set. would initial Exactly models all the solve in equivalences,Exactly identifying the equivalences would require us to solve all models in the initial set.,9378
"models, an We solving approximation. of this avoid selectively by some the introducing thereby","We avoid this by selectively solving some of the models, thereby introducing an approximation.",9379
"Not are domains commonly dynamic, Markov where used. in however, stochastic processes so, decision","Not so, however, in dynamic, stochastic domains where Markov decision processes are commonly used.",9380
domains and exploit. method which the can of assess We our features analyse,We assess and analyse the features of domains which our method can exploit.,9381
and delete both for planning. lists relaxation optimal paramount The importance satisficing ignoring of is,The ignoring delete lists relaxation is of paramount importance for both satisficing and optimal planning.,9382
establish between connections topology. causal structure and graph We h+,We establish connections between causal graph structure and h+ topology.,9383
"methods, call we This polynomial results TorchLight. in time a analysis low-order implemented tool in","This results in low-order polynomial time analysis methods, implemented in a tool we call TorchLight.",9384
"domains way, can distinguish TorchLight ones. hard easy from In this","In this way, TorchLight can distinguish easy domains from hard ones.",9385
path planning problem is challenging real-life with Multi-agent a numerous applications.,Multi-agent path planning is a challenging problem with numerous real-life applications.,9386
"tractable algorithm introduce on MAPP, a undirected for multi-agent graphs. path We planning","We introduce MAPP, a tractable algorithm for multi-agent path planning on undirected graphs.",9387
basic version extensions. and a several We present,We present a basic version and several extensions.,9388
on grid were game Experiments maps. run realistic,Experiments were run on realistic game grid maps.,9389
be across can instances re-used map. Parts on same MAPPs computation of the,Parts of MAPPs computation can be re-used across instances on the same map.,9390
reasoning to investigate an approach through argumentation. about causes We,We investigate an approach to reasoning about causes through argumentation.,9391
"a about consider model We causal system, look for a facts. arguments and physical for","We consider a causal model for a physical system, and look for arguments about facts.",9392
work Argumentation explaining from various then links. at these is,Argumentation is then at work from these various explaining links.,9393
based search. LAMA on is classical a heuristic system forward planning,LAMA is a classical planning system based on heuristic forward search.,9394
We into that transformed the spatial be can visual observe domain/task-specific domain/task-independent perceptions/actions problems.,We observe that the domain/task-specific visual perceptions/actions can be transformed into domain/task-independent spatial problems.,9395
The the efficacy idea complexity shown of is analysis. by time,The efficacy of the idea is shown by time complexity analysis.,9396
of constraint formulations problem. We present programming the several,We present several constraint programming formulations of the problem.,9397
Boolean linear present maximum formulations satisfiability also and weighted using partial programming. We integer mixed,We also present formulations using partial weighted maximum Boolean satisfiability and mixed integer linear programming.,9398
on discuss We different coevolutionary effect the the deletion mechanisms spacer of dynamics.,We discuss the effect of different spacer deletion mechanisms on the coevolutionary dynamics.,9399
lineage are rates. a genes restricted and to Orphans emerge at phylogenetic high single,Orphans are genes restricted to a single phylogenetic lineage and emerge at high rates.,9400
been not paradox resolved. has This yet,This paradox has not yet been resolved.,9401
"to be older more Interestingly, ones. are recently likely orphans than lost emerged","Interestingly, recently emerged orphans are more likely to be lost than older ones.",9402
"strong highly male-bias likely with to Furthermore, are orphans expressed be a retained. more","Furthermore, highly expressed orphans with a strong male-bias are more likely to be retained.",9403
important task an in processing. Thresholding image is,Thresholding is an important task in image processing.,9404
"image is recognition, in and detection edge segmentation, tool analysis. a It main scene pattern","It is a main tool in pattern recognition, image segmentation, edge detection and scene analysis.",9405
"we Tsallis In present this based paper, thresholding a on new entropy. technique two-dimensional","In this paper, we present a new thresholding technique based on two-dimensional Tsallis entropy.",9406
method the result that achieve the than proposed Experimental method. Shannon results better demonstrate,Experimental results demonstrate that the proposed method achieve better result than the Shannon method.,9407
We calculate networks cancer tissue in gene of breast and hierarchy patients. measures of,We calculate measures of hierarchy in gene and tissue networks of breast cancer patients.,9408
"and face LPQ, for tLBP, identification LTP, verification both face tasks. LBP, and LGXP) POEM,","LBP, LTP, LPQ, POEM, tLBP, and LGXP) for both face identification and face verification tasks.",9409
solutions. being is It in extensively used security,It is being used extensively in security solutions.,9410
This reduced-rank scheme novel process proposes paper regression. for a Gaussian,This paper proposes a novel scheme for reduced-rank Gaussian process regression.,9411
"with a Indeed, it model parameters. few produces additional dependency only a meaningful","Indeed, it produces a meaningful dependency model with only a few additional parameters.",9412
"likelihood, by maximum algorithm. by of means The an estimated, are EM parameters","The parameters are estimated, by maximum likelihood, by means of an EM algorithm.",9413
biological relevance on the and sets model. medical data new applications show Two this of,Two applications on medical and biological data sets show the relevance of this new model.,9414
"much have literature, systems been playing AI one limited In the game. to of","In much of the literature, AI systems have been limited to playing one game.",9415
proposed This of completion by model us. is a an paper earlier,This paper is a completion of an earlier model proposed by us.,9416
connections. that intracellular realized should these tubulin We a role play in,We realized that tubulin should play a role in these intracellular connections.,9417
the of tubulin switch configuration. switch the is This way realized dimer by the fast,This way the fast switch is realized by the switch of the tubulin dimer configuration.,9418
"search timetabling. course applied to algorithm harmony In is this paper, curriculum-based","In this paper, harmony search algorithm is applied to curriculum-based course timetabling.",9419
location the be new will in any available select solution. harmony feasible scheduled to,will select any feasible location available to be scheduled in the new harmony solution.,9420
variable-selection in of investigates global regression paper make analysis. use procedures This sensitivity that,This paper investigates variable-selection procedures in regression that make use of global sensitivity analysis.,9421
dealing in of used many with uncertainty. The is notion situations causality,The notion of causality is used in many situations dealing with uncertainty.,9422
network enhance use plate capability. license to computing vehicle recognition neural its Most techniques,Most vehicle license plate recognition use neural network techniques to enhance its computing capability.,9423
An adapted solution algorithm management system. is for into the parking,An algorithm is adapted into the solution for parking management system.,9424
concept solution algorithm. of to as then is implemented The proof the,The solution then is implemented as proof of concept to the algorithm.,9425
"computational as on processes treating natural phenomena the within info-computational informational of structures. constructivism, framework","within the framework of info-computational constructivism, treating natural phenomena as computational processes on informational structures.",9426
fusion image pyramid paper transform wavelet based on using and This method. laplacian is,This paper is based on image fusion using laplacian pyramid and wavelet transform method.,9427
size Images experimentation. of for same used are,Images of same size are used for experimentation.,9428
"local Immigration to populations a stabilise rescue metapopulation. can from helping extinction,","Immigration can rescue local populations from extinction, helping to stabilise a metapopulation.",9429
for real-time hard have problems search used Heuristics regions with depressions. solving,Heuristics used for solving hard real-time search problems have regions with depressions.,9430
"depressed do however, search Those algorithms, not guide towards regions. avoiding","Those algorithms, however, do not guide search towards avoiding depressed regions.",9431
which depression mark-and-avoid implemented: ways can avoidance in be move-to-border. propose We and two,We propose two ways in which depression avoidance can be implemented: mark-and-avoid and move-to-border.,9432
computer is common of concept in a science. objects representations Compact,Compact representations of objects is a common concept in computer science.,9433
"always plans where Some representations, are compact for example, have macros. known using cases the","Some cases are known where the plans always have compact representations, for example, using macros.",9434
relevant are results finally discussed in relation contexts. to other Our,Our results are finally discussed in relation to other relevant contexts.,9435
classification. of is universal interest objects for framework locating This a through,This is a universal framework for locating objects of interest through classification.,9436
critical modeling a applications. component Background various is vision-based for,Background modeling is a critical component for various vision-based applications.,9437
solving large-scale problems. be when traditional to tend inefficient methods Most,Most traditional methods tend to be inefficient when solving large-scale problems.,9438
our (SOIR) two Iterative algorithm. combine Outlier form to parts The Sparse Removal,The two parts combine to form our Sparse Outlier Iterative Removal (SOIR) algorithm.,9439
the in videos. space data natural algorithm of The obey structure tensor operates to,The algorithm operates in tensor space to obey the natural data structure of videos.,9440
determine results discriminative of performance that show a background few the extraction. the frames Experimental,Experimental results show that a few discriminative frames determine the performance of the background extraction.,9441
"has Thus, in advantage solving large-scale an SOIR tasks.","Thus, SOIR has an advantage in solving large-scale tasks.",9442
of studied. is A majorizing image approximation of approximations hierarchical by optimal sequence sequence,A majorizing of optimal approximation sequence by hierarchical sequence of image approximations is studied.,9443
also is different give NP an Co-NP. from simpler We that alternative and assuming proof,We also give an alternative and simpler proof assuming that NP is different from Co-NP.,9444
the constraints. encoding This give allows polynomial arc-consistent ROBDD-based us for to generalized first Pseudo-Boolean,This allows us to give the first polynomial generalized arc-consistent ROBDD-based encoding for Pseudo-Boolean constraints.,9445
package. describes SPRINT in R bootstrapping This inclusion for the parallelization the of paper,This paper describes the parallelization of bootstrapping for inclusion in the SPRINT R package.,9446
and multiple batteries with Efficient problem practical wide use is application. growing of a,Efficient use of multiple batteries is a practical problem with wide and growing application.,9447
cast under a The can be uncertainty. as problem problem planning,The problem can be cast as a planning problem under uncertainty.,9448
dimension. This depends of the temporal discretisation carefully approach on handled,This approach depends on carefully handled discretisation of the temporal dimension.,9449
and approach are its described the in The paper. generality,The approach and its generality are described in the paper.,9450
"detection image-components, data. methods, poorly on connected based very this Heuristic perform on","Heuristic detection methods, based on connected image-components, perform very poorly on this data.",9451
"transform. here, method is Hough introduced An based alternative, on geometrically-principled the","An alternative, geometrically-principled method is introduced here, based on the Hough transform.",9452
"axis-aligned coordinates. transform Hough A projective the in of each to two applied is clusters,","A projective Hough transform is applied to each of the two clusters, in axis-aligned coordinates.",9453
"sweeping line by through dual the transform. easily detected, is This a pattern","This pattern is easily detected, by sweeping a dual line through the transform.",9454
an of based error. geometric on photometric conclusion is analysis and This both,This conclusion is based on an analysis of both geometric and photometric error.,9455
and of understood. is origin development The consciousness poorly,The origin and development of consciousness is poorly understood.,9456
to steps its evolution. in be propose taken that I two had Here,Here I propose that two steps had to be taken in its evolution.,9457
"evolved an First, fitness as goal-directedness fitness. systems intrinsic by self-estimated living internalizing a Darwinian","First, living systems evolved an intrinsic goal-directedness by internalizing Darwinian fitness as a self-estimated fitness.",9458
consciousness. form in primary change a subjectively is as intrinsic of meaning The resulting experienced,The resulting change in intrinsic meaning is experienced subjectively as a primary form of consciousness.,9459
tools spherical (SD). Among such deconvolution is,Among such tools is spherical deconvolution (SD).,9460
from Playing Game matches and modules. Quality slow state-switching knowledge General suffers weak (GGP) of,Quality of General Game Playing (GGP) matches suffers from slow state-switching and weak knowledge modules.,9461
easily to mGDL to map rules C++ allows functions.,mGDL allows to easily map rules to C++ functions.,9462
of processing. plays contemporary vital world Image identification the role a Character in,Character identification plays a vital role in the contemporary world of Image processing.,9463
It many work makes and easier. can composite solve problems humans,It can solve many composite problems and makes humans work easier.,9464
Tamil system Handwritten Character to uses the Neural Networks them. recognition distinguish,Tamil Handwritten Character recognition system uses the Neural Networks to distinguish them.,9465
structural used characters. Neural characteristics written Network recognize and to are instruct and,Neural Network and structural characteristics are used to instruct and recognize written characters.,9466
"online via approach is in for MDPs. recent, Replanning a popular determinization planning","Replanning via determinization is a recent, popular approach for online planning in MDPs.",9467
as to safe We as do so. it this is plan execute long,We execute this plan as long as it is safe to do so.,9468
"longer this case, we no the When replan. is","When this is no longer the case, we replan.",9469
techniques is for sampling Our of overcoming contribution introduction main these the two state problems.,Our main contribution is the introduction of state sampling techniques for overcoming these two problems.,9470
into system modern face focal vigorous enhanced the skill In have detection time point. the,In modern time the skill have enhanced face detection system into the vigorous focal point.,9471
toughness on to of focuses mainly paper lighting the This variations. issue,This paper focuses mainly on the issue of toughness to lighting variations.,9472
"as we In first an article, this model. linear programming the problem formulate integer","In this article, we first formulate the problem as an integer linear programming model.",9473
"we the branch-and-bound solve a Next, problem. devise to algorithm","Next, we devise a branch-and-bound algorithm to solve the problem.",9474
"some are of different parameter the impacts Finally, additional experiments. disclosed by settings","Finally, the impacts of different parameter settings are disclosed by some additional experiments.",9475
the in Domain-independent is of foundational of Intelligence. areas the planning field Artificial one,Domain-independent planning is one of the foundational areas in the field of Artificial Intelligence.,9476
"a a planning difficult question. choose which task for is given to heuristic Thus,","Thus, which heuristic to choose for a given planning task is a difficult question.",9477
motivation the for take. characters they that actions This means have perceivable,This means that characters have perceivable motivation for the actions they take.,9478
computation equilibrations. in tropical methods these The is of the crux tropicalization of,The crux of these tropicalization methods is in the computation of tropical equilibrations.,9479
powerful Pan-tilt-zoom to far-field in object recognition support are scenes. cameras (PTZ) and identification,Pan-tilt-zoom (PTZ) cameras are powerful to support object identification and recognition in far-field scenes.,9480
"extracted visual directly of the frames. calibration accurate the content So, from be must","So, accurate calibration must be directly extracted from the visual content of the frames.",9481
calls computation maximum. the at approximation likelihood Laplace second derivatives The for of the,The Laplace approximation calls for the computation of second derivatives at the likelihood maximum.,9482
that the study arise sectorisation. of constraints airspace We a make in,We make a study of the constraints that arise in airspace sectorisation.,9483
"transitions. subsampling We and by parallel tempering present algorithms, tempering tempered subsampled subsampled two","We present two tempering by subsampling algorithms, subsampled parallel tempering and subsampled tempered transitions.",9484
framework This the common trial. demonstrated on procedure spontaneous is a critical breathing care: in,This framework is demonstrated on a common procedure in critical care: the spontaneous breathing trial.,9485
"had models simple focused to be of development on interactions. Previous very few, regulatory","Previous models of development had to be focused on very few, simple regulatory interactions.",9486
tasks. recognition preferred in and codebooks visual many are discriminative visual Compact,Compact and discriminative visual codebooks are preferred in many visual recognition tasks.,9487
are of in this them specifically work. Three focused,Three of them are specifically focused in this work.,9488
being One of the is human upright. walk defining ability of to characteristic their,One of the defining characteristic of human being is their ability to walk upright.,9489
is features. having in functions modern machine treadmill automated built and an A predefined,A modern treadmill is an automated machine having built in functions and predefined features.,9490
extract Remote performed image sensing classification meaningful can features. be to many in different ways,Remote sensing image classification can be performed in many different ways to extract meaningful features.,9491
to common approach perform One detection. edge is,One common approach is to perform edge detection.,9492
"be results, can approaches get combined. two To these optimal","To get optimal results, these two approaches can be combined.",9493
an important problem remote Feature selection the classification. in optimization sensing is image combinatorial,Feature selection is an important combinatorial optimization problem in the remote sensing image classification.,9494
"in solution sensing our is However, the to more remote image. robust illumination changes","However, our solution is more robust to illumination changes in the remote sensing image.",9495
balancing of simulations ancient our confirm of the Coalescent approximations. accuracy selection,Coalescent simulations of ancient balancing selection confirm the accuracy of our approximations.,9496
We epidemic an $M$ population propagation study between centra.,We study an epidemic propagation between $M$ population centra.,9497
migration becomes with parameters. epidemic a This in model evident vanishing purely,This becomes evident in a purely migration model with vanishing epidemic parameters.,9498
process article for data In Lasik suggested. the for this classifier a is best getting,In this article a process for getting the best classifier for Lasik data is suggested.,9499
quantification limits is limits Biomarker and/or subject often detection. of to of data,Biomarker data is often subject to limits of quantification and/or limits of detection.,9500
"to left- corresponds this Statistically, or interval-censoring.","Statistically, this corresponds to left- or interval-censoring.",9501
"in variable and high-dimensional is about setting. paper estimation This an unsupervised clustering selection,","This paper is about variable selection, clustering and estimation in an unsupervised high-dimensional setting.",9502
"location and expression, rearrangements, gene formatting controls and genetic the genetic of repair. timing DNA","DNA formatting controls the timing and location of genetic rearrangements, gene expression, and genetic repair.",9503
is of the of these circuits. events control cellular precise under Each,Each of these events is under the control of precise cellular circuits.,9504
computer are variation representations of in pervasive appearance Linear vision. subspace,Linear subspace representations of appearance variation are pervasive in computer vision.,9505
"barriers. the structural plausible diffusion organization Additionally, explore of we","Additionally, we explore the structural organization of plausible diffusion barriers.",9506
a stronger of form We show the result. above,We show a stronger form of the above result.,9507
The clever tricks of induction. proof employs,The proof employs clever tricks of induction.,9508
networks. gene/transcription used signaling simulate The is factor to dynamics asymmetric Hopfield model in,The asymmetric Hopfield model is used to simulate signaling dynamics in gene/transcription factor networks.,9509
into a direct gene of model states. allows pattern expression for attractor a The mapping,The model allows for a direct mapping of a gene expression pattern into attractor states.,9510
and subject. literature on the review paper a comprehensive this timely presents This of,This paper presents a comprehensive and timely review of the literature on this subject.,9511
for (DBNs) Bayesian observed networks processes with are a states. Dynamic partially stochastic model general,Dynamic Bayesian networks (DBNs) are a general model for stochastic processes with partially observed states.,9512
DBNs inferring (i.e. is state Belief the belief filtering of task in the,Belief filtering in DBNs is the task of inferring the belief state (i.e.,9513
observations. over and probability process based incomplete distribution the states) on noisy,the probability distribution over process states) based on incomplete and noisy observations.,9514
large with spaces. can complex processes state hard in This be problem a,This can be a hard problem in complex processes with large state spaces.,9515
"constraint the In inference lifted this language. from we article, decouple","In this article, we decouple lifted inference from the constraint language.",9516
each to the assigning Scene of is object problem labeling an pixel. label,Scene labeling is the problem of assigning an object label to each pixel.,9517
the and It segmentation problems. image object recognition unifies,It unifies the image segmentation and object recognition problems.,9518
horse demonstrate outperforms state-of-the-art CHM that Stanford and on datasets. Weizmann background We,We demonstrate that CHM outperforms state-of-the-art on Stanford background and Weizmann horse datasets.,9519
to Major were needed play in Answering Watson IBM Question advances for technology Jeopardy!,Major advances in Question Answering technology were needed for IBM Watson to play Jeopardy!,9520
framework seen can general be as abstract Dung's for non-monotonic theory argumentation reasoning. a,Dung's abstract argumentation theory can be seen as a general framework for non-monotonic reasoning.,9521
maximal to an function of knowledge the which consistent inconsistent returns base. subsets,to the function which returns maximal consistent subsets of an inconsistent knowledge base.,9522
also are attack define relations of which members We new class. this,We also define new attack relations which are members of this class.,9523
problems costs) participating (or produce different gains multi-agent constraints agents. Many include for the that,Many multi-agent problems include constraints that produce different gains (or costs) for the participating agents.,9524
gains model. cannot naturally constrained represented the of by be Asymmetric agents standard DCOP,Asymmetric gains of constrained agents cannot be naturally represented by the standard DCOP model.,9525
present DCOPs general a proposes paper (ADCOPs). The framework for Asymmetric,The present paper proposes a general framework for Asymmetric DCOPs (ADCOPs).,9526
in. constraints for ADCOPs valuations they may agents different In that have are involved different,In ADCOPs different agents may have different valuations for constraints that they are involved in.,9527
present converge quality The to local high the paper proposed search solutions. in algorithms,The local search algorithms proposed in the present paper converge to high quality solutions.,9528
call highly the We effect this `arrival non-ergodic of the frequent'.,We call this highly non-ergodic effect the `arrival of the frequent'.,9529
near are: spectrum. channels light infrared spectrum visible The the and,The channels are: the visible light spectrum and near infrared spectrum.,9530
Nowadays very techniques are image and decompression compression much digital important.,Nowadays digital image compression and decompression techniques are very much important.,9531
"objects Image locate to typically and segmentation etc. is (lines, boundaries used curves","Image segmentation is typically used to locate objects and boundaries (lines, curves etc.",9532
changed something is the analyze. is which segmentation After into to image more meaningful,After segmentation the image is changed into something which is more meaningful to analyze.,9533
"an measures. in paper parameter-free consider compression-based Hence, we distance this alternative approach:","Hence, in this paper we consider an alternative parameter-free approach: compression-based distance measures.",9534
"p-Sulu We problem. a the planning solve then which Planner, CCQSP develop can tractably","We then develop the p-Sulu Planner, which can tractably solve a CCQSP planning problem.",9535
bounded. is is of to the failure first probability capability ensures that The,The first capability is to ensures that the probability of failure is bounded.,9536
"theory, applications. systematically this properties, two In and compare metrics we on paper, these","In this paper, we systematically compare these two metrics on theory, properties, and applications.",9537
train and spike and conducting in statistical inferences the mean space. variance),mean and variance) and conducting statistical inferences in the spike train space.,9538
when possible. beneficial little is This is clustering particularly,This is particularly beneficial when little clustering is possible.,9539
source reference to There different algorithms for genome. alignment the sequence available are open,There are different open source algorithms available for sequence alignment to the reference genome.,9540
"tests nearly their ago a introduction since controversial. been Nevertheless, significance have century","Nevertheless, since their introduction nearly a century ago significance tests have been controversial.",9541
criticisms NHST. numerous are raised against Many aware researchers the not of,Many researchers are not aware of the numerous criticisms raised against NHST.,9542
NHST is reason why is misunderstood. often a undoubtedly This major very,This is undoubtedly a major reason why NHST is very often misunderstood.,9543
We architecture. a in noise ratio (SNR) compressive signal the lensless imaging analyze (LCI) to,We analyze the signal to noise ratio (SNR) in a lensless compressive imaging (LCI) architecture.,9544
of images approach. immunofluorescence the this classification Experiments the of show usefulness on indirect,Experiments on the classification of indirect immunofluorescence images show the usefulness of this approach.,9545
is of This the the to simplify encoding constraints. then information CNF applied,This information is then applied to simplify the CNF encoding of the constraints.,9546
Cellular and common of parallel Automata computations. most models simple are (CA),Cellular Automata (CA) are common and most simple models of parallel computations.,9547
image in applied processing. successfully can CA be,CA can be successfully applied in image processing.,9548
this detection with compared techniques. existing The performance some of edge approach is,The performance of this approach is compared with some existing edge detection techniques.,9549
prepared MATLAB. and in used in the All this paper algorithms results are,All the algorithms and results used in this paper are prepared in MATLAB.,9550
for with multiple surveys objectives. decision-making sequential This algorithms problems designed article,This article surveys algorithms designed for sequential decision-making problems with multiple objectives.,9551
"on learning. methods survey for we taxonomy, planning Using multi-objective this the literature and","Using this taxonomy, we survey the literature on multi-objective methods for planning and learning.",9552
"discuss key for and Finally, applications such future we of outline work. opportunities methods","Finally, we discuss key applications of such methods and outline opportunities for future work.",9553
"number Here, of into target the is a patches. partitioned","Here, the target is partitioned into a number of patches.",9554
"each the patches In occluded when dictionary. excluded are updating frame,","In each frame, occluded patches are excluded when updating the dictionary.",9555
proposed several results that shows Extensive state-of-the-art challenging on method trackers. the outperforms experimental sequences,Extensive experimental results on several challenging sequences shows that the proposed method outperforms state-of-the-art trackers.,9556
GAC-Schema performs also supports. better find that than We on full-length HaggisGAC,We also find that HaggisGAC performs better than GAC-Schema on full-length supports.,9557
are of All for propagating excellent constraints. disjunctions algorithms the proposed,All the proposed algorithms are excellent for propagating disjunctions of constraints.,9558
is parts. into split modelling The two,The modelling is split into two parts.,9559
and The vision level the relies computer on enable low real-time of footages analysis surveillance.,The analysis of the footages relies on low level computer vision and enable real-time surveillance.,9560
"trimap corresponding original size, segmentation a to built. upsampled and is This the is","This segmentation is upsampled to the original size, and a corresponding trimap is built.",9561
binding. protein organized network The domain the is specific core in of skeleton,The core skeleton of the network is organized in protein specific domain binding.,9562
settings. Such mechanism life in common is regulatory quite real a,Such a regulatory mechanism is quite common in real life settings.,9563
a present Monte-Carlo co-clusters. approximation polynomial-time mining thus for lagged We fuzzy algorithm,We thus present a polynomial-time Monte-Carlo approximation algorithm for mining fuzzy lagged co-clusters.,9564
"Moreover, overlapping missing algorithm and patterns. anti-correlations, values noise, the handles","Moreover, the algorithm handles noise, anti-correlations, missing values and overlapping patterns.",9565
real artificial algorithm The evaluated was extensively using datasets. and both,The algorithm was extensively evaluated using both artificial and real datasets.,9566
"the in method However, identifying network. in the be effective can hubs pattern-matching","However, the pattern-matching method can be effective in identifying hubs in the network.",9567
analysis new statistical that simple Markdown technology fully-reproducible a R and creating is makes painless.,R Markdown is a new technology that makes creating fully-reproducible statistical analysis simple and painless.,9568
or increasing be can also multimodal. monotonically relationships Diversity-dispersal,Diversity-dispersal relationships can also be monotonically increasing or multimodal.,9569
that our experimental algorithms. outperforms The demonstrate surveyed results algorithm,The experimental results demonstrate that our algorithm outperforms surveyed algorithms.,9570
constructed cost series of bitwise The is through operations strings. binary a volume on,The cost volume is constructed through bitwise operations on a series of binary strings.,9571
in problems recognition. recognition Single-sample the challenging face one face is of most,Single-sample face recognition is one of the most challenging problems in face recognition.,9572
This the the is present research of work. objective,This is the objective of the present research work.,9573
"compressed research work, Specifically work to proposal is document images. the run-length on this in","Specifically in this research work, the proposal is to work on run-length compressed document images.",9574
"activate thoughts, (pgACC), also and Pain loss to and sad hurt linked regrets. pregenual ACC","Pain and loss activate also pregenual ACC (pgACC), linked to sad thoughts, hurt and regrets.",9575
"(LHb) The stimulated by habenula outcome, lateral aversive and is pain, predicted loss. received discomfort,","The lateral habenula (LHb) is stimulated by predicted and received pain, discomfort, aversive outcome, loss.",9576
"LHb. lateral pgACC, found and caudolateral DTI Using adjacent dACC, OFC projections I to AI,","Using DTI I found dACC, pgACC, AI, adjacent caudolateral and lateral OFC projections to LHb.",9577
likely lack vACC This for projections for confirmed of mOFC. LHb was and,This lack of LHb projections was confirmed for vACC and likely for mOFC.,9578
disease can model. control mixtures population how using dynamics investigate improve We a host,We investigate how host mixtures can improve disease control using a population dynamics model.,9579
mixing The ratios. severity be a minimized disease over of range can,The disease severity can be minimized over a range of mixing ratios.,9580
of that findings diseases efficiently may in designing These help host crops. control mixtures,These findings may help in designing host mixtures that efficiently control diseases of crops.,9581
to many describe the with then generalize mixtures We model host components.,We then generalize the model to describe host mixtures with many components.,9582
"the specialization increases, becomes in the decrease larger. degree disease As severity of","As the degree of specialization increases, the decrease in disease severity becomes larger.",9583
in of change implementations algorithm both and ready-to-use random XPP include Matlab. We the time,We include ready-to-use implementations of the random time change algorithm in both XPP and Matlab.,9584
A central that of neural recording activities of issue the many in distinguishing is neurons.,A central issue in neural recording is that of distinguishing the activities of many neurons.,9585
We present the parallel efficiency analyses that speedup and prove of the implementation. the efficacy,We present the speedup and efficiency analyses that prove the efficacy of the parallel implementation.,9586
"used parallel be framework. each can methods different step at our Noteworthy, of","Noteworthy, different methods can be used at each step of our parallel framework.",9587
An the methods. proposed study approach existing superiority of over extensive proves state-of-the-art the experimental,An extensive experimental study proves the superiority of the proposed approach over existing state-of-the-art methods.,9588
"are index. multi-dimensional coupled a features inverted Basically, into complementary","Basically, complementary features are coupled into a multi-dimensional inverted index.",9589
"of exploit we color Specifically, c-MI. fusion local the into feature","Specifically, we exploit the fusion of local color feature into c-MI.",9590
"well Importantly, prior complementary show we is c-MI many that to techniques.","Importantly, we show that c-MI is well complementary to many prior techniques.",9591
for as new in and a entity biometrics emerging multi-modal verification. Palmprints are human identification,Palmprints are emerging as a new entity in multi-modal biometrics for human identification and verification.,9592
and evolution mutations fix. of are by The outcomes which occur determined,The outcomes of evolution are determined by which mutations occur and fix.,9593
"specific samplers space. other can sample specific In of for designed parts words, be the","In other words, specific samplers can be designed for specific parts of the sample space.",9594
total speedup and experiments. decrease of in observed Significant variation is,Significant speedup and decrease of total variation is observed in experiments.,9595
techniques prior benefits using performance highlight the dataset. our over We this of algorithm,We highlight the performance benefits of our algorithm over prior techniques using this dataset.,9596
"on applications. classification we focus binary Here,","Here, we focus on binary classification applications.",9597
dependence vestigate ways. We in structure samplers in- MCMC different using the several,We in- vestigate several MCMC samplers using the dependence structure in different ways.,9598
particle of densities in filtering with density kernel filter. deals the The paper estimates,The paper deals with kernel density estimates of filtering densities in the particle filter.,9599
of of The the analysis. means estimates by convergence is investigated Fourier,The convergence of the estimates is investigated by means of Fourier analysis.,9600
of time. persistence condition A Sobolev is presented sufficient character this for the over,A sufficient condition is presented for the persistence of this Sobolev character over time.,9601
of Both results to are derivatives the filtering densities. partial estimates and extended,Both results are extended to partial derivatives of the estimates and filtering densities.,9602
variation validity methods ecology been in and applying questioned. effectiveness of The community partitioning has,The effectiveness and validity of applying variation partitioning methods in community ecology has been questioned.,9603
"the role we partitioning. on analyzed influencing variation of species covariance slight Therefore,","Therefore, we analyzed the role of slight covariance on influencing species variation partitioning.",9604
binary and has label images been component (CCA) segments. used Connected analysis to heavily classify,Connected component analysis (CCA) has been heavily used to label binary images and classify segments.,9605
"to been it has multi-valued images. segment not well-exploited natural However,","However, it has not been well-exploited to segment multi-valued natural images.",9606
multi-value novel color images. to segmentation work utilizes proposes that CCA algorithm a segment This,This work proposes a novel multi-value segmentation algorithm that utilizes CCA to segment color images.,9607
labelled raw segmented the consists algorithm of of The distinctly output regions.,The raw output of the algorithm consists of distinctly labelled segmented regions.,9608
structures nucleated form  as centrosomes referred (MTs)  Microtubules star-shaped asters. to by,  Microtubules (MTs) nucleated by centrosomes form star-shaped structures referred to as asters.,9609
"cell and dynamics division, motility stability, vital polarization genome Aster and differentiation. is for","Aster motility and dynamics is vital for genome stability, cell division, polarization and differentiation.",9610
away cell or towards center Asters it. either move the from,Asters move either towards the cell center or away from it.,9611
drift with velocity The profiles and a walk directionality DNA. random towards directed suggest,The velocity and directionality profiles suggest a random walk with drift directed towards DNA.,9612
"ring theory been segmentation. has very with related image not However,","However, ring theory has not been very related with image segmentation.",9613
the tool to new that obtained The proved suitable index compare images. a is results,The obtained results proved that the new index is a suitable tool to compare images.,9614
compatible model is g-factor totally with The proposed formulations.,The proposed model is totally compatible with g-factor formulations.,9615
proposed. An images image super-resolution of low-resolution observation method is multiple from,An image super-resolution method from multiple observation of low-resolution images is proposed.,9616
the world. getting deployed Both farms rapidly over are as well offshore onshore as wind,Both onshore as well as offshore wind farms are getting deployed rapidly over the world.,9617
was variety feed forward network a of type used. simple trained algorithms of A using,A simple feed forward type of network trained using a variety of algorithms was used.,9618
hidden based on of number The was trials. nodes,The number of hidden nodes was based on trials.,9619
significant ARIMA showed lag. predicted the were values used also models but,ARIMA models were also used but the predicted values showed significant lag.,9620
observations free rigorous calculations. energy The by quantified are,The observations are quantified by rigorous free energy calculations.,9621
"SQMC replace Monte algorithm. SMC (particle In Carlo) PMCMC may chain particular, a Markov within","In particular, SQMC may replace SMC within a PMCMC (particle Markov chain Monte Carlo) algorithm.",9622
that SQMC outperform scenarios. in evidence provide numerical We SMC significantly practical may,We provide numerical evidence that SQMC may significantly outperform SMC in practical scenarios.,9623
function. to shown distribution an is DNA that fitted codon Others best have exponential the,Others have shown that the DNA codon distribution is best fitted to an exponential function.,9624
"reason been scaling behavior such However, addressed. for has yet the not","However, the reason for such scaling behavior has not yet been addressed.",9625
interact It individuals is cooperatively that the and competitively. assumed,It is assumed that the individuals interact cooperatively and competitively.,9626
"Moreover, ecological interactions. be behavior seen behaviors of such as other can the emergent","Moreover, other ecological behaviors can be seen as the emergent behavior of such interactions.",9627
empirical is data formulation. model its tested justify with to The,The model is tested with empirical data to justify its formulation.,9628
global characteristic It is humanity's interesting this population in is that observed growth.,It is interesting that this characteristic is observed in humanity's global population growth.,9629
"of show We certain for of results classes this architectures. form, network","We show results of this form, for certain classes of network architectures.",9630
experiment. thought the discuss brain prosthesis We,We discuss the brain prosthesis thought experiment.,9631
"simulated. is branching In a the paper, Owl this of model Spotted process Northern","In this paper, a branching process model of the Northern Spotted Owl is simulated.",9632
time We on focus extinction. the until,We focus on the time until extinction.,9633
established of samples processes. asymptotic is from when behaviour computed drawn HSIC The random from,The asymptotic behaviour of HSIC is established when computed from samples drawn from random processes.,9634
It which earlier that shown in is procedures i.i.d. worked bootstrap the,It is shown that earlier bootstrap procedures which worked in the i.i.d.,9635
to We this unlikely. find underlying analyze and be the assumptions scenario,We analyze this scenario and find the underlying assumptions to be unlikely.,9636
autonomous social instincts be for as that robots may hypothesize We such attachment learning. effective,We hypothesize that social instincts for autonomous robots may be effective such as attachment learning.,9637
information interest about (e.g. to a MetaboX library includes of methods The extract resource,The MetaboX library includes methods to extract information about a resource of interest (e.g.,9638
"unipartite reactants enzyme enzyme) networks, and build and metabolite, and networks. bipartite to reaction enzyme-metabolite","metabolite, reaction and enzyme) and to build reactants networks, bipartite enzyme-metabolite and unipartite enzyme networks.",9639
networks with tools. be visualization standard These for in data exported formats different can,These networks can be exported in different formats for data visualization with standard tools.,9640
be with Knowledge should real-time time learned constraints. in,Knowledge should be learned in real-time with time constraints.,9641
multi-scale The is tool vesselness filter in imaging. Frangi established an (retinal) vascular,The multi-scale Frangi vesselness filter is an established tool in (retinal) vascular imaging.,9642
"only or However, bifurcations, looks it structures. crossings it cope with cannot elongated since for","However, it cannot cope with crossings or bifurcations, since it only looks for elongated structures.",9643
"image invertible scale) in (multiple scores. the orientation structures we via crossing Therefore, disentangle","Therefore, we disentangle crossing structures in the image via (multiple scale) invertible orientation scores.",9644
on methods public are Both evaluated dataset. a,Both methods are evaluated on a public dataset.,9645
hybrid It on major of paradigms ToM that combines the a the relies two domain.,It relies on a hybrid ToM that combines the two major paradigms of the domain.,9646
step image. thinning the fingerprint is in input system identification fingerprint critical A of,A critical step in fingerprint identification system is thinning of the input fingerprint image.,9647
"proposed. fast is fingerprint a So, thinning algorithm","So, a fast fingerprint thinning algorithm is proposed.",9648
technique matching A proposed. is new fingerprint minutiae-based,A new minutiae-based fingerprint matching technique is proposed.,9649
"algorithm is rotation storage needs invariant, and size. and less The translation","The algorithm is rotation and translation invariant, and needs less storage size.",9650
"a case, more the individuals the peer. of successful strategy imitate former In","In the former case, individuals imitate the strategy of a more successful peer.",9651
under a dynamics dynamics. imitation Therefore can favored aspiration under be strategy disfavored,Therefore a strategy favored under imitation dynamics can be disfavored under aspiration dynamics.,9652
"biological principles be the of underlying principles Thus, operations. should also mental organization","Thus, biological principles of organization should also be the principles underlying mental operations.",9653
"advance called occurs, of such organizational When traverse. an is it a a system","When such an organizational advance of a system occurs, it is called a traverse.",9654
"model some widespread is A which, state-space shortcomings. however, representation, has elaborated and","A widespread and elaborated model is state-space representation, which, however, has some shortcomings.",9655
this heuristic knowledge. We can defining give some problem to remedy of kind by,We can give remedy to this problem by defining some kind of heuristic knowledge.,9656
"after unspecified semiparametric Gaussian. monotone the are transformations, marginally multivariate that, The assumes distributions model","The semiparametric model assumes that, after unspecified marginally monotone transformations, the distributions are multivariate Gaussian.",9657
datasets. outperforms real-world synthetic PCA both COCA on confirm Careful sparse experiments and that,Careful experiments confirm that COCA outperforms sparse PCA on both synthetic and real-world datasets.,9658
there are of on is currently roads that millions It the vehicles worldwide. believed,It is believed that there are currently millions of vehicles on the roads worldwide.,9659
"system. security against automatic activities, order these criminal an we In need public restrict","In order restrict against these criminal activities, we need an automatic public security system.",9660
identifier. vehicle Each (VIN) has Identification their as their primary own Number Vehicle,Each vehicle has their own Vehicle Identification Number (VIN) as their primary identifier.,9661
targets preferential adaptive The chromosomes them of evolution. pattern of inheritance unique makes X,The unique inheritance pattern of X chromosomes makes them preferential targets of adaptive evolution.,9662
We investigate of X selection here all in great chromosome natural on the species apes.,We here investigate natural selection on the X chromosome in all species of great apes.,9663
"data, Gene in actual the are data, protein for transcription or cells. expression concentrations surrogates","Gene expression data, or transcription data, are surrogates for actual protein concentrations in the cells.",9664
cell. the protein-protein in In all interactions diagrams interactions the addition static of protein-protein are,In addition protein-protein interactions are static diagrams of all the protein-protein interactions in the cell.,9665
cycle to to this apply yeast and We cell cancer.,We apply this to yeast cell cycle and to cancer.,9666
few drive mutations Precancerous growth and population rapid acquire key a cells carcinogenesis. that,Precancerous cells acquire a few key mutations that drive rapid population growth and carcinogenesis.,9667
passengers is unexploited of Collective currently therapeutic an target. effect deleterious,Collective deleterious effect of passengers is currently an unexploited therapeutic target.,9668
future therapies. be both effects exacerbated and We how by current discuss their might,We discuss how their effects might be exacerbated by both current and future therapies.,9669
"which a tracking technique, attractive as tracking treats an problem. category detection become has Tracking-by-detection","Tracking-by-detection has become an attractive tracking technique, which treats tracking as a category detection problem.",9670
ensemble of detectors. linear tracker exemplar-based discriminant is (ELDA) The an analysis proposed,The proposed tracker is an ensemble of exemplar-based linear discriminant analysis (ELDA) detectors.,9671
"we background both update and improve adaptivity, object models. its To","To improve its adaptivity, we update both object and background models.",9672
The model that of is msBayes. extension implemented in an,The model is an extension of that implemented in msBayes.,9673
most Image in challenging applications. the issue computer is vision segmentation,Image segmentation is the most challenging issue in computer vision applications.,9674
The to nearest is that so minimized. learned each its face image line subspaces are,The subspaces are learned so that each face image to its nearest line is minimized.,9675
showed results that The outperforms other algorithm proposed ones. experiment the,The experiment results showed that the proposed algorithm outperforms other ones.,9676
"this histogram Automatic Image method proposed. Registration for (AIR) a is paper, In through","In this paper, a method for Automatic Image Registration (AIR) through histogram is proposed.",9677
This of be to the pair combines registered. images new several segmentations of approach,This new approach combines several segmentations of the pair of images to be registered.,9678
A parameter on relaxation is histogram modes introduced. delineation the,A relaxation parameter on the histogram modes delineation is introduced.,9679
objects matched and estimation. translation The rotation are for used,The matched objects are used for rotation and translation estimation.,9680
differences registration with It images for and the of in translation. of rotation pairs allows,It allows for the registration of pairs of images with differences in rotation and translation.,9681
high-dimensional computer vision in applications. has many become use practice features of The normal a,The use of high-dimensional features has become a normal practice in many computer vision applications.,9682
image processing an interpolation field. Feature image in preserving is area active,Feature preserving image interpolation is an active area in image processing field.,9683
the Also previous proposed of methods images. outperforms JPEG method in case noisy and compressed,Also the proposed method outperforms previous methods in case of noisy and JPEG compressed images.,9684
"can process, optimization speed. achieve the in higher without Furthermore, for algorithm need the the","Furthermore, without the need for optimization in the process, the algorithm can achieve higher speed.",9685
component filtering image important is Pornographic recognition an system. in each,Pornographic image recognition is an important component in each filtering system.,9686
"is images In paper, detecting introduced. new for this pornographic approach a","In this paper, a new approach for detecting pornographic images is introduced.",9687
"suggested. are features approach, two new this In","In this approach, two new features are suggested.",9688
related Achieved verify proposed system results of works. the other versus performance,Achieved results verify the performance of proposed system versus other related works.,9689
"idling was Also, epochs. low during correlation the","Also, the correlation was low during idling epochs.",9690
gift-exchange-game from an a presents experiment. This paper analysis data of,This paper presents an analysis of data from a gift-exchange-game experiment.,9691
zero. Gift conditions no under was rate,Gift rate was under no conditions zero.,9692
"findings calculate some derive special Further, and we their significance.","Further, we derive some special findings and calculate their significance.",9693
algorithm literature method delivers. for Rarely a the of in after cares edit segmentation the,Rarely in literature a method of segmentation cares for the edit after the algorithm delivers.,9694
no They solution segmentation when provide goes wrong.,They provide no solution when segmentation goes wrong.,9695
of centripetal-parameterized formulate spline. to Catmull-Rom terms model distribution in propose We point,We propose to formulate point distribution model in terms of centripetal-parameterized Catmull-Rom spline.,9696
"brings edit fusion that so interactivity handled. is model-based better Such segmentation, to","Such fusion brings interactivity to model-based segmentation, so that edit is better handled.",9697
"the points shifts delivered user segment simply is When the unsatisfactory, to curve. vary","When the delivered segment is unsatisfactory, user simply shifts points to vary the curve.",9698
"an this In present we music two paper, study of experiments. EEG improvisation","In this paper, we present an EEG study of two music improvisation experiments.",9699
"mode. of modes: music ""let-go"" Each different piece mode was in strict two performed and","Each piece of music was performed in two different modes: strict mode and ""let-go"" mode.",9700
listeners. musicians both and measured was data EEG from Synchronized,Synchronized EEG data was measured from both musicians and listeners.,9701
structure the a from recover of silhouettes. to a compound We multiple object method propose,We propose a method to recover the structure of a compound object from multiple silhouettes.,9702
consider relevant cognitive We computational sequential idealized device. the and an concurrent in minimal concepts,We consider the relevant sequential and concurrent computational concepts in an idealized minimal cognitive device.,9703
cognition. concept of formalizes The limits chain fundamental temporal of cognitive,The concept of fundamental cognitive chain formalizes temporal limits of cognition.,9704
analysis categorical Latent clustering based perform responses. to model for is class multivariate used,Latent class analysis is used to perform model based clustering for multivariate categorical responses.,9705
and simulated is real data. tested approach The on,The approach is tested on simulated and real data.,9706
images is Humans' on and successful. objects and ability remarkably locate salient fast detect to,Humans' ability to detect and locate salient objects on images is remarkably fast and successful.,9707
simulation efficiently kinetic problem We consider for inference performing the and stochastic models. of,We consider the problem of efficiently performing simulation and inference for stochastic kinetic models.,9708
accuracy is Evaluation the of selected of the made. features,Evaluation of the accuracy of the selected features is made.,9709
neural The back by was system tested dataset. and propagation network trained CENPRMI with,The system was trained and tested by back propagation neural network with CENPRMI dataset.,9710
of behavior the quantitative for analyzing rational theory a provides agents. Game framework,Game theory provides a quantitative framework for analyzing the behavior of rational agents.,9711
comparing for chess A method openings. is described quantitative,A quantitative method is described for comparing chess openings.,9712
expert appealing in are and agree cases with they intuitively results opinion. some The,The results are intuitively appealing and in some cases they agree with expert opinion.,9713
The twofold. of aim is paper this,The aim of this paper is twofold.,9714
support dynamic the at decision system extending thesis with process. making aims This for,This thesis aims at extending the system with support for dynamic decision making process.,9715
"foundation is as data history industry, Appending Macau introduced. DCMS, the extending of of","Appending history data of Macau industry, as the foundation of extending DCMS, is introduced.",9716
"in may disappear. circumstances particularly the unfavorable Furthermore, ecosystem","Furthermore, in particularly unfavorable circumstances the ecosystem may disappear.",9717
Many solutions. notions formalisms some of of provide these,Many of these formalisms provide some notions of solutions.,9718
beings across  multiple Human correspondence stereoscopic scales.  process,  Human beings process stereoscopic correspondence across multiple scales.,9719
"dense state-of-the-art is this bio-inspiration correspondence. for ignored stereo However, aggregation cost by methods","However, this bio-inspiration is ignored by state-of-the-art cost aggregation methods for dense stereo correspondence.",9720
the Thyroid of the hormones body. control metabolism,Thyroid hormones control the metabolism of the body.,9721
"the Therefore, correct is diagnosis diseases very of these important topic.","Therefore, the correct diagnosis of these diseases is very important topic.",9722
is by performance evaluation estimated accuracy classification of and this cross-validation. using The system k-fold,The performance evaluation of this system is estimated by using classification accuracy and k-fold cross-validation.,9723
Calculus for topological Connection relations. is well-known The a Region (RCC) calculus representing part-whole and,The Region Connection Calculus (RCC) is a well-known calculus for representing part-whole and topological relations.,9724
"role in spatial science, and It plays qualitative information reasoning, important ontology. geographical an","It plays an important role in qualitative spatial reasoning, geographical information science, and ontology.",9725
the focus constraint of networks. consistency Most works these qualitative on of,Most of these works focus on the consistency of qualitative constraint networks.,9726
"the this redundant qualitative problem we of consider paper, important constraints. In","In this paper, we consider the important problem of redundant qualitative constraints.",9727
Bag-of-Words recent to works. image state-of-the-art retrieval The applied well is (BoW) representation,The Bag-of-Words (BoW) representation is well applied to recent state-of-the-art image retrieval works.,9728
"and correct multiple Typically, quantization improve vocabularies artifacts generated recall. are to","Typically, multiple vocabularies are generated to correct quantization artifacts and improve recall.",9729
"However, by routine this corrupted vocabulary different vocabularies. overlapping i.e., correlation, among is","However, this routine is corrupted by vocabulary correlation, i.e., overlapping among different vocabularies.",9730
datasets. experiments our extensive through We on method evaluate three benchmark,We evaluate our method through extensive experiments on three benchmark datasets.,9731
in analysis the computer field and presents vision. image challenging of recognition Face problem a,Face recognition presents a challenging problem in the field of image analysis and computer vision.,9732
security The becoming very significant difficult. information of is and,The security of information is becoming very significant and difficult.,9733
Face security. is system in used Recognition,Face Recognition system is used in security.,9734
image. a able to system detect should be an Face recognition in face automatically,Face recognition system should be able to automatically detect a face in an image.,9735
"is counting simple does used. but related metadata, instead, It mechanism require a not","It does not require related metadata, but instead, a simple counting mechanism is used.",9736
concept root to. that a is it is linked That plus concept -,That is - root concept plus a concept that it is linked to.,9737
weak model as a can learner a be used pedestrian in detection The framework. boosting-based,The model can be used as a weak learner in a boosting-based pedestrian detection framework.,9738
attributed influenza. contribution B make to Influenza viruses to morbidity a seasonal considerable,Influenza B viruses make a considerable contribution to morbidity attributed to seasonal influenza.,9739
phylogenetic across investigate genomic trees We segments. inter-lineage reassortments by comparing,We investigate inter-lineage reassortments by comparing phylogenetic trees across genomic segments.,9740
and that for guidelines then personalized oscillations. We these treatment options diabetic suggest patients maintain,We then suggest guidelines and personalized treatment options for diabetic patients that maintain these oscillations.,9741
understanding is level. the human Our intelligence of primarily directed at,Our understanding of intelligence is directed primarily at the human level.,9742
looking slightly different. suggests as The consciousness at being intelligence also and paper,The paper also suggests looking at intelligence and consciousness as being slightly different.,9743
"same. that, purposes, qualitatively find versions these for are We most practical the","We find that, for most practical purposes, these versions are qualitatively the same.",9744
"Penicillium in culture Mortierella spp., spp., while were spp., predominant Geomyces spp. Ganoderma","were predominant in culture while Ganoderma spp., Geomyces spp., Mortierella spp., Penicillium spp.",9745
from was data structure analyses community that diverse. fungal CI diversity Alpha revealed highly,Alpha diversity analyses from CI data revealed that fungal community structure was highly diverse.,9746
"to true the remains sampling. species under due diversity undetermined However,","However, the true species diversity remains undetermined due to under sampling.",9747
The of to that indicated recovery adapted frequent has the Pd habitats. pathogen WNS-afflicted,The frequent recovery of Pd indicated that the pathogen has adapted to WNS-afflicted habitats.,9748
"species. this an Further, the introduced that supports Pd is study hypothesis","Further, this study supports the hypothesis that Pd is an introduced species.",9749
"Molecular and with link functions,  evolution. genomic  information organismic phenotypes fitness,","  Molecular phenotypes link genomic information with organismic functions, fitness, and evolution.",9750
complex multiple that loci. phenotypes Quantitative genomic are traits on depend,Quantitative traits are complex phenotypes that depend on multiple genomic loci.,9751
"adaptive quantitative infer method on solution, to a evolution this of traits. Based we develop","Based on this solution, we develop a method to infer adaptive evolution of quantitative traits.",9752
system. pose adaptive for paper presents This the framework recognition face component-based,This paper presents the framework for pose adaptive component-based face recognition system.,9753
all proposed issues. mentioned deals above framework the with The,The framework proposed deals with all the above mentioned issues.,9754
"results is to simple, holistic feasible when and compared provides better Our method other methods.","Our method is simple, feasible and provides better results when compared to other holistic methods.",9755
technology attracted in analysis interface. attention has human-machine Facial for the much,Facial analysis has attracted much attention in the technology for human-machine interface.,9756
"from However, most a whole of these view methods face treat standpoint.","However, most of these methods treat face from a whole view standpoint.",9757
of disease infectious networks type. of We spread through contact the Configuration consider Model,We consider the spread of infectious disease through contact networks of Configuration Model type.,9758
we existing process offer the In some models. the of of simplifications,In the process we offer simplifications of some of the existing models.,9759
"appropriate models equivalent. Indeed, the under conditions are","Indeed, under appropriate conditions the models are equivalent.",9760
modeling. discuss Finally network-based for epidemic we challenges ongoing,Finally we discuss ongoing challenges for network-based epidemic modeling.,9761
promise for recognition. their networks Deep have neural as universal convolutional (CNN) a representation shown,Deep convolutional neural networks (CNN) have shown their promise as a universal representation for recognition.,9762
"and/or similar are many proposed that communities solutions, cases, different Different complementary. in","Different communities proposed different solutions, that are in many cases, similar and/or complementary.",9763
"exploration/exploitation, solutions active learning. These online-learning social include and learning,","These solutions include active learning, exploration/exploitation, online-learning and social learning.",9764
inter-disciplinary research discussion applications. consider We that will improve an such and integrated,We consider that such an integrated discussion will improve inter-disciplinary research and applications.,9765
competing resources. populations for by The interact,The populations interact by competing for resources.,9766
proof is provided. boundedness the the A for of populations,A proof for the boundedness of the populations is provided.,9767
kind A of competitive metapopulation obtained. exclusion is systems for principle,A kind of competitive exclusion principle for metapopulation systems is obtained.,9768
structure. individual different signal heart being to human unique is ECG due for an,ECG signal for an individual human being is different due to unique heart structure.,9769
"such useful parameter. we strong to model Therefore, a extract mathematical need","Therefore, we need a strong mathematical model to extract such useful parameter.",9770
adaptive an mathematical Hilbert-Huang is transform (HHT). model Here analysis,Here an adaptive mathematical analysis model is Hilbert-Huang transform (HHT).,9771
"nonstationary the is and the data. analyze This implemented non-linear new transform, approach, to Hilbert-Huang","This new approach, the Hilbert-Huang transform, is implemented to analyze the non-linear and nonstationary data.",9772
verified of is the the scheme through effectiveness simulation. The proposed,The effectiveness of the proposed scheme is verified through the simulation.,9773
"populations we all to with a variable Finally, generalize results effective size.","Finally, we generalize all results to populations with a variable effective size.",9774
have for multimodal fusion Recent studies information on classification. demonstrated sparsity advantages of based models,Recent studies have demonstrated advantages of information fusion based on sparsity models for multimodal classification.,9775
for task. tree-structured multimodal reformulates sparse the paper This model classification,This paper reformulates the tree-structured sparse model for multimodal classification task.,9776
to also problems these versions norms. employ describe solve We centered of regularization how which,We also describe how to solve regularization problems which employ centered versions of these norms.,9777
"datasets. multitask different to matrix regularizers matrix completion and learning apply the Finally, we","Finally, we apply the matrix regularizers to different matrix completion and multitask learning datasets.",9778
"and a convolutional adaptive translations computes scattering with rotations, A network. along invariants rigid-motion deep","A rigid-motion scattering computes adaptive invariants along translations and rotations, with a deep convolutional network.",9779
"joint preserves information, rotation while at desired any global translation It and invariants scale. providing","It preserves joint rotation and translation information, while providing global invariants at any desired scale.",9780
"characterization of stationary classification processes a is realization. through single from Texture studied, the","Texture classification is studied, through the characterization of stationary processes from a single realization.",9781
"important rotation results with on are texture and bases, scaling variabilities. State-of-the-art data obtained multiple","State-of-the-art results are obtained on multiple texture data bases, with important rotation and scaling variabilities.",9782
biases influence motor in perception how We visual whether might and actions. tested,We tested whether and how biases in visual perception might influence motor actions.,9783
"social (e.g. for image which labels), a is natural setting collections image-level","image-level labels), which is a natural setting for social image collections (e.g.",9784
"an weakly to efficient such Furthermore, develop solve sparse supervised learning algorithm problem. we","Furthermore, we develop an efficient algorithm to solve such weakly supervised sparse learning problem.",9785
our experimental on The results two the benchmark of show approach. effectiveness datasets,The experimental results on two benchmark datasets show the effectiveness of our approach.,9786
The potato. ratio the worse higher the the is,The higher the ratio the worse is the potato.,9787
potato the part by shown de-serializing also the Green of is output.,Green part of the potato is also shown by de-serializing the output.,9788
"probably higher give Additionally, camera feature images quality resolution better tracking. would","Additionally, higher resolution camera images would probably give better quality feature tracking.",9789
"Moreover, efficiently calculated optimization a problem. by it be convex solving can","Moreover, it can be efficiently calculated by solving a convex optimization problem.",9790
This the towards is memberships critical a points. step data evaluating between,This is a critical step towards evaluating the memberships between data points.,9791
"a the tasks. cognitive we systematic enabling techniques on in key Furthermore, tutorial provide involved","Furthermore, we provide a systematic tutorial on key enabling techniques involved in the cognitive tasks.",9792
"issues we and ahead. least, Last not open but present challenges research the","Last but not least, we present the research challenges and open issues ahead.",9793
"(with demand, social behavior, social world etc. and the human","and the social world (with human demand, social behavior, etc.",9794
"provisioning. allocation, resource and network enhance intelligent automatic and operation, service smart ),","), and enhance smart resource allocation, automatic network operation, and intelligent service provisioning.",9795
the (Kimura) is model. systems onto ``coarse--grained'' odd--even The neutral two--species mapped,The odd--even ``coarse--grained'' systems is mapped onto the two--species neutral (Kimura) model.,9796
"In four long all survive times. (evolutionary) the for regime, species ``neutral''","In the ``neutral'' regime, all four species survive for long (evolutionary) times.",9797
computer the checked results of simulations against model. The analytical are,The analytical results are checked against computer simulations of the model.,9798
species. for generalized model is The n,The model is generalized for n species.,9799
"generalized is of model version analysed. the a Volterra Also,","Also, a generalized version of the Volterra model is analysed.",9800
"criteria. into partitioned Within food various groups species to according can webs, be","Within food webs, species can be partitioned into groups according to various criteria.",9801
"trophic novel detection. we develop Here, group a for algorithm","Here, we develop a novel algorithm for trophic group detection.",9802
"descent. variables In rules cyclic develop coordinate for during particular, discarding we","In particular, we develop rules for discarding variables during cyclic coordinate descent.",9803
by a database Users can of operating variety access the web browser any systems. using,Users can access the database by any web browser using a variety of operating systems.,9804
sequencing. recently gain detailed become concerning this possible process through It high-throughput has information to,It has recently become possible to gain detailed information concerning this process through high-throughput sequencing.,9805
relevant were today. they and are how still accurate notes how It they,It notes how accurate they were and how relevant they still are today.,9806
combinations search to matched desired crop were that for methods specifications. Optimization then GEM used,Optimization methods were then used to search for GEM combinations that matched desired crop specifications.,9807
experiments filter Our competitive show methods. that to recently compared proposed is our,Our experiments show that our filter is competitive compared to recently proposed methods.,9808
is problem applied of to The the assessing similarity approach genetic practical in genetics. population,The approach is applied to the practical problem of assessing genetic similarity in population genetics.,9809
many challenging in tracking applications. is and important task an Object in video,Object tracking in video is an important and challenging task in many applications.,9810
changed WiSARD was algorithm original The to training the tracking. allow,The original WiSARD training algorithm was changed to allow the tracking.,9811
"reduce space, To have approaches exploiting been priors. proposed many various by the solution","To reduce the solution space, many approaches have been proposed by exploiting various priors.",9812
"solution. unsuitable In easily priors to practice, these would some lead","In practice, these priors would easily lead to some unsuitable solution.",9813
large image compact a descriptor search. novel for present scale image We,We present a novel compact image descriptor for large scale image search.,9814
clustering presented. is the angles function membership for also learning A by technique principled,A principled technique for learning the membership function by clustering angles is also presented.,9815
Homology; Homology Removal Maps; Keywords: Contraction Cellular Operations. and Generalized $n$D Generators;,Keywords: $n$D Generalized Maps; Cellular Homology; Homology Generators; Contraction and Removal Operations.,9816
the applicability We QM opinion in basic properties two to analyze polling. of,We analyze the applicability of QM to two basic properties in opinion polling.,9817
"for occur a properties together, it these Whenever two problem QM. poses","Whenever these two properties occur together, it poses a problem for QM.",9818
The among biology. problems developmental the major of multicellularity evolutionary programs emergence are of and,The emergence of multicellularity and developmental programs are among the major problems of evolutionary biology.,9819
an phenotypic individuals variation. of are differences important in Heritable expression source between gene,Heritable differences in gene expression between individuals are an important source of phenotypic variation.,9820
hundreds for in differences were genes. of translation Strain observed,Strain differences in translation were observed for hundreds of genes.,9821
"these dimensional be high simulations of can data. prohibitive cost the computational However, for","However, the computational cost of these simulations can be prohibitive for high dimensional data.",9822
"Potts important commonly image An analysis. which example in is the is model, used","An important example is the Potts model, which is commonly used in image analysis.",9823
method estimating by also remotely the smoothing sensed the parameter We for imagery. illustrate satellite,We also illustrate the method by estimating the smoothing parameter for remotely sensed satellite imagery.,9824
"scale. inference Bayesian for that impractical of precomputation, is Without datasets","Without precomputation, Bayesian inference is impractical for datasets of that scale.",9825
"shape in are for each kept. good polygon frame, a matches local Only","Only good matches for each local shape polygon in a frame, are kept.",9826
topic research through state being speech is interesting nowadays. a recognition Emotional very,Emotional state recognition through speech is being a very interesting research topic nowadays.,9827
"modification. frequency we For a purpose, gender this dependent scaling propose","For this purpose, we propose a gender dependent frequency scaling modification.",9828
outcomes. parameter and data to incorporate distributions We population-level clinical estimate,We incorporate clinical data to estimate population-level parameter distributions and outcomes.,9829
distributions posterior when highest are improvements obtained highly correlated are The considered.,The highest improvements are obtained when highly correlated posterior distributions are considered.,9830
chains. present We reversible Markov prior a over nonparametric,We present a nonparametric prior over reversible Markov chains.,9831
"cluster and plants compete invasive particular, are most propagation, through In clonal, locally. vegetative","In particular, most invasive plants are clonal, cluster through vegetative propagation, and compete locally.",9832
interface random in of produces invasive extent the along variation the competition Interspecific growth.,Interspecific competition along the interface produces random variation in the extent of invasive growth.,9833
"clover (Lolium perenne) we invade these test experimentally. white predictions, To repens) (Trifolium let ryegrass","To test these predictions, we let white clover (Trifolium repens) invade ryegrass (Lolium perenne) experimentally.",9834
"differed, However, clover's scaling the exponents consequence of growth a likely morphology.","However, the scaling exponents differed, likely a consequence of clover's growth morphology.",9835
molecular alternative shown all-atom to conventional to CABS-flex (MD). be dynamics an efficient was,CABS-flex was shown to be an efficient alternative to conventional all-atom molecular dynamics (MD).,9836
"and between ensembles. that is NMR average, correlation On than MD this stronger","On average, this correlation is stronger than that between MD and NMR ensembles.",9837
(CovDs) spaces. comparing and introduce infinite-dimensional in approach to Covariance Descriptors We an computing,We introduce an approach to computing and comparing Covariance Descriptors (CovDs) in infinite-dimensional spaces.,9838
vision. to address have become problems computer CovDs increasingly classification in popular,CovDs have become increasingly popular to address classification problems in computer vision.,9839
exploit these then purposes. for divergences classification We,We then exploit these divergences for classification purposes.,9840
"annotating images. RGBD SmartAnnotator, interactive facilitate We system to an present","We present SmartAnnotator, an interactive system to facilitate annotating RGBD images.",9841
"process speeds objects resolve. finalized, to up ambiguities fewer As the with are","As objects are finalized, the process speeds up with fewer ambiguities to resolve.",9842
have years boosted the focus. healthy into recent In the recognition technologies system face,In recent years the technologies have boosted face recognition system into the healthy focus.,9843
to on issue mainly lighting of toughness focuses paper This the variations.,This paper focuses mainly on the issue of toughness to lighting variations.,9844
are subordinate in important animals. of and the group living behaviours Dominance social organizations ingredients,Dominance and subordinate behaviours are important ingredients in the social organizations of group living animals.,9845
"recent visual the neural research, remain intense of conscious correlates Despite elusive. perception","Despite intense recent research, the neural correlates of conscious visual perception remain elusive.",9846
"power. phase, in waveform, of and correlate SCP perception conscious manifests its strongly The","The SCP correlate of conscious perception manifests strongly in its waveform, phase, and power.",9847
"contrast, confidence performance relatively In objective by contributed transient activity. brain both were and","In contrast, objective performance and confidence were both contributed by relatively transient brain activity.",9848
"light and metacognitive conscious, shed mechanisms the brain on processing. These unconscious, new results of","These results shed new light on the brain mechanisms of conscious, unconscious, and metacognitive processing.",9849
all several sparse We possible learning further issues in over discuss unnoticed subgraph features.,We further discuss several unnoticed issues in sparse learning over all possible subgraph features.,9850
computational show We that in data al. all et Schweizer,We show that all computational data in Schweizer et al.,9851
All drawn are void. conclusions model the from computational,All conclusions drawn from the computational model are void.,9852
within Program of this one process. crucial is the activities debugging,Program debugging is one of the crucial activities within this process.,9853
of ASP explanations of efficient a allow suggested fault. approaches computation possible debugging Recently,Recently suggested ASP debugging approaches allow efficient computation of possible explanations of a fault.,9854
of A interference. of the with sensors presence major Time multipath issue Flight is,A major issue with Time of Flight sensors is the presence of multipath interference.,9855
credibility is the to carried this analysis of judge historical out bibliographical A thesis. and,A historical and bibliographical analysis is carried out to judge the credibility of this thesis.,9856
possible then makes to final it the bring report of a physical The inspection answer.,The physical inspection of the report then makes it possible to bring a final answer.,9857
this (TTMCMC). Transdimensional Chain Transformation Monte based methodology as Carlo We Markov refer to,We refer to this methodology as Transdimensional Transformation based Markov Chain Monte Carlo (TTMCMC).,9858
"application such RJMCMC of for To exists high-dimensional no knowledge, there mixtures. our","To our knowledge, there exists no application of RJMCMC for such high-dimensional mixtures.",9859
to convergence our a also propose Based of algorithms. method we variable-dimensional criterion assess on,Based on our method we also propose a criterion to assess convergence of variable-dimensional algorithms.,9860
dictionary concrete gives bounds and the learning. more sample complexity of of,and gives more concrete bounds of the sample complexity of dictionary learning.,9861
variety the learning a We procedure. of cover sparsity employed measures in,We cover a variety of sparsity measures employed in the learning procedure.,9862
"largely diagram, thus, However, synaptic this and polarities are circuit unknown. in functions","However, synaptic polarities in this diagram, and thus, circuit functions are largely unknown.",9863
excitatory to and distant similar connections of they probably targets. have inhibitory number,they have probably similar number of inhibitory and excitatory connections to distant targets.,9864
the category of under dynamic texture. fall,fall under the category of dynamic texture.,9865
Researchers textures investigated nineties. to ways analyze have since different early dynamic,Researchers have investigated different ways to analyze dynamic textures since early nineties.,9866
and Both (image based approaches appearance are investigated. based intensities) motion,Both appearance based (image intensities) and motion based approaches are investigated.,9867
approaches more Motion turn effective. out be to based,Motion based approaches turn out to be more effective.,9868
"recognition, using computation, texture dynamic motion synthesis. block includes motion segmentation, analysis and Research into","Research into dynamic texture analysis using block motion includes recognition, motion computation, segmentation, and synthesis.",9869
a these We of approaches. comprehensive provide review,We provide a comprehensive review of these approaches.,9870
The developed fragmentary. proteomic always is data almost that this are paper in point,The point developed in this paper is that proteomic data are almost always fragmentary.,9871
can such effective technique image produce fusion An remotely images. sensed,An effective image fusion technique can produce such remotely sensed images.,9872
"after of For samples, the progressively distributions sequence a each are sample. adapted","For a sequence of samples, the distributions are progressively adapted after each sample.",9873
in Decision-Making and Laboratory (DEMATEL) real Evaluation many widely Trial applications. is used method,Decision-Making Trial and Evaluation Laboratory (DEMATEL) method is widely used in many real applications.,9874
"some that we paper, this in show short not situations. reasonable In it is","In this short paper, we show that it is not reasonable in some situations.",9875
"with of dealing involving information. incapable uncertain However, is the and cases CFPR incomplete the","However, the CFPR is incapable of dealing with the cases involving uncertain and incomplete information.",9876
much improved fit recent observations. a into empirical the to Incorporating provides model growth,Incorporating recent growth into the model provides a much improved fit to empirical observations.,9877
larger sample marked becomes This sizes. for phenomenon more,This phenomenon becomes more marked for larger sample sizes.,9878
"to satisfy. such hard In is real applications, however, separability assumption","In real applications, however, such separability assumption is hard to satisfy.",9879
performance on data the the synthetic of show and real We algorithm sets.,We show the performance of the algorithm on real and synthetic data sets.,9880
strictly advantage transitory. We on landscapes rugged is adaptive recombination of static the that show,We show that the adaptive advantage of recombination on static rugged landscapes is strictly transitory.,9881
"learning. sparse and reduction, are basic components analysis, cluster Dimensionality representation in machine","Dimensionality reduction, cluster analysis, and sparse representation are basic components in machine learning.",9882
"yet been fully relationships investigated. their However, have not","However, their relationships have not yet been fully investigated.",9883
"is efficient to codes solve SSR. An developed algorithm, of the sparse NSCrt,","An efficient algorithm, NSCrt, is developed to solve the sparse codes of SSR.",9884
"various Scut. of sets Experiments SSR, on data the and and NSCrt, demonstrate properties strengths","Experiments on various data sets demonstrate the properties and strengths of SSR, NSCrt, and Scut.",9885
silhouettes views. seed We separate from in two shape the seeds quantified the of,We quantified seed shape from the silhouettes of the seeds in two separate views.,9886
"weeds shape but by varied phylogenetic Trajectory separated view, from and groupings. change crops of","Trajectory of shape change varied by view, but separated crops from weeds and phylogenetic groupings.",9887
evidence exploitation. of further reflecting gave adaptation Three-dimensional models differences in environmental shape for,Three-dimensional models gave further evidence of differences in shape reflecting adaptation for environmental exploitation.,9888
also taxa. of a may niche Seed separation shape between mechanism be,Seed shape may also be a mechanism of niche separation between taxa.,9889
adds mounting paper This the evidence this to is indeed the that case.,This paper adds to the mounting evidence that this is indeed the case.,9890
dataset. for memory outperforms low retrieval For sculptures consistently it except instance footprint methods,For instance retrieval it consistently outperforms low memory footprint methods except for sculptures dataset.,9891
are representations simple e.g. The augmentation modified further using techniques,The representations are further modified using simple augmentation techniques e.g.,9892
scenes. also of technique complicated sort This extracting allows objects some from,This technique also allows extracting objects from some sort of complicated scenes.,9893
approach step choices. selection environmental movement (SSF) function to uncover use We drivers a behind,We use a step selection function (SSF) approach to uncover environmental drivers behind movement choices.,9894
data. adapt to work with compressive problem sensing first this We minimization,We first adapt this minimization problem to work with compressive sensing data.,9895
extension reconstruction. measurement the That problem brings out of,That extension brings out the problem of measurement reconstruction.,9896
patients this These often neurological features modify healthy diseases structure. appear whereas in,These features appear in healthy patients whereas neurological diseases often modify this structure.,9897
"distribution. of composite performing tests asymptotic because likelihood inconvenient Unfortunately, awkward is their ratio","Unfortunately, performing composite likelihood ratio tests is inconvenient because of their awkward asymptotic distribution.",9898
and The Wald-type is for true counterparts. score-type same,The same is true for Wald-type and score-type counterparts.,9899
in increasingly Markov recent years. inference become for Bayesian processes has relevant,Bayesian inference for Markov processes has become increasingly relevant in recent years.,9900
by this identify of paper gait. is to analyzing individuals The their goal,The goal of this paper is to identify individuals by analyzing their gait.,9901
"pyramidal the gait of motion. representation Thus, obtaining a","Thus, obtaining a pyramidal representation of the gait motion.",9902
the is to components unity. The retained of amplitude normalized,The amplitude of the retained components is normalized to unity.,9903
"for a mostly computational MAPK Ordinary Differential model are Partial using Equations. networks simulation, and","MAPK networks are a model for computational simulation, mostly using Ordinary and Partial Differential Equations.",9904
"have switch-like bistability that networks and Key showed oscillations. behavior, these results can","Key results showed that these networks can have switch-like behavior, bistability and oscillations.",9905
result et. present apply theoretical We the therefore P\'erez Mill\'an in,We therefore apply the theoretical result present in P\'erez Mill\'an et.,9906
information improvements The of with extraordinary to unprecedented devices modern data content. imaging offer access,The extraordinary improvements of modern imaging devices offer access to data with unprecedented information content.,9907
"measurement In below applications, is keep necessary it many threshold. intensities a desired to","In many applications, it is necessary to keep measurement intensities below a desired threshold.",9908
"that have and, limited are applications. slow, state-of-the-art thus, accuracy methods Most achieve","Most methods that achieve state-of-the-art accuracy are slow, and, thus, have limited applications.",9909
a This of on achieved regression of with ensembles based sequence is estimators trees.,This is achieved with a sequence of estimators based on ensembles of regression trees.,9910
has that value. Experimental our results show method practical,Experimental results show that our method has practical value.,9911
movement patterns contrasting between by adopted a reflect balance these should processes. organisms The,The movement patterns adopted by organisms should reflect a balance between these contrasting processes.,9912
Thus food exposure stronger for to trajectories present predators. between a separation exploration and real,Thus real trajectories present a stronger separation between exploration for food and exposure to predators.,9913
with aura. Familial hemiplegic of subtype rare migraine migraine is (FHM) a,Familial hemiplegic migraine (FHM) is a rare subtype of migraine with aura.,9914
inactivation genetic the This gate. defect affects,This genetic defect affects the inactivation gate.,9915
spp. FoxPatch represents include Setaria models which weedy evolutionary Trait-based,Trait-based evolutionary models include FoxPatch which represents weedy Setaria spp.,9916
rules algorithms. with behavior process explicit seed life and prediction history,seed behavior with explicit life history process prediction rules and algorithms.,9917
"computing that indents image compressed decompressed, be this the additional implies resources. However, which should","However, this implies that the compressed image should be decompressed, which indents additional computing resources.",9918
in is increasingly interested detectors. object of becoming rapid Computer vision estimation the,Computer vision is increasingly becoming interested in the rapid estimation of object detectors.,9919
of a we the How difficulty puzzle? predict can Sudoku,How can we predict the difficulty of a Sudoku puzzle?,9920
model activity. results human computational of a using are solving best The obtained,The best results are obtained using a computational model of human solving activity.,9921
regression of comparison. A has for been models variety considered,A variety of regression models has been considered for comparison.,9922
the What and is brain behavior? between relationship,What is the relationship between brain and behavior?,9923
between characterizing The necessitates answer mapping function. question this to and structure the,The answer to this question necessitates characterizing the mapping between structure and function.,9924
"becomes noise, presence regularization. of an ill-posed However, it requires in problem that the","However, in the presence of noise, it becomes an ill-posed problem that requires regularization.",9925
that method. are the Numerical merits demonstrate presented experiments our of,Numerical experiments are presented that demonstrate the merits of our method.,9926
"computational this paper cost. however, demonstrate, comes a In efficiency at that we this","In this paper we demonstrate, however, that this computational efficiency comes at a cost.",9927
fast arise a as so can variation to sometimes ecosystem population in Genetic modify dynamics.,Genetic variation in a population can sometimes arise so fast as to modify ecosystem dynamics.,9928
protein folding one molecular Understanding of great biochemistry has the challenges and been biophysics. in,Understanding protein folding has been one of the great challenges in biochemistry and molecular biophysics.,9929
"in advances protein In far behind. the field comparison, membrane lag folding","In comparison, advances in the membrane protein folding field lag far behind.",9930
"no are these Furthermore, folding available biomolecules systematic for experimental in strategies vitro.","Furthermore, no systematic experimental strategies are available for folding these biomolecules in vitro.",9931
"Additional being multidomain one proteins, the interpretation data critical. membrane emerge complexities in most of","Additional complexities emerge in multidomain membrane proteins, data interpretation being one of the most critical.",9932
Many low-dimensional by integrals the produced methods are to designs. existing tailored nested,Many existing methods are tailored to the low-dimensional integrals produced by nested designs.,9933
code online in supplementary the R is material. Source provided in,Source code in R is provided in the online supplementary material.,9934
divisible distributions connect on We infinitely $\mathbb{R}^{d}$. to shift-invariant characteristic kernels,We connect shift-invariant characteristic kernels to infinitely divisible distributions on $\mathbb{R}^{d}$.,9935
paper is of The contribution this two-fold.,The contribution of this paper is two-fold.,9936
of the object process play recognition. an of information critical in a role object Shape,Shape information of an object play a critical role in the process of object recognition.,9937
an while boundary. object Rasters important computing role play,Rasters play an important role while computing object boundary.,9938
common to the understand structure the most used better. Hierarchies world are,Hierarchies are the most common structure used to understand the world better.,9939
"for a are system. in hierarchical instance, multiple-star systems galaxies, In organised","In galaxies, for instance, multiple-star systems are organised in a hierarchical system.",9940
approach to propose this We framework-based a remedy situation.,We propose a framework-based approach to remedy this situation.,9941
"directions conclude we findings for work. present our Finally, and with future","Finally, we conclude with our findings and present directions for future work.",9942
"novel propose this a framework, problem. this ROML, To termed to address we end,","To this end, we propose a novel framework, termed ROML, to address this problem.",9943
contributions of key as our summarized are Two follows.,Two of our key contributions are summarized as follows.,9944
source interoperability. purpose human communication and facilitate A is whose to major,A source whose major purpose is to facilitate human communication and interoperability.,9945
want The when problem data/knowledge. share their they occurs to,The problem occurs when they want to share their data/knowledge.,9946
merge into Thus to ontologies one. tools we need,Thus we need tools to merge ontologies into one.,9947
as ontology This task matching. termed is,This task is termed as ontology matching.,9948
paper In graphs. ontologies have this we framework a using to matching shown,In this paper we have shown a framework to matching ontologies using graphs.,9949
"outliers. be structure may underlying sparse by and/or errors the Nevertheless, affected","Nevertheless, the underlying structure may be affected by sparse errors and/or outliers.",9950
"yeast clusters, of higher Shortly of cell evolution evolved after the rates death. the","Shortly after the evolution of clusters, the yeast evolved higher rates of cell death.",9951
trouble Bioinformatic pipelines are keeping having pace. analysis,Bioinformatic analysis pipelines are having trouble keeping pace.,9952
large. to screened is space the chemical be Moreover prohibitively,Moreover the chemical space to be screened is prohibitively large.,9953
the to filtering promising is of prior docking Therefore a concept. application,Therefore the application of filtering prior to docking is a promising concept.,9954
"analysis. component principal a binary our applies classification Additionally, program using","Additionally, our program applies a binary classification using principal component analysis.",9955
the extensions approach. standard propose to We two stacking,We propose two extensions to the standard stacking approach.,9956
second parts. two extension consists of The,The second extension consists of two parts.,9957
accordance the Second thermodynamics of This restated a is in with stability Law as criterion.,This is in accordance with the restated Second Law of thermodynamics as a stability criterion.,9958
"introduce To from astrophysics limitation, address flux-limited this we the domain. diffusion, a technique","To address this limitation, we introduce flux-limited diffusion, a technique from the astrophysics domain.",9959
three have of roles ATP RecA. of established features the and We,We have established three features of the roles of ATP and RecA.,9960
for activity or found ATPase autoregulatory mechanism polymerase. similar previously DNA a been No has,No similar ATPase activity or autoregulatory mechanism has previously been found for a DNA polymerase.,9961
viral RV studies agents. evolutionary proteins This other of support or different may,This may support other evolutionary studies of different RV proteins or viral agents.,9962
"implies compressed decompressed, However, image which indents computing be additional should this resources. the that","However, this implies that the compressed image should be decompressed, which indents additional computing resources.",9963
in directly research compressed motivation This from features limitation image. induces the to the extracting,This limitation induces the motivation to research in extracting features directly from the compressed image.,9964
The identical feature those uncompressed values to images. exactly are so extracted extracted from,The feature values so extracted are exactly identical to those extracted from uncompressed images.,9965
often In by logistic cohort outcomes binary very regression. studies analyzed are,In cohort studies binary outcomes are very often analyzed by logistic regression.,9966
"regression a preferable. cases, these is model In log-binomial","In these cases, a log-binomial regression model is preferable.",9967
"new, invariant scaling derive a measure. I","I derive a new, scaling invariant measure.",9968
"and to the deal the colonizations system. I using surreal with measure number extend extinctions,","I extend the measure to deal with colonizations and extinctions, using the surreal number system.",9969
them. to to methods for look and realize,and to look for methods to realize them.,9970
"imaginative constructivism. based on which Brouwer the called are machines generally, is theoretical More work","More generally, the theoretical work on which Brouwer machines are based is called imaginative constructivism.",9971
a model. bias B-spline correction based supports on MBIS multi-channel field,MBIS supports multi-channel bias field correction based on a B-spline model.,9972
channel. each bias first validate field and We of estimated the for segmentation the accuracy,We first validate the accuracy of segmentation and the estimated bias field for each channel.,9973
evaluation. a a segmentation widely MBIS tool used in outperforms cross-comparison,MBIS outperforms a widely used segmentation tool in a cross-comparison evaluation.,9974
literature either algorithms of but discrete in Most Gaussian recovery continuous the assume data.,Most recovery algorithms in the literature assume either discrete of continuous but Gaussian data.,9975
the convolution we For cases theorem. these present,For these cases we present the convolution theorem.,9976
"a changes, study. and To provide reflect we these enlarged modernized","To reflect these changes, we provide a modernized and enlarged study.",9977
from of Szeliski insights with the Key al. study our agree results et,Key insights from our study agree with the results of Szeliski et al.,9978
models the they of studied. types for,for the types of models they studied.,9979
"both the from (ATR) challenge platforms, Automatic For of specific Recognition channel (e.g. Target radar","For the specific challenge of Automatic Target Recognition (ATR) from radar platforms, both channel (e.g.",9980
spatial critical useful specific information task. a and polarization) diversity and provide such for can,polarization) and spatial diversity can provide useful information for such a specific and critical task.,9981
various This can paper seemingly presents that formulae conundrums. philosophical solve hopeless,This paper presents formulae that can solve various seemingly hopeless philosophical conundrums.,9982
"between we DAGs. Furthermore, also (potentially allow unknown) the dependency for structure","Furthermore, we allow also for (potentially unknown) dependency structure between the DAGs.",9983
fMRI data presented are both multiple obtained and simulated Results subjects. on data from,Results are presented on both simulated data and fMRI data obtained from multiple subjects.,9984
of theory from this evolution. to dynamics Applications social range cell,Applications of this theory range from cell dynamics to social evolution.,9985
"Zernike rotation moments invariant to Orthogonality identify its property. describes due shape,","Zernike moments describes shape, identify rotation invariant due to its Orthogonality property.",9986
ancient cursive an characters. of and complex script had is India MODI of representation,MODI is an ancient script of India had cursive and complex representation of characters.,9987
"the Unfortunately, GMM turns to due non-convex to out be objective.","Unfortunately, GMM turns out to be non-convex due to the objective.",9988
of COCA the illustrate distributions. then in Gaussian synthetic structured with compound advantages We simulations,We then illustrate the advantages of COCA in synthetic simulations with structured compound Gaussian distributions.,9989
"of we (in only require Fourier Instead, the smoothness process. domain) properties the certain","Instead, we only require certain smoothness properties (in the Fourier domain) of the process.",9990
"a and than previous more simpler efficient algorithms. search new is minimax MTD(f) algorithm,","MTD(f) is a new minimax search algorithm, simpler and more efficient than previous algorithms.",9991
randomly delay discuss from accessible is element how the on also information any line. We,We also discuss how information is randomly accessible from any element on the delay line.,9992
chemical for existing integration systems. system's allows with The modularity,The system's modularity allows for integration with existing chemical systems.,9993
"The drug and as delivery. smart line delay has in biomedical treatment, such diagnosis applications","The delay line has applications in biomedical diagnosis and treatment, such as smart drug delivery.",9994
Rubin is Donald Loeb Statistics of University. Bruce John Harvard L. Professor at,Donald Bruce Rubin is John L. Loeb Professor of Statistics at Harvard University.,9995
object classification. Networks provide (CNNs) Convolutional accurate Neural can,Convolutional Neural Networks (CNNs) can provide accurate object classification.,9996
DenseNet efficient object descriptors. involve feature will training work detectors Future with,Future work will involve training efficient object detectors with DenseNet feature descriptors.,9997
defined on $\Omega$. Our function piecewise density of constant a a estimate binary partition is,Our density estimate is a piecewise constant function defined on a binary partition of $\Omega$.,9998
"rate. resulting has The and simple, is efficient, convergence provable algorithm a","The resulting algorithm is simple, efficient, and has a provable convergence rate.",9999
estimation a as We demonstrate its efficiency empirically density method.,We empirically demonstrate its efficiency as a density estimation method.,10000
search This a algo- paper rithms. paradigm new introduces minimax for game-tree,This paper introduces a new paradigm for minimax game-tree search algo- rithms.,10001
version a Test MT of procedure. is Pearls memory-enhanced,MT is a memory-enhanced version of Pearls Test procedure.,10002
search algorithms minimax assessments of simulations. have been on of Most the based,Most of the assessments of minimax search algorithms have been based on simulations.,10003
to guide is aid. designed a without a vehicle Autopilot system,Autopilot is a system designed to guide a vehicle without aid.,10004
flight. a in becomes system consistency Autopilot reliability crucial and Thus a of an role,Thus reliability and consistency of an Autopilot system becomes a crucial role in a flight.,10005
vague Rough Applications containing sets a for of potent data. be can kind such tool,Rough sets can be a potent tool for such kind of Applications containing vague data.,10006
determination factor an approach using proposes Set paper Consistency towards This Rough Theory.,This paper proposes an approach towards Consistency factor determination using Rough Set Theory.,10007
"through Set evaluated Consistency Theory. Factor using Rough is payloads, these","Consistency Factor is evaluated through these payloads, using Rough Set Theory.",10008
"it perceived solves a SSS*, algorithm. of practical all reformulation making three The drawbacks","The reformulation solves all three perceived drawbacks of SSS*, making it a practical algorithm.",10009
results. This runs simulation counter published to,This runs counter to published simulation results.,10010
solving a There of under-determined inverse for are linear number problem. large methods,There are a large number of methods for solving under-determined linear inverse problem.,10011
have datasets. them of time Many very high large complexity for,Many of them have very high time complexity for large datasets.,10012
a tackle new problem. Representation Two-Stage method to called this We propose Sparse (TSSR),We propose a new method called Two-Stage Sparse Representation (TSSR) to tackle this problem.,10013
approximate concentration its distribution The a exploit to dictionary designed sub-Gaussian property. is to,The dictionary is designed to approximate a sub-Gaussian distribution to exploit its concentration property.,10014
The over representing coefficients testing encoded are and the are basis coefficients obtained. the final,The testing coefficients are encoded over the basis and the final representing coefficients are obtained.,10015
"Since conducted the is a on reduced. is space, greatly projection sparser runtime much the","Since the projection is conducted on a much sparser space, the runtime is greatly reduced.",10016
"instance provide the proposed realization, for we an TSSR. concrete For","For concrete realization, we provide an instance for the proposed TSSR.",10017
Statistical author's The program is on work of based Interactions (SMNI). Neocortical in the Mechanics,The program is based on the author's work in Statistical Mechanics of Neocortical Interactions (SMNI).,10018
other as widely required histograms like windows. and by applied Parzen methods,as required by other widely applied methods like histograms and Parzen windows.,10019
"game-playing optimal trees. build search Hence, appears nearly it programs","Hence, it appears game-playing programs build nearly optimal search trees.",10020
"graph conventional definition the of minimal wrong. However, the is","However, the conventional definition of the minimal graph is wrong.",10021
graph the just conventional of a minimal is definition left-most The approximation.,The conventional definition of the minimal graph is just a left-most approximation.,10022
size computationally of Calculating too intensive. minimal the the real is graph,Calculating the size of the real minimal graph is too computationally intensive.,10023
"significantly graph. bound be approximations than it However, minimal show to upper smaller the left-most","However, upper bound approximations show it to be significantly smaller than the left-most minimal graph.",10024
"is widely it programs are that appears as as believed. Hence, game-playing efficiently searching not","Hence, it appears that game-playing programs are not searching as efficiently as is widely believed.",10025
"of search shown reduce them, significantly cutoffs, size. tree is transposition enhanced One to","One of them, enhanced transposition cutoffs, is shown to significantly reduce search tree size.",10026
Alternative such methods as - maximum the explicitly -- parsimony structure tree account. take into,Alternative methods -- such as maximum parsimony - explicitly take the tree structure into account.,10027
"document spotting like retrieval. obtained entropies are The establishing word equivalence, in useful and applications","The entropies obtained are useful in applications like establishing equivalence, word spotting and document retrieval.",10028
"guide. as this we on of touch our a means debate framing Below,","Below, we touch on this debate as a means of framing our guide.",10029
usage are on FATIGUE. as The as NNN based case well its study evaluated formalization,The NNN formalization as well as its usage are evaluated based on case study FATIGUE.,10030
"image mostly however, For lacking. deblurring, approaches discriminative have been","For image deblurring, however, discriminative approaches have been mostly lacking.",10031
"discriminative this features suitable is Second, prediction. to variability for construct it difficult quite given","Second, given this variability it is quite difficult to construct suitable features for discriminative prediction.",10032
"our Each cascade i.e. is stage of semi-parametric,","Each stage of our cascade is semi-parametric, i.e.",10033
"parameters depends on the instance-specific restoration it of as blur the the such problem, kernel.","it depends on the instance-specific parameters of the restoration problem, such as the blur kernel.",10034
model minimization by with train synthetically generated our training loss data. We,We train our model by loss minimization with synthetically generated training data.,10035
system the we of mutant of strains. study the the under behavior Finally injection,Finally we study the behavior of the system under the injection of mutant strains.,10036
"computing alternative Here, power. with an present classical inspired quantum method a I by","Here, I present an alternative inspired by a classical computing method with quantum power.",10037
distributed called relies hyperstrings. on This representations special method,This method relies on special distributed representations called hyperstrings.,10038
cooperation conditional mechanisms are Direct prevent to reciprocity dilemmas. social riding free in and important,Direct reciprocity and conditional cooperation are important mechanisms to prevent free riding in social dilemmas.,10039
population of the relationships are ecology. one interactions in Predator-prey studied most,Predator-prey relationships are one of the most studied interactions in population ecology.,10040
potentially higher pathways metabolite concentrations. infeasible to lead that could intermediate Alternative often flux carry,Alternative pathways that could potentially carry higher flux often lead to infeasible intermediate metabolite concentrations.,10041
The on video content. depends performance algorithm tracking,The tracking algorithm performance depends on video content.,10042
First Lucas-Tomasi is feature the detection Kanade- (KLT) improved tracking. object using,First the object detection is improved using Kanade- Lucas-Tomasi (KLT) feature tracking.,10043
tracking This selection evaluation. online by is supported an,This selection is supported by an online tracking evaluation.,10044
been experimented The approach public has video datasets. three on,The approach has been experimented on three public video datasets.,10045
model. model relaxation LP between establish a further and our LP conventional bound We the,We further establish a relaxation bound between our LP model and the conventional LP model.,10046
"we of experiments, the method for segmentation. interactive object task our In the demonstrate","In the experiments, we demonstrate our method for the task of interactive object segmentation.",10047
theory briefly for connections We discuss and existing the between LASSO. selection permutation,We briefly discuss connections between permutation selection and existing theory for the LASSO.,10048
versions conditions. more Further different could coexistence introduce ecological under models,Further versions could introduce more coexistence models under different ecological conditions.,10049
genes the entropy. studied and dimension were on Y-chromosome Shannon fractal human All using,All genes on the human Y-chromosome were studied using fractal dimension and Shannon entropy.,10050
the enriched also sequences lncRNA among were outliers.,lncRNA sequences were also enriched among the outliers.,10051
entropy; Shannon Keywords: centromeric . dimension; genes; gene degredation; lncRNA di-nucleotide Y-chromosome; fractal,Keywords: Y-chromosome; Shannon di-nucleotide entropy; fractal dimension; centromeric genes; gene degredation; lncRNA .,10052
node sets. This a gives model-based clustering of the,This gives a model-based clustering of the node sets.,10053
of to number real demonstrate bipartite Application the networks algorithms discussed. a observed,Application to a number of real observed bipartite networks demonstrate the algorithms discussed.,10054
"discriminativeness. maintaining inter-person while reduces It the intra-person dramatically variances, the","It dramatically reduces the intra-person variances, while maintaining the inter-person discriminativeness.",10055
"As is an verification. approach face recovery used face this application, for","As an application, this face recovery approach is used for face verification.",10056
on LFW dataset. the achieves state-of-the-art Our the approach performance,Our approach achieves the state-of-the-art performance on the LFW dataset.,10057
same may of loci these tameness and colocalize the reflect loci with genetic Several variants.,Several of these loci colocalize with tameness loci and may reflect the same genetic variants.,10058
"Emergency (EMS) dynamic paper, Service systems. address Medical we the In this","In this paper, we address the dynamic Emergency Medical Service (EMS) systems.",10059
and to tries presented relocate A that dynamic is locate ambulances. location the model,A dynamic location model is presented that tries to locate and relocate the ambulances.,10060
Pattern This topic challenges Recognition: main the year opportunities. interdisciplinary and is:,This year the main topic is: Pattern Recognition: interdisciplinary challenges and opportunities.,10061
"the model images. for computational Inspired a by saliency theory, develop in detection we","Inspired by the theory, we develop a computational model for saliency detection in images.",10062
"constitute a original First, is downsampled to to the image different pyramid. scales","First, the original image is downsampled to different scales to constitute a pyramid.",10063
An special graphs. of analysis mutant non-directed on of fixation classes probability a of the,An analysis of the fixation probability of a mutant on special classes of non-directed graphs.,10064
"can fast for inference. neural circuit and perform decoding plausible dynamics efficient sampling Thus,","Thus, plausible neural circuit dynamics can perform fast sampling for efficient decoding and inference.",10065
"and easily-parallelized a on Building ProPPR. for algorithm we develop weight-learning fast this,","Building on this, we develop a fast and easily-parallelized weight-learning algorithm for ProPPR.",10066
"been methods popularity gaining for AD Since learning related machine first introduced, have research. increasing","Since first introduced, machine learning methods have been gaining increasing popularity for AD related research.",10067
requires domain costly This and knowledge tedious. is procedure and,This procedure requires domain knowledge and is costly and tedious.,10068
"shortcomings, numbers a D theory proposed. these overcome theory To is novel called","To overcome these shortcomings, a novel theory called D numbers theory is proposed.",10069
proposed is An provided the function. demonstrate case to effectiveness illustrative the of,An illustrative case is provided to demonstrate the effectiveness of the proposed function.,10070
"However, to it is difficult solve. underconstrained and severely","However, it is severely underconstrained and difficult to solve.",10071
on resolution limits practical the uncomfortable can of creates we matte. This that images,This creates uncomfortable practical limits on the resolution of images that we can matte.,10072
"super-linear mitigates Current literature complexity. problem, the all they but in remain","Current literature mitigates the problem, but they all remain super-linear in complexity.",10073
properties How production understood. poorly environments varying is stiffness in control with force their,How their properties control force production in environments with varying stiffness is poorly understood.,10074
"a values properties Here, we isoforms into literature myosin incorporated II model. of cross-bridge for","Here, we incorporated literature values for properties of myosin II isoforms into a cross-bridge model.",10075
actin gliding from force-velocity curves speeds were Similar previous observed. expected and experiments,Similar actin gliding speeds and force-velocity curves expected from previous experiments were observed.,10076
"this action In a novel paper, propose framework. we recognition","In this paper, we propose a novel action recognition framework.",10077
robust complex The accurately. view interactions point to proposed SODA+MRF transformations is detect model and,The proposed SODA+MRF model is robust to view point transformations and detect complex interactions accurately.,10078
the (HH) classical neglects dynamics. model spiking Hodgkin--Huxley concentrations time-dependence in of The ion,The classical Hodgkin--Huxley (HH) model neglects the time-dependence of ion concentrations in spiking dynamics.,10079
can modulatory spiking by the that be regulate bifurcations. explained rate mechanisms Also,Also modulatory mechanisms that regulate the spiking rate can be explained by bifurcations.,10080
candidates A compatible by non-maximal suppression. subset then is obtained of,A compatible subset of candidates is then obtained by non-maximal suppression.,10081
biological the in  an interpretation play important Mathematical models experiments. of role increasingly ,  Mathematical models play an increasingly important role in the interpretation of biological experiments.,10082
Commonly simple symmetries. underlying patterns from observed derive,Commonly observed patterns derive from simple underlying symmetries.,10083
potential Ontology paper for ontologies rules. (Web OWL Language) the discloses of This of generation,This paper discloses the potential of OWL (Web Ontology Language) ontologies for generation of rules.,10084
Parameter discretely for processes observed estimation problem. is Markov a challenging,Parameter estimation for discretely observed Markov processes is a challenging problem.,10085
"straightforward Markov However, Gillespie using simulation the algorithm. is processes of","However, simulation of Markov processes is straightforward using the Gillespie algorithm.",10086
theory. of an the open in application issue management Conflict is evidence Shafer Dempster still,Conflict management is still an open issue in the application of Dempster Shafer evidence theory.,10087
of works been A have issue. this to presented address lot,A lot of works have been presented to address this issue.,10088
"this theory (GET), paper, theory, called proposed. evidence new as is In generalized a","In this paper, a new theory, called as generalized evidence theory (GET), is proposed.",10089
the handled framework evidence The conflicting is of under GET.,The conflicting evidence is handled under the framework of GET.,10090
annotated axoplasmic region are morphological a using algorithm. growing These then reticula,These axoplasmic reticula are then annotated using a morphological region growing algorithm.,10091
This method axoplasmic annotates reticula with high precision.,This method annotates axoplasmic reticula with high precision.,10092
is a challenging task. similarity Learning fine-grained image,Learning fine-grained image similarity is a challenging task.,10093
needs between-class image capture differences. and within-class It to,It needs to capture between-class and within-class image differences.,10094
A effectively. the has been developed describe multiscale network images structure to novel,A novel multiscale network structure has been developed to describe the images effectively.,10095
well-known. sparse The of representations is and capability potential,The potential and capability of sparse representations is well-known.,10096
"but correlation discriminative the ability some the of overlooks considers sparsity. Besides, work","Besides, some work considers the correlation but overlooks the discriminative ability of sparsity.",10097
"introduce adjusting detail. By different parameter, levels we of the smoothness","By adjusting the smoothness parameter, we introduce different levels of detail.",10098
The of is demonstrated object detection approach effectiveness with framework. Regionlets the proposed the,The effectiveness of the proposed approach is demonstrated with the Regionlets object detection framework.,10099
resulting result increase improved to metabolism. capability desired is The predictive understanding of in,The desired result is to increase predictive capability resulting in improved understanding of metabolism.,10100
"also implemented FALCON of Toolbox, in MATLAB. requires COBRA the use","FALCON requires use of the COBRA Toolbox, also implemented in MATLAB.",10101
Online area years. become feature features an selection in with dynamic active has recent research,Online feature selection with dynamic features has become an active research area in recent years.,10102
intra-group consists selection. inter-group and stages: approach online Our selection proposed online of two,Our proposed approach consists of two stages: online intra-group selection and online inter-group selection.,10103
"selection, we of to subset the use In inter-group globally optimal Lasso features. select a","In the inter-group selection, we use Lasso to select a globally optimal subset of features.",10104
Dempster-Shafer tool fusion. information is in powerful a evidence theory,Dempster-Shafer evidence theory is a powerful tool in information fusion.,10105
"presented. are results be conflicting, counter-intuitive will highly the the When evidence","When the evidence are highly conflicting, the counter-intuitive results will be presented.",10106
to Dempster's The is rule used evidence. the combine new combination,The Dempster's combination rule is used to combine the new evidence.,10107
"existing method is with proposed the the Comparing methods, new efficient.","Comparing with the existing methods, the new proposed method is efficient.",10108
"weighted edges the term. and implements infinite a smoothness set non-terminal Furthermore, of","Furthermore, a set of infinite weighted and non-terminal edges implements the smoothness term.",10109
"of is segmentation the determined Subsequently, source-set. result out the","Subsequently, the segmentation result is determined out of the source-set.",10110
"solutions In by efficiently dynamic be programming. satisfactory this way, can computed","In this way, satisfactory solutions can be efficiently computed by dynamic programming.",10111
the our show Experimental two approach. benchmark datasets of results performance state-of-the-art on,Experimental results on two benchmark datasets show the state-of-the-art performance of our approach.,10112
"oscillations . herbivory, stage-structure, insects, Keywords: pollination, Abstract mutualism,","Abstract Keywords: mutualism, pollination, herbivory, insects, stage-structure, oscillations .",10113
"those policies previous which it of produces methods, Additionally, such LSPI. as outperform","Additionally, it produces policies which outperform those of previous methods, such as LSPI.",10114
arises behavioural many interplay through of a different complicated features. It,It arises through a complicated interplay of many different behavioural features.,10115
notions in depth important analysis. data are regions Tukey nonparametric multivariate,Tukey depth regions are important notions in nonparametric multivariate data analysis.,10116
"very are algorithms practically but available, slow. Feasible","Feasible algorithms are available, but practically very slow.",10117
ones. that Data the runs indicate proposed algorithm examples the much faster than existing,Data examples indicate that the proposed algorithm runs much faster than the existing ones.,10118
ecology of Identifying the and in climate critical research change conservation. for effects is future,Identifying the effects of climate change is critical for future research in ecology and conservation.,10119
an focus paper. reconstruction to method this referred Bayesian in on We interpolation,We focus on an interpolation method referred to Bayesian reconstruction in this paper.,10120
topic is important systems (GRNs) gene networks in regulatory Modeling biology. an,Modeling gene regulatory networks (GRNs) is an important topic in systems biology.,10121
we of core large GRNs. motifs for the dynamic that Then the searched govern behavior,Then we searched for the core motifs that govern the dynamic behavior of large GRNs.,10122
provide design networks. study and insights of Our genetic for results the,Our results provide insights for the study and design of genetic networks.,10123
better and enabling advances methodological and of are processes extinction patterns. Recent speciation examination,Recent methodological advances are enabling better examination of speciation and extinction processes and patterns.,10124
affect Yet it and differences seems the lineages mass same that groups. extinctions both probable,Yet it seems probable that both lineages differences and mass extinctions affect the same groups.,10125
"we shifts are find lineage extinctions. Overall, detected that than better mass","Overall, we find that lineage shifts are better detected than mass extinctions.",10126
"al. presented and (Dai algorithm et TurKontrol, in","TurKontrol, and algorithm presented in (Dai et al.",10127
TurKPF perform TurKontrol of that and see series experiments we a Through similarly.,Through a series of experiments we see that TurKPF and TurKontrol perform similarly.,10128
then hypotheses of for restraint. classical We consider the evolution,We then consider classical hypotheses for the evolution of restraint.,10129
to use same concepts. for ontologies different tend same different Two terms knowledge the assimilating,Two different ontologies assimilating same knowledge tend to use different terms for the same concepts.,10130
which works technique used core system. have the graph the at matching We a of,We have used a graph matching technique which works at the core of the system.,10131
"relies bulk different which nucleic properties acid exhibits BOINC, from sequencing. FISSEQ-BOINC on","FISSEQ-BOINC exhibits different properties from BOINC, which relies on bulk nucleic acid sequencing.",10132
rich with could annotations. a FISSEQ-BOINC scalable molecular for approach connectomes whole-mammalian-brain mapping become,FISSEQ-BOINC could become a scalable approach for mapping whole-mammalian-brain connectomes with rich molecular annotations.,10133
and are physical governed fluxes Metabolic economic by principles.,Metabolic fluxes are governed by physical and economic principles.,10134
"penalise To larger methods terms. some prediction heuristic flux by fluxes cost capture this,","To capture this, some flux prediction methods penalise larger fluxes by heuristic cost terms.",10135
"from must lead Generally, fluxes to economic potentials. lower higher","Generally, fluxes must lead from lower to higher economic potentials.",10136
"analysis. complement can thermodynamic flux constraints, and resembles stationarity thermodynamic in This principle, which constraints","This principle, which resembles thermodynamic constraints, can complement stationarity and thermodynamic constraints in flux analysis.",10137
Enzymes of potential in likely that modes targets participate are regulation. futile,Enzymes that participate in potential futile modes are likely targets of regulation.,10138
"interest constraints. there information growing with Recently, has been modeling planning in a","Recently, there has been a growing interest in modeling planning with information constraints.",10139
method global solution is thus and efficient. The the proposed gives analytically computationally,The proposed method gives the global solution analytically and thus is computationally efficient.,10140
the existing method that results methods. clustering further outperforms proposed Experimental indicate,Experimental results further indicate that the proposed method outperforms existing clustering methods.,10141
examples. give describe reference of desirable properties structures the Here we and,Here we describe the desirable properties of reference structures and give examples.,10142
produces that results burden. The results EP show comparable less computational with,The results show that EP produces comparable results with less computational burden.,10143
Metabolic a systems compromise cost. between metabolic enzyme benefit and governed by are,Metabolic systems are governed by a compromise between metabolic benefit and enzyme cost.,10144
"is lagging. models these networks study general more analytical However, of Boolean","However, analytical study of these more general Boolean networks models is lagging.",10145
"In addition, exists between and knowledge $\mathcal{ALC}$+T$^+_{\mathsf{min}}$ we connection close concept-circumscribed the that bases. show","In addition, we show the close connection that exists between $\mathcal{ALC}$+T$^+_{\mathsf{min}}$ and concept-circumscribed knowledge bases.",10146
"in reasoning the we Finally, study the tasks complexity deciding classical $\mathcal{ALC}$+T$^+_{\mathsf{min}}$. of","Finally, we study the complexity of deciding the classical reasoning tasks in $\mathcal{ALC}$+T$^+_{\mathsf{min}}$.",10147
"Where, when remains this integration performed is and unclear. how","Where, when and how this integration is performed remains unclear.",10148
clustering area an data important mining. Data of is,Data clustering is an important area of data mining.,10149
means very on Fuzzy C logic. fuzzy a based is clustering technique important,Fuzzy C means is a very important clustering technique based on fuzzy logic.,10150
have techniques clustering we K-means like popular hard some Also available among ones. the,Also we have some hard clustering techniques available like K-means among the popular ones.,10151
vision in recognition for field is Activity computer sport attractive an research.,Activity recognition in sport is an attractive field for computer vision research.,10152
single with emphasis activities recognition on paper of volleyball. describes player in sport This special,This paper describes recognition of single player activities in sport with special emphasis on volleyball.,10153
neuroscience. established becoming more approaches modeling are ever Mathematical in clinical Background:,Background: Mathematical modeling approaches are becoming ever more established in clinical neuroscience.,10154
Our approach measure interaction of information the on relies (II).,Our approach relies on the measure of interaction information (II).,10155
Filters Correlation accurate a localization. for designed pattern which of are (CFs) classifiers class are,Correlation Filters (CFs) are a class of classifiers which are designed for accurate pattern localization.,10156
multi-spectral (Meteosat zones data by Second Generation). exploiting satellite these implemented is of Correction,Correction of these zones is implemented by exploiting multi-spectral satellite data (Meteosat Second Generation).,10157
two metaheuristic for This the presents algorithms article CluVRP. hybrid alternative,This article presents two alternative hybrid metaheuristic algorithms for the CluVRP.,10158
choices cluster average size. on promising algorithm to relatively are Recommendations provided,Recommendations on promising algorithm choices are provided relatively to average cluster size.,10159
"Recent much by works low attract attention. sparse representation extensions using representation, rank their and","Recent works by using sparse representation, low rank representation and their extensions attract much attention.",10160
their among are functions. objective differences main them The,The main differences among them are their objective functions.,10161
"orthogonal. sampling be can the Furthermore, data are if the subspaces insufficient","Furthermore, the data sampling can be insufficient if the subspaces are orthogonal.",10162
special cases. all are methods Some existing,Some existing methods are all special cases.,10163
present subspace segmentation. Regression Squares the method Least we for Then (LSR),Then we present the Least Squares Regression (LSR) method for subspace segmentation.,10164
"It is advantage which data data. common in takes real of correlation,","It takes advantage of data correlation, which is common in real data.",10165
together. tends effect encourages grouping which a group data LSR correlated highly to,LSR encourages a grouping effect which tends to group highly correlated data together.,10166
"accuracy, more experiments much Beyond all efficient. that LSR segmentation demonstrate is","Beyond segmentation accuracy, all experiments demonstrate that LSR is much more efficient.",10167
mechanical of the waves potentials. complicated of than those movements The were evoked more,The waves of mechanical movements were more complicated than those of the evoked potentials.,10168
The the in electrical mechanical stimulation mechanisms propagated of discussed. brain by movements are,The mechanisms of mechanical movements propagated in the brain by electrical stimulation are discussed.,10169
technological architectural feature Bow-tie and common a or biological found networks. in hourglass is structure,Bow-tie or hourglass structure is a common architectural feature found in biological and technological networks.,10170
"bow-tie is Little how however, evolve. architectures about known,","Little is known, however, about how bow-tie architectures evolve.",10171
robustly These lead conditions bow-tie structures. to,These conditions robustly lead to bow-tie structures.,10172
(FACS); action information; action detection; unit system relative coding Keywords: . temporal facial facial,Keywords: facial action coding system (FACS); relative facial action unit detection; temporal information; .,10173
"than exception, were With less compounds potent WIN. significantly against one the CIT","With one exception, the compounds were significantly less potent against CIT than WIN.",10174
State-of-the-art lower-bound scenarios for methods of problems estimation generate batches the SAA independently. for,State-of-the-art methods for lower-bound estimation generate batches of scenarios for the SAA problems independently.,10175
liquid-air The and applied method can to recognition liquid-liquid surfaces. both be of,The method can be applied to recognition of both liquid-air and liquid-liquid surfaces.,10176
number the of knowledge No is phases prior required. of,No prior knowledge of the number of phases is required.,10177
requires travel the CM-NPC-DNA Development messenger ligands sites. from to binding target,Development requires messenger ligands to travel from the CM-NPC-DNA target binding sites.,10178
"maximum approaches Thus non-equilibrium, condition life. information for one","Thus maximum information approaches non-equilibrium, one condition for life.",10179
were in Introgressions detected Neanderthals from modern and Denisovans humans.,Introgressions from Neanderthals and Denisovans were detected in modern humans.,10180
history among admixture suggested hominins. Above complex a results,Above results suggested a complex admixture history among hominins.,10181
could approach The proposed research species. facilitate across also admixture,The proposed approach could also facilitate admixture research across species.,10182
"morphology of providing a graphs. anthology at survey on This ""literary"" paper aims mathematical","This survey paper aims at providing a ""literary"" anthology of mathematical morphology on graphs.",10183
and Under is inconsistent. information knowledge the everything is standard if entailed semantics two-valued,Under the standard two-valued semantics everything is entailed if knowledge and information is inconsistent.,10184
logic offers of LP the semantics paraconsistent a The solution.,The semantics of the paraconsistent logic LP offers a solution.,10185
three-valued Preferring difference. eliminates interpretations this conflict-minimal,Preferring conflict-minimal three-valued interpretations eliminates this difference.,10186
"paper To the argumentation an proposes handle non-monotonicity, this assumption-based system.","To handle the non-monotonicity, this paper proposes an assumption-based argumentation system.",10187
the branches of tableaux a Assumptions arguments. needed form close semantic to,Assumptions needed to close branches of a semantic tableaux form the arguments.,10188
modelling Deontic shown reasoning. human is logic to be applicable for,Deontic logic is shown to be applicable for modelling human reasoning.,10189
discussed detail. task in the this For task and selection Wason are suppression the,For this the Wason selection task and the suppression task are discussed in detail.,10190
"of Most the (e.g. recovery fixed of approaches, a conventional CS exploited however, bases set","Most of the conventional CS recovery approaches, however, exploited a set of fixed bases (e.g.",10191
"resulting quantiza-tion the (SQ). is residual Then, scalar processed by","Then, the resulting residual is processed by scalar quantiza-tion (SQ).",10192
are studied AI. the argumentation frameworks most of formalisms in one (AFs) Abstract,Abstract argumentation frameworks (AFs) are one of the most studied formalisms in AI.,10193
"certain this a call work, of subclass which compact. we we In AFs introduce","In this work, we introduce a certain subclass of AFs which we call compact.",10194
"controllers are generic, application-specific. Logics are whereas","Logics are generic, whereas controllers are application-specific.",10195
"nonetheless, an does, for Every nonmonotonic controller algorithm revision. belief provide","Every controller does, nonetheless, provide an algorithm for nonmonotonic belief revision.",10196
"conventional CS set (e.g. however, a fixed bases recovery exploited approaches, Most of of","Most of conventional CS recovery approaches, however, exploited a set of fixed bases (e.g.",10197
"with the a model Dead. consider B compartments: three-state we A, and Here","Here we consider a three-state model with the compartments: A, B and Dead.",10198
develop We model. Methods: simplifies an equation the three-state of that the use,Methods: We develop an equation that simplifies the use of the three-state model.,10199
has and The in a public applicability broad equation health. epidemiology,The equation has a broad applicability in epidemiology and public health.,10200
"waves spontaneous, stimuli, developmental receiving called activity Prior correlated activity-dependent drives to programs. retinal visual","Prior to receiving visual stimuli, spontaneous, correlated activity called retinal waves drives activity-dependent developmental programs.",10201
"mediated bursts potentials. waves of spreading manifest Early-stage slow, as (ACh) action by acetylcholine","Early-stage waves mediated by acetylcholine (ACh) manifest as slow, spreading bursts of action potentials.",10202
of by that modulated concentration may be cholinergic evidence the waves Recent suggests ACh. extracellular,Recent evidence suggests that cholinergic waves may be modulated by the extracellular concentration of ACh.,10203
"address to this over represent we introduce of trust. degrees pseudometrics states To problem, differing","To address this problem, we introduce pseudometrics over states to represent differing degrees of trust.",10204
(ASP) is Set programming Answer Programming a declarative paradigm.,Answer Set Programming (ASP) is a declarative programming paradigm.,10205
improvements ASP WASP. the paper the on reports This recent of solver,This paper reports on the recent improvements of the ASP solver WASP.,10206
competitive The on not of often with class. was solutions alternative version WASP previous this,The previous version of WASP was often not competitive with alternative solutions on this class.,10207
substantial shows new a increase The performance. in of version WASP,The new version of WASP shows a substantial increase in performance.,10208
random are used multivariate modeling commonly models tools Graphical for variables.,Graphical models are commonly used tools for modeling multivariate random variables.,10209
propose less We than images. using two an recovering algorithm depth for,We propose an algorithm for recovering depth using less than two images.,10210
without is the aspect all. key cameras The at that send information the communicating,The key aspect is that the cameras send the information without communicating at all.,10211
"Hence, reduced. bandwidth communication the significantly is required","Hence, the required communication bandwidth is significantly reduced.",10212
placed burden The on is decoding decoder of the (host).,The burden of decoding is placed on the decoder (host).,10213
"depth be can Using the map right and the synthesized. image image, the left","Using the depth map and the left image, the right image can be synthesized.",10214
"was framework open-source. made To further tracking developments, our encourage","To encourage further developments, our tracking framework was made open-source.",10215
is on essay In the stance  robots discussed.  this,  In this essay the stance on robots is discussed.,10216
given. explanations uncanny The some valley possible and are,The uncanny valley and some possible explanations are given.,10217
"of stable. accompanied grounded, variety by a It is and semantics including preferred complete,","It is accompanied by a variety of semantics including grounded, complete, preferred and stable.",10218
"shortcomings, led which numerous have enrichments. their Although of powerful, development AFs to","Although powerful, AFs have their shortcomings, which led to development of numerous enrichments.",10219
"the ADFs. known also as general abstract dialectical the Among frameworks, are most ones the","Among the most general ones are the abstract dialectical frameworks, also known as the ADFs.",10220
of use arbitrary relations. They to make so-called represent the acceptance conditions,They make use of the so-called acceptance conditions to represent arbitrary relations.,10221
"controversial argumentation, issues, in cycles. most support of the the concerns only recognized not One","One of the most controversial issues, recognized not only in argumentation, concerns the support cycles.",10222
"Moreover, known Dung ADF from versions setting. of properties the provide the we","Moreover, we provide ADF versions of the properties known from the Dung setting.",10223
Several the variants have of been chase proposed.,Several variants of the chase have been proposed.,10224
"probabilistic implemented Prolog. in been knowledge which has bases,","probabilistic knowledge bases, which has been implemented in Prolog.",10225
many biological fundamental are to processes. Protein-protein interactions,Protein-protein interactions are fundamental to many biological processes.,10226
of information record. protein interactions the An alternative sequence is about source rich evolutionary,An alternative rich source of information about protein interactions is the evolutionary sequence record.,10227
"learns training, NLS. of During that sequence minimize maps decent a the SDM","During training, SDM learns a sequence of decent maps that minimize the NLS.",10228
the the under which guaranteed is prove We to conditions SDM converge.,We prove the conditions under which the SDM is guaranteed to converge.,10229
www.humansensing.cs.cmu.edu/intraface. has code been available at The made,The code has been made available at www.humansensing.cs.cmu.edu/intraface.,10230
Examples illustrate introduced future at topics point for and concepts and research. issues the,Examples illustrate the introduced concepts and point at issues and topics for future research.,10231
tree. are likelihood widely (MLE) used for estimation methods Maximum evolutionary,Maximum likelihood estimation (MLE) methods are widely used for evolutionary tree.,10232
Other under proofs were proposed. of models various consistency also,Other proofs of consistency under various models were also proposed.,10233
is MEASP solver programs. ASP ground multi-engine a for,MEASP is a multi-engine solver for ground ASP programs.,10234
rank-SVM ranked We this a then train data. using,We then train a rank-SVM using this ranked data.,10235
member distinct 'wholes.' multiple allows of This be to for 'parts' a individual,This allows for individual 'parts' to be a member of multiple distinct 'wholes.',10236
"operator define possible Moreover, the it this to Harper thanks contraction a makes to identity.","Moreover, this makes it possible to define a contraction operator thanks to the Harper identity.",10237
in aid Objects labeled to per-instance precise are using localization. object segmentations,Objects are labeled using per-instance segmentations to aid in precise object localization.,10238
"for of multiple comparison batch requiring effects needed Therefore, are of samples. mechanisms detection","Therefore, mechanisms for detection of batch effects are needed requiring comparison of multiple samples.",10239
DNA (HC) hierarchical RNA-seq derived apply clustering of We multiple k-mer files. counts Fastq on,We apply hierarchical clustering (HC) on DNA k-mer counts of multiple RNA-seq derived Fastq files.,10240
"reads the human present samples, in to batch mapped genome. our In effects also were","In our samples, batch effects were also present in reads mapped to the human genome.",10241
to used can net sentence. a A semantic be represent,A semantic net can be used to represent a sentence.,10242
"positivity, neutrality truth, and and negativity. triads These and include falsehood, also so indeterminacy","These triads include truth, indeterminacy and falsehood, and so also positivity, neutrality and negativity.",10243
representation paper demonstrates method The computers in a memory. a polarity the of for,The paper demonstrates a method for the representation of polarity in a computers memory.,10244
"augmentation. P\'olya-Gamma from sampling element is ${PG}(b,z)$, an essential the P\'olya-Gamma of distribution, Efficiently data","Efficiently sampling from the P\'olya-Gamma distribution, ${PG}(b,z)$, is an essential element of P\'olya-Gamma data augmentation.",10245
The separately density with points dimensions. are in computed twice-projected ratio local,The points are computed with local density ratio separately in twice-projected dimensions.,10246
"outliers the values of largest points process, are weight. the those After scoring","After the process, outliers are those points scoring the largest values of weight.",10247
inverted using into hashing These locality-sensitive index are an (LSH). then hashed,These are then hashed into an inverted index using locality-sensitive hashing (LSH).,10248
"low difficult. contrast We footage, Airborne makes examine and more whose extraction motion CCTV","We examine CCTV and Airborne footage, whose low contrast makes motion extraction more difficult.",10249
"contrasts handles low long occlusion. Our routes, approach and","Our approach handles long routes, low contrasts and occlusion.",10250
Programming (TPLP). appear in Practice and Logic   To of Theory,  To appear in Theory and Practice of Logic Programming (TPLP).,10251
Trait differences species selection. be to may natural between attributable,Trait differences between species may be attributable to natural selection.,10252
"multiple-line to a scenarios. allows different distinguish lineage-specific Second, selection test","Second, a multiple-line test allows to distinguish different lineage-specific selection scenarios.",10253
by analytical extensive numerical complemented Our results simulations. are,Our analytical results are complemented by extensive numerical simulations.,10254
functional and an cartoon of the texture components energy of Minimizers derived are image.,Minimizers of the derived energy functional are cartoon and texture components of an image.,10255
XSB the in implemented integrated Prolog. of A system prototype is,A prototype of the integrated system is implemented in XSB Prolog.,10256
tool Recently face becomes analysis recognition. for regression a popular,Recently regression analysis becomes a popular tool for face recognition.,10257
that error lead illumination to generally We low-rank observe occlusion image. changes a and,We observe that occlusion and illumination changes generally lead to a low-rank error image.,10258
"regression, and for matrix classification. face representation","matrix regression, for face representation and classification.",10259
for has application. developed system of Various wide expert range been,Various expert system has been developed for wide range of application.,10260
with Protein-protein of prominence have as associated area an research. networks gained diseases interaction,Protein-protein interaction networks associated with diseases have gained prominence as an area of research.,10261
"families (e.g. protein Moreover, several we identify","Moreover, we identify several protein families (e.g.",10262
"in many ITG, PIK, AKT cancer motifs of that are families) repeated pathways. the","PIK, ITG, AKT families) that are repeated motifs in many of the cancer pathways.",10263
"symmetry rather central are often of sources than these peripheral. Interestingly,","Interestingly, these sources of symmetry are often central rather than peripheral.",10264
in Our drugs. for can promising of aide targets anti-cancer results identification,Our results can aide in identification of promising targets for anti-cancer drugs.,10265
"disorders). neurodegenerative viral diseases, substance abuse diseases,","neurodegenerative diseases, viral diseases, substance abuse disorders).",10266
rod-shaped many bacteria homogenous. are far proteins of The from in distributions,The distributions of many proteins in rod-shaped bacteria are far from homogenous.,10267
accumulate the poles cell they Often or in the center. cell at,Often they accumulate at the cell poles or in the cell center.,10268
"distortion caused Furthermore, examined. imperfect the by rendering is also","Furthermore, the distortion caused by imperfect rendering is also examined.",10269
of native be languages. must their regardless accessible Knowledge everyone to,Knowledge must be accessible to everyone regardless of their native languages.,10270
a for machine translation. language from semantic UNL extracts data Interlingua native,UNL extracts semantic data from a native language for Interlingua machine translation.,10271
the UNL represents as sentence semantics conceptual a of hyper-graph. a,UNL represents the semantics of a sentence as a conceptual hyper-graph.,10272
"tracking across Specifically, neural is precisely slices. processes difficult, many especially image","Specifically, precisely tracking neural processes is difficult, especially across many image slices.",10273
is author's available the computations made to carry the website. out code Computer on,Computer code to carry out the computations is made available on the author's website.,10274
We humans image using pose address of of the RGB the problem input. estimating,We address the problem of estimating the pose of humans using RGB image input.,10275
"using i.e. input, However, are we pure RGB","However, we are using pure RGB input, i.e.",10276
make encodings functions. number used system for SAT to Transalg cryptographic the of We a,We used the Transalg system to make SAT encodings for a number of cryptographic functions.,10277
problem recognition Point pattern common a constellation is with many applications. matching,Point constellation recognition is a common problem with many pattern matching applications.,10278
"mainly by fingerprint useful contexts, in is matching. Whilst motivated work many this","Whilst useful in many contexts, this work is mainly motivated by fingerprint matching.",10279
of as oriented traditionally minutiae. are called Fingerprints modelled constellations points,Fingerprints are traditionally modelled as constellations of oriented points called minutiae.,10280
constellations. point task The verifier's comparing fingerprint in two consists,The fingerprint verifier's task consists in comparing two point constellations.,10281
This new paper constellation algorithms. presents three matching,This paper presents three new constellation matching algorithms.,10282
two an The and Despiegel. Bringer first generalize by algorithm methods,The first two methods generalize an algorithm by Bringer and Despiegel.,10283
"and is this, In image. added the noise in resolution input low contrast low the","In this, the noise is added in the input low resolution and low contrast image.",10284
the remove used input The from filter noise median image. is,The median filter is used remove noise from the input image.,10285
enhance SWT image. used input is to,SWT is used to enhance input image.,10286
All high using resolution these reconstructed image as by are sub-bands DWT (IDWT). inverse,All these sub-bands are reconstructed as high resolution image by using inverse DWT (IDWT).,10287
the SVD. uses DWT proposed contrast technique and To increase the,To increase the contrast the proposed technique uses DWT and SVD.,10288
equalize used is image. GHE to an,GHE is used to equalize an image.,10289
over the conventional superiority techniques. proposed result experimental technique of shows The,The experimental result shows superiority of the proposed technique over conventional techniques.,10290
background The study evaluates subtraction techniques. three,The study evaluates three background subtraction techniques.,10291
levels ranges from and algorithms used in of computational the varying study The accuracy complexity.,The algorithms used in the study ranges from varying levels of accuracy and computational complexity.,10292
under logic (ASP) answer Set stable Answer is model set the or semantics. Programming programming,Answer Set Programming (ASP) is logic programming under the stable model or answer set semantics.,10293
"quantifiers, HEX atoms, Among and are constraints. there abstract atoms, these, aggregate generalized","Among these, there are aggregate atoms, HEX atoms, generalized quantifiers, and abstract constraints.",10294
refer we generalized In these to atoms. paper as constructs this collectively,In this paper we refer to these constructs collectively as generalized atoms.,10295
"patient this into only microphone. speak computer Using a a technology, need","Using this technology, a patient need only speak into a computer microphone.",10296
"the a machine what's on, of presented from going learning gist perspective. Here's","Here's the gist of what's going on, presented from a machine learning perspective.",10297
"is inference daunting and task. particularly learning a performing domains in hybrid However,","However, performing inference and learning in hybrid domains is a particularly daunting task.",10298
scenarios. We real validate method artificial our on and world,We validate our method on artificial and real world scenarios.,10299
networks: networks. of types We hybridization consider two HGT and reticulation networks,We consider two types of reticulation networks: hybridization networks and HGT networks.,10300
"HGT we analogous demonstrate For more results. some but networks, intricate","For HGT networks, we demonstrate some analogous but more intricate results.",10301
this prove consistent its population fixed-$N$ estimator that for counterpart. We is,We prove that this estimator is consistent for its fixed-$N$ population counterpart.,10302
results example. a on The illustrated are numerical,The results are illustrated on a numerical example.,10303
using both validated method synthetically . The is,The method is validated using both synthetically .,10304
of will output image union mosaic the of two the be images. input The,The output of the image mosaic will be the union of two input images.,10305
Image-mosaicing mosaiced get to are algorithms image. used,Image-mosaicing algorithms are used to get mosaiced image.,10306
"Homography Blending computation, registration, extraction, Which Feature point Warping includes; and Image. Image if","Which includes; Feature point extraction, Image registration, Homography computation, Warping and Blending if Image.",10307
algorithm Various Feature detection for being extraction. used corner is,Various corner detection algorithm is being used for Feature extraction.,10308
This mosaiced produces output and an image. informative corner efficient,This corner produces an efficient and informative output mosaiced image.,10309
"the ie. issue small of estimating probabilities Considering p,","Considering the issue of estimating small probabilities p, ie.",10310
an of the nested However appropriate issue. sequence of remains subsets open definition an,However the definition of an appropriate sequence of nested subsets remains an open issue.,10311
"flow. to behavior each implements Besides, determine species migration optimal reflexive the strategy","Besides, each species implements reflexive behavior strategy to determine the optimal migration flow.",10312
to The be idea for planning. search can memorize during used visited naturally states same,The same idea can be naturally used to memorize visited states during search for planning.,10313
effective the and learn parameters of proposes to This paper MST-AOG. methods structure,This paper proposes effective methods to learn the structure and parameters of MST-AOG.,10314
action The from MST-AOG based on novel inference recognition views. enables,The inference based on MST-AOG enables action recognition from novel views.,10315
and verified experimentally benchmark depth database. combination face The of using is intensity map,The combination of intensity and depth map is verified experimentally using benchmark face database.,10316
method multimodal modality. experimental proposed show is the The results individual better that than,The experimental results show that the proposed multimodal method is better than individual modality.,10317
"huge the originating result readings in from Data and datasets. increasingly sensor Web, media social","Data originating from the Web, sensor readings and social media result in increasingly huge datasets.",10318
"propose specifically, More approach evaluate the using a and parallel framework. MapReduce we","More specifically, we propose and evaluate a parallel approach using the MapReduce framework.",10319
and Theory of Logic To Programming Practice (TPLP). appear in,To appear in Theory and Practice of Logic Programming (TPLP).,10320
a problem convex under variational constrained The formulation. is addressed,The problem is addressed under a variational convex constrained formulation.,10321
balanced term total is term regularizer. a as This variation with,This term is balanced with a total variation term as regularizer.,10322
in many has effective factorization classification proved tasks. (NMF) clustering and Non-negative matrix,Non-negative matrix factorization (NMF) has proved effective in many clustering and classification tasks.,10323
"use nonlinear cases measures. handled these are However, not when we error properly","However, nonlinear cases are not properly handled when we use these error measures.",10324
"alternative measures on As as nonlinear consequence, based kernels, a proposed. correntropy, such are","As a consequence, alternative measures based on nonlinear kernels, such as correntropy, are proposed.",10325
"data. corresponding feature Meanwhile, the kernel each learn can from","Meanwhile, each feature can learn corresponding kernel from the data.",10326
is in Feature video task computer in a tracking crucial vision.,Feature tracking in video is a crucial task in computer vision.,10327
used is biometric widely a approach. Face recognition,Face recognition is a widely used biometric approach.,10328
to face are attacks non-real vulnerable systems But by spoof faces. made recognition,But face recognition systems are vulnerable to spoof attacks made by non-real faces.,10329
secure Liveness to such against spoofing. system needs order detection guard A in,A secure system needs Liveness detection in order to guard against such spoofing.,10330
their the helps categorization relation developed to attacks This solutions. scenarios different understanding spoof and,This categorization helps understanding different spoof attacks scenarios and their relation to the developed solutions.,10331
the liveness A detection works is presented. of regarding latest face works review,A review of the latest works regarding face liveness detection works is presented.,10332
(e.g. link hierarchical and the processing between organisation Understanding the,Understanding the link between the hierarchical organisation and processing (e.g.,10333
remains of the one field. the challenges main of,remains one of the main challenges of the field.,10334
"or morphology, pattern. vs. expression gene inhibitory excitatory),","inhibitory vs. excitatory), morphology, or gene expression pattern.",10335
It health cell environment influences well that in and disease. is functions mechanical established the,It is well established that the mechanical environment influences cell functions in health and disease.,10336
are to approaches neighbor Product quantization-based approximate effective for nearest points search. high-dimensional encode data,Product quantization-based approaches are effective to encode high-dimensional data points for approximate nearest neighbor search.,10337
traditional more distortion provide lower methods. and flexibility This than can errors,This can provide more flexibility and lower distortion errors than traditional methods.,10338
"populations. for systems in example, coupled of neuronal Activity often firing pattern oscillatory, the is","Activity in coupled systems is often oscillatory, for example, the firing pattern of neuronal populations.",10339
"occurred. peak ratios individual subsequent neurons, For between multiple with non-integer frequencies peaks","For individual neurons, multiple peak frequencies with non-integer ratios between subsequent peaks occurred.",10340
representation with The $\mathcal{A}log$ presents which knowledge aggregates. ASP a extends paper language,The paper presents a knowledge representation language $\mathcal{A}log$ which extends ASP with aggregates.,10341
also and shallow that aspects We successfully methods can shared. deep identify be of,We also identify aspects of deep and shallow methods that can be successfully shared.,10342
models experiments Source made the the publicly in available. code paper is and to reproduce,Source code and models to reproduce the experiments in the paper is made publicly available.,10343
"focus our portfolio and framework, black-box we adaptive selection. on online In algorithm","In our framework, we focus on black-box algorithm portfolio and online adaptive selection.",10344
classes. problem provide can across selection a confirm even naive improved that We strategy performance,We confirm that even a naive selection strategy can provide improved performance across problem classes.,10345
arguments counterarguments. offers and evaluating representing of way and an Abstract appealing argumentation,Abstract argumentation offers an appealing way of representing and evaluating arguments and counterarguments.,10346
This each enhanced assignment be by approach to argument. can probability a,This approach can be enhanced by a probability assignment to each argument.,10347
assignment. can this various There are be to interpretations that ascribed,There are various interpretations that can be ascribed to this assignment.,10348
assignment. consider constraints probability We on various the,We consider various constraints on the probability assignment.,10349
factors. added Two for GC-FOVE new are heterogeneous to operators treating,Two new operators are added to GC-FOVE for treating heterogeneous factors.,10350
"the very popular. ACT-R In science, is cognitive cognitive computational architecture","In computational cognitive science, the cognitive architecture ACT-R is very popular.",10351
implementation results first in such of mechanism. the one This,This results in the first implementation of one such mechanism.,10352
an statistics. important in models role play semiparametric Additive,Additive models play an important role in semiparametric statistics.,10353
for models. learning kernel gives rates based This regularized for paper methods additive,This paper gives learning rates for regularized kernel based methods for additive models.,10354
images natural of Traditional usually suffer patch-based modeling from two sparse problems. representation,Traditional patch-based sparse representation modeling of natural images usually suffer from two problems.,10355
still Efficient issue. is an world information open of modeling uncertain in real,Efficient modeling of uncertain information in real world is still an open issue.,10356
of Dempster-Shafer theory commonly most one is methods. evidence used the,Dempster-Shafer evidence theory is one of the most commonly used methods.,10357
is numbers combination of D two presented. rule The,The combination rule of two D numbers is presented.,10358
presented. of is combination D numbers rule The two,The combination rule of two D numbers is presented.,10359
"Finally, D transformation linguistic of decision. variables make presented a numbers a is to","Finally, a linguistic variables transformation of D numbers is presented to make a decision.",10360
The expression of networks by regulation of genes. is transcriptional interacting complex orchestrated gene,The transcriptional regulation of gene expression is orchestrated by complex networks of interacting genes.,10361
"ESmodels. we algorithms basic in describe and Then, optimization approaches used","Then, we describe basic algorithms and optimization approaches used in ESmodels.",10362
"we applications After constraint of in satisfaction. and ESmodels conformant discuss planning possible that,","After that, we discuss possible applications of ESmodels in conformant planning and constraint satisfaction.",10363
Each environmentally is is plan assessed and its footprint evaluated.,Each plan is environmentally assessed and its footprint is evaluated.,10364
neural convolutional speeding this focus paper the evaluation of The up is networks. of,The focus of this paper is speeding up the evaluation of convolutional neural networks.,10365
algorithm large regression propose problems. two for stage LING scale We a new,We propose a new two stage algorithm LING for large scale regression problems.,10366
"to that contributes to this finding patterns study some meaningful are Additionally, management.","Additionally, this study contributes to finding some patterns that are meaningful to management.",10367
"the existence, of stable and cumulativity. The lacks relevance properties model (SM) semantics","The stable model (SM) semantics lacks the properties of existence, relevance and cumulativity.",10368
dealt this of with paper. is the issue type in That,That is the type of issue dealt with in this paper.,10369
for transcription various develop factor gene a framework detailed oscillators. of types We theoretical,We develop a detailed theoretical framework for various types of transcription factor gene oscillators.,10370
image. a the enlarging is spatial zooming Image resolution digital process of given the of,Image zooming is the process of enlarging the spatial resolution of a given digital image.,10371
replication We technique zooming. classical pixel novel that present a the method intelligently modifies for,We present a novel technique that intelligently modifies the classical pixel replication method for zooming.,10372
and by mutations are eliminated natural selection. eventually new Most are deleterious,Most new mutations are deleterious and are eventually eliminated by natural selection.,10373
of To in (TPLP). Practice appear Logic Programming and Theory,To appear in Theory and Practice of Logic Programming (TPLP).,10374
tissue experimental to cortical curves data. Real and folds according,Real cortical tissue curves and folds according to experimental data.,10375
"only unfolded However, simulations use our layers. current cortical","However, our current simulations only use unfolded cortical layers.",10376
"reasoning syllogistic prove to Barston effect. out a Pollard Evans, this task carried psychological and","Evans, Barston and Pollard carried out a psychological syllogistic reasoning task to prove this effect.",10377
whether they were would asked reject a Participants accept given syllogism. or,Participants were asked whether they would accept or reject a given syllogism.,10378
formalization about reasoning. extensions abductive in Our new reveals possible questions,Our formalization reveals new questions about possible extensions in abductive reasoning.,10379
weaker support context. the by consequences prior some of indicates abductive relevant a notion A,A weaker notion indicates the support of some relevant consequences by a prior abductive context.,10380
"a $\mathcal{EL_{\bot}}$. on present new method paper, In adaptation this the we for DL based","In this paper, we present a new method for adaptation based on the DL $\mathcal{EL_{\bot}}$.",10381
"various very remains spurious which this a uncertainties, imagery observation challenging noises cause task. Given","Given various spurious imagery noises which cause observation uncertainties, this remains a very challenging task.",10382
"are approach our Results of competition in with comparison presented, benchmark. the","Results of our approach are presented, in comparison with the competition benchmark.",10383
"framework the are regarding discussed. Finally, future developments","Finally, future developments regarding the framework are discussed.",10384
"pertinence 'sociotype' theoretically work the and construct, The both present oriented. empirically discusses a of","The present work discusses the pertinence of a 'sociotype' construct, both theoretically and empirically oriented.",10385
based the algorithm states simulation SMS phenomenon. on matter The of of is the,The SMS algorithm is based on the simulation of the states of matter phenomenon.,10386
one state considering at algorithm is each matter exploration/exploitation ratio. The by different devised of,The algorithm is devised by considering each state of matter at one different exploration/exploitation ratio.,10387
"exhibit In each molecules state, capacities. movement different (individuals)","In each state, molecules (individuals) exhibit different movement capacities.",10388
of We optional graphs. study simple on stochastic games evolution,We study stochastic evolution of optional games on simple graphs.,10389
social Optional relaxed games dilemmas. to lead,Optional games lead to relaxed social dilemmas.,10390
we spatial and explore games. structure the interaction between optional Here,Here we explore the interaction between spatial structure and optional games.,10391
large for We limits various population size. and weak derive selection,We derive various limits for weak selection and large population size.,10392
we strong for selection. cases For some results derive analytic,For some cases we derive analytic results for strong selection.,10393
describes noise a the consistently state good model. channel variance with agreement discrete It in,It consistently describes the channel noise variance in a good agreement with discrete state model.,10394
also prospects constructing briefly for function. of future discusses brain integrated an It theory,It also discusses briefly future prospects for constructing an integrated theory of brain function.,10395
co-promotor dr. Denecker. and Janssens prof. Marc is Gerda my,Gerda Janssens and my co-promotor is prof. dr. Marc Denecker.,10396
robot. using can This system in computer by vision problem tackled be,This problem can be tackled by using computer vision system in robot.,10397
untangle can to used to This be then robot wire/s. guide information,This information can then be used to guide robot to untangle wire/s.,10398
"done remove preprocessing is image, an Given to noise.","Given an image, preprocessing is done to remove noise.",10399
usually views. extracted descriptors are these Multiple different to represent,Multiple descriptors are usually extracted to represent these different views.,10400
is contributing step the that every final We conclude rate. for recognition to crucial,We conclude that every step is crucial for contributing to the final recognition rate.,10401
concerned for $\alpha$-stable is volatility paper models. filtering particle with This stochastic,This paper is concerned with particle filtering for $\alpha$-stable stochastic volatility models.,10402
fruit industry economic cause losses agricultural in Diseases and worldwide. in production devastating in problem,Diseases in fruit cause devastating problem in economic losses and production in agricultural industry worldwide.,10403
play de in suggest important that strongly protein novo results These role lncRNAs evolution. an,These results strongly suggest that lncRNAs play an important role in de novo protein evolution.,10404
for plays Computer Inter-frame role and Compression video Vision. Coding significant,Inter-frame Coding plays significant role for video Compression and Computer Vision.,10405
systems (e.g. vision life many been applications have in incorporated Computer real,Computer vision systems have been incorporated in many real life applications (e.g.,10406
"medical surveillance robot systems, identity navigation and systems). verification imaging,","surveillance systems, medical imaging, robot navigation and identity verification systems).",10407
"two until AFSA, these the algorithm termination. standard constant In remain parameters","In standard AFSA, these two parameters remain constant until the algorithm termination.",10408
evaluated on based been the proposed well-known have The four approaches functions. benchmark,The proposed approaches have been evaluated based on the four well-known benchmark functions.,10409
the AFSA. of Experimental results show positive considerable impact on performance,Experimental results show considerable positive impact on the performance of AFSA.,10410
for novel mechanism We propose a concentrating urine.,We propose a novel mechanism for concentrating urine.,10411
hypothesis feasibility the of the proposed. to is An testing approach,An approach to testing the feasibility of the hypothesis is proposed.,10412
algorithm determine circles of parameters candidates. edge to combination uses a as points three The,The algorithm uses a combination of three edge points as parameters to determine circles candidates.,10413
actually candidates are determines reinforcement the at A if signal such circle present image.,A reinforcement signal determines if such circle candidates are actually present at the image.,10414
a fast is complicated multiple-circle facing The conditions. overall approach despite detector,The overall approach is a fast multiple-circle detector despite facing complicated conditions.,10415
"we $a$ article, necessarily $b$ consider In and independent. not this","In this article, we consider $a$ and $b$ not necessarily independent.",10416
result can spatial metastable between indicates states. that transitions the Our structure accelerate,Our result indicates that spatial structure can accelerate the transitions between metastable states.,10417
"informative individuals. often features differences group-wise However, of are assessing less for","However, group-wise differences are often less informative for assessing features of individuals.",10418
"has composition been studied. Second, clusters those of the clade","Second, the clade composition of those clusters has been studied.",10419
shrinkage between step The quadratic proposed and non-local a subproblem. alternates a algorithm,The proposed algorithm alternates between a non-local shrinkage step and a quadratic subproblem.,10420
shrinkage derive penalties relevant non-local that regularization. for are several We rules analytical in,We derive analytical shrinkage rules for several penalties that are relevant in non-local regularization.,10421
resulting non-local be alternating algorithms. current considerably is than faster The algorithm observed to,The resulting algorithm is observed to be considerably faster than current alternating non-local algorithms.,10422
(DGSM) this global The measures} sensitivity in been developed {\em recently derivative-based respect. have,The {\em derivative-based global sensitivity measures} (DGSM) have been developed recently in this respect.,10423
analytical requires of PC orthonormal which of expansions. derivatives enter This polynomials the derivation the,This requires the analytical derivation of derivatives of the orthonormal polynomials which enter PC expansions.,10424
efficiency on The sensitivity in of two illustrated well-known benchmark problems analysis. approach the is,The efficiency of the approach is illustrated on two well-known benchmark problems in sensitivity analysis.,10425
The reasoning. a which algorithm paper presents employs detection such fuzzy corner,The paper presents a corner detection algorithm which employs such fuzzy reasoning.,10426
events. a bulb olfactory thus odor is an sharp The to of sequence response,The olfactory bulb response to an odor is thus a sequence of sharp events.,10427
"is discrete a an odor Thus, the signature sequence. of","Thus, the signature of an odor is a discrete sequence.",10428
"efficient clusters, a number Hierarchical an Clustering algorithm determine proposed. of is (SHC) Semi-supervised To","To determine an efficient number of clusters, a Semi-supervised Hierarchical Clustering (SHC) algorithm is proposed.",10429
"To end, is this Direct Computation Vector proposed. a algorithm Fiedler","To this end, a Direct Fiedler Vector Computation algorithm is proposed.",10430
is consider based graph data not does structure the cut and on The labels.,The graph cut is based on the data structure and does not consider labels.,10431
used to Labels define criterion only cut. are graph the for stopping,Labels are used only to define the stopping criterion for graph cut.,10432
five algorithms on image-set state-of-the-art The standard datasets. proposed algorithm classification outperformed,The proposed algorithm outperformed state-of-the-art image-set classification algorithms on five standard datasets.,10433
learning in involved understood. not processes cortical are The well,The cortical processes involved in learning are not well understood.,10434
the adaptation. findings formation during assemblies cell of have Initial suggested,Initial findings have suggested the formation of cell assemblies during adaptation.,10435
"work hierarchical Bayesian is this fact, network. based on dynamic In a","In fact, this work is based on a dynamic hierarchical Bayesian network.",10436
using semantic and vision Computer are defined tasks evaluated categories. traditionally,Computer vision tasks are traditionally defined and evaluated using semantic categories.,10437
single classification steps in and algorithm. approach clustering a unifies the Our proposed,Our proposed approach unifies the clustering and classification steps in a single algorithm.,10438
qualitatively results. this and quantitatively we significant Following approach show both,Following this approach we show both qualitatively and quantitatively significant results.,10439
distribution. the visual tail long that a We show subclasses demonstrate,We show that the visual subclasses demonstrate a long tail distribution.,10440
"detection of object state methods show we Finally, that art the (e.g.","Finally, we show that state of the art object detection methods (e.g.",10441
image stages are different processing. of  being at  techniques widely used satellite Resampling,  Resampling techniques are being widely used at different stages of satellite image processing.,10442
rules. structure of if-then non-redundant minimal and explore We sets consisting graded of the,We explore the structure of non-redundant and minimal sets consisting of graded if-then rules.,10443
plays most applications. role in important an Change detection video-based,Change detection plays an important role in most video-based applications.,10444
judgments study problem of be can how We such aggregated. multiple the,We study the problem of how multiple such judgments can be aggregated.,10445
domain. adapting for the define properties classical problem argumentation social-choice-theoretic by various We the,We define the problem by adapting various classical social-choice-theoretic properties for the argumentation domain.,10446
speckle on presents wavelet paper The transform. real and time based activity computation algorithm de-noising,The paper presents real time speckle de-noising based on activity computation algorithm and wavelet transform.,10447
illuminated from arise an an Speckles laser in surface. light when is image reflected,Speckles arise in an image when laser light is reflected from an illuminated surface.,10448
FPGA can Programmable Gate in implemented The (Field be Array) technology. project,The project can be implemented in FPGA (Field Programmable Gate Array) technology.,10449
"is detecting reliable In for method images faces proposed. human paper, this in a color","In this paper, a reliable method for detecting human faces in color images is proposed.",10450
adjusting The is are given results circuit implemented of out. simulation some and real-time,The real-time adjusting circuit is implemented and some of simulation results are given out.,10451
"suggest several focus we for Therefore, sectors. experimental SCA which shifting the to identifies proteins","Therefore, we suggest shifting the experimental focus to proteins for which SCA identifies several sectors.",10452
high created energy need in to the computational capacity. have physics increase Advances,Advances in high energy physics have created the need to increase computational capacity.,10453
composed challenge. Project this HEPGAME was to address,Project HEPGAME was composed to address this challenge.,10454
"and We have these simplify using common investigated to ways subexpression expressions, schemes elimination. Horner","We have investigated ways to simplify these expressions, using Horner schemes and common subexpression elimination.",10455
"applies search Our AI. has in been a procedure successful approach MCTS, that","Our approach applies MCTS, a search procedure that has been successful in AI.",10456
use We schemes. to near-optimal find it Horner,We use it to find near-optimal Horner schemes.,10457
"Although this rise further finds approach MCTS gives two challenges. to solutions, better","Although MCTS finds better solutions, this approach gives rise to two further challenges.",10458
manually. has to be constant This tuned,This constant has to be tuned manually.,10459
revealed connectivity wider repertoire during approach of states This control than post-psilocybin a conditions.,This approach revealed a wider repertoire of connectivity states post-psilocybin than during control conditions.,10460
is detection a Face recognition. basic task expression for,Face detection is a basic task for expression recognition.,10461
detection recognition. ways There to several are face undergo &,There are several ways to undergo face detection & recognition.,10462
"or Operations, the Processing classifiers, virtual can We filters use various machines former. for Image","We can use Image Processing Operations, various classifiers, filters or virtual machines for the former.",10463
are Facial available Various Expression strategies for Detection. being,Various strategies are being available for Facial Expression Detection.,10464
these Local popularity have gained years. Binary Pattern based texture in algorithms (LBP),Local Binary Pattern (LBP) based texture algorithms have gained popularity in these years.,10465
expression LBP is feature-based approach. effective & is a recognition an to facial approach have,LBP is an effective approach to have facial expression recognition & is a feature-based approach.,10466
mixture Bayesian models and unsupervised learning  analysis. exploratory data for applied widely  are,  Bayesian mixture models are widely applied for unsupervised learning and exploratory data analysis.,10467
techniques explored been not Their have much. registration,Their registration techniques have not been explored much.,10468
in making DEM-to-map are pyramid registration. Coarse-to-fine Methods like or DEM-to-image or common,Methods like Coarse-to-fine or pyramid making are common in DEM-to-image or DEM-to-map registration.,10469
"being error these methods lack evaluation. in apart time But and from intensive, complexity matrix","But these methods apart from being time and complexity intensive, lack in error matrix evaluation.",10470
also been same the done. analysis Error for has,Error analysis for the same has also been done.,10471
"bind with can in a e.g. covalent way, Cysteine ZnS","Cysteine can bind with ZnS in a covalent way, e.g.",10472
the forming disulfide S with by solid. in bond the,by forming the disulfide bond with S in the solid.,10473
"water is mobile. of first and layer Furthermore, disordered the","Furthermore, the first layer of water is disordered and mobile.",10474
detection based optimization. circle a Differential on introduces This paper method Evolution (DE),This paper introduces a circle detection method based on Differential Evolution (DE) optimization.,10475
and shown remarkable has advantageous robustness. convergence also It properties,It has also shown advantageous convergence properties and remarkable robustness.,10476
considered combinational to is optimization The a detection similar process problem.,The detection process is considered similar to a combinational optimization problem.,10477
if image. in objective are actually circle the The determines some function present candidates,The objective function determines if some circle candidates are actually present in the image.,10478
"candidate Specifically, loci automatically identified adaptation by associations. are assessing genome-environment to","Specifically, candidate loci to adaptation are identified by automatically assessing genome-environment associations.",10479
on the covalent cysteine atom the bond is sulfur The potential. Morse by with modeled,The covalent bond with the sulfur atom on cysteine is modeled by the Morse potential.,10480
surface properties The the on depend water near sensitively force the field. of,The properties of water near the surface sensitively depend on the force field.,10481
provide for information-aggregation a framework featuring reasoning modelling Multi-context powerful systems heterogeneous components. systems,Multi-context systems provide a powerful framework for modelling information-aggregation systems featuring heterogeneous reasoning components.,10482
"non-negligible execution cost. Their can, incur however,","Their execution can, however, incur non-negligible cost.",10483
"systems. of focus Here, on cost-complexity such we","Here, we focus on cost-complexity of such systems.",10484
We cost-complexity MCSs. the of consistency formulate reasoning in and for notion problems,We formulate the notion of cost-complexity for consistency and reasoning problems in MCSs.,10485
high-throughput depend RNA-seq protocols of sequencing Current on cDNA. short-read,Current RNA-seq protocols depend on high-throughput short-read sequencing of cDNA.,10486
"there segmenting is combined a for approach this, the In image.","In this, there is a combined approach for segmenting the image.",10487
time. In image produce it segmented way computation this with better less,In this way it produce better segmented image with less computation time.,10488
point. histogram class represents and threshold therefore the function a a pixel Each approximating Gaussian,Each Gaussian function approximating the histogram represents a pixel class and therefore a threshold point.,10489
raised community currently. is problem the by complex management Crisis a scientific,Crisis management is a complex problem raised by the scientific community currently.,10490
approach challenge. this paper The addresses this in we propose,The approach we propose in this paper addresses this challenge.,10491
here. and approach The are provided experimentations general,The general approach and experimentations are provided here.,10492
on give few a study. his We then notes,We then give a few notes on his study.,10493
"resolved. related body-check orientation issues such efficiently global Moreover, are to as","Moreover, global issues related to body-check such as orientation are efficiently resolved.",10494
Many illustrate our from a pilot the efficiency of and framework. implementation effectiveness examples,Many examples from a pilot implementation illustrate the efficiency and effectiveness of our framework.,10495
coordinate to has spatial been proposed coupling dynamics temporal across (CFC) neural Cross-frequency and scales.,Cross-frequency coupling (CFC) has been proposed to coordinate neural dynamics across spatial and temporal scales.,10496
considerable Producing time. needed simulations large of quantity requires the computing,Producing the large quantity of simulations needed requires considerable computing time.,10497
more where makes ABC applications to expensive. This simulation is scalable,This makes ABC more scalable to applications where simulation is expensive.,10498
all-atom combined involves technique. method simulations with The sampling dynamics molecular umbrella the,The method involves all-atom molecular dynamics simulations combined with the umbrella sampling technique.,10499
acids. is the for Another amino most top) at (O non-binding,Another (O at the top) is non-binding for most amino acids.,10500
the protein. events acids Various mobilities adsorption within in different patterns to lead amino of,Various adsorption events lead to different patterns in mobilities of amino acids within the protein.,10501
based paper method circle describes Electromagnetism-Like detection Optimization This a on (EMO).,This paper describes a circle detection method based on Electromagnetism-Like Optimization (EMO).,10502
principles. for complex in inspired problems heuristic EMO method electromagnetism a solving is optimization,EMO is a heuristic method for solving complex optimization problems inspired in electromagnetism principles.,10503
based the in This repulsion attraction algorithm searches and prototype among candidates. solution a,This algorithm searches a solution based in the attraction and repulsion among prototype candidates.,10504
the if such actually are present circle An image. function in objective candidates determines,An objective function determines if such circle candidates are actually present in the image.,10505
"strategies the efficient minimize we to continuation local convergence to Additionally, minima. of risk employ","Additionally, we employ efficient continuation strategies to minimize the risk of convergence to local minima.",10506
then minimax to are These guide values used simulations. future,These minimax values are then used to guide future simulations.,10507
problems. reduction related result dimensionality is to strictly clustering The and,The result is strictly related to dimensionality reduction and clustering problems.,10508
integral. an estimator of This multiple randomized unbiased in lattice results rule the,This randomized lattice rule results in an unbiased estimator of the multiple integral.,10509
modeling image. representation multiscale local approach an involves in The patterns of a,The approach involves modeling local patterns in a multiscale representation of an image.,10510
of properties non-local image a image. coarsened Local reflect properties original of the,Local properties of a coarsened image reflect non-local properties of the original image.,10511
relatively of parameters. framework leads a to that expressive small This depend priors on number,This framework leads to expressive priors that depend on a relatively small number of parameters.,10512
example evaluate the approach applications. with We two,We evaluate the approach with two example applications.,10513
of data. and account numerical large a vectors portion Numbers for,Numbers and numerical vectors account for a large portion of data.,10514
"increased dramatically. string amount generated has the of However, data recently","However, recently the amount of string data generated has increased dramatically.",10515
"problem string is Consequently, fields. in classifying common many data a","Consequently, classifying string data is a common problem in many fields.",10516
tracking we Replication (R. with programming. F.)for dynamic this used In a Function paper improvement,In this paper we used a Replication Function (R. F.)for improvement tracking with dynamic programming.,10517
F. R. in more regions. are images and The resulting striking skin visible of,The resulting images of R. F. are more striking and visible in skin regions.,10518
hand Programming face. improves (D. P.) Dynamic The R. in F. overlapping and,The R. F. improves Dynamic Programming (D. P.) in overlapping hand and face.,10519
this literature Here as we from optimization programs matrix manifolds. disparate survey methods over,Here we survey methods from this disparate literature as optimization programs over matrix manifolds.,10520
statistics patch-based image from operation that local representations involve descriptors. computed State-of-the-art a aggregates pooling,State-of-the-art patch-based image representations involve a pooling operation that aggregates statistics computed from local descriptors.,10521
include Standard max-pooling. pooling sum- and operations,Standard pooling operations include sum- and max-pooling.,10522
widely Spirtes The condition et al. of faithfulness used,The widely used faithfulness condition of Spirtes et al.,10523
the on natural This challenging letter solving focuses detecting image of boundaries. problem,This letter focuses on solving the challenging problem of detecting natural image boundaries.,10524
usually border A between to different meanings. regions with the boundary two refers semantic,A boundary usually refers to the border between two regions with different semantic meanings.,10525
characterize available second usual mostly phenomenon order are on the based The to tools correlations.,The usual tools available to characterize the phenomenon are mostly based on second order correlations.,10526
This the (BoW) for search. reranking based Bag-of-Words method improved introduces image an paper,This paper introduces an improved reranking method for the Bag-of-Words (BoW) based image search.,10527
"advantage enhanced. nature the Taking the is of further complementary reranking various of performance features,","Taking advantage of the complementary nature of various features, the reranking performance is further enhanced.",10528
"information. method Particularly, the and combining the we BoW exploit color reranking","Particularly, we exploit the reranking method combining the BoW and color information.",10529
extracted lower to are higher layers layers transmitted on codes recursively. Sparse,Sparse codes extracted on lower layers are transmitted to higher layers recursively.,10530
"different With from fused. scales this mechanism, cues are","With this mechanism, cues from different scales are fused.",10531
in integrated The existing package phylogenetic inference is BEAST. procedure the,The procedure is integrated in the existing phylogenetic inference package BEAST.,10532
"of are a keypoint employed. cues local Conventionally,","Conventionally, local cues of a keypoint are employed.",10533
is an inconsistency related inconsistencies. as issue viewed handling Measuring important to,Measuring inconsistency is viewed as an important issue related to handling inconsistencies.,10534
of are supposed properties. set Good to satisfy a rational measures,Good measures are supposed to satisfy a set of rational properties.,10535
"However, sound sometimes is problematic. properties defining","However, defining sound properties is sometimes problematic.",10536
"propose end, we this which more existing independent than To intuitive proposals. an decomposition is","To this end, we propose an independent decomposition which is more intuitive than existing proposals.",10537
inconsistency Such MUSdecomposition satisfies an number a of measure properties. that in particular desired results,Such particular MUSdecomposition results in an inconsistency measure that satisfies a number of desired properties.,10538
Proximity based for image an method Distribution effective representation. is Kernel bag-of-featues,Proximity Distribution Kernel is an effective method for bag-of-featues based image representation.,10539
Visual is contribution model function proposed proximity distributions. ambiguous to word,Visual word contribution function is proposed to model ambiguous proximity distributions.,10540
developed distributions proximity by functions. ambiguous Three contribution three is ambiguous,Three ambiguous proximity distributions is developed by three ambiguous contribution functions.,10541
and experiments image medical on both retrieval are classification data The conducted sets. of,The experiments are conducted on both classification and retrieval of medical image data sets.,10542
"analysis presented. (medical) this   algorithm In semi-automatic is segmentation image for a contribution,","  In this contribution, a semi-automatic segmentation algorithm for (medical) image analysis is presented.",10543
"object appearances between the background, or to the object. homogeneous and noise inside the due","due to homogeneous appearances between the object and the background, or noise inside the object.",10544
cases needs additional For user these the support. difficult still algorithm,For these difficult cases the algorithm still needs additional user support.,10545
resistance elusive. arises of The after antibiotic methods has reliable for restoring proven development susceptibility,The development of reliable methods for restoring susceptibility after antibiotic resistance arises has proven elusive.,10546
developments discusses recent this This in mini-review area.,This mini-review discusses recent developments in this area.,10547
"recognition intensive challenging After a face research, heterogenous problem. still is","After intensive research, heterogenous face recognition is still a challenging problem.",10548
to the main owing relationship difficulties face spaces. The complex between heterogenous are image,The main difficulties are owing to the complex relationship between heterogenous face image spaces.,10549
"RBMs processed by local are shared connected the and PCA. Finally, of together representations","Finally, the shared representations of local RBMs are connected together and processed by PCA.",10550
problems databases and proposed Two method. and to three evaluate are the selected (Sketch-Photo NIR-VIS),Two problems (Sketch-Photo and NIR-VIS) and three databases are selected to evaluate the proposed method.,10551
"Sketch-Photo CUFS database. For problem, perfect results on obtain the we","For Sketch-Photo problem, we obtain perfect results on the CUFS database.",10552
signal driven the last decade. popularity tremendous processing has in Sparsity gained,Sparsity driven signal processing has gained tremendous popularity in the last decade.,10553
of evidence open still Decision application making the is Dempster-Shafer issue an theory. in,Decision making is still an open issue in the application of Dempster-Shafer evidence theory.,10554
presented A have works of it. been lot for,A lot of works have been presented for it.,10555
images theorems to automatically. from approach generate an propose geometric diagrams electronic of We,We propose an approach to generate geometric theorems from electronic images of diagrams automatically.,10556
algorithms. essential become classifiers have for Pixel superpixel EM and tools segmentation,Pixel and superpixel classifiers have become essential tools for EM segmentation algorithms.,10557
of segmentation We provide these on multiple results classifiers. strength show datasets to the,We provide segmentation results on multiple datasets to show the strength of these classifiers.,10558
We estimating cluster two the procedures tree for $f$. given present samples from,We present two procedures for estimating the cluster tree given samples from $f$.,10559
is single first of hierarchical linkage clustering. algorithm a robust variant the The for,The first is a robust variant of the single linkage algorithm for hierarchical clustering.,10560
$k$-nearest the is the based of graph The samples. on neighbor second,The second is based on the $k$-nearest neighbor graph of the samples.,10561
"population data are be could whose sample or indeterminate. there Also, individuals","Also, there are population or sample individuals whose data could be indeterminate.",10562
proteins lipid waves on bilayers. coli Escherichia traveling The can into supported from self-organize Min,The Min proteins from Escherichia coli can self-organize into traveling waves on supported lipid bilayers.,10563
We introduced reproducing observed patterns. two-dimensional model the effective an,We introduced an effective two-dimensional model reproducing the observed patterns.,10564
"correlated (multiple effective. splits decision be highly oblique feature) trees data, Given with can","Given highly correlated data, decision trees with oblique (multiple feature) splits can be effective.",10565
"oblique expense. comes of however, Use considerable splits, at computational","Use of oblique splits, however, comes at considerable computational expense.",10566
framework novel proposes This agglomerative segmentation. for EM a paper,This paper proposes a novel agglomerative framework for EM segmentation.,10567
"minimization of initialization. a Next, we with convergence good alternating global prove","Next, we prove global convergence of alternating minimization with a good initialization.",10568
the the dimensionality dependence indeed is tight. sample suggest $n$ on Simulations of that size,Simulations suggest that the dependence of the sample size on dimensionality $n$ is indeed tight.,10569
subject in independent random often changes to the dynamics are environment. Population,Population dynamics are often subject to random independent changes in the environment.,10570
by space. atom Markov time a continuous in chain a finite modeled This is,This is modeled by a continuous time Markov chain in a finite atom space.,10571
"when first states. To chain has consider intuition, case create we two the the Markov","To create intuition, we first consider the case when the Markov chain has two states.",10572
constraints joint between structural exploit as The architecture relationships can such domain body locations. geometric,The architecture can exploit structural domain constraints such as geometric relationships between body joint locations.,10573
to estimation action step is pose a recognition. key Human,Human pose estimation is a key step to action recognition.,10574
We length anthropomorphically constraints to enforce skeletons. limb eliminate (ii) implausible,(ii) We enforce limb length constraints to eliminate anthropomorphically implausible skeletons.,10575
(ADM) method We alternating the to problem efficiently. solve direction optimization the use,We use the alternating direction method (ADM) to solve the optimization problem efficiently.,10576
on the state-of-the-arts approach benchmark Our outperforms datasets. three,Our approach outperforms the state-of-the-arts on three benchmark datasets.,10577
than error measure also scale. help We rather a scale-invariant relations apply depth to,We also apply a scale-invariant error to help measure depth relations rather than scale.,10578
novel clustering compact a normalization algorithm for We a propose pose learning space. graph-based,We propose a novel graph-based clustering algorithm for learning a compact pose normalization space.,10579
discriminatively trained of Convolutional recognition for Networks investigate video. action deep architectures (ConvNets) We in,We investigate architectures of discriminatively trained deep Convolutional Networks (ConvNets) for action recognition in video.,10580
performing hand-crafted to within features learning also a framework. generalise aim We best the data-driven,We also aim to generalise the best performing hand-crafted features within a data-driven learning framework.,10581
"two-stream spatial a propose First, temporal incorporates architecture we ConvNet which and networks.","First, we propose a two-stream ConvNet architecture which incorporates spatial and temporal networks.",10582
"we problem the this consider paper, In when cost. this are network fuzzy links traffic","In this paper, we consider this problem when the traffic network links are fuzzy cost.",10583
"experiments performance of test this used to some Eventually, the are method.","Eventually, some experiments are used to test the performance of this method.",10584
existing results compared that The with other demonstrate is approach competitive when algorithms. our,The results demonstrate that our approach is competitive when compared with other existing algorithms.,10585
shaping Understanding and the forces science is communities ecological to conservation. crucially basic important,Understanding the forces shaping ecological communities is crucially important to basic science and conservation.,10586
"progress Finally, benchmark datasets we on report existing of segmenting three salient . significant objects","Finally, we report significant benchmark progress on three existing datasets of segmenting salient objects .",10587
"involving a sparse remains problem, Solving convolutional problems however. computational constraints difficult and","Solving problems involving sparse and convolutional constraints remains a difficult computational problem, however.",10588
this paper in In of framework. an consistent overview sparse coding convolutional present a we,In this paper we present an overview of convolutional sparse coding in a consistent framework.,10589
undirected models. foundational results estimation This theoretical for parameter probabilistic on graphical presents distributed paper,This paper presents foundational theoretical results on distributed parameter estimation for undirected probabilistic graphical models.,10590
we for present recognition of framework this scene In natural text. the work a,In this work we present a framework for the recognition of natural scene text.,10591
"segments enforces same consistency appearance part. within particular, between the In model our","In particular, our model enforces appearance consistency between segments within the same part.",10592
a task. networks vital in  telecommunication Failure  detection is,  Failure detection in telecommunication networks is a vital task.,10593
is approaches them no Among has attention more required. since unsupervised attracted data label,Among them unsupervised approaches has attracted more attention since no label data is required.,10594
"not to devices information type are provide Often, able network about of the failure.","Often, network devices are not able to provide information about the type of failure.",10595
"their important properties data nature. of temporal the However, of network one is sequential","However, one of the important properties of network data is their temporal sequential nature.",10596
the original paper we tempering to In this present extensions adaptive algorithm. parallel,In this paper we present extensions to the original adaptive parallel tempering algorithm.,10597
"introduce second the one, of In we the number temperatures. online of adjustment","In the second one, we introduce online adjustment of the number of temperatures.",10598
the the method. demonstrate proposed experiments effectiveness Numerical of,Numerical experiments demonstrate the effectiveness of the proposed method.,10599
algorithms Fuzzy suitable clustering performing based soft for tasks. are logic always,Fuzzy logic based algorithms are always suitable for performing soft clustering tasks.,10600
(FCM) C Means based logic is fuzzy Fuzzy a algorithm very algorithm. popular,Fuzzy C Means (FCM) algorithm is a very popular fuzzy logic based algorithm.,10601
components critical two time many systems. are and real of Space world,Space and time are two critical components of many real world systems.,10602
"this has reason, of a data in interest. anomalies of For spatiotemporal great been analysis","For this reason, analysis of anomalies in spatiotemporal data has been a great of interest.",10603
"the arguments. from ranking values arguments of strength is on In derived the general, the","In general, the ranking on arguments is derived from the strength values of the arguments.",10604
"fixed we issues this point paper, technique. will In via cope these with","In this paper, we will cope with these issues via fixed point technique.",10605
low effectively Crowdsourcing labels paradigm a for at popular cost. is collecting,Crowdsourcing is a popular paradigm for effectively collecting labels at low cost.,10606
"crowd In we efficient paper, a two-stage algorithm multi-class problems. this propose for labeling","In this paper, we propose a two-stage efficient algorithm for multi-class crowd labeling problems.",10607
to estimate spectral an the first parameters. stage initial obtain The method uses of,The first stage uses the spectral method to obtain an initial estimate of parameters.,10608
achieves the convergence up We logarithmic to show that rate factor. optimal algorithm our a,We show that our algorithm achieves the optimal convergence rate up to a logarithmic factor.,10609
synthetic conduct experiments on extensive datasets. We real and,We conduct extensive experiments on synthetic and real datasets.,10610
"freely together available, examples. source with code some is The","The source code is freely available, together with some examples.",10611
"covariates. not or observed typically the in corruptions to these However, outliers robust algorithms are","However, these algorithms are typically not robust to outliers or corruptions in the observed covariates.",10612
"has quantifying such effects to relationship been polygenic their risk disease challenging. and However,","However, quantifying such polygenic effects and their relationship to disease risk has been challenging.",10613
a For this number purpose of large literature. in methods exist,For this purpose a large number of methods exist in literature.,10614
a proposed. shapes has fusion In this recognition based method been paper,In this paper a fusion based shapes recognition method has been proposed.,10615
of a It applications from number to security has entertainment. ranging,It has a number of applications ranging from security to entertainment.,10616
"time communication and need machines. applications applications These real fast, generally are accurate with","These applications generally are real time applications and need fast, accurate communication with machines.",10617
based before has done with calculation gloves/devices. sensor been Angle,Angle calculation has been done before with sensor based gloves/devices.,10618
level-set geometrical using detected were using analysis. and method angles are Fingertips calculated,Fingertips are detected using level-set method and angles were calculated using geometrical analysis.,10619
to This system task. technique the for perform no training requires,This technique requires no training for system to perform the task.,10620
low-rank estimation The a autoregressive the key is matrices. of find to,The key is to find a low-rank estimation of the autoregressive matrices.,10621
processes further exact are chosen details often consideration. of these much without The,The exact details of these processes are often chosen without much further consideration.,10622
choices this micro-level view robust the really are. these on challenge We and investigate how,We challenge this view and investigate how robust these choices on the micro-level really are.,10623
can that assumes scales images be at (zoom levels). only It taken different,It only assumes that images can be taken at different scales (zoom levels).,10624
are this causes gene possible Many incongruence tree of known.,Many possible causes of this gene tree incongruence are known.,10625
"is process. commonly The studied best the lineage sorting, modeled by which coalescent incomplete is","The best studied is incomplete lineage sorting, which is commonly modeled by the coalescent process.",10626
"fundamental metric for learning. critical are problems Two feature person representation and re-identification,","Two fundamental problems are critical for person re-identification, feature representation and metric learning.",10627
"well computation regularization. XQDA, also method as as for We present practical a its","We also present a practical computation method for XQDA, as well as its regularization.",10628
"hierarchical set Granular It in naturally problem, realized classification that any is arises.","It is realized that in any hierarchical classification problem, Granular set naturally arises.",10629
This of set. the concept generalizes rough,This generalizes the concept of rough set.,10630
theory of finds a It hoped sets that is graded granular/ applications. several detailed,It is hoped that a detailed theory of granular/ graded sets finds several applications.,10631
Such advantages has as many argued in modelling literature.,Such modelling has many advantages as argued in literature.,10632
"the computational increasing population, overall Approximating repeated cost. these of requires the samplings statistics drastically","Approximating these statistics requires repeated samplings of the population, drastically increasing the overall computational cost.",10633
estimation video major developing the problems one in coding Motion of is applications.,Motion estimation is one of the major problems in developing video coding applications.,10634
belonging the solutions. approach vectors to The window uses potential search as motion,The approach uses motion vectors belonging to the search window as potential solutions.,10635
matching of candidate. motion vector the each function evaluates A quality fitness,A fitness function evaluates the matching quality of each motion vector candidate.,10636
Evolutionary counterintuitive interesting dynamics and many findings. to lead can on graphs,Evolutionary dynamics on graphs can lead to many interesting and counterintuitive findings.,10637
"of Remarkably, the all single mutant fixation on is a the regular probability networks. same","Remarkably, the fixation probability of a single mutant is the same on all regular networks.",10638
increase networks fixation decrease But or probability. can the non-regular,But non-regular networks can increase or decrease the fixation probability.,10639
and introduced methods in the on Vision techniques literature. other based Several Computer are,Several methods based on Computer Vision and other techniques are introduced in the literature.,10640
Metropolis-Hastings define sampler We random-walk for updating diffusion a type bridges.,We define a random-walk type Metropolis-Hastings sampler for updating diffusion bridges.,10641
diffusion proposals sampling guided for Our methods are illustrated bridges. using,Our methods are illustrated using guided proposals for sampling diffusion bridges.,10642
examples the performance of Numerical our methods. demonstrate,Numerical examples demonstrate the performance of our methods.,10643
valuable this in are situation. particularly Diagnostics,Diagnostics are particularly valuable in this situation.,10644
"tool Gaussian large-scale updates, plot'. a visualisation the and We diagnostic `medal present for","We present a diagnostic and visualisation tool for large-scale Gaussian updates, the `medal plot'.",10645
"new, or can abnormal, novel, unusual be They information. noisy","They can be novel, new, abnormal, unusual or noisy information.",10646
the majority data. of are interesting more Outliers the sometimes than,Outliers are sometimes more interesting than the majority of the data.,10647
texture. is classification of mammographic characteristics major One mass for the,One of the major mammographic characteristics for mass classification is texture.,10648
the The Haralick characterizing in features Run features. textural length are masses and statistical used,The statistical textural features used in characterizing the masses are Haralick and Run length features.,10649
bias comparable laboratories the same measurement amongst were for measurement also process. Ratio and variability,Ratio measurement variability and bias were also comparable amongst laboratories for the same measurement process.,10650
biases were for using different enrichment observed processes measurement Different mRNA protocols.,Different biases were observed for measurement processes using different mRNA enrichment protocols.,10651
in significance the discuss our of We results real life context the of networks.,We discuss the significance of our results in the context of real life networks.,10652
complete dengue This epidemic would the epidemic. most to the proposed be describe model,This would be the most complete epidemic model proposed to describe the dengue epidemic.,10653
"fuzzy paradigm. This guide, formalized was in write logic using language, natural","This guide, write in natural language, was formalized using fuzzy logic paradigm.",10654
"are system project. including validation, All expert TERAPERS part these researches, of","All these researches, including expert system validation, are part of TERAPERS project.",10655
"matrix-vector sparse easy and is to products DFacTo, only parallelize. algorithm, two requires Our","Our algorithm, DFacTo, only requires two sparse matrix-vector products and is easy to parallelize.",10656
"new a regardless network called SPP-net, of can The representation fixed-length generate size/scale. image structure,","The new network structure, called SPP-net, can generate a fixed-length representation regardless of image size/scale.",10657
Pyramid to is pooling robust deformations. also object,Pyramid pooling is also robust to object deformations.,10658
"CNN-based With improve advantages, general image should classification these all in SPP-net methods.","With these advantages, SPP-net should in general improve all CNN-based image classification methods.",10659
also SPP-net detection. significant in object is power of The,The power of SPP-net is also significant in object detection.,10660
avoids This computing convolutional method the features. repeatedly,This method avoids repeatedly computing the convolutional features.,10661
introduces improvement the This this made also for manuscript competition.,This manuscript also introduces the improvement made for this competition.,10662
"experimental linear are the model. also of Several reported, which the suitability justify results","Several experimental results are also reported, which justify the suitability of the linear model.",10663
agreement. We with single molecule find strong experiments compare our available predictions and manipulation,We compare our predictions with available single molecule manipulation experiments and find strong agreement.,10664
and elucidate constrained In particular overstretching. we difference unconstrained the between angularly,In particular we elucidate the difference between angularly constrained and unconstrained overstretching.,10665
key vision. a actions human is Understanding computer problem in,Understanding human actions is a key problem in computer vision.,10666
"is of only recognizing what step actions first However, is a person understanding doing. the","However, recognizing actions is only the first step of understanding what a person is doing.",10667
make contributions main to three this. We achieve,We make three main contributions to achieve this.,10668
a population have a The environment evolves selection. crucial can in on impact which,The environment in which a population evolves can have a crucial impact on selection.,10669
in environment. finite dynamics changing size We a fixed evolutionary populations of in study,We study evolutionary dynamics in finite populations of fixed size in a changing environment.,10670
death events. population driven dynamics by The and birth are,The population dynamics are driven by birth and death events.,10671
the varies structure which payoff our evolutionary games We in theory to for time. apply,We apply our theory to evolutionary games for which the payoff structure varies in time.,10672
surprising interpretation provide We this an intuitive effect. of,We provide an intuitive interpretation of this surprising effect.,10673
distributions dynamics. We in also investigate stationary present when are the mutations,We also investigate stationary distributions when mutations are present in the dynamics.,10674
"measure. regime, approximations stationary In the we find this of two","In this regime, we find two approximations of the stationary measure.",10675
"well for works for One switching, rapid the slowly other fluctuating environments.","One works well for rapid switching, the other for slowly fluctuating environments.",10676
Neural Convolutional (CNN) tasks. demonstrated Network in performance single-label classification image promising has,Convolutional Neural Network (CNN) has demonstrated promising performance in single-label image classification tasks.,10677
the affecting representations. investigates such transferability factors introduces This paper and several of,This paper introduces and investigates several factors affecting the transferability of such representations.,10678
"problem In study a we detector, the i.e. this `test-driving' of paper,","In this paper, we study the problem of `test-driving' a detector, i.e.",10679
We on of here their analyses plants. focus,We focus here on their analyses of plants.,10680
the sampled value are components using algorithm. These MSR down enhanced,These down sampled value components are enhanced using the MSR algorithm.,10681
wavelet the and computes in This scheme image enhanced of both original energy domain.,This scheme computes the energy of both original and enhanced image in wavelet domain.,10682
method state-of-the-art for action Our and gives prediction. results keypoint,Our method gives state-of-the-art results for keypoint and action prediction.,10683
citation. however excessive this Metrics through are such potentially as self impacted,Metrics such as this are however potentially impacted through excessive self citation.,10684
insights how to differ. rigorous on they leading We both comparison a provide of perspectives,We provide a rigorous comparison of both perspectives leading to insights on how they differ.,10685
"local such of well-known of structure, forms lines Patches T-junctions. or as edges straight exhibit","Patches of edges exhibit well-known forms of local structure, such as straight lines or T-junctions.",10686
estimate We a criterion of derive information number selection clusters. to model the,We derive a model selection information criterion to estimate the number of clusters.,10687
and have To developed for been methods in introduced various images. date assessing textures,To date various methods have been developed and introduced for assessing textures in images.,10688
"studies, success- use. has ful weaknesses in TEM Despite number several of a serious","Despite several success- ful studies, TEM has a number of serious weaknesses in use.",10689
image. therefore masks cannot they is; adapted The drawback the predefined be are major to,The major drawback is; the masks are predefined therefore they cannot be adapted to image.,10690
"assess To TEM. aTEM, with it the compared of applicability is","To assess the applicability of aTEM, it is compared with TEM.",10691
for this texture aTEM results a analysis. method of indicate that study is The successful,The results of this study indicate that aTEM is a successful method for texture analysis.,10692
with extensive range testing wide and performed results. a We models have the of report,We have performed extensive testing with a wide range of models and report the results.,10693
"three the truncation. measurements to flow-cytometry first We the of modeling problems, consider being subject","We consider three problems, the first being the modeling of flow-cytometry measurements subject to truncation.",10694
of where problems latter inference Markov Carlo is Monte are chain instances doubly-intractable. two The,The latter two are instances of problems where Markov chain Monte Carlo inference is doubly-intractable.,10695
Our algorithms for such demonstrate problems. state-of-the-art experiments superior performance over sampling,Our experiments demonstrate superior performance over state-of-the-art sampling algorithms for such problems.,10696
"views face factors, Various coupled in (poses), illuminations, identities, images. are as such and","Various factors, such as identities, views (poses), and illuminations, are coupled in face images.",10697
challenge identity recognition. major a face in is and representations Disentangling the view,Disentangling the identity and view representations is a major challenge in face recognition.,10698
is of from different human the behavior brain. This,This is different from the behavior of human brain.,10699
superior achieve The features MultiPIE dataset. the identity of performance on MVP,The identity features of MVP achieve superior performance on the MultiPIE dataset.,10700
well-performing exist clustering network to algorithms partitions. infer Several topological,Several well-performing clustering algorithms exist to infer topological network partitions.,10701
techniques factorization Such non-negative can with for clustering of be graphs. together useful obtaining discretization,Such discretization together with non-negative factorization techniques can be useful for obtaining clustering of graphs.,10702
shape (Hermite a overall are global The modifications and RBF). by surface defined variational implicit,The overall shape and global modifications are defined by a variational implicit surface (Hermite RBF).,10703
properties the We of the NIR also physical with discuss the response. to respect results,We also discuss the results with respect to the physical properties of the NIR response.,10704
abbreviated Optical is Character HOCR Handwritten as Recognition.,HOCR is abbreviated as Handwritten Optical Character Recognition.,10705
"rotation shape, property. Zernike to due Orthogonality moments describes identify invariant its","Zernike moments describes shape, identify rotation invariant due to its Orthogonality property.",10706
representation characters. of India script is an had ancient of cursive and complex MODI,MODI is an ancient script of India had cursive and complex representation of characters.,10707
image processing. is one the of Segmentation tasks important in most,Segmentation is one of the most important tasks in image processing.,10708
segmentation the depends applied The the on select the threshold. to of quality method,The quality of the segmentation depends on the method applied to select the threshold.,10709
"of been conditions, Under extended. such approaches optimization evolutionary use the has","Under such conditions, the use of optimization evolutionary approaches has been extended.",10710
"search overhead. Different a low to algorithms, EMO capabilities exhibits whereas maintains interesting computational other","Different to other algorithms, EMO exhibits interesting search capabilities whereas maintains a low computational overhead.",10711
"is multilevel EMO introduced. In on this (MT) paper, based algorithm a thresholding the","In this paper, a multilevel thresholding (MT) algorithm based on the EMO is introduced.",10712
search image random feasible the from The space a samples takes histogram. algorithm inside,The algorithm takes random samples from a feasible search space inside the image histogram.,10713
for Experimental show segmentation. the results of performance implementation evidence image EMO of digital,Experimental results show performance evidence of the implementation of EMO for digital image segmentation.,10714
Automatic by facial received significant recognition representations. specialised improvement developing has performance face image,Automatic face recognition has received significant performance improvement by developing specialised facial image representations.,10715
"to object generic On face rarely the been the recognition. has applied hand, recognition other","On the other hand, generic object recognition has rarely been applied to the face recognition.",10716
in order changes. to to Populations can external adapt evolve,Populations can evolve in order to adapt to external changes.,10717
to successful and of evolve difficult. infectious cancer adapt and diseases treatment The capacity makes,The capacity to evolve and adapt makes successful treatment of infectious diseases and cancer difficult.,10718
"for quickly become challenge a Indeed, global resistance health. therapy has key","Indeed, therapy resistance has quickly become a key challenge for global health.",10719
"been models sparse many have in signal vision successfully Moreover, applications. used","Moreover, sparse signal models have successfully been used in many vision applications.",10720
contributions. the major paper authors of work extends of the two This by some,This paper extends the work of some of the authors by two major contributions.,10721
intuition the information. carries modalities accounts in most the that contrast This for image,This accounts for the intuition that contrast in image modalities carries the most information.,10722
"structured pages newspapers and about databases. to everywhere, from web Statements entities occur","Statements about entities occur everywhere, from newspapers and web pages to structured databases.",10723
"a We for description"", by model ""reference references. resolving formal present","We present ""reference by description"", a formal model for resolving references.",10724
novel genomes. homology-based for results applying utility methods vertebrate combined These annotating the of illustrate,These combined results illustrate the utility of applying homology-based methods for annotating novel vertebrate genomes.,10725
statistical methodology research employs of The Bayesian use to perform calibration. the presented,The research presented employs the use of Bayesian methodology to perform statistical calibration.,10726
this models based Two paper. are separate DLM presented in,Two separate DLM based models are presented in this paper.,10727
radiometry These to methods set. dynamic applied data a microwave are,These dynamic methods are applied to a microwave radiometry data set.,10728
"this new clustering framework. algorithm on we the a propose fuzzy based In mode-seeking paper,","In this paper, we propose a new fuzzy clustering algorithm based on the mode-seeking framework.",10729
"can be Using order. any in approach, generated this samples","Using this approach, samples can be generated in any order.",10730
"algorithms so high. exhibit reuse Fortunately, rendering most cache is good coherence,","Fortunately, most rendering algorithms exhibit good coherence, so cache reuse is high.",10731
data a density tool Estimation versatile of in analysis. derivatives is statistical,Estimation of density derivatives is a versatile tool in statistical data analysis.,10732
A first approach is its compute estimate the to and naive derivative. then density,A naive approach is to first estimate the density and then compute its derivative.,10733
"robustly major functional distinguish challenge to noise. transcriptional a However, outputs is from","However, a major challenge is to robustly distinguish functional outputs from transcriptional noise.",10734
"close a comparative noncoding Using thousand RNAs. identify highly-expressed tools, to we candidate","Using comparative tools, we identify close to a thousand highly-expressed candidate noncoding RNAs.",10735
"The are spatial together with priors, learned and CRF. incorporated contexts, geometric in","The learned priors, together with spatial and geometric contexts, are incorporated in CRF.",10736
dataset Our complicated road contains validated is that diverse scenarios. challenging KITTI the approach on,Our approach is validated on the challenging KITTI dataset that contains diverse complicated road scenarios.,10737
and evaluations qualitative Both demonstrate its effectiveness. quantitative,Both quantitative and qualitative evaluations demonstrate its effectiveness.,10738
"one to On the problems. one declaratively allow Problems model Constraint Satisfaction hand,","On the one hand, Constraint Satisfaction Problems allow one to declaratively model problems.",10739
"propositional huge SAT (SAT) can hand, solvers the problem handle other instances. satisfiability On","On the other hand, propositional satisfiability problem (SAT) solvers can handle huge SAT instances.",10740
"and than Our improved simpler, is declarative, direct hand modeling. technique less error-prone more and","Our technique is simpler, more declarative, and less error-prone than direct and improved hand modeling.",10741
or not. same the to belong class),belong to the same class) or not.,10742
"general, method us a model-based propose leads statistical to This instead.","This leads us to propose a general, statistical model-based method instead.",10743
"Pandemic have responsible humans outbreaks recent clones, globally. for gastroenteritis spread in of","Pandemic clones, responsible for recent outbreaks of gastroenteritis in humans have spread globally.",10744
of profile the acquired have isolates recombination genetic location. current their most that means Frequent,Frequent recombination means that most isolates have acquired the genetic profile of their current location.,10745
distant genome found of between a example strong regions. We single epistasis,We found a single example of strong epistasis between distant genome regions.,10746
bacterial A with secretion majority strains a VI of killing. type system associated had,A majority of strains had a type VI secretion system associated with bacterial killing.,10747
with strains remaining by signaling. biofilm genes formation and regulated The c-di-GMP associated had,The remaining strains had genes associated with biofilm formation and regulated by c-di-GMP signaling.,10748
two also HRFM parameters: initiation rate and The ribosomal includes (constant) positive elongation the rate.,The HRFM also includes two positive parameters: ribosomal initiation rate and the (constant) elongation rate.,10749
to of the translation-elongation. Our may results better contribute mechanisms of and understanding evolution a,Our results may contribute to a better understanding of the mechanisms and evolution of translation-elongation.,10750
the The underlying mechanism. optimized that assumption is translation evolution,The underlying assumption is that evolution optimized the translation mechanism.,10751
Nonlocal simple for image powerful techniques filters and denoising. are,Nonlocal filters are simple and powerful techniques for image denoising.,10752
a methods Currently in environment. popular consuming computing very time are desktop typically estimation,Currently popular estimation methods are typically very time consuming in a desktop computing environment.,10753
"solution algorithm a solution. design greedy a for fast Further, we","Further, we design a greedy algorithm solution for a fast solution.",10754
Two and are (n-conorm) conorm of neutrosophic classes (n-norm) neutrosophic defined. norm,Two classes of neutrosophic norm (n-norm) and neutrosophic conorm (n-conorm) are defined.,10755
listed of logic of are Examples neutrosophic to last section. in physics the applications,Examples of applications of neutrosophic logic to physics are listed in the last section.,10756
for human on resulting detection model performance to the PASCAL state-of-the-art The leads datasets.,The resulting model leads to state-of-the-art performance for human detection on the PASCAL datasets.,10757
can problem Learning be an on then as optimization manifold. expressed Grassmann a,Learning can then be expressed as an optimization problem on a Grassmann manifold.,10758
pooling. on spatial basis low-level Our built new visual features and features the of are,Our new features are built on the basis of low-level visual features and spatial pooling.,10759
translational spatial and the invariance robustness improves of detection thus pooling process. Incorporating the the,Incorporating spatial pooling improves the translational invariance and thus the robustness of the detection process.,10760
networks. to guarantee be discrete/continuous wide can theoretical applied a range Markov Our of,Our theoretical guarantee can be applied to a wide range of discrete/continuous Markov networks.,10761
"exist. have focused models absolute on which competitions, Ecological optima in","Ecological models have focused on absolute competitions, in which optima exist.",10762
time. manuscript proposes to a describe the at same This both new way,This manuscript proposes a new way to describe both at the same time.,10763
a with of theory. version begins r/K-selection revised It,It begins with a revised version of r/K-selection theory.,10764
surveillance in is intelligent video system. object the Moving challenging most task recognition,Moving object recognition is the most challenging task in intelligent video surveillance system.,10765
"been different have many regard, this on proposed methods. techniques based In","In this regard, many techniques have been proposed based on different methods.",10766
"All all, these to develop necessary in techniques. exceedingly make robust it","All in all, these make it necessary to develop exceedingly robust techniques.",10767
results proposed recognition rate. Experimental achieves an show approach that our excellent,Experimental results show that our proposed approach achieves an excellent recognition rate.,10768
principled a and framework inference understanding Probabilistic computation. both for offers behaviour cortical,Probabilistic inference offers a principled framework for understanding both behaviour and cortical computation.,10769
"and inhibition and were inhibition excitation. balanced, Second, lagged excitation","Second, excitation and inhibition were balanced, and inhibition lagged excitation.",10770
is tooth deposition biofilm microbial the a of Dental material substratum. over,Dental biofilm is the deposition of microbial material over a tooth substratum.,10771
"delineation, current expert to bias. consuming is standard, and Manual the time gold subject","Manual delineation, the current gold standard, is time consuming and subject to expert bias.",10772
"the gamboostLSS, models. we fit method to R provide boosting these a package With","With the R package gamboostLSS, we provide a boosting method to fit these models.",10773
available model and regularized choice regression within selection framework. this are Variable naturally,Variable selection and model choice are naturally available within this regularized regression framework.,10774
hierarchical learning. This algorithm supervised recognition novel based feature face proposes on a large-scale paper,This paper proposes a novel face recognition algorithm based on large-scale supervised hierarchical feature learning.,10775
feature and learning two parts: large-scale The learning. of model consists hierarchical approach,The approach consists of two parts: hierarchical feature learning and large-scale model learning.,10776
The feature levels learning supervised feature in searches granularity of in a three hierarchical way.,The hierarchical feature learning searches feature in three levels of granularity in a supervised way.,10777
by supervised few channels learning. activate distinguish most We a,We activate a few most distinguish channels by supervised learning.,10778
that show training samples accuracy more notable Experiments will improvement. bring,Experiments show that more training samples will bring notable accuracy improvement.,10779
and FRGC experiments on LFW. We conduct,We conduct experiments on FRGC and LFW.,10780
protocol approach Results the notably. proposed that same outperforms the algorithms show under existing,Results show that the proposed approach outperforms existing algorithms under the same protocol notably.,10781
fish-eye cameras and two conventional a camera).,two conventional cameras and a fish-eye camera).,10782
experiments proposed demonstrate the of method. real Simulation effectiveness the and,Simulation and real experiments demonstrate the effectiveness of the proposed method.,10783
"Moreover, develop which is online. calibration the available toolbox, we camera","Moreover, we develop the camera calibration toolbox, which is available online.",10784
between SPD eigenvalues. It their evaluates similarity through two matrices the,It evaluates the similarity between two SPD matrices through their eigenvalues.,10785
classification of optimizing values by optimal are The parameter proxy performance. sought a,The optimal parameter values are sought by optimizing a proxy of classification performance.,10786
multi-channel research. a plays vital in of biological Visualization role Background: microscopy data,Background: Visualization of multi-channel microscopy data plays a vital role in biological research.,10787
managed of detail structure. data an levels The at is by octree different volumetric,The volumetric data is managed at different levels of detail by an octree structure.,10788
"Thus, time idle of biologists is reduced. the being","Thus, the time of biologists being idle is reduced.",10789
an unsupervised the combine processes with of results. visualization We intuitive accurate,We combine accurate unsupervised processes with an intuitive visualization of the results.,10790
progression. Various models to epidemic infer details of the that proposed been describe have dynamics,Various models have been proposed to infer details of the dynamics that describe epidemic progression.,10791
include derived coalescent These theory. Kingman's inference approaches from,These include inference approaches derived from Kingman's coalescent theory.,10792
simulated genealogies known epidemiological by analyzing Comparisons sets parameters. made of under are precisely,Comparisons are made by analyzing sets of genealogies simulated under precisely known epidemiological parameters.,10793
in indispensable analysis is data Clustering scientific disciplines. many for,Clustering is indispensable for data analysis in many scientific disciplines.,10794
"remains clusters sparse particularly from for heavy challenging, data. high-dimensional Detecting noise","Detecting clusters from heavy noise remains challenging, particularly for high-dimensional sparse data.",10795
"issue. graph-theoretic address algorithm this framework, the present paper Based a novel to proposes on","Based on graph-theoretic framework, the present paper proposes a novel algorithm to address this issue.",10796
effectiveness and of the real-world Extensive data experiments on validate our algorithms. scientific toy,Extensive experiments on toy and real-world scientific data validate the effectiveness of our algorithms.,10797
sequence and of empirical findings are variation. to related These epigenetic studies,These findings are related to empirical studies of sequence and epigenetic variation.,10798
Simultaneous and (SDS). Segmentation call We this Detection task,We call this task Simultaneous Detection and Segmentation (SDS).,10799
"a bounding requires just a box. box Unlike detection, segmentation SDS classical and not","Unlike classical bounding box detection, SDS requires a segmentation and not just a box.",10800
"individual classical require we object segmentation, semantic Unlike instances.","Unlike classical semantic segmentation, we require individual object instances.",10801
"top- We then our proposals. refine category-specific, bottom-up use to figure-ground predictions down","We then use category-specific, top- down figure-ground predictions to refine our bottom-up proposals.",10802
"tools provide unpack for provide directions that and work. we Finally, diagnostic performance future","Finally, we provide diagnostic tools that unpack performance and provide directions for future work.",10803
on vide and our pro- as examples. We illustrate argument justification several methodology a heuristic,We pro- vide a heuristic argument as justification and illustrate our methodology on several examples.,10804
of as We font-recognition address categorization problem the analysis and textures.,We address the font-recognition problem as analysis and categorization of textures.,10805
We machines for wavelet using complex transform use and features support vector classification. extract,We extract features using complex wavelet transform and use support vector machines for classification.,10806
discovery for miscommunication. and the This opportunities process slows provides both,This both slows the discovery process and provides opportunities for miscommunication.,10807
can fields which macroscopic be techniques with Neurons such generate as magneto-encephalography. recorded magnetic,Neurons generate magnetic fields which can be recorded with macroscopic techniques such as magneto-encephalography.,10808
findings. these test to We numerical illustrate properties and suggest by experiments simulations these,We illustrate these properties by numerical simulations and suggest experiments to test these findings.,10809
the of philosophical a knowledge. question purpose of analysis of is technologization possibility Article of,Article purpose is the analysis of a question of possibility of technologization of philosophical knowledge.,10810
corresponding to technologization. results parameters) as precisely applied set to,to precisely corresponding set parameters) to applied results as technologization.,10811
Transformation of its technologization. philosophy of possibilities allows revealing sense of,Transformation of sense of philosophy allows revealing possibilities of its technologization.,10812
investigated. theory making in application decision An is of this,An application of this theory in decision making is investigated.,10813
is sets soft discussed. of a neutrosophic Also introduced possibility and measure similarity two,Also a similarity measure of two possibility neutrosophic soft sets is introduced and discussed.,10814
methods and Analytic computational developed have numerous found physics statistical disciplines. within applications in,Analytic and computational methods developed within statistical physics have found applications in numerous disciplines.,10815
"solve a such problem to statistical In genetics. this in long-standing we letter, use methods","In this letter, we use such methods to solve a long-standing problem in statistical genetics.",10816
"posed and [J.B.S. Waddington by problem, Haldane The","The problem, posed by Haldane and Waddington [J.B.S.",10817
and may applications population framework Extensions in have of beyond. the genetics,Extensions of the framework may have applications in population genetics and beyond.,10818
stage of population of reproducing the evolution is individuals. The,The stage of evolution is the population of reproducing individuals.,10819
"small-world behave networks similarly. networks, scale-free Simulations and lattices, of perturbed","Simulations of perturbed lattices, small-world networks, and scale-free networks behave similarly.",10820
"information may Therefore, clinically corrupted. be important quantitative","Therefore, clinically important quantitative information may be corrupted.",10821
"observed patterns far, successfully predicted model So simple this spread. of has","So far, this simple model has successfully predicted observed patterns of spread.",10822
current allows invasion. risk according The to structure of their spatial territories ranking,The spatial structure allows ranking territories according to their current risk of invasion.,10823
"standard, often and subject time annotation, is gold current bias. human Manual consuming to","Manual annotation, current gold standard, is time consuming and often subject to human bias.",10824
decision frameworks reasoning such Existing simple decision-theoretic networks structures and data processes. as use,Existing decision-theoretic reasoning frameworks such as decision networks use simple data structures and processes.,10825
results large very our problems spaces. show algorithm on with Experimental information working,Experimental results show our algorithm working on problems with very large information spaces.,10826
"alteration Furthermore, macrophage important the functions of (e.g.","Furthermore, the alteration of important macrophage functions (e.g.",10827
zinc intracellular toxicity. to of The is studied dissolution/uptake also characterize further zinc,The intracellular dissolution/uptake of zinc is also studied to further characterize zinc toxicity.,10828
"unaffected zinc at phagocytosis moderately toxic Conversely, appear concentrations. levels glutathione and","Conversely, glutathione levels and phagocytosis appear unaffected at moderately toxic zinc concentrations.",10829
is formation Biofilm on tooth a of microbial material substrata.,Biofilm is a formation of microbial material on tooth substrata.,10830
show quantification degree results precision. consistency and high of a The,The quantification results show a high degree of consistency and precision.,10831
"method fast multipliers of to alternating achieve an we (ADMM) propose Second, direction reconstruction.","Second, we propose an alternating direction method of multipliers (ADMM) to achieve fast reconstruction.",10832
warm up is to procedure multi-scale A start speed convergence. the proposed,A multi-scale warm start procedure is proposed to speed up the convergence.,10833
are Applications demonstrated. to data real matching in stereo,Applications to real data in stereo matching are demonstrated.,10834
PSNR tested Performance calculating SSIM by is images. of and,Performance is tested by calculating PSNR and SSIM of images.,10835
be exploited natural up by template matching. The speed blurring target can fast to search,The natural blurring can be exploited to speed up target search by fast template matching.,10836
algorithm transitive blur. We matching incorporating template image controlled extend fast for by,We extend transitive algorithm for fast template matching by incorporating controlled image blur.,10837
A less speedup. larger efficient group computations more guarantees size and,A larger efficient group size guarantees less computations and more speedup.,10838
EGS auto-correlation our is as algorithm proposed component Optimizing in used (OptA) a algorithm.,EGS algorithm is used as a component in our proposed Optimizing auto-correlation (OptA) algorithm.,10839
and algorithm. satellite proposed effectiveness demonstrate image of the aerial on Experiments the datasets,Experiments on satellite and aerial image datasets demonstrate the effectiveness of the proposed algorithm.,10840
In complete explanatory et induction al. Inverse for subsumption Yamamoto,In Inverse subsumption for complete explanatory induction Yamamoto et al.,10841
"path to structured from this In semantic instantiations. logical shallow consider context, to the we","In this context, we consider the path from structured to logical to shallow semantic instantiations.",10842
JZ-extension traditional approaches. well-justified diverges The more resulting from semantics,The resulting well-justified JZ-extension semantics diverges from more traditional approaches.,10843
recorded segments. nerves from homologous more anterior posterior and we in Simultaneously,Simultaneously we recorded from homologous nerves in more anterior and posterior segments.,10844
the released. when were promptly was limb reversed These changes,These changes were promptly reversed when the limb was released.,10845
"(e.g. present, mandatory no At prophylaxis so exists, control procedures","At present, no prophylaxis exists, so mandatory control procedures (e.g.",10846
These spread limiting the suggest disease. for of intervention lines results the some of,These results suggest some lines of intervention for limiting the spread of the disease.,10847
"seeds. the of good choice on depends segmentation A result set a ""correct"" for","A good segmentation result depends on a set of ""correct"" choice for the seeds.",10848
erroneous segmentation can results. This to lead,This can lead to erroneous segmentation results.,10849
"is cellular automatic growing image In region algorithm seeded segmentation. paper, an for this proposed","In this paper, an automatic seeded region growing algorithm is proposed for cellular image segmentation.",10850
"First, regions (ROIs) the from image. the of interest extracted preprocessed","First, the regions of interest (ROIs) extracted from the preprocessed image.",10851
"initial the based automatically seeds on extracted image. the ROIs Second, from selected are","Second, the initial seeds are automatically selected based on ROIs extracted from the image.",10852
"selected Third, reprehensive the algorithm. most learning a using machine are seeds","Third, the most reprehensive seeds are selected using a machine learning algorithm.",10853
"corresponds Finally, where into segmented region a is cellular regions seed. image to the each","Finally, the cellular image is segmented into regions where each region corresponds to a seed.",10854
"can systems molecular of dynamic all networks viewed Indeed, as interactions. biological be","Indeed, all biological systems can be viewed as dynamic networks of molecular interactions.",10855
made by DCT is setting zero. the coefficients to smallest sparse The image,The image is made sparse by setting the smallest DCT coefficients to zero.,10856
of image. be same the transformation algorithm can used considering whole The,The same algorithm can be used considering transformation of the whole image.,10857
Reconstructions black are colour of white and considered. and images,Reconstructions of black and white and colour images are considered.,10858
"districts most defined are skeleton curves. described more accurately piecewise by However,","However, most skeleton districts are more accurately described by piecewise defined curves.",10859
a accuracy challenging resolution issue Preserving in super images. is,Preserving accuracy is a challenging issue in super resolution images.,10860
"having images The different be inclination focal retrieved size. and illumination, could length,","The retrieved images could be having different focal length, illumination, inclination and size.",10861
in areas. devices average PM limited are large measuring monitoring Current over,Current monitoring devices are limited in measuring average PM over large areas.,10862
present two and extraction between gap which algorithms the bridge We environmental haze monitoring. image,We present two algorithms which bridge the gap between image haze extraction and environmental monitoring.,10863
compare three algorithm We methods. removal to our known haze,We compare our algorithm to three known haze removal methods.,10864
under-researched a sensing problem. Accurate images digital from and atmospheric challenging is,Accurate atmospheric sensing from digital images is a challenging and under-researched problem.,10865
pedestrian. Pedestrian image classifiers windows contain decide which a,Pedestrian classifiers decide which image windows contain a pedestrian.,10866
reasoning image An applies analogous for sequences.,An analogous reasoning applies for image sequences.,10867
"(SSL) paradigm. we a stacked learning specifically, More classifiers train using sequential pedestrian","More specifically, we train pedestrian classifiers using a stacked sequential learning (SSL) paradigm.",10868
Caltech. on known test well dataset: We also a,We also test on a well known dataset: Caltech.,10869
"Interestingly, the accuracy most the dangerous i.e. improves more situations, at SSL","Interestingly, SSL improves more the accuracy at the most dangerous situations, i.e.",10870
the to when close pedestrian a camera. is,when a pedestrian is close to the camera.,10871
"selection. deleterious, rates be should mostly by mutation are natural mutations reduced Because","Because mutations are mostly deleterious, mutation rates should be reduced by natural selection.",10872
"However, the provide also raw adaptation. mutations material for","However, mutations also provide the raw material for adaptation.",10873
are These represented model. by spatial mixture relationships a,These spatial relationships are represented by a mixture model.,10874
useful applications. The and graphics computer field light camera vision for is,The light field camera is useful for computer graphics and vision applications.,10875
these applications. is Calibration step an essential for,Calibration is an essential step for these applications.,10876
"the rectify image we After calibrated calibration, camera can captured using parameters. the by","After calibration, we can rectify the captured image by using the calibrated camera parameters.",10877
fixed. usually viewpoints The assumes and technique calibration working multi-camera volume are that the,The multi-camera calibration technique usually assumes that the working volume and viewpoints are fixed.,10878
simultaneously. the parameters applied is optimization refine Global all to then,Global optimization is then applied to refine all the parameters simultaneously.,10879
some animals caveats for the domestic of future. in propose mtDNA We studies,We propose some caveats for mtDNA studies of domestic animals in the future.,10880
object functional matching joint for segmentation variational developed. and shape A is,A functional for joint variational object segmentation and shape matching is developed.,10881
w.r.t. transport The on formulation optimal is based,The formulation is based on optimal transport w.r.t.,10882
"a is confined non-convex, to low-dimensional the is variable. non-convexity overall While functional","While the overall functional is non-convex, non-convexity is confined to a low-dimensional variable.",10883
has DNA emerged a for cost-effective identification. barcoding species approach as,DNA barcoding has emerged as a cost-effective approach for species identification.,10884
is release the LV The performance BOLD. of assessed Barcoding on data of,The performance of LV Barcoding is assessed on the data release of BOLD.,10885
"decades, the rate. few a fast has been at optimization For last developing","For the last few decades, optimization has been developing at a fast rate.",10886
Bio-inspired by algorithms optimization nature. metaheuristics are inspired,Bio-inspired optimization algorithms are metaheuristics inspired by nature.,10887
"to engineering, problems solve and been applied in economics, algorithms domains. have different other These","These algorithms have been applied to solve different problems in engineering, economics, and other domains.",10888
protein dynamics. its the information Determining folding core process about yields and a of folding,Determining the folding core of a protein yields information about its folding process and dynamics.,10889
The linked been has psychological of flow human state to optimizing performance.,The psychological state of flow has been linked to optimizing human performance.,10890
model propose Intelligence of for computational We (AI) flow a simple agents. Artificial,We propose a simple computational model of flow for Artificial Intelligence (AI) agents.,10891
the Maximizing control a serves meta for agent. as flow the,Maximizing the flow serves as a meta control for the agent.,10892
open interesting for in work. promising testbed synthetic directions Results and are future a,Results in a synthetic testbed are promising and open interesting directions for future work.,10893
traditionally by of seizure-initiating comprised pathologic cortical epilepsy tissue `foci'. Drug-resistant characterized is,Drug-resistant epilepsy is traditionally characterized by pathologic cortical tissue comprised of seizure-initiating `foci'.,10894
particle algorithms to and filtering fast Existing easy are extremely implement.,Existing particle filtering algorithms are extremely fast and easy to implement.,10895
both and of time accuracy finite asymptotic are Theory on established. convergence L-CCA the,Theory on both the asymptotic convergence and finite time accuracy of L-CCA are established.,10896
the Thresholds in communities. many biological of dynamics occur,Thresholds occur in the dynamics of many biological communities.,10897
branches model transcritical crossing the The scenario. of curve in critical bifurcation two a exhibits,The model exhibits two branches of the critical curve crossing in a transcritical bifurcation scenario.,10898
possible the a this The ecosystem. result indicates role for significant in estuarine microparasite,The result indicates a possible significant role for the microparasite in this estuarine ecosystem.,10899
highest output. The scores final with proposals our are,The proposals with highest scores are our final output.,10900
of regions different and Biological functionally topologically identifies networks species. between alignment network conserved,Biological network alignment identifies topologically and functionally conserved regions between networks of different species.,10901
both methods use ASs. different NCFs and Different different,Different methods use both different NCFs and different ASs.,10902
"to evaluate alignment improve quality. MI-GRAAL newer GHOST Here, we further against potentially","Here, we evaluate MI-GRAAL against newer GHOST to potentially further improve alignment quality.",10903
"approach that important several have asked thus Also, systematically we questions been not far.","Also, we approach several important questions that have not been asked systematically thus far.",10904
"resulting Existing methods determine alignment(s). this more-less affect arbitrarily, the could which","Existing methods determine this more-less arbitrarily, which could affect the resulting alignment(s).",10905
assume better. sizes neighborhood larger are Existing methods that,Existing methods assume that larger neighborhood sizes are better.,10906
"general for alignment methods. several evaluation network our give fair recommendations Together, results a of","Together, our results give several general recommendations for a fair evaluation of network alignment methods.",10907
"After subtraction, forehead and detected. is background face","After background subtraction, face and forehead is detected.",10908
"step, space In skin in forehead HSV H-S the plane calculated. histograms is second","In the second step, forehead skin H-S plane histograms in HSV space is calculated.",10909
hand algorithm. The of using extracted are Freeman contours code chain,The contours of hand are extracted using Freeman chain code algorithm.,10910
target. drug ligand-binding a Cyclophilin a significant and A is multi-functional protein,Cyclophilin A is a multi-functional ligand-binding protein and a significant drug target.,10911
"the are only labels available task. detection of those for a fraction Unfortunately, small","Unfortunately, only a small fraction of those labels are available for the detection task.",10912
and Models are at available software .,Models and software are available at .,10913
Intelligent Road  recognition in one  Systems. the Transport sign core of technologies is,  Road sign recognition is one of the core technologies in Intelligent Transport Systems.,10914
energy moves. by efficiently model fusion minimized This can be,This energy model can be efficiently minimized by fusion moves.,10915
this So functions have do called not resolution-limit-free limit. quality,So called resolution-limit-free quality functions do not have this limit.,10916
"previously introduced graph for This partitioning. property that was is, clustering, hard","This property was previously introduced for hard clustering, that is, graph partitioning.",10917
is resulting to strongly linked The Potts Constant function Model. the,The resulting function is strongly linked to the Constant Potts Model.,10918
In out this too turns property. a be strong to setting resolution-limit-free,In this setting resolution-limit-free turns out to be too strong a property.,10919
under We report hypergravitational bacterial growth stress.,We report bacterial growth under hypergravitational stress.,10920
under Experiments to also hypergravitational carried conditions. consumption were investigate nutrient out,Experiments were also carried out to investigate nutrient consumption under hypergravitational conditions.,10921
to difference between considerable subjected conditions. hypergravity show normal Our samples and results,Our results show considerable difference between samples subjected to hypergravity and normal conditions.,10922
train deep are a classifier. Neural These to (CNN) Convolutional used Network views random,These random views are used to train a deep Convolutional Neural Network (CNN) classifier.,10923
the number reduces The FP/vol. proposed of method,The proposed method reduces the number of FP/vol.,10924
results previous work. The marked improvement show upon,The results show marked improvement upon previous work.,10925
the the represent theory Recently developed Waddington's quasi-potential to landscape. researchers,Recently researchers developed the quasi-potential theory to represent the Waddington's landscape.,10926
this to address a two-step aggregation problem. paradigm propose We,We propose a two-step aggregation paradigm to address this problem.,10927
as of The regularized discrete aggregation energy. the designed step is optimization a global,The aggregation step is designed as the discrete optimization of a global regularized energy.,10928
with map flow two the field The the occlusion is jointly throughout steps. estimated,The occlusion map is estimated jointly with the flow field throughout the two steps.,10929
We motion approach generic vectors. exemplar-based a filling propose with for occlusion,We propose a generic exemplar-based approach for occlusion filling with motion vectors.,10930
concepts ways. These adapted in two are,These concepts are adapted in two ways.,10931
"of semirings. to correspondence ""invariant-polymorphism"" is generalized (min,max) Firstly, the","Firstly, the correspondence of ""invariant-polymorphism"" is generalized to (min,max) semirings.",10932
"Moreover, not operator of such is the guarantee necessary. even existence an for a","Moreover, even a guarantee for the existence of such an operator is not necessary.",10933
or algorithm the solution The the finds discards either problem.,The algorithm either finds the solution or discards the problem.,10934
only The problem majority possible if no the has is polymorphism. latter,The latter is possible only if the problem has no majority polymorphism.,10935
influenza dynamics. the of We patterns global study spatio-temporal,We study the global spatio-temporal patterns of influenza dynamics.,10936
significance. health implications of important public and understanding has basic scientific these A patterns,A basic understanding of these patterns has important public health implications and scientific significance.,10937
The the of a model. provided state-transition standard semantics language is by,The semantics of the language is provided by a standard state-transition model.,10938
of structure The physics provide for rich spatiotemporal a the experience. our environment,The physics of the environment provide a rich spatiotemporal structure for our experience.,10939
"optimization penalization TV hard non-smooth minimize. are to to that problems However, leads","However, TV penalization leads to non-smooth optimization problems that are hard to minimize.",10940
or be can losses used with penalties. algorithm other This,This algorithm can be used with other losses or penalties.,10941
ADNI a The on applied is problem on dataset. the classification algorithm,The algorithm is applied on a classification problem on the ADNI dataset.,10942
in vision. used Bag-of-Words is representation computer The (BoW) widely,The Bag-of-Words (BoW) representation is widely used in computer vision.,10943
"of by in .The which optimization, swarm group. performs highly behavior influenced the is creature,",".The swarm optimization, which is highly influenced by the behavior of creature, performs in group.",10944
is prior drawn A ever-increasing likelihood uniformly from above replacement an point threshold. the,A replacement point is drawn uniformly from the prior above an ever-increasing likelihood threshold.,10945
"case. whether is test to develop this a I work, In check the statistical this","In this work, I develop a statistical test to check whether this is the case.",10946
"Test"" sampling verify ""Shrinkage algorithms nested in to controlled This is a useful environment.","This ""Shrinkage Test"" is useful to verify nested sampling algorithms in a controlled environment.",10947
"comparison algorithm however, MULTINEST. inefficient RADFRIENDS to This is, in","This RADFRIENDS algorithm is, however, inefficient in comparison to MULTINEST.",10948
holds infrequently. CAR condition We that rather the show,We show that the CAR condition holds rather infrequently.,10949
and results This literature CAR in obtained previous on MRE. interconnects the generalizes and,This generalizes and interconnects previous results obtained in the literature on CAR and MRE.,10950
Precise e.g. of spatio-temporal underly patterns neuronal action potentials,Precise spatio-temporal patterns of neuronal action potentials underly e.g.,10951
sensory of activities. control and representations muscle,sensory representations and control of muscle activities.,10952
License on parking monitoring and recognition management Plate role important systems. an traffic the plays,License Plate recognition plays an important role on the traffic monitoring and parking management systems.,10953
"using plate number morphology Then, by has detection been specified. some and location operations, edge","Then, by using some edge detection and morphology operations, plate number location has been specified.",10954
"plat is its away. tilt, tilt has Finally, removed the if","Finally, if the plat has tilt, its tilt is removed away.",10955
shows the clustering We found that important coefficient changes the database. of this during growth,We found that the clustering coefficient shows important changes during the growth of this database.,10956
face in research is Human computer currently community. the perception an vision active area,Human face perception is currently an active research area in the computer vision community.,10957
important Skin and primary one most is stages this purpose. for the of detection,Skin detection is one of the most important and primary stages for this purpose.,10958
"are proposed far, this to many done case. So approaches","So far, many approaches are proposed to done this case.",10959
kind these show cannot an that Results skin. of for methods every provide approach accurate,Results show that these methods cannot provide an accurate approach for every kind of skin.,10960
equalization) technique images. image the An to enhancement and is enhance fingerprint face used (histogram,An image enhancement technique (histogram equalization) is used to enhance the face and fingerprint images.,10961
fingerprint face the the extracted of technique. Salient features were Gabor using filter and,Salient features of the face and fingerprint were extracted using the Gabor filter technique.,10962
and approach performance the is effective. validated The proposed of is,The performance of the proposed approach is validated and is effective.,10963
(LPR) License an management. on parking Plate plays the Recognition important monitoring role and traffic,License Plate Recognition (LPR) plays an important role on the traffic monitoring and parking management.,10964
as granulation. in uncertainty aspect input of information recognized a is crucial Modeling data the,Modeling of the input data uncertainty is recognized as a crucial aspect in information granulation.,10965
Recognition plays parking important on and monitoring management. Plate an traffic License the role,License Plate Recognition plays an important role on the traffic monitoring and parking management.,10966
those essential. their of very transportation service and better restriction Administration becomes tools for,Administration and restriction of those transportation tools for their better service becomes very essential.,10967
its Then image using are techniques. distinguished characters processing,Then its characters are distinguished using image processing techniques.,10968
"Nearest classifier for recognition. character used was Finally, K Neighbour (KNN)","Finally, K Nearest Neighbour (KNN) classifier was used for character recognition.",10969
control presents a substations. fuzzy in system for volt/var (VVC) paper distribution integrated This inference,This paper presents a fuzzy inference system for integrated volt/var control (VVC) in distribution substations.,10970
"approaches case. are this So to done many far, proposed","So far, many approaches are proposed to done this case.",10971
an for accurate these methods approach Results skin. every provide of show cannot kinds,Results show these methods cannot provide an accurate approach for every kinds of skin.,10972
"paper, an statistical technique. using problem this this features approach In is to proposed solve","In this paper, an approach is proposed to solve this problem using statistical features technique.",10973
advantages. are of other noise computational Low complexity low sensitivity and some,Low noise sensitivity and low computational complexity are some of other advantages.,10974
to higher enables novel further A contribution model effectively in dimensional spaces. the work,A further novel contribution enables the model to work effectively in higher dimensional spaces.,10975
reveals from be Equalizing which bias for this identified. co-evolutionary local signals native-contacts can true,Equalizing for this bias reveals true co-evolutionary signals from which local native-contacts can be identified.,10976
"Importantly, obtained other significantly interactions are values for not all equalization. by affected the parameter","Importantly, parameter values obtained for all other interactions are not significantly affected by the equalization.",10977
This the demand the low possibilities opens models. for new computational extremely improving of quality,This extremely low computational demand opens new possibilities for improving the quality of the models.,10978
requires reconstruction. as a microarray FASTCORMICS and data input Genome-scale,FASTCORMICS requires as input microarray data and a Genome-scale reconstruction.,10979
in applications. analysis Texture image important plays an many role,Texture plays an important role in many image analysis applications.,10980
paper. for this let information please and updating changing me this,please let me for changing this information and updating this paper.,10981
for situations the which in prediction include is These at the task relevant hand. information,These include situations in which the information is relevant for the prediction task at hand.,10982
probability Expectation is a theory. central notion in,Expectation is a central notion in probability theory.,10983
The of makes expectation sense other for notions uncertainty. also of notion,The notion of expectation also makes sense for other notions of uncertainty.,10984
stage Translation is an expression. in important gene,Translation is an important stage in gene expression.,10985
important production. question An maximize is how to protein,An important question is how to maximize protein production.,10986
for or Each various providing be can of viewed these evidence against as hypotheses. observations,Each of these observations can be viewed as providing evidence for or against various hypotheses.,10987
orders. not for when Complications that orders are total partial examining present arise,Complications arise when examining partial orders that are not present for total orders.,10988
of important power properties necessary evidence to seems expressive capture This .,This expressive power seems necessary to capture important properties of evidence .,10989
tracking-by-detection. tracking multi-target most of popular is the approaches to One,One of the most popular approaches to multi-target tracking is tracking-by-detection.,10990
and work empirically. both we Here extend that theoretically,Here we extend that work both theoretically and empirically.,10991
"explanation methods dataset. a this applies Furthermore, give whether to particular also we verifying for","Furthermore, we also give methods for verifying whether this explanation applies to a particular dataset.",10992
defined. conditional measures notion is plausibility of algebraic general A,A general notion of algebraic conditional plausibility measures is defined.,10993
all MDPs (DM) states the assume that and maker Traditional decision knows actions.,Traditional MDPs assume that the decision maker (DM) knows all states and actions.,10994
"in may true be of many this not interest. situations However,","However, this may not be true in many situations of interest.",10995
"we can found polynomial in In when time. solution characterize near-optimal a be particular,","In particular, we characterize when a near-optimal solution can be found in polynomial time.",10996
actually of heterogeneity the a from determine that the typical information condition. follows This fault,This follows from the heterogeneity of the information that actually determine a typical fault condition.,10997
characters and is are plate segmented. localized Vehicle license,Vehicle license plate is localized and characters are segmented.,10998
"method of finding real and a the is place time expressed. plate, For novel","For finding the place of plate, a novel and real time method is expressed.",10999
A character on based chain used technique recognition. new directional robust code is for and,A new and robust technique based on directional chain code is used for character recognition.,11000
system model and is implemented a is developed. hardware integrated prototype software and The working,The hardware and software integrated system is implemented and a working prototype model is developed.,11001
fundamental calculated from The theory. the rates transitional are,The transitional rates are calculated from the fundamental theory.,11002
its studied. and is state the transporter steady stability is The found of,The steady state of the transporter is found and its stability is studied.,11003
directions translocation The compared. rates two in glucose and are steps in different (xylose),The glucose (xylose) translocation rates in two directions and in different steps are compared.,11004
time predicted. transition temperature is The non-Arrhenius transport the mean rate dependence of the and,The non-Arrhenius temperature dependence of the transition rate and the mean transport time is predicted.,11005
"become cluster distribution after rescaling. function, this species-independent distance statistics nearest-neighbor Correlation and","Correlation function, cluster statistics and nearest-neighbor distance distribution become species-independent after this rescaling.",11006
interrelated. (species local to spatial structure observables Global appear frequencies) be and,Global observables (species frequencies) and local spatial structure appear to be interrelated.,11007
goal is query. satisfaction DL of given a of plan terms The in,The plan goal is given in terms of satisfaction of a DL query.,11008
Such the often system. adaptations of the architecture involve software reconfiguring,Such adaptations often involve reconfiguring the software architecture of the system.,11009
the simple Cellular computation. most and common parallel of are Automata models (CA),Cellular Automata (CA) are the most common and simple models of parallel computation.,11010
"some the experimental presented. methodologies Finally, of results are","Finally, the experimental results of some methodologies are presented.",11011
evolutionary to We underlying like the would identify structures.,We would like to identify the underlying evolutionary structures.,11012
This displays overhead yet desirable theoretical parallelisation highly the approach minimises properties.,This approach minimises the parallelisation overhead yet displays highly desirable theoretical properties.,11013
"of consequence work, demonstrates this that variable our [...]. analysis In importances","In consequence of this work, our analysis demonstrates that variable importances [...].",11014
from the differs work by others. validation method used each that the found We,We found that the validation method used by each work differs from the others.,11015
"comparison direct cannot made. works among a be So,","So, a direct comparison among works cannot be made.",11016
vision. processing area is in an computer research important Image,Image processing is an important research area in computer vision.,11017
the rule plays image in research. segmentation vital Image processing,Image segmentation plays the vital rule in image processing research.,11018
There many exist so image methods segmentation. for,There exist so many methods for image segmentation.,11019
image used be also for Clustering segmentation. can,Clustering can also be used for image segmentation.,11020
for Scientific applications require perception engineering Several images of and sampled development. complex merging,Several Scientific and engineering applications require merging of sampled images for complex perception development.,11021
"at for images intensity merged most requirements, are level. cases, such In","In most cases, for such requirements, images are merged at intensity level.",11022
arising being of of out phase incoherent The properties main modulation is lost. problem intensity,The main problem is incoherent modulation of intensity arising out of phase properties being lost.,11023
"compensate amplitude combined is phase merge losses, to and In order demanded. these","In order to compensate these losses, combined phase and amplitude merge is demanded.",11024
outperforms PMF constrained to show that MAP simple treatment the model estimation. Bayesian We this,We show that this Bayesian treatment to the constrained PMF model outperforms simple MAP estimation.,11025
"truncated precision propose model. a this, of light In we","In light of this, we propose a truncated precision model.",11026
Our suggest that to delay model this overfitting. results experimental tends,Our experimental results suggest that this model tends to delay overfitting.,11027
modular updates. satisfy that call updates We property this,We call updates that satisfy this property modular updates.,11028
define in discuss two inappropriate expert We updates their and use influential systems. modular,We define modular updates and discuss their inappropriate use in two influential expert systems.,11029
of found the dengue hotspots center Machala. transmission of near We significant,We found significant hotspots of dengue transmission near the center of Machala.,11030
for analysis. techniques This thesis proposes spatio-spectral hyperspectral image,This thesis proposes spatio-spectral techniques for hyperspectral image analysis.,11031
"paragraphs. text to Typography in and hierarchical layout lead organisation words, text of the lines,","Typography and layout lead to the hierarchical organisation of text in words, text lines, paragraphs.",11032
perspective. problem This paper of segmentation in scenes from a the hierarchical text natural addresses,This paper addresses the problem of text segmentation in natural scenes from a hierarchical perspective.,11033
"examined we by induced can Here, whether oscillations acoustic be stimulation. slow effectively rhythmic","Here, we examined whether slow oscillations can be effectively induced by rhythmic acoustic stimulation.",11034
"can results if of be estimation classes. However, biased there the the are latent","However, the results of the estimation can be biased if there are latent classes.",11035
problem. Bayesian solves new estimation We then the propose that a procedure,We then propose a new Bayesian estimation procedure that solves the problem.,11036
a patches. is two-tiered on image The propagation method information by classifying based hierarchical,The method is based on a hierarchical two-tiered information propagation by classifying image patches.,11037
"another two interesting obvious levels in results. These interactive qualitatively with ways, are producing one","These two levels are interactive with one another in obvious ways, producing qualitatively interesting results.",11038
of is conceive a structure. cross-linked correlated of every We piece that knowledge and cluster,We conceive that every piece of knowledge is a cluster of cross-linked and correlated structure.,11039
the This abstract theory framework an on is based concepts. by provided,This theory is based on an abstract framework provided by the concepts.,11040
tumor validated using Lemon-Tree candidate by pathway were and analyses. predicted survival driver Novel genes,Novel candidate driver genes predicted by Lemon-Tree were validated using tumor pathway and survival analyses.,11041
"of for strong extended solutions calculated and them. We these both games, re-examined","We re-examined both of these games, and calculated extended strong solutions for them.",11042
"are Several to negligible animal called i.e. senescence, exhibit is considered species what","Several animal species are considered to exhibit what is called negligible senescence, i.e.",11043
tissues stresses. with exceptional NMR of genotoxic endurance various These correlate well phenotypic to observations,These phenotypic observations correlate well with exceptional endurance of NMR tissues to various genotoxic stresses.,11044
"organism be an of lifelong Therefore, longevity. key the transcriptional may of a determinant stability","Therefore, the lifelong transcriptional stability of an organism may be a key determinant of longevity.",11045
"the aging network has and relation genetic been stress-resistance stability, However, between defined. not exact","However, the exact relation between genetic network stability, stress-resistance and aging has not been defined.",11046
data Analysis common a dimensional high of is task.,Analysis of high dimensional data is a common task.,11047
"aim work, at this blank. we filling this In","In this work, we aim at filling this blank.",11048
"of a the multiple However, combined such diseases result cancer genes. as of are effect","However, diseases such as cancer are a result of the combined effect of multiple genes.",11049
mapping the establish the rigorous of and thermodynamics. We thermodynamics a ensemble develop to,We develop the thermodynamics of the ensemble and establish a rigorous mapping to thermodynamics.,11050
the is fact case that This an the widely despite available. effective vaccine is,This is the case despite the fact that an effective vaccine is widely available.,11051
the study Numerical of the simulations are qualitative aid to dynamics. system's generated,Numerical simulations are generated to aid the study of the system's qualitative dynamics.,11052
"the resources Consequently, and relatively computational at fovea at periphery. are fewer the allocated more","Consequently, more computational resources are allocated at the fovea and relatively fewer at the periphery.",11053
by modeling this improve masking. contrast our observer further work numerical In we,In this work we further improve our numerical observer by modeling contrast masking.,11054
in structure. same share payoff the Cooperative those agents both games which are,Cooperative games are those in which both agents share the same payoff structure.,11055
alternative methods a environments. search reasonable are partially Policy to for value-based observable methods,Policy search methods are a reasonable alternative to value-based methods for partially observable environments.,11056
quality across We dimensions. noise with good and a present sampling method different blue performance,We present a blue noise sampling method with good quality and performance across different dimensions.,11057
"selects prior rays samples spoke-dart and sampling, samples these from from method, The shoots rays.","The method, spoke-dart sampling, shoots rays from prior samples and selects samples from these rays.",11058
"optimization, and Delaunay spoke-dart graph motion robotic applications planning. construction, demonstrate approximate for global We","We demonstrate spoke-dart applications for approximate Delaunay graph construction, global optimization, and robotic motion planning.",11059
"best this the always not penalty. However, with is performance estimator achieved","However, the best estimator performance is not always achieved with this penalty.",11060
We two course the solutions the problems. final believe proposal our of is of that,We of course believe that our proposal is the final solutions of the two problems.,11061
"the Indeed when cast, won. word was battle already the was","Indeed when the word was cast, the battle was already won.",11062
"and Jean-Charles Editors: Sanchez. Bini Guest Juan Hochstrasser Calvete, Luca , Natacha J. Denis Turck,","Guest Editors: Luca Bini , Juan J. Calvete, Natacha Turck, Denis Hochstrasser and Jean-Charles Sanchez.",11063
affine affine parameter feature without achieves needing over space. to invariance The full simulate,The feature achieves full affine invariance without needing to simulate over affine parameter space.,11064
"achieved and could ways in SIFT. invariance similar Rotation, scaling to be translation","Rotation, translation and scaling invariance could be achieved in ways similar to SIFT.",11065
"efficiently. recent Furthermore, owing be parameter computed convex in breakthrough optimization, such to could","Furthermore, owing to recent breakthrough in convex optimization, such parameter could be computed efficiently.",11066
place demonstrate recognition our major as We effectiveness its in will application.,We will demonstrate its effectiveness in place recognition as our major application.,11067
network help parameters. influential them identify engineers for important results knowledge are they as These,These results are important for knowledge engineers as they help them identify influential network parameters.,11068
"This particular networks. for problem, is of fundamental and a interest Bayesian","This is a fundamental problem, and of particular interest for Bayesian networks.",11069
We a updating new for incomplete rule with observations. propose probabilities,We propose a new rule for updating probabilities with incomplete observations.,11070
"posterior rule general, well as probabilities, produces determinate In our lower as partially decisions.","In general, our rule produces lower posterior probabilities, as well as partially determinate decisions.",11071
logical the ignorance of about This incompleteness is the consequence mechanism. a,This is a logical consequence of the ignorance about the incompleteness mechanism.,11072
Asymptotic distribution the are approximations of proposed.,Asymptotic approximations of the distribution are proposed.,11073
"is First, Normalized Pixel Difference a called new image feature (NPD) proposed.","First, a new image feature called Normalized Pixel Difference (NPD) is proposed.",11074
"is the new scale image. bounded, and able feature original reconstruct The to invariant, is","The new feature is scale invariant, bounded, and is able to reconstruct the original image.",11075
"a This soft-cascade only classifier face way, detection. single is to needed unconstrained handle","This way, only a single soft-cascade classifier is needed to handle unconstrained face detection.",11076
the with to smart face element key are human gesture Hand and system. interact the,Hand gesture and human face are the key element to interact with the smart system.,11077
also used games reality. the and be gesture could computer in based virtual method planned,also the planned method could be used in gesture based computer games and virtual reality.,11078
"of the are but style as partially in analyzed Savage, Preferences acts L. ordered. among","Preferences among acts are analyzed in the style of L. Savage, but as partially ordered.",11079
weaker than The counts. postulates considered three on are Savage's rationality,The rationality postulates considered are weaker than Savage's on three counts.,11080
Principle derived this Thing in is The Sure setting.,The Sure Thing Principle is derived in this setting.,11081
popular re-identification have algorithms for person metric learning Several as evaluated baselines. been,Several popular metric learning algorithms for person re-identification have been evaluated as baselines.,11082
"naturally type from This knowledge arises of data, others. prior observational experimental among and","This type of knowledge naturally arises from prior experimental and observational data, among others.",11083
"to such knowledge. addition, take prior search-operator novel a In advantage of is proposed","In addition, a novel search-operator is proposed to take advantage of such prior knowledge.",11084
sequences. Sequential decision solvable often simulating action are possible problems approximately by future,Sequential decision problems are often approximately solvable by simulating possible future action sequences.,11085
stages. Such processing achieve compaction computation also subsequent can a information for,Such a computation can also achieve information compaction for subsequent processing stages.,11086
"presented, is results Feature possible detailed. are a implementation being VLSI and extraction","Feature extraction results are being presented, and a possible VLSI implementation is detailed.",11087
of This leads imprecise an the to Markov definition chain.,This leads to the definition of an imprecise Markov chain.,11088
of to to generalisation Perron-Frobenius leads Theorem non-trivial a This the chains. Markov classical imprecise,This leads to a non-trivial generalisation of the classical Perron-Frobenius Theorem to imprecise Markov chains.,11089
environmental theoretically also various guarantee practical its under improves We active performance that sensing conditions.,We also theoretically guarantee its active sensing performance that improves under various practical environmental conditions.,11090
algorithm We MDPs (FOMDPs). heuristic search a present first-order solving for,We present a heuristic search algorithm for solving first-order MDPs (FOMDPs).,11091
"state the directly FOMDP Firstly, on propositionalization. we abstraction avoiding apply","Firstly, we apply state abstraction directly on the FOMDP avoiding propositionalization.",11092
state Such abstraction. kind abstraction as of referred to is firstorder,Such kind of abstraction is referred to as firstorder state abstraction.,11093
"general, intractable. correction summation terms over In all is full","In general, full summation over all correction terms is intractable.",11094
algorithm in for the approach al. Chertkov an We presented develop et,We develop an algorithm for the approach presented in Chertkov et al.,11095
are underlie in Congenital to synaptic due plasticity. molecular that pathways dysfunctions deficits cognitive frequently,Congenital cognitive dysfunctions are frequently due to deficits in molecular pathways that underlie synaptic plasticity.,11096
"mutations cbp models long-term late-phase LTM of potentiation (LTP). RTS, In and of impair animal","In animal models of RTS, mutations of cbp impair LTM and late-phase long-term potentiation (LTP).",11097
"by were effects changes, parameter LTP. simulated and increased drug Plausible many","Plausible drug effects were simulated by parameter changes, and many increased LTP.",11098
synaptic Importantly parameter paired not changes alter these weight. basal did,Importantly these paired parameter changes did not alter basal synaptic weight.,11099
A phosphodiesterase acetyltransferase effective. simulated similarly pair was and activator an a that inhibitor,A pair that simulated a phosphodiesterase inhibitor and an acetyltransferase activator was similarly effective.,11100
pairs strong present. For additive was synergism both,For both pairs strong additive synergism was present.,11101
expression that the the the intense layer. in more granular Thresholding signal shows is profiles,Thresholding the expression profiles shows that the signal is more intense in the granular layer.,11102
between insightful symmetries and establish chains. rapid of orbital model mixing connection Markov an We,We establish an insightful connection between model symmetries and rapid mixing of orbital Markov chains.,11103
"models. the MCMC we first probabilistic present for lifted graphical algorithm Thus,","Thus, we present the first lifted MCMC algorithm for probabilistic graphical models.",11104
the empirical approach. Both of effectiveness the demonstrate and analytical results efficiency and,Both analytical and empirical results demonstrate the effectiveness and efficiency of the approach.,11105
We for bandit tree propose search. alternative algorithms,We propose alternative bandit algorithms for tree search.,11106
aspect is sensory but of active sensing. understood poorly the important An role processing of,An important but poorly understood aspect of sensory processing is the role of active sensing.,11107
in developmental deep place progressive takes under architectures Learning scheme. a,Learning takes place in deep architectures under a progressive developmental scheme.,11108
in the (FIR) domain. filters are designed Finite-impulse-response implemented and frequency,Finite-impulse-response (FIR) filters are designed and implemented in the frequency domain.,11109
"computer gesture extensive recognition, recognition applications games. language virtual and sign Hand possesses reality, in","Hand gesture recognition possesses extensive applications in virtual reality, sign language recognition, and computer games.",11110
In presented. real-time novel is this hand system paper a for and gesture recognition approach,In this paper a novel and real-time approach for hand gesture recognition system is presented.,11111
to applied the stage Cross-correlation extraction the is coefficient feature gesture recognize on it. In,In feature extraction stage the Cross-correlation coefficient is applied on the gesture to recognize it.,11112
"susceptible handles. is the to However, close in artifacts, this deformation to method particular","However, this method is susceptible to artifacts, in particular close to the deformation handles.",11113
"linear do and we extend To so, generic regularization introduced on a approach. build recently","To do so, we build on and extend a recently introduced generic linear regularization approach.",11114
minimizes runtime. complexity only a implementation This and small has the impact on,This minimizes the implementation complexity and has only a small impact on runtime.,11115
of due images to mixed pixels resolution hyperspectral Hyperspectral contain spatial sensors. low,Hyperspectral images contain mixed pixels due to low spatial resolution of hyperspectral sensors.,11116
this for multilayer of the (MLNMF) unmixing. In purpose using letter NMF hyperspectral proposed we,In this letter we proposed using multilayer NMF (MLNMF) for the purpose of hyperspectral unmixing.,11117
"signature matrices. sparse a as of product be this In modeled can approach, spectral matrix","In this approach, spectral signature matrix can be modeled as a product of sparse matrices.",11118
of number In iteratively the fact decomposes a observation in matrix MLNMF layers.,In fact MLNMF decomposes the observation matrix iteratively in a number of layers.,11119
is real and The on applied algorithm datasets. proposed synthetic,The proposed algorithm is applied on synthetic and real datasets.,11120
Synthetic endmembers generated spectral data USGS based library. on is from,Synthetic data is generated based on endmembers from USGS spectral library.,11121
proposed dataset AVIRIS for of as dataset real used method. evaluation a Cuprite been has,AVIRIS Cuprite dataset has been used as a real dataset for evaluation of proposed method.,11122
measures. SAD based and Results are AAD quantified of on experiments,Results of experiments are quantified based on SAD and AAD measures.,11123
are and available data research videos purposes. publicly All for source the,All source videos and the data are publicly available for research purposes.,11124
migration. Results: migration seasonally improves rates the of varying Incorporating modeling,Results: Incorporating seasonally varying migration rates improves the modeling of migration.,11125
"Consequently, not approaches deconvolution cameras current for blind are applicable. traditional","Consequently, current blind deconvolution approaches for traditional cameras are not applicable.",11126
"entropy-driven, would facilitating peptides, this homoaggregation so identical their Predominantly co-localize presentation. collective","Predominantly entropy-driven, this homoaggregation would co-localize identical peptides, so facilitating their collective presentation.",11127
are fine-tuned time over avoid Concentrations self-proteins to of this. evolutionary,Concentrations of self-proteins are fine-tuned over evolutionary time to avoid this.,11128
of in described. enhancement targets technique clutter the point is for A,A technique for the enhancement of point targets in clutter is described.,11129
inverse in are processing. commonplace image biomedical problems Ill-posed,Ill-posed inverse problems are commonplace in biomedical image processing.,11130
about typically prior imposing Their truth. the knowledge solution ground latent requires,Their solution typically requires imposing prior knowledge about the latent ground truth.,11131
empirical in an or hoc} ad fashion. chosen are hence {\em Priors mostly,Priors are hence mostly chosen in an {\em ad hoc} or empirical fashion.,11132
interpretations prior. possible by We and conclude discussing the limitations of the,We conclude by discussing the limitations and possible interpretations of the prior.,11133
present with characteristics. a architecture We two main cerebellar,We present a cerebellar architecture with two main characteristics.,11134
sensory errors. in that respond The first increases to one is spikes complex,The first one is that complex spikes respond to increases in sensory errors.,11135
been trees learning. used Decision widely have in machine,Decision trees have been widely used in machine learning.,11136
tree fuzzy should to able The handle decision such be data.,The decision tree should be able to handle such fuzzy data.,11137
method construct a presents fuzzy paper tree. This to decision,This paper presents a method to construct fuzzy decision tree.,11138
likely in significant on Progress challenging question to is this be near the future.,Progress on this challenging question is likely to be significant in the near future.,11139
analogous way. In efforts under case of the are well height,In the case of height analogous efforts are well under way.,11140
iid. are These where not two training are data scenarios typical,These are two typical scenarios where training data are not iid.,11141
with points an experiment a permissible Consider finite set design of conditions. trial representing,Consider an experiment with a finite set of design points representing permissible trial conditions.,11142
also forgetting. and use functions the to remembering explain We,We also use the functions to explain remembering and forgetting.,11143
our algorithm with on compare benchmark evaluate five and We ten datasets. algorithms state-of-the-art,We evaluate and compare our algorithm with ten state-of-the-art algorithms on five benchmark datasets.,11144
"results on algorithm algorithms. better Experimental accuracy average, that, than show gives ten our these","Experimental results show that, on average, our algorithm gives better accuracy than these ten algorithms.",11145
Our consists technique detection algorithm. and descriptor of keypoint a novel,Our technique consists of a novel descriptor and keypoint detection algorithm.,11146
"approach teaching to e.g. A a concept, visual natural","A natural approach to teaching a visual concept, e.g.",11147
"is to images. bird species, a relevant show","a bird species, is to show relevant images.",11148
"not concept However, a relevant well. equally images all represent","However, not all relevant images represent a concept equally well.",11149
"not words, they other are In necessarily iconic.","In other words, they are not necessarily iconic.",11150
exactly makes an image iconic? And what,And what exactly makes an image iconic?,11151
We images. combine predict to them iconicity of new unseen,We combine them to predict iconicity of new unseen images.,11152
a that iconicity discriminatively predictor ratings. also trained iconicity with direct is We propose,We also propose a direct iconicity predictor that is discriminatively trained with iconicity ratings.,11153
"both systems, get performance. approaches prediction By that combining an human we iconicity","By combining both systems, we get an iconicity prediction that approaches human performance.",11154
our reduction. variance variates are well-defined that prove control and provide a We positive,We prove that our control variates are well-defined and provide a positive variance reduction.,11155
reduction. these optimise variance for variates control we how Furthermore to show,Furthermore we show how to optimise these control variates for variance reduction.,11156
Kalman its in situations this have and paper. to extensions non-linear/non-Gaussian filter used been,Kalman filter and its extensions to non-linear/non-Gaussian situations have been used in this paper.,11157
method and compared both has to off-line on-line existing methods. been Performance,Performance has been compared to both existing on-line method and off-line methods.,11158
been methods. not has possible previously extraction these features with The existing of,The extraction of these features has previously not been possible with existing methods.,11159
and filament The existing of outperforms runtime methods in accuracy. extraction computational terms clearly FS,The FS clearly outperforms existing methods in terms of computational runtime and filament extraction accuracy.,11160
benchmark and FS implementation database The available the source. open are as the of,The implementation of the FS and the benchmark database are available as open source.,11161
"appearances. and include Some human variation in environments, illumination variation in challenges changes, common","Some common challenges include illumination changes, variation in environments, and variation in human appearances.",11162
need robust method human Thus a pose estimation. for there for a is,Thus there is a need for a robust method for human pose estimation.,11163
operations performed subtraction. are from on then background blobs foreground Morphological,Morphological operations are then performed on foreground blobs from background subtraction.,11164
the However the conditions. relies of on performance initial ICP heavily,However the performance of ICP relies heavily on the initial conditions.,11165
"by occluded the Principal done regions Analysis restoration After is detected, are Component (PCA).","After the occluded regions are detected, restoration is done by Principal Component Analysis (PCA).",11166
form face Features reconstructed face from the images in normals. non-occluded are extracted the of,Features are extracted from the reconstructed non-occluded face images in the form of face normals.,11167
conclusions key Our also. are these hold found in to examples,Our key conclusions are found to hold in these examples also.,11168
theoretical illustrated examples; approach compared by with The general are findings numerical is calculations.,The general approach is illustrated by examples; theoretical findings are compared with numerical calculations.,11169
Mammographic Experiments Society on (MIAS) dataset. Analysis Image held,Experiments held on Mammographic Image Analysis Society (MIAS) dataset.,11170
is in science. The Hastings computational tool key a algorithm,The Hastings algorithm is a key tool in computational science.,11171
"it conceptually be detailed While balance, by mathematically grasp. difficult to justified can","While mathematically justified by detailed balance, it can be conceptually difficult to grasp.",11172
"present intuitive and understand complementary the ways algorithm. we derive Here, and to two","Here, we present two complementary and intuitive ways to derive and understand the algorithm.",11173
capability confer a We a possibly network. stochastic note fluctuations dynamical can to new that,We note that stochastic fluctuations can possibly confer a new dynamical capability to a network.,11174
structures set theory Algebraic very developed. rapidly are using soft,Algebraic structures using soft set theory are very rapidly developed.,11175
all subjects Samples fourth and the anchors were concatenated for along dimension.,Samples for all anchors and subjects were concatenated along the fourth dimension.,11176
inference mechanism is on An Birds Angry domain. the ensemble proposed,An ensemble inference mechanism is proposed on the Angry Birds domain.,11177
model online the is to according performed parameters. Learning an estimation strategy procedure for,Learning procedure is performed according to an online estimation strategy for the model parameters.,11178
changes on the the environment in We exerting which selection case stochastically. focus,We focus on the case in which the environment exerting selection changes stochastically.,11179
therefore detect propose We it. experimental observable an to,We therefore propose an experimental observable to detect it.,11180
genetic linked one site at selection shapes sites. of patterns variation at Natural,Natural selection at one site shapes patterns of genetic variation at linked sites.,11181
from substantial linked of a inference suggests effect selection non-classic Our further sweeps.,Our inference further suggests a substantial effect of linked selection from non-classic sweeps.,11182
"face pre-processing, performance with data anti-spoofing some Combined drastically. the improves","Combined with some data pre-processing, the face anti-spoofing performance improves drastically.",11183
provides memory patterns. basis conclusion maintaining Another mathematical concept or for a,Another mathematical conclusion provides a basis for maintaining memory or concept patterns.,11184
can itself The space processes. through automatic self-organisation,The self-organisation can space itself through automatic processes.,11185
might The structures afterwards. help with also memory finding process,The process might also help with finding memory structures afterwards.,11186
model conclusions. and new version This further cognitive the extended with some integrates provides existing,This extended version integrates further with the existing cognitive model and provides some new conclusions.,11187
analysis sex of of revealed the a number Our in recombination. distribution differences,Our analysis revealed a number of sex differences in the distribution of recombination.,11188
commonly images. Complex exist natural in structures,Complex structures commonly exist in natural images.,11189
methods. forms issue prior The fundamental challenge a for,The issue forms a fundamental challenge for prior methods.,11190
"scales. patch varying sizes downsizing region-based Different or images, measure we from","Different from varying patch sizes or downsizing images, we measure region-based scales.",11191
"model, inference is map. single-scale a selected obtain Through saliency our to information","Through our inference model, single-scale information is selected to obtain a saliency map.",11192
improves cannot on images well traditionally. detection that handled many Our be method quality,Our method improves detection quality on many images that cannot be handled well traditionally.,11193
method based bootstrap embeddings hypothesis for on kernel nonparametric proposed. tests is distribution wild A,A wild bootstrap method for nonparametric hypothesis tests based on kernel distribution embeddings is proposed.,11194
makes potentially object for number them learning classes. and recognizing a higher suitable of This,This makes them potentially suitable for learning and recognizing a higher number of object classes.,11195
takes learns The contour approach spatial frequent simple and their configurations. fragments,The approach takes simple contour fragments and learns their frequent spatial configurations.,11196
effective Shannon method entropy. fuzzy and facial using theory suggests paper for This recognition an,This paper suggests an effective method for facial recognition using fuzzy theory and Shannon entropy.,11197
entropy the theory of Combination methods. fuzzy other complication and Shannon eliminates of,Combination of fuzzy theory and Shannon entropy eliminates the complication of other methods.,11198
"scheduling programming. cumulative environments, uncertain in using investigate constraint We","We investigate cumulative scheduling in uncertain environments, using constraint programming.",11199
algorithm filtering the FlexC this in sweep detail paper constraint. of the global We dynamic,We detail in this paper the dynamic sweep filtering algorithm of the FlexC global constraint.,11200
graph. eigenmapping graph of deriving for Laplacian the the the Then adopted is Laplacian,Then the Laplacian eigenmapping is adopted for deriving the graph Laplacian of the graph.,11201
"plugging by SGC graph the can GT the obtained be Finally, framework. into Laplacian conventional","Finally, SGC can be obtained by plugging the graph Laplacian into the conventional GT framework.",11202
and SR. merits SGC GT inherits the both of,SGC inherits the merits of both GT and SR.,11203
"of SGC robustness to power the SR, discriminating the and GT. Compared improves","Compared to SR, SGC improves the robustness and the discriminating power of GT.",11204
"exploits the SGC GT, sufficiently to data. Compared whole","Compared to GT, SGC sufficiently exploits the whole data.",11205
"study model. stochastic in Here, Moran we tunneling a","Here, we study stochastic tunneling in a Moran model.",11206
recognition ways countless are the define to for task. measures There,There are countless ways to define measures for the recognition task.,11207
feature be used textural can for Co-occurrence extraction. matrix,Co-occurrence matrix can be used for textural feature extraction.,11208
models of form into Two from fit original to our updated were their framework. the,Two of the models were updated from their original form to fit into our framework.,11209
observed biological molecules phenomenon. on Clustering widely of membranes a is,Clustering of molecules on biological membranes is a widely observed phenomenon.,11210
"mechanism form systems and unclear. in these the remains by Yet, are which clusters maintained","Yet, the mechanism by which clusters form and are maintained in these systems remains unclear.",11211
"it activation. promotes has that further been discovered Recently, Ras Ras activated","Recently, it has been discovered that activated Ras promotes further Ras activation.",11212
the such inference. two to of with propose We deal ideas challenges novel computational,We propose two novel ideas to deal with the computational challenges of such inference.,11213
"a empirically. approach The introduces which investigate we bias,","The approach introduces a bias, which we investigate empirically.",11214
two care intensive at looked We delivery. and applications: detection malware,We looked at two applications: malware detection and intensive care delivery.,11215
environments recurring techniques. planning pair Operating requires new vehicles between adversarial a in origin-destination,Operating vehicles in adversarial environments between a recurring origin-destination pair requires new planning techniques.,11216
game introduced. zero-sum A two players is,A two players zero-sum game is introduced.,11217
this of damage. second maximize goal the is to The player,The goal of the second player is to maximize this damage.,11218
Different routing strategies environments. unstructured over defined in are order to environment models compute,Different environment models are defined in order to compute routing strategies over unstructured environments.,11219
Finally is world real tested on framework the environments.,Finally the framework is tested on real world environments.,11220
"dominant localized around these populations space, sequence In are transiently strains. two","In sequence space, these two populations are localized around transiently dominant strains.",11221
as spurious outcomes are to solutions. commonly These referred,These outcomes are commonly referred to as spurious solutions.,11222
"aim In this we issue. this at addressing contribution,","In this contribution, we aim at addressing this issue.",11223
"and of is sets Characterizations these discovered. an inclusion are given, relationship","Characterizations of these sets are given, and an inclusion relationship is discovered.",11224
"we occur. In spurious possible part, the the second investigate where scenarios solutions","In the second part, we investigate the possible scenarios where spurious solutions occur.",11225
the of advices Some choice practical given for nonlinearity are function.,Some advices are given for the practical choice of nonlinearity function.,11226
of oceanography Commonplace in is drifter the collection ocean positions.,Commonplace in oceanography is the collection of ocean drifter positions.,11227
"posterior flow. on data, drifter Using obtain the underlying is to it possible a","Using drifter data, it is possible to obtain a posterior on the underlying flow.",11228
"ocean underlying drifter, Through controlling flow. an to attempt the our we improve knowledge of","Through controlling an ocean drifter, we attempt to improve our knowledge of the underlying flow.",11229
better is as of variance a the flow. smaller interpreted A understanding,A smaller variance is interpreted as a better understanding of the flow.,11230
numerically the study We birth-death process. also associated stochastic,We also study numerically the associated stochastic birth-death process.,11231
approach. in Numerical general analytical agreement simulations with our are,Numerical simulations are in general agreement with our analytical approach.,11232
paper Information for selection model This in (KIC) introduces Kernel-based analysis. Criterion regression,This paper introduces Kernel-based Information Criterion (KIC) for model selection in regression analysis.,11233
"""right-handed"" individuals and every ""left-handed"" are assayed. in common strongly line biased Very","Very strongly biased ""left-handed"" and ""right-handed"" individuals are common in every line assayed.",11234
role of The noise NO and of effect investigated. are,The role of NO and effect of noise are investigated.,11235
stress apoptosis induce the to system. Excess,Excess stress induce apoptosis to the system.,11236
of urban relatively small amount movements. is freight covering a There research,There is a relatively small amount of research covering urban freight movements.,11237
searching selecting involve thought and is discrete Creativity amongst to candidates. multiple idea,Creativity is thought to involve searching and selecting amongst multiple discrete idea candidates.,11238
study these paper reports This on tests a that between theories.,This paper reports on a study that tests between these theories.,11239
beneficial by Representing has codes proven many tasks. visual images recognition for compact,Representing images by compact codes has proven beneficial for many visual recognition tasks.,11240
ways assess evaluated. the describe systems are AI critically will different and We,We will describe and critically assess the different ways AI systems are evaluated.,11241
We approach. traditional the on evaluation task-oriented first focus,We first focus on the traditional task-oriented evaluation approach.,11242
"Human of benchmarks identify We and kinds problem three peer confrontation. evaluation: discrimination,","We identify three kinds of evaluation: Human discrimination, problem benchmarks and peer confrontation.",11243
results from those numerical are mean with compared field The calculations. obtained,The numerical results are compared with those obtained from mean field calculations.,11244
in-painting video completion solve methods data to and aim image tensor-based problems. Many,Many tensor-based data completion methods aim to solve image and video in-painting problems.,11245
"for were a only But, methods all developed dataset. single","But, all methods were only developed for a single dataset.",11246
temporal This recognition. representations proposes for characteristics video to paper equivalent in action encode,This paper proposes to encode equivalent temporal characteristics in video representations for action recognition.,11247
"To (TSP). achieve called scale pyramid temporal a we temporal develop invariance, method scale","To achieve temporal scale invariance, we develop a method called temporal scale pyramid (TSP).",11248
"paper, line a method In text is this proposed. identification","In this paper, a text line identification method is proposed.",11249
"variable. gaps But and the is non-uniform are interline handwritten documents, line in","But in handwritten documents, the line is non-uniform and interline gaps are variable.",11250
pre-processing stage consists Our of to a image. document method clean the,Our method consists of a pre-processing stage to clean the document image.,11251
separate region. manuscript doodles we from textual the get Then to the,Then we separate doodles from the manuscript to get the textual region.,11252
text the manuscript. we lines identify that on the After,After that we identify the text lines on the manuscript.,11253
role in wide hyperspectral a of (HU) plays unmixing a fundamental range applications. Hyperspectral,Hyperspectral unmixing (HU) plays a fundamental role in a wide range of hyperspectral applications.,11254
"this relieved. greatly are effects side channels of the outlier way, In","In this way, the side effects of outlier channels are greatly relieved.",11255
"over varies Besides, each level mixed observe pixel of we image grids. that the","Besides, we observe that the mixed level of each pixel varies over image grids.",11256
"so and The it and to highly resulted is optimize. hard is non-smooth, non-convex objective","The resulted objective is highly non-convex and non-smooth, and so it is hard to optimize.",11257
"algorithm a profound to propose contribution, we As it. efficient theoretical solve an","As a profound theoretical contribution, we propose an efficient algorithm to solve it.",11258
"provided. convergence computational the the proof analysis Meanwhile, complexity systematically are and","Meanwhile, the convergence proof and the computational complexity analysis are systematically provided.",11259
Artificial of in increased with recognition (AI) Use potential. has its environmental modelling Intelligence of,Use of Artificial Intelligence (AI) in environmental modelling has increased with recognition of its potential.,11260
of better the complexity required. However are with increase methods,However with the increase of complexity better methods are required.,11261
with Reasoning sets fuzzy achieved measures distance. be through such similarity can as and,Reasoning with fuzzy sets can be achieved through measures such as similarity and distance.,11262
comparison measure sets. of alleviates The aids and ambiguous combined automatic in results the fuzzy,The combined measure alleviates ambiguous results and aids in the automatic comparison of fuzzy sets.,11263
"mixed model. ), the resulting under phylogeny other and the then and parameters estimating model","), and then estimating the phylogeny and other model parameters under the resulting mixed model.",11264
Carlo Markov expensive. (MCMC) to Applying Monte sets algorithms computationally data is standard chain large,Applying standard Markov chain Monte Carlo (MCMC) algorithms to large data sets is computationally expensive.,11265
"In single mobility. human this including the spread work, for we a study Dengue serotype","In this work, we study the Dengue spread for a single serotype including human mobility.",11266
"We closely for infection certain and model intensity. observed rates, CA peaks that find reproduces","We find that for certain infection rates, CA model closely reproduces observed peaks and intensity.",11267
object a recognition known is domain. well in bias Dataset problem,Dataset bias is a well known problem in object recognition domain.,11268
"face issue, research. nonetheless, in explored rarely alignment is This","This issue, nonetheless, is rarely explored in face alignment research.",11269
"that plays integral part an we face show of study, performance. In dataset this alignment","In this study, we show that dataset plays an integral part of face alignment performance.",11270
"confidence. the body In with frame, are part mode the chosen positions each as highest","In each frame, body part positions are chosen as the mode with the highest confidence.",11271
This move outlined a complex elsewhere. reasons necessary in detail for is,This is a necessary move for complex reasons outlined in detail elsewhere.,11272
thought. paradox This creative to respect not with does exist,This paradox does not exist with respect to creative thought.,11273
"notice such many that is We be penetrations utilized, if untangled. information can","We notice that if such information is utilized, many penetrations can be untangled.",11274
"they are the decreases when applied. In operators performance scenarios, of these the","In these scenarios, the performance of the operators decreases when they are applied.",11275
images. and from multi-orientation filters Gabor features extract can multiscale face,Gabor filters can extract multi-orientation and multiscale features from face images.,11276
"regard paper, as smooth In GMPs surfaces. this we","In this paper, we regard GMPs as smooth surfaces.",11277
methods use we classification. Finally for joint construct subspace and histograms,Finally we construct joint histograms and use subspace methods for classification.,11278
Recent adaptive consistently inferred of Drosophila between high studies have evolution molecular rates species.,Recent studies have consistently inferred high rates of adaptive molecular evolution between Drosophila species.,11279
linkage adaptive affect correlations Here evolution. analyze how we,Here we analyze how linkage correlations affect adaptive evolution.,11280
"these yet cannot However, the between discriminated difference two be models experimentally.","However, the difference between these two models cannot yet be discriminated experimentally.",11281
"a process from timing. this investigated the perspective, We novel","We investigated this process from a novel perspective, the timing.",11282
$\alpha\theta$ beta The process with and $\beta\theta$. of density parameters equilibrium this is,The equilibrium density of this process is beta with parameters $\alpha\theta$ and $\beta\theta$.,11283
regions separated. spatial Two temporal are and two,Two temporal and two spatial regions are separated.,11284
to groupings in natural objects. of cluster goal is primary The discover analysis,The primary goal in cluster analysis is to discover natural groupings of objects.,11285
"shortcomings in is bioinformatics. its clustering dominant hierarchical in accuracy, method clustering the Despite","Despite its shortcomings in accuracy, hierarchical clustering is the dominant clustering method in bioinformatics.",11286
simultaneously. on multiple scales clustering operates Hierarchical,Hierarchical clustering operates on multiple scales simultaneously.,11287
"accommodates parameters, and clustering information separates data, convex on Our missing supports relationships. prior software","Our convex clustering software separates parameters, accommodates missing data, and supports prior information on relationships.",11288
for The processing (GPUs) ATI is speed. implemented nVidia software graphics and units maximal on,The software is implemented on ATI and nVidia graphics processing units (GPUs) for maximal speed.,11289
application local Most vision correspondences different on computer rely images. finding between algorithms,Most computer vision application rely on algorithms finding local correspondences between different images.,11290
algorithms at stable descriptors keypoints. centered and detect local scale-invariant compare These invariant,These algorithms detect and compare stable local invariant descriptors centered at scale-invariant keypoints.,11291
diverse methods. the raises very question comparison fair between a This of,This raises the question of a fair comparison between very diverse methods.,11292
"has far, entity in as been functional viewed whole cancer one studies. tumor a Thus","Thus far, one tumor has been viewed as a whole entity in cancer functional studies.",11293
technologies Emerging single-cell as clones. RNA-Seq such profile and genome tumor sequencing could,Emerging technologies such as single-cell genome sequencing and RNA-Seq could profile tumor clones.,11294
apply on from approach tissue. brain electron microscopy data this invertebrate We fly,We apply this approach on electron microscopy data from invertebrate fly brain tissue.,11295
et recent James paper by A al.,A recent paper by James et al.,11296
the that interactions model decrease finds mutualistic biodiversity ecosystems. of,finds that mutualistic interactions decrease the biodiversity of model ecosystems.,11297
on in visual application We propagation the simulation of outdoor and algorithm's sound highlight scenes.,We highlight the algorithm's application on simulation of sound and visual propagation in outdoor scenes.,11298
consists subpopulations often cell (clones). of multiple tumor A,A tumor often consists of multiple cell subpopulations (clones).,11299
a clone often one target of chemo-treatments Current tumor.,Current chemo-treatments often target one clone of a tumor.,11300
"reoccurs. the tumor kills clone, drug and other the that it clones Although overtake","Although the drug kills that clone, other clones overtake it and the tumor reoccurs.",11301
We call for proposed (GCFF). the Graph-Cuts F-formation method,We call the proposed method Graph-Cuts for F-formation (GCFF).,11302
noise from fluctuations study rates. arising residue-specific translational in elongation This analyzes,This study analyzes translational noise arising from fluctuations in residue-specific elongation rates.,11303
the contingency models Log-linear tables. workhorses are of analyzing popular,Log-linear models are the popular workhorses of analyzing contingency tables.,11304
"represent subtype-specific which identified, potentially mechanisms underlying Two tumorigenesis. core-signaling were networks","Two subtype-specific networks were identified, which potentially represent core-signaling mechanisms underlying tumorigenesis.",11305
"targets, effectively predicted subtype-specific networks validated. experimentally which drug Further, the were","Further, the networks effectively predicted subtype-specific drug targets, which were experimentally validated.",11306
for optical presented work. edge in in segmentation detector image is this quantum A environments,A quantum edge detector for image segmentation in optical environments is presented in this work.,11307
is presented A same version detector the Boolean of too.,A Boolean version of the same detector is presented too.,11308
"edge detector computational new with the states, version quantum basis The exclusively. works of","The quantum version of the new edge detector works with computational basis states, exclusively.",11309
outside so far quantum Said too. to processing algorithms is image extended,Said so far is extended to quantum algorithms outside image processing too.,11310
"However, although are they the equally are results quantum valid. different,","However, although they are different, the quantum results are equally valid.",11311
the is simulations clearly in seen latter . The computer,The latter is clearly seen in the computer simulations .,11312
"as have generality more explored markers reduced. consciousness specificity Moreover, was stimulus here the","Moreover, the consciousness markers explored here have more generality as stimulus specificity was reduced.",11313
and systems. feature ubiquitous a complex is in nature fundamental Variability of,Variability is ubiquitous in nature and a fundamental feature of complex systems.,11314
"genetic itself variance a studies, investigated have as control. however, trait Few under","Few studies, however, have investigated variance itself as a trait under genetic control.",11315
the degree heritable. demonstrate that itself We variability of is,We demonstrate that the degree of variability is itself heritable.,11316
"validated We alleles, deficiencies, genetic have transgenes. and results inducible null RNAi using these","We have validated these results using genetic deficiencies, null alleles, and inducible RNAi transgenes.",11317
"for other data types of However, (e.g.","However, for other types of data (e.g.",11318
involves Protein changes. function conformational binding often and,Protein binding and function often involves conformational changes.,11319
was two shown The earlier. existence two rigid structures dimensional of,The existence of two rigid two dimensional structures was shown earlier.,11320
"detection. networks for multi-stage deformable In deep object paper, this propose neural and convolutional we","In this paper, we propose multi-stage and deformable deep convolutional neural networks for object detection.",11321
has learning aspects. multiple This diagram innovations detection new deep in object,This new deep learning object detection diagram has innovations in multiple aspects.,11322
extensive component-wise also is experimental Detailed analysis evaluation. provided through,Detailed component-wise analysis is also provided through extensive experimental evaluation.,11323
image analysis from useful method to information basic matching in images. Template is extract a,Template matching is a basic method in image analysis to extract useful information from images.,11324
"paper, In new a we method pattern for matching. suggest this","In this paper, we suggest a new method for pattern matching.",11325
dimensional dimensional template from two the image into method image Our one transform vector.,Our method transform the template image from two dimensional image into one dimensional vector.,11326
movie viewing of data. presented real-world on results art the are State,State of the art results are presented on real-world movie viewing data.,11327
interplay speciation trophic structure interactions. and emerges the food-web between from The dynamical thereby,The food-web structure thereby emerges from the dynamical interplay between speciation and trophic interactions.,11328
coupled introduce by webs on We of metacommunity several evolutionary food multitrophic an habitats migration.,We introduce an evolutionary metacommunity of multitrophic food webs on several habitats coupled by migration.,11329
"between Additionally, to able migrate species the are habitats.","Additionally, species are able to migrate between the habitats.",11330
detection the problem binary each as This classification of target IR pixel. study a formulates,This study formulates the IR target detection as a binary classification problem of each pixel.,11331
of labeled models parameters learned these are from The MRF examples.,The parameters of these MRF models are learned from labeled examples.,11332
is incorporating performance improved detection further temporal background by The information subtraction. through,The detection performance is further improved by incorporating temporal information through background subtraction.,11333
this benchmark datasets for IR method effectiveness target detection. of demonstrate performances High on,High performances on benchmark datasets demonstrate effectiveness of this method for IR target detection.,11334
this automated The machinery contribution mediation is the does that. which paper of,The contribution of this paper is the automated mediation machinery which does that.,11335
method study case used comparison an it the with earlier in a For work. illustrates,For comparison it illustrates the method with a case study used in an earlier work.,11336
descriptor image property structure. not give ownership This does the of local,This descriptor property does not give ownership of the local image structure.,11337
to has clinical improve tests for osteoporosis method the potential This pathologies. diagnostics,This method has the potential to improve clinical diagnostics tests for osteoporosis pathologies.,11338
requires variables estimation completely of s-sparse The measurements. just with many independent as signals,The estimation of s-sparse signals with completely independent variables requires just as many measurements.,11339
fundamental metrics analysis network Network toolbox. form the of a part,Network metrics form a fundamental part of the network analysis toolbox.,11340
"learning. machine In this metrics probabilistic work, connect network we modern to","In this work, we connect network metrics to modern probabilistic machine learning.",11341
"formulate we node eigenvector-based importance. Bayesian determining centrality First, for an model","First, we formulate an eigenvector-based Bayesian centrality model for determining node importance.",11342
Gaussian bridge We model processes. to then this sparse,We then bridge this model to sparse Gaussian processes.,11343
process. jump the methods from ability Such require trajectories conditioned to the simulate ostensibly,Such methods ostensibly require the ability to simulate trajectories from the conditioned jump process.,11344
therefore improving conditioned processes. the jump efficiency three for of methods We propose simulating,We therefore propose three methods for improving the efficiency of simulating conditioned jump processes.,11345
We sequential also to our problem. adapt recently Monte Carlo a scheme proposed,We also adapt a recently proposed sequential Monte Carlo scheme to our problem.,11346
"We approach, the of MJP. two this two based implementations approximations of continuous consider on","We consider two implementations of this approach, based on two continuous approximations of the MJP.",11347
and state-of-the-art. of the validate real-world the against performance method our on We synthetic data,We validate the performance of our method on synthetic and real-world data against the state-of-the-art.,11348
edge to matching an of We use potential region. technique the unknown structure infer the,We use an edge matching technique to infer the potential structure of the unknown region.,11349
"structure, the texture curves. estimated by along synthesis the automatically Driven estimated performed is","Driven by the estimated structure, texture synthesis is performed automatically along the estimated curves.",11350
results approach favorable this techniques. show that existing state-of-the-art experiments outperform that Our achieves,Our experiments show that this approach achieves favorable results that outperform existing state-of-the-art techniques.,11351
"the an for in we algorithm raw deflectometric devise of Furthermore, interest object images. detecting","Furthermore, we devise an algorithm for detecting the object of interest in raw deflectometric images.",11352
gain to Experiments of previous methods. significant measurements a suggest in compared coverage single,Experiments suggest a significant gain of coverage in single measurements compared to previous methods.,11353
"on data To will research facilitate publicly our set reconstruction, make available. specular surface we","To facilitate research on specular surface reconstruction, we will make our data set publicly available.",11354
"Carlo introduce estimation which new class of exact Monte a we algorithms. methods, call We","We introduce a new class of Monte Carlo methods, which we call exact estimation algorithms.",11355
"differing parameters. structures In optimization, of often Bayesian with we over practical search numbers must","In practical Bayesian optimization, we must often search over structures with differing numbers of parameters.",11356
to our on adapted case. min-max principle is Courant-Fischer It the based,It is based on the Courant-Fischer min-max principle adapted to our case.,11357
causal The is shown techniques proposed state-of-the-art inference. method be to competitive for with,The proposed method is shown to be competitive with state-of-the-art techniques for causal inference.,11358
generates embedded agglomerative an idea self-taught hypotheses. that into localization clustering technique is This,This idea is embedded into an agglomerative clustering technique that generates self-taught localization hypotheses.,11359
function Tukey one robust multivariate the purposes. serving depth famous most is of tools,Tukey depth function is one of the most famous multivariate tools serving robust purposes.,11360
"combinatorial In issue algorithms. this by address this we paper, presenting computing two","In this paper, we address this computing issue by presenting two combinatorial algorithms.",11361
faster run existing minimal ones. and Both very than memory the much require,Both require very minimal memory and run much faster than the existing ones.,11362
the All Tukey they that indicate depth. experiments exact compute,All experiments indicate that they compute the exact Tukey depth.,11363
the a straightforward based of problem presented. study on analysis theory this probability In is,In this study a straightforward analysis of the problem based on probability theory is presented.,11364
"computer features shown vision. performance in have mid-level promising Recently,","Recently, mid-level features have shown promising performance in computer vision.",11365
"limb's are combined Secondly, part and limbs into the are parts mapped into states states.","Secondly, limbs are combined into parts and the limb's states are mapped into part states.",11366
as or parts Local Frequent FLPs. These referred to are Parts,These parts are referred to as Frequent Local Parts or FLPs.,11367
powerful The bag-of-FLP-based to representation. build us action allow FLPs,The FLPs allow us to build powerful bag-of-FLP-based action representation.,11368
"these this gaps. With fill to we intend paper,","With this paper, we intend to fill these gaps.",11369
can recognition? the How represented be for efficiently in memory object,How can the object be efficiently represented in memory for recognition?,11370
extracted? How information be semantic can reliably,How reliably can semantic information be extracted?,11371
the to from finally: are query the pipeline What And in factors limiting semantics? resulting,And finally: What are the limiting factors in the resulting pipeline from query to semantics?,11372
"other higher than method, the accuracy we fuzzy obtain methods. Using","Using the fuzzy method, we obtain higher accuracy than other methods.",11373
experimentally and VOC. validate We it PASCAL DPM the model call DeepPyramid resulting on,We call the resulting model DeepPyramid DPM and experimentally validate it on PASCAL VOC.,11374
"we the wave find that Additionally, of spontaneous generation activity. supports the spiral system","Additionally, we find that the system supports the spontaneous generation of spiral wave activity.",11375
are from experiments. in vitro both with qualitatively and quantitatively observations results Our consistent,Our results are both qualitatively and quantitatively consistent with observations from in vitro experiments.,11376
imagery matching different face across to (HFR) face domains. recognition refers Heterogeneous,Heterogeneous face recognition (HFR) refers to matching face imagery across different domains.,11377
in recent techniques HFR. a comprehensive established developments of review survey This and provides,This survey provides a comprehensive review of established techniques and recent developments in HFR.,11378
"of a detailed commonly and account we datasets benchmarks Moreover, offer for used evaluation.","Moreover, we offer a detailed account of datasets and benchmarks commonly used for evaluation.",11379
"a we In paper, present alignment. this approach regression deep face for","In this paper, we present a deep regression approach for face alignment.",11380
global consists of a deep and The layers. local architecture multi-stage layer,The deep architecture consists of a global layer and multi-stage local layers.,11381
the parameters. back-propagation with dropout We optimize to regression apply jointly the algorithm strategy the,We apply the back-propagation algorithm with the dropout strategy to jointly optimize the regression parameters.,11382
Experimental state-of-the-art show the . our approach achieves results that,Experimental results show that our approach achieves the state-of-the-art .,11383
the a study self-replicating We problem scheduling of factory.,We study the scheduling problem of a self-replicating factory.,11384
is scheduled several allowing lines in self-replication production to run parallel. Optimally by facilitated,Optimally scheduled self-replication is facilitated by allowing several production lines to run in parallel.,11385
is of analysis Object datasets objects. statistical (OODA) data oriented the analysis of complex,Object oriented data analysis (OODA) is the statistical analysis of datasets of complex objects.,11386
"contact data function news, be rate can easily of LPMs estimated. From the closing","From data of LPMs closing news, the contact rate function can be easily estimated.",11387
"Horner's method To be can end, used. this","To this end, Horner's method can be used.",11388
"optimal MCTS reported Recently, to to was find be able near schemes.","Recently, MCTS was reported to be able to find near optimal schemes.",11389
"had parameters However, to fine-tuned manually. several be","However, several parameters had to be fine-tuned manually.",11390
"only the a as We neighborhood time allowed computation leaving structure, suitable a found parameter.","We found a suitable neighborhood structure, leaving only the allowed computation time as a parameter.",11391
obtained by results are better similar by obtained those or SLS than The MCTS.,The results obtained by SLS are similar or better than those obtained by MCTS.,11392
"of population a consequence a drift. Over genetic substitutions as random acquires neutral time,","Over time, a population acquires neutral genetic substitutions as a consequence of random drift.",11393
"then across rates u$. are birth uniform \leq If $K sites,","If birth rates are uniform across sites, then $K \leq u$.",11394
Our to a model variety population of applied can structures. be,Our model can be applied to a variety of population structures.,11395
"genetic we in example, In the of the accumulation small one investigate intestine. mutations","In one example, we investigate the accumulation of genetic mutations in the small intestine.",11396
in are probability. Parse reported then decreasing trees order of,Parse trees are then reported in order of decreasing probability.,11397
Forecasting and may health inform these efforts. simulation interventions public of,Forecasting and simulation of these interventions may inform public health efforts.,11398
resources address will to commitment A and support long-term of the necessary outbreak. be,A long-term commitment of resources and support will be necessary to address the outbreak.,11399
 multispectral and  based This hyperspectral paper variational approach images. fusing presents a to,  This paper presents a variational based approach to fusing hyperspectral and multispectral images.,11400
Corrections expectation about the propagation are explained. expansion by then,Corrections by expansion about the expectation propagation are then explained.,11401
advanced sections. explored in are some and final topics applications the Finally inference,Finally some advanced inference topics and applications are explored in the final sections.,11402
are This between capturing mind-wandering FC suggests fluctuations that the effects. and SC,This suggests that the fluctuations between FC and SC are capturing mind-wandering effects.,11403
memory It also certain problems of limitations. overcomes,It also overcomes certain problems of memory limitations.,11404
"renewed approaches multiobjective optimization. Recently, has decomposition-based in there a evolutionary been interest for","Recently, there has been a renewed interest in decomposition-based approaches for evolutionary multiobjective optimization.",11405
"we this the In parameters. functions investigate and behavior different scalarizing of their paper,","In this paper, we investigate the behavior of different scalarizing functions and their parameters.",11406
main them encoding difference schemes. among their is The,The main difference among them is their encoding schemes.,11407
"instead considers sparsity. Different group of among points from LRR the effect SC, data","Different from SC, LRR considers the group effect among data points instead of sparsity.",11408
"offer (i.e., Benefiting LrrSPM) property, can from the better performance. a method proposed this","Benefiting from this property, the proposed method (i.e., LrrSPM) can offer a better performance.",11409
\emph{Entity naturally videos Discovery} in as The tracklet posed of task be can clustering.,The task of \emph{Entity Discovery} in videos can be naturally posed as tracklet clustering.,11410
major TC models Bayesian for are contributions at the Our tracklet-level. first nonparametric,Our major contributions are the first Bayesian nonparametric models for TC at tracklet-level.,11411
"functions rate. Given improve object the features, these are category-specific recognition learned combination to","Given these features, category-specific combination functions are learned to improve the object recognition rate.",11412
key dictionary representation. for the of a A is sparse success proper,A proper dictionary is a key for the success of sparse representation.,11413
sparse in dictionaries The caculated learned classification representation (SRC). are based,The learned dictionaries are caculated in sparse representation based classification (SRC).,11414
dictionary proposed accuracy reconstruction error the evaluate to learning and The used are classification method.,The classification accuracy and reconstruction error are used to evaluate the proposed dictionary learning method.,11415
q-value that outperform stratified thresholds substantially E-values. We show,We show that stratified q-value thresholds substantially outperform E-values.,11416
of The this is twofold. paper contribution,The contribution of this paper is twofold.,11417
"annealing solving single-stage the method we First, and problem. effective robust simulated propose for an","First, we propose an effective and robust single-stage simulated annealing method for solving the problem.",11418
"the and an apply design parameter we Secondly, tuning for statistically-principled methodology extensive procedure. and","Secondly, we design and apply an extensive and statistically-principled methodology for the parameter tuning procedure.",11419
illustrated and The trains. method proposed on experimental simulated spike is,The proposed method is illustrated on simulated and experimental spike trains.,11420
"we described from that the Moreover, domain task benefits adaptation. show","Moreover, we show that the described task benefits from domain adaptation.",11421
"applied Secondly, on graph layers to framework many reduction dimensionality based cascaded embedding GENet. are","Secondly, many cascaded dimensionality reduction layers based on graph embedding framework are applied to GENet.",11422
"classify used classifier is linear Finally, dimension-reduced features. a to SVM","Finally, a linear SVM classifier is used to classify dimension-reduced features.",11423
"for In and views. pointclouds unseen cross-view action from contrast, process recognition unknown directly we","In contrast, we directly process pointclouds for cross-view action recognition from unknown and unseen views.",11424
The show that Experimental methods. provide techniques over improvement significant state-of-the-art results our,The Experimental results show that our techniques provide significant improvement over state-of-the-art methods.,11425
novel localisation lip proposed face includes method. scheme approach The a a localisation and automatic,The proposed approach includes a novel automatic face localisation scheme and a lip localisation method.,11426
data large sets methods. other of proposed to on Simulations method performance the superior show,Simulations on large data sets show superior performance of the proposed method to other methods.,11427
"state-of-the-art the by is and LSEs of laborious tuning parameter application computation. expensive compromised However,","However, the application of state-of-the-art LSEs is compromised by laborious parameter tuning and expensive computation.",11428
"to contrast faster. LSEs are the existing In methods, proposed significantly","In contrast to existing methods, the proposed LSEs are significantly faster.",11429
"they parameters importantly, while better Most involve performance. achieving fewer much","Most importantly, they involve much fewer parameters while achieving better performance.",11430
(CNNs) Networks gained Neural detection. Convolutional have and image great object success classification Deep in,Deep Convolutional Neural Networks (CNNs) have gained great success in image classification and object detection.,11431
"studies on focus deserving issue. this However, fewer","However, fewer studies focus on this deserving issue.",11432
selected the to concepts in unfriendly These correspond category. image vectors each,These selected vectors correspond to the unfriendly concepts in each image category.,11433
"compressive to cascaded sensing from feature we network use the Firstly, data. learn","Firstly, we use cascaded compressive sensing network to learn feature from the data.",11434
"CSNet generates feature Secondly, by binary histograms. and hashing block-wise the","Secondly, CSNet generates the feature by binary hashing and block-wise histograms.",11435
"classify features. to is used linear classifier a Finally, these SVM","Finally, a linear SVM classifier is used to classify these features.",11436
together the thesis. topic projects this connects This presented in,This topic connects together the projects presented in this thesis.,11437
transcriptional on regulation. first predominantly chapter The focuses,The first chapter predominantly focuses on transcriptional regulation.,11438
by the major in high-throughput thesis is challenges The sequencing analysis. concluded listing,The thesis is concluded by listing the major challenges in high-throughput sequencing analysis.,11439
discuss I demanding the also questions top data. sequencing of current the integration,I also discuss the current top questions demanding the integration of sequencing data.,11440
in understanding image holistic provide insight We and deep this into a deblurring review.,We provide a holistic understanding and deep insight into image deblurring in this review.,11441
years. of  has lot the attention a drawn recent identification  Biometric-based in,  Biometric-based identification has drawn a lot of attention in the recent years.,11442
"Among all features. a rich biometrics, of is known palmprint to possess set","Among all biometrics, palmprint is known to possess a rich set of features.",11443
result accurate in introduced here identification. near-perfectly The a features,The features introduced here result in a near-perfectly accurate identification.,11444
of significantly recent large-scale image Image years the in with availability classification sets. advanced has,Image classification has advanced significantly in recent years with the availability of large-scale image sets.,11445
Attributes the Caltech-UCSD datasets. substantially establish a on and state-of-the-art Animals with Birds improved We,We establish a substantially improved state-of-the-art on the Animals with Attributes and Caltech-UCSD Birds datasets.,11446
"By combining results. improve different output embeddings, we further","By combining different output embeddings, we further improve results.",11447
cumulative measures interference. to the of used amount quantify We linkage disequilibrium,We used cumulative linkage disequilibrium measures to quantify the amount of interference.,11448
are by experimental These from data one predictions HIV-infected patient. supported,These predictions are supported by experimental data from one HIV-infected patient.,11449
"Thus, escapes interference explain why later could are slower.","Thus, interference could explain why later escapes are slower.",11450
experiments well with two Networks very literature. the We known from made Bayesian,We made experiments with two very well known Bayesian Networks from the literature.,11451
"most first we contribute problem. and review In work, comprehensive on this recent the this","In this work, we contribute the first comprehensive and most recent review on this problem.",11452
interesting by We also these results. to analyzing point some discoveries,We also point to some interesting discoveries by analyzing these results.,11453
In of spite population the the \emph{Ae. of dynamics,In spite of the population dynamics of the \emph{Ae.,11454
stability of was and analysis The equilibrium performed.,The analysis of equilibrium and stability was performed.,11455
follows of climate Cenozoic. evolution evolution the the biodiversity Their,Their biodiversity evolution follows the climate evolution of the Cenozoic.,11456
"Earth is the evolution understand. the of hard Cenozoic climate to considering When processes, only","When considering only Earth processes, the climate evolution of the Cenozoic is hard to understand.",11457
losses in other conditions. Adaptation and environment often gains fitness to in one results,Adaptation to one environment often results in fitness gains and losses in other conditions.,11458
"developments accessible. becoming more sequencing technologies, in novel data time-sampled and are With available","With novel developments in sequencing technologies, time-sampled data are becoming more available and accessible.",11459
"in these parallel to from datasets. been population Naturally, parameters infer there genetic efforts have","Naturally, there have been efforts in parallel to infer population genetic parameters from these datasets.",11460
number on sequences. We system of evaluate motion a monocular human the large,We evaluate the system on a large number of monocular human motion sequences.,11461
to Simulation results used illustrate our are findings.,Simulation results are used to illustrate our findings.,11462
control of absence in the ground points).,in the absence of ground control points).,11463
"monocular single we a depth from Next, image. estimate key body measurements","Next, we estimate key body measurements from a single monocular depth image.",11464
squares Levenberg-Marquardt least flexible is algorithm procedure problems. solve The iterative a to used non-linear,The Levenberg-Marquardt algorithm is a flexible iterative procedure used to solve non-linear least squares problems.,11465
show approach. and stability to efficiency of compositional the this simulated and data Applications real,Applications to real and simulated compositional data show the stability and efficiency of this approach.,11466
"effectiveness demonstrate comparisons and its \emph{state-of-the-art} through a real example. with simulations, We the data","We demonstrate its effectiveness through simulations, comparisons with the \emph{state-of-the-art} and a real data example.",11467
matrix are based features Textural gray image level extracted of fabric co-occurrence (GLCM). on,Textural features of fabric image are extracted based on gray level co-occurrence matrix (GLCM).,11468
on output real this The shown time monitor. of implementation a was,The real time output of this implementation was shown on a monitor.,11469
This dynamics. with a introduces state-space time-varying paper linear model,This paper introduces a linear state-space model with time-varying dynamics.,11470
Bayesian approximation. The posterior using is inference performed variational,The posterior inference is performed using variational Bayesian approximation.,11471
years from has In learning attention the machine significant aggregation received recent rank community.,In recent years rank aggregation has received significant attention from the machine learning community.,11472
study ranking. a known the collaborative as problem version We of,We study a version of the problem known as collaborative ranking.,11473
The connection a to the standard completion matrix very here interesting have results problem.,The results here have a very interesting connection to the standard matrix completion problem.,11474
"a pulse-based, transfer may is dynamically through Because routed neural circuit. information be","Because transfer is pulse-based, information may be dynamically routed through a neural circuit.",11475
"effective and describe protein temperatures. higher complementary respectively These at of aspects lower activity,","These describe complementary aspects of protein activity, at respectively higher and lower effective temperatures.",11476
"To series the of approximations analysis, CRLBs representations computation introduced. and and are facilitate","To facilitate computation and analysis, series representations and approximations of the CRLBs are introduced.",11477
previous a in conclusion the This result case. generalizes narrowband,This conclusion generalizes a previous result in the narrowband case.,11478
of the perceived color to subject be appears with The respect to illumination. different,The perceived color of the subject appears to be different with respect to illumination.,11479
color also photometric spaces. analyze different invariant We,We also analyze different photometric invariant color spaces.,11480
(ADMM). using We optimization direction of this problem alternating the multipliers solve method,We solve this optimization problem using the alternating direction method of multipliers (ADMM).,11481
"to require smooth we spectral the in process be sufficiently Instead, domain. the","Instead, we require the process to be sufficiently smooth in the spectral domain.",11482
related foraging have Animal where self-similar optimal to are trajectories movements been strategies central.,Animal movements have been related to optimal foraging strategies where self-similar trajectories are central.,11483
"approaches. also However, very similar tackled has since with been SLAM monocular","However, monocular SLAM has since also been tackled with very similar approaches.",11484
"limits also to and this scale is However, reconstructions. representation memory-intensive applicability small the","However, this representation is also memory-intensive and limits the applicability to small scale reconstructions.",11485
for have been limitation. explored avenues Several this overcoming,Several avenues have been explored for overcoming this limitation.,11486
"explicit expression require and attempts unavailable are often recover. difficult which to Previous components,","Previous attempts require explicit expression components, which are often unavailable and difficult to recover.",11487
than rate. terms positive of in SRC performs C-HiSLR true better even,C-HiSLR performs even better than SRC in terms of true positive rate.,11488
model an errors. methods fit a shape synthesis minimizing to image by AAM-based texture,AAM-based methods fit a shape model to an image by minimizing texture synthesis errors.,11489
a image directly facial mapping function Regression-based from points. appearance methods to feature facial learn,Regression-based methods directly learn a mapping function from facial image appearance to facial feature points.,11490
"starvation, the subtilis forms durable spores. To bacterium survive Bacillus","To survive starvation, the bacterium Bacillus subtilis forms durable spores.",11491
"in osmolarity line experiments in medium Indeed, prevents with engulfment predictions. the our decreasing","Indeed, decreasing the medium osmolarity in experiments prevents engulfment in line with our predictions.",11492
"threshold can clustering that is determine needed. empirically an Thus, algorithm","Thus, an algorithm that can empirically determine clustering threshold is needed.",11493
regions that of the clustering We accuracy. observe repeat presence affects,We observe that the presence of repeat regions affects clustering accuracy.,11494
does this things Gabora paper. two in,Gabora does two things in this paper.,11495
"Second, it whole views is central evolve. that the world that concept advances she","Second, she advances the central concept that it is whole world views that evolve.",11496
of is Each her argument convincing. completely section,Each section of her argument is completely convincing.,11497
"is survival. for rapid especially learning, critical Learning,","Learning, especially rapid learning, is critical for survival.",11498
"computer can algebra analyzed. availability of be With games packages, the these","With the availability of computer algebra packages, these games can be analyzed.",11499
"Recently, several to introduced have been plants. by approaches identify many researchers","Recently, many approaches have been introduced by several researchers to identify plants.",11500
"of and are applications Now, vein texture, practices. common features color shape,","Now, applications of texture, shape, color and vein features are common practices.",11501
"several had been Therefore, in experiments research. this conducted","Therefore, several experiments had been conducted in this research.",11502
"together objects. objects with In appear generally natural scenes, other","In natural scenes, objects generally appear together with other objects.",11503
"mixing be unharmful Finally, stimuli. for we satisfied derive of that conditions must simple","Finally, we derive simple conditions that must be satisfied for unharmful mixing of stimuli.",11504
is Ebola spreading virus rapidly Significance statement: in Africa.,Significance statement: Ebola virus is spreading rapidly in Africa.,11505
is here. and amino The connection acid between sequence mortality identified protein,The connection between protein amino acid sequence and mortality is identified here.,11506
resource intensive companies. of game The is animation a process production in,The production of animation is a resource intensive process in game companies.,11507
"been have to synthesize Therefore, animations techniques developed.","Therefore, techniques to synthesize animations have been developed.",11508
"artists. techniques adaptability offer procedural animation However, these by limited","However, these procedural techniques offer limited adaptability by animation artists.",11509
intelligence. in and remote plays indispensable image Aerial  artificial  an role categorization sensing,  Aerial image categorization plays an indispensable role in remote sensing and artificial intelligence.,11510
geometric and distribution the property The reflect aerial image. subgraphs of color an the both,The subgraphs reflect both the geometric property and the color distribution of an aerial image.,11511
"image, a are sub-RCGs extracted. the refined structures to corresponding aerial its new given Lastly,","Lastly, given a new aerial image, its sub-RCGs corresponding to the refined structures are extracted.",11512
They a SVM discriminative quantized for further classification. into are vector,They are further quantized into a discriminative vector for SVM classification.,11513
the validate results of the proposed experimental effectiveness method. Thorough,Thorough experimental results validate the effectiveness of the proposed method.,11514
algorithm function. stereo presents genetic evaluation paper a fuzzy with matching This,This paper presents a genetic stereo matching algorithm with fuzzy evaluation function.,11515
system. face step first Face detection in recognition is the,Face detection is the first step in face recognition system.,11516
"paper, will be In fixed spread value used. this","In this paper, fixed spread value will be used.",11517
skipped; be errors and need to detected Base-level structural (ii) be errors can corrected.,(ii) Base-level errors can be skipped; structural errors need to be detected and corrected.,11518
during the population. issue a walking health elderly are Falls major in,Falls during walking are a major health issue in the elderly population.,11519
"impaired falling dynamic They their explains which risk. often have increased balance,","They often have impaired dynamic balance, which explains their increased falling risk.",11520
"variable. performed predictor, were age regressions a as for Using each linear dependent","Using age as a predictor, linear regressions were performed for each dependent variable.",11521
while all Noise major communication. a of electronic through kinds transferring is issue images,Noise is a major issue while transferring images through all kinds of electronic communication.,11522
suppressing can these while details image noise. impulsive All methods preserve primarily,All these methods can primarily preserve image details while suppressing impulsive noise.,11523
minimum Distributed of with communication data from increasingly models multiple learning repositories is probabilistic important.,Distributed learning of probabilistic models from multiple data repositories with minimum communication is increasingly important.,11524
interaction of among agents many years. investigation subject been extensive Communication the have and since,Communication and interaction among agents have been the subject of extensive investigation since many years.,11525
"simple inter-agent beyond commitments, forms of event-based provide, DACMASs Notably, communication.","Notably, DACMASs provide, beyond commitments, simple forms of inter-agent event-based communication.",11526
approaches. properties the that DACmMCMASs show original retain nice of the We formal,We show that DACmMCMASs retain the nice formal properties of the original approaches.,11527
One the surface that remains area constant. is total bole,One is that the total bole surface area remains constant.,11528
"in not However, presence is of higher its established. neocortex mammals the well","However, its presence in the neocortex of higher mammals is not well established.",11529
"Remarkably, for many these scales. are temporal seen different fluctuations correlated","Remarkably, these correlated fluctuations are seen for many different temporal scales.",11530
"this LSEPSO,has paper, algorithm,called optimization (MMO) been an In improved proposed. multimodal","In this paper, an improved multimodal optimization (MMO) algorithm,called LSEPSO,has been proposed.",11531
local and global ability shown of improve the finding algorithm. It to optima has been,It has been shown to improve global and local optima finding ability of the algorithm.,11532
method its particle attenuation prevented a following This particle's and specific neighbours. by,This method prevented particle's attenuation and following a specific particle by its neighbours.,11533
"address needed curricular learning additional shift to outcomes. Consequently, is a","Consequently, a curricular shift is needed to address additional learning outcomes.",11534
studies seven from We provide institutions. case,We provide case studies from seven institutions.,11535
approaches to science innovations needs. teaching varied curricular address to demonstrate new These data,These varied approaches to teaching data science demonstrate curricular innovations to address new needs.,11536
of solved problem then learned the optimization tangent based subspaces. set is on The,The optimization problem is then solved based on the learned set of tangent subspaces.,11537
"as this follows. In paper, recent our work we build on","In this paper, we build on our recent work as follows.",11538
"to of Second, subset optimal we min-max tangent algorithm an a select subspaces. develop","Second, we develop a min-max algorithm to select an optimal subset of tangent subspaces.",11539
"validity clustering. we Grassmann-distance of analytically the prove based Finally, the proposed","Finally, we analytically prove the validity of the proposed Grassmann-distance based clustering.",11540
networks. alignment be Network between transfer knowledge to used regions conserved can functional different of,Network alignment can be used to transfer functional knowledge between conserved regions of different networks.,11541
"of after but alignments measures edges, conserved produced. Typically, amount one are only the","Typically, one measures the amount of conserved edges, but only after alignments are produced.",11542
a the hyperspectral Laplacian graph regularization This paper introduces unmixing in formulation.,This paper introduces a graph Laplacian regularization in the hyperspectral unmixing formulation.,11543
of the hyperspectral graph construction relies the proposed The regularization representation upon image. of a,The proposed regularization relies upon the construction of a graph representation of the hyperspectral image.,11544
the The Alternating using problem Method is resulting (ADMM). Direction solved optimization of convex Multipliers,The resulting convex optimization problem is solved using the Alternating Direction Method of Multipliers (ADMM).,11545
proposed and adopted to Both verify the simulated algorithm. datasets of performance the are real,Both simulated and real datasets are adopted to verify the performance of the proposed algorithm.,11546
susceptible the interactions and these dynamic The infected countries. the of populations model describes of,The model describes the dynamic interactions of the susceptible and infected populations of these countries.,11547
also level size investigated. application line font at text an Finally to is detect,Finally an application to detect font size at text line level is also investigated.,11548
proposed experimentally the documents. set validated of All compressed data with are algorithms sufficient,All the proposed algorithms are validated experimentally with sufficient data set of compressed documents.,11549
"this framework propose very a for task. FAMILY, We general","We propose FAMILY, a very general framework for this task.",11550
these use CAPTCHAs. To prevent registrations kinds false many websites of,To prevent these kinds of false registrations many websites use CAPTCHAs.,11551
kinds Among OCR-Based are visual all CAPTCHAs very of common. CAPTCHAs or,Among all kinds of CAPTCHAs OCR-Based or visual CAPTCHAs are very common.,11552
is containing sequence an visual of Actually image characters. CAPTCHA a,Actually visual CAPTCHA is an image containing a sequence of characters.,11553
species null community which models between models provide Neutral assume assembly. ecological equivalence for,Neutral models which assume ecological equivalence between species provide null models for community assembly.,11554
Bayesian multi-site strategy we Using UNTB. for efficient fitting an approximation the developed this,Using this approximation we developed an efficient Bayesian fitting strategy for the multi-site UNTB.,11555
not higher was true This at ranks. taxonomic,This was not true at higher taxonomic ranks.,11556
human is of communication. one Gesturing of natural modes the,Gesturing is one of the natural modes of human communication.,11557
"a by al. Earlier we of a published critique scathing year, this Mendez paper et","Earlier this year, we published a scathing critique of a paper by Mendez et al.",11558
"sarcastic inflamed."" for censored our being was Unfortunately, ""too reply and","Unfortunately, our reply was censored for being ""too sarcastic and inflamed.""",11559
"removed, the a was and dedication were castrated, Acknowledgments meanings were deleted. References in","References were removed, meanings were castrated, and a dedication in the Acknowledgments was deleted.",11560
compared context. approaches We satisfying to state-of-the-art in results when obtain the standard DTW,We obtain satisfying results when compared to standard state-of-the-art approaches in the DTW context.,11561
"is simplification Fisher vector so propose we Sparse A FV attractive, (SFV). of","A simplification of FV is attractive, so we propose Sparse Fisher vector (SFV).",11562
"how of SFV features. in representation consistence preserves demonstrate the In we similar addition, local","In addition, we demonstrate how SFV preserves the consistence in representation of similar local features.",11563
attributed distinct circuits. phenomena have previously The diverse to neural been,The diverse phenomena have been previously attributed to distinct neural circuits.,11564
for and real-time a tracking. ball approach paper This presents single-colored detection,This paper presents a real-time approach for single-colored ball detection and tracking.,11565
The two approach of main phases. consists,The approach consists of two main phases.,11566
"meaning models psychology, In a tradition. have geometric of long","In psychology, geometric models of meaning have a long tradition.",11567
cate- types for can of which algorithms. gories distinguish six We used database different be,We distinguish six database cate- gories which can be used for different types of algorithms.,11568
cursive Arabic consonantal is writing. handwriting and a,Arabic handwriting is a consonantal and cursive writing.,11569
its morphology syntax rich Arabic which large. makes and is lexicon in,Arabic is rich in morphology and syntax which makes its lexicon large.,11570
for new preprocessing technique the A strokes introduced. handling is delayed,A new preprocessing technique for handling the delayed strokes is introduced.,11571
different is system small two databases large covering The lexicons. performance using evaluated and,The system performance is evaluated using two different databases covering small and large lexicons.,11572
small database. the Our outperforms system state-of-art the systems for lexicon,Our system outperforms the state-of-art systems for the small lexicon database.,11573
is step Translation a crucial expression. gene in,Translation is a crucial step in gene expression.,11574
most of consume known is Translation cell's to energy. the,Translation is known to consume most of the cell's energy.,11575
bounded constraint An is these on cellular rates affine model the used to budget.,An affine constraint on these rates is used to model the bounded cellular budget.,11576
decrease ribosomal chain. The rates densities optimal to along monotonically the the corresponding,The ribosomal densities corresponding to the optimal rates monotonically decrease along the chain.,11577
We results. some discuss implications of of biological these the,We discuss some of the biological implications of these results.,11578
performed. task to be more if complex The becomes comparisons are,The task becomes more complex if comparisons are to be performed.,11579
to mining undertaken problems implementing by overcome is verification study physical data technique. the This,This study is undertaken to overcome the physical verification problems by implementing data mining technique.,11580
frame of major correlate the GRI. with that work the conceptual findings is The accomplishment,The major accomplishment is that the findings correlate with the conceptual frame work of GRI.,11581
This training strategy distinct into partitions stages. two,This strategy partitions training into two distinct stages.,11582
output of is a test image. The semantic labeling the reconstruction,The output reconstruction is a semantic labeling of the test image.,11583
deep research avenues sparse The of approach for representations. our opens on effectiveness new,The effectiveness of our approach opens new avenues for research on deep sparse representations.,11584
a in manner. Our representation novel this utilize classifiers,Our classifiers utilize this representation in a novel manner.,11585
We cell phenotypes. dynamics  different  consider n the population with,  We consider the cell population dynamics with n different phenotypes.,11586
through or phenotype Cells conversions one cells asymmetric in produce in can other divisions. phenotypes,Cells in one phenotype can produce cells in other phenotypes through conversions or asymmetric divisions.,11587
We under give which a necessary and assumption condition valid. is sufficient this,We give a sufficient and necessary condition under which this assumption is valid.,11588
"also ""phenotypic assumption. without prove the equilibrium"" such We","We also prove the ""phenotypic equilibrium"" without such assumption.",11589
We non-Markov to extend also our cases. results,We also extend our results to non-Markov cases.,11590
feasibility datasets available publicly of the the results Experimental approach. proposed demonstrate on,Experimental results on publicly available datasets demonstrate the feasibility of the proposed approach.,11591
multgee ordinal clinical through demonstrate multinomial with We of trial the clustered application responses. a,We demonstrate the application of multgee through a clinical trial with clustered ordinal multinomial responses.,11592
for suitable learn building We representations. to local word image features mid-level propose,We propose to learn local mid-level features suitable for building word image representations.,11593
mid-level word produce to global then aggregated features signature. local a image can be Our,Our local mid-level features can then be aggregated to produce a global word image signature.,11594
Errors are implicative in from coming binary theories data studied.,Errors in implicative theories coming from binary data are studied.,11595
"that theories two errors First, may classes singled implicative are of affect out.","First, two classes of errors that may affect implicative theories are singled out.",11596
The Duquenne-Guigues) first implication (canonical the cardinality or approach base. minimal uses,The first approach uses the cardinality minimal (canonical or Duquenne-Guigues) implication base.,11597
is computationally such construction of a The base intractable.,The construction of such a base is computationally intractable.,11598
"about approaches certain implications. on are based questions of the interactive, Both validity","Both approaches are interactive, based on questions about the validity of certain implications.",11599
are and discussed. experiments Results computer presented of,Results of computer experiments are presented and discussed.,11600
Index distribution group size; collective Terms-swarm; hyper-geometric decision . making; noise;,Index Terms-swarm; collective decision making; noise; group size; hyper-geometric distribution .,11601
"are mechanisms neither However, nor occurs, the molecular involved to known. this the which extent","However, neither the extent to which this occurs, nor the molecular mechanisms involved are known.",11602
foraging a this problem formalize problem. monitoring We as,We formalize this monitoring problem as a foraging problem.,11603
summarization. exploited The is for to approach multi-stream be suitable proposed here video,The approach proposed here is suitable to be exploited for multi-stream video summarization.,11604
"e.g. a sophisticated as step video Meanwhile, it serve for preliminary more can surveillance,","Meanwhile, it can serve as a preliminary step for more sophisticated video surveillance, e.g.",11605
misfit to the based approach on The function. inverse problems classical a optimization is of,The classical approach to inverse problems is based on the optimization of a misfit function.,11606
"In paper, spike this new binless are trains for proposed. similarity measures two","In this paper, two new binless similarity measures for spike trains are proposed.",11607
kinds one possible We two these approach representations. explore to combining of,We explore one possible approach to combining these two kinds of representations.,11608
We a theory/semantics of logic first model present based vectors reals. order on for,We present a model theory/semantics for first order logic based on vectors of reals.,11609
a Identity-by-descent concept genetics (IBD) many with in is fundamental applications.,Identity-by-descent (IBD) is a fundamental concept in genetics with many applications.,11610
"the affect inferred This length segment so biases inferences. IBD distribution, and can downstream","This biases the inferred IBD segment length distribution, and so can affect downstream inferences.",11611
detail effect its conflation make the will in tractable. correction methods Understanding future in more,Understanding the conflation effect in detail will make its correction in future methods more tractable.,11612
able generalize unseen to to dense is Our datasets. also model,Our dense model is also able to generalize to unseen datasets.,11613
"a implementation of named more Correspondingly, adaptive MSRS, presented. is ""SO-SA"",","Correspondingly, a more adaptive implementation of MSRS, named ""SO-SA"", is presented.",11614
"map as formulate we this regression saliency problem. a computation In paper,","In this paper, we formulate saliency map computation as a regression problem.",11615
map. the produce across to finally multiple fused Saliency are scores levels saliency,Saliency scores across multiple levels are finally fused to produce the saliency map.,11616
propose detection. discriminate integration is we object salient approach a feature One that regional for,One is that we propose a discriminate regional feature integration approach for salient object detection.,11617
"fast that our algorithms. most as we method Meanwhile, existing demonstrate as runs","Meanwhile, we demonstrate that our method runs as fast as most existing algorithms.",11618
that our a certain sense in strategy optimal. pruning We theoretically is show,We show that our pruning strategy is in a certain sense theoretically optimal.,11619
OBDD augments with a This called generalization paper conjunctive to decomposition OBDD[$\wedge$]. propose,This paper augments OBDD with conjunctive decomposition to propose a generalization called OBDD[$\wedge$].,11620
of $i$ strictly ROBDD[$\wedge_{\widehat{i}}$] show when increasing increases. We is that succinctness the,We show that the succinctness of ROBDD[$\wedge_{\widehat{i}}$] is strictly increasing when $i$ increases.,11621
"for is utilized problem. as in a feature CPD method unsupervised an extraction Nowadays, classification","Nowadays, CPD is utilized as an unsupervised method for feature extraction in a classification problem.",11622
"a is machine, that, vector After required features. as such support classify those classifier, to","After that, a classifier, such as support vector machine, is required to classify those features.",11623
"classification achieved this task isolated two In steps. in is manner,","In this manner, classification task is achieved in two isolated steps.",11624
"with these To propose visual difficulties, model. co-occurrence a deal we novel word","To deal with these difficulties, we propose a novel visual word co-occurrence model.",11625
are Linear then SVMs classifiers descriptors. these trained co-occurrence using as,Linear SVMs are then trained as classifiers using these co-occurrence descriptors.,11626
Region object search is localization. for widely used,Region search is widely used for object localization.,11627
The parameter flexibly. a be controlled density by can region,The region density can be controlled by a parameter flexibly.,11628
"can the efficiently belief by optimized LS-MWCS Moreover, propagation. be","Moreover, the LS-MWCS can be efficiently optimized by belief propagation.",11629
"a in and Hypergraph problems. is several machine powerful representation computer learning pattern recognition vision,","Hypergraph is a powerful representation in several computer vision, machine learning and pattern recognition problems.",11630
"researchers keen the to models. last develop many been In hypergraph different decade, have","In the last decade, many researchers have been keen to develop different hypergraph models.",11631
"of much hyperedge no design attention to weights. been the paid has In contrast,","In contrast, no much attention has been paid to the design of hyperedge weights.",11632
hypegraphs. applies also that argue to this We,We argue that this also applies to hypegraphs.,11633
algorithm a is path Conflict-Based finding multi-agent problem. Meta-Agent Search~(MA-CBS) the The for proposed recently,The Meta-Agent Conflict-Based Search~(MA-CBS) is a recently proposed algorithm for the multi-agent path finding problem.,11634
number a a key in of applications. is Optimization task,Optimization is a key task in a number of applications.,11635
"article, focus propositional this we In constraints. on","In this article, we focus on propositional constraints.",11636
video. for in are techniques Background detection modeling used moving object,Background modeling techniques are used for moving object detection in video.,11637
with of in the algorithms detection Many object exist purposes. different field,Many algorithms exist in the field of object detection with different purposes.,11638
"based paper, moving improvement this of an we on segmentation. propose In object detection codebook","In this paper, we propose an improvement of moving object detection based on codebook segmentation.",11639
codebook with detection the We algorithm. edge original algorithm an associate,We associate the original codebook algorithm with an edge detection algorithm.,11640
"of In metrics evaluation each case, detection. the use we the frame-based for","In each case, we use frame-based metrics for the evaluation of the detection.",11641
The are presented results different analyzed. and,The different results are presented and analyzed.,11642
subjects findings demonstrate human increase These that acquiring their skill. horizon planning motor when a,These findings demonstrate that human subjects increase their planning horizon when acquiring a motor skill.,11643
two an exists imino as one. tautomers solution which in is of Deoxycitidine one uncanonical,Deoxycitidine in solution exists as two tautomers one of which is an uncanonical imino one.,11644
able pair. base a potentially hoosteen tautomer is imino The form to GC,The imino tautomer potentially is able to form a hoosteen GC base pair.,11645
"durable. reproduction Otherwise, resistance quantitative was only that depressed pathogen more","Otherwise, quantitative resistance that depressed only pathogen reproduction was more durable.",11646
"The performed an which lieu of is $s(x)$, sensitivity analysis is on $f(x)$.","The sensitivity analysis is performed on $s(x)$, which is an lieu of $f(x)$.",11647
"that consequences. very show has we Moreover, probabilistic underlying interpretation interesting the","Moreover, we show that the underlying probabilistic interpretation has very interesting consequences.",11648
of for of procedure dynamics. the The results network's summary and are the interpretation useful,The results of the procedure are useful for interpretation and summary of the network's dynamics.,11649
"such some directions, propose We as unsupervised-learning. future","We propose some future directions, such as unsupervised-learning.",11650
software available download. (ILPEACE) The is freely for,The software (ILPEACE) is freely available for download.,11651
simulation method studies. then performance investigated proposed of using is our Numerical aggregation,Numerical performance of our proposed aggregation method is then investigated using simulation studies.,11652
disease. require wearable wireless and of automatic systems detection Such cardiovascular ECG,Such wearable and wireless ECG systems require automatic detection of cardiovascular disease.,11653
"few have detection fully developed. (MI) been well infarction However, myocardial automatic algorithms disease","However, few fully automatic myocardial infarction (MI) disease detection algorithms have well been developed.",11654
method to The diagnostic within ECG Physionet. database the was PTB proposed applied,The proposed method was applied to the PTB diagnostic ECG database within Physionet.,11655
signals bispectrum is We for introduce nonzero. which in such fractional bispectrum fractional the,We introduce fractional bispectrum in which for such signals the fractional bispectrum is nonzero.,11656
controlled propose directional We locally a domain edge-preserving with kernel filter for smoothing. bilateral a,We propose a bilateral filter with a locally controlled domain kernel for directional edge-preserving smoothing.,11657
kernel oriented bilateral combined forms kernel filter. with the domain a range directional The,The oriented domain kernel combined with a range kernel forms the directional bilateral filter.,11658
Systems Recognition based ocular- reasons. primarily devices are used security Iris for biometric,Iris Recognition Systems are ocular- based biometric devices used primarily for security reasons.,11659
are The commonly with recognition used literature classifiers. directional using features several compared results in,The recognition results are compared with commonly used directional features in literature using several classifiers.,11660
our sophisticated that benchmark approach datasets. new baselines outperforms We on demonstrate,We demonstrate that our new approach outperforms sophisticated baselines on benchmark datasets.,11661
"Tabu in out In this carried approach, parallel. searches two are","In this approach, two Tabu searches are carried out in parallel.",11662
our goal research twofold. of is The,The goal of our research is twofold.,11663
"wish clear we definitions. to to First, way competing a provide compare","First, we wish to provide a clear way to compare competing definitions.",11664
or data stochastic data. temporal Realizations are observed process functional often of,Realizations of stochastic process are often observed temporal data or functional data.,11665
interests data. or functional dynamic of There in classification growing are,There are growing interests in classification of dynamic or functional data.,11666
ODE in We the classifiers. as the parameters use for features,We use the parameters in the ODE as features for classifiers.,11667
"smoothing for paper. estimator, revisited namely this in is A piecewise-constant M-smoother, robust","A robust estimator, namely M-smoother, for piecewise-constant smoothing is revisited in this paper.",11668
the depth Experiments show denoising effectiveness framework. on map our of,Experiments on depth map denoising show the effectiveness of our framework.,11669
rules dependencies express between data. fuzzy object-attribute The incidence in attributes,The rules express dependencies between fuzzy attributes in object-attribute incidence data.,11670
interactions of systems central a Discovering biology. protein-protein problem patterns in in networks (PPIs) is,Discovering patterns in networks of protein-protein interactions (PPIs) is a central problem in systems biology.,11671
does Matrix using this It by Non-negative (NMTF). Tri-Factorization,It does this by using Non-negative Matrix Tri-Factorization (NMTF).,11672
protein the and demonstrates patterns complementarity sequence networks. wiring of in This their PPI,This demonstrates complementarity of protein sequence and their wiring patterns in the PPI networks.,11673
"FUSE than network is aligners. multiple efficient other Also, more computationally","Also, FUSE is more computationally efficient than other multiple network aligners.",11674
Extensive that it outperforms show algorithms. experimental subspace state-of-the-art results clustering,Extensive experimental results show that it outperforms state-of-the-art subspace clustering algorithms.,11675
It does LASSO is grouping effect. known not already have the,It is already known LASSO does not have the grouping effect.,11676
West outbreak virus of is Ebola experiencing a currently severe disease (EVD). Africa,West Africa is currently experiencing a severe outbreak of Ebola virus disease (EVD).,11677
strength this a major Borrowing groups in the across challenge is context.,Borrowing strength across the groups is a major challenge in this context.,11678
"Inference Bayesian is performed via a novel (ABC), functional regression approximate computation adjustment. through","Inference is performed through approximate Bayesian computation (ABC), via a novel functional regression adjustment.",11679
employed and tasks. data driven for Visual object have classification successfully recognition been various dictionaries,Visual data driven dictionaries have been successfully employed for various object recognition and classification tasks.,11680
also among of domain-shift domains. the We pair minimize from the each data,We also minimize the domain-shift among the data from each pair of domains.,11681
"common a Simultaneously, we adaptive dictionary. learn","Simultaneously, we learn a common adaptive dictionary.",11682
"have to used public interventions. models recent health epidemic years, guide In been","In recent years, epidemic models have been used to guide public health interventions.",11683
outcome depends that epidemic hospital capacity Results and on behavior. individual both suggest,Results suggest that epidemic outcome depends on both hospital capacity and individual behavior.,11684
strategy using significant A is (RMS). recursive a achieved search minimum speedup,A significant speedup is achieved using a recursive minimum search strategy (RMS).,11685
is of algorithm method. Then the RMS essential an EDP the part,Then the RMS algorithm is an essential part of the EDP method.,11686
that First existence of necessary for parameters assumptions are about clarify terraces. the we model,First we clarify assumptions about model parameters that are necessary for the existence of terraces.,11687
"discuss Finally, of some strategies problems. we these for of remediation","Finally, we discuss strategies for remediation of some of these problems.",11688
that optimal. show We this minimax is estimator also,We also show that this estimator is minimax optimal.,11689
consider extinction given a on random We of processes with structure. networks the problem,We consider the problem of extinction processes on random networks with a given structure.,11690
"the probable use theory to In find path to extinction. leading the particular, we most","In particular, we use the theory to find the most probable path leading to extinction.",11691
"differential Genes often modelled connected in equations. ordinary regulatory by networks, are","Genes are connected in regulatory networks, often modelled by ordinary differential equations.",11692
of Changes the in along propagate genes in a expression network. to paths other gene,Changes in expression of a gene propagate to other genes along paths in the network.,11693
"matrix stable the confers about network Jacobian a system's At information connectivity. state,","At a stable state, the system's Jacobian matrix confers information about network connectivity.",11694
"disclose is properties knowledge of essential. the functional network connections of To genes,","To disclose the functional properties of genes, knowledge of network connections is essential.",11695
proposes choices paper evaluates extract This alternative and densely. to patches,This paper proposes and evaluates alternative choices to extract patches densely.,11696
fine-grain and approaches benchmarks. classification image different recent on are evaluated The retrieval,The different approaches are evaluated on recent image retrieval and fine-grain classification benchmarks.,11697
  Kinetics. review Chemical of Book Stochastic,  Book review of Stochastic Chemical Kinetics.,11698
"Erdi, (Mostly) P. G. Applications, and Biological Lente. Systems Theory","Theory and (Mostly) Systems Biological Applications, P. Erdi, G. Lente.",11699
historical of theorisations of a occurrences We distil creativity. serendipitous and unified from description serendipity,We distil a unified description of serendipitous occurrences from historical theorisations of serendipity and creativity.,11700
modelling a framework models. mixture cluster based Finite Gaussian on provides for analysis mixture parsimonious,Finite mixture modelling provides a framework for cluster analysis based on parsimonious Gaussian mixture models.,11701
This model-based clustering. clustvarsel which for paper R subset selection package describes performs the,This paper describes the R package clustvarsel which performs subset selection for model-based clustering.,11702
through the of presented package is discussion of examples. several Usage data the,Usage of the package is presented through the discussion of several data examples.,11703
stop to Which we genes target should cancer?,Which genes should we target to stop cancer?,11704
"in MNL learning with mixture two even infeasible However, general. is components","However, even learning mixture with two MNL components is infeasible in general.",11705
"The high proposed also for resolution, cluttered method almost images. linearly scales","The proposed method also scales almost linearly for high resolution, cluttered images.",11706
etc. applications Its range to ophthalmology dermatology from to coronary imaging,Its applications range from ophthalmology to dermatology to coronary imaging etc.,11707
"Here, and on retinal ophthalmology is OCT in the focus applications of imaging.","Here, the focus is on applications of OCT in ophthalmology and retinal imaging.",11708
Therefore use importance. different techniques of are high of the segmentation,Therefore the use of different segmentation techniques are of high importance.,11709
"of image helpful. use can be registration Thus, techniques very","Thus, use of image registration techniques can be very helpful.",11710
"will overview techniques In work, this of be given. analysis image such an","In this work, an overview of such image analysis techniques will be given.",11711
due hyperspectral contain low pixels resolution images spatial sensors. to Hyperspectral of mixed,Hyperspectral images contain mixed pixels due to low spatial resolution of hyperspectral sensors.,11712
called more Mixed endmembers. containing material than one pixels pixels distinct are,Mixed pixels are pixels containing more than one distinct material called endmembers.,11713
in presence mixed pixels The called are percentages endmembers abundance fractions. of,The presence percentages of endmembers in mixed pixels are called abundance fractions.,11714
it space. in representing low while dimensional structure preserves data This method of the geometrical,This method preserves the geometrical structure of data while representing it in low dimensional space.,11715
real algorithm applied and datasets. Proposed synthetic is on,Proposed algorithm is applied on synthetic and real datasets.,11716
endmembers generated library. from is Synthetic data spectral based USGS on,Synthetic data is generated based on endmembers from USGS spectral library.,11717
dataset dataset for AVIRIS evaluation used is Cuprite proposed method. of as real,AVIRIS Cuprite dataset is used as real dataset for evaluation of proposed method.,11718
distance spectral angle (SAD) measures. and Results are distance quantified (AAD) angle abundance based on,Results are quantified based on spectral angle distance (SAD) and abundance angle distance (AAD) measures.,11719
all astronomical Here statements these factors. show that are by wrong I,Here I show that all these statements are wrong by astronomical factors.,11720
The logic. based mathematical are errors extravagant paper's claims of on,The paper's extravagant claims are based on errors of mathematical logic.,11721
has. Further establishing how importance odor of analysis space highlights the many the perceptual dimensions,Further analysis highlights the importance of establishing how many dimensions the perceptual odor space has.,11722
and answer. experimental on I towards an arguments topic propose avenues some the review,I review some arguments on the topic and propose experimental avenues towards an answer.,11723
High repeats. depth to used be read short can sequence assemble,High read depth can be used to assemble short sequence repeats.,11724
repetitive fail in existing genome read. of longer regions than average The assemblers,The existing genome assemblers fail in repetitive regions of longer than average read.,11725
first the variational for Bayesian fully present We Gaussian-process-modulated scheme Poisson inference processes. continuous,We present the first fully variational Bayesian inference scheme for continuous Gaussian-process-modulated Poisson processes.,11726
"number limit of consider windows. features, the we non overlapping the To only","To limit the number of features, we consider only the non overlapping windows.",11727
"work and shown methods. to The are derived LBP fairly well fisherface features eigenface, over","The derived features are shown to work fairly well over eigenface, fisherface and LBP methods.",11728
change in of Classifications organismal insights. inferences phylogenetic new and of groups light,Classifications and phylogenetic inferences of organismal groups change in light of new insights.,11729
knowledge diverse in semantic from ontological Developments have technologies encoding promoted web domains. of,Developments in semantic web technologies have promoted ontological encoding of knowledge from diverse domains.,11730
circumscription. networks extend the with We DL SROIQ and grounded constraint,We extend the DL SROIQ with constraint networks and grounded circumscription.,11731
introducing expressive on constraint approaches modelling paper existing restrictions overcomes by constructs. This,This paper overcomes restrictions on existing constraint modelling approaches by introducing expressive constructs.,11732
minimization Grounded is circumscription role DL. allows for decidable and and concept,Grounded circumscription allows concept and role minimization and is decidable for DL.,11733
"a decision present for GC-SROIQ(C), tableau the logic: resulting and procedure We describe it.","We present the resulting logic: GC-SROIQ(C), and describe a tableau decision procedure for it.",11734
to methods for nonlinear We estimator state learning an use construct statistical systems. adaptive stochastic,We use statistical learning methods to construct an adaptive state estimator for nonlinear stochastic systems.,11735
"often into search satisficing several have with Recently, found that problems. researchers cost-based runs A*","Recently, several researchers have found that cost-based satisficing search with A* often runs into problems.",11736
"in `mind-reading' mental are complex. processes involved However, the","However, the mental processes involved in `mind-reading' are complex.",11737
results we with the preliminary proposed some architecture present Together promising .,Together with the proposed architecture we present some promising preliminary results .,11738
"kinds apes, the all Among as have human. slim of body gibbons only","Among all kinds of apes, only gibbons have the slim body as human.",11739
the of bipedalism. to account Such locomotion mode can transition for reasonably,Such mode of locomotion can account reasonably for the transition to bipedalism.,11740
"feet secondly, ancestor the evolved arche. longitudinal of with our it","secondly, it evolved the feet of our ancestor with longitudinal arche.",11741
an procedure We detection of two enhanced consists detail further that steps.,We further detail an enhanced detection procedure that consists of two steps.,11742
"further feature clinical extraction based As introduce novel we approach observations. a a improvement, on","As a further improvement, we introduce a novel feature extraction approach based on clinical observations.",11743
the as for detailed detectors These input to regions serve will analysis. lesion specific,These regions will serve as input to the specific lesion detectors for detailed analysis.,11744
system. of increase This performance can computational procedure the a screening,This procedure can increase the computational performance of a screening system.,11745
proposes sets matching paper method. robust point a This,This paper proposes a robust point sets matching method.,11746
robust algorithm iterative present an noise is to case. that We,We present an iterative algorithm that is robust to noise case.,11747
"all Firstly, transformations between calculate two points. we","Firstly, we calculate all transformations between two points.",11748
that the are computed transformation two measure Then matrix to possibility are true. both similarity,Then similarity matrix are computed to measure the possibility that two transformation are both true.,11749
matrix. the matrix score iteratively the matching by update similarity using We,We iteratively update the matching score matrix by using the similarity matrix.,11750
"we graph, obtain using matching By the algorithm on matching result.","By using matching algorithm on graph, we obtain the matching result.",11751
approach and results Experimental by to show obtained jitter. robustness outlier our,Experimental results obtained by our approach show robustness to outlier and jitter.,11752
"nonparametric The model. process mixture Dirichlet is (DPM) a flexible ubiquitous, Bayesian statistical","The Dirichlet process mixture (DPM) is a ubiquitous, flexible Bayesian nonparametric statistical model.",11753
"rigorous (MAP) statistically a-posteriori inference approximate for develop algorithms yet we DPMs. simplified Here, maximum","Here, we develop simplified yet statistically rigorous approximate maximum a-posteriori (MAP) inference algorithms for DPMs.",11754
DPM the This well-posed a is solution model. to probabilistic of the MAP approximation,This is a well-posed approximation to the MAP solution of the probabilistic DPM model.,11755
"visualization, functionality of and introduce We series. monitoring surveillance for the the modelling time","We introduce the functionality for the visualization, modelling and monitoring of surveillance time series.",11756
"we series time look modelling. Furthermore, and at categorical Dirichlet-Multinomial beta-binomial overdispersion address using or","Furthermore, we look at categorical time series and address overdispersion using beta-binomial or Dirichlet-Multinomial modelling.",11757
"optical has character problem recognition, of literature. the in OCR, been The widely discussed","The problem of optical character recognition, OCR, has been widely discussed in the literature.",11758
"aims hand-written text. at program text, a the Having the recognizing","Having a hand-written text, the program aims at recognizing the text.",11759
"issue, is are open there it to approaches still this several an though problem. Even","Even though there are several approaches to this issue, it is still an open problem.",11760
short. run is training The very time also and,The training and run time is also very short.,11761
of key clutter systems. visual for ability objects Searching amongst is a,Searching for objects amongst clutter is a key ability of visual systems.,11762
depend appearance the scene and target trade-off on How complexity? does,How does the trade-off depend on target appearance and scene complexity?,11763
"procedure the when is this cluttered. However, computationally too scene becomes visual intensive","However, this procedure is computationally intensive when the visual scene becomes too cluttered.",11764
"scales Therefore, of also that conjecture to large optimal an clutters. number strategy we","Therefore, we also conjecture an optimal strategy that scales to large number of clutters.",11765
"We orthant computation the probabilities, Gaussian study of i.e.","We study the computation of Gaussian orthant probabilities, i.e.",11766
probability inside a falls a Gaussian quadrant. that the,the probability that a Gaussian falls inside a quadrant.,11767
propose we (PF). filter a using As improvement an particle,As an improvement we propose using a particle filter (PF).,11768
empirically We this improvements currently show can on to algorithms. lead drastic that used,We show empirically that this can lead to drastic improvements on currently used algorithms.,11769
"orthants We to mixture framework etc. (Student, of also of extend Gaussians the Cauchy","We also extend the framework to orthants of mixture of Gaussians (Student, Cauchy etc.",11770
"and truncated Gaussians. of to ), the simulation","), and to the simulation of truncated Gaussians.",11771
systems. physical noise ubiquitous are biological and in Oscillations and,Oscillations and noise are ubiquitous in physical and biological systems.,11772
"of the the phase well no defined. presence In asymptotic longer noise, is","In the presence of noise, the asymptotic phase is no longer well defined.",11773
classical in noise. limit phase asymptotic the to the vanishing of It reduces,It reduces to the classical asymptotic phase in the limit of vanishing noise.,11774
Simulating brain a the challenge. is entire daunting,Simulating the entire brain is a daunting challenge.,11775
tolerance determining factor plants. the major is Ion in of salinity transport,Ion transport is the major determining factor of salinity tolerance in plants.,11776
knockout salinity overexpress described. ion attempts tolerance changing are to transporters Several or for,Several attempts to overexpress or knockout ion transporters for changing salinity tolerance are described.,11777
method with classification a Gaussian a is appealing process of number properties. popular,Gaussian process classification is a popular method with a number of appealing properties.,11778
"presented. in for this composition a is locomotion footprint-driven Therefore, flexible paper approach","Therefore, in this paper a flexible approach for footprint-driven locomotion composition is presented.",11779
is A for learning MLDANet algorithm described.,A learning algorithm for MLDANet is described.,11780
"been in exchangeable methods have However, settings. such independent data or largely studied","However, such methods have largely been studied in independent or exchangeable data settings.",11781
image an of information. are considered crucial a of Edges type,Edges of an image are considered a crucial type of information.,11782
can by These applying edge with detectors different be methodology. extracted,These can be extracted by applying edge detectors with different methodology.,11783
our detection competitive Extensive experimental proved edge that a significantly method algorithm. is evaluations,Extensive experimental evaluations proved that our edge detection method is significantly a competitive algorithm.,11784
category We an of fitness find for recombination of landscapes. complex a advantage,We find an advantage of recombination for a category of complex fitness landscapes.,11785
recombination the However largely effect complex gene work ignores of classical on interactions.,However classical work on recombination largely ignores the effect of complex gene interactions.,11786
tends accelerate if is or to adaptation. standard the question decelerate A about recombination mechanism,A standard question about recombination is if the mechanism tends to accelerate or decelerate adaptation.,11787
"than that effects how the important more falls. we argue majority However, may be extreme","However, we argue that extreme effects may be more important than how the majority falls.",11788
"oscillations Neural a related variety of functions, including attention. to wide cognitive are","Neural oscillations are related to a wide variety of cognitive functions, including attention.",11789
input is the show capacity-achieving that distribution We IID.,We show that the capacity-achieving input distribution is IID.,11790
"not increase of that show we the does Further, this capacity channel. feedback","Further, we show that feedback does not increase the capacity of this channel.",11791
a registration this is fast method In mesh based  proposed. paper  triangular,  In this paper a fast triangular mesh based registration method is proposed.,11792
is evaluating used results. the metric Mean Difference (MSD) a Squared as for,Mean Squared Difference (MSD) is used as a metric for evaluating the results.,11793
MATLAB done implementation specific without was in optimization. any The,The implementation was done in MATLAB without any specific optimization.,11794
be speeds can achieved Higher implementations. using C/C++,Higher speeds can be achieved using C/C++ implementations.,11795
that is finite It a shown mixture representation. any restricted satisfy such must distribution,It is shown that any such distribution must satisfy a restricted finite mixture representation.,11796
rates. hypothetical Applicability tested disease and incidence a in with mortality realistic is,Applicability is tested in a hypothetical disease with realistic incidence and mortality rates.,11797
a information? lattices how the also in represents play Do role brain,Do lattices also play a role in how the brain represents information?,11798
"how resolution optimize lattice theory, of information the we to ask spatial codes. Using neuronal","Using information theory, we ask how to optimize the spatial resolution of neuronal lattice codes.",11799
"flying monkeys, prediction quantitative bats, tested This be or arboreal in experimentally could cetaceans.","This quantitative prediction could be tested experimentally in flying bats, arboreal monkeys, or cetaceans.",11800
recognized abnormality. reasons any the abnormalities can Our report of main recognize model and,Our model can recognize abnormalities and report the main reasons of any recognized abnormality.,11801
central cellular Allosteric at many processes. to sites is regulation distant,Allosteric regulation at distant sites is central to many cellular processes.,11802
knowledge The a (CNN). prior Network with pre-trained is Convolutional modeled Neural,The prior knowledge is modeled with a pre-trained Convolutional Neural Network (CNN).,11803
be concept Then the based fine-tuned new on instances. these detector can,Then the concept detector can be fine-tuned based on these new instances.,11804
cellular have recognized messengers they pathways. also in been signalling important Recently as,Recently they have been also recognized as important messengers in cellular signalling pathways.,11805
metabolism tissue. of model computational sphingolipid comprehensive We propose in human first,We propose first comprehensive computational model of sphingolipid metabolism in human tissue.,11806
during with size. organism often development scale generated Biological regeneration patterns and,Biological patterns generated during development and regeneration often scale with organism size.,11807
Our features plan regeneration in as of model in captures observed experiments. essential body flatworms,Our model captures essential features of body plan regeneration in flatworms as observed in experiments.,11808
  new datasets described We describe two by images humans. with our,  We describe our two new datasets with images described by humans.,11809
"using a Amazon were Both the collected datasets crowdsourcing platform. Turk, Mechanical","Both the datasets were collected using Amazon Mechanical Turk, a crowdsourcing platform.",11810
significantly The contain two more than existing datasets other descriptions datasets. per image,The two datasets contain significantly more descriptions per image than other existing datasets.,11811
been in Africa. proposed West epidemiological to Multiple have spread models the predict of Ebola,Multiple epidemiological models have been proposed to predict the spread of Ebola in West Africa.,11812
displacement and strain monitoring engineering. important and an of in fields problem is Reconstruction,Reconstruction and monitoring of displacement and strain fields is an important problem in engineering.,11813
"comprehensive presented. is comparative of photogrammetric and analysis review solutions a literature study, this In","In this study, a comprehensive literature review and comparative analysis of photogrammetric solutions is presented.",11814
representation. contributions the CNN of makes to two inference video This paper,This paper makes two contributions to the inference of CNN video representation.,11815
"of regularly we with associate its object material. shape As built humans, an","As humans, we regularly associate shape of an object with its built material.",11816
the a user study. also We of appraise our using output algorithm,We also appraise the output of our algorithm using a user study.,11817
datasets. proposed method is The evaluated video benchmark using,The proposed method is evaluated using benchmark video datasets.,11818
using discuss We $X$. check model the how to,We discuss how to check the model using $X$.,11819
same for the PTA likelihood belonging terrace. trees avoids computations to,PTA avoids likelihood computations for trees belonging to the same terrace.,11820
sizes Speedups independent correlate amount the on but terrace with data. of are missing,Speedups are independent on terrace sizes but correlate with the amount of missing data.,11821
"Thus, well phylogenomic for PTA analyses. is suited structure data the","Thus, the PTA data structure is well suited for phylogenomic analyses.",11822
concentrated solution. a This fibrils in when protein highly forms amyloid,This protein forms amyloid fibrils when in a highly concentrated solution.,11823
"of provide simulation, Monte on fibril crowding-induced Carlo scenarios we Based formation.","Based on Monte Carlo simulation, we provide scenarios of crowding-induced fibril formation.",11824
"of microscopy. I image report, introduce correlative this registration In methods for two","In this report, I introduce two methods of image registration for correlative microscopy.",11825
fiducials on (beads). method first is The based,The first method is based on fiducials (beads).,11826
is the afterwards to least-squares further matching registration. applied process A refine,A least-squares matching process is applied afterwards to further refine the registration.,11827
analogies approach. method the second inspired is image by The,The second method is inspired by the image analogies approach.,11828
into introduce the analogies. representation I sparse model image,I introduce the sparse representation model into image analogies.,11829
a one. method The problem transforms a to registration mono-modal multi-modal,The method transforms a multi-modal registration problem to a mono-modal one.,11830
test my Electron approaches (TEM) Transmission microscopy on confocal I and images. Microscopy,I test my approaches on Transmission Electron Microscopy (TEM) and confocal microscopy images.,11831
the report. of shown are Experimental in results methods this also,Experimental results of the methods are also shown in this report.,11832
gradient for minimizing stochastic objective functions. methods Classical well suited expected-value are,Classical stochastic gradient methods are well suited for minimizing expected-value objective functions.,11833
The two iterations the interplay convergence scales. of time with different involves,The convergence involves the interplay of two iterations with different time scales.,11834
"dynamic find wide in SCGD applications proposed estimation, methods etc. The programming, learning,","The proposed SCGD methods find wide applications in learning, estimation, dynamic programming, etc.",11835
"network on create studies this focus Frequently, drive inputs how structure. plasticity theoretical to feedforward","Frequently, theoretical studies focus on how feedforward inputs drive plasticity to create this network structure.",11836
an image-dependent combining Their metric. construction with distance contrast into spatial information based on is,Their construction is based on combining spatial distance with contrast information into an image-dependent metric.,11837
"information image local hold computing texture. are that In graph about generated amoebas, structures","In computing amoebas, graph structures are generated that hold information about local image texture.",11838
subject. the directions and extensions out presents future on for It some points investigation,It presents some extensions and points out directions for future investigation on the subject.,11839
networks by Most based on firing rates. of are models computation neural population,Most models of computation by neural networks are based on population firing rates.,11840
"to approach extended Here, can be the spike-based show plausible networks. biophysically how we","Here, we show how the spike-based approach can be extended to biophysically plausible networks.",11841
These increase the biological spike-based approach computation. to significantly plausibility the results of network,These results significantly increase the biological plausibility of the spike-based approach to network computation.,11842
"games. paper, we such the targeted of positions this In for problem generating starting address","In this paper, we address the problem of generating targeted starting positions for such games.",11843
that We approximating is undecidable. optimal show the costs integer cost for,We show that for integer costs approximating the optimal cost is undecidable.,11844
different several detailed classifiers. experimental descriptors a show We of with evaluation,We show a detailed experimental evaluation of different descriptors with several classifiers.,11845
well. very experiments that clearly perform they show Our,Our experiments clearly show that they perform very well.,11846
of Deontic logic branch very mathematical  is and philosophy.  a researched logic well,  Deontic logic is a very well researched branch of mathematical logic and philosophy.,11847
"quite accurate, and model which verify both is we often quantitatively. Our qualitatively","Our model is often quite accurate, which we verify both qualitatively and quantitatively.",11848
computation. are Neural processing during task study a brain to central cognitive information correlations and,Neural correlations during a cognitive task are central to study brain information processing and computation.,11849
"currently We approaches, detection observe exist quality. families that reaching of similar there all three","We observe that there exist three families of approaches, all currently reaching similar detection quality.",11850
of the robust estimator kernel popular present a (KDE). version density We,We present a robust version of the popular kernel density estimator (KDE).,11851
a KDE This scaled yields and (SPKDE). projected,This yields a scaled and projected KDE (SPKDE).,11852
series. widely in is time Simple forecasting smoothing exponential economic used,Simple exponential smoothing is widely used in forecasting economic time series.,11853
forecasts. quick compute This generally it delivers because accurate is to it is and,This is because it is quick to compute and it generally delivers accurate forecasts.,11854
series The for forecasting. be can method used time,The method can be used for time series forecasting.,11855
Such complex data are are concepts models limited. training may target when advantages and/or have,Such models may have advantages when target concepts are complex and/or training data are limited.,11856
incorporated the state are long-term when dependencies possible is network nonlinearities Learning into updates.,Learning long-term dependencies is possible when nonlinearities are incorporated into the network state updates.,11857
that This same of input works technique data the B-scans of consists location. on several,This technique works on input data that consists of several B-scans of the same location.,11858
tasks. for We performance three all on benchmarks state-of-the-art achieve,We achieve state-of-the-art performance on benchmarks for all three tasks.,11859
"outermost level of taken is simplicity, the graphs into account. For only","For simplicity, only the outermost level of graphs is taken into account.",11860
"systematic the diagnostic of preferences, and ""conflict conflict considered. Localization area"", bases are generation for","Localization of the ""conflict area"", diagnostic preferences, and bases for systematic conflict generation are considered.",11861
are visual networks powerful hierarchies yield models of that features. Convolutional,Convolutional networks are powerful visual models that yield hierarchies of features.,11862
"fast crowd fully paper, propose for segmentation. network this In we (FCNN) neural a convolutional","In this paper, we propose a fast fully convolutional neural network (FCNN) for crowd segmentation.",11863
attractive it other These segmentation make properties problems. general to image extendable,These attractive properties make it extendable to other general image segmentation problems.,11864
motion then stage-by-stage filers appearance optimized. pretrained are and jointly and Both filters,Both appearance filters and motion filers are pretrained stage-by-stage and then jointly optimized.,11865
released be datasets They are to the the largest currently public. and crowd segmentation will,They are currently the largest crowd segmentation datasets and will be released to the public.,11866
"ROSS includes The the language called representational for classes. specification scheme a of ""Star"" ontology","The ROSS representational scheme includes a language called ""Star"" for the specification of ontology classes.",11867
"also called a The model"". formal method the ROSS scheme ""instance includes","The ROSS method also includes a formal scheme called the ""instance model"".",11868
area of models the in are natural used situations. to representation language Instance represent meaning,Instance models are used in the area of natural language meaning representation to represent situations.,11869
for method. ROSS the the and both provides paper This background the rationale philosophical,This paper provides both the rationale and the philosophical background for the ROSS method.,11870
(CNNs). accelerate convolutional This deep networks test-time computation the to paper neural of aims,This paper aims to accelerate the test-time computation of deep convolutional neural networks (CNNs).,11871
nonlinear We effective problem. develop optimization to an constrained solution this,We develop an effective solution to this constrained nonlinear optimization problem.,11872
algorithm approximated. is also reducing An presented error layers for are when accumulated multiple the,An algorithm is also presented for reducing the accumulated error when multiple layers are approximated.,11873
what But the we should right architecture is use?,But what is the right architecture we should use?,11874
limited. asynchronous results therefore show Our that is fundamentally reducibility of the networks,Our results therefore show that the reducibility of asynchronous networks is fundamentally limited.,11875
keypoint identify images from viewpoint. candidates training good We in the multiple first taken same,We first identify good keypoint candidates in multiple training images taken from the same viewpoint.,11876
"this In HPE. for \emph{latent} and propose we clothing a shall paper, explore attribute approach","In this paper, we shall propose and explore a \emph{latent} clothing attribute approach for HPE.",11877
our show approach. benchmarks on proposed extensive real-world experiments performance two state-of-the-art of Our the,Our extensive experiments on two real-world benchmarks show the state-of-the-art performance of our proposed approach.,11878
time. significant perceived variations There outcomes over in were,There were significant variations in perceived outcomes over time.,11879
good went report on group to outcomes. this most Nevertheless ultimately in,Nevertheless most in this group went on to ultimately report good outcomes.,11880
"carers. overall considered so DBS, life Many quality much than more recipients of better following","Many considered overall quality of life much better following DBS, more so recipients than carers.",11881
varied greatly Post-DBS experiences non-motor cases. between of and motor symptoms both,Post-DBS experiences of both motor and non-motor symptoms varied greatly between cases.,11882
"worse symptoms to show no Some DBS. or reported following be were improvement,","Some symptoms were reported to show no improvement, or be worse following DBS.",11883
on been Recent with pose estimation Networks achieved has Convolutional Deep performance (ConvNets). human-body state-of-the-art,Recent state-of-the-art performance on human-body pose estimation has been achieved with Deep Convolutional Networks (ConvNets).,11884
of at reduced benefits cost of come the pooling These localization accuracy.,These benefits of pooling come at the cost of reduced localization accuracy.,11885
unbounded support. examine possible discrete of with all We cases possibly carefully continuous and variables,We carefully examine all possible cases of discrete and continuous variables with possibly unbounded support.,11886
wave phenomenon Spreading is a depression in tissue. gray (SD) matter,Spreading depression (SD) is a wave phenomenon in gray matter tissue.,11887
"characterized cell of membranes. Locally, it massive ions across re-distribution by is","Locally, it is characterized by massive re-distribution of ions across cell membranes.",11888
categorical when mixture numerical observations arise and described a by data Mixed are of variables.,Mixed data arise when observations are described by a mixture of numerical and categorical variables.,11889
standard package type R to of PCAmixdata extends incorporate analysis this multivariate The data. methods,The R package PCAmixdata extends standard multivariate analysis methods to incorporate this type of data.,11890
as querying this effectively constraint be thus performed. implemented checking SPARQL Further can and,Further this constraint checking can be implemented as SPARQL querying and thus effectively performed.,11891
"use who entirely. fail i.e., severe Transform, In many more extraction cases, the Hough can","In more severe cases, i.e., many who use the Hough Transform, extraction can fail entirely.",11892
"In propose a this we method these tackles paper, issues. that","In this paper, we propose a method that tackles these issues.",11893
maps localized. originates connected edge and robust well that This are,This originates connected edge maps that are robust and well localized.,11894
sp. in the of strings Transfer Synechocystis nucleotide,Transfer of nucleotide strings in the Synechocystis sp.,11895
an strings. reveals intrinsic It the dynamics transfer of in partly nucleotide,It partly reveals an intrinsic dynamics in the transfer of nucleotide strings.,11896
with The are thus basic results mobile endowed the structures. on correlation the elements,The mobile elements are thus endowed with the basic results on the correlation structures.,11897
individual of linear data variables Factor a set. factors relationships between describe provides analysis that,Factor analysis provides linear factors that describe relationships between individual variables of a data set.,11898
and have achieved Parts in performance Networks Convolutional Models detection. each notable Deformable object,Deformable Parts Models and Convolutional Networks each have achieved notable performance in object detection.,11899
"we of the effectiveness demonstrate experiments, the Through the method. proposed","Through the experiments, we demonstrate the effectiveness of the proposed method.",11900
"a argumentation-theoretic Trees in literal i.e. justify Attack terms,","Attack Trees justify a literal in argumentation-theoretic terms, i.e.",11901
We bipartite based on investigate solutions weighted graphs.,We investigate solutions based on weighted bipartite graphs.,11902
can show real-world method coarsely also that provide high-quality We from products our datasets. registered,We also show that our method can provide high-quality products from coarsely registered real-world datasets.,11903
"implementation future is facilitate research. MATLAB provided a to Finally,","Finally, a MATLAB implementation is provided to facilitate future research.",11904
of question study performance of medium-sized for improving classes). (thousands database the We recognition,We study the question of improving recognition performance for medium-sized database (thousands of classes).,11905
"an paper in On we statistically the algorithm. contrary, optimal propose this greedy alternative,","On the contrary, in this paper we propose an alternative, statistically optimal greedy algorithm.",11906
with likelihood. maximal class next model selected The the to the is from check,The next model to check is selected from the class with the maximal likelihood.,11907
"novel a propose algorithm reconstruction. sparsity this paper, for In analysis-based we","In this paper, we propose a novel algorithm for analysis-based sparsity reconstruction.",11908
a are application: on imaging. conducted compressive practical sensing Experiments magnetic resonance,Experiments are conducted on a practical application: compressive sensing magnetic resonance imaging.,11909
structures changes of involved iron large binding/dissociation. crystal different hTf conformational delineate Three in,Three different crystal structures of hTf delineate large conformational changes involved in iron binding/dissociation.,11910
"lobes delivery. open, pH, endosomal available irons readily At for making both","At endosomal pH, both lobes readily open, making irons available for delivery.",11911
"counting inter-observer low subjective and However, agreement. is laborious, from suffer mitosis may","However, mitosis counting is laborious, subjective and may suffer from low inter-observer agreement.",11912
descriptions evaluation of and results are eleven from presented. methods Short the,Short descriptions and results from the evaluation of eleven methods are presented.,11913
"such implementation model direct more for stable semantics of For allows solving. efficient problems,","For such problems, direct implementation of stable model semantics allows for more efficient solving.",11914
"information too allow in be this may coarse However, precise the layer localization. to","However, the information in this layer may be too coarse to allow precise localization.",11915
"contrary, may On localization semantics. earlier in will precise capture but layers not the be","On the contrary, earlier layers may be precise in localization but will not capture semantics.",11916
"Here, we base scene on a understanding representation. to object propose high-resolution","Here, we propose to base scene understanding on a high-resolution object representation.",11917
"the or rotation usually more Variances in problem pose, the difficult. make scale","Variances in the pose, scale or rotation usually make the problem more difficult.",11918
"or avoid Importantly, like expensive information part we using box from annotations bounding end-to-end.","Importantly, we avoid using expensive annotations like bounding box or part information from end-to-end.",11919
easier constraint generalize. The supervision makes weak work our to,The weak supervision constraint makes our work easier to generalize.,11920
achieved improvements weakest supervision and significant Our the pipeline condition. under best delivered accuracy the,Our pipeline delivered significant improvements and achieved the best accuracy under the weakest supervision condition.,11921
that competitive annotations. The is performance additional rely on other methods against,The performance is competitive against other methods that rely on additional annotations.,11922
zoonotic of are emerging Ticks important vectors diseases.,Ticks are important vectors of emerging zoonotic diseases.,11923
"immature avian Furthermore, taxonomic ticks ecological hosts. to evidence that we show found or specificity","Furthermore, we found evidence that immature ticks show taxonomic or ecological specificity to avian hosts.",11924
several characteristics main The neighbors. similar normal point is with that idea has,The main idea is that normal point has similar characteristics with several neighbors.,11925
. real performance this of approach illustrated is datasets The through,The performance of this approach is illustrated through real datasets .,11926
work. consider a We depth of monocular the single estimation from image this problem in,We consider the problem of depth estimation from a single monocular image in this work.,11927
of address We videos. the in action detection problem,We address the problem of action detection in videos.,11928
two and in appearance motion incorporate We ways.,We incorporate appearance and motion in two ways.,11929
"using build feature to spatio-temporal Convolutional strong classifiers extract we Networks. representations Neural Second,","Second, we extract spatio-temporal feature representations to build strong classifiers using Convolutional Neural Networks.",11930
"We we produce consistent link our to call which action time, tubes. detections in predictions","We link our predictions to produce detections consistent in time, which we call action tubes.",11931
other outperforms the action task our approach techniques show in of detection. We that,We show that our approach outperforms other techniques in the task of action detection.,11932
subgraphs matrix connected We combinatorics of connectivity the linear into embedding overcome (LMI). inequalities by,We overcome the combinatorics of connectivity by embedding connected subgraphs into linear matrix inequalities (LMI).,11933
for numerous in algorithms. in improving shown search advances its Recent has search usefulness metareasoning,Recent advances in metareasoning for search has shown its usefulness in improving numerous search algorithms.,11934
paper IDA* applies heuristics This several available. when rational to are admissible metareasoning,This paper applies rational metareasoning to IDA* when several admissible heuristics are available.,11935
and in Biodiversity issues central evolution. extinction are,Biodiversity and extinction are central issues in evolution.,11936
"we ""border BN present our algorithm,"" converts directed chain. Here a which a first into","Here we first present our ""border algorithm,"" which converts a BN into a directed chain.",11937
All in this probabilistic quantities interpretations. have paper,All quantities in this paper have probabilistic interpretations.,11938
using or matches tasks. also outperforms works for CNN Our these recent approach,Our approach also outperforms or matches recent works using CNN for these tasks.,11939
code the system complete online. is Source available of,Source code of the complete system is available online.,11940
at guarantees Our decreases original iteration. energy that each the algorithm,Our algorithm guarantees that the original energy decreases at each iteration.,11941
attenuation features and ill-conditioned bias resulting to generates matrices. feature introduces the This,This attenuation introduces bias to the resulting features and generates ill-conditioned feature matrices.,11942
for by information from at lost MIFS using compensates scales. recapturing coarse differential operators information,MIFS compensates for information lost from using differential operators by recapturing information at coarse scales.,11943
allows ranges This recaptured information different actions us match to and motion. speeds at of,This recaptured information allows us to match actions at different speeds and ranges of motion.,11944
exponentially. differential-based that the MIFS We enhances features learnability of prove,We prove that MIFS enhances the learnability of differential-based features exponentially.,11945
detection event action performance show and on results challenging tasks. improved Experimental significantly recognition,Experimental results show significantly improved performance on challenging action recognition and event detection tasks.,11946
are systems and splines available surface in with designs. curves the B\'ezier various widely,B\'ezier splines are widely available in various systems with the curves and surface designs.,11947
determines points number B\'ezier of polynomial. the of control degree the The,The number of control points determines the degree of the B\'ezier polynomial.,11948
cubic B\'ezier segments. from cubic is The B\'ezier curve obtained spline,The cubic B\'ezier spline is obtained from cubic B\'ezier curve segments.,11949
"To various compared. proposed demonstrate the the experimental algorithm, of are performances results","To demonstrate the various performances of the proposed algorithm, experimental results are compared.",11950
"state-of-the-art mate- D-CNN recognition. improves rial in the texture, substantially and scene","D-CNN substantially improves the state-of-the-art in texture, mate- rial and scene recognition.",11951
"arbitrary can seamlessly regions multi-scale describe D-CNN shapes and incorporate and information Furthermore, sizes. of","Furthermore, D-CNN can seamlessly incorporate multi-scale information and describe regions of arbitrary shapes and sizes.",11952
multi-group case. the in These faster best are rates the known significantly than results,These rates are significantly faster than the best known results in the multi-group case.,11953
"with two-group the coincide minimax optimal Moreover, rates they case. the for","Moreover, they coincide with the optimal minimax rates for the two-group case.",11954
analysis. with validate numerical We results theoretical our,We validate our theoretical results with numerical analysis.,11955
of once The again in theoretical effect foundations. Baldwin is convincing need,The Baldwin effect is once again in need of convincing theoretical foundations.,11956
"experiments to show carried out describe reliability the the Moreover, the method. of we","Moreover, we describe the experiments carried out to show the reliability of the method.",11957
"direct proposes classification image tags, without paper This from user-supplied filtering. of learning","This paper proposes direct learning of image classification from user-supplied tags, without filtering.",11958
user is who supplied tag Each the the online. by shared image,Each tag is supplied by the user who shared the image online.,11959
"Our manual the is to approach costly. conventional extremely is complementary annotation, which of approach","Our approach is complementary to the conventional approach of manual annotation, which is extremely costly.",11960
"problems. based methods problems, efficiently can to For solve cuts be large-scale submodular employed","For submodular problems, cuts based methods can be employed to efficiently solve large-scale problems.",11961
"nonsubmodular solve. However, more are general challenging significantly problems to","However, general nonsubmodular problems are significantly more challenging to solve.",11962
"to its Spectral easy is simple is and relaxation implement, loose. bound but","Spectral relaxation is simple and easy to implement, but its bound is loose.",11963
"work, In for with BQPs, formulation desirable we present a this SDP properties. two new","In this work, we present a new SDP formulation for BQPs, with two desirable properties.",11964
"First, SDP conventional relaxation it a formulations. to bound similar has","First, it has a similar relaxation bound to conventional SDP formulations.",11965
"and for quasi-Newton the two solvers, Newton We namely, then dual smoothing problem. propose methods,","We then propose two solvers, namely, quasi-Newton and smoothing Newton methods, for the dual problem.",11966
are than efficiently them more significantly interior-point Both of standard methods.,Both of them are significantly more efficiently than standard interior-point methods.,11967
models Single computational a tradition long neuroscience. neuron in have,Single neuron models have a long tradition in computational neuroscience.,11968
can then overfitting. data fitted be show that efficiently to without We this model,We then show that this model can be efficiently fitted to data without overfitting.,11969
"as very such basic still utilize detections. However, information distance between they","However, they still utilize very basic information such as distance between detections.",11970
are propose approach novel training detectors. images a for effective to object synthesizing that We,We propose a novel approach to synthesizing images that are effective for training object detectors.,11971
multivariate model propose statistics. underlying Gaussian We generalized analyze and the distribution to the (MGGD),We propose and analyze the multivariate generalized Gaussian distribution (MGGD) to model the underlying statistics.,11972
"configuration is color measure. KLD to with Furthermore, CIELAB the associated space obtained deviation best","Furthermore, the best configuration is obtained with CIELAB color space associated to KLD deviation measure.",11973
"via nanoparticles, that depends size. uptake on cellular endocytosis, show of receptor-mediated nanoparticle Experiments strongly","Experiments show that cellular uptake of nanoparticles, via receptor-mediated endocytosis, strongly depends on nanoparticle size.",11974
proposed this a is by paper cross-convolutional-layer technique This in new achieved called adopting pooling.,This is achieved by adopting a new technique proposed in this paper called cross-convolutional-layer pooling.,11975
"differences or rates across loci mutation range expansions, background in selection).","range expansions, differences in mutation rates across loci or background selection).",11976
mechanics We methods. scope and its other its compare to existing discuss,We discuss its scope and compare its mechanics to other existing methods.,11977
"vision, problem learning recognition. selection and essential Feature in category is computer important for an","Feature selection is an essential problem in computer vision, important for category learning and recognition.",11978
"Different classification a that we with solutions. discriminant start linear sparse from encourages boosting, formulation","Different from boosting, we start with a discriminant linear classification formulation that encourages sparse solutions.",11979
equivalent we that diverse unsupervised an Then discovers of problem clustering jointly ensembles obtain features.,Then we obtain an equivalent unsupervised clustering problem that jointly discovers ensembles of diverse features.,11980
valuable They united more but are of in cluster a powerful when independently even classifiers.,They are independently valuable but even more powerful when united in a cluster of classifiers.,11981
"Mixture on distribution. which The the based Laplacian Laplacian (LMM), is Model a first is","The first is a Laplacian Mixture Model (LMM), which is based on the Laplacian distribution.",11982
variations. the due face Predicting attributes is face challenging wild in complex to,Predicting face attributes in the wild is challenging due to complex face variations.,11983
learning attribute novel for deep We propose framework in a prediction the wild.,We propose a novel deep learning framework for attribute prediction in the wild.,11984
of well sparse be concepts. can Each these combination a attribute linear with explained,Each attribute can be well explained with a sparse linear combination of these concepts.,11985
"mass reporting market, cameras sales in the already GoPro increase year. penetrated every substantial","GoPro cameras already penetrated the mass market, reporting substantial increase in sales every year.",11986
few captured photographer recognized of The video when from be walking. seconds a reliably can,The photographer can be reliably recognized from a few seconds of video captured when walking.,11987
may shot also YouTube) egocentric a specific videos for photographer become by possible.,YouTube) for egocentric videos shot by a specific photographer may also become possible.,11988
Solving problem cues. motion both difficult combining a such and requires appearance,Solving such a difficult problem requires combining both appearance and motion cues.,11989
"inevitable due shared In to networks, input. typically correlations presynaptic finite are","In finite networks, correlations are typically inevitable due to shared presynaptic input.",11990
Syllogism is deductive type statements. quantified reasoning a of involving,Syllogism is a type of deductive reasoning involving quantified statements.,11991
to statistic data. of the Invariance that with uninformative is transformations guarantees respect constant the,Invariance guarantees that the statistic is constant with respect to uninformative transformations of the data.,11992
"gap. we closing first toward Here, a this take step","Here, we take a first step toward closing this gap.",11993
not coding benefit stimuli triplet similar correlations. both occur This when elicit does,This coding benefit does not occur when both stimuli elicit similar triplet correlations.,11994
higher-order on have coding. correlations effect results that population could a indicate strong These,These results indicate that higher-order correlations could have a strong effect on population coding.,11995
task. in real-world materials a challenging images is Recognizing,Recognizing materials in real-world images is a challenging task.,11996
"not this did total variation apply regularization. to analysis algorithms based Unfortunately, on","Unfortunately, this analysis did not apply to algorithms based on total variation regularization.",11997
variation reason Our on works. why the an reveals total very based analysis algorithm,Our analysis reveals the very reason why an algorithm based on total variation works.,11998
numerical utility of through demonstrate simulated approaches experiments We these data. on the,We demonstrate the utility of these approaches through numerical experiments on simulated data.,11999
"methods In approach kernel paper, this data. develop with an to exploiting manifold-valued we","In this paper, we develop an approach to exploiting kernel methods with manifold-valued data.",12000
be representation. distinguished high-level ambiguity This can by,This ambiguity can be distinguished by high-level representation.,12001
"special of has considered. For case been processes, the Dirichlet streaming inference","For the special case of Dirichlet processes, streaming inference has been considered.",12002
"demonstrate text corpora. the the of efficacy in We documents algorithm on streaming clustering large,","We demonstrate the efficacy of the algorithm on clustering documents in large, streaming text corpora.",12003
"small. might be examples number when Therefore, is of training unreliable the they","Therefore, they might be unreliable when the number of training examples is small.",12004
"is robust overfitting to it small-size training Hence, data. on","Hence, it is robust to overfitting on small-size training data.",12005
"of asymmetric. is majority Unfortunately, models the graphical real-world","Unfortunately, the majority of real-world graphical models is asymmetric.",12006
for the This relational case representations is when given. evidence even is,This is even the case for relational representations when evidence is given.,12007
"existing Markov such over-symmetric as First, networks. a logic relational all representation require approximations","First, all existing over-symmetric approximations require a relational representation such as Markov logic networks.",12008
"often computed the symmetries significantly, distribution highly the probabilities biased. making change Second, induced the","Second, the induced symmetries often change the distribution significantly, making the computed probabilities highly biased.",12009
"The while leads estimates probability therefore, unbiased. remaining framework, to improved","The framework, therefore, leads to improved probability estimates while remaining unbiased.",12010
existing the approach Experiments demonstrate MCMC outperforms that algorithms.,Experiments demonstrate that the approach outperforms existing MCMC algorithms.,12011
Analyzing corporations to Big Data im-prove help efficiency. can their,Analyzing Big Data can help corporations to im-prove their efficiency.,12012
"Semantic Machine-Learning proposea techniques HMC process, scalable We Rule-based reasoning. and also using","We also proposea Semantic HMC process, using scalable Machine-Learning techniques and Rule-based reasoning.",12013
on We egocentric from task pose of estimation viewpoints. everyday focus hand the,We focus on the task of everyday hand pose estimation from egocentric viewpoints.,12014
"understanding limited. of them remains Nevertheless, our","Nevertheless, our understanding of them remains limited.",12015
to contribute framework answer invert general this a question we To representations.,To answer this question we contribute a general framework to invert representations.,12016
"In compact learn is binary their embeddings manifolds to on work, intrinsic how this considered.","In this work, how to learn compact binary embeddings on their intrinsic manifolds is considered.",12017
macromolecules of noisy structural central in is in variability projections biological a Cryo-EM. Classifying problem,Classifying structural variability in noisy projections of biological macromolecules is a central problem in Cryo-EM.,12018
"very easy and Finally, train are to conceptually and use. simple they","Finally, they are conceptually very simple and easy to train and use.",12019
"utility our of Lastly, the on through Internet experiments demonstrate collections. we formulations photo","Lastly, we demonstrate the utility of our formulations through experiments on Internet photo collections.",12020
"everyday objects. pose during interactions estimation we Importantly, with consider","Importantly, we consider pose estimation during everyday interactions with objects.",12021
performance. work robust depth-based priors Past features for and are pose+viewpoint crucial shows that strong,Past work shows that strong pose+viewpoint priors and depth-based features are crucial for robust performance.,12022
egocentric workspace. this call volume We an,We call this volume an egocentric workspace.,12023
appearance property notable A workspace with location. is correlates hand that,A notable property is that hand appearance correlates with workspace location.,12024
the This simplify performance. and improves architecture greatly,This greatly simplify the architecture and improves performance.,12025
state-of-the-art method images performance pose hand RGB-D Our recognition egocentric provides from real-time. in,Our method provides state-of-the-art hand pose recognition performance from egocentric RGB-D images in real-time.,12026
We architecture semantic purely for a introduce segmentation. feed-forward,We introduce a purely feed-forward architecture for semantic segmentation.,12027
by classified a network. multilayer feedforward are Instead superpixels,Instead superpixels are classified by a feedforward multilayer network.,12028
distortion a Results for of great type. its effectiveness variety demonstrates,Results demonstrates its effectiveness for a great variety of distortion type.,12029
uses original other little the benefits about very only this Among image. information measure,Among other benefits this measure uses only very little information about the original image.,12030
"paper, we cost. accuracy time constrained the CNNs this of investigate In under","In this paper, we investigate the accuracy of CNNs under constrained time cost.",12031
also This the in importance is designs. helpful of for network the factors understanding,This is also helpful for understanding the importance of the factors in network designs.,12032
"state-of-the-art our throughout. we of performance pipeline, stages Analysing show the","Analysing the stages of our pipeline, we show state-of-the-art performance throughout.",12033
involved cost reduces runtime the in memory in learned turn CNNs. This and deploying the,This in turn reduces the memory and runtime cost involved in deploying the learned CNNs.,12034
"Besides, proposed on the results. more densities significant experiment tissues bring we approach various to","Besides, we experiment the proposed approach on various tissues densities to bring more significant results.",12035
"Base. explored from At challenging Data end, we breast medical some BIRADS this images","At this end, we explored some challenging breast images from BIRADS medical Data Base.",12036
methods. promising with segmentation mass results the to first Our showed edges regard experimentations,Our first experimentations showed promising results with regard to the edges mass segmentation methods.,12037
works This this in the discusses first area. paper achieved main,This paper discusses first the main works achieved in this area.,12038
evaluated concluding paper. our and The results main are before showed,The main results are showed and evaluated before concluding our paper.,12039
"viewpoints. cases, of methods with as large Appearance-based appearances such change changes in fail","Appearance-based methods fail in such cases, as appearances change with large changes of viewpoints.",12040
"computed common different robust efficiently. This be can occlusions viewpoints, under measure to and is","This measure is robust to occlusions common under different viewpoints, and can be computed efficiently.",12041
is retrieval cameras. stationary held demonstrated challenging Event using from and videos hand,Event retrieval is demonstrated using challenging videos from stationary and hand held cameras.,12042
in detection image component fundamental and been detection Contour object segmentation has many a systems.,Contour detection has been a fundamental component in many image segmentation and object detection systems.,12043
"related and However, predicting mutually that claim we are recognizing objects two tasks. contours","However, we claim that recognizing objects and predicting contours are two mutually related tasks.",12044
the This network scales of image different the to is applied four of section input.,This section of the network is applied to four different scales of the image input.,12045
"Why addressed complete? is we here Turing are why, question The","The question addressed here is why, Why are we Turing complete?",12046
theory we theories. computing fusing build and a problem So by set,So we build a problem theory by fusing set and computing theories.,12047
is conditions complete. Turing be to The the of last,The last of the conditions is to be Turing complete.,12048
the most problems. answer would question to be: our to solve Then,Then the answer to our question would be: to solve most problems.,12049
"gradient descriptor, This efficient introduces spatiotemporal called local high histograms a (GBH). boundary paper","This paper introduces a high efficient local spatiotemporal descriptor, called gradient boundary histograms (GBH).",12050
"to GBH spatio-temporal on compute. is descriptor proposed are built gradients, fast which The simple","The proposed GBH descriptor is built on simple spatio-temporal gradients, which are fast to compute.",12051
the of model tractable inference For marginal likelihood the approximations made. to be must,For tractable inference approximations to the marginal likelihood of the model must be made.,12052
paper with idea this In compression. a we nested extend this variational,In this paper we extend this idea with a nested variational compression.,12053
"to Thereby, we able characteristic are the objects. properties determine of","Thereby, we are able to determine characteristic properties of the objects.",12054
"a a direct recognition comparison to similar rates. superior algorithm Moreover, shows","Moreover, a direct comparison to a similar algorithm shows superior recognition rates.",12055
firing and evoked by learn can replay sequences stimuli. Neuronal circuits of patterns sensory,Neuronal circuits can learn and replay firing patterns evoked by sequences of sensory stimuli.,12056
both a precise order sequences. propose of the timing We for learning event and mechanism,We propose a mechanism for learning both the order and precise timing of event sequences.,12057
population Learned between another. populations synaptic to time one determine the necessary weights for activate,Learned synaptic weights between populations determine the time necessary for one population to activate another.,12058
how of inherited The playback stochasticity is timings. determine process dynamics the sequence in learned,The dynamics of the playback process determine how stochasticity is inherited in learned sequence timings.,12059
cooperation pressure Relatedness affect and selection the and on altruism. synergy,Relatedness and synergy affect the selection pressure on cooperation and altruism.,12060
"complementary ongoing contribute to we but ways. Here, this synthesis two in distinct","Here, we contribute to this ongoing synthesis in two distinct but complementary ways.",12061
helping. on of This relatedness (or spatial qualitative broadens evolution of the the effects structure),This broadens the qualitative effects of relatedness (or spatial structure) on the evolution of helping.,12062
"this exploit convolutional to features. propose masking we In via information shape paper,","In this paper, we propose to exploit shape information via masking convolutional features.",12063
"segments The treated (e.g., convolutional as super-pixels) on are proposal feature masks maps. the","The proposal segments (e.g., super-pixels) are treated as masks on the convolutional feature maps.",12064
take will for emerge? long it to How them,How long will it take for them to emerge?,12065
our firing-rate self-organization adaptation. questions these on based model address within We,We address these questions within our self-organization model based on firing-rate adaptation.,12066
significant presents an to occlusion. humans there parsing This paper when is approach,This paper presents an approach to parsing humans when there is significant occlusion.,12067
a We call parts. connected object flexible subtree composition of each,We call each connected subtree a flexible composition of object parts.,12068
method a cues. occlusion novel for involves This learning,This involves a novel method for learning occlusion cues.,12069
we a of inference search During over to models. flexible need mixture different,During inference we need to search over a mixture of different flexible models.,12070
archaeologists. uninitiated put These off can aspects,These aspects can put off uninitiated archaeologists.,12071
object spurred research methods. This recent improving in proposal,This spurred recent research in improving object proposal methods.,12072
"stable much Level iso-intensity ones. such more or are under especially the longer curves conditions,","Level or iso-intensity curves are much more stable under such conditions, especially the longer ones.",12073
"stable In corners on on iso-curves we them. this detect portions and long paper, identify","In this paper, we identify stable portions on long iso-curves and detect corners on them.",12074
the signal. identification-verification It supervisory learned with is,It is learned with the identification-verification supervisory signal.,12075
Markov the two-parameter overly Kimura complex or model). (the general model),the Kimura two-parameter model) or overly complex (the general Markov model).,12076
"i.e. eyes salient can recognize on person based identities Human small regions,","Human eyes can recognize person identities based on small salient regions, i.e.",12077
distinctive is reliable across human in pedestrian matching views. saliency and disjoint camera,human saliency is distinctive and reliable in pedestrian matching across disjoint camera views.,12078
person methods outperforms approach on both Our state-of-the-art the re-identification datasets.,Our approach outperforms the state-of-the-art person re-identification methods on both datasets.,12079
of unknown parameters. used compute These estimators are samples to the the Bayesian finally generated,These generated samples are finally used to compute the Bayesian estimators of the unknown parameters.,12080
We a algorithm Problem. have concrete Traveling for Salesman created the,We have created a concrete algorithm for the Traveling Salesman Problem.,12081
images We model of regions. generates descriptions their and a that language natural present,We present a model that generates natural language descriptions of images and their regions.,12082
for importance of of classification. the parts investigate action and the tasks We attribute,We investigate the importance of parts for the tasks of action and attribute classification.,12083
biomolecules important most Proteins for organisms. the are living,Proteins are the most important biomolecules for living organisms.,12084
"are characterization, for utilized protein classification. identification MTFs and","MTFs are utilized for protein characterization, identification and classification.",12085
track proposed of origin geometric of the method protein topological slicing The to invariants. is,The method of slicing is proposed to track the geometric origin of protein topological invariants.,12086
and all-atom coarse-grained representations of MTFs are constructed. Both,Both all-atom and coarse-grained representations of MTFs are constructed.,12087
"end, filtration based correlation developed. a this To is matrix","To this end, a correlation matrix based filtration is developed.",12088
excellent simulation and between homology prediction An molecular consistence is persistent found. our dynamics,An excellent consistence between our persistent homology prediction and molecular dynamics simulation is found.,12089
topology-function the This reveals work of proteins. relationship,This work reveals the topology-function relationship of proteins.,12090
"and essentially qualitative limited data (CIA). a such is analysis identification characterization, However, to success","However, such a success is essentially limited to qualitative data characterization, identification and analysis (CIA).",12091
"a we objective-oriented protocol homology general outline methods. this persistent work, to construct In","In this work, we outline a general protocol to construct objective-oriented persistent homology methods.",12092
EGG generator waveform assembly those neuron based modeled gastric data. using on morphological,EGG waveform generator based on gastric morphological neuron assembly modeled using those data.,12093
correlation to our based further filtration introduced findings. matrix verify A is,A correlation matrix based filtration is introduced to further verify our findings.,12094
therapy lumbar as injection images These guidance ultrasound for of act treatment for radiculopathy.,These ultrasound images act as guidance for injection therapy for treatment of lumbar radiculopathy.,12095
"resorting protein the protein one predict without can to As flexibility Hamiltonian. such, interaction","As such, one can predict protein flexibility without resorting to the protein interaction Hamiltonian.",12096
This algorithm fast work analysis for macromolecules. (fFRI) the a introduces of large FRI flexibility,This work introduces a fast FRI (fFRI) algorithm for the flexibility analysis of large macromolecules.,12097
The proposed the complexity O(N). further fFRI to reduces computational,The proposed fFRI further reduces the computational complexity to O(N).,12098
"for protein analysis (aFRI) Additionally, the FRI of algorithms anisotropic we propose dynamics. collective","Additionally, we propose anisotropic FRI (aFRI) algorithms for the analysis of protein collective dynamics.",12099
The FRI accuracy method tested. the of is,The accuracy of the FRI method is tested.,12100
"of aFRI the and domain demonstrate FRI application protein Finally, we analysis. to","Finally, we demonstrate the application of FRI and aFRI to protein domain analysis.",12101
images. to transform applied is distorted reference and This both,This transform is applied to both reference and distorted images.,12102
information graphical of individuality in consider theoretic and terms principles. We biological,We consider biological individuality in terms of information theoretic and graphical principles.,12103
Our boundaries is decomposition an system-environment supporting algorithmic purpose extract individuality. to through,Our purpose is to extract through an algorithmic decomposition system-environment boundaries supporting individuality.,12104
than that exist. assume infer they individuals or detect evolved We rather,We infer or detect evolved individuals rather than assume that they exist.,12105
"structural introduced reviewed. of by equations, Halpern definition Pearl, which causality and uses A is","A definition of causality introduced by Halpern and Pearl, which uses structural equations, is reviewed.",12106
takes into account of an epistemic state. blame agent's Degree,Degree of blame takes into account an agent's epistemic state.,12107
"Wright's compared structural-equations the test. Finally, definition of causality NESS is to","Finally, the structural-equations definition of causality is compared to Wright's NESS test.",12108
in often is retical a to theo- hard observation The noise treat analysis.,The noise observation is often hard to treat in a theo- retical analysis.,12109
"exploitation, arbitration mechanism between utilizes that it an i.e. balances Additionally,","Additionally, it utilizes an arbitration mechanism that balances between exploitation, i.e.",12110
of sequencing cytosine Bisulphite the Motivation: detection methylation. enables,Motivation: Bisulphite sequencing enables the detection of cytosine methylation.,12111
indicates fluctuations. analysis an oscillatory on statistical clearly Our dependence rainfall of levels EPG,Our statistical analysis clearly indicates an oscillatory dependence of EPG levels on rainfall fluctuations.,12112
"smaller practice. and in this more model is preferable achieves a complexity However,","However, this model achieves a smaller complexity and is more preferable in practice.",12113
framework for linear semiparametric inferential models. dimensional ratio generalized likelihood We based a high propose,We propose a likelihood ratio based inferential framework for high dimensional semiparametric generalized linear models.,12114
"likelihood. post-regularization is inferential on based novel approach a existing Unlike directional methods, our","Unlike existing post-regularization inferential methods, our approach is based on a novel directional likelihood.",12115
proposed illustrate provided approach. and to our simulation analysis are data studies Extensive real,Extensive simulation studies and real data analysis are provided to illustrate our proposed approach.,12116
whith feedback neuron is considered. A delayed binding (BN),A binding neuron (BN) whith delayed feedback is considered.,12117
The with neuron externally $\lambda$. stream intensity Poisson a of is fed,The neuron is fed externally with a Poisson stream of intensity $\lambda$.,12118
$\Delta$. into are input fed neuron's delay its spikes time output with The,The neuron's output spikes are fed into its input with time delay $\Delta$.,12119
can presence firing is radically output feedback alter neuronal concluded that statistics. It delayed,It is concluded that delayed feedback presence can radically alter neuronal output firing statistics.,12120
factor variation to background. is image One cluttered the within-class contributing,One contributing factor to the within-class variation is cluttered image background.,12121
notions the by extended of Halpern and definition introducing Chockler blame. of causality responsibility and,Chockler and Halpern extended the definition of causality by introducing notions of responsibility and blame.,12122
"This comes benefit, high a computational cost. typically however, at","This benefit, however, typically comes at a high computational cost.",12123
"we have In aims. writing this survey, three distinct","In writing this survey, we have three distinct aims.",12124
of homomorphisms We study graphs. connectome to propose,We propose to study homomorphisms of connectome graphs.,12125
some connectome results for are computation described. graphs Initial,Initial computation results for some connectome graphs are described.,12126
"we additional can the variables, By disambiguate adding stories.","By adding additional variables, we can disambiguate the stories.",12127
is Bayesian for inference automatic with Biips systems. interacting particle a software platform,Biips is a software platform for automatic Bayesian inference with interacting particle systems.,12128
"with Octave. the softwares developed interfaces software C++ with in R, is The and Matlab","The software is developed in C++ with interfaces with the softwares R, Matlab and Octave.",12129
a systems. step general most requirement is inventive law or patentability Non-obviousness patent for in,Non-obviousness or inventive step is a general requirement for patentability in most patent law systems.,12130
"it operator to contraction this define possible the makes thanks a Moreover, identity. Harper to","Moreover, this makes it possible to define a contraction operator thanks to the Harper identity.",12131
"standard filters particle dimensions, high (e.g. However, in","However, in high dimensions, standard particle filters (e.g.",12132
analytically illustrate i.i.d. of a a simple We for this model,We illustrate this analytically for a model of a simple i.i.d.,12133
"Similar general expected structure more hold, a results are the i.i.d.~one. under to than","Similar results are expected to hold, under a more general structure than the i.i.d.~one.",12134
models structures. numerical of by complex simulations supported practical on Our are results also theoretical,Our theoretical results are also supported by numerical simulations on practical models of complex structures.,12135
"the How smart on house, are in kitchen, things neighborhood, your your internet? your","How smart are things in your kitchen, your house, your neighborhood, on the internet?",12136
through way the The standard evaluating Turing Test. is AI of,The standard way of evaluating AI is through the Turing Test.,12137
"systems evaluation Ideally, these specifically would measures. our be tuned to","Ideally, our systems would be tuned specifically to these evaluation measures.",12138
in task Image is registration imaging. the medical a of world challenging,Image registration is a challenging task in the world of medical imaging.,12139
"variety role of a plays central conditions. accurate edge clinical a registration in Particularly,","Particularly, accurate edge registration plays a central role in a variety of clinical conditions.",12140
"this edge appears accurate less to method However, registration. regarding be","However, this method appears to be less accurate regarding edge registration.",12141
"Possibilistic is Adaptive clustering possibilistic novel this paper c-means, called presented. In a c-means algorithm,","In this paper a novel possibilistic c-means clustering algorithm, called Adaptive Possibilistic c-means, is presented.",12142
algorithm indicative convergence of results behavior provided. the also are are Theoretical the of that,Theoretical results that are indicative of the convergence behavior of the algorithm are also provided.,12143
detection algorithms characterize and road road level at areas accordingly. group Current pixels pixel,Current road detection algorithms characterize road areas at pixel level and group pixels accordingly.,12144
"algorithms and variations. presence in of fail However, shadows lighting these strong","However, these algorithms fail in presence of strong shadows and lighting variations.",12145
"alignment. on detection a video propose we based road algorithm Therefore,","Therefore, we propose a road detection algorithm based on video alignment.",12146
aims road The former detection. on-line perform to,The former aims to perform on-line road detection.,12147
algorithm that valid detection the quantitative and Qualitative is prove evaluations approach. a proposed road,Qualitative and quantitative evaluations prove that the proposed algorithm is a valid road detection approach.,12148
allows framework practice. clinical model use This in for direct of NTCP a the,This framework allows for a direct use of the NTCP model in clinical practice.,12149
"not problems. related but We optimization do solve, formulate,","We formulate, but do not solve, related optimization problems.",12150
"seeded an for cellular image is algorithm paper, In this segmentation. growing proposed region automatic","In this paper, an automatic seeded region growing algorithm is proposed for cellular image segmentation.",12151
"image. interest extracted of regions (ROIs) the First, from the preprocessed","First, the regions of interest (ROIs) extracted from the preprocessed image.",12152
"selected seeds ROIs from the based extracted initial are Second, on image. automatically the","Second, the initial seeds are automatically selected based on ROIs extracted from the image.",12153
"Third, machine the selected using learning most reprehensive algorithm. are a seeds","Third, the most reprehensive seeds are selected using a machine learning algorithm.",12154
"Finally, the corresponds into cellular seed. image regions a is region to each where segmented","Finally, the cellular image is segmented into regions where each region corresponds to a seed.",12155
is programming the approach. on A problem quadratic method based box-constrained of solving new,A method of solving the new problem is based on box-constrained quadratic programming approach.,12156
is forces. search by two Our strategy guided,Our search strategy is guided by two forces.,12157
jump (e.g. to in enables image This regions distant the across,This enables to jump across distant regions in the image (e.g.,12158
publications number  the (OCR). large are Character available Recognition  Optical A of for,  A large number of publications are available for the Optical Character Recognition (OCR).,12159
"articles researches, as Latin, for as well are Japanese Chinese the and scripts. present Significant","Significant researches, as well as articles are present for the Latin, Chinese and Japanese scripts.",12160
Arabic OCR one perspective. also is script of from mature script,Arabic script is also one of mature script from OCR perspective.,12161
methods researchers are used process OCR by techniques presented. and various,OCR process techniques and methods used by various researchers are presented.,12162
last for discussed. is The also dedicated is and conclusion section future work,The last section is dedicated for future work and conclusion is also discussed.,12163
"classification. designed However, originally forward backward propagation was for and whole-image","However, forward and backward propagation was originally designed for whole-image classification.",12164
the patch-by-patch It scanning. exactly as same results those by generates,It generates exactly the same results as those by patch-by-patch scanning.,12165
the speedup The sizes increases of and with patches. images,The speedup increases with the sizes of images and patches.,12166
popular Positive (SPD) Definite become matrices information. Symmetric have encode image to,Symmetric Positive Definite (SPD) matrices have become popular to encode image information.,12167
"useful of need of is analyze the development, information the increasing. technology With extraction and","With the technology development, the need of analyze and extraction of useful information is increasing.",12168
this rules To extract Bayesian algorithm. genetic problem network from overcome we using,To overcome this problem we extract rules from Bayesian network using genetic algorithm.,12169
and the the network. matching graphical generate extracted chain we rules Bayesian by mutually Then,Then we generate the graphical chain by mutually matching the extracted rules and Bayesian network.,12170
of presents neutrosophic information. The representation steps for multi-valued paper some,The paper presents some steps for multi-valued representation of neutrosophic information.,12171
"net score. was addition, defined definedness neutrosophic it and truth, In","In addition, it was defined net truth, definedness and neutrosophic score.",12172
explores the the advances digital field document report latest recognition. This in of,This report explores the latest advances in the field of digital document recognition.,12173
"we document recognition. for hand-written discuss review addition, the technologies and available In","In addition, we review and discuss the available technologies for hand-written document recognition.",12174
"some benchmark provide this engines. company-accumulated also results available OCR we In report, on","In this report, we also provide some company-accumulated benchmark results on available OCR engines.",12175
"algorithms of also certain Besides, global proposed under convergence the conditions. investigate we the","Besides, we also investigate the global convergence of the proposed algorithms under certain conditions.",12176
and is assessment aesthetic analysis study of images. Aesthetic of properties the image the,Aesthetic image analysis is the study and assessment of the aesthetic properties of images.,12177
either or analysis computational results. to aesthetic image approaches Current provide interpretable accurate,Current computational approaches to aesthetic image analysis either provide accurate or interpretable results.,12178
computer increasingly vision. methods on Riemannian manifolds in popular become have Kernel recently,Kernel methods on Riemannian manifolds have recently become increasingly popular in computer vision.,12179
"definite of known positive very However, kernels remain on number limited. the manifolds","However, the number of known positive definite kernels on manifolds remain very limited.",12180
"kernel, A drop-off. of value, poor choice yield or performance of significant may parameter","A poor choice of kernel, or of parameter value, may yield significant performance drop-off.",12181
"optimize is the trained loss PASCAL to Importantly, function. corpus-level specifically SegNet IOU","Importantly, SegNet is specifically trained to optimize the corpus-level PASCAL IOU loss function.",12182
activity synchrony index A cross-correlation proposed. new for is neural,A new cross-correlation synchrony index for neural activity is proposed.,12183
on function. the estimation the integration cross-correlation kernel of the The index based is of,The index is based on the integration of the kernel estimation of the cross-correlation function.,12184
approximate Two of proposed distribution are the resampling statistics. the bootstrap plans to test,Two bootstrap resampling plans are proposed to approximate the distribution of the test statistics.,12185
new of as present a mixture represents model simplices. data a We which,We present a new model which represents data as a mixture of simplices.,12186
that are geometric generalize triangles. Simplices structures,Simplices are geometric structures that generalize triangles.,12187
We a give to structure us understanding learn a geometric simple efficiently. simplicial allows that,We give a simple geometric understanding that allows us to learn a simplicial structure efficiently.,12188
"which data sources. modeling of dimensions, can different Simplices inhomogeneous have facilitates","Simplices can have different dimensions, which facilitates modeling of inhomogeneous data sources.",12189
reconstruction provides method regularization. The good and,The method provides good reconstruction and regularization.,12190
models neighbor supports good constructed. to be nearest classification and It allows realistic it generative,It supports good nearest neighbor classification and it allows realistic generative models to be constructed.,12191
data become has multiple analysis. an task modalities in datasets Matching important of,Matching datasets of multiple modalities has become an important task in data analysis.,12192
"joint a Specifically, all for is built modalities. nearest-neighbor graph","Specifically, a joint nearest-neighbor graph is built for all modalities.",12193
"economical is today, In of life. human an aspect using cash of inseparable societies","In economical societies of today, using cash is an inseparable aspect of human life.",12194
"for use cashes marketing, services, People entertainments, and on. so operations bank","People use cashes for marketing, services, entertainments, bank operations and so on.",12195
"this new processing. paper, for image a model propose mathematical we In","In this paper, we propose a new mathematical model for image processing.",12196
"<x>. two and scalar operations: we real <+> Firstly, define multiplication addition","Firstly, we define two operations: addition <+> and real scalar multiplication <x>.",12197
"gray real the With of levels vector space. becomes operations, these a set","With these operations, the set of gray levels becomes a real vector space.",12198
"an of gray obtain Euclidean the space ||, levels. we","||, we obtain an Euclidean space of the gray levels.",12199
"we images. and functions these for Secondly, operations extend color","Secondly, we extend these operations and functions for color images.",12200
an image. finally on effect simple show the of various operations We,We finally show the effect of various simple operations on an image.,12201
logarithmic the context images. for authors questions such to in of models answer The,The authors answer to such questions in the context of logarithmic models for images.,12202
are The be detected categories novel continuously must and then added. datasets dynamic and,The datasets are dynamic and novel categories must be continuously detected and then added.,12203
"trained to has At prediction a unseen myriad time, categories. system with deal","At prediction time, a trained system has to deal with myriad unseen categories.",12204
"systems time, learn. require minimum even to Operational down","Operational systems require minimum down time, even to learn.",12205
existing for extends world algorithms recognition. theory Our open,Our theory extends existing algorithms for open world recognition.,12206
We evaluation a recognition systems. for of world protocol open present,We present a protocol for evaluation of open world recognition systems.,12207
superior yields world NNO recognition. on results consistently open,NNO consistently yields superior results on open world recognition.,12208
"irreversible. reduction be of essentially In was to found addition, riboflavin","In addition, reduction of riboflavin was found to be essentially irreversible.",12209
It followed second riboflavin. a of reduced is intermediate and charge-transfer by NADP+ enzyme-bound,It is followed by a second charge-transfer intermediate of enzyme-bound NADP+ and reduced riboflavin.,12210
bursting We leakage due show effects. the translational noise have opposite and that transcriptional to,We show that transcriptional leakage and noise due to translational bursting have the opposite effects.,12211
"number derived. strict bound region a a the is facets Tukey on of Further, of","Further, a strict bound on the number of facets of a Tukey region is derived.",12212
a method enhancing the level This paper gray images. for presents,This paper presents a method for enhancing the gray level images.,12213
interpolation these The nodes using the calculated Bernstein of are functions polynomials.,The interpolation nodes of these functions are calculated using the Bernstein polynomials.,12214
algebraical this In images. a the paper model for new gray we level propose,In this paper we propose a new algebraical model for the gray level images.,12215
for image digital processing. can used It be,It can be used for digital image processing.,12216
"structure enhancement. in image the used defined be could Also,","Also, the defined structure could be used in image enhancement.",12217
results experimental The general image with presented approach enhancement. to demonstrate is,The general approach is presented with experimental results to demonstrate image enhancement.,12218
image in applications processing. and Background/foreground video segmentation a lot of has,Background/foreground segmentation has a lot of applications in image and video processing.,12219
The problems new are computational beauty reviewing neuroscience three and discussed teaching of by textbooks.,The problems and beauty of teaching computational neuroscience are discussed by reviewing three new textbooks.,12220
The for image offers model tools new logarithmic processing.,The logarithmic model offers new tools for image processing.,12221
presents images. method level gray for This paper enhancing the a,This paper presents a method for enhancing the gray level images.,12222
lines. graphic latter polygonal The a represented by have,The latter have a graphic represented by polygonal lines.,12223
measures an top-down in distortion for image important Therefore are processing. the local,Therefore local distortion measures in an image are important for the top-down processing.,12224
"perceptual for To this, global account features can be used.","To account for this, global perceptual features can be used.",12225
"approach thoroughly discussed, behind been by an accompanied has intuitive proposed analysis. motivation The the","The motivation behind the proposed approach has been thoroughly discussed, accompanied by an intuitive analysis.",12226
"to of concept task learning. Specifically, the","Specifically, to the task of concept learning.",12227
"for we image paper, In propose model mathematical processing. color a this","In this paper, we propose a mathematical model for color image processing.",12228
real two <+> addition define We multiplication <x>. scalar and operations:,We define two operations: addition <+> and real scalar multiplication <x>.,12229
operations With these of real becomes a vector colors space. the space,With these operations the space of colors becomes a real vector space.,12230
"Euclidean ||, we a (logarithmic) obtain space.","||, we obtain a (logarithmic) Euclidean space.",12231
A perceptual to evaluate the in approach sharpness blind is proposed. a present image natural,A blind approach to evaluate the perceptual sharpness present in a natural image is proposed.,12232
microtubules attachments detachments. with highly frequent and dynamic is kinetochores interaction The between and,The interaction between kinetochores and microtubules is highly dynamic with frequent attachments and detachments.,12233
theory. into web The model food present introduced parameterized realities,The present model introduced parameterized realities into food web theory.,12234
"broad studies However, values of theoretical predict a $b$. of dynamics of range population","However, theoretical studies of population dynamics predict a broad range of values of $b$.",12235
on our support results. analyses theoretical Empirical datasets two,Empirical analyses on two datasets support our theoretical results.,12236
established Proving of Theorem an (ATP) Automated Intelligence. branch is Artificial,Automated Theorem Proving (ATP) is an established branch of Artificial Intelligence.,12237
This is optimization the distribution. form the convex regardless of of target problem,This optimization problem is convex regardless of the form of the target distribution.,12238
with results detection We edge our classical comparison the present using operators. results of the,We present the comparison of our results with the results using classical edge detection operators.,12239
of many in cycle central Iron-peroxide are intermediates biomolecules. the reaction iron-containing,Iron-peroxide intermediates are central in the reaction cycle of many iron-containing biomolecules.,12240
to Then sets. is idea interval neutrosophic the extended,Then the idea is extended to interval neutrosophic sets.,12241
classification and supervised algorithms training. rely Current state-of-the-art on detection,Current state-of-the-art classification and detection algorithms rely on supervised training.,12242
by auto-encoder convolutional This to assumption regularized slowness and is sparsity. a train exploited pooling,This assumption is exploited to train a convolutional pooling auto-encoder regularized by slowness and sparsity.,12243
"feedback actionable we features tool personalized, with surgeons developed provide regarding motion. The their","The features we developed provide surgeons with personalized, actionable feedback regarding their tool motion.",12244
"unfortunately Second, time. which for this algorithm Hungarian cubic the calls usurps","Second, this calls for the Hungarian algorithm which unfortunately usurps cubic time.",12245
deeper CNNs advantages with compare We further different brought by architecture. show and depths the,We further compare CNNs with different depths and show the advantages brought by deeper architecture.,12246
"that improves solutions the the combination of the two show results. we Finally,","Finally, we show that the combination of the two solutions improves the results.",12247
are We results standard on report two datasets. state-of-the-art that equivalent the to,We report results that are equivalent to the state-of-the-art on two standard datasets.,12248
"the we study for In paper, segmentation semantic this of problem part animals.","In this paper, we study the problem of semantic part segmentation for animals.",12249
"model. semantic incorporate And the and into appearance, cues edge, we part compositional","And we incorporate edge, appearance, and semantic part cues into the compositional model.",12250
Experimental effectiveness the of results our demonstrate method.,Experimental results demonstrate the effectiveness of our method.,12251
"only Knowledge good it is complete. is if and consistent sound,","Knowledge is only good if it is sound, consistent and complete.",12252
"true its holds knowledge, holds knowledge same concepts about for which and association. The conceptual","The same holds true for conceptual knowledge, which holds knowledge about concepts and its association.",12253
"situations, In interval-valued practical frequently are encountered. fuzzy sets","In practical situations, interval-valued fuzzy sets are frequently encountered.",12254
"sets sets. paper, firstly, shadowed understanding interpreting interval fuzzy this present for and we In","In this paper, firstly, we present shadowed sets for interpreting and understanding interval fuzzy sets.",12255
"interval-valued we sets. construct approximations three-way of fuzzy errors-based Secondly,","Secondly, we construct errors-based three-way approximations of interval-valued fuzzy sets.",12256
"Thirdly, using functions. sets loss by three-way errors-based of compute interval-valued approximations interval-valued fuzzy we","Thirdly, we compute errors-based three-way approximations of interval-valued fuzzy sets by using interval-valued loss functions.",12257
is Bayesian challenging. large methods The fully variable data sets application to computationally selection of,The application of fully Bayesian variable selection methods to large data sets is computationally challenging.,12258
time. the data in measured Temporal information of context is,Temporal data is information measured in the context of time.,12259
longitudinal dependence. within we between and data In expect also subject,In longitudinal data we also expect within and between subject dependence.,12260
"series using plotted longitudinal analyzed often although differently, and displays. are data, Time similar","Time series and longitudinal data, although analyzed differently, are often plotted using similar displays.",12261
the methods of and allele genomes. among These archaic pattern sharing study modern,These methods study the pattern of allele sharing among modern and archaic genomes.,12262
of depends bias. and two factors: bias The admixture branch-length effect ghost population-size on,The effect of ghost admixture depends on two factors: branch-length bias and population-size bias.,12263
branch in Population-size differences in gene because genealogy. occurs size bias population distort the lengths,Population-size bias occurs because differences in population size distort branch lengths in the gene genealogy.,12264
"absence In admixture, of the small. these ghost are effects","In the absence of ghost admixture, these effects are small.",12265
"ghost presence in the of become admixture. however, They important,","They become important, however, in the presence of ghost admixture.",12266
differ the pattern in Estimators of response.,Estimators differ in the pattern of response.,12267
may estimator one deflate another. a given but parameter inflate Increasing,Increasing a given parameter may inflate one estimator but deflate another.,12268
"informative. this are estimators For among reason, comparisons","For this reason, comparisons among estimators are informative.",12269
It of identifies also Melanesia. an into inconsistency in archaic estimates admixture,It also identifies an inconsistency in estimates of archaic admixture into Melanesia.,12270
In this enhancement. model paper use we for a gray image logarithmic level,In this paper we use a logarithmic model for gray level image enhancement.,12271
important dynamics neural for are efficient to t These be functioning. thought,These dynamics are thought to be important t for efficient neural functioning.,12272
"for a new representation paper, we feature this present In videos. first-person","In this paper, we present a new feature representation for first-person videos.",12273
confirm first-person these findings. various datasets activity to Multiple tested were under settings,Multiple first-person activity datasets were tested under various settings to confirm these findings.,12274
"decidable. condition, under that run-boundedness we verification Specifically, a show stays","Specifically, we show that under a run-boundedness condition, verification stays decidable.",12275
Object computer of research important the in detection field vision. is an area,Object detection is an important research area in the field of computer vision.,12276
"appearance imaging assumptions detector the object of conditions. on object relies specific and However, each","However, each object detector relies on specific assumptions of the object appearance and imaging conditions.",12277
"can universal. no As consequence, as algorithm a considered be","As a consequence, no algorithm can be considered as universal.",12278
"propose combine to to paper, we object learn In framework this how detectors. a","In this paper, we propose a framework to learn how to combine object detectors.",12279
"the these predictive concept of affects drift application in where future. maintenance), models","predictive maintenance), where concept drift affects application of these models in the future.",12280
error-prone is construction models such an context of elaborate task. Manual and,Manual construction of such context models is an elaborate and error-prone task.,12281
of scenario need example and Our information. context tracking the change shows distribution for effective,Our example scenario shows the need for distribution and effective change tracking of context information.,12282
In results implementation. present first this the CKR such evaluation paper we for,In this paper we present the first evaluation results for such CKR implementation.,12283
"experiment scalability In study respect different first its to reasoning regimes. particular, with in we","In particular, in first experiment we study its scalability with respect to different reasoning regimes.",12284
are recognition techniques used for English for documents language. printed widely Character,Character recognition techniques for printed documents are widely used for English language.,12285
"been to accuracy Sinhala attention given improve little character of a recognition. has Therefore, the","Therefore, a little attention has been given to improve the accuracy of Sinhala character recognition.",12286
inference a paper This describes recognize characters. use Sinhala fuzzy to system the of,This paper describes the use of a fuzzy inference system to recognize Sinhala characters.,12287
used iterative is convex smoothing proximal to corresponding optimization algorithm the A solve problems.,A proximal iterative smoothing algorithm is used to solve the corresponding convex optimization problems.,12288
"intervals the develop inverse for confidence Furthermore, elements we matrix. of point-wise covariance conditional individual","Furthermore, we develop point-wise confidence intervals for individual elements of the conditional inverse covariance matrix.",12289
"natural addressed. problem texture of synthetic In paper, images is the this classifying and","In this paper, the problem of classifying synthetic and natural texture images is addressed.",12290
and was databases. Vistex method tested the on Brodatz well-known The,The method was tested on the well-known Brodatz and Vistex databases.,12291
indicates the employing natural potential in other methods on phenomena of inspired fields. This,This indicates the potential of employing methods inspired on natural phenomena in other fields.,12292
How students are them we do learning the learn? what we to know want,How do we know the students are learning what we want them to learn?,12293
over the time? we improve program How do,How do we improve the program over time?,12294
they with examples illustrative as programs for moving new guidelines. serve the hope We forward,We hope they serve as illustrative examples for programs moving forward with the new guidelines.,12295
Examples as basic usage illustrate performance. improving as well for methods,Examples illustrate basic usage as well as methods for improving performance.,12296
low-level typically in features eye prediction integrating researches (e.g. on fixation While rely previous,While previous researches in eye fixation prediction typically rely on integrating low-level features (e.g.,12297
Forage used were: cv. decumbens} crops \textit{Brachiaria,Forage crops used were: \textit{Brachiaria decumbens} cv.,12298
Brizantha} \textit{Brachiaria (Hochst. Germain \textit{Brachiaria IPEAN; ruziziensis} \& Evrard;,IPEAN; \textit{Brachiaria ruziziensis} Germain \& Evrard; \textit{Brachiaria Brizantha} (Hochst.,12299
algorithms present reliable yet to calibrate and the do proposed We the inference. system simple,We present simple yet reliable algorithms to calibrate the proposed system and do the inference.,12300
in low-dimensional often to lie different belong corresponding they subspaces High-dimensional to. classes data,High-dimensional data often lie in low-dimensional subspaces corresponding to different classes they belong to.,12301
an vision. role in plays computer important Texture,Texture plays an important role in computer vision.,12302
"for analysis. novel a texture complexity characterization, approach based paper This on presents","This paper presents a novel approach for texture characterization, based on complexity analysis.",12303
the images the from method An album Brodatz using experiment illustrates performance.,An experiment using images from the Brodatz album illustrates the method performance.,12304
classification. focused work the on scenarios binary Most of of those were,Most of those work were focused on the scenarios of binary classification.,12305
"of linear preceding individually researches the presence data, sparse recommended analysis(LDA). multi-class In pairwise discriminant","In the presence of multi-class data, preceding researches recommended individually pairwise sparse linear discriminant analysis(LDA).",12306
properties It among enjoys features. allows which non-asymptotic appealing insignificant correlations,It enjoys appealing non-asymptotic properties which allows insignificant correlations among features.,12307
This exhibits superior estimator both data. capability simulated real and on,This estimator exhibits superior capability on both simulated and real data.,12308
morphology. semitic is a complex Arabic by a characterized language rich and,Arabic is a semitic language characterized by a complex and rich morphology.,12309
for online Several building handwriting research have systems. efforts recognition Arabic devoted been,Several research efforts have been devoted for building online Arabic handwriting recognition systems.,12310
"Arabic paper, online In this database. we large AltecOnDB, scale handwriting a introduce","In this paper, we introduce AltecOnDB, a large scale online Arabic handwriting database.",12311
include and complete The digits punctuation marks. that samples sentences are collected,The collected samples are complete sentences that include digits and punctuation marks.,12312
and tools phases. to verification verification are and facilitate Annotation annotation developed the,Annotation and verification tools are developed to facilitate the annotation and verification phases.,12313
considerable the promise and to geometry topology. between gap persistence bears bridge Multidimensional,Multidimensional persistence bears considerable promise to bridge the gap between geometry and topology.,12314
"However, practical a has robust construction and been challenge. its","However, its practical and robust construction has been a challenge.",12315
"introduce families multiscale namely two multidimensional We and persistence. persistence of multidimensional pseudo-multidimensional persistence,","We introduce two families of multidimensional persistence, namely pseudo-multidimensional persistence and multiscale multidimensional persistence.",12316
unfolded between proteins multidimensional folded observed partial transition has persistence. and been Topological in,Topological transition between partial folded and unfolded proteins has been observed in multidimensional persistence.,12317
signals The smart sensors observing from come a environment.,The signals come from sensors observing a smart environment.,12318
central a biochemistry. Allosteric of proteins structural in and problem communication yet unsolved is,Allosteric communication in proteins is a central and yet unsolved problem of structural biochemistry.,12319
based on a proposes fractal descriptor work This theory. texture,This work proposes a texture descriptor based on fractal theory.,12320
descriptors. The the based on is method Bouligand-Minkowski,The method is based on the Bouligand-Minkowski descriptors.,12321
"average from and features extract entropy Thus, we both deviation.","Thus, we extract entropy features from both average and deviation.",12322
the of concatenation measures. by proposed provided are descriptors such The,The proposed descriptors are provided by the concatenation of such measures.,12323
process. renewal point neuron A fed with is a,A neuron is fed with a point renewal process.,12324
another three given. provided to the calculate This any pdfs are of the allows two,This allows to calculate any of the three pdfs provided the another two are given.,12325
corresponding is derived. between Similar relation variances and means,Similar relation between corresponding means and variances is derived.,12326
model. for relations binding neuron are exactly The checked the,The relations are checked exactly for the binding neuron model.,12327
"propose we novel texture for obtaining a gray-level images. work, technique this of In descriptors","In this work, we propose a novel technique for obtaining descriptors of gray-level texture images.",12328
be detection as approach object the that used general purpose can toolbox. We a believe,We believe that the approach can be used as a general purpose object detection toolbox.,12329
"car achieved and We dataset, on algorithm UIUC tested the an result. outstanding","We tested the algorithm on UIUC car dataset, and achieved an outstanding result.",12330
is such concerned shape situations. The the this challenging problem extraction work in in,The problem concerned in this work is the shape extraction in such challenging situations.,12331
shape alignment We approach and through extraction recovery. the shape,We approach the shape extraction through shape alignment and recovery.,12332
contour Our shape energies. active idea deformation in to model of minimizing use general is,Our idea is to use general model of shape deformation in minimizing active contour energies.,12333
evolution evolution variation the (PVSE). The shape curve called is prior,The curve evolution is called the prior variation shape evolution (PVSE).,12334
derive for active minimizing energy-minimizing energies. also the contour We PVSE,We also derive the energy-minimizing PVSE for minimizing active contour energies.,12335
"choosing PVSE, theory For shape a of preservability of PVSE established. shape-preserving such the is","For choosing such shape-preserving PVSE, a theory of shape preservability of the PVSE is established.",12336
the the contain five at were presented papers and notes which accepted workshop. These,These notes contain the five papers which were accepted and presented at the workshop.,12337
can hypothesis confidence be conduct it construct tests. intervals to Hence and used valid,Hence it can be used to construct valid confidence intervals and conduct hypothesis tests.,12338
"with desired convergence rate and it first-stage any allows estimator sparsity. for Instead, empirical","Instead, it allows for any first-stage estimator with desired convergence rate and empirical sparsity.",12339
"results the verify theory. Finally, empirical provide we to","Finally, we provide empirical results to verify the theory.",12340
Synchronized cortical to were in ex-vivo scale large responses stimuli network studied networks.,Synchronized network responses to stimuli were studied in ex-vivo large scale cortical networks.,12341
"as of responsiveness in, Network the or residency plane. is amongst, shifting areas latency-decay reflected","Network responsiveness is reflected as residency in, or shifting amongst, areas of the latency-decay plane.",12342
for dimensional dimensional models. the high assessment We uncertainty consider of in problem components low,We consider the problem of uncertainty assessment for low dimensional components in high dimensional models.,12343
both consider and for M-estimators. We confidence hypothesis generic penalized tests regions,We consider both hypothesis tests and confidence regions for generic penalized M-estimators.,12344
broaden framework this applications. generalize its also to We,We also generalize this framework to broaden its applications.,12345
"model theory for we establish Second, misspecification. the","Second, we establish the theory for model misspecification.",12346
back theoretical numerical the are Thorough developed up conducted to studies results.,Thorough numerical studies are conducted to back up the developed theoretical results.,12347
a to back numerical our example are real up data Thorough provided results. simulations and,Thorough numerical simulations and a real data example are provided to back up our results.,12348
spurred of improving the research This has the accuracy shortcoming into localization.,This shortcoming has spurred the research into improving the accuracy of localization.,12349
"out Until now, been and investigations carried of independently. scientific however, memory have humanistic","Until now, however, humanistic and scientific investigations of memory have been carried out independently.",12350
process. and overview a provide stimulating perspectives these diverse memory Together the broad of,Together these diverse perspectives provide a broad and stimulating overview of the memory process.,12351
regime. matrix have guarantees Recently sampling for the non-uniform completion been obtained in theoretical,Recently theoretical guarantees have been obtained for matrix completion in the non-uniform sampling regime.,12352
To deviation and of large employ measure a a bound concentration this bound. accomplish we,To accomplish this we employ a concentration of measure bound and a large deviation bound.,12353
"E. D. In Sontag, Biophys. [J. and Barton P.","In [J. P. Barton and E. D. Sontag, Biophys.",12354
used order recognition improve in been rate. to techniques Various decision have the fusion,Various decision fusion techniques have been used in order to improve the recognition rate.,12355
"Bosphorus databases. proposed FERET, has on system and The been HP, face the tested","The proposed system has been tested on the FERET, HP, and Bosphorus face databases.",12356
compared The and proposed conventional is system techniques. thestate-of-the-art with,The proposed system is compared with conventional and thestate-of-the-art techniques.,12357
results. Our by is thorough theory numerical supported,Our theory is supported by thorough numerical results.,12358
for solution We numerical problem. procedures discuss of the our efficient,We discuss efficient numerical procedures for the solution of our problem.,12359
"been missing. but general framework a literature, still have Many in constraints is the proposed","Many constraints have been proposed in the literature, but a general framework is still missing.",12360
We investigate this constraint programming use as framework for task. general of the,We investigate the use of constraint programming as general framework for this task.,12361
applicable sequence are four that categories mining. constraints We first of identify to,We first identify four categories of constraints that are applicable to sequence mining.,12362
propose programming two then We formulations. constraint,We then propose two constraint programming formulations.,12363
called new first global constraint exists-embedding. The introduces a formulation,The first formulation introduces a new global constraint called exists-embedding.,12364
formulation constraint. one support of type efficient the is most not This but does,This formulation is the most efficient but does not support one type of constraint.,12365
database specialised formulations can technique in use used algorithms. projected Both the,Both formulations can use the projected database technique used in specialised algorithms.,12366
flexibility the existing and to compare approach Experiments methods. towards constraint-based settings demonstrate the,Experiments demonstrate the flexibility towards constraint-based settings and compare the approach to existing methods.,12367
the Gene co-located on chromosome. genes Motivation: blocks are,Motivation: Gene blocks are genes co-located on the chromosome.,12368
of is common Consecutive behavioral actions repetition sequences. in,Consecutive repetition of actions is common in behavioral sequences.,12369
to feedback sequences how in auditory Here songbirds. contributes we study repetitive syllable generating,Here we study how auditory feedback contributes to generating repetitive syllable sequences in songbirds.,12370
distributions repeat models mechanism song. in that this Computational finch explains show Bengalese observed,Computational models show that this mechanism explains repeat distributions observed in Bengalese finch song.,12371
"connectomes, Brain widely or today. being graphs, are explored","Brain graphs, or connectomes, are being widely explored today.",12372
"flow trajectories that instantaneous For we movements. vectors down indicate into first the phase, break","For the first phase, we break down trajectories into flow vectors that indicate instantaneous movements.",12373
"terms reachability set and of in Next, similarity. components' we create motion proximity spatial motion","Next, we create motion components' reachability set in terms of spatial proximity and motion similarity.",12374
bilateral reducing the filter. begins specular and reflection a The small using method noise with,The method begins with reducing the specular reflection and small noise using a bilateral filter.,12375
highest technique accuracy. shows achieves The experiment that our result,The experiment result shows that our technique achieves highest accuracy.,12376
"a estimating correspondences across We method propose scenes. for accurate matching fast, dense pixel","We propose a fast, accurate matching method for estimating dense pixel correspondences across scenes.",12377
match- are into seamlessly The embedded features multi-layer learned a framework. ing then,The learned features are then seamlessly embedded into a multi-layer match- ing framework.,12378
in fast presented algorithm A feature selection paper. is forward this,A fast forward feature selection algorithm is presented in this paper.,12379
classifier. Gaussian a It based mixture model on is (GMM),It is based on a Gaussian mixture model (GMM) classifier.,12380
GMM used images. hyperspectral for are classifying,GMM are used for classifying hyperspectral images.,12381
The maximizes estimation spectral classification that an iteratively rate. of features selects algorithm the,The algorithm selects iteratively spectral features that maximizes an estimation of the classification rate.,12382
is using validation. k-fold The the estimation cross done,The estimation is done using the k-fold cross validation.,12383
"fast proposed. efficient order implementation to time, is terms in In of an perform computing","In order to perform fast in terms of computing time, an efficient implementation is proposed.",12384
"the Furthermore, contains channels. model extracted very few spectral","Furthermore, the extracted model contains very few spectral channels.",12385
"\textit{Group algorithm propose simple To the effective problem, we $K$-Means}. yet a solve","To solve the problem, we propose a simple yet effective algorithm \textit{Group $K$-Means}.",12386
"hierarchical a non-convex we propose for problem. such Besides, a initialization","Besides, we propose a hierarchical initialization for such a non-convex problem.",12387
effectiveness Experimental the results the of validate approach. well,Experimental results well validate the effectiveness of the approach.,12388
such the achieved penalizations. the by improvements data-driven illustrate experiments of use Numerical significant,Numerical experiments illustrate the significant improvements achieved by the use of such data-driven penalizations.,12389
detection most challenges the vision. Object of in computer important is one,Object detection is one of the most important challenges in computer vision.,12390
are trained Object usually on images. from detectors bounding-boxes still,Object detectors are usually trained on bounding-boxes from still images.,12391
"Recently, an used data. as video source alternative has been of","Recently, video has been used as an alternative source of data.",12392
"behind examine reasons we the performance In gap. this this paper,","In this paper, we examine the reasons behind this performance gap.",12393
We location for a propose in ImageNet. method annotating objects of the,We propose a method for annotating the location of objects in ImageNet.,12394
quantum fully that induction argue is to applicable Solomonoff mechanics. also We,We also argue that Solomonoff induction is fully applicable to quantum mechanics.,12395
"distributions constant, When is held confirming produce of neighborhood the similar size patterns isolation-by-distance, predictions.","When neighborhood size is held constant, the distributions produce similar patterns of isolation-by-distance, confirming predictions.",12396
underlying relationships. Conscious experience awash is with,Conscious experience is awash with underlying relationships.,12397
probability bias the define Representing can expected shows that quantities. this system distribution a using,Representing this bias using a probability distribution shows that the system can define expected quantities.,12398
"defines the parameters the efe, minimise the Under that requirement brain relationship relationships.","Under the requirement that the relationship parameters minimise efe, the brain defines relationships.",12399
approximation measurements. tracking with detection-based This facilitates also efficient multi-sensor,This approximation also facilitates efficient multi-sensor tracking with detection-based measurements.,12400
are presented approach. results verify proposed to Simulation the,Simulation results are presented to verify the proposed approach.,12401
"Big shapes, sizes. comes in ways, data forms types, various and","Big data comes in various ways, types, shapes, forms and sizes.",12402
is vision. problem tracking standing a in long Object,Object tracking is a long standing problem in vision.,12403
"visual object tracking. motivated, in Thus this adapt paper for we propose objectness to","Thus motivated, in this paper we propose to adapt objectness for visual object tracking.",12404
"named benefit ADOBING, We the is objectness, emphasize generic. of that proposed the adaptive","We emphasize that the benefit of the proposed adaptive objectness, named ADOBING, is generic.",12405
"in evaluations. combine this, recent with To show trackers we seven performed ADOBING top","To show this, we combine ADOBING with seven top performed trackers in recent evaluations.",12406
"methods better obtained suggested Local results in than recent global have years, methods.","Local methods suggested in recent years, have obtained better results than global methods.",12407
"parts should data considered be for The first, the clustering? of which","The first, which parts of the data should be considered for clustering?",12408
"de-noising. is method and for what suitable clustering the second, data","and the second, what data clustering method is suitable for de-noising.",12409
learn to over exploited Then is an dictionary. complete clustering,Then clustering is exploited to learn an over complete dictionary.,12410
the problem. to efficient also is proposed An solve algorithm WSNM,An efficient algorithm is also proposed to solve the WSNM problem.,12411
"freely package R by available the procedure ""poweRlaw"". analytical software was implemented The","The analytical procedure was implemented by the freely available R software package ""poweRlaw"".",12412
"of exploitation the calculations Computing of by was High-Throughput Time technology, considerably a.k.a. shorten","Time of calculations was considerably shorten by the exploitation of High-Throughput Computing technology, a.k.a.",12413
represented. of however is widely Log-normal distribution ISI,Log-normal distribution of ISI is however widely represented.,12414
stage log-normal The distributions SE of final and ISI. only by appears dominated power-law of,The final stage of SE appears dominated only by power-law and log-normal distributions of ISI.,12415
element is boundary moving problem finite The model adaptive method. the solved by,The moving boundary model problem is solved by the adaptive finite element method.,12416
obtained of low explain The previous improvements ions models availability. would failures of any for,The obtained improvements would explain any failures of previous models for ions of low availability.,12417
"simpler burden. our Therefore, could to low its be due a computational model alternative","Therefore, our model could be a simpler alternative due to its low computational burden.",12418
patterns these mark changes). of covalent can across generations (epigenetic Some transmitted be,Some of these covalent mark patterns can be transmitted across generations (epigenetic changes).,12419
be maintained memory. a stable metabolic dynamics Hopfield-like are and can long-term The as biochemical,The Hopfield-like metabolic dynamics are stable and can be maintained as a long-term biochemical memory.,12420
manifestation structural processes Epigenetic metabolic this of cellular memory. the be appear to,Epigenetic processes appear to be the structural manifestation of this cellular metabolic memory.,12421
noiseless show noisy and settings. in improved both performance experiments Our,Our experiments show improved performance in both noiseless and noisy settings.,12422
FDB the method and implementation of made The are publicly benchmark available. the,The benchmark and the implementation of the FDB method are made publicly available.,12423
iron a superoxide SOR reductase radical that microorganisms. detoxifies metalloenzyme non-heme in Superoxide is,Superoxide reductase SOR is a non-heme iron metalloenzyme that detoxifies superoxide radical in microorganisms.,12424
"available of context-less, interaction protein-protein networks the majority Nevertheless, are i.e.","Nevertheless, the majority of available protein-protein interaction networks are context-less, i.e.",12425
bioinformatic Conclusion: of analyses. biological interactomes confirms work that the contextualizing significance improves Our,Conclusion: Our work confirms that contextualizing interactomes improves the biological significance of bioinformatic analyses.,12426
"noisy gives it scores similar types. different to of That relationship equally is,","That is, it gives similar scores to equally noisy relationship of different types.",12427
"oxidation. There, oxygen molecular flavin the activates substrate reduced for","There, the reduced flavin activates molecular oxygen for substrate oxidation.",12428
vision work parallel subsystem subsystem and formation control to accomplish in The control.,The vision subsystem and control subsystem work in parallel to accomplish formation control.,12429
is for framework. being the follower in the leader time first proposed This,This is being proposed for the first time in the leader follower framework.,12430
simulation in results is practical in demonstrated the and experiments. The performance of,The performance of the results is demonstrated in simulation and in practical experiments.,12431
is for System real-time Short-Messaging-Service system monitoring and any Auto-Monitoring critical operational environments. a,Auto-Monitoring and Short-Messaging-Service System is a real-time monitoring system for any critical operational environments.,12432
bridge the user. is to monitoring GSM system network between communication and A utilized,A GSM network is utilized to bridge the communication between monitoring system and user.,12433
This details the prototyping presents paper system. of,This paper presents details of prototyping the system.,12434
a Experiments on Mammographic images. held real,Experiments held on a real Mammographic images.,12435
classifiers. and classical other high state-of-art to SOM Results showed compared accuracy,Results showed high accuracy compared to classical SOM and other state-of-art classifiers.,12436
adaptation. rapid selection results may driving be used Our to polygenic understand,Our results may be used to understand polygenic selection driving rapid adaptation.,12437
disorder. The system basin lake as modeled quenched is,The lake basin system is modeled as quenched disorder.,12438
time Linear extinction estimate of is the computed.,Linear estimate of the extinction time is computed.,12439
computational Detecting careful photometric the and modeling message decoding recovery. message for requires,Detecting and decoding the message requires careful photometric modeling for computational message recovery.,12440
networks. interaction eukaryotic the model evolution protein-protein We (PPI) of,We model the evolution of eukaryotic protein-protein interaction (PPI) networks.,12441
evolution. present-day into in way which biological provide data insights suggests This could proteomics a,This suggests a way in which present-day proteomics data could provide insights into biological evolution.,12442
"the Often, results. industrious author's or cannot validate upon expand reader the","Often, the industrious reader cannot expand upon or validate the author's results.",12443
to upon results struggle own Even revisiting reproduce may them. their author the,Even the author may struggle to reproduce their own results upon revisiting them.,12444
"and In entry, we this these quantities. define introduce","In this entry, we introduce and define these quantities.",12445
"existing categorical are interpret. that parameters have to unordered models difficult Furthermore,","Furthermore, existing unordered categorical models have parameters that are difficult to interpret.",12446
"through analysing is e.g.video accomplished surveillance features. tasks, imagerybased visual Many summarisation, conventionally","Many visual surveillance tasks, e.g.video summarisation, is conventionally accomplished through analysing imagerybased features.",12447
"of supporting Science software the collaboration, reproduction, depends tools. development on and result","Science depends on collaboration, result reproduction, and the development of supporting software tools.",12448
these versions. management of requires of careful software Each,Each of these requires careful management of software versions.,12449
"equation derived is for Here, map proteins. stiffness gives that the an","Here, an equation is derived that gives the stiffness map for proteins.",12450
"we effectiveness the of demonstrate algorithms. examples several give to Finally, our","Finally, we give several examples to demonstrate the effectiveness of our algorithms.",12451
"state-of-the-art Image, using image deep developed learning. system, We a present end-to-end recognition Deep","We present a state-of-the-art image recognition system, Deep Image, developed using end-to-end deep learning.",12452
challenging Our method benchmarks. excellent vision on results multiple achieves computer,Our method achieves excellent results on multiple challenging computer vision benchmarks.,12453
this show function a polytope. on convex that We is enzyme cost,We show that enzyme cost is a convex function on this polytope.,12454
show on distributions. results simulation these as well We as some study theoretical,We show some theoretical results as well as simulation study on these distributions.,12455
not to this I be the show case.,I show this not to be the case.,12456
This registration with area-based subpixel image under deals paper hypothesis. rotation-isometric scaling-translation transformation,This paper deals with area-based subpixel image registration under rotation-isometric scaling-translation transformation hypothesis.,12457
"fBm The fully are satisfied, assumptions the underlying when model benefits observed e.g. are main","The main benefits are observed when assumptions underlying the fBm model are fully satisfied, e.g.",12458
for normally stationary textures distributed increments. with isotropic,for isotropic normally distributed textures with stationary increments.,12459
Schwarzkopf our data. hope of and understand responses our We these others will help interpretation,We hope our responses will help Schwarzkopf and others understand our interpretation of these data.,12460
a We Analysis develop recency Factors account. into new that Recent-Performance data takes model,We develop a new Recent-Performance Factors Analysis model that takes data recency into account.,12461
"models impulse data noise or bursts sparse outliers, loss. Typically,","Typically, sparse noise models outliers, impulse bursts or data loss.",12462
combined matrix. of precision a diagonal is The modeled noise by,The precision of combined noise is modeled by a diagonal matrix.,12463
"is the be Further, modeling computationally efficient. to combined found more","Further, the combined modeling is found to be computationally more efficient.",12464
are conducted the Experiments isolated for handwritten of recognition automatic strokes.,Experiments are conducted for the automatic recognition of isolated handwritten strokes.,12465
lanes biggest curvy the accidents blind and of road One reasons is for turns.,One of the biggest reasons for road accidents is curvy lanes and blind turns.,12466
"shapes can ellipse, using predefined like detected available presently circle, etc, few Only functions. be","Only few predefined shapes like circle, ellipse, etc, can be detected using presently available functions.",12467
are the two between The behavioral distributed reports possible choices. that derivation assumed evenly,The derivation assumed that behavioral reports are distributed evenly between the two possible choices.,12468
"the data ideal, deviates from scenario. the zero-signal, unbiased as","as the data deviates from the ideal, zero-signal, unbiased scenario.",12469
and made. This result classical a is no claim is to originality well-known,This is a well-known classical result and no claim to originality is made.,12470
noise? subject coupling regulatory gene Can spatial for such of be to signals benefit conveying,Can such spatial coupling be of benefit for conveying signals subject to gene regulatory noise?,12471
extended extend we Here to gene framework regulation systems. our for information-theoretic spatially,Here we extend our information-theoretic framework for gene regulation to spatially extended systems.,12472
are models expression Stochastic able to of fluctuations. kinetic describe protein genetic,Stochastic kinetic models of genetic expression are able to describe protein fluctuations.,12473
of of the Carlo in reducing applications. variance A estimators quasi-Monte discussion on possibility the,A discussion on the possibility of reducing the variance of quasi-Monte Carlo estimators in applications.,12474
"provided Further Carlo"". Reduction are for paper in Quasi-Monte details the ""Variance accompanying","Further details are provided in the accompanying paper ""Variance Reduction for Quasi-Monte Carlo"".",12475
"least, et but of al. a debiasing as not result Last Cheng","Last but not least, as a result of debiasing Cheng et al.",12476
with considered. blocks threshold Within are than a this entropies domain approach only greater,Within this approach only domain blocks with entropies greater than a threshold are considered.,12477
images. algorithm for examined been has well-known The some,The algorithm has been examined for some well-known images.,12478
"To classifier hyper-parameters and this, privately. accuracy best for releasing address the methods we introduce","To address this, we introduce methods for releasing the best hyper-parameters and classifier accuracy privately.",12479
help we these that review theory empirical clarify examples and conflicts. Here,Here we review theory and empirical examples that help clarify these conflicts.,12480
"problem is describe procedure. formulation, from a multi-stage INRC-I, which, differently we the First,","First, we describe the problem formulation, which, differently from INRC-I, is a multi-stage procedure.",12481
"rules the Finally, the we competition. of state","Finally, we state the rules of the competition.",12482
genomics. the Sequence modern mapping cornerstone Motivation: is of,Motivation: Sequence mapping is the cornerstone of modern genomics.,12483
"insufficiently However, algorithms are existing sequence general. mapping most","However, most existing sequence mapping algorithms are insufficiently general.",12484
mapped will not length) mapping read positions. the previously of increasing alter by,by increasing read length) will not alter the mapping of previously mapped positions.,12485
schemes Context general several in are senses.,Context schemes are general in several senses.,12486
"to reference. novel of rearrangements detection complex, natively relative the support arbitrary They the","They natively support the detection of arbitrary complex, novel rearrangements relative to the reference.",12487
scale length. of magnitude query can orders They in over sequence,They can scale over orders of magnitude in query sequence length.,12488
proposed The multi-channel to contrast phase is angiogram. also extended method,The proposed method is also extended to multi-channel phase contrast angiogram.,12489
"costs, a inconvenience discomfort. Ultimately, such might method reduce patient and healthcare","Ultimately, such a method might reduce healthcare costs, patient inconvenience and discomfort.",12490
numerical integration solution to is approximation. functional with trade-off One,One solution is to trade-off numerical integration with functional approximation.,12491
shown herein well-suited to statistical be This and computation. strategy modern explored to is,This strategy is explored herein and shown to be well-suited to modern statistical computation.,12492
"reaction This condition the of boundary for analysis enzymes new treatment ""imperfect"" with rates. allows","This new boundary condition treatment allows for the analysis of enzymes with ""imperfect"" reaction rates.",12493
discrete triangulated surface are geometry terms differential All explained for meshes.,All discrete differential geometry terms are explained for triangulated surface meshes.,12494
These for utilities methods. serve as basis feature line the,These utilities serve as basis for the feature line methods.,12495
method. re-implement line knowledge with provide feature the every reader We to all,We provide the reader with all knowledge to re-implement every feature line method.,12496
"restricted work medical models. to, is by, and not surface but Our motivated biological","Our work is motivated by, but not restricted to, medical and biological surface models.",12497
of classically for The centrality has as a data. robust multivariate used been estimator EM,The EM has classically been used as a robust estimator of centrality for multivariate data.,12498
using novel binary We strings. describing means image local a appearances present of,We present a novel means of describing local image appearances using binary strings.,12499
"this Learned end LATCH. Three or Patch Codes descriptors, propose To of we the Arrangements","To this end we propose the Learned Arrangements of Three Patch Codes descriptors, or LATCH.",12500
results  symmetric produce methods  part on bilaterally mirror Do localization object images?,  Do object part localization methods produce bilaterally symmetric results on mirror images?,12501
into look paper issue. closer we In a this this take,In this paper we take a closer look into this issue.,12502
of both theoretically upper our experimentally. show We tightness bounds and,We show tightness of our upper bounds both theoretically and experimentally.,12503
reasoning. The a did device develop brain not dedicated for,The brain did not develop a dedicated device for reasoning.,12504
problem regression is as a This representation the a of from formulated holistic image.,This is formulated as a regression problem from a holistic representation of the image.,12505
"are desired. much Thus, inference methods reconstruction-free for","Thus, reconstruction-free methods for inference are much desired.",12506
are This compressive for measurements. search to us preserved properties that such leads in,This leads us to search for such properties that are preserved in compressive measurements.,12507
"achieved results of Recently, approach deep various learning promising computer vision. fields in has","Recently, deep learning approach has achieved promising results in various fields of computer vision.",12508
simulated and We the experiments sensor). on results report experiments (using both physical Kinect,We report results on both simulated experiments and physical experiments (using the Kinect sensor).,12509
"ground depth quantitative truth, results, a demonstrate Our against accuracy. in evaluated improvement","Our results, evaluated against ground truth, demonstrate a quantitative improvement in depth accuracy.",12510
"performance. impacts on the recognition report how In we observations our data big this paper,","In this paper, we report our observations on how big data impacts the recognition performance.",12511
"security in a Furthermore, performance scenario. certification report real-world the we","Furthermore, we report the performance in a real-world security certification scenario.",12512
clear between exists human There performance. gap recognition and still machine a,There still exists a clear gap between machine recognition and human performance.,12513
ahead experiments three present summarize and our recent lying challenges recognition. in We face,We summarize our experiments and present three challenges lying ahead in recent face recognition.,12514
these towards indicate solutions possible we challenges. And several,And we indicate several possible solutions towards these challenges.,12515
the Sion's characterize adopted to condition. theorem saddle is value minimax,Sion's minimax theorem is adopted to characterize the saddle value condition.,12516
"the likelihood rule Least are favorable ratio decision and test robust derived. distributions, the robust","Least favorable distributions, the robust decision rule and the robust likelihood ratio test are derived.",12517
evaluate derivations. the and The exemplify theoretical simulations performed,The simulations performed evaluate and exemplify the theoretical derivations.,12518
"proposed In a hybrid generative-descriptive the model. using vocabulary learning is approach, performed","In the proposed approach, vocabulary learning is performed using a hybrid generative-descriptive model.",12519
"Clustering Entropy statistical a relationships parts algorithm. using are Conditional Minimum learned First, between","First, statistical relationships between parts are learned using a Minimum Conditional Entropy Clustering algorithm.",12520
"representation part substructures. internal compressing Finally, the with discovered by are data compositions constructed","Finally, part compositions are constructed by compressing the internal data representation with discovered substructures.",12521
"better methods. than performance retrieval CHOP shape state-of-the-art Additionally, provides retrieval shape the","Additionally, CHOP provides better shape retrieval performance than the state-of-the-art shape retrieval methods.",12522
widely and theory function. Task regarding current cited is difficulty fronto-parietal in cognitive control,Task difficulty is widely cited in current theory regarding cognitive control and fronto-parietal function.,12523
for of recognition vessels valuable in numerous transparent material is Visual applications. boundaries,Visual recognition of material boundaries in transparent vessels is valuable for numerous applications.,12524
boundary. properties image material were for Several indicators the examined as,Several image properties were examined as indicators for the material boundary.,12525
The boundary found curve algorithm. using Dijkstra's was optimal,The optimal boundary curve was found using Dijkstra's algorithm.,12526
"particularly with esotropia e.g. the multisystem infantile ones syndrome), disorders, associated","infantile esotropia syndrome), particularly the ones associated with multisystem disorders, e.g.",12527
paper the to of the part The package. explains use main how,The main part of the paper explains how to use the package.,12528
"in illustrate real we functionality and package's artificial both examples. the Furthermore,","Furthermore, we illustrate the package's functionality in both real and artificial examples.",12529
"visual literature, very BOW has the drawbacks. Although the influential in model distinct traditional two","Although very influential in the literature, the traditional visual BOW model has two distinct drawbacks.",12530
instead We to the of robustness focus noises. the model on non-Gaussian,We instead focus on the robustness of the model to non-Gaussian noises.,12531
rows/features. Also further we the for method handling data with outlier extend the,Also we further extend the method for handling the data with outlier rows/features.,12532
the studies problem. subspace This segmentation paper,This paper studies the subspace segmentation problem.,12533
The is method clustering framework. used as spectral the,The spectral clustering method is used as the framework.,12534
LSR. can a regarded which adaptively be as and It method balances SSC,It can be regarded as a method which adaptively balances SSC and LSR.,12535
CASS. and experimental effectiveness theoretical show the Both results of,Both theoretical and experimental results show the effectiveness of CASS.,12536
similarities for seems claim reasonable not but others. adaptive some This for,This claim seems reasonable for some adaptive similarities but not for others.,12537
biology. evolutionary to evolution and modern theory Darwin's both results apply of to Our,Our results apply both to Darwin's theory of evolution and to modern evolutionary biology.,12538
consider the setting linear We regression in dimension. of high,We consider the setting of linear regression in high dimension.,12539
"equations, Lotka-Volterra the matrix. necessitates of interaction nonzero the For coexistence determinant sustainable generalized","For the generalized Lotka-Volterra equations, sustainable coexistence necessitates nonzero determinant of the interaction matrix.",12540
Speech to  automatically.  predict words spoken Recognition searches the,  Speech Recognition searches to predict the spoken words automatically.,12541
"will that minimizes interesting. very building the a be Hence, the recognizer model cost of","Hence, building a model that minimizes the cost of the recognizer will be very interesting.",12542
"the HMM the greatly minimize using of can cost recognizer these Consequently, systems. belief","Consequently, using the belief HMM recognizer can greatly minimize the cost of these systems.",12543
"compared to recent convnets their pedestrian underperform Despite successes, diverse historically detectors. other","Despite their recent diverse successes, convnets historically underperform compared to other pedestrian detectors.",12544
(e.g. network We into explicitly problem modelling omit the deliberately the,We deliberately omit explicitly modelling the problem into the network (e.g.,12545
the present on Caltech convnet dataset. best and We KITTI detectors the,We present the best convnet detectors on the Caltech and KITTI dataset.,12546
experimentally filter and on this a framework different explore unifying observation we families. propose Based,Based on this observation we propose a unifying framework and experimentally explore different filter families.,12547
analysis. a systematic report results We enabling extensive,We report extensive results enabling a systematic analysis.,12548
"solved is commonly most by heuristics. this complexity, problem Due to its","Due to its complexity, this problem is most commonly solved by heuristics.",12549
of a consists be to unpromising prevents filtering moves mechanism that evaluated. Such strategy,Such strategy consists of a filtering mechanism that prevents unpromising moves to be evaluated.,12550
"the other well-known two the in proposed metaheuristics. addition, analyzed In effectiveness of strategy we","In addition, we analyzed the effectiveness of the proposed strategy in two other well-known metaheuristics.",12551
We tracking novel automatically propose to populations time-lapse cell in a images. approach,We propose a novel approach to automatically tracking cell populations in time-lapse images.,12552
such a Handwritten digit recognition is example of problem. one typical,Handwritten digit recognition is a typical example of one such problem.,12553
GA a local optimize the developed To has of regions based been selection approach here.,To optimize the selection of local regions a GA based approach has been developed here.,12554
calculated the functions Mass theorem. are by Bayes from using functions generalized plausibility,Mass functions are calculated from plausibility functions by using the generalized Bayes theorem.,12555
belief a Bayesian The compared classifier approach. the with of performances are,The performances of the belief classifier are compared with a Bayesian approach.,12556
"implemented on distributions been far, So based various for distributions. continuous have generating techniques discrete","So far, various techniques have been implemented for generating discrete distributions based on continuous distributions.",12557
characteristics of have this kind been distributions The and of probability studied. properties,The characteristics and properties of this kind of probability distributions have been studied.,12558
"been methods. the estimation related have trough Furthermore, computed parameters classical of","Furthermore, the estimation of related parameters have been computed trough classical methods.",12559
"However, parameter the through studies issue a distributions methods. of these addressed few estimate Bayesian","However, a few studies addressed the parameter estimate issue of these distributions through Bayesian methods.",12560
The patterns vocal a are elements of in birdsong example. prime,The patterns of vocal elements in birdsong are a prime example.,12561
rules The such circuits neural understood. encoding poorly of in is,The encoding of such rules in neural circuits is poorly understood.,12562
"player in key birdsong a together, these findings as determining syntax. HVC Taken implicate","Taken together, these findings implicate HVC as a key player in determining birdsong syntax.",12563
new Generalized were A Model developed. Rectangular,A new Generalized Rectangular Model were developed.,12564
of are Some the generalizations existing offered. models,Some generalizations of the existing models are offered.,12565
"typical is general colocalization, more tasks. than or localization far weakly-supervised cosegmentation, This","This is far more general than typical colocalization, cosegmentation, or weakly-supervised localization tasks.",12566
The software chain the of using a modular rule derivatives. follows implementation,The software follows a modular implementation using the chain rule of derivatives.,12567
"reinforce ecological migratory resilience, MPAs particularly also species. for","MPAs also reinforce ecological resilience, particularly for migratory species.",12568
even of result existing deterioration or Unsuitable condition. in planning the low effectiveness MPA might,Unsuitable MPA planning might result in low effectiveness or even deterioration of the existing condition.,12569
Today areas image of many support decision are processing. and networks used more in Bayesian,Today Bayesian networks are more used in many areas of decision support and image processing.,12570
"image Network the Bayesian In our to quality. this uses approach segmented proposed modelize way,","In this way, our proposed approach uses Bayesian Network to modelize the segmented image quality.",12571
on of a is This calculated measures. that quality represent set local attributes evaluation,This quality is calculated on a set of attributes that represent local evaluation measures.,12572
"by we a Bayesian In Network. the phase, model superpixels second the","In the second phase, we model the superpixels by a Bayesian Network.",12573
"messages. is other negatively taken However, as of side the web inflammatory an such posting","However, an other side of the web is negatively taken such as posting inflammatory messages.",12574
"warning this phenomenon. this enough to of reduce But, kind is not","But, this kind of warning is not enough to reduce this phenomenon.",12575
"be is uncertain defined an more To proposal framework. realistic, within our","To be more realistic, our proposal is defined within an uncertain framework.",12576
infer technique an to edge We the region. potential of structure matching unknown the use,We use an edge matching technique to infer the potential structure of the unknown region.,12577
"estimated along curves. the estimated Driven structure, by synthesis texture automatically performed the is","Driven by the estimated structure, texture synthesis is performed automatically along the estimated curves.",12578
sampling. in ranked estimation entropy the This of set is with non-parametric concerned paper,This paper is concerned with non-parametric estimation of the entropy in ranked set sampling.,12579
are properties studied. of estimator Theoretical the proposed,Theoretical properties of the proposed estimator are studied.,12580
proposed simple estimator rival sampling. estimator The with random in compared the is,The proposed estimator is compared with the rival estimator in simple random sampling.,12581
the are studies examine Monte-Carlo simulation conducted performance the estimator. to Several of,Several Monte-Carlo simulation studies are conducted to examine the performance of the estimator.,12582
disease We model developed a stage-structured (EVD). Ebola virus of,We developed a stage-structured model of Ebola virus disease (EVD).,12583
"to spread, due cases, interventions, asymptomatic immunity, spatiotemporal patterns, etc. pre-existing contact","due to contact patterns, interventions, spatiotemporal spread, pre-existing immunity, asymptomatic cases, etc.",12584
"independance on for made. usually asumption is each For fusion, technique data the used","For each technique used for data fusion, the asumption on independance is usually made.",12585
Markov recognition. Models Hidden (HMMs) for pattern methods are learning,Hidden Markov Models (HMMs) are learning methods for pattern recognition.,12586
"a we using Model present belief In paper, second-order functions. this Markov Hidden","In this paper, we present a second-order Hidden Markov Model using belief functions.",12587
first-order in HMMs on the HMMs. works Previous been belief have focused,Previous works in belief HMMs have been focused on the first-order HMMs.,12588
We second-order the model. them extend to,We extend them to the second-order model.,12589
is semantically fundamental an task from for image. meaningful Segmentation a extracting regions,Segmentation is a fundamental task for extracting semantically meaningful regions from an image.,12590
to each goal object of location. accurately The to labels image is segmentation assign algorithms,The goal of segmentation algorithms is to accurately assign object labels to each image location.,12591
"of cause assignment. ambiguities and However, shortcomings algorithms, label image in uncertainty image-noise,","However, image-noise, shortcomings of algorithms, and image ambiguities cause uncertainty in label assignment.",12592
"other estimate computationally to methods label On the fail efficient hand, most uncertainty.","On the other hand, most computationally efficient methods fail to estimate label uncertainty.",12593
"approach a For evaluation, dataset. quantitative we apply to the our icgbench","For a quantitative evaluation, we apply our approach to the icgbench dataset.",12594
utilize between usually tree body of Mixture representing structure model parts. for relations parts,Mixture of parts model usually utilize tree structure for representing relations between body parts.,12595
We other features. incorporate cues input from,We incorporate other cues from input features.,12596
"although extracted easily be relative surveillance human For environments, example, not silhouettes can flawlessly. in","For example, in surveillance environments, human silhouettes can be extracted relative easily although not flawlessly.",12597
our under framework. other method can easily extend be The generalized features to candidate,The method can be easily extend to other candidate features under our generalized framework.,12598
The approach is of pedestrian experiments through the demonstrated on datasets. efficacy standard several,The efficacy of the approach is demonstrated through experiments on several standard pedestrian datasets.,12599
"features are not easily discriminatory measurable. However,","However, discriminatory features are not easily measurable.",12600
experimentation is for discriminatory Investigative features. necessary identifying,Investigative experimentation is necessary for identifying discriminatory features.,12601
"measurements, scale and statistical descriptors, contrast structures.","statistical descriptors, contrast measurements, and scale structures.",12602
"we results under configurations. obtain various reasonable This way,","This way, we obtain reasonable results under various configurations.",12603
unconstrained of setting. albums the We an photo identities in task peoples' in explore recognizing,We explore the task of recognizing peoples' identities in photo albums in an unconstrained setting.,12604
phenomena Earth's to surface. study the by used acquired are Multispectral satellites on images,Multispectral images acquired by satellites are used to study phenomena on the Earth's surface.,12605
(MD) Macular field central visual loss. causes degeneration,Macular degeneration (MD) causes central visual field loss.,12606
with along structures the is associated degeneration pathways. MD of visual,MD is associated with degeneration of structures along the visual pathways.,12607
morphometry . field degeneration visual voxel-based visual macular Keywords: - - pathway -,Keywords: macular degeneration - visual pathway - visual field - voxel-based morphometry .,12608
The is former for latter. the usually the basis,The former is usually the basis for the latter.,12609
hull are numbers. few of in character features convex based recognition handwritten for Experiments,Experiments of convex hull based features for handwritten character recognition are few in numbers.,12610
"are these combined similarity consid-ered of sources information, as measures, establish Perfectible links. to","Perfectible similarity measures, consid-ered as sources of information, are combined to establish these links.",12611
powerful uncertain is tool for functions belief of a combining such mathematical theory The information.,The theory of belief functions is a powerful mathematical tool for combining such uncertain information.,12612
for characterizing important trajectory recognition object. Motion moving of an property the is,Motion trajectory recognition is important for characterizing the moving property of an object.,12613
this trajectories. method paper novel In extraction we for a propose feature,In this paper we propose a novel feature extraction method for trajectories.,12614
"feature trajectory Experimental show the improvement than results shows methods. that, proposed state-of-art encouraging","Experimental results show that, the proposed trajectory feature shows encouraging improvement than state-of-art methods.",12615
structure the normal a in of cells. change Cancer the with network progresses in gene,Cancer progresses with a change in the structure of the gene network in normal cells.,12616
way flexible to function a information combine provides provided by Belief sources. different theory,Belief function theory provides a flexible way to combine information provided by different sources.,12617
Some to help choose hypothesis. the most rules likely,Some rules help to choose the most likely hypothesis.,12618
made a a of Others set hypotheses. that is on allow decision,Others allow that a decision is made on a set of hypotheses.,12619
Consequently the speed system true vehicles our and accurately. fast computes of moving,Consequently our system computes the true speed of moving vehicles fast and accurately.,12620
effective selective proposal We sampling using object propose techniques. two methods,We propose two effective selective sampling methods using object proposal techniques.,12621
classification. in the how activities terms of decomposed be to,how the activities to be decomposed in terms of classification.,12622
partially the be activated to is configuration allowed network inference. during,the network configuration is allowed to be partially activated during inference.,12623
"outperforms methods. approach in state-of-the-art Our scenarios, and is challenging validated","Our approach is validated in challenging scenarios, and outperforms state-of-the-art methods.",12624
large human activity addition. A database videos RGB-D in of is presented,A large human activity database of RGB-D videos is presented in addition.,12625
on hard cases failed to able before. now are for They they succeed which,They are now able to succeed on hard cases for which they failed before.,12626
"whose scope from algorithm The local steps, three consists global. changes of to","The algorithm consists of three steps, whose scope changes from local to global.",12627
a local from In it step features of (e.g. images local pair extracts the,In the local step it extracts from a pair of images local features (e.g.,12628
"of order belief the function. develop this will study relation, consonant case In we to","In order to develop this relation, we will study the case of consonant belief function.",12629
into modularity as We develop study of the how brain changes children human adults.,We study how modularity of the human brain changes as children develop into adults.,12630
is fitness extracted from model. function A the,A fitness function is extracted from the model.,12631
exploring suggested. modularity performance are Experiments the on cognitive of effect,Experiments exploring the effect of modularity on cognitive performance are suggested.,12632
"or be biomarker a Modularity disease. for injury, may rehabilitation, potential","Modularity may be a potential biomarker for injury, rehabilitation, or disease.",12633
systems. become active inattention has very  Driver intelligent transportation in assessment  a field,  Driver inattention assessment has become a very active field in intelligent transportation systems.,12634
Each relevant information inattention. module driver assessing produces for,Each module produces relevant information for assessing driver inattention.,12635
"well ignorant information. of being own information, their agents' as other maintaining private as","being ignorant of other agents' information, as well as maintaining their own private information.",12636
for year. multi-scale This one over synthetic on a NC-Florida implemented data-set was,This was implemented on a multi-scale synthetic data-set over NC-Florida for one year.,12637
it decision nature subtleties of dynamic is menu-dependent regret-minimization The applied to problems. creates when,The menu-dependent nature of regret-minimization creates subtleties when it is applied to dynamic decision problems.,12638
"with more Secondly, to is arise sophistication subtleties inconsistency. dynamic deal used when","Secondly, more subtleties arise when sophistication is used to deal with dynamic inconsistency.",12639
in role support to computation. modern plays undisputedly vital a GPU) massive implementations CNN,GPU) undisputedly plays a vital role in modern CNN implementations to support massive computation.,12640
"In the to a model process. clean mathematical define paper, framework we this AB-test","In this paper, we define a clean mathematical framework to model the AB-test process.",12641
interesting sketches of line unique Freehand are representation. visual form an and,Freehand line sketches are an interesting and unique form of visual representation.,12642
"of sketches end process. product Typically, the are such an utilized studied as sketching and","Typically, such sketches are studied and utilized as an end product of the sketching process.",12643
category-epitomes analyze belonging to and epitome-scores freehand and categories. We object sketches for construct various,We construct and analyze category-epitomes and epitome-scores for freehand sketches belonging to various object categories.,12644
object Our novel provides for analysis categories. of a the semantic viewpoint nature studying,Our analysis provides a novel viewpoint for studying the semantic nature of object categories.,12645
decades developed for purpose. been this standard algorithms Throughout have already couple,Throughout couple decades standard algorithms have already been developed for this purpose.,12646
image algorithms Some these also. to of are applicable degraded,Some of these algorithms are applicable to degraded image also.,12647
"picture, each part. detect we Given for a test hypotheses","Given a test picture, we detect hypotheses for each part.",12648
resulting function. alternating of apply method minimize We multipliers convex the (ADMM) the direction to,We apply the alternating direction method of multipliers (ADMM) to minimize the resulting convex function.,12649
of video summaries sharing keyframe-based Compact popular are on video way platforms. a viewership generating,Compact keyframe-based video summaries are a popular way of generating viewership on video sharing platforms.,12650
associations. is co-regularization to An meaningful used discover subject-scene scheme original,An original co-regularization scheme is used to discover meaningful subject-scene associations.,12651
The to highly-relevant multimodal resulting keyframes. then select used representations are,The resulting multimodal representations are then used to select highly-relevant keyframes.,12652
The lead for more even smaller is significant summaries.,The lead is even more significant for smaller summaries.,12653
developed an also to is objective An iterative minimize and algorithm we constructed function it.,An objective function is constructed and we also developed an iterative algorithm to minimize it.,12654
based problem. evaluated new bag-of-features is method The image on classification,The new method is evaluated on bag-of-features based image classification problem.,12655
designed This whose is within candidates algorithm (i.e. vertices graph a are candidacy matching,This algorithm is designed within a candidacy graph whose vertices are matching candidates (i.e.,12656
"of and convergence. for two in a target iterates source parts), pair steps body and","a pair of source and target body parts), and iterates in two steps for convergence.",12657
(i) relations on It based competitive body partial and matches among compatible possible generates parts.,(i) It generates possible partial matches based on compatible and competitive relations among body parts.,12658
match of The with SRC. mechanism are optimized the well projections and,The projections are optimized and match well with the mechanism of SRC.,12659
"Therefore, transformed well SRC space. the in performs OP-SRC","Therefore, SRC performs well in the OP-SRC transformed space.",12660
natural content. problems in and fundamental scene classification are understanding two texture Dynamic video,Dynamic texture and scene classification are two fundamental problems in understanding natural video content.,12661
a is crucial problems. step these towards Extracting effective robust and solving features,Extracting robust and effective features is a crucial step towards solving these problems.,12662
term as two-level a Feature We feature ConvNet extraction (TCoF). this Transferred scheme,We term this two-level feature extraction scheme as a Transferred ConvNet Feature (TCoF).,12663
visual contain often sparse detail. Freehand sketches,Freehand sketches often contain sparse visual detail.,12664
"the However, has images explored. been domain sketch not of","However, the domain of sketch images has not been explored.",12665
"""deep"" a paper CNNs. introduces sketch based freehand from recognition on extracted This features framework","This paper introduces a freehand sketch recognition framework based on ""deep"" features extracted from CNNs.",12666
is understanding. an Action multimedia problem important in recognition,Action recognition is an important problem in multimedia understanding.,12667
model. action This an addresses building this by expressive paper compositional problem,This paper addresses this problem by building an expressive compositional action model.,12668
"are domains. between contextual leaf-nodes the and in temporal interactions both Moreover, spatial defined","Moreover, the contextual interactions are defined between leaf-nodes in both spatial and temporal domains.",12669
production optimization the leaf-nodes associated multi-layer or-nodes) with parameters. the of of with the along,the production of leaf-nodes associated with the or-nodes) along with the optimization of multi-layer parameters.,12670
"information, work In complete cases, under knowledge building bases. agents most complete","In most cases, agents work under complete information, building complete knowledge bases.",12671
of memory the prohibitive is multi-region limiting requirements. the One adaption segmentation issue large-scale sometimes,One issue limiting the adaption of large-scale multi-region segmentation is the sometimes prohibitive memory requirements.,12672
"instead any This sub-sampling, but method does over all sub-windows. perform optimizes not","This method does not perform any sub-sampling, but instead optimizes over all sub-windows.",12673
this substantial performance we datasets. available publicly gains on three show approach Using,Using this approach we show substantial performance gains on three publicly available datasets.,12674
of characterize Information a performance popularity theory tool is gaining to as systems. biological,Information theory is gaining popularity as a tool to characterize performance of biological systems.,12675
as domain the considered Only entropy are greater than with pool. a domain threshold blocks,Only the domain blocks with entropy greater than a threshold are considered as domain pool.,12676
significant detection benchmark advances resulted on supervised has Recent datasets. in performance object in salient,Recent advances in supervised salient object detection has resulted in significant performance on benchmark datasets.,12677
"annotations pixel-wise Training however, expensive salient models, of objects. requires such","Training such models, however, requires expensive pixel-wise annotations of salient objects.",12678
approach. the on our of datasets results Experimental proposed validate benchmark effectiveness,Experimental results on benchmark datasets validate the effectiveness of our proposed approach.,12679
"demonstrate the of variability Furthermore, strong the temporal spread balance and processes. in we spatial","Furthermore, we demonstrate the strong spatial and temporal variability in the balance of spread processes.",12680
"other demic, times more cultural diffusion. more sometimes region and shows at Each","Each region shows sometimes more demic, and at other times more cultural diffusion.",12681
"Only dominance a of areas few, diffusion. environmentally possibly marginal, demic show","Only few, possibly environmentally marginal, areas show a dominance of demic diffusion.",12682
"use such computationally normalising however, of The hampered constant. models intractable by is, a","The use of such models is, however, hampered by a computationally intractable normalising constant.",12683
fully a discrete treatment makes of parameter Markov This Bayesian difficult. estimation fields random and,This makes parameter estimation and a fully Bayesian treatment of discrete Markov random fields difficult.,12684
a problem   representations effective is understanding. but critical challenging Constructing multimedia in,  Constructing effective representations is a critical but challenging problem in multimedia understanding.,12685
"domain features traditional limiting on knowledge, performances rely the often The exiting handcraft of methods.","The traditional handcraft features often rely on domain knowledge, limiting the performances of exiting methods.",12686
features into manner. a in layer-wise compositional wavelets) Gabor,Gabor wavelets) into compositional features in a layer-wise manner.,12687
"number produce (i.e. a In each base of classifiers layer, we","In each layer, we produce a number of base classifiers (i.e.",12688
lot of in A research applications. and been many conducted resulted has has,A lot of research has been conducted and has resulted in many applications.,12689
recent provide a field. We progress this review of in comprehensive to aim the,We aim to provide a comprehensive review of the recent progress in this field.,12690
"addition, evaluation also and we In existing datasets the review influential metrics.","In addition, we also review the existing influential datasets and evaluation metrics.",12691
"design in research segmentation. suggest we directions for image future research and some Finally, flavors","Finally, we suggest some design flavors and research directions for future research in image segmentation.",12692
structures (e.g. model the This determines iteratively algorithm,This algorithm iteratively determines the model structures (e.g.,12693
parameter along learning. and the with the their layouts) nodes,the nodes and their layouts) along with the parameter learning.,12694
"variables structures This ""switch"" by tree hierarchical the introducing model traditional the extends (i.e.","This model extends the traditional hierarchical tree structures by introducing the ""switch"" variables (i.e.",12695
variations. that or-nodes) explicitly capture production specify shape to the rules,the or-nodes) that explicitly specify production rules to capture shape variations.,12696
"bricks, sequence At video i.e. the scene, we location of a regular extract of any","At any location of the scene, we extract a sequence of regular video bricks, i.e.",12697
and spatial temporal over domain. both spanning video volumes,video volumes spanning over both spatial and temporal domain.,12698
parameter well. empirical studies The analysis and setting as component of are presented,The empirical studies of parameter setting and component analysis are presented as well.,12699
(i) Our exemplars retrieving a target its with framework steps: two image (i.e. parses,Our framework parses a target image with two steps: (i) retrieving its exemplars (i.e.,12700
labels and the target of the references). assigned the,the assigned labels of the target and the references).,12701
"Proc. Therefore, observation 'sliding window' et Renz a (M. applied al., we approach","Therefore, we applied a 'sliding observation window' approach (M. Renz et al., Proc.",12702
the results of demonstrate The effectiveness experimental the proposed method.,The experimental results demonstrate the effectiveness of the proposed method.,12703
information. that is of issue we role the address epigenetic Another,Another issue that we address is the role of epigenetic information.,12704
the in our records. senate simulations We modeling demonstrate of and application method for voting,We demonstrate our method in simulations and for the application of modeling senate voting records.,12705
common is face problem. yet in challenging a recognition Occlusion,Occlusion in face recognition is a common yet challenging problem.,12706
representation in (SRC) conditions sparse laboratory shown based classification While promising been has (i.e. performance,While sparse representation based classification (SRC) has been shown promising performance in laboratory conditions (i.e.,12707
"scenarios. noiseless pixel in or much random it corrupted), worse practical performs","noiseless or random pixel corrupted), it performs much worse in practical scenarios.",12708
We occlusion coding the address (SOC) structured to occlusion problems. propose,We propose the structured occlusion coding (SOC) to address occlusion problems.,12709
coding in The two here structured lies folds.,The structured coding here lies in two folds.,12710
"dictionary we recognition. one hand, structured On for a employ","On one hand, we employ a structured dictionary for recognition.",12711
"formulation. other hand, structured the use this the sparsity propose we On in to","On the other hand, we propose to use the structured sparsity in this formulation.",12712
"the Specifically, image. SOC simultaneously occlusion and the separates classifies","Specifically, SOC simultaneously separates the occlusion and classifies the image.",12713
"large world. in in improvement occlusion handling achieves real significant Moreover, SOC","Moreover, SOC achieves significant improvement in handling large occlusion in real world.",12714
inference. consisting a two propose of data-driven of We framework phases,We propose a data-driven framework consisting of two phases of inference.,12715
Graph label can the algorithm. efficient The using joint assignment solved be Cuts,The joint label assignment can be solved using the efficient Graph Cuts algorithm.,12716
"not of whole do allow algorithms existing analyses spectra. However, for automatic","However, existing algorithms do not allow for automatic analyses of whole spectra.",12717
automatic spectral is partitioning The of into fragments. idea protein signal mass main,The main idea is automatic partitioning of protein mass spectral signal into fragments.,12718
Gaussian obtained models. mixture separately into decomposed fragments The are,The obtained fragments are separately decomposed into Gaussian mixture models.,12719
network. gives of an accurate the connectivity of modeling neuronal This way the to,This gives way to an accurate modeling of the connectivity of the neuronal network.,12720
The important feature pathway. peripheral is the the neuronal all timing auditory in,The timing is the all important feature in the peripheral neuronal auditory pathway.,12721
build solid basis a provide tools mathematical for to The processes. auditory peripheral models,The mathematical tools provide a solid basis to build models for peripheral auditory processes.,12722
key in high-resolution a remote of interpretation imagery. classification sensing is Scene the problem,Scene classification is a key problem in the interpretation of high-resolution remote sensing imagery.,12723
"have {e.g. been literature, In strategies the many used, sampling different","In the literature, many different sampling strategies have been used, {e.g.",12724
"sampling, saliency-based keypoint-based sampling, etc. and dense random sampling sampling,","dense sampling, random sampling, keypoint-based sampling and saliency-based sampling, etc.",12725
support use machines additional We perform sparse model to vector reduction.,We use sparse support vector machines to perform additional model reduction.,12726
of classifier any gives form of reliable regularization. disparities pixel without estimations already Proposed,Proposed classifier gives reliable estimations of pixel disparities already without any form of regularization.,12727
to algorithm very comparable promising the Our results state-of-the-art. obtained,Our algorithm obtained very promising results comparable to the state-of-the-art.,12728
probabilistic (MCMC) achieved Chain data-driven by algorithm. inference is a Markov Carlo The Monte,The probabilistic inference is achieved by a data-driven Markov Chain Monte Carlo (MCMC) algorithm.,12729
"kinetic is unclear. mechanism However, its underlying","However, its underlying kinetic mechanism is unclear.",12730
trajectories single-year most cohorts. is with consistent birth prediction This reported mortality for recently the,This prediction is consistent with the most recently reported mortality trajectories for single-year birth cohorts.,12731
while of rate protecting the increases senescence decreases. The the energy organism of,The rate of senescence of the organism increases while the protecting energy decreases.,12732
the mortality protecting plateau The minimal energy when is decreases reached levels. its to,The mortality plateau is reached when the protecting energy decreases to its minimal levels.,12733
protecting is the energy The rate of decreasing temperature-dependent.,The decreasing rate of the protecting energy is temperature-dependent.,12734
study and biochemical exploring This connection is demography. the between mechanism,This study is exploring the connection between biochemical mechanism and demography.,12735
approximations and scalable Heuristic high-performance empirically implementations tested. are described for,Heuristic approximations for scalable high-performance implementations are described and empirically tested.,12736
done are step. and for experimental The analysis each evaluation,The experimental evaluation and analysis are done for each step.,12737
We a that it real-world both believe research to application. is service academic useful and,We believe that it is a useful service to both academic research and real-world application.,12738
classifier It sub-interval in the maximum score the time. linear identifies with,It identifies the sub-interval with the maximum classifier score in linear time.,12739
two-fold. The advantages of our are method,The advantages of our method are two-fold.,12740
"in lack training usually labelled of this requirement is by the practice, impaired samples. However,","However, in practice, this requirement is usually impaired by the lack of labelled training samples.",12741
"a binarization In this is images framework multiple-expert multispectral work, proposed. for","In this work, a multiple-expert binarization framework for multispectral images is proposed.",12742
binarization enhance a uses binarization. performance the framework to wrapper the The gray-level of,The framework uses a binarization wrapper to enhance the performance of the gray-level binarization.,12743
of bands used the spectral preprocessing the textual enhance Nonlinear is information. individual to,Nonlinear preprocessing of the individual spectral bands is used to enhance the textual information.,12744
is framework truth results. a ground multispectral with promising The applied dataset to,The framework is applied to a ground truth multispectral dataset with promising results.,12745
presented The as and well. empirical analysis implementation details are,The implementation details and empirical analysis are presented as well.,12746
been of emergence the deep advanced The face state-of-the-art has recognition significantly learning. of by,The state-of-the-art of face recognition has been significantly advanced by the emergence of deep learning.,12747
effectiveness their motivates investigate to on This recognition. face us,This motivates us to investigate their effectiveness on face recognition.,12748
in LFW face given is of end. verification result the discussion further A,A further discussion of LFW face verification result is given in the end.,12749
"classifiers compositions, alternative making specifying for or-nodes) (i.e. and (ii) local","or-nodes) for specifying alternative compositions, and (ii) making local classifiers (i.e.",12750
for capabilities structural account These preserving variabilities while us to the model compact. enable well,These capabilities enable us to account well for structural variabilities while preserving the model compact.,12751
vectors for arc fitting tangent at specified isovalues. the Hermite Circular the curves defines,Circular arc fitting defines the tangent vectors for the Hermite curves at specified isovalues.,12752
mixture low-dimensional robust Gaussian action models each using action approach using This a features.,This approach models each action using a Gaussian mixture using robust low-dimensional action features.,12753
"Using an design MCMC to the that optimum. we converges techniques, global algorithm","Using MCMC techniques, we design an algorithm that converges to the global optimum.",12754
IHT. a called method uses The greedy reconstruction procedure Pseudo-inverse fast,The reconstruction procedure uses a fast greedy method called Pseudo-inverse IHT.,12755
truncation in iteration. an is efficient embedded The step power via achieved tensor the sparsity,The sparsity is achieved via an efficient truncation step embedded in the tensor power iteration.,12756
further is investigation theoretical A thorough conducted.,A thorough theoretical investigation is further conducted.,12757
"inapplicable. work medial renders community's of segmentation the has much that been solved, axis figure-ground","that figure-ground segmentation has been solved, renders much of the medial axis community's work inapplicable.",12758
"al. in article, computational et we framework, In this reported previously review Lee a","In this article, we review a computational framework, previously reported in Lee et al.",12759
method We scenes. images cluttered on of the evaluate,We evaluate the method on images of cluttered scenes.,12760
tasks. plays and Image essential representation related recognition image feature role in an,Image feature representation plays an essential role in image recognition and related tasks.,12761
is The data. learning paradigm state-of-the-art feature current labeled learning supervised from,The current state-of-the-art feature learning paradigm is supervised learning from labeled data.,12762
"data media. social Instead, user learn behavior from we on collected","Instead, we learn from user behavior data collected on social media.",12763
from large-scale collect behavior dataset user image a and We Behance.net.,We collect a large-scale image and user behavior dataset from Behance.net.,12764
benchmarks. learned performs on that We also the competitively various recognition feature show,We also show that the learned feature performs competitively on various recognition benchmarks.,12765
"For frontiers. the new than for half a transform more is Hough ever-expanding century,","For more than half a century, the Hough transform is ever-expanding for new frontiers.",12766
and research Thousands over have the of papers decades. applications evolved numerous,Thousands of research papers and numerous applications have evolved over the decades.,12767
space-demanding. out is all-inclusive and possible an survey Carrying enormously hardly,Carrying out an all-inclusive survey is hardly possible and enormously space-demanding.,12768
and solution approaches. clarified high computation time is The for storage with demand different,The high demand for storage and computation time is clarified with different solution approaches.,12769
Various explanations the findings discussed. for are,Various explanations for the findings are discussed.,12770
"SBtab types, of types data. define also adjusting user own table The to can other","The user can also define own table types, adjusting SBtab to other types of data.",12771
"be found further information tools, www.sbtab.net. at code, and Software can","Software code, tools, and further information can be found at www.sbtab.net.",12772
data have density in and (GMM) many Mixture applications Models clustering. found Gaussian estimation,Gaussian Mixture Models (GMM) have found many applications in density estimation and data clustering.,12773
"to curved and data. model not the However, does strongly nonlinear well adapt","However, the model does not adapt well to curved and strongly nonlinear data.",12774
a introduce for to We identify materials. using novel method reflectance,We introduce a novel method for using reflectance to identify materials.,12775
material information discriminative disks reflectance encode recognition. efficient These and for accurate,These reflectance disks encode discriminative information for efficient and accurate material recognition.,12776
"tasks. are image re-identification search person two For time, separately studied long and","For long time, person re-identification and image search are two separately studied tasks.",12777
"two paper claims this contributions. Specifically, major","Specifically, this paper claims two major contributions.",12778
"settings, are perspectives new closer to dataset and reaches provided. realistic Our","Our dataset reaches closer to realistic settings, and new perspectives are provided.",12779
"datasets, three on we competitive compared results with report state-of-the-art methods. Moreover, the","Moreover, on three datasets, we report competitive results compared with the state-of-the-art methods.",12780
the this state-of-the-art surveys This topic. on paper techniques,This paper surveys the state-of-the-art techniques on this topic.,12781
available related and knowledge the first features crowded the to scenes. background We provide,We first provide the background knowledge and the available features related to crowded scenes.,12782
the performance also We datasets available outline evaluation. for,We also outline the available datasets for performance evaluation.,12783
"discussions. presented Finally, some research future and with problems directions are promising","Finally, some research problems and promising future directions are presented with discussions.",12784
number is action of The rapidly. for growing categories recognition,The number of categories for action recognition is growing rapidly.,12785
popular the increasingly be 'zero-shot (ZSL) by issue ameliorated may learning' paradigm. This,This issue may be ameliorated by the increasingly popular 'zero-shot learning' (ZSL) paradigm.,12786
"data, attribute-based Existing on and semantic image representations. studies focus ZSL primarily","Existing ZSL studies focus primarily on image data, and attribute-based semantic representations.",12787
propagation a segmentation. novel presents constraint This selective constrained for image method paper,This paper presents a novel selective constraint propagation method for constrained image segmentation.,12788
graph-based problem then efficient a Such constraint selective is learning an algorithm. by propagation solved,Such a selective constraint propagation problem is then solved by an efficient graph-based learning algorithm.,12789
the limitation predictive reliability thus power This downstream of overall and biological inference. reduces,This limitation reduces overall predictive power and thus the reliability of downstream biological inference.,12790
high objects. of an with contain are bounding ensemble potential to boxes Object proposals,Object proposals are an ensemble of bounding boxes with high potential to contain objects.,12791
with to offers dimensional functions inputs. It multivariate high the ability model also random,It also offers the ability to model multivariate random functions with high dimensional inputs.,12792
approached recognition counting can by visual be problems instances. Many,Many visual recognition problems can be approached by counting instances.,12793
to parameters of many how can instances occur in a encode tend Learned interest. class,Learned parameters can encode how many instances tend to occur in a class of interest.,12794
tackle of to or levels encoded soft diverse be ambiguity. relations Hard cardinality can,Hard or soft cardinality relations can be encoded to tackle diverse levels of ambiguity.,12795
data resonance process. an with medical imaging slow tool (MRI) Magnetic inherently essential acquisition is,Magnetic resonance imaging (MRI) is an essential medical tool with inherently slow data acquisition process.,12796
process to scanning apparatus. patient time acquisition requires Slow long be exposed to,Slow acquisition process requires patient to be long time exposed to scanning apparatus.,12797
emerging is processing. Sensing Compressive in theory signal an,Compressive Sensing is an emerging theory in signal processing.,12798
reduce reconstruction. It signal acquired for successful aims to required amount of the data,It aims to reduce the amount of acquired data required for successful signal reconstruction.,12799
"acquisition of lower time, i.e. to leads acquired the coefficients amount image Reducing","Reducing the amount of acquired image coefficients leads to lower acquisition time, i.e.",12800
the MRI exposition of to time apparatus.,time of exposition to the MRI apparatus.,12801
Face procedure. application detection was above explain chosen to,Face detection application was chosen to explain above procedure.,12802
"method subgraph. multiplier Furthermore, propose to a the U-statistic we confidence bootstrap construct","Furthermore, we propose a U-statistic multiplier bootstrap method to construct the confidence subgraph.",12803
and include filters feature distributions. of The statistics velocity spatio-temporal histograms,The feature statistics include histograms of spatio-temporal filters and velocity distributions.,12804
Epistasis a of adaptation. theory the in concept is key,Epistasis is a key concept in the theory of adaptation.,12805
depend recent Some information approaches theory. on,Some recent approaches depend on information theory.,12806
can loci misleading. entropy show shared for We that of be pairs considering,We show that considering shared entropy for pairs of loci can be misleading.,12807
reason is shared entropy epistasis does pair. imply for that The the not,The reason is that shared entropy does not imply epistasis for the pair.,12808
observation holds order absence epistasis. the in also of true This higher,This observation holds true also in the absence of higher order epistasis.,12809
interactions pairwise discuss We identifying using a entropy. refined approach for,We discuss a refined approach for identifying pairwise interactions using entropy.,12810
semantically in challenge. a notoriously first-order logic is Reasoning,Reasoning semantically in first-order logic is notoriously a challenge.,12811
share means knowledge. Ontologies widely accepted are to,Ontologies are widely accepted means to share knowledge.,12812
"as in central semantic role such segmentation, labelling Pixel-level tasks, a play understanding. image","Pixel-level labelling tasks, such as semantic segmentation, play a central role in image understanding.",12813
widely systems. such method One tight is used the frame wavelet representative,One such representative method is the widely used wavelet tight frame systems.,12814
multi-view we In of paper this consider the face problem detection.,In this paper we consider the problem of multi-view face detection.,12815
"faces fully capture in dozens also models to of all training They require orientations, e.g.","They also require training dozens of models to fully capture faces in all orientations, e.g.",12816
"verbal memory, fluencies reasoning, short-term and speed, processing auditory-verbal assessed. memory, Long-term non-verbal were","Long-term memory, auditory-verbal short-term memory, processing speed, non-verbal reasoning, and verbal fluencies were assessed.",12817
"verbal and effects. For age fluencies, scored musicians different controls displayed than higher","For verbal fluencies, musicians scored higher than controls and displayed different age effects.",12818
gap. paper The at filling aims the,The paper aims at filling the gap.,12819
noise repressor cause transcriptional binding. may a RNA repression between competition and Transcriptional by polymerase,Transcriptional repression may cause transcriptional noise by a competition between repressor and RNA polymerase binding.,12820
"In transcriptional particular, also substantial allows promoters give noise. this weak to","In particular, this allows also weak promoters to give substantial transcriptional noise.",12821
dynamics migration. and Mathematical cell cytoskeleton modeling of,Mathematical modeling of cytoskeleton dynamics and cell migration.,12822
to scale-based modeling. from multiscale mechanics: Cell single models,Cell mechanics: from single scale-based models to multiscale modeling.,12823
the causal is method based The constraint algorithm for discovery. PC state-of-the-art,The PC algorithm is the state-of-the-art constraint based method for causal discovery.,12824
and a range We high dimensional to real-world of our synthetic method apply datasets.,We apply our method to a range of synthetic and real-world high dimensional datasets.,12825
"parallel-PC relationships. inference regulatory for causal inferring into miRNA-mRNA Furthermore, we a integrate method","Furthermore, we integrate parallel-PC into a causal inference method for inferring miRNA-mRNA regulatory relationships.",12826
observations. of forward to the is admitted explored form model The the likelihoods well,The well admitted forward model is explored to form the likelihoods of the observations.,12827
leads Sylvester solving Maximizing equation. likelihoods a the to,Maximizing the likelihoods leads to solving a Sylvester equation.,12828
closed-form Efficient with inference memory footprint. straightforward computations parallelisation a distributed for and small allows,Efficient closed-form inference allows for straightforward parallelisation and distributed computations with a small memory footprint.,12829
GP resources data can sufficient handle our distributed computing arbitrarily sets. large With model,With sufficient computing resources our distributed GP model can handle arbitrarily large data sets.,12830
Different situations. required of clusterings different characteristics are in,Different characteristics of clusterings are required in different situations.,12831
"with tractability such fail CSPs most restrictions to global However, guarantee for constraints.","However, most such restrictions fail to guarantee tractability for CSPs with global constraints.",12832
therefore constraints. instances structural restrictions applicability the with We study of to such,We therefore study the applicability of structural restrictions to instances with such constraints.,12833
"w.r.t. ensemble to of shaped For simultaneously this, policies, an learn propose we","For this, we propose to simultaneously learn an ensemble of policies, shaped w.r.t.",12834
a scales. many range and heuristics of on,many heuristics and on a range of scales.,12835
then policy by is obtained voting. The target,The target policy is then obtained by voting.,12836
shape multi-scale to a sketch-based approach We  present retrieval. ,  We present a multi-scale approach to sketch-based shape retrieval.,12837
image blocks using connected into The preprocessed input component and segmented analysis. is,The input image is preprocessed and segmented into blocks using connected component analysis.,12838
block the obtain based morphology the approach. of using segmented skeleton We,We obtain the skeleton of the segmented block using morphology based approach.,12839
"weight is task in an non-trivial, important matching generation however a technique. instance Automatic","Automatic weight generation is a non-trivial, however an important task in instance matching technique.",12840
"Thus solve far, issues. techniques some different proposed have various pressing of the to researchers","Thus far, various researchers have proposed different techniques to solve some of the pressing issues.",12841
medical segmentation representative image techniques. paper an overview This seminar presents of,This seminar paper presents an overview of representative medical image segmentation techniques.,12842
The is condition valid. the measurement primary necessary a condition and sufficient that is,The primary condition is the necessary and sufficient condition that a measurement is valid.,12843
"reconstruction We conjecture, when i.e., threshold. impossible of this below half prove one is the","We prove one half of this conjecture, i.e., reconstruction is impossible when below the threshold.",12844
"direction, to weighted In the introduce label the a exploit positive we information. graph","In the positive direction, we introduce a weighted graph to exploit the label information.",12845
"sets paper, advantages. to we use these and this several data In real illustrate simulated","In this paper, we use several simulated and real data sets to illustrate these advantages.",12846
of We extensions also will model. our some discuss briefly,We will also briefly discuss some extensions of our model.,12847
of principles Extending perception i.e. the image,Extending the principles of image perception i.e.,12848
feature a represent interactions central of human all exchanges. Reciprocating,Reciprocating interactions represent a central feature of all human exchanges.,12849
steps into the future. different of plan Agents may also numbers,Agents may also plan different numbers of steps into the future.,12850
"However, complexities inordinately IPOMDPs challenging. the computationally various make","However, the various complexities make IPOMDPs inordinately computationally challenging.",12851
elucidate richness of interactive and behaviour use inference. generated We sophistication the to,We use generated behaviour to elucidate the richness and sophistication of interactive inference.,12852
of evaluation the This complete most correctors collection has so high-performance error far.,This evaluation has the most complete collection of high-performance error correctors so far.,12853
presented. is in A models for B-splines mixed algorithm fast,A fast algorithm for B-splines in mixed models is presented.,12854
"because sparse. attractive, B-splines have computational and are are support the local corresponding matrices","B-splines have local support and are computational attractive, because the corresponding matrices are sparse.",12855
in We discovering for approach novel interactions videos. a human present,We present a novel approach for discovering human interactions in videos.,12856
"carried datasets, VIRAT, Activity, been Extensive Collective UT-interaction. experiments three and over out have challenging","Extensive experiments have been carried out over three challenging datasets, Collective Activity, VIRAT, and UT-interaction.",12857
retrieval. representing We for method for and motion video classification a propose information,We propose a method for representing motion information for video classification and retrieval.,12858
"on datasets. Experimental model achieves state-of-the-arts several simple, our that, results albeit show benchmark results","Experimental results show that, albeit simple, our model achieves state-of-the-arts results on several benchmark datasets.",12859
segmentation maximum is and principle entropy used images. of The in thresholding largely,The maximum entropy principle is largely used in thresholding and segmentation of images.,12860
"in of Here, entropic index determining discuss role its we the thresholds. the","Here, we discuss the role of its entropic index in determining the thresholds.",12861
"framework. configurable also incorporates solvers, and providing usable It state-of-the-art a","It incorporates state-of-the-art solvers, providing also a usable and configurable framework.",12862
high-dimension datasets. size Our size traditional and high-sample method low-sample both proposed low-dimensional handles,Our proposed method handles both high-dimension low-sample size and traditional low-dimensional high-sample size datasets.,12863
conflicting involving a Multicriteria decision a person problem aims analysis decision criteria. at facing supporting,Multicriteria decision analysis aims at supporting a person facing a decision problem involving conflicting criteria.,12864
the first can be that the general case. show unbounded We length in of explanations,We show first that the length of explanations can be unbounded in the general case.,12865
include the D-criteria. criteria popular most A- The and,The most popular criteria include the A- and D-criteria.,12866
from for idea ontologies. an paper generation use presents This learning of rule inductive,This paper presents an idea of inductive learning use for rule generation from ontologies.,12867
"cross-modal we corresponded focus with paper, on partially In this hashing data.","In this paper, we focus on cross-modal hashing with partially corresponded data.",12868
full correspondence performance. The enhance without hashing to made are in data use the,The data without full correspondence are made in use to enhance the hashing performance.,12869
"experiments With such systems, to those post-stimulus well-timed similar obtained. from be responses can","With such systems, well-timed post-stimulus responses similar to those from experiments can be obtained.",12870
way. seamless aspects a related) process in,process related) aspects in a seamless way.,12871
to is the when the challenging video becomes dynamic. used capture problem especially camera The,The problem especially becomes challenging when the camera used to capture the video is dynamic.,12872
"approaches pros/cons, representative motivations, are and strategies, compared. The of and described performance","The motivations, strategies, pros/cons, and performance of representative approaches are described and compared.",12873
"for research promising future are directions Moreover, discussed.","Moreover, promising directions for future research are discussed.",12874
A the book given is follows. of overview as brief,A brief overview of the book is given as follows.,12875
meaningful learning users' aims interactions enhance through them. experience It with to providing,It aims to enhance users' learning experience through providing meaningful interactions with them.,12876
evidence the original conclude there for claim. no We is that,We conclude that there is no evidence for the original claim.,12877
optimization in black-box expensive of constraints arise problems. many Unknown types,Unknown constraints arise in many types of expensive black-box optimization problems.,12878
"pathologies when to lead can constraints. EI with However, used","However, EI can lead to pathologies when used with constraints.",12879
"approaches suffer failure from not these modes. By do contrast, information-based","By contrast, information-based approaches do not suffer from these failure modes.",12880
selection annealing estimates AIS problem an is for obtaining of accurate schedule. important the One,One important problem for obtaining accurate AIS estimates is the selection of an annealing schedule.,12881
belief A of an basic fusion for using assignments information layer approach. combination,A layer for combination of basic belief assignments using an information fusion approach.,12882
the decision Ranking is to system used making. ultimately decisions support,Ranking the system decisions is ultimately used to support decision making.,12883
"have problems. to Bayesian networks learning neural In approach not does these the principle,","In principle, the Bayesian approach to learning neural networks does not have these problems.",12884
"network to dataset techniques and lack sizes. scalability large existing However, Bayesian","However, existing Bayesian techniques lack scalability to large dataset and network sizes.",12885
"are written For by consistent with each some people multiple other. descriptions images,","For some images, descriptions written by multiple people are consistent with each other.",12886
"images, considerably. But people other for vary descriptions across","But for other images, descriptions across people vary considerably.",12887
"the notion image this paper, In we of introduce specificity.","In this paper, we introduce the notion of image specificity.",12888
to ecological are communities. to interactions fundamental interspecific thought structures of contribute Dynamic,Dynamic interspecific interactions are thought to contribute to fundamental structures of ecological communities.,12889
proposed be intransitive is structure. an referred arrangement as that to an such It,It is proposed that such an arrangement be referred to as an intransitive structure.,12890
Gustafson-Kessel be for clustering. a generalization This algorithm the of considered can algorithm fuzzy,This algorithm can be considered a generalization of the Gustafson-Kessel algorithm for fuzzy clustering.,12891
algorithm using learning. Segmentation A proposed Fusion distributed is consensus Semi-supervised and,A Semi-supervised Segmentation Fusion algorithm is proposed using consensus and distributed learning.,12892
a (SSSF). Semi-supervised algorithm USF called Semi-supervision Segmentation using is in new incorporated Fusion,Semi-supervision is incorporated in USF using a new algorithm called Semi-supervised Segmentation Fusion (SSSF).,12893
a study. in efficiency the is simulation The algorithm illustrated of,The efficiency of the algorithm is illustrated in a simulation study.,12894
to the pose neural object. use the a of predict (CNN) network convolutional We,We use a convolutional neural network (CNN) to predict the pose of the object.,12895
pixel using trained normals synthetic containing images rendered objects. CNN is in This,This CNN is trained using pixel normals in images containing rendered synthetic objects.,12896
"it algorithms tested data. on real outperforms real data, When on alternative trained","When tested on real data, it outperforms alternative algorithms trained on real data.",12897
in the domain. is Image important stage enhancement image-processing an,Image enhancement is an important stage in the image-processing domain.,12898
image is the The equalization. known enhancement most method histogram,The most known image enhancement method is the histogram equalization.,12899
obtained within are and can logarithmic affine efficient be used. Simple transforms models methods if,Simple and efficient methods can be obtained if affine transforms within logarithmic models are used.,12900
"In has due in years, construction data distinctive-dictionary recent his processing. usefulness gained to importance","In recent years, distinctive-dictionary construction has gained importance due to his usefulness in data processing.",12901
construction A is dictionary introduced. new algorithm,A new dictionary construction algorithm is introduced.,12902
that generic and construction general applications. dictionary different fits is The,The dictionary construction is generic and general that fits different applications.,12903
"classification, decomposition, content-based randomized computer LU detection, file decomposition, security. LU }$ construction, Dictionary","}$ Dictionary construction, classification, LU decomposition, randomized LU decomposition, content-based file detection, computer security.",12904
prototype camera a A using is built modulator. liquid-crystal-on-silicon,A prototype camera is built using a liquid-crystal-on-silicon modulator.,12905
techniques reduction Therefore need importance. speckle high for is noise of the,Therefore the need for speckle noise reduction techniques is of high importance.,12906
system a automatically position. information applied appropriate be can correct this to to further This,This information can be further applied to a appropriate system to automatically correct this position.,12907
"proposed In activity is algorithm new paper, group heat-map-based (HMB) this for recognition. a","In this paper, a new heat-map-based (HMB) algorithm is proposed for group activity recognition.",12908
a for recognizing also activities. group surface-fitting And method (SF) proposed is,And a surface-fitting (SF) method is also proposed for recognizing group activities.,12909
results algorithms. our the effectiveness demonstrate of Experimental proposed,Experimental results demonstrate the effectiveness of our proposed algorithms.,12910
"common specifically, matching More correspondences graphs. among the node","More specifically, matching the common node correspondences among graphs.",12911
concept of photographs. the importance the analyzes in group paper of This individuals,This paper analyzes the concept of the importance of individuals in group photographs.,12912
that hold We the rules. relationships inclusion the study between,We study the inclusion relationships that hold between the rules.,12913
"substantially methods. baseline than achieve performance After better we finetuning, the","After finetuning, we achieve substantially better performance than the baseline methods.",12914
with long bias' evolution `turnover and favors of times. the strategies inevitably This exists generation,This `turnover bias' inevitably exists and favors the evolution of strategies with long generation times.,12915
We learning chess a a tasks framework the in domain. series of with evaluate rule,We evaluate the framework with a series of tasks in a chess rule learning domain.,12916
for the optimization is methodology of Bayesian an expensive optimization functions with evaluations. global effective,Bayesian optimization is an effective methodology for the global optimization of functions with expensive evaluations.,12917
querying surrogate model. a on by It functions cheap over a relatively distribution defined relies,It relies on querying a distribution over functions defined by a relatively cheap surrogate model.,12918
of one oldest show world. game running Countdown TV the The is in the,The Countdown game is one of the oldest TV show running in the world.,12919
"conjunction, introduced. and Basic operations of negation disjunction are","Basic operations of conjunction, disjunction and negation are introduced.",12920
localization video role text major in a detection and The analysis understanding. and plays,The text detection and localization plays a major role in video analysis and understanding.,12921
input edge is and The detect text is done preprocessed the emphasis to clusters. image,The input image is preprocessed and edge emphasis is done to detect the text clusters.,12922
ideas hexa-valued logics. from combine fuzzy and and sets four-valued Both,Both combine ideas from fuzzy sets and four-valued and hexa-valued logics.,12923
objects. imprecise or represent contradictory whose accomplished unknown is some Both properties for degree,Both represent imprecise properties whose accomplished degree is unknown or contradictory for some objects.,12924
is and the discussed. possible in processing analysis color image application A color,A possible application in the color analysis and color image processing is discussed.,12925
"sparsity-based end, this To two metrics, i.e.","To this end, two sparsity-based metrics, i.e.",12926
imprecise set or Here novel viz. vague to a idea handle,Here a novel idea to handle imprecise or vague set viz.,12927
triplet is fuzzy of functions. two membership set element Pseudo a and its,Pseudo fuzzy set is a triplet of element and its two membership functions.,12928
may membership the may not functions dependent. be Both or,Both the membership functions may or may not be dependent.,12929
positive sense The has sense. is hypothesis every negative some that,The hypothesis is that every positive sense has some negative sense.,12930
"and So, another negative. has considered function positive membership one as as been","So, one membership function has been considered as positive and another as negative.",12931
models error-free NTB entire an algorithm scene The network. proposed as the,The proposed NTB algorithm models the entire scene as an error-free network.,12932
"detected. ""package these activities analyzing can By effectively transmission"" specific various processes, be","By analyzing these specific ""package transmission"" processes, various activities can be effectively detected.",12933
results demonstrate Experimental proposed effectiveness of our algorithm. the,Experimental results demonstrate the effectiveness of our proposed algorithm.,12934
today factual agents can answer Artificial questions.,Artificial agents today can answer factual questions.,12935
questions sense fall But require short they that common on reasoning.,But they fall short on questions that require common sense reasoning.,12936
semantic In - learned paper images leverage common we knowledge from i.e. sense this,In this paper we leverage semantic common sense knowledge learned from images - i.e.,12937
visual in paraphrasing. textual and two - visual tasks: fill-in-the-blank common sense,visual common sense - in two textual tasks: fill-in-the-blank and visual paraphrasing.,12938
visual scenes We as the abstraction. a imagine,We imagine the scenes as a visual abstraction.,12939
a on baseline strong text-only approach these outperforms tasks. Our,Our approach outperforms a strong text-only baseline on these tasks.,12940
Our publicly available. datasets code are and,Our code and datasets are publicly available.,12941
the comparison tracker difficult. This cross-paper makes,This makes the cross-paper tracker comparison difficult.,12942
"non-spatial running such found REM during results sleep. and behaviors as we Remarkably, similar wheel","Remarkably, we found similar results during non-spatial behaviors such as wheel running and REM sleep.",12943
used visualize detectors. feature algorithms We introduce spaces to object by,We introduce algorithms to visualize feature spaces used by object detectors.,12944
by natural images. method works multiple a feature to inverting Our back visual,Our method works by inverting a visual feature back to multiple natural images.,12945
"can of understanding gain intuitive spaces, visualizing systems. a By feature we more recognition","By visualizing feature spaces, we can gain a more intuitive understanding of recognition systems.",12946
"problem, i.e. determination the This highly to related is action-attribution","This determination is highly related to the action-attribution problem, i.e.",12947
the an The proof result support constructing one well-known of alternative theorem. latter deficiency might,The latter result might support constructing an alternative proof of the well-known deficiency one theorem.,12948
are illustrated notions examples. through developed and The results,The developed notions and results are illustrated through examples.,12949
be the MIIS Metropolis-within-Gibbs this more a method than efficient approach. case In can,In this case the MIIS method can be more efficient than a Metropolis-within-Gibbs approach.,12950
coarse-grained four within proteins a We citrate structure-based study model. synthase homodimeric,We study four citrate synthase homodimeric proteins within a structure-based coarse-grained model.,12951
tool efficient to mathematical Dempster-Shafer with uncertain is theory evidence deal information. an,Dempster-Shafer evidence theory is an efficient mathematical tool to deal with uncertain information.,12952
an is Dempster-Shafer in Decision-making issue on open theory. still BPA based evidence,Decision-making based on BPA is still an open issue in Dempster-Shafer evidence theory.,12953
to the Numerical given examples proposed show are approach.,Numerical examples are given to show the proposed approach.,12954
"of gene evolutionary expression have the modes remained However, controversial.","However, the evolutionary modes of gene expression have remained controversial.",12955
we of the dominates expression flies. evidence Here evolution in adaptation gene present that levels,Here we present evidence that adaptation dominates the evolution of gene expression levels in flies.,12956
years. progress shown recognition recent has Activity in impressive,Activity recognition has shown impressive progress in recent years.,12957
hand-centric activity detection. We benefits our the classification fine-grained for of and show approach,We show the benefits of our hand-centric approach for fine-grained activity classification and detection.,12958
them. can having data without recognize Using composites novel for training script data we,Using script data we can recognize novel composites without having training data for them.,12959
bi-level often for used of The or maximum multi-level thresholding images. entropy is principle,The maximum entropy principle is often used for bi-level or multi-level thresholding of images.,12960
"this For and on are Tsallis methods entropies. some available Shannon based purpose,","For this purpose, some methods are available based on Shannon and Tsallis entropies.",12961
"Kaniadakis and a them propose based on paper, method this entropy. In discuss we","In this paper, we discuss them and propose a method based on Kaniadakis entropy.",12962
regimes different which are different genetic phases favored. These strategies define in,These phases define different regimes in which different genetic strategies are favored.,12963
"it demonstrates temperature. real most has thermodynamic system importantly, the that a Perhaps","Perhaps most importantly, it demonstrates that the system has a real thermodynamic temperature.",12964
"temperature. normal to For is related replication, linearly temperature effective this","For normal replication, this temperature is linearly related to effective temperature.",12965
linear of does immune but relationship. response rescales The temperature not the change observed strength,The strength of immune response rescales temperature but does not change the observed linear relationship.,12966
"In study in we case. problem MBL the the more latter this paper, realistic","In this paper, we study the MBL problem in the latter more realistic case.",12967
gestures These system analyzed transformed will the be and by of as indicator participation.,These gestures will be analyzed by the system and transformed as indicator of participation.,12968
We use fingers. between ambiguities how context deal with efficiently show to also to,We also show how to use context efficiently to deal with ambiguities between fingers.,12969
for similarity generalization of a propose information We networks. measure SimRank heterogeneous,We propose a generalization of SimRank similarity measure for heterogeneous information networks.,12970
to in convex contrast which is has used rank. the previous This relaxations for work,This is in contrast to previous work which has used convex relaxations for the rank.,12971
the an upper bound this provide We error estimation model. matrix on under,We provide an upper bound on the matrix estimation error under this model.,12972
rank low algorithm is of $M^*$. This based factorization on,This algorithm is based on low rank factorization of $M^*$.,12973
performance real existing over with validate improved the data on and method methods. We synthetic,We validate the method on synthetic and real data with improved performance over existing methods.,12974
"the when are code are However, impractical high runs time-consuming. they","However, they are impractical when the code runs are high time-consuming.",12975
dimensions several such assess efficiency strategies. the in of illustrations sequential Numerical,Numerical illustrations in several dimensions assess the efficiency of such sequential strategies.,12976
understanding. and localization text for important video The and detection analysis is,The text detection and localization is important for video analysis and understanding.,12977
"of most single the or still in frame. approaches images text video However, detect scene","However, most of the approaches detect scene text in still images or single video frame.",12978
redundancy. differ in images from Videos temporal,Videos differ from images in temporal redundancy.,12979
"mechanostability theoretically. been far, assessed experimentally has So its or either not","So far, its mechanostability has not been assessed either experimentally or theoretically.",12980
The character initially type. to according automation widgets creating is the achieved by,The automation is achieved initially by creating widgets according to the character type.,12981
"effectiveness real experiments algorithm of the via on and our simulated demonstrate Finally, we data.","Finally, we demonstrate the effectiveness of our algorithm via experiments on simulated and real data.",12982
"self-improving intelligent security from software. address Finally, implications we","Finally, we address security implications from self-improving intelligent software.",12983
"both desirable But, methods for settings hospital are measuring in non-contact (e.g. vital signs","But, non-contact methods for measuring vital signs are desirable both in hospital settings (e.g.",12984
(e.g. for tracking ubiquitous and in-situ in NICU) health,in NICU) and for ubiquitous in-situ health tracking (e.g.,12985
phone on and mobile webcams). computers with,on mobile phone and computers with webcams).,12986
"Recently, vital non-contact camera-based monitoring feasible. to be have been sign shown","Recently, camera-based non-contact vital sign monitoring have been shown to be feasible.",12987
parameter estimation We static models. consider state-space Bayesian online for,We consider Bayesian online static parameter estimation for state-space models.,12988
"algorithm the to of under addition, provide assumptions, expectations estimates shown w.r.t. consistent In is","In addition, under assumptions, the algorithm is shown to provide consistent estimates of expectations w.r.t.",12989
is The implemented Bayesian algorithm statistical on several models.,The algorithm is implemented on several Bayesian statistical models.,12990
"updating Finally, is a yet designed for training. lazy scheme effective CNN","Finally, a lazy yet effective updating scheme is designed for CNN training.",12991
solve to developed involved. algorithm A the optimization practical is,A practical algorithm is developed to solve the optimization involved.,12992
"straightforward convergent. This and monotonic fast, is algorithm to implement,","This algorithm is fast, straightforward to implement, and monotonic convergent.",12993
"the to well sense. theoretically be defined Furthermore, shown a proposed in is stable methodology","Furthermore, the proposed methodology is theoretically shown to be stable in a well defined sense.",12994
"experiments, demonstrated Its sequence is inpainting. empirical image simulation including promising performance of via a","Its promising empirical performance is demonstrated via a sequence of simulation experiments, including image inpainting.",12995
techniques deblurring roles processing in important play applications. image Image many,Image deblurring techniques play important roles in many image processing applications.,12996
algorithm. proposed effectiveness the the demonstrate results of Experimental,Experimental results demonstrate the effectiveness of the proposed algorithm.,12997
in arising be subtleties the Computational case discussed. will former,Computational subtleties arising in the former case will be discussed.,12998
"clustering seem like For may such natural a approach. purposes,","For such purposes, clustering may seem like a natural approach.",12999
"in validate $\delta$-medoids namely empirically We music and motion domains, analysis. analysis two","We empirically validate $\delta$-medoids in two domains, namely music analysis and motion analysis.",13000
regularization of surface functional. area used function The as was the,The function of surface area was used as the regularization functional.,13001
dynamic because estimate previous variations. image current regularization on character depends has and The,The regularization has dynamic character because depends on current and previous image estimate variations.,13002
the distribution Monte Currently Carlo probability used coincidences. of methods approximate,Currently used Monte Carlo methods approximate the probability distribution of coincidences.,13003
an is LSTM for recognition. ideal candidate text Deep,Deep LSTM is an ideal candidate for text recognition.,13004
"becomes difficult range intractable. segmentation, context long learning very is and computationally Without","Without segmentation, learning very long range context is difficult and becomes computationally intractable.",13005
"decisions alternative Therefore, are soft at the level. pre-processing needed","Therefore, alternative soft decisions are needed at the pre-processing level.",13006
filters errors. verification language outputs segmentation framework models using eliminates wrong Thus recognition and the,Thus the verification framework using language models eliminates wrong segmentation outputs and filters recognition errors.,13007
sequential standard Resampling is generally Monte Carlo in filters step methods. a more particle and,Resampling is a standard step in particle filters and more generally sequential Monte Carlo methods.,13008
"particles. resampling We for algorithm, called chopthin, an present weighted","We present an algorithm, called chopthin, for resampling weighted particles.",13009
resampling algorithm outperforms studies standard Simulation chopthin the consistently that show methods.,Simulation studies show that the chopthin algorithm consistently outperforms standard resampling methods.,13010
implicitly a bound size. the sample guarantees effective lower on It,It implicitly guarantees a lower bound on the effective sample size.,13011
"making The can useful. implemented it algorithm practically efficiently, be","The algorithm can be implemented efficiently, making it practically useful.",13012
expected the of show the effort We particles. linear in that computational number is,We show that the expected computational effort is linear in the number of particles.,13013
"available. CRAN), R for (on Python Matlab Implementations C++, are and","Implementations for C++, R (on CRAN), Python and Matlab are available.",13014
"article, of estimation we problem images. this tackle monocular depth the from In single","In this article, we tackle the problem of depth estimation from single monocular images.",13015
are common field processing life. Image every our in applications daily of,Image processing applications are common in every field of our daily life.,13016
"Therefore, solutions processing. hardware heterogeneous on are application-specialized image for based used architectures","Therefore, application-specialized hardware solutions based on heterogeneous architectures are used for image processing.",13017
"scratch which and error-prone systems processing inflexible. time-consuming, Nowadays, image from started is are","Nowadays, image processing systems are started from scratch which is time-consuming, error-prone and inflexible.",13018
This systems. introduces paper top new approach image down a for holistic processing,This paper introduces a new holistic top down approach for image processing systems.,13019
"architecture. to want and close Second, between system abstract we the gap system the the","Second, we want to close the gap between the abstract system and the system architecture.",13020
Information data with and research field such are visualization two mining goal.,Information visualization and data mining are two research field with such goal.,13021
"deep improved trained the considerably performance Recently, object from recognition. have CNNs of large datasets","Recently, deep CNNs trained from large datasets have considerably improved the performance of object recognition.",13022
"using non-uniform image motion prior. a is removed patch-level blur Finally, model by deblurring","Finally, motion blur is removed by a non-uniform deblurring model using patch-level image prior.",13023
or computer-generated with The results drift only are surrogate containing compared only time series microsaccades.,The results are compared with computer-generated surrogate time series containing only drift or only microsaccades.,13024
analyses components drift. to between due Our show a binocular dependence FEM predominantly velocity,Our analyses show a binocular dependence between FEM velocity components predominantly due to drift.,13025
SVMs Ensemble calibrating Exemplar the We method of for present a model.,We present a method for calibrating the Ensemble of Exemplar SVMs model.,13026
with We on descriptors. trained experiment state-of-the-art CNN EE-SVM,We experiment with EE-SVM trained on state-of-the-art CNN descriptors.,13027
"normalized algorithm. fast this a first purpose, For we develop cuts","For this purpose, we first develop a fast normalized cuts algorithm.",13028
effective propose a that high-performance multiscale information. makes then of use segmenter hierarchical We,We then propose a high-performance hierarchical segmenter that makes effective use of multiscale information.,13029
their animals on Humans sensory noisy and input. base other decisions,Humans and other animals base their decisions on noisy sensory input.,13030
been the computations that work such Much decisions. devoted understanding underly has to therefore,Much work has therefore been devoted to understanding the computations that underly such decisions.,13031
"understood. is impact well the judgments of However, in correlations on sensory perceptual noise not","However, the impact of correlations in sensory noise on perceptual judgments is not well understood.",13032
impact correlations Here stimulus decision we noise in examine correlations sensory together how with making.,Here we examine how stimulus correlations together with correlations in sensory noise impact decision making.,13033
distractors are of correlated. the and We measurements assume both the the that observer's stimuli,We assume that both the distractors and the observer's measurements of the stimuli are correlated.,13034
impact structured in strongly therefore can to responses perceptual Correlations judgments. stimuli neural,Correlations in neural responses to structured stimuli can therefore strongly impact perceptual judgments.,13035
$T$ for means unique that Here phylogeny is perfect characters. these the `capture',Here `capture' means that $T$ is the unique perfect phylogeny for these characters.,13036
combinatorial of method. Our this proof on probabilistic result based is the short,Our short proof of this combinatorial result is based on the probabilistic method.,13037
a category computer localization Object problem is vision. in challenging,Object category localization is a challenging problem in computer vision.,13038
annotations object supervised requires box instances. training bounding Standard of,Standard supervised training requires bounding box annotations of object instances.,13039
is annotation time-consuming in learning. process sidestepped weakly supervised This,This time-consuming annotation process is sidestepped in weakly supervised learning.,13040
"information. and In layer adopt appearance feature motion information which concatenation combines we a addition,","In addition, we adopt a feature concatenation layer which combines appearance information and motion information.",13041
an the for propose problem. colony We ant optimization algorithm,We propose an ant colony optimization algorithm for the problem.,13042
role plays industry. an research and important in HFR both biometrics,HFR plays an important role in both biometrics research and industry.,13043
degree overestimation when significant of spatial marked. heterogeneity The becomes is,The degree of overestimation becomes significant when spatial heterogeneity is marked.,13044
. on those examples accelerating illustrate a of We series features,We illustrate those accelerating features on a series of examples .,13045
"and techniques, fast memory to efficient. is ConF accurate, Compared nearest-neighbour more standard","Compared to standard nearest-neighbour techniques, ConF is more accurate, fast and memory efficient.",13046
"discarding inaccurate frequencies, clonal etc. computation specific genotypes clone of","inaccurate computation of clonal frequencies, discarding clone specific genotypes etc.",13047
tested also our We simulated different on model data.,We also tested our model on different simulated data.,13048
heterogeneity. methods that results shows HetFHMM outperforms tumor infer recent which The,The results shows that HetFHMM outperforms recent methods which infer tumor heterogeneity.,13049
"research novel area. heterogeneity tumor Therefore, approach is in HetFHMM a","Therefore, HetFHMM is a novel approach in tumor heterogeneity research area.",13050
This around ability sustained distracting of addition object. dots be with the can,This ability can be sustained with addition of distracting dots around the object.,13051
"has capability such been algorithmically computers. reproduced on However, not","However, such capability has not been reproduced algorithmically on computers.",13052
into new a set subsets. algorithm We non-disjoint introduce that multiple of groups dots a,We introduce a new algorithm that groups a set of dots into multiple non-disjoint subsets.,13053
a spanning into It the cue. dots connects the using tree proximity,It connects the dots into a spanning tree using the proximity cue.,13054
is O($n$). of number generated polygons The,The number of polygons generated is O($n$).,13055
both these effectiveness show algorithms. and We of to images synthetic used natural,We used both natural and synthetic images to show effectiveness of these algorithms.,13056
"considerations. introduce normative a formal distinction between we statistical Further, and","Further, we introduce a formal distinction between statistical and normative considerations.",13057
 single-target tracker paper  problem addresses This performance the evaluation. of,  This paper addresses the problem of single-target tracker performance evaluation.,13058
statistical equivalence ranking-based The significance and in methodology tracker practical terms differences. of addresses,The ranking-based methodology addresses tracker equivalence in terms of statistical significance and practical differences.,13059
introduced. annotations dataset several visual A with with fully-annotated is per-frame attributes,A fully-annotated dataset with per-frame annotations with several visual attributes is introduced.,13060
the This to sophistically most dataset makes annotated constructed date. and it,This makes it the most sophistically constructed and annotated dataset to date.,13061
as presented of third-party easy system trackers is allowing multi-platform evaluation A well. integration,A multi-platform evaluation system allowing easy integration of third-party trackers is presented as well.,13062
difficulty the is of perspective An of analysis dataset the from exhaustive carried out. tracking,An exhaustive analysis of the dataset from the perspective of tracking difficulty is carried out.,13063
new comparison a To facilitate proposed. technique is tracker performance visualization,To facilitate tracker comparison a new performance visualization technique is proposed.,13064
Discrete-time a of approach for capturing system. models are dynamics population traditional host-parasitoid the,Discrete-time models are the traditional approach for capturing population dynamics of a host-parasitoid system.,13065
and debate. diversification have Sino-Tibetan populations been The hot a of long-standing origin,The origin and diversification of Sino-Tibetan populations have been a long-standing hot debate.,13066
"However, Tibetan of the populations information genetic topic from this clear. limited far keeps","However, the limited genetic information of Tibetan populations keeps this topic far from clear.",13067
identification. for also dataset valuable are generated The here forensic,The dataset generated here are also valuable for forensic identification.,13068
"the degraded when However, dealing images. performance color PCANet be with may of","However, the performance of PCANet may be degraded when dealing with color images.",13069
method CEoptim optimization. the the implementation package R for new The CE of provides,The new package CEoptim provides the R implementation of the CE method for optimization.,13070
We some describe optimization useful for CE general as the and modifications. well methodology,We describe the general CE methodology for optimization and well as some useful modifications.,13071
compromising orders show results two of without of performance. complexity reduction Simulation a unmixing magnitude,Simulation results show a complexity reduction of two orders of magnitude without compromising unmixing performance.,13072
"is wide-spread Bistability bacteria considered e.g. cells, and eukaryotic among useful","Bistability is considered wide-spread among bacteria and eukaryotic cells, useful e.g.",13073
"for switching. induction, epigenetic and bet hedging, enzyme","for enzyme induction, bet hedging, and epigenetic switching.",13074
"However, deterministic well-mixed has described with this stochastic or been dynamic models. mostly phenomenon","However, this phenomenon has mostly been described with deterministic dynamic or well-mixed stochastic models.",13075
withdrawn the by authour. This been has paper,This paper has been withdrawn by the authour.,13076
on synthetic real-world well results as data. encouraging as Experiments show,Experiments show encouraging results on synthetic as well as real-world data.,13077
cut proposes of This algorithm for a clustering. variant paper the spectral normalized,This paper proposes a variant of the normalized cut algorithm for spectral clustering.,13078
to performance the of is presented the empirical study examine algorithm. An,An empirical study is presented to examine the performance of the algorithm.,13079
of data machine applications occur categorical learning. many Multivariate in,Multivariate categorical data occur in many applications of machine learning.,13080
variables is difficulties of of vectors categorical One sparsity. these main with the,One of the main difficulties with these vectors of categorical variables is sparsity.,13081
data. have significant gained improvement in tasks supervised models with Recent this,Recent models have gained significant improvement in supervised tasks with this data.,13082
models between continuous them. to observations a similarities embed in capture These space,These models embed observations in a continuous space to capture similarities between them.,13083
sampling for variational based inference We recent in developments based derive our model on inference.,We derive inference for our model based on recent developments in sampling based variational inference.,13084
storage in stored that are table All hash possible minimum space. a requires transitions state,All possible state transitions are stored in a hash table that requires minimum storage space.,13085
The each is individual the of predicted of probability. the confidence player's product team's formation,The confidence of the predicted team's formation is the product of each individual player's probability.,13086
is low-level framework visual and on high-level hand-crafted the basis multiple Our features. built of,Our framework is built on the basis of multiple low-level hand-crafted and high-level visual features.,13087
is proposals basic training and iterate generating idea networks. The to between automatically region convolutional,The basic idea is to iterate between automatically generating region proposals and training convolutional networks.,13088
"These vise masks the for improving gradually and versa. segmentation recover steps two networks,","These two steps gradually recover segmentation masks for improving the networks, and vise versa.",13089
problems optimization. Many formulated high sparse nonconvex are as dimensional learning,Many high dimensional sparse learning problems are formulated as nonconvex optimization.,13090
"the limits relaxations. In paper, this of we statistical study convex","In this paper, we study the statistical limits of convex relaxations.",13091
illustrated Bayesian is regression spline by example. performance Its free-knot a,Its performance is illustrated by a Bayesian free-knot spline regression example.,13092
unordered experiment sequentially this and on both Internet We captured method images. images,We experiment this method on both sequentially captured images and unordered Internet images.,13093
"and robustness, efficiency. in accuracy, The demonstrate experiments its strength","The experiments demonstrate its strength in robustness, accuracy, and efficiency.",13094
systems recognition increasing Datasets steadily object for training size. in are,Datasets for training object recognition systems are steadily increasing in size.,13095
the We discriminatively features. focus templates oriented defined on gradient paradigm popular trained of on,We focus on the popular paradigm of discriminatively trained templates defined on oriented gradient features.,13096
Most think works. of basic us about we some how know facts time,Most of us think we know some basic facts about how time works.,13097
excellent of These our ordered understanding are events. ones intuitions for conception conscious,These intuitions are excellent ones for understanding our conscious conception of ordered events.,13098
"motor proteins most transported In cells, of by along microtubule. cargos are","In cells, most of cargos are transported by motor proteins along microtubule.",13099
"address particle is this problem, this TASEP discussed in with species two To study.","To address this problem, TASEP with two particle species is discussed in this study.",13100
each to get species methods Theoretical densities particle are provided. of,Theoretical methods to get densities of each particle species are provided.,13101
"are phase related transition particle Using methods, of properties obtained. these densities","Using these methods, phase transition related properties of particle densities are obtained.",13102
diagrams parameters for presented. are typical also Phase model,Phase diagrams for typical model parameters are also presented.,13103
the The of models. digestion the complex development complexity has anaerobic of motivated process,The complexity of the anaerobic digestion process has motivated the development of complex models.,13104
"it emergent intractable stability However, and complexity to makes pin-point this behaviour.","However, this complexity makes it intractable to pin-point stability and emergent behaviour.",13105
"Anaerobic Model No. widely Here, used Digestion the","Here, the widely used Anaerobic Digestion Model No.",13106
"stable were degradation, always non-oscillatory. and chlorophenol Parameterised phenol and steady-states for","Parameterised for phenol and chlorophenol degradation, steady-states were always stable and non-oscillatory.",13107
Monte samples property chain This distribution posterior Carlo from (MCMC) non-identifiable. makes Markov the simulated,This property makes Markov chain Monte Carlo (MCMC) samples simulated from the posterior distribution non-identifiable.,13108
"\pkg{label.switching} this paper, is the package In introduced.","In this paper, the \pkg{label.switching} package is introduced.",13109
"using the can reordered parametric be of sample. inferred interest Then, any function MCMC","Then, any parametric function of interest can be inferred using the reordered MCMC sample.",13110
Obstacle role (USV). plays important unmanned surface vehicles in detection an,Obstacle detection plays an important role in unmanned surface vehicles (USV).,13111
of by the This constrained unsupervised paper problem segmentation. online addresses detection,This paper addresses the problem of online detection by constrained unsupervised segmentation.,13112
"of for segmentation. histogram-based about the distances is regularized optimal-transport convex, This use work image","This work is about the use of regularized optimal-transport distances for convex, histogram-based image segmentation.",13113
"for documents shredding is several the damaged reasons, reason hands. Out of main by getting","Out of several reasons, the main reason for documents getting damaged is shredding by hands.",13114
recovery is and time tedious such damaged consuming task. of Manual document,Manual recovery of such damaged document is tedious and time consuming task.,13115
"a technique using simple order to we this, In called Image are Mosaicing. implement","In order to implement this, we are using a simple technique called Image Mosaicing.",13116
new is complete image torn technique a using this developed fragments. or In more two,In this technique a complete new image is developed using two or more torn fragments.,13117
useful facial information can provide expressions. for recognizing features Temporal,Temporal information can provide useful features for recognizing facial expressions.,13118
"effort. a However, design useful of manually lot to requires features","However, to manually design useful features requires a lot of effort.",13119
different on Our two deep models. network based is,Our deep network is based on two different models.,13120
"with several other. we each Through showed that experiments, two models cooperate the","Through several experiments, we showed that the two models cooperate with each other.",13121
discriminative space. into of the It a data special case the transforming/projecting subsumes,It subsumes the special case of transforming/projecting the data into a discriminative space.,13122
paper facial state-of-the-art expression In recognition. two this person-independent descriptors we for propose,In this paper we propose two state-of-the-art descriptors for person-independent facial expression recognition.,13123
vectors Neighbor expressions. the the the recognize After used K-Nearest to was extracting feature classifier,After extracting the feature vectors the K-Nearest Neighbor classifier was used to recognize the expressions.,13124
interactions in key role a cell Protein-peptide play functions.,Protein-peptide interactions play a key role in cell functions.,13125
"is discovery the characterization, of though drugs. for challenging, important new Their structural","Their structural characterization, though challenging, is important for the discovery of new drugs.",13126
two model. improve contributions to We and analyze RCPN further novel propose the,We analyze RCPN and propose two novel contributions to further improve the model.,13127
We develop mathematical these combines discovery a of model that all of scientific elements.,We develop a mathematical model of scientific discovery that combines all of these elements.,13128
bring research. results the Our about communication also to clarity verbal debates of,Our results also bring clarity to verbal debates about the communication of research.,13129
plasticity Hebbian the that to example. show learn by allows the networks We model consistency,We show that Hebbian plasticity allows the networks to learn the consistency model by example.,13130
well approximated the We that by dynamics sampling process. show stochastic a are search,We show that the search dynamics are well approximated by a stochastic sampling process.,13131
"if the then positively diminishing habitats are autocorrelated, fishery with increases profit returns. However,","However, if habitats are positively autocorrelated, then the fishery profit increases with diminishing returns.",13132
"Consequently, methods. are and analysis basis of respective quantitative the on qualitative results compared of","Consequently, results are compared on the basis of quantitative and qualitative analysis of respective methods.",13133
"predicted More networks. specifically, by are the outputs separate two structure","More specifically, the structure outputs are predicted by two separate networks.",13134
robustness state-of-the-art in speed. and the Experiments show MODS outperforms the that,Experiments show that the MODS outperforms the state-of-the-art in robustness and speed.,13135
"the size are neighborhoods of the By compared, method that scale-dependent. is implicitly the adjusting","By adjusting the size of the neighborhoods that are compared, the method is implicitly scale-dependent.",13136
on or small able to also gradients. changes focus is large either It,It is also able to focus on either small changes or large gradients.,13137
are processes a Branching with (CTMCs) chains ubiquitous applications. of Markov continuous-time class,Branching processes are a class of continuous-time Markov chains (CTMCs) with ubiquitous applications.,13138
challenging We dataset. demonstrate the our a on accuracy and approach of robustness,We demonstrate the robustness and accuracy of our approach on a challenging dataset.,13139
a I to relations priori experienced candidates answer former cover questions. for,I cover a priori relations to experienced answer candidates for former questions.,13140
"concentration novel missing We for i.e. concerned are inequalities obtaining the with mass,","We are concerned with obtaining novel concentration inequalities for the missing mass, i.e.",13141
outcomes mass the of observed the probability not total sample. in the,the total probability mass of the outcomes not observed in the sample.,13142
difference have in sums terms whose large magnitude.,sums whose terms have large difference in magnitude.,13143
is should dependent and a parameter application of Which errors be the measure. tolerable are,Which errors are tolerable is application dependent and should be a parameter of the measure.,13144
The should measure. be error so reflected by do time the to needed,The time needed to do so should be reflected by the error measure.,13145
solve and pose simultaneously allows for The method shape parameters. to,The method allows to simultaneously solve for shape and pose parameters.,13146
"the our for it applicable speed method, of applications. of is Because realtime","Because of the speed of our method, it is applicable for realtime applications.",13147
pseudo-input functions cannot the sparse to Standard handle process (GP) Gaussian well. complex approximations,Standard sparse pseudo-input approximations to the Gaussian process (GP) cannot handle complex functions well.,13148
to to are alternatives but over-fit. attempt spectrum known Sparse this answer,Sparse spectrum alternatives attempt to answer this but are known to over-fit.,13149
"on which The distribution has random is function a variational posterior, covariance placed. a","The random covariance function has a posterior, on which a variational distribution is placed.",13150
function transforms the The random fit variational to data. covariance the distribution,The variational distribution transforms the random covariance function to fit the data.,13151
standard than complex functions avoids and over-fitting. approaches approximation better captures Our,Our approximation captures complex functions better than standard approaches and avoids over-fitting.,13152
of method the instead structure plants their analyses using silhouettes. Our only,Our method instead analyses the structure of plants using only their silhouettes.,13153
"may appear. feedback, selectivity a in sharp strong For neuronal resonance the spectral","For strong feedback, a sharp resonance in the spectral neuronal selectivity may appear.",13154
reconstruct from sensing scenes static Existing systems SMC techniques (CS). compressive using,Existing SMC systems reconstruct static scenes using techniques from compressive sensing (CS).,13155
"quality. poor however, For acquisition methods and videos, existing deliver recovery","For videos, however, existing acquisition and recovery methods deliver poor quality.",13156
Machine for are recognition problems. methods most used today learning,Machine learning methods are used today for most recognition problems.,13157
Artificial size of to data We use enhance set. training small System Principles Immune,We use Artificial Immune System Principles to enhance small size of training data set.,13158
to further model emphasize statistics. The atoms parts both in dictionary weighted adaptively are their,The dictionary atoms in both parts are further adaptively weighted to emphasize their model statistics.,13159
"selectivity. stimulate or sparsity under are networks of To this constraint behavior, trained mathematical networks","To stimulate this behavior, networks are trained networks under mathematical constraint of sparsity or selectivity.",13160
"address does mature, rich literature, driving the and rash not Car though processing Image problem.","Car Image processing literature, though rich and mature, does not address the rash driving problem.",13161
"(FEP) on the neuroscience ""free-energy principle'' so-called increasingly in cognitive is high-profile. becoming Research","Research on the so-called ""free-energy principle'' (FEP) in cognitive neuroscience is becoming increasingly high-profile.",13162
transmitting as flavivirus. known Aedes vector dengue is aegypti responsible the,Aedes aegypti is known as the responsible vector transmitting dengue flavivirus.,13163
basic equilibria offspring mosquito shown. The analysis and obtained was was number of,The basic mosquito offspring number was obtained and analysis of equilibria was shown.,13164
two-valued introduced fitness quasispecies classical is landscape the for model. A Eigen's,A two-valued fitness landscape is introduced for the classical Eigen's quasispecies model.,13165
sequences for for of given. threshold error are Sufficient so-called orbits conditions the,Sufficient conditions for the so-called error threshold for sequences of orbits are given.,13166
optimization problem. Vehicles dispersed Air is targets an to (AVs) geographically prosecute important Multiple,Multiple Air Vehicles (AVs) to prosecute geographically dispersed targets is an important optimization problem.,13167
"each classification, attack successively viz., are performed multiple target. on verification tasks target and Associated","Associated multiple tasks viz., target classification, attack and verification are successively performed on each target.",13168
tasks are Coupled involving time task order constraints addressed. and,Coupled tasks involving time and task order constraints are addressed.,13169
"have sufficient existence When solution AVs of is optimal endurance, guaranteed.","When AVs have sufficient endurance, existence of optimal solution is guaranteed.",13170
annotations Yet are require expensive bounding-box to learn. expensive and computationally these still,Yet these require expensive bounding-box annotations and are still computationally expensive to learn.,13171
well-known monocular is tracking images surface Deformable from to under-constrained. be,Deformable surface tracking from monocular images is well-known to be under-constrained.,13172
"state-of-the-art framework for Our textured, achieves both surfaces. poorly and well occluded results","Our framework achieves state-of-the-art results for both well and poorly textured, occluded surfaces.",13173
face system. for a present approach recognition We new,We present a new approach for face recognition system.,13174
active decades. of in topic has past a Integer Optimization research been Mixed,Mixed Integer Optimization has been a topic of active research in past decades.,13175
Statistical data. regression massive solve of and problems used been It to has involving classification,It has been used to solve Statistical problems of classification and regression involving massive data.,13176
"However, huge vagueness real inherent of degree life is in there present data. an","However, there is an inherent degree of vagueness present in huge real life data.",13177
impreciseness handled Fuzzy Sets. This is by,This impreciseness is handled by Fuzzy Sets.,13178
discrete The of exploits problem. character methodology,The methodology exploits discrete character of problem.,13179
way solved limits. scale are within large practical problems In this,In this way large scale problems are solved within practical limits.,13180
images and The vertical are analyzing unfolding these horizontal images. before of done,The vertical and horizontal unfolding of images are done before analyzing these images.,13181
results The the spatial of confirms heterogeneity in mismatch domain.,The mismatch of results confirms the heterogeneity in spatial domain.,13182
experiments. by accuracy and the demonstrate real our of synthetic of means We solution,We demonstrate the accuracy of our solution by means of synthetic and real experiments.,13183
paper means studies efficient This intra-category for detection. with object dealing in diversity,This paper studies efficient means for dealing with intra-category diversity in object detection.,13184
fast features detection AdaBoost detection. is employed lookup pixel scheme for An with,An AdaBoost detection scheme is employed with pixel lookup features for fast detection.,13185
is optimization on the via the a coarse step. Augmenting details mesh fine final done,Augmenting the fine details on the coarse mesh is done via a final optimization step.,13186
bearings using acquired is data from test faulty a Vibration rig.,Vibration data is acquired from faulty bearings using a test rig.,13187
smoothing feature distributions. The involves of processing,The processing involves smoothing of feature distributions.,13188
edge-preserving to have spread due Bilateral properties. use filters their wide,Bilateral filters have wide spread use due to their edge-preserving properties.,13189
construction the efficient We build on lattice for permutohedral filtering.,We build on the permutohedral lattice construction for efficient filtering.,13190
This architectures. into view provides new ways structure encode to network model,This view provides new ways to encode model structure into network architectures.,13191
usage filters. A validates of of forms set experiments empirically diverse the general of,A diverse set of experiments empirically validates the usage of general forms of filters.,13192
"that a Second, consideration independence. we degree takes of of new rule combination propose sources'","Second, we propose a new combination rule that takes consideration of sources' degree of independence.",13193
functions. illustrated generated mass The is proposed on method,The proposed method is illustrated on generated mass functions.,13194
parametric address and models to considered hyperspectral been Different have unmixing problems. nonparametric,Different parametric and nonparametric models have been considered to address hyperspectral unmixing problems.,13195
one is The the mixing linear model. simplest,The simplest one is the linear mixing model.,13196
"has nonlinear. also been phenomena it recognized can be that Nevertheless, mixing","Nevertheless, it has been recognized that mixing phenomena can also be nonlinear.",13197
"propose mixed we In pixels. for technique a this nonlinearly detecting paper,","In this paper, we propose a technique for detecting nonlinearly mixed pixels.",13198
given. {\sigma}_BS and analysis I_B of been A detailed has,A detailed analysis of I_B and {\sigma}_BS has been given.,13199
A both and coarse-to-fine segmentation approach of structures. smooth accurate ensures a,A coarse-to-fine approach ensures a smooth and accurate segmentation of both structures.,13200
segmentation A performed to structures. snake-based is of the segmentation both refine,A snake-based segmentation is performed to refine the segmentation of both structures.,13201
logical ability reasoning cognition with of Higher-level and common sense. answering includes question the,Higher-level cognition includes logical reasoning and the ability of question answering with common sense.,13202
"specificity, and deontic model normative reasoning, construction. logic,","specificity, deontic logic, normative reasoning, and model construction.",13203
"obtained are mass, concentration missing inequalities i.e. Novel the for","Novel concentration inequalities are obtained for the missing mass, i.e.",13204
sample. mass not observed total the of the probability outcomes the in,the total probability mass of the outcomes not observed in the sample.,13205
The converted colored gray algorithm been processes without to scale. images developed directly,The developed algorithm processes colored images directly without been converted to gray scale.,13206
The evolution two mutation of selection. and are classical forces,The two classical forces of evolution are mutation and selection.,13207
mutation interactions also allow Rather rates. epistatic to we affect,Rather we allow epistatic interactions to also affect mutation rates.,13208
dynamics. can lead This qualitatively non-trivial to,This can lead to qualitatively non-trivial dynamics.,13209
"valleys are negligible overcome fitness accessible. become in SSWM-regime, Pathways, can the that and","Pathways, that are negligible in the SSWM-regime, can overcome fitness valleys and become accessible.",13210
"found between and were Various chromosomes both non-random within genome, chromosomes. associations the across","Various non-random associations were found across the genome, both within chromosomes and between chromosomes.",13211
"That and recognize to animation to why, is characters. develop need we detect method a","That is why, we need to develop a method to detect and recognize animation characters.",13212
important steps one most the is Skin this detection way. in of,Skin detection is one of the most important steps in this way.,13213
"leather, etc. example hair other (for shirt, Moreover, things many","Moreover, many other things (for example leather, shirt, hair etc.",13214
"similar skin, which are color have ), skin. can to not","), which are not skin, can have color similar to skin.",13215
relations. on values based Our RGB their comparative are methods and,Our methods are based on RGB values and their comparative relations.,13216
strategies. species exhibit two fighting The different considered,The two considered species exhibit different fighting strategies.,13217
experiments fluctuations a observed in model algorithm. by Gillespie of We means the standard the,We model the fluctuations observed in the experiments by means of a standard Gillespie algorithm.,13218
make steps implementations of available. We these also publicly preprocessing,We make implementations of these preprocessing steps also publicly available.,13219
of important is Estimation imaging. dynamic response medical in task an functions,Estimation of response functions is an important task in dynamic medical imaging.,13220
"estimated source the response and deconvolution. are Therefore, separation functions blind via","Therefore, the response functions are estimated via blind source separation and deconvolution.",13221
response the used heavily Performance depends the this functions. of algorithm of models on,Performance of this algorithm heavily depends on the used models of the response functions.,13222
"of In using functions Bayesian this we priors. non-parametric study paper, the response estimation","In this paper, we study estimation of the response functions using non-parametric Bayesian priors.",13223
source within used These hierarchical deconvolution priors blind are the of assumptions algorithm. separation and,These assumptions are used within hierarchical priors of the blind source separation and deconvolution algorithm.,13224
that of estimation both is response priors shown flexible It improve in functions non-parametric cases.,It is shown that flexible non-parametric priors improve estimation of response functions in both cases.,13225
freely resulting for of algorithms download. MATLAB is implementation the available,MATLAB implementation of the resulting algorithms is freely available for download.,13226
an the in All framework's of to fashion. pipeline automatic be designed stages used are,All stages of the framework's pipeline are designed to be used in an automatic fashion.,13227
"First, the focus selected gradient optimal using the is method. absolute","First, the optimal focus is selected using the absolute gradient method.",13228
a model style line the flexible data. of for streamline visualization illustrative present We,We present a flexible illustrative line style model for the visualization of streamline data.,13229
way generic add There system. to to is any feature a new a,There is a generic way to add any new feature to a system.,13230
a yield non-intuitive is or and may brute-force method a meaningful This result.,This is a brute-force method and may yield a non-intuitive or meaningful result.,13231
on translate. we we depends get Of how what course,Of course what we get depends on how we translate.,13232
is methodology conclusion. general Our the in in discussed,Our methodology in general is discussed in the conclusion.,13233
"exhibit chain pseudo-marginal towards its mixing target. and However, Markov the can convergence poor slow","However, the pseudo-marginal Markov chain can exhibit poor mixing and slow convergence towards its target.",13234
article w.r.t. expectations we In the approximation consider of this,In this article we consider the approximation of expectations w.r.t.,13235
"i.i.d. In problems practical an of one many cannot interest, achieve","In many practical problems of interest, one cannot achieve an i.i.d.",13236
the associated sequence sampling probability of distributions. of,sampling of the associated sequence of probability distributions.,13237
"positive-definite (SPDMs). of symmetric, matrices set the","the set of symmetric, positive-definite matrices (SPDMs).",13238
"execution requires one invariant rates, classes trajectories. Since metrics to action for comparing rate-invariant are","Since action classes are invariant to execution rates, one requires rate-invariant metrics for comparing trajectories.",13239
of trajectories re-parameterization. comparison invariant to the This their resulting makes,This makes the resulting comparison of trajectories invariant to their re-parameterization.,13240
results literature. than or achieve comparable to either current we both the cases In better,In both cases we achieve results either comparable to or better than the current literature.,13241
nature. throughout Territoriality phenomenon is exhibited a,Territoriality is a phenomenon exhibited throughout nature.,13242
to has mathematical relied area. are upon animals attracted central assuming Previous analysis a,Previous mathematical analysis has relied upon assuming animals are attracted to a central area.,13243
intrinsic same of object low-dimensional spaces the in images on manifolds lie Multi-view (e.g. descriptor,Multi-view images of the same object lie on intrinsic low-dimensional manifolds in descriptor spaces (e.g.,13244
topology the manifolds These same geometrically share object despite being different.,These object manifolds share the same topology despite being geometrically different.,13245
be manifold of version represented manifold. unified Each deformed a can a object as,Each object manifold can be represented as a deformed version of a unified manifold.,13246
be its by manifold. unified mapping/reconstruction object manifolds homeomorphic from the thus The can parameterized,The object manifolds can thus be parameterized by its homeomorphic mapping/reconstruction from the unified manifold.,13247
We several achieve experiments extensive on and perform datasets results. state-of-the-art challenging,We perform extensive experiments on several challenging datasets and achieve state-of-the-art results.,13248
"feature, a markings For to Lane this edge robust purpose Laplacian detected are using shadowing.","For this purpose Lane markings are detected using a Laplacian edge feature, robust to shadowing.",13249
to fitted and model are parabolic assumed robust be using Lanes as RANSAC.,Lanes are assumed to be as parabolic model and fitted using robust RANSAC.,13250
recently We mechanism. as on an theory interpretation Misrepair-accumulation introduced aging,We recently introduced Misrepair-accumulation theory as an interpretation on aging mechanism.,13251
these to covariance construct convolutions. model and coregionalization process of Alternatives functions are linear the,Alternatives to construct these covariance functions are the linear model of coregionalization and process convolutions.,13252
play very driving a and avoiding vital signs accident. safe in Traffic role,Traffic signs play a very vital role in safe driving and avoiding accident.,13253
"improved is Our classification accuracy. sign and shape algorithm two parts, of consist classification for","Our algorithm is consist of two parts, shape classification and sign classification for improved accuracy.",13254
prevalent vision. computer become have Superpixels in,Superpixels have become prevalent in computer vision.,13255
have superpixels Markov random with field also models. combined People (MRF),People have also combined superpixels with Markov random field (MRF) models.,13256
resultant superpixel-level for submodular MRF. energy pixel-level also remains a The submodular MRF,The resultant superpixel-level MRF energy also remains submodular for a submodular pixel-level MRF.,13257
hence in way a to formulate MRF energy gives superpixel-level. handy The derived formula us,The derived formula hence gives us a handy way to formulate MRF energy in superpixel-level.,13258
"the vision we In our the computer demonstrate approach of experiments, efficacy on problems. several","In the experiments, we demonstrate the efficacy of our approach on several computer vision problems.",13259
"to the variances. of is enigmatic structural revealed the be previously covariance Furthermore, difference","Furthermore, the previously enigmatic structural covariance is revealed to be the difference of variances.",13260
Dissimilarity the We anti/symmetric specifically term for SSIM (DQ) the derived Quotient use NVF.,We use the term Dissimilarity Quotient (DQ) for the specifically anti/symmetric SSIM derived NVF.,13261
is mechanism of intentionality The shared evolution the unknown. behind,The mechanism behind the evolution of shared intentionality is unknown.,13262
results the new evolution shed Our collaborative light on of behaviours.,Our results shed new light on the evolution of collaborative behaviours.,13263
increases An. senescence with magnitude The of,The magnitude of senescence increases with An.,13264
"implying mosquitoes can cohorts gambiae exhibit senescence. before that wild long strong lifespan, die","gambiae lifespan, implying that wild mosquitoes die long before cohorts can exhibit strong senescence.",13265
reviewed studies published spp. available Anopheles We of mortality,We reviewed available published mortality studies of Anopheles spp.,13266
prediction aging fundamental populations. confirm this in wild to of,to confirm this fundamental prediction of aging in wild populations.,13267
or functions. statistics experimentally identified states power-law through are sometimes universal scaling Critical,Critical states are sometimes identified experimentally through power-law statistics or universal scaling functions.,13268
these macroscopically. large-scale that universal features when rather considered are We networks suggest of,We rather suggest that these are universal features of large-scale networks when considered macroscopically.,13269
put the laws found of These in results caution scaling interpretation on nature.,These results put caution on the interpretation of scaling laws found in nature.,13270
concepts of and This note revisits difficulty. the task,This note revisits the concepts of task and difficulty.,13271
notion introduce asynchronous-time of stochastic tasks. We a,We introduce a notion of asynchronous-time stochastic tasks.,13272
modification of images reconstruction. TwIST Compressive Sensing paper MRI the algorithm for This presents,This paper presents modification of the TwIST algorithm for Compressive Sensing MRI images reconstruction.,13273
in of Sensing the importance. found application Compressive imaging The has great biomedical,The application of the Compressive Sensing in biomedical imaging has found great importance.,13274
It the Total based is on Variation minimization.,It is based on the Total Variation minimization.,13275
"different different tested. coefficients selection domains In paper, are masks for this and transformation","In this paper, different masks and different transformation domains for coefficients selection are tested.",13276
The theory experimental with results. is proved,The theory is proved with experimental results.,13277
the preserving its while filters preserve image. Edge blurring an and edges information,Edge preserving filters preserve the edges and its information while blurring an image.,13278
"etc. diffusion filter, Examples anisotropic guided filter bilateral filter, are trilateral filter,","Examples are bilateral filter, anisotropic diffusion filter, guided filter, trilateral filter etc.",13279
during clear these not are It the is also network how development. formed patterns,It is also not clear how these patterns are formed during the network development.,13280
"a function which score introduce curve. surrogate We parametric an using the integral model, approximates","We introduce a parametric surrogate model, which approximates the score function using an integral curve.",13281
"as known likelihood, of model heteroskedasticity properties and incorporates Our such the surrogate critical temperature.","Our surrogate model incorporates known properties of the likelihood, such as heteroskedasticity and critical temperature.",13282
"available algorithm open source ""bayesImageS."" of package is our the R implementation in An","An open source implementation of our algorithm is available in the R package ""bayesImageS.""",13283
of life huge Speciation Earth. to is the on diversity fundamental,Speciation is fundamental to the huge diversity of life on Earth.,13284
address not in this does regime. weak important theory mutation Current dependence the,Current theory does not address this dependence in the important weak mutation regime.,13285
"the Hausdorff almost In an distribution. uniform averaged particular, prefers indicator","In particular, the averaged Hausdorff indicator prefers an almost uniform distribution.",13286
"above paper, for described forward the extracting we strategy put In subset. a this","In this paper, we put forward a strategy for extracting the above described subset.",13287
(CRF) applied in have widely segmentations. Rields Conditional been Random image,Conditional Random Rields (CRF) have been widely applied in image segmentations.,13288
Then structured machine vector using a the (SSVM). parameters learnt CRF are support,Then the CRF parameters are learnt using a structured support vector machine (SSVM).,13289
of segmentation binary on demonstrate experiments multi-class the promise the proposed method. and benchmarks Extensive,Extensive experiments on binary and multi-class segmentation benchmarks demonstrate the promise of the proposed method.,13290
Theoretical empirical confirmed of result. plausibility modeling this the,Theoretical modeling confirmed the plausibility of this empirical result.,13291
"over to Pose that Boosting. a Trained ensemble Index Feature, regressor is is CPR similar","Trained over Pose Index Feature, CPR is a regressor ensemble that is similar to Boosting.",13292
paper Network. can how show CPR be Neural represented a In we this as,In this paper we show how CPR can be represented as a Neural Network.,13293
BP global outperforms with (pre-)training. that training We show layer-wise empirically,We empirically show that global training with BP outperforms layer-wise (pre-)training.,13294
full We length of methods capable videos. propose handling two,We propose two methods capable of handling full length videos.,13295
models an frames. second ordered proposed of explicitly The video sequence as method the,The second proposed method explicitly models the video as an ordered sequence of frames.,13296
these All interactive. but were factors highly significant,All these factors were significant but highly interactive.,13297
likelihood via The is estimation maximum EM-algorithm a procedure presented.,The maximum likelihood estimation procedure via a EM-algorithm is presented.,13298
Computation Monte likelihood-free are methods. Carlo Approximate Bayesian (ABC),Approximate Bayesian Computation (ABC) are likelihood-free Monte Carlo methods.,13299
"methods other The kernel ABC nonparametric likelihood or ABC including ABC, outperforms synthetic ABC.","The nonparametric ABC outperforms other methods including ABC, kernel ABC or synthetic likelihood ABC.",13300
of data Synthetic our data test method. were inference and the used real to Bayesian,Synthetic data and real data were used to test the Bayesian inference of our method.,13301
"computationally intractable is large data sets. for However, it","However, it is computationally intractable for large data sets.",13302
"tackling work, introduce data multiresolution we sets. for persistent large this In homology","In this work, we introduce multiresolution persistent homology for tackling large data sets.",13303
method RNA fingerprints DNA further and molecules. We for topological from proposed demonstrate extracting the,We further demonstrate the proposed method for extracting topological fingerprints from DNA and RNA molecules.,13304
"of linear commonly To problem variability, the averaging. smoothed address are before data group","To address the problem of variability, data are commonly smoothed before group linear averaging.",13305
smooth convex guarantees. problem an algorithm convergence a leads optimization with strong It to and,It leads to a smooth convex optimization problem and an algorithm with strong convergence guarantees.,13306
"their needs attributes one associations. objects, such to a description and generate automatically, model To","To generate such a description automatically, one needs to model objects, attributes and their associations.",13307
"object Conventional attribute methods require less and them strong annotation scalable. locations, making of","Conventional methods require strong annotation of object and attribute locations, making them less scalable.",13308
is This a Bayesian model. by achieved weakly introducing supervised novel non-parametric,This is achieved by introducing a novel weakly supervised non-parametric Bayesian model.,13309
knowledge framework hypothesis called virtues so present for formalising a graphs. We within,We present a framework for formalising so called hypothesis virtues within knowledge graphs.,13310
proposed by experiments the validate literature-based framework discovery. in We,We validate the proposed framework by experiments in literature-based discovery.,13311
and its experiments work demonstrated w.r.t. our The utility superiority have the of,The experiments have demonstrated the utility of our work and its superiority w.r.t.,13312
were Mass tested of screen-and-treat for achieving drug likelihood campaigns elimination.,Mass screen-and-treat drug campaigns were tested for likelihood of achieving elimination.,13313
of Conclusions reservoir. the density Low infectious portion a infections substantial comprise,Conclusions Low density infections comprise a substantial portion of the infectious reservoir.,13314
drugs control for Antimalarial a powerful are malaria and tool elimination.,Antimalarial drugs are a powerful tool for malaria control and elimination.,13315
setting. therapies can reduce combination widely campaign (ACTs) when transmission a in Artemisinin-based distributed,Artemisinin-based combination therapies (ACTs) can reduce transmission when widely distributed in a campaign setting.,13316
of killing to clinical gametocytes parasites data. asexual and calibrated Drug was,Drug killing of asexual parasites and gametocytes was calibrated to clinical data.,13317
We introduce an the between compatibility that a embedding. and function label measures a image,We introduce a function that measures the compatibility between an image and a label embedding.,13318
game computer terminals. system's in The runs,The game runs in computer system's terminals.,13319
"sciences, is computational fundamental computer saliency Visual vision. a problem in including both and cognitive","Visual saliency is a fundamental problem in both cognitive and computational sciences, including computer vision.",13320
saliency of to coherence then refinement propose enhance the We a our method results. spatial,We then propose a refinement method to enhance the spatial coherence of our saliency results.,13321
large-scale more common (LGDM) decision making group  are  and more nowadays. Linguistic problems,  Linguistic large-scale group decision making (LGDM) problems are more and more common nowadays.,13322
Distributed topic have signal during years. processing a become algorithms past the hot,Distributed signal processing algorithms have become a hot topic during the past years.,13323
attention One class special received that (PFs). of filters algorithms have particles are,One class of algorithms that have received special attention are particles filters (PFs).,13324
"To the computer illustrate theoretical findings, tracking out target simulations carry a we problem. for","To illustrate the theoretical findings, we carry out computer simulations for a target tracking problem.",13325
attack a models two species. on between account predator's prey the These rates trade-off for,These models account for a trade-off between the predator's attack rates on two prey species.,13326
"permanence strong lost. is at trade-offs, However,","However, at strong trade-offs, permanence is lost.",13327
"coexistence. attractors loss this supporting there Despite of permanence, be can","Despite this loss of permanence, there can be attractors supporting coexistence.",13328
"may however, coincide the These is at attractors, excluded. which predator with attractors","These attractors, however, may coincide with attractors at which the predator is excluded.",13329
Real bipartisanship on pairwise also congress US analyzing are in and DNA/time-series reported. alignments applications,Real applications on analyzing bipartisanship in US congress and pairwise DNA/time-series alignments are also reported.,13330
pipelines. Aligning is a to in a bioinformatics reference reads fundamental numerous sequence step,Aligning reads to a reference sequence is a fundamental step in numerous bioinformatics pipelines.,13331
positions the genome. reads about Simulation their source is accompanied information by in of,Simulation of reads is accompanied by information about their positions in the source genome.,13332
the evaluate by information mapper. then is used alignments produced to This,This information is then used to evaluate alignments produced by the mapper.,13333
"successful alignments read are of reports containing created. statistics Finally,","Finally, reports containing statistics of successful read alignments are created.",13334
"components. Futhermore, have software we an associated developed RNF package principal two containing","Futhermore, we have developed an associated software package RNF containing two principal components.",13335
"tools CuReSim simulating of MIShmash Mason, etc.) one popular applies DwgSim, Art, (among read","MIShmash applies one of popular read simulating tools (among DwgSim, Art, Mason, CuReSim etc.)",13336
the RNF and generated reads into transforms format.,and transforms the generated reads into RNF format.,13337
using a reads evaluates simulated given RNF format. LAVEnder read in mapper then,LAVEnder evaluates then a given read mapper using simulated reads in RNF format.,13338
"traditional to these solve with addition, methods. and unstable are In problems hard numerically","In addition, these problems are numerically unstable and hard to solve with traditional methods.",13339
"the in embedding We low-dimensional consider k-nearest of problem Euclidean space. directed graphs neighbor unweighted,","We consider the problem of embedding unweighted, directed k-nearest neighbor graphs in low-dimensional Euclidean space.",13340
develop target and cells. dynamics model biophysical particles We a contact-mediated virus of involving,We develop a biophysical model of contact-mediated dynamics involving virus particles and target cells.,13341
approximation word derive we a for counts. normal Here such,Here we derive a normal approximation for such word counts.,13342
manifold. as is naturally on a clustering Grassmann problem This described,This is naturally described as a clustering problem on Grassmann manifold.,13343
tasks. computer in applications many have new The methods vision,The new methods have many applications in computer vision tasks.,13344
two penta-valued variants paper of neutrosophic for representation This presents entropy.,This paper presents two variants of penta-valued representation for neutrosophic entropy.,13345
"at the targeted applications. primarily One PETS dataset, well-known surveillance of the few exceptions is","One of the few exceptions is the well-known PETS dataset, targeted primarily at surveillance applications.",13346
"we are and here all of The provide passages. translations French, letters in key","The letters are all in French, and here we provide translations of key passages.",13347
"as stored, at synapses. least Memories strong of patterns partly, are","Memories are stored, at least partly, as patterns of strong synapses.",13348
"molecular for years how synapses maintain the strong can Given that turnover, persist? can memories","Given molecular turnover, how can synapses maintain strong for the years that memories can persist?",13349
postulate bistability maintains strong models synapses. Some that biochemical,Some models postulate that biochemical bistability maintains strong synapses.,13350
Bistability single has never synapses also empirically of been demonstrated.,Bistability of single synapses has also never been empirically demonstrated.,13351
it persistence. models to distributions is Thus important for memory simulate both and unimodal long-term,Thus it is important for models to simulate both unimodal distributions and long-term memory persistence.,13352
"are clusters for years. competition, these stable With","With competition, these clusters are stable for years.",13353
applications tomatic intelligent recognition vehicle plate in systems. management license several Au- traffic has,Au- tomatic vehicle license plate recognition has several applications in intelligent traffic management systems.,13354
"this system. phone mobile license describe we In paper, based, client-server architected, a plate recognition","In this paper, we describe a mobile phone based, client-server architected, license plate recognition system.",13355
in end end system architecture describe We to detail. the,We describe the end to end system architecture in detail.,13356
proposed lab environment. in been the system of the has working A prototype implemented,A working prototype of the proposed system has been implemented in the lab environment.,13357
a problem is of ill-posed online and The conversion problem. an to challenging offline script,The problem of offline to online script conversion is a challenging and an ill-posed problem.,13358
kinect system This using sensor. face a real-time presents recognition paper,This paper presents a real-time face recognition system using kinect sensor.,13359
speed GPU implemented observed. is using on are and significant opencl improvements The algorithm,The algorithm is implemented on GPU using opencl and significant speed improvements are observed.,13360
The mainly steps. composed is of algorithm three,The algorithm is mainly composed of three steps.,13361
the step First in is video to faces detect using jones all viola algorithm.,First step is to detect all faces in the video using viola jones algorithm.,13362
database on tracking The window generation face. using step is the online second a,The second step is online database generation using a tracking window on the face.,13363
classifier. train svm This feature used a is vector to,This feature vector is used to train a svm classifier.,13364
multiple recognition vector. faces modified Third involves of our step based feature on,Third step involves recognition of multiple faces based on our modified feature vector.,13365
"development present of design camera and paper, this In an the we system. undersea","In this paper, we present the design and development of an undersea camera system.",13366
ROVs There our system are AUVs. comparing with two or using advantages main,There are two main advantages comparing our system with using ROVs or AUVs.,13367
"can continuous undersea our the of monitoring First, habitat. system provide","First, our system can provide continuous monitoring of the undersea habitat.",13368
"a system has low our cost. Second, hardware","Second, our system has a low hardware cost.",13369
Gabor depends This filtering. the dominant extraction using on of selection features,This selection depends on the extraction of dominant features using Gabor filtering.,13370
images images. and to as processes non-landmark classify landmark these method them then The,The method then processes these images to classify them as landmark and non-landmark images.,13371
highly the landmark. features The candidate number on classification performance depends of of the,The classification performance highly depends on the number of candidate features of the landmark.,13372
generalizations are Tensors or of multiarray matrices. data,Tensors or multiarray data are generalizations of matrices.,13373
researched. vectorizing multiarray clustering Subspace has been on extensively data based,Subspace clustering based on vectorizing multiarray data has been extensively researched.,13374
"exploit complete However, structure not information. of tensorial vectorization data does","However, vectorization of tensorial data does not exploit complete structure information.",13375
"clustering vectorization paper, we adopting this process. subspace propose In without a algorithm any","In this paper, we propose a subspace clustering algorithm without adopting any vectorization process.",13376
a on novel heterogeneous Our based approach Tucker decomposition model. is,Our approach is based on a novel heterogeneous Tucker decomposition model.,13377
but closed-form updates. last mode the have All,All but the last mode have closed-form updates.,13378
latter to by two restaurant extending task The is the process accomplished Chinese dimensions. well-known,The latter task is accomplished by extending the well-known Chinese restaurant process to two dimensions.,13379
the approach. that issues performance affect We also future discuss this of,We also discuss issues that affect the future performance of this approach.,13380
vision represent models subspace Many data. algorithms employ computer to,Many computer vision algorithms employ subspace models to represent data.,13381
paper. manifold extending in explored on LRR this is The of possibility Grassmann,The possibility of extending LRR on Grassmann manifold is explored in this paper.,13382
proposed designed the is implemented. LRR solving model for algorithm Grassmannian new and A,A new algorithm for solving the proposed Grassmannian LRR model is designed and implemented.,13383
method outperforms The existing show a of experimental methods. number our results,The experimental results show our method outperforms a number of existing methods.,13384
"novo assembly also challenging, De high-quality only data. computationally is not but requires","De novo assembly is not only computationally challenging, but also requires high-quality data.",13385
"we implemented In we describe in approach this PopIns. paper, the","In this paper, we describe the approach we implemented in PopIns.",13386
for Image stack-based is HDR challenging. registration photography,Image registration for stack-based HDR photography is challenging.,13387
analytical model an function. Our is continues,Our model is an analytical continues function.,13388
features to semantics. in known inadequate be hand-crafted complex are analyzing video Traditional,Traditional hand-crafted features are known to be inadequate in analyzing complex video semantics.,13389
"very video Based attained on on evaluations, are benchmarks. two the classification popular competitive results","Based on the evaluations, very competitive results are attained on two popular video classification benchmarks.",13390
by embedding features. a of the The graph linear is locally construction detection motivated,The graph construction is motivated by a locally linear embedding of the detection features.,13391
to decompose into global sub-problems. function propose We the node-wise objective,We propose to decompose the global objective function into node-wise sub-problems.,13392
"opens the possibility it parallel of Moreover, implementation.","Moreover, it opens the possibility of parallel implementation.",13393
the units). recognition automatic for (sub-character primitives handwritten of are Experiments isolated conducted character,Experiments are conducted for the automatic recognition of isolated handwritten character primitives (sub-character units).,13394
using sequence protein k-means is information extracted optimised algorithm. this work In,In this work protein sequence information is extracted using optimised k-means algorithm.,13395
used of The swarm technique is optimisation particle optimisation one frequently method. the,The particle swarm optimisation technique is one of the frequently used optimisation method.,13396
current motif k-means extraction. In information the the PSO work is for used,In the current work the PSO k-means is used for motif information extraction.,13397
the structure protein acquired homogeneity the The motif of sequence. information is on based,The motif information acquired is based on the structure homogeneity of the protein sequence.,13398
can spike make A approximately simple trains. optimal network decisions given Bayes stochastic,A simple network can make approximately Bayes optimal decisions given stochastic spike trains.,13399
be this arbitrarily to by that Bayes enlarging optimal can obtained Decisions are network. close,Decisions that are arbitrarily close to Bayes optimal can be obtained by enlarging this network.,13400
with a is cerebellar-like of a particle analog architecture. neural filter Bayesian This,This is a neural analog of a Bayesian particle filter with cerebellar-like architecture.,13401
is of problem associated from data. high-dimensional The classification one statistical,The associated statistical problem is one of classification from high-dimensional data.,13402
representations the various and cues. concepts attempt visual capture learned combination to The of,The learned representations attempt to capture the combination of various visual concepts and cues.,13403
"for fractals. dimensional work novel real-time introduces This a two of transformations tool interactive, IFS","This work introduces a novel tool for interactive, real-time transformations of two dimensional IFS fractals.",13404
and grounds of the tool software give theoretical application. then We the,We give theoretical grounds of the tool and then the software application.,13405
the in In study we appearance-based this wild. estimation work gaze,In this work we study appearance-based gaze estimation in the wild.,13406
more significantly is respect to illumination. Our existing dataset ones variable with appearance and than,Our dataset is significantly more variable than existing ones with respect to appearance and illumination.,13407
from obstacles. of shortest algorithm addresses finding avoiding source problem path Path to destination finding,Path finding algorithm addresses problem of finding shortest path from source to destination avoiding obstacles.,13408
"optimization. algorithms Dijkstra's ant colony A*, namely search various There exist and","There exist various search algorithms namely A*, Dijkstra's and ant colony optimization.",13409
"essential in location part applications. gaming become The can network, based, algorithm proposed","The proposed algorithm can become essential part in location based, network, gaming applications.",13410
"Intelligence. mobile traversal, Artificial robot gaming navigation, applications, and grid","grid traversal, navigation, gaming applications, mobile robot and Artificial Intelligence.",13411
"the any aim we sampling at step. More specifically posteriors, avoiding with close-form achieving inference","More specifically we aim at achieving the inference with close-form posteriors, avoiding any sampling step.",13412
"sampling, focus highly avoid exact Carlo convergence To demanding and on we inference. issues Monte","To avoid convergence issues and highly demanding Monte Carlo sampling, we focus on exact inference.",13413
"sampling any aim with step. we posteriors, achieving at the close-form More specifically avoiding inference","More specifically we aim at achieving the inference with close-form posteriors, avoiding any sampling step.",13414
"graphs we To mixtures of set aim, restrict this of trees. to the considered spanning","To this aim, we restrict the set of considered graphs to mixtures of spanning trees.",13415
two consists proposed The method stages. of,The proposed method consists of two stages.,13416
noise (DMF) Filter stage base The Decision removal preliminary as acts algorithm. Median first the,The first stage Decision base Median Filter (DMF) acts as the preliminary noise removal algorithm.,13417
as Ecosystems interacting commonly conceptualized are networks of species.,Ecosystems are commonly conceptualized as networks of interacting species.,13418
"allows direction. from the This problem"" ""species formalism approaching opposite the","This formalism allows approaching the ""species problem"" from the opposite direction.",13419
diverse discussed. theoretical is The describing ecosystems for highly construct of natural relevance this,The relevance of this theoretical construct for describing highly diverse natural ecosystems is discussed.,13420
have over last driving Systems the safer made (ADAS) Advanced Assistance decade. Driver,Advanced Driver Assistance Systems (ADAS) have made driving safer over the last decade.,13421
a this anticipate we occur. few they before seconds driving work In maneuvers,In this work we anticipate driving maneuvers a few seconds before they occur.,13422
to Input-Output alongwith information We model an propose contextual the Autoregressive the HMM maneuvers.,We propose an Autoregressive Input-Output HMM to model the contextual information alongwith the maneuvers.,13423
human machine intelligence can Whether the issue. the is intelligence a surpass controversial,Whether the machine intelligence can surpass the human intelligence is a controversial issue.,13424
"whole-body navigation. unclear BMI whether ensembles is However, a operate cortical could it for","However, it is unclear whether cortical ensembles could operate a BMI for whole-body navigation.",13425
"paper, one from interpret to aim scenes indoor RGBD image. we this In","In this paper, we aim to interpret indoor scenes from one RGBD image.",13426
connection need! all An they is Internet,An Internet connection is all they need!,13427
"lactate is acids. and and Low abundance high characterized diversity, of by lactobacilli, amino relative","Low diversity, and high relative abundance of lactobacilli, is characterized by lactate and amino acids.",13428
"Near-Online present another we Tracking As Multi-target (NOMT) algorithm. contribution, a","As another contribution, we present a Near-Online Multi-target Tracking (NOMT) algorithm.",13429
"on datasets. challenging run and comprehensive tracking two evaluation We datasets, experimental KITTI a MOT","We run a comprehensive experimental evaluation on two challenging tracking datasets, KITTI and MOT datasets.",13430
"To efficiency, is steps. improve Newton-Raphson employed also algorithm the search in local","To improve the efficiency, Newton-Raphson algorithm is also employed in local search steps.",13431
to resolve is Conformal the solving instability numerical used classical equations. mapping geodesic associated with,Conformal mapping is used to resolve numerical instability associated with solving the classical geodesic equations.,13432
local by are numerically constructed under Geodesic using approximation. conformal mapping quadratic flows,Geodesic flows under conformal mapping are constructed numerically by using local quadratic approximation.,13433
a is approximation. Gaussian The copula using marginal adjustment approach extended,The marginal adjustment approach is extended using a Gaussian copula approximation.,13434
method examples. is This illustrated with several,This method is illustrated with several examples.,13435
organ important diagnosis computer-aided systems. for an segmentation is many Automatic prerequisite,Automatic organ segmentation is an important prerequisite for many computer-aided diagnosis systems.,13436
based method image of hierarchical on coarse-to-fine local regions The (superpixels). classification is,The method is based on hierarchical coarse-to-fine classification of local image regions (superpixels).,13437
abdominal Clustering Superpixels Simple Iterative extracted region using (SLIC). are from the Linear,Superpixels are extracted from the abdominal region using Simple Linear Iterative Clustering (SLIC).,13438
"can't rendering. the real-time it complexity, requirement of reach of the Because high computation","Because of the high computation complexity, it can't reach the requirement of real-time rendering.",13439
discuss the technique. further of We applications,We discuss further applications of the technique.,13440
"of a high-resolution reach most those in As imaging remains the result, consumers. beyond spectrum","As a result, high-resolution imaging in those spectrum remains beyond the reach of most consumers.",13441
image is reconstruction from these measurements. the computational A algorithm to coded then used recover,A computational reconstruction algorithm is then used to recover the image from these coded measurements.,13442
"Furthermore, methods and two state-of-the-art outperform complementary combined. the are previous significantly when","Furthermore, two methods are complementary and significantly outperform the previous state-of-the-art when combined.",13443
levels This the affected may of moderate and be in of noise. estimate presence severe,This estimate may be affected in the presence of moderate and severe levels of noise.,13444
"values is variable the of Nonetheless, the task. of crisp a exact, determination context hard","Nonetheless, the determination of exact, crisp values of the context variable is a hard task.",13445
"this we novel for methods paper, determination. that propose two fuzzy using approaches In","In this paper, we propose two novel methods using fuzzy approaches for that determination.",13446
the given to example methods. the is of illustrate uses proposed A numerical,A numerical example is given to illustrate the uses of the proposed methods.,13447
analysis fish visual abundance increasing with attention. estimation aid Non-extractive drawn the has of,Non-extractive fish abundance estimation with the aid of visual analysis has drawn increasing attention.,13448
autonomous Ensuring difficult. is that and both complex work ethically systems,Ensuring that autonomous systems work ethically is both complex and difficult.,13449
a vital Retrieval System. a playing is role as part CBIR of Image Trademark,Trademark Image Retrieval is playing a vital role as a part of CBIR System.,13450
carries status because is Trademark company. it of great any value the of significance,Trademark is of great significance because it carries the status value of any company.,13451
it nature further matching manually. challenging features imprecise to of This makes choose,This imprecise nature of matching further makes it challenging to choose features manually.,13452
on The well function is cross-domain as loss within-domain the defined similarities. as the,The loss function is defined on the within-domain as well as the cross-domain similarities.,13453
"is performance a gain computational in in Situations discussed. which are expected,","Situations in which a gain in computational performance is expected, are discussed.",13454
Richard's mathematics are and medicine statistics their and interests research in and to applications biology.,Richard's research interests are in statistics and mathematics and their applications to medicine and biology.,13455
"individuals of on analyze where nodes a graphs, population. represent We the dynamics evolutionary","We analyze evolutionary dynamics on graphs, where the nodes represent individuals of a population.",13456
most among set selected sequences patterns frequent protein The and of is concatenated.,The set of most frequent patterns among protein sequences is selected and concatenated.,13457
family sequences alignment necessary. of species phylogenetic the tree are and No,No sequences alignment and phylogenetic tree of the species family are necessary.,13458
universe. to ideas apply briefly limits computation of We our our in physical intelligent the,We briefly apply our ideas to the physical limits of intelligent computation in our universe.,13459
localize Beyond GWAS: alternatives QTLs animals. to in farm,Beyond GWAS: alternatives to localize QTLs in farm animals.,13460
a Similarity-preserving nearest hashing method for neighbour widely-used image search large-scale is tasks. retrieval in,Similarity-preserving hashing is a widely-used method for nearest neighbour search in large-scale image retrieval tasks.,13461
"of is the vocabularies of dimensionality method particular, joint reduction considered. In multiple","In particular, the method of joint dimensionality reduction of multiple vocabularies is considered.",13462
one is in analysis. steps (BS) the of key Background video Subtraction,Background Subtraction (BS) is one of the key steps in video analysis.,13463
performance on sets. data public promising and have models proposed background achieved Many been,Many background models have been proposed and achieved promising performance on public data sets.,13464
"background illumination to etc. However, change, due challenges dynamic such as","However, due to challenges such as illumination change, dynamic background etc.",13465
as noise. foreground background the as segmentation well often consists holes resulted of,the resulted foreground segmentation often consists of holes as well as background noise.,13466
"generalized structured this regard, quest fused we foregrounds. regularization lasso to intact consider for In","In this regard, we consider generalized fused lasso regularization to quest for intact structured foregrounds.",13467
is text localization method and end-to-end recognition unconstrained presented. An,An unconstrained end-to-end text localization and recognition method is presented.,13468
"estimation on based introduced. Additionally, a area stroke feature novel character is","Additionally, a novel feature based on character stroke area estimation is introduced.",13469
hunter-prey a extinction in mathematical of mechanistic coexistence ecosystem. study We and generic a model,We study a mechanistic mathematical model of extinction and coexistence in a generic hunter-prey ecosystem.,13470
"to concomitant transmit systems biochemical dissipate signals, energy In with regulatory entropy biological order production.","In order to transmit biochemical signals, biological regulatory systems dissipate energy with concomitant entropy production.",13471
"signaling place often Additionally, environmental in conditions. takes challenging","Additionally, signaling often takes place in challenging environmental conditions.",13472
"Negative are characteristic dissipation. of at responses, feedback optimal high loops, shock","Negative feedback loops, characteristic of shock responses, are optimal at high dissipation.",13473
"informative. positive Close loops, stability, for to known more their become equilibrium feedback","Close to equilibrium positive feedback loops, known for their stability, become more informative.",13474
human an is many classification diagnosis Automated important of computer-aided for prerequisite systems. anatomy,Automated classification of human anatomy is an important prerequisite for many computer-aided diagnosis systems.,13475
makes of anatomy human complexity the classification throughout and difficult. variability spatial The body,The spatial complexity and variability of anatomy throughout the human body makes classification difficult.,13476
"methods have various analytical characterize regions selection, genomic To natural detecting candidate developed. for been","To characterize natural selection, various analytical methods for detecting candidate genomic regions have been developed.",13477
genome-wide perform to We component analysis. scans natural propose selection of using principal,We propose to perform genome-wide scans of natural selection using principal component analysis.,13478
distinctions. subtle to with objects aims categorization image Fine-grained classify,Fine-grained image categorization aims to classify objects with subtle distinctions.,13479
of the The expensive annotations methods. usage for wide these object need prevents,The need for expensive object annotations prevents the wide usage of these methods.,13480
"In detectors work, models. part-based as object this to we propose Fisher-Vector-based interpret","In this work, we propose to interpret Fisher-Vector-based object detectors as part-based models.",13481
forms models. of graphical challenges of Structure inference learning the one statistical,Structure learning forms one of the inference challenges of statistical graphical models.,13482
with be combined the to reversal can further. edge moves sampler Finally the method improve,Finally the method can be combined with edge reversal moves to improve the sampler further.,13483
more artifacts increases. These view unnatural of the become more field as and,These artifacts become more and more unnatural as the field of view increases.,13484
"consider We a more setup, viz. realistic","We consider a more realistic setup, viz.",13485
Triander visualization. is an software package for nucleotide The sequence interactive program,The Triander program is an interactive software package for nucleotide sequence visualization.,13486
can Triander produce types four of plots.,Triander can produce four types of plots.,13487
position to DNA independently is nucleotide build in triplets. three each for walks It possible,It is possible to build three DNA walks independently for each nucleotide position in triplets.,13488
recently. Neural Convolutional Networks (CNNs) success have in Deep remarkable applications enjoyed vision,Deep Convolutional Neural Networks (CNNs) have enjoyed remarkable success in vision applications recently.,13489
the that data. the within primary training highly reason co-occur The attributes is,The primary reason is that the attributes highly co-occur within the training data.,13490
"salient describe, detail, this aspects in We of dataset.","We describe, in detail, salient aspects of this dataset.",13491
standard This approach state-of-the-art achieves dataset a performance of RAW on images.,This approach achieves state-of-the-art performance on a standard dataset of RAW images.,13492
human introduction recognition. in of has promoted The research the skeleton-based action low-cost sensors RGB-D,The introduction of low-cost RGB-D sensors has promoted the research in skeleton-based human action recognition.,13493
provide challenge. two this insights here into We,We here provide two insights into this challenge.,13494
action-level trajectorylet is by pooling The encodings. representation obtained,The action-level representation is obtained by pooling trajectorylet encodings.,13495
"results. VLAD Vectors or have FV, Existing achieved excellent Super methods such as","Existing methods such as FV, VLAD or Super Vectors have achieved excellent results.",13496
"DTVD robustly. estimate to robust classifier-based method is Furthermore, proposed a","Furthermore, a robust classifier-based method is proposed to estimate DTVD robustly.",13497
"applications. task, Texture  many to crucial image constitutes processing a standard segmentation ","  Texture segmentation constitutes a standard image processing task, crucial to many applications.",13498
been has packing studied. protein thoroughly Single-domain side-chain,Single-domain protein side-chain packing has been thoroughly studied.,13499
"protein designed methods exist specifically few side-chain that packing are complexes. To date, for","To date, few side-chain packing methods exist that are specifically designed for protein complexes.",13500
"more iTreePack SCWRL Even monomeric is and much for more proteins, slightly accurate. efficient than","Even for monomeric proteins, iTreePack is much more efficient than SCWRL and slightly more accurate.",13501
the require frequently arise contours of parameterization computer Problems vision closed applications. that in,Problems that require the parameterization of closed contours arise frequently in computer vision applications.,13502
point from five numerous to source by subsets ellipses The fitting algorithm the starts data.,The algorithm starts by fitting ellipses to numerous five point subsets from the source data.,13503
median set The by perimeter of the ellipses. is the determining curve closed parameterized of,The closed curve is parameterized by determining the median perimeter of the set of ellipses.,13504
"parameterized. closed not curve The contours is an arbitrary ellipse, be resulting allowing to","The resulting curve is not an ellipse, allowing arbitrary closed contours to be parameterized.",13505
explored. of the is perimeter use rather than median also perimeter modal the The,The use of the modal perimeter rather than the median perimeter is also explored.,13506
datasets research years. in visual been benchmark Several have tracking for proposed recent,Several benchmark datasets for visual tracking research have been proposed in recent years.,13507
"discrepant tracking some Surprisingly, findings beliefs common research the community. our visual in with are","Surprisingly, our findings are discrepant with some common beliefs in the visual tracking research community.",13508
most find feature tracker. plays role that the in extractor We the important a,We find that the feature extractor plays the most important role in a tracker.,13509
"that could affect the and model many Moreover, details updater the model result. motion contain","Moreover, the motion model and model updater contain many details that could affect the result.",13510
introgressive broaden strategy is a pivotal crops. the Breeding to of genetic hybridization by basis,Breeding by introgressive hybridization is a pivotal strategy to broaden the genetic basis of crops.,13511
"IIT is outright aim, functionalism champion particular. reject The to nor in to though, neither","The aim, though, is neither to reject IIT outright nor to champion functionalism in particular.",13512
"attention? p our happen escaped may, of rest the What that to however, initially","What may, however, happen to the rest of p that initially escaped our attention?",13513
"of the were beliefs set hidden what As may visible. become changes,","As the set of beliefs changes, what were hidden may become visible.",13514
a We beliefs. that present incorporates perception-based theory latent belief,We present a perception-based belief theory that incorporates latent beliefs.,13515
distributions. undirected graphs problem estimating triangle-free dimensional of consider the of high We,We consider the problem of estimating undirected triangle-free graphs of high dimensional distributions.,13516
"Fermat's to tractability, we For family. a distribution the propose principle graphical regularize inferential","For inferential tractability, we propose a graphical Fermat's principle to regularize the distribution family.",13517
"and graph nonparametric consider different discrete including under estimation distribution settings, We families. problems","We consider graph estimation problems under different settings, including discrete and nonparametric distribution families.",13518
illustrate provided the to the method. numerical Thorough results of proposed usefulness are,Thorough numerical results are provided to illustrate the usefulness of the proposed method.,13519
first the by trajectory to motion camera popular improve features estimation. dense We explicit propose,We first propose to improve the popular dense trajectory features by explicit camera motion estimation.,13520
to RANSAC. matches a are homography used estimate The with,The matches are used to estimate a homography with RANSAC.,13521
"removed. thus motion, Trajectories consistent and are with due camera the considered as to homography","Trajectories consistent with the homography are considered as due to camera motion, and thus removed.",13522
the camera the from use to homography out motion cancel We also flow. optical,We also use the homography to cancel out camera motion from the optical flow.,13523
on in motion-based results improvement MBH This significant descriptors. and HOF,This results in significant improvement on motion-based HOF and MBH descriptors.,13524
"three we over tasks, substantial results. all In improvements the show state-of-the-art","In all three tasks, we show substantial improvements over the state-of-the-art results.",13525
image of time paper separated the This task aerial addresses registration.,This paper addresses the task of time separated aerial image registration.,13526
detectors extractor contain important object components: Most and two object classifier. feature an a,Most object detectors contain two important components: a feature extractor and an object classifier.,13527
classification that for is paper object carefully deep as This designing important. just demonstrates networks,This paper demonstrates that carefully designing deep networks for object classification is just as important.,13528
"classifier shared, region-independent that networks features. region-wise with experiment convolutional We use","We experiment with region-wise classifier networks that use shared, region-independent convolutional features.",13529
"We feature ""Networks Convolutional (NoCs). maps"" call on them","We call them ""Networks on Convolutional feature maps"" (NoCs).",13530
"local and the parameters. obtained global calculate values, we Using these","Using these obtained values, we calculate the local and global parameters.",13531
of (propagators) classes algorithms Specialised for propagation exist many constraints.,Specialised propagation algorithms (propagators) exist for many classes of constraints.,13532
is design support of concept pervasive The the in propagators. of,The concept of support is pervasive in the design of propagators.,13533
a support literal. tuple each GAC-Schema of GAC to values algorithms use as such,GAC algorithms such as GAC-Schema use a tuple of values to support each literal.,13534
in of ways. generalize two notions support We these,We generalize these notions of support in two ways.,13535
"set we act tuples a as to of allow support. First,","First, we allow a set of tuples to act as support.",13536
support. We a using for propagators methodology developing correct design generalized,We design a methodology for developing correct propagators using generalized support.,13537
is produced. carefully be The to to allow designed framework algorithms efficient,The framework is carefully designed to allow efficient algorithms to be produced.,13538
for dynamic literals efficiency. make literal use of or triggers may watched Derived algorithms,Derived algorithms may make use of dynamic literal triggers or watched literals for efficiency.,13539
"deriving case given. algorithms are efficient two studies of Finally,","Finally, two case studies of deriving efficient algorithms are given.",13540
these challenging. Identifying high-dimension faces is in,Identifying these faces in high-dimension is challenging.,13541
composite the concerns second Our maximum contribution likelihood estimate.,Our second contribution concerns the composite maximum likelihood estimate.,13542
an analysis to applied is of the Then comfort. passengers' model,Then the model is applied to an analysis of passengers' comfort.,13543
and of measured driving. movements in drivers The passengers were head slalom,The head movements of drivers and passengers were measured in slalom driving.,13544
the epidemics networks. migration of of patterns investigate in complex We role spread on the,We investigate the role of migration patterns on the spread of epidemics in complex networks.,13545
diffusion. We nonlinear the enhance to a model SIS-diffusion metapopulations on,We enhance the SIS-diffusion model on metapopulations to a nonlinear diffusion.,13546
to the confirms complementary is result the learning-based features. This that LOAD,This result confirms that the LOAD is complementary to the learning-based features.,13547
"of given temporal key-poses, i.e. are Interval occurrences limits by","Interval limits are given by temporal occurrences of key-poses, i.e.",13548
often difficult It subtle directly. such to features is detect,It is often difficult to detect such subtle features directly.,13549
tasks. been Deep image networks classification successfully applied to neural convolutional have,Deep convolutional neural networks have been successfully applied to image classification tasks.,13550
"highly parallelizable. time-efficient is The and resulting algorithm convex,","The resulting algorithm is convex, time-efficient and highly parallelizable.",13551
validated SegSALSA-STR of The real in hyperspectral images. algorithm segmentation the is,The SegSALSA-STR algorithm is validated in the segmentation of real hyperspectral images.,13552
has the been novel A WxBS algorithm problem matching addressing introduced. for,A novel matching algorithm for addressing the WxBS problem has been introduced.,13553
"version In paper, is this of new limitations. to a these overcome EFIS proposed","In this paper, a new version of EFIS is proposed to overcome these limitations.",13554
"For a we use Field approximation Mean based algorithm. inference,","For inference, we use a Mean Field approximation based algorithm.",13555
"cannot the small is invasive It for endoscopies traditional important intestine, especially where reach.","It is especially important for the small intestine, where traditional invasive endoscopies cannot reach.",13556
"appearance to true from of the Intuitively, image object image. boundaries varies","Intuitively, the appearance of true object boundaries varies from image to image.",13557
"our substantially method [Hariharan al. Additionally, outperforms et","Additionally, our method substantially outperforms [Hariharan et al.",13558
paper disorder This presents (OCD). a account coding of predictive obsessive-compulsive,This paper presents a predictive coding account of obsessive-compulsive disorder (OCD).,13559
thus our knowledge literature. has To this connection the far in unremarked gone,To our knowledge this connection has thus far gone unremarked in the literature.,13560
this a of we this In literature topic. substantial the on give generalization paper,In this paper we give a substantial generalization of the literature on this topic.,13561
then where extra an a We characterize guarantee obtainable. such condition is,We then characterize an extra condition where such a guarantee is obtainable.,13562
of method conceptual our properties use proposed to derive of analysis system. the We the,We use the method of conceptual analysis to derive the properties of our proposed system.,13563
literature. with existing compare critically and discuss the related applications We,We discuss applications and compare critically with the existing related literature.,13564
variant is Illumina pipeline Summary: FermiKit data. a calling for,Summary: FermiKit is a variant calling pipeline for Illumina data.,13565
here were et Kemp The al. models introduced in considered,The models considered here were introduced in Kemp et al.,13566
The structures has over presented here explicit prior an formula. graph,The prior over graph structures presented here has an explicit formula.,13567
estimation rapidly Hand in recent matured has pose years.,Hand pose estimation has matured rapidly in recent years.,13568
"with cluttered further scenes. challenging new we diverse, To progress dataset spur introduce a","To spur further progress we introduce a challenging new dataset with diverse, cluttered scenes.",13569
"evaluation experiments. human a define consistent We criteria, rigorously motivated by","We define a consistent evaluation criteria, rigorously motivated by human experiments.",13570
beyond systems that training sets. not do their This generalize implies most,This implies that most systems do not generalize beyond their training sets.,13571
with progress. for future directions conclude We,We conclude with directions for future progress.,13572
are query shape. shapes each The for retrieved top-n,The top-n shapes are retrieved for each query shape.,13573
tested accuracy by Bulls is score means The standard of eye method.,The accuracy is tested by means of standard Bulls eye score method.,13574
"for build tool, natural a Using large this a of saliency we dataset number images.","Using this tool, we build a saliency dataset for a large number of natural images.",13575
process. a global-based matching further introduce We,We further introduce a global-based matching process.,13576
various effectiveness demonstrate of approach. our results datasets the Experimental on,Experimental results on various datasets demonstrate the effectiveness of our approach.,13577
problem when used sub-optimal starting occurs This to values algorithm. are the often initialize,This problem often occurs when sub-optimal starting values are used to initialize the algorithm.,13578
also obtained The implications local different by the are solutions discussed. maxima clustering of,The implications of the different clustering solutions obtained by local maxima are also discussed.,13579
choice This actions. an to LSTM learn of dynamics complex ideal makes various the property,This property makes LSTM an ideal choice to learn the complex dynamics of various actions.,13580
first a penalty solves sparse term. PCA Robust with the issue,Robust PCA solves the first issue with a sparse penalty term.,13581
"with is handled can matrix The however be model, factorization second the non-convex. issue which","The second issue can be handled with the matrix factorization model, which is however non-convex.",13582
"also based using graph clustering enhanced PCA Besides, a be similarity. can by of data","Besides, PCA based clustering can also be enhanced by using a graph of data similarity.",13583
use the This helpful case applying is for in these techniques. choosing specific,This is helpful in choosing the specific use case for applying these techniques.,13584
"In visual a object HMMTxD. novel for paper, tracking this propose called we method","In this paper, we propose a novel method for visual object tracking called HMMTxD.",13585
"the on almost datasets all state-of-the-art, The all outperforms in criteria. often HMMTxD significantly,","The HMMTxD outperforms the state-of-the-art, often significantly, on all datasets in almost all criteria.",13586
"compression effects artifacts, Lossy artifacts, the compression introduces ringing particularly blocking and blurring. complex","Lossy compression introduces complex compression artifacts, particularly the blocking artifacts, ringing effects and blurring.",13587
introduced. novel for extraction object proposals A efficient of is method,A novel efficient method for extraction of object proposals is introduced.,13588
We investigate computational analysis issues. related PMCS inconsistency and complexity in also,We also investigate inconsistency analysis in PMCS and related computational complexity issues.,13589
"operates the are generated using local on which method keypoints, algorithm. interest presented The SURF","The presented method operates on local interest keypoints, which are generated using the SURF algorithm.",13590
from phenotypes egg? cells Furthermore different how single differentiate do of reproducibly a fertilized,Furthermore how do cells of different phenotypes differentiate reproducibly from a single fertilized egg?,13591
problems These fundamental are developmental in biology.,These are fundamental problems in developmental biology.,13592
phenotypes. histone different of play role an cell modifications the in important maintenance Epigenetic,Epigenetic histone modifications play an important role in the maintenance of different cell phenotypes.,13593
cell mechanism for inheritance molecular generations patterns modification of remains elusive. exact the The over,The exact molecular mechanism for inheritance of the modification patterns over cell generations remains elusive.,13594
to work object classify networks. using previous efficiently R-CNN builds on proposals deep convolutional Fast,Fast R-CNN builds on previous work to efficiently classify object proposals using deep convolutional networks.,13595
timescales. with symbolic We separate biochemical networks the dynamics of discuss,We discuss the symbolic dynamics of biochemical networks with separate timescales.,13596
actions. our objects apply anticipate to and algorithms on representation then recognition predicted We,We then apply recognition algorithms on our predicted representation to anticipate objects and actions.,13597
applications useful from concepts organization of many in are Hierarchies objects. of to navigation,Hierarchies of concepts are useful in many applications from navigation to organization of objects.,13598
generally rise by levels. It life considered was is the that oxygen animal of triggered,It is generally considered that animal life was triggered by the rise of oxygen levels.,13599
"the some stationary be distribution cases, effect on substantial. the In can","In some cases, the effect on the stationary distribution can be substantial.",13600
several even scalable This instances. to large with method our million makes training datasets,This makes our method scalable even to large datasets with several million training instances.,13601
Neural image classification (CNN) state-of-the-art Networks models Convolutional many tasks. for are,Convolutional Neural Networks (CNN) are state-of-the-art models for many image classification tasks.,13602
agreement our accuracy similar between the is pathologists. method inter-observer classification to of The,The classification accuracy of our method is similar to the inter-observer agreement between pathologists.,13603
data task. challenging mining is a mining pattern under constraints Sequential,Sequential pattern mining under constraints is a challenging data mining task.,13604
"to what This interpretable our is us and object representation easily visualize allows detector ""sees"".","This representation is easily interpretable and allows us to visualize what our object detector ""sees"".",13605
correspondences estimation between The of resp. two images,The estimation of correspondences between two images resp.,13606
core point problem in vision. sets a computer is,point sets is a core problem in computer vision.,13607
so to second have this been methods Several proposed solve called problem. order,Several so called second order methods have been proposed to solve this problem.,13608
no guarantees order theoretical are known. third algorithms these most of For,For most of these third order algorithms no theoretical guarantees are known.,13609
"cameras. behaviors of using front are work, individuals' In web in this computers studied","In this work, individuals' behaviors in front of computers are studied using web cameras.",13610
for data in non-Gaussian. of the problem analysis that are general We factor discriminative consider,We consider the problem of discriminative factor analysis for data that are in general non-Gaussian.,13611
the proposed. model based data the Bayesian of is on A ranks,A Bayesian model based on the ranks of the data is proposed.,13612
rank-likelihood. We first {\em introduce the of new max-margin} a version,We first introduce a new {\em max-margin} version of the rank-likelihood.,13613
"developed robust technique. we work, In and have this departure a detection lane warning","In this work, we have developed a robust lane detection and departure warning technique.",13614
on single sensor. system camera based Our is,Our system is based on single camera sensor.,13615
using fitted to curves cubic and RANSAC. robust assumed are be Lanes,Lanes are assumed to be cubic curves and fitted using robust RANSAC.,13616
the and method This can detect lanes road boundary. its reliably of,This method can reliably detect lanes of the road and its boundary.,13617
flow lane angle optical an For departure based used. method were,For lane departure angle an optical flow based method were used.,13618
"for probabilistic this we spoken systems. a framework In paper, goal-driven present dialog","In this paper, we present a probabilistic framework for goal-driven spoken dialog systems.",13619
that also We DM efficient are DM. based more demonstrate than entropy-based non-entropy strategies,We also demonstrate that entropy-based DM strategies are more efficient than non-entropy based DM.,13620
powerful limited from Deep learning and flexibility. suffer expensive tools but methods computation often are,Deep learning methods are powerful tools but often suffer from expensive computation and limited flexibility.,13621
deep alternative is An models representations. combine with light-weight to,An alternative is to combine light-weight models with deep representations.,13622
"As absent. in problems, visual several exist unified cases is framework a successful","As successful cases exist in several visual problems, a unified framework is absent.",13623
boosting pre-trained transfers feed CCF CNN features to from forest low-level models model. the,CCF transfers low-level features from pre-trained CNN models to feed the boosting forest model.,13624
"increasingly is popular, very is egocentric browsing video While it becoming difficult.","While egocentric video is becoming increasingly popular, browsing it is very difficult.",13625
segment Recognizing (index) and to egocentric activities us unstructured temporally videos. long-term long enables,Recognizing long-term activities enables us to temporally segment (index) long and unstructured egocentric videos.,13626
"input, flow a sparse camera optical classifies activity. Given as wearer's the volume CNN our","Given a sparse optical flow volume as input, our CNN classifies the camera wearer's activity.",13627
widely cancer in Bladder the world. spread is,Bladder cancer is widely spread in the world.,13628
exploration remains clinical surface. internal for of procedure bladder Video-endoscopy visual the standard the,Video-endoscopy remains the standard clinical procedure for visual exploration of the bladder internal surface.,13629
"And, lesions over several typically, images. spread are,","And, lesions are, typically, spread over several images.",13630
"problems. addresses classes different, optimization two of paper yet This interrelated","This paper addresses two classes of different, yet interrelated optimization problems.",13631
We consider (aka existential for specifying formalism ontologies. Datalog+) as rules a,We consider existential rules (aka Datalog+) as a formalism for specifying ontologies.,13632
"inconsistent imply Taken tree together, under inference mild results these conditions.","Taken together, these results imply inconsistent tree inference under mild conditions.",13633
convergence world significant and Our demonstrates sets. on data algorithm synthetic improvements real,Our algorithm demonstrates significant convergence improvements on synthetic and real world data sets.,13634
demonstrate neural ensembles that of the modeled. can activity be quantitatively we Here,Here we demonstrate that the activity of neural ensembles can be quantitatively modeled.,13635
"this bias in computer been ignored mostly However, has vision. volumetric","However, this volumetric bias has been mostly ignored in computer vision.",13636
significant We segmentation to this methods due artifacts demonstrate reconstruction and in standard bias.,We demonstrate significant artifacts in standard segmentation and reconstruction methods due to this bias.,13637
of centers dynamics Epidemic is stochastic a interacting network considered. epidemic in,Epidemic dynamics in a stochastic network of interacting epidemic centers is considered.,13638
The migration epidemic modelled by processes and are chains. Markov's,The epidemic and migration processes are modelled by Markov's chains.,13639
probability the derived. Explicit formulas migration process for are of distribution,Explicit formulas for probability distribution of the migration process are derived.,13640
"and numerically. Dependence parameters population, of outbreak parameters coupling analytically initial is parameters, on examined","Dependence of outbreak parameters on initial parameters, population, coupling parameters is examined analytically and numerically.",13641
The for process general approximation a derived. is mean migration field,The mean field approximation for a general migration process is derived.,13642
in from are We cues. interested geometric automatic scene understanding,We are interested in automatic scene understanding from geometric cues.,13643
"real-time semantic we the reconstruction. of aim To loop segmentation to this end, bring in","To this end, we aim to bring semantic segmentation in the loop of real-time reconstruction.",13644
"world segment any real without scenes Importantly, to noise our network is modelling. able","Importantly, our network is able to segment real world scenes without any noise modelling.",13645
corner presents a descriptors. paper and novel shape This scale multi gradient based point a,This paper presents a novel multi scale gradient and a corner point based shape descriptors.,13646
large for method runtime The it the images. quasilinear suitable makes of segmenting,The quasilinear runtime of the method makes it suitable for segmenting large images.,13647
are medicine. of frequently in new special of today's Discoveries importance biomarkers for occurring diseases,Discoveries of new biomarkers for frequently occurring diseases are of special importance in today's medicine.,13648
interaction. of social strategy a Ethnocentrism behavioral scale seen on every is,Ethnocentrism is a behavioral strategy seen on every scale of social interaction.,13649
"selects models boosts demonstrate it which increases that Game-theory ethnocentrism reproductive evolution because cooperation, fitness.","Game-theory models demonstrate that evolution selects ethnocentrism because it boosts cooperation, which increases reproductive fitness.",13650
"simulations we test this hypothesis. Here, use agent-based to computer","Here, we use agent-based computer simulations to test this hypothesis.",13651
end-to-end This optimization. as lends opportunities as advanced introspection to for well,This lends to advanced introspection as well as opportunities for end-to-end optimization.,13652
HOG state-of-the-art applications improve Both on approaches. respective the,Both applications improve on the respective state-of-the-art HOG approaches.,13653
good necessary learning supervision representation? strong for visual a Is,Is strong supervision necessary for learning a good visual representation?,13654
Neural images millions a Convolutional semantically-labeled Do to train need of (CNN)? really Network we,Do we really need millions of semantically-labeled images to train a Convolutional Neural Network (CNN)?,13655
visual Our provides the key idea supervision. that tracking is,Our key idea is that visual tracking provides the supervision.,13656
a function Siamese-triplet network to representation. this We a design CNN loss train with ranking,We design a Siamese-triplet network with a ranking loss function to train this CNN representation.,13657
toward the surveys detection. This current review made anomaly article video-based progresses,This review article surveys the current progresses made toward video-based anomaly detection.,13658
"aspect fundamental for that address anomaly most is, feature video video representation. the We detection,","We address the most fundamental aspect for video anomaly detection, that is, video feature representation.",13659
obtained Large data of Internet. quantities can the easily video from unlabeled be,Large quantities of unlabeled video data can be easily obtained from the Internet.,13660
trainable solution document a images. this binarization fully present In for paper degraded we,In this paper we present a fully trainable binarization solution for degraded document images.,13661
"out-of-domain generalization data. good capabilities shows the trained binarization Additionally, classifier on document","Additionally, the trained document binarization classifier shows good generalization capabilities on out-of-domain data.",13662
"it direction will a Given move, scene, move? in is to and what what going","Given a scene, what is going to move, and in what direction will it move?",13663
could non-semantic action form a be considered question Such a prediction. of,Such a question could be considered a non-semantic form of action prediction.,13664
"this convolutional network approach present prediction. we (CNN) for motion based a In neural work,","In this work, we present a convolutional neural network (CNN) based approach for motion prediction.",13665
large previous approaches all margins. outperform We by,We outperform all previous approaches by large margins.,13666
is a Particle Swarm (PSO) problems. solving optimization continuous Optimization nature-inspired meta-heuristic for,Particle Swarm Optimization (PSO) is a nature-inspired meta-heuristic for solving continuous optimization problems.,13667
"purpose, the is dimension mathematically. in each analyzed For particles' potential this","For this purpose, the particles' potential in each dimension is analyzed mathematically.",13668
"the reasonable particles' made. some assumptions if Additionally, are behavior on potential the","Additionally, some reasonable assumptions on the behavior if the particles' potential are made.",13669
KITTI on We model also for estimation dataset. the car train distance a,We also train a model for car distance estimation on the KITTI dataset.,13670
real can approach to direct Results well our perception that show generalize driving images.,Results show that our direct perception approach can generalize well to real driving images.,13671
on website. project available data and Source code are our,Source code and data are available on our project website.,13672
regularity. presents and examples intriguing processes proceed high multiple precision at of which Nature,Nature presents multiple intriguing examples of processes which proceed at high precision and regularity.,13673
activity among (ST). are precancerous of studied The tissues S-transform through and evolution development,The evolution and development of precancerous activity among tissues are studied through S-transform (ST).,13674
"are Additionally, particularly environmental non-bilaterian high not levels. bodyplans expected oxygen require to","Additionally, non-bilaterian bodyplans are not expected to require particularly high environmental oxygen levels.",13675
the panicea. with is demonstrating consistent of low-oxygen Halichondria sponge experimental evidence tolerance the This,This is consistent with experimental evidence demonstrating the low-oxygen tolerance of the sponge Halichondria panicea.,13676
analyze We positions. of choosing strategies propose these several and,We propose and analyze several strategies of choosing these positions.,13677
This plant for video paper proposes novel framework fluorescence processing. a,This paper proposes a novel framework for fluorescence plant video processing.,13678
within plant interested research The photosynthetic in analysis community plant. the is a leaf-level,The plant research community is interested in the leaf-level photosynthetic analysis within a plant.,13679
"as We tracking identify alignment, segmentation, this problem. and multi-leaf a joint","We identify this as a joint multi-leaf segmentation, alignment, and tracking problem.",13680
of also The limitation algorithm studied. is our,The limitation of our algorithm is also studied.,13681
"and show of results the robustness effectiveness, Experimental method. proposed the efficiency,","Experimental results show the effectiveness, efficiency, and robustness of the proposed method.",13682
great from still Event image images for understanding. is importance recognition of,Event recognition from still images is of great importance for image understanding.,13683
"called a new we Neural Object-Scene Specifically, Convolutional architecture, (OS-CNN). design Network","Specifically, we design a new architecture, called Object-Scene Convolutional Neural Network (OS-CNN).",13684
"Furthermore, are complementary to networks and very-deep other. find deep we that the each","Furthermore, we find that the deep and very-deep networks are complementary to each other.",13685
to Potential Application: of development methodology include of reduce study carsickness. a applications this,Application: Potential applications of this study include development of a methodology to reduce carsickness.,13686
significant It problem. potentially is yet under-researched a,It is a potentially significant yet under-researched problem.,13687
"dimension of spatial signal, Independently image etc. image, (one-dimensional images, volume of the","Independently of the image spatial dimension (one-dimensional signal, image, volume of images, etc.",13688
can for problematic These learning established many requirements for to act. methods be,These requirements can be problematic for many established methods for learning to act.,13689
"time free However, settings. planning is most not in real-world","However, planning time is not free in most real-world settings.",13690
Markov and for We the analyze (MDPs). problem Decision formalize Processes metareasoning,We formalize and analyze the metareasoning problem for Markov Decision Processes (MDPs).,13691
"approximations. discuss, motivating optimal we be general metareasoning impractical, to reasons turns out For","For reasons we discuss, optimal general metareasoning turns out to be impractical, motivating approximations.",13692
two and cons approaches pros are elucidated. the The of,The pros and cons of the two approaches are elucidated.,13693
can work view The and non-overlapping have layout. in the of independently cameras larger field,The cameras work independently and can have larger field of view in the non-overlapping layout.,13694
"with. dealt be a should scale ambiguity However, factor","However, a scale factor ambiguity should be dealt with.",13695
"the Kalman filter extended is For approaches, real-time recursive as a estimator. both used","For both approaches, the extended Kalman filter is used as a real-time recursive estimator.",13696
with The are and studied alike. verified real experiments approaches synthetic,The approaches studied are verified with synthetic and real experiments alike.,13697
enumerate provide notes. We a features few and its implementation novel,We enumerate its novel features and provide a few implementation notes.,13698
weighted the also We a algorithm filtering. fast provide for,We also provide a fast algorithm for the weighted filtering.,13699
is to expression. standard differential High-throughput RNA method gene determine sequencing (RNA-seq) now the,High-throughput RNA sequencing (RNA-seq) is now the standard method to determine differential gene expression.,13700
differentially estimates read on variability. count depends of expressed crucially Identifying genes,Identifying differentially expressed genes crucially depends on estimates of read count variability.,13701
"techniques. spectral Traditional with approximately, minimization methods for problem solve this instance","Traditional methods solve this minimization problem approximately, for instance with spectral techniques.",13702
"for occluded Machines, algorithm our object classification. When shows compared performance superior Restricted Boltzmann to","When compared to Restricted Boltzmann Machines, our algorithm shows superior performance for occluded object classification.",13703
based is and method mathematical Gaussian on interest scale based The morphology. map space,The method is based on Gaussian scale space based interest map and mathematical morphology.,13704
It luminance efficiently exudates. sized variation and can suitable it is handle for varied,It can efficiently handle luminance variation and it is suitable for varied sized exudates.,13705
an action There performing. cues reveal what multiple a which person are image is in,There are multiple cues in an image which reveal what action a person is performing.,13706
"an be road, the information. can of presence additional and joggers other trail) source of","road, trail) and the presence of other joggers can be an additional source of information.",13707
"recognition. limited R*CNN that show action is we Last, not to","Last, we show that R*CNN is not limited to action recognition.",13708
"such classification. particular, be R*CNN In used tasks can as fine-grained tackle to attribute also","In particular, R*CNN can also be used to tackle fine-grained tasks such as attribute classification.",13709
People of by on Berkeley this claim reporting performance the dataset. We Attributes state-of-the-art validate,We validate this claim by reporting state-of-the-art performance on the Berkeley Attributes of People dataset.,13710
control experiments acceptable process that Negative the specificity has and and sensitivity. positive screening indicate,Negative and positive control experiments indicate that the screening process has acceptable specificity and sensitivity.,13711
large present web an approach learning amounts data of CNNs. We utilize for to,We present an approach to utilize large amounts of web data for learning CNNs.,13712
"learning, we Specifically by CNN present inspired a curriculum two-step approach for training.","Specifically inspired by curriculum learning, we present a two-step approach for CNN training.",13713
"to images visual representation. First, train easy use an initial we","First, we use easy images to train an initial visual representation.",13714
"with more offers search. search strategy, the reportedly its efficient for potential best-first SSS*,","SSS*, with its best-first search strategy, reportedly offers the potential for more efficient search.",13715
"search. using In algorithm depth-first effect, formulate we a best-first","In effect, we formulate a best-first algorithm using depth-first search.",13716
"SSS*. practice, less typically variants Alpha-Beta than In evaluate nodes","In practice, Alpha-Beta variants typically evaluate less nodes than SSS*.",13717
These have hierarchy. policies to form a,These policies have to form a hierarchy.,13718
"method the proposed increases, the if state space of the ones. However, existing outperforms size","However, if the size of the state space increases, the proposed method outperforms existing ones.",13719
of known collagen repairs fibers. tissue fibrosis to is by of be Tissue result a,Tissue fibrosis is known to be a result of repairs of tissue by collagen fibers.,13720
"""Misrepair"". a of collagen manner A is with fibers repair","A repair with collagen fibers is a manner of ""Misrepair"".",13721
stiffness the organ. of results or of Fibrosis in progressive atrophy organ and an failure,Fibrosis results in stiffness or atrophy of an organ and progressive failure of the organ.,13722
hair-whitening of probably and hair-loss fibrosis. Senile are consequence dermal,Senile hair-loss and hair-whitening are probably consequence of dermal fibrosis.,13723
"foot-and-mouth disease animal Phylogeography, South Key-words: America, trade. virus,","Key-words: Phylogeography, foot-and-mouth disease virus, South America, animal trade.",13724
"are Finally, the by labels super-pixel initial contextual refined further smoothing.","Finally, the initial super-pixel labels are further refined by contextual smoothing.",13725
gene synthetic problem Bayesian the using address optimization. design of We,We address the problem of synthetic gene design using Bayesian optimization.,13726
deal propose We this to issue. three-step with a approach,We propose a three-step approach to deal with this issue.,13727
"the we First, of to Gaussian the behavior cell. process model use a emulate","First, we use a Gaussian process model to emulate the behavior of the cell.",13728
edge-aware The filtering. image filter is guided a for technique,The guided filter is a technique for edge-aware image filtering.,13729
improve We filter. acceleration and this applications will hope current further this performance of popularize,We hope this acceleration will improve performance of current applications and further popularize this filter.,13730
be back-propagation can algorithm. gradients the efficiently pixel-wise computed using The,The pixel-wise gradients can be efficiently computed using the back-propagation algorithm.,13731
"maps use image this the work, In we saliency segmentation. for computed","In this work, we use the computed saliency maps for image segmentation.",13732
assignment method. protein (LBS) of Computational ligand function is prediction a well-established binding sites,Computational prediction of ligand binding sites (LBS) is a well-established protein function assignment method.,13733
"apply (Reyes, the specific LBS algorithm we we recently Here described detection V.M.","Here we apply the specific LBS detection algorithm we recently described (Reyes, V.M.",13734
of our believe LBS this level We confidence increases the criterion predictions. further,We believe this further criterion increases the confidence level of our LBS predictions.,13735
a learning local structured edges the structures exploit CNN of using We by shadow framework.,We exploit the local structures of shadow edges by using a structured CNN learning framework.,13736
"huge In numbers with massive modern research, frequently of are encountered. observations scientific datasets","In modern scientific research, massive datasets with huge numbers of observations are frequently encountered.",13737
real both supported simulation examples. method of The promising data preference by and the is,The promising preference of the method is supported by both simulation and real data examples.,13738
include and changes two point DNA DNA types: changes. mutations chromosome,DNA changes include two types: point DNA mutations and chromosome changes.,13739
are injuries the Severe DNA DNA causes for mutations.,Severe DNA injuries are the causes for DNA mutations.,13740
"source In DNA is cells, of of main the mutations. somatic Misrepair DNA","In somatic cells, Misrepair of DNA is the main source of DNA mutations.",13741
in that can develop the Tumors only tissues are regenerable.,Tumors can only develop in the tissues that are regenerable.,13742
two regenerations tumor-development. repeated the are for cell injuries preconditions cell Repeated and,Repeated cell injuries and repeated cell regenerations are the two preconditions for tumor-development.,13743
"For is prevention, a living style cancer moderate flexible and advised.","For cancer prevention, a moderate and flexible living style is advised.",13744
in characterized the development of Atherosclerosis (APs) atherosclerotic a plaques disease is arterial by endothelium.,Atherosclerosis is a disease characterized by the development of atherosclerotic plaques (APs) in arterial endothelium.,13745
APs part an on on and wall distribution. size The inhomogeneous arterial are of in,The APs in part of an arterial wall are inhomogeneous on size and on distribution.,13746
"of an repair injured of AP development a of result endothelium. general, is In","In general, development of an AP is a result of repair of injured endothelium.",13747
part endothelium damage-sensitivity (Misrepair) and have Altered the local reduced makes increased of repair-efficiency. remodeling,Altered remodeling (Misrepair) makes the local part of endothelium have increased damage-sensitivity and reduced repair-efficiency.,13748
"injuries, and endothelium for of Misrepairs. lipid-infusion, increased will this have Thus, risk part","Thus, this part of endothelium will have increased risk for injuries, lipid-infusion, and Misrepairs.",13749
of development deposition of in and Focalized a lipids of plaque. focalized result accumulation Misrepairs,Focalized accumulation of Misrepairs and focalized deposition of lipids result in development of a plaque.,13750
"self-accelerating. of AP a is endothelium-Misrepair, and circle between lipid-infusion By an viscous growing","By a viscous circle between lipid-infusion and endothelium-Misrepair, growing of an AP is self-accelerating.",13751
"faster grow part APs than arterial wall, younger ones. Within an of older","Within part of an arterial wall, older APs grow faster than younger ones.",13752
"factor an atherosclerosis. growing Therefore, the in self-accelerated of is AP fatal a","Therefore, the self-accelerated growing of an AP is a fatal factor in atherosclerosis.",13753
"scale, manual global parameters, minimal user provides SICP and the using rotation translation intervention.","SICP provides the global scale, rotation and translation parameters, using minimal manual user intervention.",13754
"a proposed. explores to In originally his Gibson, of affordances as this paper tribute theory","In a tribute to Gibson, this paper explores his theory of affordances as originally proposed.",13755
how specific about Representation   can reason we problem. determines a,  Representation determines how we can reason about a specific problem.,13756
more a us easily others. representation helps find Sometimes proof one than,Sometimes one representation helps us find a proof more easily than others.,13757
Most on one reasoning representation. automated within tools focus current reasoning,Most current automated reasoning tools focus on reasoning within one representation.,13758
alternating spectral paper unmixing a on based projection. algorithm fast Dykstra's presents This,This paper presents a fast spectral unmixing algorithm based on Dykstra's alternating projection.,13759
to thus sought obtained projection. The to is converge guaranteed sequence the,The sequence thus obtained is guaranteed to converge to the sought projection.,13760
"small better subformulas when algorithms perform unsatisfiable are MaxSAT early. these found Usually,","Usually, these MaxSAT algorithms perform better when small unsatisfiable subformulas are found early.",13761
"In we formula MaxSAT graph representation. propose to using paper, a resolution-based the partition this","In this paper, we propose to partition the MaxSAT formula using a resolution-based graph representation.",13762
one The and is ends optimal found. the algorithm remains partition only solution when,The algorithm ends when only one partition remains and the optimal solution is found.,13763
"between parts. constraints to Still, approach this agnostic high-level remains object","Still, this approach remains agnostic to high-level constraints between object parts.",13764
"preemptive desirable are which in allow schedules, almost less always practice. algorithms for Existing evacuation","Existing algorithms almost always allow for preemptive evacuation schedules, which are less desirable in practice.",13765
"this recognition In face we scale. paper, study at","In this paper, we study face recognition at scale.",13766
"at and how problems, scale. We and consider both pose recognition evaluate identification verification affects","We consider both verification and identification problems, and evaluate how pose affects recognition at scale.",13767
"plant different consider gradients altitude In we (i.e. particular, the at","In particular, we consider the plant at different altitude gradients (i.e.",13768
"expensive. more training difficult layers makes However, and adding computationally","However, adding layers makes training more difficult and computationally expensive.",13769
simple thumb rule of We be these to where a determine branches added. should formulate,We formulate a simple rule of thumb to determine where these branches should be added.,13770
It beliefs. by to knowledge discover means new completion simple facilitates operations vector of,It facilitates knowledge completion by means of simple vector operations to discover new beliefs.,13771
Extensive experimental proposed improvements. show results the with outperforms significant that model the state-of-the-arts,Extensive experimental results show that the proposed model outperforms the state-of-the-arts with significant improvements.,13772
Decision (i.e. tuples study the and Diagrams relations (MDD) We Multi-valued between,We study the relations between Multi-valued Decision Diagrams (MDD) and tuples (i.e.,13773
the variables). of Cartesian elements of Product,elements of the Cartesian Product of variables).,13774
"tuples an and algorithms in-place we from present some MDD. Then, adding for deleting","Then, we present some in-place algorithms for adding and deleting tuples from an MDD.",13775
Table show experiments competitive are with MDD Some constraints constraints. that,Some experiments show that MDD constraints are competitive with Table constraints.,13776
of or performed genes. differentially over- number the expressed Two under-predicting tools poorly;,Two tools performed poorly; over- or under-predicting the number of differentially expressed genes.,13777
approach and good shown results. evaluation performed Thorough yields the has that has been,Thorough evaluation has been performed and has shown that the approach yields good results.,13778
for We evaluated our several learning machine portfolio. techniques,We evaluated several machine learning techniques for our portfolio.,13779
good on also We results confirm domain. yields our that approach SAT,We also confirm that our approach yields good results on SAT domain.,13780
"equally with PC PC-GES another effective. GES), is algorithm, (a of For combination strong-linearity,","For strong-linearity, another algorithm, PC-GES (a combination of PC with GES), is equally effective.",13781
"to quick Moreover, training a the updating process is derived improve speed. dictionary further","Moreover, a quick dictionary updating process is derived to further improve the training speed.",13782
"approach a proposals. argue We data-driven, object semantic ranking for for","We argue for a data-driven, semantic approach for ranking object proposals.",13783
"nature of and degree Bernstein Lupa\c{s} $(p,q)$-B$\acute{e}$zier reduction for We degree functions. elevation the study","We study the nature of degree elevation and degree reduction for Lupa\c{s} $(p,q)$-B$\acute{e}$zier Bernstein functions.",13784
"$(p,q)$-Bernstein Lupa\c{s} using are represented curves basis. Parametric","Parametric curves are represented using Lupa\c{s} $(p,q)$-Bernstein basis.",13785
"curves. introduce de Lupa\c{s} for We type Casteljau affine algorithm B$\acute{e}$zier $(p,q)$-Bernstein","We introduce affine de Casteljau algorithm for Lupa\c{s} type $(p,q)$-Bernstein B$\acute{e}$zier curves.",13786
similar curves. properties The new have curves to $q$-B$\acute{e}$zier some,The new curves have some properties similar to $q$-B$\acute{e}$zier curves.,13787
are and measured of occur data disciplines. angles Circular in variety scientific in data a,Circular data are data measured in angles and occur in a variety of scientific disciplines.,13788
promise circular Bayesian of flexible to for methods analysis allow data.,Bayesian methods promise to allow for flexible analysis of circular data.,13789
these phases the of specialized involve structures. Some formation large-scale of,Some of these phases involve the formation of specialized large-scale structures.,13790
We then be independence perturbations. through describe evolution manipulated can various how of external the,We then describe how the evolution of independence can be manipulated through various external perturbations.,13791
UCB algorithm of modified applied application Mi-UCT the The to the algorithm trees. is,The Mi-UCT algorithm is the application of the modified UCB algorithm applied to trees.,13792
as Vector or using Finally the Machines(SVM). eye is classified Support open closed state,Finally the eye state is classified as open or closed using Support Vector Machines(SVM).,13793
illumination carried using variation is in (BHE). Histogram Bi Equalization out Compensation,Compensation in illumination variation is carried out using Bi Histogram Equalization (BHE).,13794
actual found to The under is be robust conditions. system driving,The system is found to be robust under actual driving conditions.,13795
"release KOS of versions and both the However, to users. pose new publishers KOS challenges","However, the release of new KOS versions pose challenges to both KOS publishers and users.",13796
random are network neural views deep convolutional (ConvNet) These to train classifiers. used,These random views are used to train deep convolutional neural network (ConvNet) classifiers.,13797
proposed methods cases. markedly CADe performance Our all improve in,Our proposed methods improve CADe performance markedly in all cases.,13798
accuracies been in networks visual problems. many have of neural splendid Deep exhibiting pattern classification,Deep neural networks have been exhibiting splendid accuracies in many of visual pattern classification problems.,13799
an paper rule augmented issue addresses for of decision with classifiers This data. trained,This paper addresses an issue of decision rule for classifiers trained with augmented data.,13800
focus. Discussion of data methods augmentation primary our is of not,Discussion of methods of data augmentation is not our primary focus.,13801
blur done problem. and solving any estimation This explicitly inverse subsequent without is,This is done without explicitly solving any blur estimation and subsequent inverse problem.,13802
weights can insight representations. provide structure the of model into The neural,The structure of the model weights can provide insight into neural representations.,13803
"the contributing to Often, relationships input-output with sparse, are inputs neural few output. a only","Often, neural input-output relationships are sparse, with only a few inputs contributing to the output.",13804
"structured account into part regularizers to for incorporated fitting such model In sparsity, are optimization.","In part to account for such sparsity, structured regularizers are incorporated into model fitting optimization.",13805
"priors, by model However, interpret to difficult make structured imposing regularizers parameters. it learned can","However, by imposing priors, structured regularizers can make it difficult to interpret learned model parameters.",13806
"sampling. stationary efficient Gibbs to schemes generalizations efficient of relaxation lead Hence, deterministic","Hence, efficient deterministic stationary relaxation schemes lead to efficient generalizations of Gibbs sampling.",13807
Eye are crucial understanding in scenes. movements complex,Eye movements are crucial in understanding complex scenes.,13808
learned were via rule. Hebbian generalized weights Place-to-grid a,Place-to-grid weights were learned via a generalized Hebbian rule.,13809
non-negativity constraint Without the a output to the lattice. square converged,Without the non-negativity constraint the output converged to a square lattice.,13810
about active been in for research area two has Saliency decades. computer modeling an vision,Saliency modeling has been an active research area in computer vision for about two decades.,13811
this and basic some models. compare dataset We analyze properties of some successful,We analyze some basic properties of this dataset and compare some successful models.,13812
implementation are For fixed the iteration. equation numerical integro-differential we a an solving point by,For the numerical implementation we are solving an integro-differential equation by a fixed point iteration.,13813
second performance must efficiency and The be balanced. scenario,The second scenario performance and efficiency must be balanced.,13814
"transmission, processing always presents acquisition, image images is Noise and coding, during digital steps. in","Noise is always presents in digital images during image acquisition, coding, transmission, and processing steps.",13815
"this of various express models. noise paper, overview In we brief a","In this paper, we express a brief overview of various noise models.",13816
noise be their analysis by of These models origin. can selected,These noise models can be selected by analysis of their origin.,13817
The fast-to-calculate method shapes. a Hausdorff fractal to complexity has estimate of dimension been fractal,The Hausdorff fractal dimension has been a fast-to-calculate method to estimate complexity of fractal shapes.,13818
"propose this for descriptor recognition. a study, novel visual scene In we place","In this study, we propose a novel scene descriptor for visual place recognition.",13819
We network neural sparse for a designed three-dimensional have implemented data. processing input convolutional,We have implemented a convolutional neural network designed for processing sparse three-dimensional input data.,13820
have cuts-based success many in minimization achieved vision great computer for energy Graph algorithms applications.,Graph cuts-based algorithms have achieved great success in energy minimization for many computer vision applications.,13821
approach. solutions via energy approximated for These multi-label functions provide algorithms move-making,These algorithms provide approximated solutions for multi-label energy functions via move-making approach.,13822
approach This solution lower-energy with generate proposal fuses the to current a a solution.,This approach fuses the current solution with a proposal to generate a lower-energy solution.,13823
"is proposals necessary for the move-making appropriate generating the success approach. of the Thus,","Thus, generating the appropriate proposals is necessary for the success of the move-making approach.",13824
"and approach generate application-independent In ""good"" proposals. an to energy-based this paper, propose we","In this paper, we propose an application-independent and energy-based approach to generate ""good"" proposals.",13825
is effective classes Extensive of functions. across energy different experiments our generation support proposal that,Extensive experiments support that our proposal generation is effective across different classes of energy functions.,13826
real algorithm The on and others proposed both synthetic outperforms problems.,The proposed algorithm outperforms others both on real and synthetic problems.,13827
made the direction Several have approaches been detection. in face of,Several approaches have been made in the direction of face detection.,13828
"a robust Haar-like is features method method. based them, Among","Among them, Haar-like features based method is a robust method.",13829
"work with - of spite In limitations. features the robustness, like some Haar","In spite of the robustness, Haar - like features work with some limitations.",13830
to instance-level ordering. It aims segmentation predict and depth accurate,It aims to predict accurate instance-level segmentation and depth ordering.,13831
One-against-one method adopted classification these is using features.,One-against-one classification method is adopted using these features.,13832
show databases the system. expression on and facial the Experiments proposed effectiveness JAFFE CK+ of,Experiments on CK+ and JAFFE facial expression databases show the effectiveness of the proposed system.,13833
available. and Several estimation routines simulation for are,Several routines for simulation and estimation are available.,13834
of function. the the based empirical matching Estimation the theoretical on with is autocorrelation,Estimation is based on the matching of the empirical with the theoretical autocorrelation function.,13835
Lamellipodium Based of Filament this (FBLM). work in treatment Model numerical We the describe the,We describe in this work the numerical treatment of the Filament Based Lamellipodium Model (FBLM).,13836
two space. dimensional Lagrange-Hermite It of dimensional over composite two is defined elements comprised,It is comprised of composite Lagrange-Hermite two dimensional elements defined over two dimensional space.,13837
in-vitro also novel experiments moving numerical present We compare and of cells. simulations to,We also present novel numerical simulations and compare to in-vitro experiments of moving cells.,13838
is feature or extraction manual required. No preprocessing step,No preprocessing step or manual feature extraction is required.,13839
available developed we descriptors. hand-crafted utilize for can enhancing techniques feature Thus,Thus we can utilize available feature enhancing techniques developed for hand-crafted descriptors.,13840
improved Our results experimental performance. on significantly recognition datasets action benchmark show four,Our experimental results on four benchmark action recognition datasets show significantly improved performance.,13841
of novel framework hybrid involves facial patches. active approach features The from proposed a extracting,The proposed framework involves a novel approach of extracting hybrid features from active facial patches.,13842
time an development of on depends its structural The complexity. organism,The development time of an organism depends on its structural complexity.,13843
complexity. is structural built in of potential The functionality,The potential of functionality is built in structural complexity.,13844
An limited limited animal longevity complexity. has it has structural because,An animal has limited longevity because it has limited structural complexity.,13845
complexity structural of species. and a are survival essential longevity for Limited limited the,Limited structural complexity and limited longevity are essential for the survival of a species.,13846
a ontology knowledge sociocultural paper presents modeling (OntoSOC) approach. This,This paper presents a sociocultural knowledge ontology (OntoSOC) modeling approach.,13847
Human approach (HAT). Theory modeling Activity based is on OntoSOC Engestrom,OntoSOC modeling approach is based on Engestrom Human Activity Theory (HAT).,13848
concepts allowed relationships and between Theory fundamental to them. identify That us,That Theory allowed us to identify fundamental concepts and relationships between them.,13849
differents top-down The define been sub-concepts. used has precess to,The top-down precess has been used to define differents sub-concepts.,13850
algorithms dynamic nature. PTZ difficulty tracking from in evaluating The arises their,The difficulty in evaluating PTZ tracking algorithms arises from their dynamic nature.,13851
online. be type of tracking only can performed This,This type of tracking can only be performed online.,13852
"this new PTZ In paper, a a camera. we framework propose virtual on based evaluation","In this paper, we propose a new evaluation framework based on a virtual PTZ camera.",13853
the setting synaptic the best for weight). possible,the best possible setting for the synaptic weight).,13854
"sampling of all Synaptic four experimental support. makes some which predictions, have","Synaptic sampling makes four predictions, all of which have some experimental support.",13855
"it the more LTP a change the is, variable protocols. synapse should First more during","First the more variable a synapse is, the more it should change during LTP protocols.",13856
"should presynpatic increase Second, rate firing falls. variability as the","Second, variability should increase as the presynpatic firing rate falls.",13857
"variance mean. to PSP Third, PSP should be proportional","Third, PSP variance should be proportional to PSP mean.",13858
"with soma. Fourth, from variability distance increase should the cell","Fourth, variability should increase with distance from the cell soma.",13859
non-redundant between study of We graded sets if-then of attributes. rules describing dependencies properties particular,We study properties of particular non-redundant sets of if-then rules describing dependencies between graded attributes.,13860
smartphone. for tablet computer a or unlocking,for unlocking a tablet computer or smartphone.,13861
"acquisition processing step segmentation, first after i.e. The is fingerprint image","The first processing step after fingerprint image acquisition is segmentation, i.e.",13862
"a repair, Misrepair. abnormality defective a links This the defective also development and","This abnormality links also a defective development and a defective repair, the Misrepair.",13863
"for maintaining of Mis-reconstructions, Misrepairs, the result tissues/organs. of Aging the is a structure","Aging is a result of Mis-reconstructions, the Misrepairs, for maintaining the structure of tissues/organs.",13864
essential processes construction aging-like Mis- the (Misrepair) development for are and Mis-reconstruction of feathers. the,Mis- construction and Mis-reconstruction (Misrepair) are the essential processes for the development of aging-like feathers.,13865
of may components comprehensive examine We important two aid understanding that processes. crucially neurodynamical said,We examine two important components that may crucially aid comprehensive understanding of said neurodynamical processes.,13866
mathematical mechanism we the show finding solutions. this Using that models accelerates right,Using mathematical models we show that this mechanism accelerates finding the right solutions.,13867
"i.e. synaptic we The structural is that plasticity, factor incorporate second","The second factor that we incorporate is structural synaptic plasticity, i.e.",13868
optimal reach that solutions. suffice not Hebbian to show learning does We alone,We show that Hebbian learning alone does not suffice to reach optimal solutions.,13869
"of networks, resulting connected fully the to In subsets topologies components. converge","In the resulting networks, topologies converge to subsets of fully connected components.",13870
"costs on although reduces the Imposing synapses the of number robust. connectivity, components remains connected","Imposing costs on synapses reduces the connectivity, although the number of connected components remains robust.",13871
adaptive Markov A Monte algorithm chain Carlo is novel presented.,A novel adaptive Markov chain Monte Carlo algorithm is presented.,13872
"structure a estimates the Further, correlation density partial that proposed. of algorithm an is","Further, an algorithm that estimates the partial correlation structure of a density is proposed.",13873
two with method empirical for examples. compared is adaption The regular covariance,The method is compared with regular empirical covariance adaption for two examples.,13874
happiness reinforcement What is agents? learning for,What is happiness for reinforcement learning agents?,13875
a satisfying of definition We a desiderata. seek formal list,We seek a formal definition satisfying a list of desiderata.,13876
"is definition i.e. difference proposed error, the Our happiness temporal of","Our proposed definition of happiness is the temporal difference error, i.e.",13877
our of compatible humans. on This is with and desiderata research satisfies most empirical definition,This definition satisfies most of our desiderata and is compatible with empirical research on humans.,13878
discuss state implications examples. and We several,We state several implications and discuss examples.,13879
"expansion, of ability increasing Intelligence an therefore, sense-making. means","Intelligence expansion, therefore, means an increasing ability of sense-making.",13880
SSs present. estimation rough where the and of locations a potential non-selective This the is,This is a rough and non-selective estimation of the locations where the potential SSs present.,13881
"extraction parallel local flow pathway. selective executed in the another Secondly, along of is feature","Secondly, another flow of local feature extraction is executed in parallel along the selective pathway.",13882
of The features size model objects. is to proposed and invariant,The proposed model is invariant to size and features of objects.,13883
task unsuitability overall for analysed Colour-based models their and are is summarised. this,Colour-based models are analysed and their overall unsuitability for this task is summarised.,13884
"but weights determined, analytically cannot also output be only transferrable. The network","The network output weights cannot only be analytically determined, but also transferrable.",13885
"a propose referred multiple model also Extensively, views, as MvEDA. of we","Extensively, we also propose a model of multiple views, referred as MvEDA.",13886
for these research. applications of of active is Development algorithms an field,Development of algorithms for these applications is an active field of research.,13887
"insufficient. availability However, database algorithms such standard of is validate to","However, availability of standard database to validate such algorithms is insufficient.",13888
paper Infra-Red creation database of discusses under illumination. Near (NIR) This the such created a,This paper discusses the creation of such a database created under Near Infra-Red (NIR) illumination.,13889
"driven trains. plasticity usually is In systems, neural spike by synaptic","In neural systems, synaptic plasticity is usually driven by spike trains.",13890
"the variability influences trains spike of how remains efficacy variability unclear. the However,","However, how the variability of spike trains influences the efficacy variability remains unclear.",13891
for still indicates personal authentication. fingerprints that biometric Market popular are research most the modality,Market research indicates that fingerprints are still the most popular biometric modality for personal authentication.,13892
of new (e.g. modalities Even onset the with,Even with the onset of new modalities (e.g.,13893
"the digits) human hand sections, of structure.","sections, digits) of the human hand structure.",13894
not infrastructure iris) is possible or economical to limits. due,iris) is not possible due to economical or infrastructure limits.,13895
network. semantic a We segmentation deconvolution learning by propose novel algorithm a,We propose a novel semantic segmentation algorithm by learning a deconvolution network.,13896
a We complementary discovery as the processes: tracking. of combination and problem formulate two,We formulate the problem as a combination of two complementary processes: discovery and tracking.,13897
for We variety nearest baseline captioning. a image of neighbor explore approaches,We explore a variety of nearest neighbor baseline approaches for image captioning.,13898
attribute of We to propose discover meaning. shades,We propose to discover shades of attribute meaning.,13899
both and resulting visually The semantic are models precise.,The resulting models are both semantic and visually precise.,13900
"catering on users' prediction attribute to By interpretations, images. improve novel accuracy they","By catering to users' interpretations, they improve attribute prediction accuracy on novel images.",13901
"in proposed were processing/reasoning different languages engines different Recently, communities. and stream query","Recently, different query languages and stream processing/reasoning engines were proposed in different communities.",13902
This the a new in reducing method paper classifier. describes error for a,This paper describes a new method for reducing the error in a classifier.,13903
method the that classify show some benchmark datasets. successfully Tests can,Tests show that the method can successfully classify some benchmark datasets.,13904
are with also an shape made. earlier wave paper Some comparisons,Some comparisons with an earlier wave shape paper are also made.,13905
Reliable exact safe and visibility assessment traffic. air for of is essential,Reliable and exact assessment of visibility is essential for safe air traffic.,13906
validity demonstrate images proposed of on method. the the Results foggy,Results on foggy images demonstrate the validity of the proposed method.,13907
multi-scale image explore nets (CNNs) convolutional We for classification. neural,We explore multi-scale convolutional neural nets (CNNs) for image classification.,13908
a output single approaches from Contemporary layer. features extract,Contemporary approaches extract features from a single output layer.,13909
introduced example the illustrate means with concepts by scenario rescue We an services. of dealing,We illustrate the introduced concepts by means of an example scenario dealing with rescue services.,13910
the MIS show estimator. simulations associated performance mixture} \mbox{\emph{partial of excellent Computer deterministic the,Computer simulations show the excellent performance of the associated \mbox{\emph{partial deterministic mixture} MIS estimator.,13911
"enough were slim deep for nor so proposed neither HCCR However, enough. previous far CNNs","However, previous CNNs so far proposed for HCCR were neither deep enough nor slim enough.",13912
correspondences simultaneously maximize feature multiple estimated The cycle across consistency and pairwise affinities images.,The estimated correspondences simultaneously maximize pairwise feature affinities and cycle consistency across multiple images.,13913
are non-natural Nucleic oligonucleotides (PNA) mimics. Peptide Acids,Peptide Nucleic Acids (PNA) are non-natural oligonucleotides mimics.,13914
this index multiscale a introduce method We to problem. resolve (mFRI) flexibility-rigidity,We introduce a multiscale flexibility-rigidity index (mFRI) method to resolve this problem.,13915
"results the of territory into catchment Finally, context the and analysis. put exploitation we our","Finally, we put our results into the context of the exploitation territory and catchment analysis.",13916
our open and for code publicly algorithm All is source available.,All code for our algorithm is open source and publicly available.,13917
of execution using stage The the first computations results stage. the second allows repeated of,The second stage allows repeated execution of computations using the results of the first stage.,13918
"present for video data. ""wearable"" summarization a or We approach camera egocentric","We present a video summarization approach for egocentric or ""wearable"" camera data.",13919
in-depth comparison paper between This detailed and presents both devices. a,This paper presents a detailed and in-depth comparison between both devices.,13920
of for vital in videos. importance are human understanding features action Visual,Visual features are of vital importance for human action understanding in videos.,13921
results features show and Experimental hand-crafted outperform TDDs previous features. that deep-learned,Experimental results show that TDDs outperform previous hand-crafted features and deep-learned features.,13922
presented is An unsupervised cell approach segmentation here.,An unsupervised cell segmentation approach is presented here.,13923
set was used cytoplasm segmentation. for model Level,Level set model was used for cytoplasm segmentation.,13924
We tumor range of for concentration of the shrinkage. an existence evidence cells optimal effector,We evidence the existence of an optimal concentration range of effector cells for tumor shrinkage.,13925
different a The image photographer taken viewpoint. other from by another is,The other image is taken by another photographer from a different viewpoint.,13926
view. not Our the of even overlapping when is applicable fields images method do have,Our method is applicable even when the images do not have overlapping fields of view.,13927
be images are other present photographers scene. available may These who the online from at,These images may be available online from other photographers who are present at the scene.,13928
"result, more as learning process it a the As progresses. efficient becomes","As a result, the learning process becomes more efficient as it progresses.",13929
a nonconvex framework with matrix present We unified for penalties. low-rank estimation,We present a unified framework for low-rank matrix estimation with nonconvex penalties.,13930
and Numerical our real world on both synthetic datasets corroborate experiments theory.,Numerical experiments on both synthetic and real world datasets corroborate our theory.,13931
"barcode usually annotations this proposed. Whereas annotations we mean with are descriptions, in paper textual","Whereas with annotations we usually mean textual descriptions, in this paper barcode annotations are proposed.",13932
"are introduced. barcodes (RBC) particular, In Radon","In particular, Radon barcodes (RBC) are introduced.",13933
the This neuroanatomy. gives on of bars estimate rise results error to of an computational,This gives rise to an estimate of error bars on the results of computational neuroanatomy.,13934
inference. Bayesian a for with optimization evaluate empirically stochastic annealing strategy We posterior variational,We empirically evaluate a stochastic annealing strategy for Bayesian posterior optimization with variational inference.,13935
Their partially performance annotated is usually datasets. on evaluated,Their performance is usually evaluated on partially annotated datasets.,13936
observations given interesting study. are this statistics analyzed Many and and in,Many interesting observations and statistics are given and analyzed in this study.,13937
from the level of value can found PERCLOS. be The fatigue of,The level of fatigue can be found from the value of PERCLOS.,13938
frames processed. is eye total closed of It the the to frames ratio,It is the ratio of closed eye frames to the total frames processed.,13939
used. (NIR) is Red Near night time Infra active During illumination,During night time active Near Infra Red (NIR) illumination is used.,13940
GPUs Ray becoming days. common tracing quite on is these,Ray tracing on GPUs is becoming quite common these days.,13941
on GPUs. for general some We frameworks have even ray tracing,We even have some general frameworks for ray tracing on GPUs.,13942
"system. this define a colour In distance for paper, we HSL the","In this paper, we define a distance for the HSL colour system.",13943
"fuzzy is Next, the algorithm a clustering for distance proposed colour construction. used","Next, the proposed distance is used for a fuzzy colour clustering algorithm construction.",13944
well-known algorithm. related presented c-means algorithm to is fuzzy The the,The presented algorithm is related to the well-known fuzzy c-means algorithm.,13945
"as colour Finally, algorithm is clustering the used method. reduction","Finally, the clustering algorithm is used as colour reduction method.",13946
experimental demonstrate approach. results the effectiveness to The are obtained presented of our,The obtained experimental results are presented to demonstrate the effectiveness of our approach.,13947
"completion termed Thus, tensor as our ""smooth proposed is method PARAFAC (SPC).""","Thus, our proposed method is termed as ""smooth PARAFAC tensor completion (SPC).""",13948
is process. the modelled populations using finite often in Evolution Moran classical,Evolution in finite populations is often modelled using the classical Moran process.,13949
"model, an that represents population. unstructured the which original Moran from","that from the original Moran model, which represents an unstructured population.",13950
detected are targets is the images. realized of by that Geo-reference in automatically signalized means,Geo-reference is realized by means of signalized targets that are automatically detected in the images.,13951
data We the labelled training synthesizing photo-realistic propose perfectly time. of in a fraction,We propose synthesizing perfectly labelled photo-realistic training data in a fraction of the time.,13952
To improve robustness such of the methods w.r.t.,To improve the robustness of such methods w.r.t.,13953
In of all the issues. paper our we aforementioned three address,In our paper we address all three of the aforementioned issues.,13954
"variational propose First, the model depth. a we operates that directly on Cartesian novel","First, we propose a novel variational model that operates directly on the Cartesian depth.",13955
"Moreover, property. with regulariser a second-order direct edge-preservation we employ","Moreover, we employ a direct second-order regulariser with edge-preservation property.",13956
solutions constraints. This requiring regulariser construction yields additional consistency without valid direct by,This direct regulariser yields by construction valid solutions without requiring additional consistency constraints.,13957
"minimisation also we on based an framework coarse-to-fine Finally, scheme. novel explicit propose a alternating","Finally, we also propose a novel coarse-to-fine minimisation framework based on an alternating explicit scheme.",13958
in-the-wild is task. a natural language descriptions Generating videos for challenging,Generating natural language descriptions for in-the-wild videos is a challenging task.,13959
"classification. object deep However, CNN architectures for centered-positioned are these single-label designed","However, these deep CNN architectures are designed for single-label centered-positioned object classification.",13960
is is region that Ours end-to-end multi-scale of first the capable trainable processing. architecture,Ours is the first end-to-end trainable architecture that is capable of multi-scale region processing.,13961
other support Our video flexible tasks. efficient potentially processing to be and extended can architecture,Our flexible and efficient architecture can potentially be extended to support other video processing tasks.,13962
"state-of-the-art struggle. certain algorithms common However, challenging and the still model classes for","However, for certain challenging and common model classes the state-of-the-art algorithms still struggle.",13963
"types econometrics. applications including of many models interest, of tracking, arise epidemiology, in and Both","Both types of models arise in many applications of interest, including tracking, epidemiology, and econometrics.",13964
minimization our methods. in the properties convergence asymptotic of results prove We the and,We prove the convergence and asymptotic properties of the minimization results in our methods.,13965
"rectifying this In report of heat capability a-helices. work, we the","In this work, we report the heat rectifying capability of a-helices.",13966
"many However, post-processing require data applications of offline.","However, many applications require post-processing of data offline.",13967
conditionally a linear models. smoothing consider class of the We problem Gaussian for,We consider the smoothing problem for a class of conditionally linear Gaussian models.,13968
investigates vision of This one computer most problems: the image paper fundamental segmentation.,This paper investigates one of the most fundamental computer vision problems: image segmentation.,13969
object-independent supervised hierarchical We to propose image approach a segmentation.,We propose a supervised hierarchical approach to object-independent image segmentation.,13970
globally solutions be by to inferred Final optimal then the finding can model segmentations efficiently.,Final segmentations can then be inferred by finding globally optimal solutions to the model efficiently.,13971
models reconstruct Many of network developed methods to been those have interactions.,Many methods have been developed to reconstruct network models of those interactions.,13972
"While informative, capture is to heterogeneity population. that represented the models fail in any aggregate","While informative, aggregate models fail to capture the heterogeneity that is represented in any population.",13973
we propose aggregate method network Here a engineer sample-specific networks models. reverse from to,Here we propose a method to reverse engineer sample-specific networks from aggregate network models.,13974
"impressive understanding has quite remains elusive. shown intelligence few lately, a general breakthroughs","has shown quite a few impressive breakthroughs lately, understanding general intelligence remains elusive.",13975
the paper novel intelligence. to general we understanding offer theoretical In approach a,In the paper we offer a novel theoretical approach to understanding general intelligence.,13976
of with We conceptual approach. introduction start the a brief current,We start with a brief introduction of the current conceptual approach.,13977
research. of conclude a with for future questions number We,We conclude with a number of questions for future research.,13978
The is among that results R-GoDec the show algorithms. fastest the robust,The results show that R-GoDec is the fastest among the robust algorithms.,13979
"enhancement algorithm, the hyper-kurtosis HKMDHE done is contrast application. In on based","In HKMDHE algorithm, contrast enhancement is done on the hyper-kurtosis based application.",13980
"processing, has and computational applications image bilateral vision, in The photography. computer diverse filter","The bilateral filter has diverse applications in image processing, computer vision, and computational photography.",13981
"poorly however, at filter, large The to known is noise levels. perform","The filter, however, is known to perform poorly at large noise levels.",13982
approach We semi-supervised localizes object a videos. present in multiple unknown long instances that,We present a semi-supervised approach that localizes multiple unknown object instances in long videos.,13983
integrate (LIF) used The and leaky neuron standard for numerical simulations. model neuronal fire represents,The leaky integrate and fire (LIF) neuron represents standard neuronal model used for numerical simulations.,13984
This usage of of numbers machine point inevitable in makes simulation. the the floating course,This makes inevitable the usage of machine floating point numbers in the course of simulation.,13985
"this is However, e.g. type necessary, decision of","However, decision of this type is necessary, e.g.",13986
periodic regimes figure in a to network. dynamical reverberating,to figure periodic dynamical regimes in a reverberating network.,13987
"independently decoded were a ""shirt"") ""fold"" presentation. single Constituents and (e.g., from","Constituents (e.g., ""fold"" and ""shirt"") were independently decoded from a single presentation.",13988
"accuracies. high-variability such approaches undesirably the organs report previous low pancreas, For as","For high-variability organs such as the pancreas, previous approaches report undesirably low accuracies.",13989
"stability, deviations. numerical smaller Deep reflected offer by standard patch CNN confidences labeling more","Deep CNN patch labeling confidences offer more numerical stability, reflected by smaller standard deviations.",13990
formalism characterizing latent We a new spatiotemporal sub-events. templates introduce of also,We also introduce a new formalism of spatiotemporal templates characterizing latent sub-events.,13991
challenging that under Our successfully conditions. inference address demonstrate we results above tasks,Our results demonstrate that we successfully address above inference tasks under challenging conditions.,13992
using edge-aware Typically smoothing process edge-preserving this is filters. achieved,Typically this edge-preserving smoothing process is achieved using edge-aware filters.,13993
small preserve unwanted edges. if as filters However contain they such structures well may,However such filters may preserve unwanted small structures as well if they contain edges.,13994
"of i.e. In the problem vision, point-sets, we unstructured have of computer creating out graphs","In computer vision, we have the problem of creating graphs out of unstructured point-sets, i.e.",13995
"one techniques After point-sets. apply Graph could the to Matching register building original the data-graph,","After building the data-graph, one could apply Graph Matching techniques to register the original point-sets.",13996
technique data could aims Our process. which registration graphs at creating point-set the help,Our technique aims at creating data graphs which could help the point-set registration process.,13997
Delaunay a databases of the compare different triangulation. number on our with algorithm We,We compare our algorithm on a number of different databases with the Delaunay triangulation.,13998
"While most are edge relatively they perform well on only detection images. clean fast, methods","While most edge detection methods are fast, they perform well only on relatively clean images.",13999
"only can edges filters. using reliably local in be detected images such Indeed,","Indeed, edges in such images can be reliably detected using only local filters.",14000
"that Unfortunately, achieve this goal methods slow. existing are quite","Unfortunately, existing methods that achieve this goal are quite slow.",14001
"context gained much this has years. issue in attention The of recent in ""inconsistency""","The issue of ""inconsistency"" in this context has gained much attention in recent years.",14002
is justified some of the indices only However usage by heuristics. the,However the usage of the indices is justified only by some heuristics.,14003
"is unclear still really It ""measure"". they what","It is still unclear what they really ""measure"".",14004
structural language C BNs. presented on free Graph_sampler inference fast a software We here for,We presented here Graph_sampler a fast free C language software for structural inference on BNs.,14005
much and the a new kernel jumping in faster Metropolis-Hastings proposed strategy We algorithm.,We proposed a new and much faster jumping kernel strategy in the Metropolis-Hastings algorithm.,14006
"uses source C storage. very fast, compact, The low memory distributed code is disk and","The source C code distributed is very compact, fast, uses low memory and disk storage.",14007
Maximum reconstruction frequently-discussed is parsimony the tree in methods estimation. of one phylogenetic most,Maximum parsimony is one of the most frequently-discussed tree reconstruction methods in phylogenetic estimation.,14008
is from It maximum parsimony extend necessary like phylogenetic trees networks. to therefore to concepts,It is therefore necessary to extend concepts like maximum parsimony from phylogenetic trees to networks.,14009
"parsimony In paper, concepts, this the analyze problem under i.e. big these so-called we two","In this paper, we analyze the so-called big parsimony problem under these two concepts, i.e.",14010
networks their investigate and properties. we parsimonious maximum analyze,we investigate maximum parsimonious networks and analyze their properties.,14011
"on different likelihood of phylogenetic concepts parallels we investigate networks. to parsimony some Lastly,","Lastly, we investigate some parallels of parsimony to different likelihood concepts on phylogenetic networks.",14012
"discussed. backside the coin, composition SIL Also the of the is","Also the backside of the coin, the SIL composition is discussed.",14013
"change. experienced archaic new genomes, number expansion genomic in indicating This rapid transcript copy","This new transcript experienced copy number expansion in archaic genomes, indicating rapid genomic change.",14014
"basal bodies. is of lipofuscin An cells, which age group a spot aggregated pathologically contain","An age spot is pathologically a group of aggregated basal cells, which contain lipofuscin bodies.",14015
of bodies a cell. of is aging sign of Accumulation a lipofuscin,Accumulation of lipofuscin bodies is a sign of aging of a cell.,14016
is a of of tissue spots. the age for basis development Aging,Aging of a tissue is the basis for development of age spots.,14017
cells. flat spot accumulation A of lipofuscin results from containing,A flat spot results from accumulation of lipofuscin containing cells.,14018
"in the resulting spot in the this of three-dimension, ""grows"" way, In protruding spot.","In this way, the spot ""grows"" in three-dimension, resulting in protruding of the spot.",14019
"ability better generalize to approach has without fine-tuning. Additionally, across datasets our","Additionally, our approach has better ability to generalize across datasets without fine-tuning.",14020
to is a Set Programming of Qsmodels environment. Answer gaming interactive novel application,Qsmodels is a novel application of Answer Set Programming to interactive gaming environment.,14021
phenomena. interpreting theory in novel aging theory accumulation a Misrepair is,Misrepair accumulation theory is a novel theory in interpreting aging phenomena.,14022
"species our longevity, MA also on theory evolution. aging, and adaptation, understanding improves premature","MA theory improves also our understanding on longevity, premature aging, adaptation, and species evolution.",14023
organism complexity the is longevity in of structure. The hidden of its of an potential,The potential of longevity of an organism is hidden in the complexity of its structure.,14024
and Cartesian-based crown-based explored: Pyramids. Two Spatial Spatial configurations novel Pyramid are,Two novel Spatial Pyramid configurations are explored: Cartesian-based and crown-based Spatial Pyramids.,14025
"propose architecture, pixel deep image SegNet, a wise semantic labelling. novel for We","We propose a novel deep architecture, SegNet, for semantic pixel wise image labelling.",14026
feature These map input methods layer dimensions. lack maps to to deep a mechanism,These methods lack a mechanism to map deep layer feature maps to input dimensions.,14027
"hoc e.g. features, methods to to upsample resort ad They","They resort to ad hoc methods to upsample features, e.g.",14028
map SegNet to outputs problems to these encoder image pixel by overcomes learning labels.,SegNet overcomes these problems by learning to map encoder outputs to image pixel labels.,14029
fixed-size the a we and analysis of provide for age-abundance relationships growing comprehensive communities. Here,Here we provide a comprehensive analysis of the age-abundance relationships for fixed-size and growing communities.,14030
"in factor and beta a morphogenesis, growth growth, cell (TGF$\beta$) differentiation. role central plays Transforming","Transforming growth factor beta (TGF$\beta$) plays a central role in morphogenesis, growth, and cell differentiation.",14031
simulation creativity. the particular is modelling computational challenge One in of AI and,One particular challenge in AI is the computational modelling and simulation of creativity.,14032
and creative the are key learning aspects process. Feedback experience from of,Feedback and learning from experience are key aspects of the creative process.,14033
we using investigate creative could model. in we how implement Here systems feedback a social,Here we investigate how we could implement feedback in creative systems using a social model.,14034
that Writers the argue the We creative setting of Workshop describes the process. anatomy,We argue that the Writers Workshop setting describes the anatomy of the creative process.,14035
by the We relevance to of AI. discussing aspects of this conclude model other broader,We conclude by discussing the broader relevance of this model to other aspects of AI.,14036
"In propose detection. propagation novel label this based we a paper, for method saliency","In this paper, we propose a novel label propagation based method for saliency detection.",14037
Image as Problem w.r.t. of Problem Decomposition a the (MP) Multicut Formulations,Formulations of the Image Decomposition Problem as a Multicut Problem (MP) w.r.t.,14038
graph attention. superpixel a considerable received have,a superpixel graph have received considerable attention.,14039
"w.r.t. In MP instances the of contrast,","In contrast, instances of the MP w.r.t.",14040
We (LMP). the MP terms of a generalization with long-range propose,We propose a generalization of the MP with long-range terms (LMP).,14041
Both the motion tackled. of recovery the is fields images the high-resolution them and relating,Both the recovery of the high-resolution images and the motion fields relating them is tackled.,14042
"a and leads to non-smooth problem. large-dimensional, non-convex This","This leads to a large-dimensional, non-convex and non-smooth problem.",14043
algorithmic framework to latter. address the We propose an,We propose an algorithmic framework to address the latter.,14044
Our methods fast evaluation techniques non-differentiable/non-convex and gradient on modern relies for optimization approach problems.,Our approach relies on fast gradient evaluation methods and modern optimization techniques for non-differentiable/non-convex problems.,14045
environments. process Species essentially a populations is random evolution their interaction biological and of between,Species evolution is essentially a random process of interaction between biological populations and their environments.,14046
"physical fluctuations. a to As evolution statistical parameters are models in result, some subject","As a result, some physical parameters in evolution models are subject to statistical fluctuations.",14047
implication discussed. threshold for of change randomization the the antiviral due to The is strategies,The implication of the threshold change due to the randomization for antiviral strategies is discussed.,14048
its from consider Finance customers Triprima should of the head with manager. the consent borrowing,Triprima Finance should consider borrowing from its customers with the consent of the head manager.,14049
the Decision-making Finance activities manually. process analysis carried by PT out Triprima at,Decision-making activities at PT Triprima Finance carried out by the analysis process manually.,14050
decisions to be facilitate the determine to expected is to taken. kaposko,is expected to facilitate kaposko to determine the decisions to be taken.,14051
"of the et report, work revisit we this In al. Pilleboue","In this report, we revisit the work of Pilleboue et al.",14052
subalgebras. relations example of The provides such interval convex an distributive of subclass one well-known,The well-known subclass of convex interval relations provides one such an example of distributive subalgebras.,14053
proposed model Eigen stochastic The al. by Feng et,The stochastic Eigen model proposed by Feng et al.,14054
The underlying this phenomenon well examined. been of yet has cause not,The underlying cause of this phenomenon has not yet been well examined.,14055
"derivation custom inversion implementation This the typically algorithms. requires design, and of sophisticated","This typically requires the custom design, derivation and implementation of sophisticated inversion algorithms.",14056
The availability rely on covariances. of calculations the the,The calculations rely on the availability of the covariances.,14057
"and horizontal transfers, neofunctionalization). duplication subsequent gene gene","horizontal gene transfers, gene duplication and subsequent neofunctionalization).",14058
models. intrinsic require of Many building previous such world computations,Many such computations require previous building of intrinsic world models.,14059
different how tasks. networks may solve Here we neural show spiking these,Here we show how spiking neural networks may solve these different tasks.,14060
have or contain The connectivity. constrained a synaptic dendritic and nonlinearities networks,The networks contain dendritic or synaptic nonlinearities and have a constrained connectivity.,14061
for We outputs connections. such learning networks rules or with recurrent then combine,We then combine such networks with learning rules for outputs or recurrent connections.,14062
"to process without ancillary practical information. a Moreover, images procedure allows autocalibration","Moreover, a practical autocalibration procedure allows to process images without ancillary information.",14063
efficiency computational data with method. assess the the and the Experiments real of accuracy,Experiments with real data assess the accuracy and the computational efficiency of the method.,14064
"allow space occur exploration to only approaches the most of sequentially. parameter However, the proposed","However, most proposed approaches only allow the exploration of the parameter space to occur sequentially.",14065
"explore. Often, simultaneously batches it desirable propose to is to of parameter values","Often, it is desirable to simultaneously propose batches of parameter values to explore.",14066
processing large available. when is facilities case This parallel the particularly are,This is particularly the case when large parallel processing facilities are available.,14067
of being physical facets or These could optimized. the process facilities computational be,These facilities could be computational or physical facets of the process being optimized.,14068
in ups experiments experimental samples to many be biological allow set processed. simultaneously several,in biological experiments many experimental set ups allow several samples to be simultaneously processed.,14069
"in running time, more much alternatives. resulting with elaborate well, compares The algorithm","The resulting algorithm compares well, in running time, with much more elaborate alternatives.",14070
"function is Lipschitz The a function. that interest, the assumes of continuous $f$, approach","The approach assumes that the function of interest, $f$, is a Lipschitz continuous function.",14071
both of Saccades fast and the the direction. simultaneous are eyes same in movement,Saccades are fast and simultaneous movement of both the eyes in the same direction.,14072
saccadic the of velocity duration. the peak is SR to ratio saccadic,SR is the ratio of peak saccadic velocity to the saccadic duration.,14073
method. spotting by query string paper In propose word a this segmentation-free we,In this paper we propose a segmentation-free query by string word spotting method.,14074
boost we re-ranking step performance. in Finally introduce retrieval to order a,Finally we introduce a re-ranking step in order to boost retrieval performance.,14075
Our four steps. algorithm consists basic of,Our algorithm consists of four basic steps.,14076
"First, method generates objectness map the objectness via our hypotheses.","First, our method generates the objectness map via objectness hypotheses.",14077
show that method Our outperforms our approaches. experimental results state-of-the-art extensive,Our extensive experimental results show that our method outperforms state-of-the-art approaches.,14078
"need set. For elements the systems, one in the address only many most-probable","For many systems, one need only address the most-probable elements in the set.",14079
the greatly fall is if most probabilities this threshold. problem below simplified of The,The problem is greatly simplified if most of the probabilities fall below this threshold.,14080
essential. low- that and transform events properly high-probability mixes of Finding represents probability a is,Finding a probability transform that properly represents mixes of low- and high-probability events is essential.,14081
to are the Four simplify operators defined derivations. help,Four operators are defined to help simplify the derivations.,14082
"genres. composed are even of multiple videos concepts or Often,","Often, videos are composed of multiple concepts or even genres.",14083
"etc. action, For nature, instance, sports, videos may news contain","For instance, news videos may contain sports, action, nature, etc.",14084
classification the corresponds Each a to visual space. word region in,Each visual word corresponds to a region in the classification space.,14085
"Such correlations genetic dynamics. pointed should, famously strongly Maynard influence John Smith as evolutionary out,","Such genetic correlations should, as John Maynard Smith famously pointed out, strongly influence evolutionary dynamics.",14086
particular of particle convergence follows. discussion A the of filters,A discussion of the convergence of particular particle filters follows.,14087
"interest increase detector of use image. Harris-Laplace the the to regions we identify To accuracy,","To increase the accuracy, we use Harris-Laplace detector to identify the interest regions of image.",14088
"build Retrieval (RBIR). we Image the Then, Region-Based","Then, we build the Region-Based Image Retrieval (RBIR).",14089
"signatures. image retrieval, images efficient binary the into we storage For encode and","For the efficient image storage and retrieval, we encode images into binary signatures.",14090
image is The its regions. of interest a binary from created signature,The binary signature of a image is created from its interest regions.,14091
"images. evaluate we on models COREL's created Finally, the","Finally, we evaluate the created models on COREL's images.",14092
and identifying aiming of components. design We space revisit relating its visualizations the at,We revisit the design space of visualizations aiming at identifying and relating its components.,14093
learning accuracy attribute-guided We that this retrieval improvement. show key a is factor for,We show that this attribute-guided learning is a key factor for retrieval accuracy improvement.,14094
learning makes our dataset the is work feasible. network large-scale contribution a Another of which,Another contribution of our work is a large-scale dataset which makes the network learning feasible.,14095
retrieval of networks in performance different evaluate We the extensively configurations.,We extensively evaluate the retrieval performance of networks in different configurations.,14096
phylogenetic the process Most evolutionary models stationary and that reversible. assume is,Most phylogenetic models assume that the evolutionary process is stationary and reversible.,14097
the the differ of The perturbations. imposed structure models on degree in,The models differ in the degree of structure imposed on the perturbations.,14098
the Bayesian analysis chain framework Carlo Monte using performed in methods. Markov The is,The analysis is performed in the Bayesian framework using Markov chain Monte Carlo methods.,14099
It as gradient (SBGP). patterns is termed binary structural,It is termed as structural binary gradient patterns (SBGP).,14100
 depend hypothesize detection to locations. algorithms proposal object networks State-of-the-art region on object ,  State-of-the-art object detection networks depend on region proposal algorithms to hypothesize object locations.,14101
"generated at in time step, $N_{x}$ the $\mathcal{X}$). each particles state are space","at each time step, $N_{x}$ particles are generated in the state space $\mathcal{X}$).",14102
algorithm. of the to calibrate automatically course the discuss in how $N_{x}$ We,We discuss how to automatically calibrate $N_{x}$ in the course of the algorithm.,14103
original perform used So representations can that be visual sparse of tasks. recognition data to,So that sparse representations of original data can be used to perform visual recognition tasks.,14104
demonstrate approach of standard on our We effectiveness the datasets.,We demonstrate the effectiveness of our approach on standard datasets.,14105
par competitive method better than state-of-the-art Our methods. performs on or,Our method performs on par or better than competitive state-of-the-art methods.,14106
computation problem. edge powerless path counting-based The for this is non-uniformity,The edge counting-based path computation is powerless for this non-uniformity problem.,14107
questions investigated Biological recent algorithmic have persistent been intensely concerning years. phylogeny and in,Biological and algorithmic questions concerning persistent phylogeny have been intensely investigated in recent years.,14108
research. image emerging topic image a very is Color processing segmentation for,Color image segmentation is a very emerging topic for image processing research.,14109
a process. Choosing is space image color for a proper very segmentation issue important color,Choosing a proper color space is a very important issue for color image segmentation process.,14110
spaces. are LAB and frequently chosen two the color HSV Generally,Generally LAB and HSV are the two frequently chosen color spaces.,14111
"their . and we the psnr parameters: mse consider For performance, measuring","For measuring their performance, we consider the parameters: mse and psnr .",14112
LAB. It performing that is found HSV space is better color than,It is found that HSV color space is performing better than LAB.,14113
method of the This improves queries. significantly accuracy,This method significantly improves the accuracy of queries.,14114
"However, increases also cost this the of computing.","However, this also increases the cost of computing.",14115
"COREL's our images. evaluated is Finally, work on","Finally, our work is evaluated on COREL's images.",14116
on colors approaches of This image the the retrieval of image. chapter system base the,This chapter approaches the image retrieval system on the base of the colors of image.,14117
"shape, topology, its own and distinguishing features. has It","It has its own shape, topology, and distinguishing features.",14118
"are systems challenging testing tasks. tracking As facial developing and such,","As such, developing and testing facial tracking systems are challenging tasks.",14119
"FdMiee facial of behaviors captures facial variations. under hardware, information different and types environmental Using","Using different types of hardware, FdMiee captures facial information under environmental and facial behaviors variations.",14120
images visual matrix descriptors of a high-dimensional exhibit for natural Most structure.,Most of high-dimensional visual descriptors for images exhibit a natural matrix structure.,14121
layers There LLE. structure external two to are,There are two structure layers external to LLE.,14122
over weighted a distribution probability compactly to possible formulas worlds. uses encode logic Markov,Markov logic uses weighted formulas to compactly encode a probability distribution over possible worlds.,14123
"of the size theory this exponential Unfortunately, is general. in","Unfortunately, the size of this theory is exponential in general.",14124
network article conceptor images. new This in classifying demonstrates classifier a based,This article demonstrates a new conceptor network based classifier in classifying images.,14125
faces features conditions. challenging identities in large number of a It real-world from,It features faces from a large number of identities in challenging real-world conditions.,14126
"bilinear case, fine-tuned each model the shows substantial In standard CNN. improvements over the","In each case, the fine-tuned bilinear model shows substantial improvements over the standard CNN.",14127
sequential for probabilities. estimation and splitting algorithms We of describe these resampling the importance,We describe sequential importance resampling and splitting algorithms for the estimation of these probabilities.,14128
provide proposed results of algorithm. We effectiveness the the numerical demonstrating,We provide numerical results demonstrating the effectiveness of the proposed algorithm.,14129
"on single recent most However, the even isolated a approaches focus of case hand.","However, even most recent approaches focus on the case of a single isolated hand.",14130
"colors, details. color fine-scale gradients The faithfully reproduce and prints resulting","The resulting prints faithfully reproduce colors, color gradients and fine-scale details.",14131
"though, neural activated inherently different by tasks. system same the Importantly can be","Importantly though, the same neural system can be activated by inherently different tasks.",14132
"stimuli. involving Here, two we and proof-of-principle auditory present studies visual","Here, we present two proof-of-principle studies involving visual and auditory stimuli.",14133
on cells quantum is A in pluripotency physically model induced chemically the proposed. and stem,A quantum model on the chemically and physically induced pluripotency in stem cells is proposed.,14134
the for deduced. is formulas calculated form is an rate rate The analytical and transitional,The transitional rate is calculated and an analytical form for the rate formulas is deduced.,14135
hyperspectral in the is the pixels. presence One data mixed of analysis of challenges,One of the challenges in hyperspectral data analysis is the presence of mixed pixels.,14136
pixels the spatial of sensors. are Mixed hyperspectral of low resolution result,Mixed pixels are the result of low spatial resolution of hyperspectral sensors.,14137
methods pixel of unmixing endmembers fractions. mixed a into decompose abundance a and set Spectral,Spectral unmixing methods decompose a mixed pixel into a set of endmembers and abundance fractions.,14138
and abundance fractions this in paper. both Sparseness matrices used spectral signatures are constraint on,Sparseness constraint on both spectral signatures and abundance fractions matrices are used in this paper.,14139
the show proposed methods. algorithm outperforms previously that Results other proposed,Results show that the proposed algorithm outperforms other previously proposed methods.,14140
"we structural our revisit algorithm recently Corresponding, selection proposed based on and sparsity. stability","Corresponding, we revisit our recently proposed algorithm based on stability selection and structural sparsity.",14141
time. the analysis to multi-center data applied first It for MRI the is,It is applied to the multi-center MRI data analysis for the first time.,14142
in common actions still and interactions monocular images? are recognizable human Which,Which common human actions and interactions are recognizable in monocular still images?,14143
at is time? many performing a How a person,How many is a person performing at a time?,14144
an for localization effective in spatio-temporal propose action videos. realistic We approach,We propose an effective approach for spatio-temporal action localization in realistic videos.,14145
approach. a proposals video using tracking-by-detection tracks high-scoring the throughout It then,It then tracks high-scoring proposals throughout the video using a tracking-by-detection approach.,14146
instance-level on relies simultaneously and class-level tracker detectors. Our,Our tracker relies simultaneously on instance-level and class-level detectors.,14147
"propose current the analyze to the In such wavelets images. we paper,","In the current paper, we propose the wavelets to analyze such images.",14148
successfully objects. testing continuous binary The to applied transform was wireframe the wavelet,The continuous wavelet transform was successfully applied to the testing wireframe binary objects.,14149
of restored wireframe binary the objects. locations structure correspond The to the testing,The restored locations correspond to the structure of the testing wireframe binary objects.,14150
codetection on process employ by constraint pairing the the We weak video with sentences. semantic,We employ weak semantic constraint on the codetection process by pairing the video with sentences.,14151
"and populations? Here, we ask: across escape of are rates host comparable reversion infected","Here, we ask: are rates of escape and reversion comparable across infected host populations?",14152
We estimates across escape cohorts. find the examined consistent rate,We find consistent escape rate estimates across the examined cohorts.,14153
Canadian a are population. rates infected between African South also host consistent and Reversion,Reversion rates are also consistent between a Canadian and South African infected host population.,14154
"more quickly. to Similarly, capacity mutants greatest tend viral inducing revert costs the to replicative","Similarly, mutants inducing the greatest costs to viral replicative capacity tend to revert more quickly.",14155
images. technique Relies faces using base standard Eigen Data on,Relies on Eigen faces technique using standard Data base images.,14156
"always selected image as a major topic In vision, by research segmentation is researchers. computer","In computer vision, image segmentation is always selected as a major research topic by researchers.",14157
segmentation in used clustering process. image Many researchers,Many researchers used clustering in image segmentation process.,14158
requires there still of such approaches. improvement But,But still there requires improvement of such approaches.,14159
"segmentation paper, is In based novel clustering this for image approach a proposed.","In this paper, a novel approach for clustering based image segmentation is proposed.",14160
"task. on for color and lab Here, choose importance give space we this","Here, we give importance on color space and choose lab for this task.",14161
filtered the with image sobel segmented filter. is Then,Then the segmented image is filtered with sobel filter.,14162
are observe and the PSNR MSE performance. evaluated to The values,The MSE and PSNR values are evaluated to observe the performance.,14163
planning for grids. widely used path on are,are widely used for path planning on grids.,14164
LIAN empirically. We examine both theoretically and,We examine LIAN both theoretically and empirically.,14165
it is and restrictions). complete (under some show that sound We,We show that it is sound and complete (under some restrictions).,14166
"brain of damage, We revisit idea i.e. the","We revisit the idea of brain damage, i.e.",14167
that generalized approach convolutions matrix uses to many reduce the fact implementations The multiplications. efficient,The approach uses the fact that many efficient implementations reduce generalized convolutions to matrix multiplications.,14168
"very on AlexNet, the In comparison method achieves performance. the competitive","In the comparison on AlexNet, the method achieves very competitive performance.",14169
was defined. real interval for new bounded a Also distance,Also a new distance for bounded real interval was defined.,14170
image in vision a computer Text challenging community. research the is problem super-resolution open yet,Text image super-resolution is a challenging yet open research problem in the computer vision community.,14171
"of low-resolution In character typical particular, recognition hamper systems. (OCR) performance images the optical","In particular, low-resolution images hamper the performance of typical optical character recognition (OCR) systems.",14172
kernel arbitrary with approximate precision). bounded can any,can approximate any bounded kernel with arbitrary precision).,14173
trade-off a of function cell. each type defines cell Intrinsic of,Intrinsic trade-off function of each cell defines a type of cell.,14174
for areas distribution. V. specific cholerae geographic identified We potential,We identified specific geographic areas for potential V. cholerae distribution.,14175
presents a of bifuzzy five-valued representation sets. paper This,This paper presents a five-valued representation of bifuzzy sets.,14176
reviews covered of methods. in are most complex These prediction challenges,These challenges are covered in most reviews of complex prediction methods.,14177
"an the addressed needs important to membrane be challenge of that prediction complexes. However, is","However, an important challenge that needs to be addressed is the prediction of membrane complexes.",14178
"methods. outperforms resulting Despite color its algorithm state-of-the-art the current estimation simplicity, constancy","Despite its simplicity, the resulting estimation algorithm outperforms current state-of-the-art color constancy methods.",14179
"to luminance-to-chromaticity a the learn Next, classifier ""end-to-end"". propose method we","Next, we propose a method to learn the luminance-to-chromaticity classifier ""end-to-end"".",14180
properties spice with pungent Black bitter having a and is odour. medicinal sativa) (Nigella cumin,Black cumin (Nigella sativa) is a spice having medicinal properties with pungent and bitter odour.,14181
interactions i.e. target of The such identified molecular one,The molecular interactions of one such target identified i.e.,14182
"concept based concept used generate to EventNet representation Finally, library videos. event is of the","Finally, the EventNet concept library is used to generate concept based representation of event videos.",14183
potential browsing. and event offers It for retrieval great,It offers great potential for event retrieval and browsing.,14184
"the inaccurate by still TBIR users. tags and from deficient applications provided However, the suffer","However, the TBIR applications still suffer from the deficient and inaccurate tags provided by users.",14185
refine completion tags. propose algorithm the we to further matrix a And,And we propose a matrix completion algorithm to further refine the tags.,14186
model video This generative a done training is to by frames. predict,This is done by training a generative model to predict video frames.,14187
"imperfect incorrectly consider test, might positives). be (false susceptible individuals infected to An however,","An imperfect test, however, might incorrectly consider susceptible individuals to be infected (false positives).",14188
"potentially In the reduces epidemic this in the of misclassifying individuals. testing case, expense","In this case, testing reduces the epidemic in the expense of potentially misclassifying individuals.",14189
"turn but at to deemed Susceptible $\theta$. ""susceptible individuals infected"" rate","Susceptible individuals turn to ""susceptible but deemed infected"" at rate $\theta$.",14190
"a state ""infected individuals tested rate Infected positive"" at go and $\alpha$. to","Infected individuals go to a state ""infected and tested positive"" at rate $\alpha$.",14191
functions specificity. of sensitivity rates of are test's these Both and,Both of these rates are functions of test's sensitivity and specificity.,14192
method. our Bayesian A gives results offline shows that an as study method similar simulation,A simulation study shows that our method gives similar results as an offline Bayesian method.,14193
We also over traditional Gaussian proposal proposal. the that the find improves skewed-normal,We also find that the skewed-normal proposal improves over the traditional Gaussian proposal.,14194
part essential mining mining. is an data data analysis itemset Frequent of and,Frequent itemset mining is an essential part of data analysis and data mining.,14195
encodings works problem frequent SAT-based propose of for the interesting itemsets. Recent discovering,Recent works propose interesting SAT-based encodings for the problem of discovering frequent itemsets.,14196
"effects restart, context, learning. the of and this heuristics we deeply clauses study branching In","In this context, we deeply study the effects of restart, branching heuristics and clauses learning.",14197
lower to compilation particular show we and new bounds DNNF knowledge In on SDD forms.,In particular we show new lower bounds on knowledge compilation to SDD and DNNF forms.,14198
simple be must less FBDDs. than which examples exponentially also SDDs concise We for derive,We also derive simple examples for which SDDs must be exponentially less concise than FBDDs.,14199
problem We event the address specific of video retrieval.,We address the problem of specific video event retrieval.,14200
This offers in complexity localizes videos. of parts gain significant a matching accurately the and,This offers a significant gain in complexity and accurately localizes the matching parts of videos.,14201
"is decompressing descriptors. case, without the this performed video retrieval In","In this case, video retrieval is performed without decompressing the descriptors.",14202
of a of We alignment videos. consider set the temporal also,We also consider the temporal alignment of a set of videos.,14203
given enables global synchronous videos scene. of the a alignment playback The of temporal,The global temporal alignment enables synchronous playback of the videos of a given scene.,14204
Variational for scalable a technique inference inference. is Bayesian approximate,Variational inference is a scalable technique for approximate Bayesian inference.,14205
difficult algorithms requires to automate. calculations; it makes Deriving this variational tedious model-specific inference,Deriving variational inference algorithms requires tedious model-specific calculations; this makes it difficult to automate.,14206
"automatic (ADVI). variational variational inference differentiation algorithm, propose an automatic We inference","We propose an automatic variational inference algorithm, automatic differentiation variational inference (ADVI).",14207
a a only Bayesian provides The and dataset; else. model user nothing,The user only provides a Bayesian model and a dataset; nothing else.,14208
a of no make assumptions We class and conjugacy models. support broad,We make no conjugacy assumptions and support a broad class of models.,14209
variational the appropriate and variational objective. automatically algorithm The an family optimizes determines,The algorithm automatically determines an appropriate variational family and optimizes the variational objective.,14210
"a Stan implement ADVI We available now), (code programming in probabilistic framework.","We implement ADVI in Stan (code available now), a probabilistic programming framework.",14211
a train million the mixture We images. quarter model on,We train the mixture model on a quarter million images.,14212
in we variational model use inference With ADVI on any Stan. write we can,With ADVI we can use variational inference on any model we write in Stan.,14213
new data of brought The challenges scale Bayesian has inference. modern to,The modern scale of data has brought new challenges to Bayesian inference.,14214
"are computationally particular, algorithms conventional In large expensive very sets. data for MCMC","In particular, conventional MCMC algorithms are computationally very expensive for large data sets.",14215
are Existing EP-MCMC algorithms approximation accuracy resampling. in and by limited difficulty,Existing EP-MCMC algorithms are limited by approximation accuracy and difficulty in resampling.,14216
"a that problems. new In solves EP-MCMC these propose this PART article, we algorithm","In this article, we propose a new EP-MCMC algorithm PART that solves these problems.",14217
illustrating theoretical provide and extensive We justification empirical performance. experiments,We provide theoretical justification and extensive experiments illustrating empirical performance.,14218
"present new We object YOLO, a to detection. approach","We present YOLO, a new approach to object detection.",14219
detection perform work repurposes on object to Prior detection. classifiers,Prior work on object detection repurposes classifiers to perform detection.,14220
"Finally, learns representations general YOLO objects. of very","Finally, YOLO learns very general representations of objects.",14221
the vehicle. to Recognition Systems are License number used license Plate plate of determine a,License Plate Recognition Systems are used to determine the license plate number of a vehicle.,14222
recognize number system uses the Optical Recognition current The mainly to Character plate.,The current system mainly uses Optical Character Recognition to recognize the number plate.,14223
this There are to several system. problems,There are several problems to this system.,14224
applications (IoT). include Internet Things of under several Practical areas,Practical applications include several areas under Internet of Things (IoT).,14225
revolution virus. possibility eradicating to the therapeutic consider of This us leads,This therapeutic revolution leads us to consider possibility of eradicating the virus.,14226
"cascade is required. of for However, care this, effective an","However, for this, an effective cascade of care is required.",14227
"difficult achieve. will eradication However, to be","However, eradication will be difficult to achieve.",14228
"English. for the of specifically designed ideas, are However, majority algorithms systems and existing","However, majority of the existing ideas, algorithms and systems are specifically designed for English.",14229
"are directions promising research future discussed. for Moreover,","Moreover, promising directions for future research are discussed.",14230
"Current empirical lack such demonstrate methods, but impressive Bayesian Quadrature, performance analysis. as theoretical","Current methods, such as Bayesian Quadrature, demonstrate impressive empirical performance but lack theoretical analysis.",14231
to challenge important An convergence guarantees. probabilistic rigorous reconcile integrators these with is,An important challenge is to reconcile these probabilistic integrators with rigorous convergence guarantees.,14232
"state-of-the-art and FWBQ In competitive optimisation. simulations, with Frank-Wolfe out-performs methods alternatives based is on","In simulations, FWBQ is competitive with state-of-the-art methods and out-performs alternatives based on Frank-Wolfe optimisation.",14233
"harmful, demanding involving a precise making. are decision events Emergency fire and potentially fast","Emergency events involving fire are potentially harmful, demanding a fast and precise decision making.",14234
There models. color-based using for are video methods fire on detection several,There are several methods for fire detection on video using color-based models.,14235
addressed question The how has extensively. humans been problem solve of,The question of how humans solve problem has been addressed extensively.,14236
"overlooked. seems study the effectiveness of process of the to However, this be direct","However, the direct study of the effectiveness of this process seems to be overlooked.",14237
or forces show Results that to become rescaling. to sparse multiplicative either weights noise invariant,Results show that multiplicative noise forces weights to become either sparse or invariant to rescaling.,14238
a almost in science common and disciplines. is Sphere all engineering fitting problem,Sphere fitting is a common problem in almost all science and engineering disciplines.,14239
are Most methods in behavior. of iterative available,Most of methods available are iterative in behavior.,14240
how other with popular some behaves. and comparison this methods method also have shown We,We have also shown some comparison with other popular methods and how this method behaves.,14241
are approximations gradient. two costs efficient novel by Computational this reduced to,Computational costs are reduced by two novel efficient approximations to this gradient.,14242
setups wide in challenging problem. baseline Finding is a correspondences,Finding correspondences in wide baseline setups is a challenging problem.,14243
integrates novel a deformation method a introduce We that model.,We introduce a novel method that integrates a deformation model.,14244
function a its optimization. further employ majorization-minimization and We robust for matching utilize cost for,We further utilize a robust cost function for matching and employ majorization-minimization for its optimization.,14245
more existing maps Our accurate than that approaches. significantly experiments finds indicate our method,Our experiments indicate that our method finds significantly more accurate maps than existing approaches.,14246
starting zero deploy goal system with a to high-accuracy training Our examples. is,Our goal is to deploy a high-accuracy system starting with zero training examples.,14247
"the over model improves As decreases. queries the on reliance crowdsourcing time,","As the model improves over time, the reliance on crowdsourcing queries decreases.",14248
"and on sentiment We tested image approach three datasets---named-entity our recognition, classification, classification.","We tested our approach on three datasets---named-entity recognition, sentiment classification, and image classification.",14249
part learning. form process a probabilistic Gaussian models core of machine (GP),Gaussian process (GP) models form a core part of probabilistic machine learning.,14250
paper each be this available replicate in experiment shortly. will to Code,Code to replicate each experiment in this paper will be available shortly.,14251
useful Such an is projection. invariance is but photography unique perspective to in,Such an invariance is useful in photography but is unique to perspective projection.,14252
extend We model to also slopes of analysis lines. our,We also extend our analysis to model slopes of lines.,14253
"We catadioptric and XSlit using our XSlit mirrors. real cameras, analyses validate panoramas,","We validate our analyses using real XSlit cameras, XSlit panoramas, and catadioptric mirrors.",14254
"efficient approach real to in simulation. enough interactive run rates time at our is Besides,","Besides, our approach is efficient enough to run at interactive rates in real time simulation.",14255
"In there the debate mechanisms Saccharomyces cerevisiae, underlying is an ongoing important this regarding asymmetry.","In Saccharomyces cerevisiae, there is an ongoing debate regarding the mechanisms underlying this important asymmetry.",14256
suggested. in progress partial for AGI is A measure,A measure for partial progress in AGI is suggested.,14257
"reasoning fundamental grounded proposed. on Based is a ability, that system Abstract","Abstract Based on that fundamental ability, a grounded reasoning system is proposed.",14258
"trouble developments. real-time been its large has a However, to most complexity computational","However, its large computational complexity has been a trouble to most real-time developments.",14259
this The past growing decade seen a in interest process. has automating,The past decade has seen a growing interest in automating this process.,14260
"approach automatic is proposed novel paper, of a hybrid for In this detection landmarks. cephalometric","In this paper, a novel hybrid approach is proposed for automatic detection of cephalometric landmarks.",14261
on those which edges. to tracing suggested landmarks located are is predict Edge method,Edge tracing method is suggested to predict those landmarks which are located on edges.,14262
least named last but estimation. the method not analysis The based is,The last but not the least method is named analysis based estimation.,14263
directly. relations a Therefore desired the estimating the in has suggested method novelty third,Therefore the third suggested method has a novelty in estimating the desired relations directly.,14264
individual used Deep in a recognize of scene. networks people are actions to the,Deep networks are used to recognize the actions of individual people in a scene.,14265
a step inference probabilistic mimics refinement message-passing step to in a similar model. graphical This,This refinement step mimics a message-passing step similar to inference in a probabilistic graphical model.,14266
more The that informative. the model other is tree for is clusters representing,The other is that the tree model for representing clusters is more informative.,14267
find method activations as embedding an uses images. Neural similar Convolutional semantically The Network to,The method uses Convolutional Neural Network activations as an embedding to find semantically similar images.,14268
"typical based unigram images, is on the these From frequencies. most selected caption","From these images, the most typical caption is selected based on unigram frequencies.",14269
"can and in work the verification modes. identification Furthermore, scheme both","Furthermore, the scheme can work in both identification and verification modes.",14270
linear our call We response (LRVB). method variational Bayes,We call our method linear response variational Bayes (LRVB).,14271
"posterior. about the make no Indeed, the form assumptions of true we","Indeed, we make no assumptions about the form of the true posterior.",14272
low canonical in machine multidimensional problem a in Learning data structure is learning. dimensional of,Learning of low dimensional structure in multidimensional data is a canonical problem in machine learning.,14273
"matrix we algorithm In online compute this to factorizations. propose paper, an","In this paper, we propose an online algorithm to compute matrix factorizations.",14274
The dictionary algorithm matrix. low-rank to updates performs,The algorithm performs low-rank updates to dictionary matrix.,14275
We for algorithm extend further handling the missing data.,We extend the algorithm further for handling missing data.,14276
and series online simulated The in data. tasks time evaluated real-world forecasting is method on,The method is evaluated in online time series forecasting tasks on simulated and real-world data.,14277
trees and phylogenetic (MUL-trees). a multi-labelled relationship close networks There between is,There is a close relationship between phylogenetic networks and multi-labelled trees (MUL-trees).,14278
"in $F$ properties In $U$ operations the more detail. of paper, we study this and","In this paper, we study properties of the operations $U$ and $F$ in more detail.",14279
"To analogue of develop a do graph we phylogenetic this, fibrations.","To do this, we develop a phylogenetic analogue of graph fibrations.",14280
way to coarse gives arguments. compare a This,This gives a coarse way to compare arguments.,14281
and convergence. properties: well-behaved two We that show model counting our normalization has,We show that our counting model has two well-behaved properties: normalization and convergence.,14282
theory with phenomenology are programmed consistent death. both show that Here and we,Here we show that both theory and phenomenology are consistent with programmed death.,14283
dynamics. is complicated with true domains large especially in This,This is especially true in large domains with complicated dynamics.,14284
prove policy. this bootstrapping near-optimal a process returns that We,We prove that this bootstrapping process returns a near-optimal policy.,14285
"results representations. policies learned without in with skills requiring policy complex planning better Thus,","Thus, planning with learned skills results in better policies without requiring complex policy representations.",14286
for DTRN Codes the be will available.,Codes for the DTRN will be available.,14287
work action recognition video. This targets in human,This work targets human action recognition in video.,14288
descriptor of tracks The body along human appearance and motion aggregates parts. information,The descriptor aggregates motion and appearance information along tracks of human body parts.,14289
We Cooking method and and challenging JHMDB recent MPII our datasets. evaluate on the,We evaluate our method on the recent and challenging JHMDB and MPII Cooking datasets.,14290
consistent of For shows the the method datasets both improvement state our over art.,For both datasets our method shows consistent improvement over the state of the art.,14291
"from more verification, is much face face Different identification demanding.","Different from face verification, face identification is much more demanding.",14292
The SNR is using method. the eigenvalue computed,The SNR is computed using the eigenvalue method.,14293
of to The a is algorithm our pair only input images. stereo,The only input to our algorithm is a pair of stereo images.,14294
network to We appearance classes. deep extract semantic for features use neural a the,We use a deep neural network to extract the appearance features for semantic classes.,14295
in urban Our Daimler scene competing dataset. method approaches other outperforms segmentation,Our method outperforms other competing approaches in Daimler urban scene segmentation dataset.,14296
semantic architecture using a network heterogeneous propose neural annotations. semi-supervised We for segmentation novel deep,We propose a novel deep neural network architecture for semi-supervised semantic segmentation using heterogeneous annotations.,14297
"paper, the imaging problem is addressed. pulsed of this reconstruction In and terahertz","In this paper, the problem of terahertz pulsed imaging and reconstruction is addressed.",14298
for sparsity-inducing this A approach purpose. is proposed,A sparsity-inducing approach is proposed for this purpose.,14299
"a incomplete interpolation simple is data. noisy First, applied to","First, a simple interpolation is applied to incomplete noisy data.",14300
"we representation. a achieve image present pyramidal To invariance, scale","To achieve scale invariance, we present a pyramidal image representation.",14301
entitled Visual algorithm Circular FREAK-ORB Odometry We novel present a (CFORB).,We present a novel Visual Odometry algorithm entitled Circular FREAK-ORB (CFORB).,14302
Two order been remove matches. geometric utilized feature visual descriptor have in to constraints invalid,Two visual geometric constraints have been utilized in order to remove invalid feature descriptor matches.,14303
These been algorithm. previously constraints Visual utilized have Odometry not in a,These constraints have not previously been utilized in a Visual Odometry algorithm.,14304
feature transform the convolutional space. structured to to discriminative We activations highly propose another,We propose to transform the structured convolutional activations to another highly discriminative feature space.,14305
to been that Here and used mathematical epithelia. model we approaches different have review computational,Here we review different mathematical and computational approaches that have been used to model epithelia.,14306
field imaging. via light captured data input The are,The input data are captured via light field imaging.,14307
requires This problem that is ill-posed regularization. an,This is an ill-posed problem that requires regularization.,14308
Our applies or curvature. absolute quadratic to lower algorithm bound optimization,Our lower bound optimization algorithm applies to quadratic or absolute curvature.,14309
dendritic voltage modulatory Regulation realized on of the by the terminal. tuning was applied integration,Regulation of dendritic integration was realized by tuning the voltage applied on the modulatory terminal.,14310
brain-inspired for Our neuromorphic a results approach provide new-concept building systems.,Our results provide a new-concept approach for building brain-inspired neuromorphic systems.,14311
of maps of flow gradient the divergence of the particles. amount the gives The neighboring,The gradient of the flow maps gives the amount of divergence of the neighboring particles.,14312
(University Florida). central UCF dataset of is crowd of conducted The on experiment the,The experiment is conducted on the crowd dataset of UCF (University of central Florida).,14313
feasible mRNAs. a made analysis expressed has of set of RNA-seq the whole Motivation:,Motivation: RNA-seq has made feasible the analysis of a whole set of expressed mRNAs.,14314
references. RNA-seq due reads is infeasible lack of high-quality to of sometimes Mapping-based assembly,Mapping-based assembly of RNA-seq reads sometimes is infeasible due to lack of high-quality references.,14315
"with moderate Further, efficient consumption. is computationally Bermuda memory","Further, Bermuda is computationally efficient with moderate memory consumption.",14316
Networks performance Challenges ImageNet Convolutional contest. Visual demonstrate Neural high Recognition on Large-Scale,Convolutional Neural Networks demonstrate high performance on ImageNet Large-Scale Visual Recognition Challenges contest.,14317
"show for results Nevertheless, classes. overall image the published performance all the only","Nevertheless, the published results only show the overall performance for all image classes.",14318
presentation for has path model traditional the model. This than image better single,This model has better presentation for image than the traditional single path model.,14319
to technique semantic context deep a We segmentation. adding present for global networks convolutional for,We present a technique for adding global context to deep convolutional networks for semantic segmentation.,14320
"example, the concrete consider we multicellularity. a to As transition","As a concrete example, we consider the transition to multicellularity.",14321
"we where of interacting a individuals study isolated model Specifically, stochastic communities (e.g.","Specifically, we study a stochastic model where isolated communities of interacting individuals (e.g.",14322
a (e.g. cells) individuality higher-order undergo to transition,cells) undergo a transition to higher-order individuality (e.g.,14323
address framework. in providing method this We implemented by a shortcoming a total-evidence Bayesian,We address this shortcoming by providing a total-evidence method implemented in a Bayesian framework.,14324
learning? unlabeled visual video How augment can,How can unlabeled video augment visual learning?,14325
"to We feature generalize ""steady"" to feature analysis analysis. slow propose","We propose to generalize slow feature analysis to ""steady"" feature analysis.",14326
"shapes. this computing paper, partial propose correspondence non-rigid between In functional for method a we","In this paper, we propose a method for computing partial functional correspondence between non-rigid shapes.",14327
that very settings. challenging cope our approach with correspondence We can show,We show that our approach can cope with very challenging correspondence settings.,14328
"unnecessary. such jointly, are non-maximum predictions generate as Because common suppression steps post-processing we","Because we generate predictions jointly, common post-processing steps such as non-maximum suppression are unnecessary.",14329
this a In we et proposed work Bruckstein by confirm al. theory,In this work we confirm a theory proposed by Bruckstein et al.,14330
brains characters? robustly How our recognize can,How can our brains robustly recognize characters?,14331
"and features are The corners, on. junctions so essential crosses,","The essential features are crosses, corners, junctions and so on.",14332
These for computer. useful recognition by character features be a may,These features may be useful for character recognition by a computer.,14333
"the difficult. of extraction However, is features","However, extraction of the features is difficult.",14334
"including extract graphs to skeleton and distorted, handwritten We noisy, characters. images achieved from","We achieved to extract skeleton graphs from images including distorted, noisy, and handwritten characters.",14335
"noise concern. its to always a However, and are the sensitivity resolution low","However, its low resolution and sensitivity to the noise are always a concern.",14336
robustness demonstrates the noise. sufficient It to also,It also demonstrates sufficient robustness to the noise.,14337
encouraging e.g. upsampling in The maintained case performance of large is the even,The encouraging performance is maintained even in the case of large upsampling e.g.,14338
game envelopes a host to agent. an The of two indistinguishable presents,The host of a game presents two indistinguishable envelopes to an agent.,14339
agent. is allocated randomly selected to One of the envelopes and the,One of the envelopes is randomly selected and allocated to the agent.,14340
of regression statistics of applied analysis involves Most data.,Most of applied statistics involves regression analysis of data.,14341
"presents software This a Bayesian Regression: menu-driven and stand-alone Parametric Models. Nonparametric package, paper and","This paper presents a stand-alone and menu-driven software package, Bayesian Regression: Nonparametric and Parametric Models.",14342
output by text menu and/or generated other Additional graphical can options. mouse-clicking be,Additional text and/or graphical output can be generated by mouse-clicking other menu options.,14343
cancer. third most the common cancers Colorectal are of type,Colorectal cancers are the third most common type of cancer.,14344
"inherent are aforementioned modelling generally. biological multiscale of the challenges more tissue Moreover, the to","Moreover, the aforementioned challenges are inherent to the multiscale modelling of biological tissue more generally.",14345
monocular low cameras. in with cost This presents for a environments method outdoor paper ego-positioning,This paper presents a method for ego-positioning in outdoor environments with low cost monocular cameras.,14346
facets studied temporal many are phenomena by that have The different communities.,The temporal phenomena have many facets that are studied by different communities.,14347
"data are heterogeneous and produced. large In Web, handled Semantic","In Semantic Web, large heterogeneous data are handled and produced.",14348
also explicit normalization and incorporates HuTO information. reasoning to certain rules,HuTO also incorporates normalization and reasoning rules to explicit certain information.,14349
also knowledge HuTO the a which proposes to base an temporal approach dimension content. associates,HuTO also proposes an approach which associates a temporal dimension to the knowledge base content.,14350
aspect. temporal considering not by information facilitates This retrieval or the,This facilitates information retrieval by considering or not the temporal aspect.,14351
very thus to useful from estimate the structure be data. It would dependency,It would thus be very useful to estimate the dependency structure from data.,14352
typical formulation. naturally These not by the Dirichlet-multinomial captured are dependencies,These dependencies are not naturally captured by the typical Dirichlet-multinomial formulation.,14353
elements mid-level relies construct representation. The a Bag-of-Elements visual to second on retrieved the one,The second one relies on the retrieved mid-level visual elements to construct a Bag-of-Elements representation.,14354
"chapter stereoscopic and in state-of-the-art current future directions describes This the work. cinema, of","This chapter describes the current state-of-the-art in stereoscopic cinema, and directions of future work.",14355
analysis. organ yet segmentation problem for an important medical is challenging Automatic image,Automatic organ segmentation is an important yet challenging problem for medical image analysis.,14356
an pancreas abdominal with anatomical organ high is The variability. very,The pancreas is an abdominal organ with very high anatomical variability.,14357
for learn class of assign ConvNets being Our superpixel each region probabilities pancreas. to,Our ConvNets learn to assign class probabilities for each superpixel region of being pancreas.,14358
paper but effective simple This bridge to aperture gap. coded a presents solution scene-adaptive this,This paper presents a simple but effective scene-adaptive coded aperture solution to bridge this gap.,14359
spatial Therefore and propagation a scheme. we temporal adopt partitioning,Therefore we adopt a spatial partitioning and temporal propagation scheme.,14360
"apply in these orientations following sequentially Secondly, with frames. the the scheme coding we proposed","Secondly, we sequentially apply the proposed coding scheme with these orientations in the following frames.",14361
scheme sampler. MCMC proposed sampling as the the refer We split to,We refer to the proposed sampling scheme as the MCMC split sampler.,14362
tensor of regression is extension Nonparametric proposed.,Nonparametric extension of tensor regression is proposed.,14363
"Gaussian local we To prior. estimator Bayesian a functions, process the estimate with develop","To estimate local functions, we develop a Bayesian estimator with the Gaussian process prior.",14364
of histopathological of a evaluation Glomerulus kidneys. step images microscopy in is detection key,Glomerulus detection is a key step in histopathological evaluation of microscopy images of kidneys.,14365
"to encountered. intensities immunohistochemistry extensive Moreover, in due staining of variations are their heterogeneity also","Moreover, extensive variations of their intensities due to heterogeneity in immunohistochemistry staining are also encountered.",14366
"suitable for Finally, variant a of abstract data. computations with noisy the algorithm, we develop","Finally, we develop a variant of the abstract algorithm, suitable for computations with noisy data.",14367
drawn than from alternate of significantly higher Samples are our quality model approaches.,Samples drawn from our model are of significantly higher quality than alternate approaches.,14368
It and programmed ImageJ. Java is compatible in with is,It is programmed in Java and is compatible with ImageJ.,14369
"mutations lymphoma, of In in genes proteins frequently are observed. histone modifying","In lymphoma, mutations in genes of histone modifying proteins are frequently observed.",14370
algorithms. model that improvements over object our yields show We state-of-the-art significant proposal,We show that our model yields significant improvements over state-of-the-art object proposal algorithms.,14371
is Crowd in step important surveillance many an tasks. flow segmentation video,Crowd flow segmentation is an important step in many video surveillance tasks.,14372
literature causal between The split non-dualistic and decision evidential theory theory decision theory. decision is,The non-dualistic decision theory literature is split between causal decision theory and evidential decision theory.,14373
and and important Frequency processes. dependent in ecological fluctuations roles evolutionary play demographic selection,Frequency dependent selection and demographic fluctuations play important roles in evolutionary and ecological processes.,14374
in reflected even be fluctuations of This environments. should population in the constant size,This should be reflected in fluctuations of the population size even in constant environments.,14375
"population. demographic small stochasticity the In extinction instead lead entire populations, to may of the","In small populations, demographic stochasticity may instead lead to the extinction of the entire population.",14376
"this universal variability? individuals, aforementioned Is the considering across also description","Is this description also universal across individuals, considering the aforementioned variability?",14377
applications. in Salient detection processing task object become important many image an has,Salient object detection has become an important task in many image processing applications.,14378
"genome-wide neuroanatomy. This set triggered a renaissance in data brain-wide, has","This brain-wide, genome-wide data set has triggered a renaissance in neuroanatomy.",14379
"no However, is available. of cells exhaustive taxonomy brain","However, no exhaustive taxonomy of brain cells is available.",14380
the of cell yields This on bars types. region-specificity error,This yields error bars on the region-specificity of cell types.,14381
"a novel of for this three-dimensional we paper, images. introduce method In parametric segmentation","In this paper, we introduce a novel parametric method for segmentation of three-dimensional images.",14382
which describes a of scheme parametric We propose evolution parametric surfaces. the,We propose a parametric scheme which describes the evolution of parametric surfaces.,14383
is numerical element proposed evolution scheme a of for equations. efficient the An finite approximation,An efficient finite element scheme is proposed for a numerical approximation of the evolution equations.,14384
paper tracking sequences. problem recognition of addresses aerial image automated and the from vehicle This,This paper addresses the problem of automated vehicle tracking and recognition from aerial image sequences.,14385
normalization solution which for A working data includes steps is described. extraction and,A working solution which includes steps for data extraction and normalization is described.,14386
and the are discussed. possible proposed future Directions research of improvements for method,Directions for future research and possible improvements of the proposed method are discussed.,14387
an adaptive image proposed technique improvement The in reports PSNR. filtering,The proposed adaptive filtering technique reports an improvement in image PSNR.,14388
enormous rich a There of algorithms. proposing variety literature is an,There is an enormous literature proposing a rich variety of algorithms.,14389
such algorithms measures. overwhelming of majority The to produce fail,The overwhelming majority of algorithms fail to produce such measures.,14390
"Bayesian place an provide used marginal methods inclusion p-values. of probabilities alternative, with in","Bayesian methods provide an alternative, with marginal inclusion probabilities used in place of p-values.",14391
"advantages, has but impractical beyond variable selection small Bayesian problems. is computationally","Bayesian variable selection has advantages, but is impractical computationally beyond small problems.",14392
"images. is a regression to categorize Finally, scene utilized softmax","Finally, a softmax regression is utilized to categorize scene images.",14393
"the First, the network using factors. units are redefined of centered visible","First, the visible units of the network are redefined using centered factors.",14394
"softmax Finally, scene is recognition. a regression for incorporated","Finally, a softmax regression is incorporated for scene recognition.",14395
incompletely rules the connectome human the of that topology understood. generative shape The remain,The generative rules that shape the topology of the human connectome remain incompletely understood.,14396
made and visualizations accessible. have more practical They large-scale,They have made large-scale visualizations practical and more accessible.,14397
"positions Finally, lie the on centerline. precisely are of shape tubular to optimized these the","Finally, the positions of these are optimized to lie precisely on the tubular shape centerline.",14398
"the method. robustness Detailed of our accuracy versatility, and experiments new show","Detailed experiments show the versatility, accuracy and robustness of our new method.",14399
the Thereby tractography. for subsequent we as input FOD improve,Thereby we improve the FOD as input for subsequent tractography.,14400
Additionally rendered viewpoints. we show images from novel,Additionally we show images rendered from novel viewpoints.,14401
individuals. is in interacting populations evolution game theory powerful framework for of studying Evolutionary a,Evolutionary game theory is a powerful framework for studying evolution in populations of interacting individuals.,14402
"combining model energies, regularization propose new We common e.g. a segmentation","We propose a new segmentation model combining common regularization energies, e.g.",14403
"combining balanced motion, image depth, texture, clustering color, features of arbitrary high-dimensional enforcing etc. location,","enforcing balanced clustering of arbitrary high-dimensional image features combining color, texture, location, depth, motion, etc.",14404
an Precision example system. application as considered is the agriculture for,Precision agriculture is considered as an example application for the system.,14405
also surveillance are increasing. So for the need techniques,So the need for surveillance techniques are also increasing.,14406
the detect system in developed a motion has to been The video.,The system has been developed to detect the motion in a video.,14407
"video. from system, In motion the or real this the is time webcam from detected","In this system, motion is detected from the webcam or from the real time video.",14408
detect been method differencing used moving to target. Background frames the subtraction has and,Background subtraction and frames differencing method has been used to detect the moving target.,14409
"difference method continuous two Similarly, difference frames two frames. takes between","Similarly, two frames difference method takes difference between two continuous frames.",14410
and difference thresholded is difference pixels resultant is Then frame that amount of the calculated.,Then that resultant difference frame is thresholded and the amount of difference pixels is calculated.,14411
"low stochastic Furthermore, us with enables approximation the to cost. computational optimize method a problem","Furthermore, a stochastic approximation method enables us to optimize the problem with low computational cost.",14412
asymptotically is estimated by also parameter We and show distributed. consistent normally SLSTD the,We also show the parameter estimated by SLSTD is consistent and asymptotically normally distributed.,14413
has studied for decades. Face been Recognition many,Face Recognition has been studied for many decades.,14414
occluders object   significantly recognition accuracy. presence The of impacts,  The presence of occluders significantly impacts object recognition accuracy.,14415
when run on parallel try architectures. are we efficiently heterogeneous These challenges to compounded,These challenges are compounded when we try to run efficiently on heterogeneous parallel architectures.,14416
"that common, libraries relational Different implemented data model. geometric unified, domains as are use a","Different geometric domains are implemented as libraries that use a common, unified, relational data model.",14417
via explained features spectral The intuitive circulant are of matrices. use,The intuitive use of spectral features are explained via circulant matrices.,14418
number a The results labeling and with experiments are substantial computation. reported of recognition,The recognition results are reported with a substantial number of experiments and labeling computation.,14419
"compute algorithm, correspondences to We a between novel called images. dense matching DeepMatching, introduce","We introduce a novel matching algorithm, called DeepMatching, to compute dense correspondences between images.",14420
performance obtains public on DeepFlow benchmarks optical estimation. for competitive flow,DeepFlow obtains competitive performance on public benchmarks for optical flow estimation.,14421
"however, of is, implementation these The computationally challenging. ideas","The implementation of these ideas is, however, computationally challenging.",14422
significant Finding predictive important task. features statistically interaction high-order modeling but in challenging is,Finding statistically significant high-order interaction features in predictive modeling is important but challenging task.,14423
where PSI. sampling distributions authors exact the for framework algorithmic in presented an computing,where the authors presented an algorithmic framework for computing exact sampling distributions in PSI.,14424
we in general. Then under yield conditions give should which results method good our,Then we give conditions under which our method should yield good results in general.,14425
compare method methods. this other we our context with two In,In this context we compare our method with two other methods.,14426
inference is Bayesian spectral presented. approach A to,A spectral approach to Bayesian inference is presented.,14427
the emulation pursues probability of posterior the It density.,It pursues the emulation of the posterior probability density.,14428
of likelihood be quantities can From calculated statistical interest expansion semi-analytically. this all spectral,From this spectral likelihood expansion all statistical quantities of interest can be calculated semi-analytically.,14429
the and Both expansion to the moments are model evidence coefficients. related posterior the,Both the model evidence and the posterior moments are related to the expansion coefficients.,14430
enables to performance high The complex real-time queries. responses,The high performance enables real-time responses to complex queries.,14431
are signal (i.e. then of a as decomposed of The sum the signals projected interest,The projected signals are then decomposed as the sum of a signal of interest (i.e.,14432
"model. Gaussian linear very discriminant) noise, and mixed using simple background a","discriminant) and background noise, using a very simple Gaussian linear mixed model.",14433
"is simplified. parameter the problem simplicity model, to the estimation Thanks of the corresponding","Thanks to the simplicity of the model, the corresponding parameter estimation problem is simplified.",14434
of context. Classification relevance proposed prove approach the a in results such the,Classification results prove the relevance of the proposed approach in such a context.,14435
for image attracts years. in minimization processing low-rank attention much recent Patch-based,Patch-based low-rank minimization for image processing attracts much attention in recent years.,14436
"we denoising. low-rank propose image a patch-based idea, Based on minimization method this for","Based on this idea, we propose a patch-based low-rank minimization method for image denoising.",14437
"main The three and stated is process PCA, way: SVD equivalent denoising in low-rank minimization.","The main denoising process is stated in three equivalent way: PCA, SVD and low-rank minimization.",14438
this method are also improvements given. Further of,Further improvements of this method are also given.,14439
an has structure. typically underlying communication Human,Human communication typically has an underlying structure.,14440
accomplish this language joint both in generative a visual using cues We and model.,We accomplish this using both visual and language cues in a joint generative model.,14441
is chemical broad ways. than The more models two model in kinetics typical,The model is more broad than typical chemical kinetics models in two ways.,14442
"First, faithful. more of simplifies time-dependent process the models hazard using making rates","First, using time-dependent hazard rates simplifies the process of making models more faithful.",14443
"includes state environmental the the resources. of traits use and individual of Second, system","Second, the state of the system includes individual traits and use of environmental resources.",14444
where in of multi-modal is sampler illustrated are The encountered. proposed distributions efficiency examples the,The efficiency of the proposed sampler is illustrated in examples where multi-modal distributions are encountered.,14445
powerful hidden sequential of modeling tools are data. Factorial (FHMMs) Markov models,Factorial hidden Markov models (FHMMs) are powerful tools of modeling sequential data.,14446
"log-likelihood than the marginal approximation a of we better FAB First, inference. previous offer","First, we offer a better approximation of marginal log-likelihood than the previous FAB inference.",14447
No the the face between were two area. found groups in differences,No differences between the two groups were found in the face area.,14448
gray showed groups. Structural difference analysis in imaging matter no density between,Structural imaging analysis showed no difference in gray matter density between groups.,14449
signature: motion based This temporal on lines. for barcode a improvement new is,This improvement is based on a new temporal signature: motion barcode for lines.,14450
"statistical has systematic The little however. performance received of these sophisticated relatively models attention,","The statistical performance of these sophisticated models has received relatively little systematic attention, however.",14451
"cohesive do subsets related this form most display Nonetheless, genes (clusters). implies they phylogenies; topologically","Nonetheless, most genes do display topologically related phylogenies; this implies they form cohesive subsets (clusters).",14452
The take successive up may days. entire three process,The entire process may take up three successive days.,14453
"space the This constitutes neophobia, avoidance space. behavior of coping untrodden with","This behavior constitutes coping with space neophobia, the avoidance of untrodden space.",14454
often task Solving computer for real-time is optimization on-the-fly difficult applications. a large-scale graphics,Solving large-scale optimization on-the-fly is often a difficult task for real-time computer graphics applications.,14455
"technique. is reduction well-adopted model tackle To a this challenge,","To tackle this challenge, model reduction is a well-adopted technique.",14456
high-order problems. among account regression practical covariates into many interactions valuable in is Taking,Taking into account high-order interactions among covariates is valuable in many practical regression problems.,14457
bending. cyclization into gain insight DNA a DNA technique nature to is the powerful of,DNA cyclization is a powerful technique to gain insight into the nature of DNA bending.,14458
the FRET recent in j-factor also We experiments. measured estimate dynamic,We also estimate the dynamic j-factor measured in recent FRET experiments.,14459
of the described. regions inter-individual was The brain graphs within and variability discovered different,The inter-individual variability of the graphs within different brain regions was discovered and described.,14460
Experiments challenging conducted state-of-the-art. a dataset with on outperforming are results,Experiments are conducted on a challenging dataset with results outperforming state-of-the-art.,14461
size of processing. is a image any color for Firstly input,Firstly a color image of any size is input for processing.,14462
The it to converts a algorithm grayscale image.,The algorithm converts it to a grayscale image.,14463
apply foreground to detector the boundary edge canny of we find the object. Next,Next we apply canny edge detector to find the boundary of the foreground object.,14464
to is Video a video context. stepping segmentation stone understanding,Video segmentation is a stepping stone to understanding video context.,14465
human presence environments. is to automatically important detect times secure in It at,It is at times important to detect human presence automatically in secure environments.,14466
"robust, shape fast that has rates. a error needs This low and recognition algorithm is","This needs a shape recognition algorithm that is robust, fast and has low error rates.",14467
"the is error encouraged. sity is the and minimized reconstruction this coefficient objective, spar- In","In this objective, the reconstruction error is minimized and the coefficient spar- sity is encouraged.",14468
algorithm. an alternative an strategy by in This objective opti- is mized iterative,This objective is opti- mized by an alternative strategy in an iterative algorithm.,14469
Almost all predictions. computations making neural involve,Almost all neural computations involve making predictions.,14470
the demonstrate datasets. BMA number of clustering use model-based We for of in a,We demonstrate the use of BMA in model-based clustering for a number of datasets.,14471
can BMA the in context enhanced Applying modeling performance. fast give is and model-based,Applying BMA in the model-based context is fast and can give enhanced modeling performance.,14472
ubiquitous machine in learning. are methods tools Kernel,Kernel methods are ubiquitous tools in machine learning.,14473
that show large classification be We used regression scale for tasks. framework this and can,We show that this framework can be used for large scale regression and classification tasks.,14474
configurations. the can viewed latent prior FIC over a variable be distribution as,FIC can be viewed as a prior distribution over the latent variable configurations.,14475
"result, towards components. may models be As a with less FIC biased overly","As a result, FIC may be overly biased towards models with less components.",14476
There Satisfiability the for are Modulo already a tools quite problems. Theories (SMT) solving few,There are already quite a few tools for solving the Satisfiability Modulo Theories (SMT) problems.,14477
solution integer convex solving with extends and The tool volume counting computation/estimation SMT polytopes. for,The tool extends SMT solving with integer solution counting and volume computation/estimation for convex polytopes.,14478
"interpretations, Based al. the on structure SE et Delgrande of","Based on the structure of SE interpretations, Delgrande et al.",14479
well-known revision. logic adapted AGM the to (LP) framework program,adapted the well-known AGM framework to logic program (LP) revision.,14480
identified introduced They of specific behavior rational revision LP operators. some and the,They identified the rational behavior of LP revision and introduced some specific operators.,14481
channel an plays river role studies. in networks of Quantitative important analysis,Quantitative analysis of channel networks plays an important role in river studies.,14482
"experimental several fitness completely the last In landscapes have years, been resolved.","In the last years, several experimental fitness landscapes have been completely resolved.",14483
"etc). fraction roughness/slope, components, non-additive of of peaks, number","number of peaks, roughness/slope, fraction of non-additive components, etc).",14484
objects together. formulated aggregated The terrain similar type the by aggregated is ground,The aggregated terrain type is formulated by the similar ground objects aggregated together.,14485
is triplet-based criterion. learning under a formulated Pose embedding distance,Pose embedding learning is formulated under a triplet-based distance criterion.,14486
compares parametric with Bayesian system identification. recently This classical for developed paper methods methods,This paper compares classical parametric methods with recently developed Bayesian methods for system identification.,14487
"not posterior distribution so often requires methods. known computational and analytically, The is","The posterior distribution is not often known analytically, and so requires computational methods.",14488
supervision transfers propose In from work a technique images different that we modalities. this between,In this work we propose a technique that transfers supervision between images from different modalities.,14489
the immune model computational simulates a We of the development adaptive have system. built that,We have built a computational model that simulates the development of the adaptive immune system.,14490
show realistic. model simulations self-centered The our is that,The simulations show that our self-centered model is realistic.,14491
state of detail. analyzed special is The steady case in case,The special case of steady state case is analyzed in detail.,14492
"problems, a new energy which we parsimonious call We discrete minimization of labeling. family propose","We propose a new family of discrete energy minimization problems, which we call parsimonious labeling.",14493
"and energy Specifically, clique potentials. high-order functional unary potentials of consists our","Specifically, our energy functional consists of unary potentials and high-order clique potentials.",14494
"for small each of component. number a labelings, putative with mixture one us This provides","This provides us with a small number of putative labelings, one for each mixture component.",14495
"putative of energy we the in Third, the labeling best choose terms value.","Third, we choose the best putative labeling in terms of the energy value.",14496
process fault in Gaussian This approach a data. novel paper to proposes time-series removal,This paper proposes a novel Gaussian process approach to fault removal in time-series data.,14497
on prediction error This based approach minimization. identification the guarantees stable predictors the of,This approach guarantees the identification of stable predictors based on the prediction error minimization.,14498
this we compare issue. this paper various In address propose to techniques and,In this paper we propose and compare various techniques to address this issue.,14499
will provided. these be results Simulations comparing techniques,Simulations results comparing these techniques will be provided.,14500
Approximate expensive likelihood impossible. where Bayesian approximate computations inference models computation are performs for or,Approximate Bayesian computation performs approximate inference for models where likelihood computations are expensive or impossible.,14501
enormously results and feature images. classification extraction advances for improved Recent digital CBIR in have,Recent advances in feature extraction and classification have enormously improved CBIR results for digital images.,14502
This process. reduction speeds and the of retrieval feature leads dimensionality to up,This leads to reduction of feature dimensionality and speeds up the retrieval process.,14503
be SAT Such partitionings can parallel. corresponding in instances used for solving,Such partitionings can be used for solving corresponding SAT instances in parallel.,14504
special we associate finite in each partitioning a space. With search the point,With each partitioning we associate a point in the special finite search space.,14505
The obtained time solving estimations with well the by method. proposed agrees,The solving time agrees well with estimations obtained by the proposed method.,14506
two of processes independent percentiles The to random be variables. are the assumed monitored,The percentiles of the two monitored processes are assumed to be independent random variables.,14507
Then robustness in prior information. very poor tested the we chart the spite of of,Then we tested the robustness of the chart in spite of very poor prior information.,14508
the semantic inference module efficiency segmentation. explore beyond level CRF the of image We,We explore the efficiency of the CRF inference module beyond image level semantic segmentation.,14509
tasks. classification image surprisingly worked shown was various PCANet in well,PCANet was shown worked surprisingly well in various image classification tasks.,14510
each a segmentation of to pixel an is the image. task class-label assigning Semantic in,Semantic segmentation is the task of assigning a class-label to each pixel in an image.,14511
"class which these pixel-level. However, overlapping predictions regions at the creates are conflicting","However, these regions are overlapping which creates conflicting class predictions at the pixel-level.",14512
Discriminative recognition applied variable (LVM) various are to models latent tasks. frequently visual,Discriminative latent variable models (LVM) are frequently applied to various visual recognition tasks.,14513
"on variation class. (positive) the of latent Conventionally, foreground variables are the de- fined","Conventionally, latent variables are de- fined on the variation of the foreground (positive) class.",14514
function the a scoring formalize generalized LVM We such (GLVM). of,We formalize the scoring function of such a generalized LVM (GLVM).,14515
a function. for based discuss we framework on the Then learning model a GLVM scoring,Then we discuss a framework for learning a model based on the GLVM scoring function.,14516
could objections; many that has bear results. This postulate counter-intuitive it have met since observed,This postulate has since met objections; many have observed that it could bear counter-intuitive results.,14517
the in provide recovery a to paradox this We work. solution,We provide a solution to the recovery paradox in this work.,14518
possess the property. enriched belief not recovery latent theory that does We show the will,We will show that the enriched latent belief theory does not possess the recovery property.,14519
"however. maintained is closure the by in logical theory, consequences The","The closure by logical consequences is maintained in the theory, however.",14520
theories. to in solution a problem it belief Hence open as serves the the set,Hence it serves as a solution to the open problem in the belief set theories.,14521
not this. methods did Existing explicitly consider,Existing methods did not explicitly consider this.,14522
DA by codes search learned ANN tasks. Better be the can for,Better codes can be learned by DA for the ANN search tasks.,14523
dictionaries the substantially We experiments by show our reduce error. that quantization optimized overall,We show by experiments that our optimized dictionaries substantially reduce the overall quantization error.,14524
image Convolutional (CNNs) achieve for are powerful models results Neural impressive Networks that classification.,Convolutional Neural Networks (CNNs) are powerful models that achieve impressive results for image classification.,14525
optimal CNN than less (e.g. A that is,A CNN that is less than optimal (e.g.,14526
images. success recognition great object in networks achieved have convolutional for still Deep,Deep convolutional networks have achieved great success for object recognition in still images.,14527
result. probably there could explain reasons We are this that that argue two,We argue that there are two reasons that could probably explain this result.,14528
Two-stream deep are models ConvNets) compared with those very in shallow domain relatively (e.g. image,Two-stream ConvNets) are relatively shallow compared with those very deep models in image domain (e.g.,14529
"capacity is by their GoogLeNet), their therefore constrained depth. modeling VGGNet, and","VGGNet, GoogLeNet), and therefore their modeling capacity is constrained by their depth.",14530
"recognition is easy small. the quite is action of this size extension as However, not","However, this extension is not easy as the size of action recognition is quite small.",14531
drawn attention since of the recognition a has Iris century. mid-twentieth lot,Iris recognition has drawn a lot of attention since the mid-twentieth century.,14532
"possess of set features. biometric a is known features, all rich Among to iris","Among all biometric features, iris is known to possess a rich set of features.",14533
features used have recognition perform to Different the past. been in iris,Different features have been used to perform iris recognition in the past.,14534
test matching new used sample. Minimum distance perform is for each to classifier template,Minimum distance classifier is used to perform template matching for each new test sample.,14535
neutral at and At low pH. folds a in $T_m$ Ub two-state manner,At $T_m$ Ub folds in a two-state manner at low and neutral pH.,14536
main incorporate correcting The variables is idea an to SR dictionary formulation. in,The main idea is to incorporate dictionary correcting variables in an SR formulation.,14537
realization per-iteration for of and DANSER. complexity practical tailor-designed A low simple is algorithm,A simple and low per-iteration complexity algorithm is tailor-designed for practical realization of DANSER.,14538
"unclear however, far, simultaneously it how are both is achieved. Thus","Thus far, however, it is unclear how both are simultaneously achieved.",14539
the (e.g. fermentation products suggests This for mechanisms that recycling,This suggests that mechanisms for recycling the fermentation products (e.g.,14540
to survival use form the criterion ensemble trees. of RSF an split log-rank,RSF use the log-rank split criterion to form an ensemble of survival trees.,14541
Harrell's and C overcome node by using for discrepancy evaluation. can splitting be This both,This discrepancy can be overcome by using Harrell's C for both node splitting and evaluation.,14542
"splitting in is scenarios. preferable noisy log-rank Conversely,","Conversely, log-rank splitting is preferable in noisy scenarios.",14543
Both splitting ranger. log-rank the R~package and in are implemented C-based,Both C-based and log-rank splitting are implemented in the R~package ranger.,14544
structures rewiring. The controlled variation by in is network linkage,The variation in network structures is controlled by linkage rewiring.,14545
"density, feeding more that lower results loops and show link Our i.e. trophic","Our results show that more feeding loops and lower trophic link density, i.e.",14546
logic introduce We probabilistic of the a approximation programs. lazy to approach explanation-based,We introduce a lazy approach to the explanation-based approximation of probabilistic logic programs.,14547
It part for program of explanations. searching most the the significant when only uses,It uses only the most significant part of the program when searching for explanations.,14548
state-of-the-art method inference. that show approximate this We experimentally outperforms,We experimentally show that this method outperforms state-of-the-art approximate inference.,14549
fail These scenes. in deblur methods to videos dynamic blurry,These methods fail to deblur blurry videos in dynamic scenes.,14550
energy model. efficient solvers framework also optimize provide and We to a the,We also provide a framework and efficient solvers to optimize the energy model.,14551
a to This that ANN-based MVS show potential manual alternative results grading. is an,This results show that an ANN-based MVS is a potential alternative to manual grading.,14552
"recognition robots monitor is Activity interact in very where scenarios useful assist humans. or with,","Activity recognition is very useful in scenarios where robots interact with, monitor or assist humans.",14553
detect and procedure different recognize activities formulate an We optimization types. of then to,We then formulate an optimization procedure to detect and recognize activities of different types.,14554
also present We algorithm model. computationally for simple a optimizing our,We also present a computationally simple algorithm for optimizing our model.,14555
large. head variation pose the is usually,the head pose variation is usually large.,14556
In paper we make to understanding. this several texture contributions,In this paper we make several contributions to texture understanding.,14557
"Hence, transmission explanations reproductive are unnecessary. statistically cultural of success invoking","Hence, explanations invoking cultural transmission of reproductive success are statistically unnecessary.",14558
and SIR immunity model time consider rate. a transmission with dependent We temporary,We consider a SIR model with temporary immunity and time dependent transmission rate.,14559
infection. as which We confers vaccination the natural immunity same dependent time assume,We assume time dependent vaccination which confers the same immunity as natural infection.,14560
"but be not Radon measures. these strategies In will functions general,","In general, these strategies will not be functions but Radon measures.",14561
"longer than the Obviously, hybrid the current GDAs. runs","Obviously, the hybrid runs longer than the current GDAs.",14562
last for (BoF) decade. studied bag-of-features been The has model the over thoroughly classification image,The bag-of-features (BoF) model for image classification has been thoroughly studied over the last decade.,14563
problem non-convex problem parts. posed program The that includes a two as integer is,The problem is posed as a non-convex integer program problem that includes two parts.,14564
are the and matrix. Unsupervised considered for supervised ordinal structures,Unsupervised and supervised structures are considered for the ordinal matrix.,14565
Yet suffer pain limitations. current from computational models of transmission several,Yet current computational models of pain transmission suffer from several limitations.,14566
findings far. our so reports paper This,This paper reports our findings so far.,14567
"addition, the intuitively, reliability as the In increases. activity sensory stimulation decreases of frequency","In addition, intuitively, the reliability of sensory activity decreases as the stimulation frequency increases.",14568
"to a a is performed refined labeling Then, achieve segmentation. binary","Then, a binary labeling is performed to achieve a refined segmentation.",14569
"boxes. detected prevents This increases and annotation automatically algorithms utilizing from burden, the these bounding","This increases the annotation burden, and prevents these algorithms from utilizing automatically detected bounding boxes.",14570
solved iterated an is MRF max-flow algorithm. by model then This,This MRF model is then solved by an iterated max-flow algorithm.,14571
likelihood a method. We derive sampling and the expression simple an for efficient,We derive a simple expression for the likelihood and an efficient sampling method.,14572
Humans symbolic (i.e. been using have representation,Humans have been using symbolic representation (i.e.,14573
of graffiti aesthetic/nonaesthetic forms underscore Contemporary dichotomy. the,Contemporary forms of graffiti underscore the aesthetic/nonaesthetic dichotomy.,14574
more alignments. We similarities find the to to these use accurate relationships between propose,We propose to use the similarities between these relationships to find more accurate alignments.,14575
"effective identifying Furthermore, particularly concepts. complex correspondences our is between method in","Furthermore, our method is particularly effective in identifying correspondences between complex concepts.",14576
ontology method the Our previous domains. state-of-the-art F-score achieves substantially three better than matching on,Our method achieves substantially better F-score than the previous state-of-the-art on three ontology matching domains.,14577
can Identifying and properties CSI simplify exploiting inference.,Identifying and exploiting CSI properties can simplify inference.,14578
work present one case. a we such the In show,In the present work we show one such a case.,14579
makes assumption This applications. for many them inflexible,This assumption makes them inflexible for many applications.,14580
"be loading will of column-wise matrices a generation correlated. two As (indirectly) consequence, the","As a consequence, the generation of two loading matrices will be column-wise (indirectly) correlated.",14581
about our the Networks intuition relationships capture world. in,Networks capture our intuition about relationships in the world.,14582
"However, the useful MIP motivating approaches. this first-order in disappears process of first-order structure development","However, useful first-order structure disappears in this process motivating the development of first-order MIP approaches.",14583
"we one mfoilp, approach. such present Here","Here we present mfoilp, one such approach.",14584
and its regression depend models the Linear matrix on directly design properties.,Linear regression models depend directly on the design matrix and its properties.,14585
dataset. show difficult substantive We performance a improvement thermal-visible on face,We show substantive performance improvement on a difficult thermal-visible face dataset.,14586
that (e.g. will key closed a robots in is observation many environment relatively stay Our,Our key observation is that many robots will stay in a relatively closed environment (e.g.,14587
will available be and upon Code data acceptance.,Code and data will be available upon acceptance.,14588
"refer to as ""knowledge this (KT). translation"" task We","We refer to this task as ""knowledge translation"" (KT).",14589
The target conformation. the depends on affinity,The target affinity depends on the conformation.,14590
various space and the compare explore scenarios. systematically search parameter We,We systematically explore the parameter space and compare various search scenarios.,14591
a that necessary DNA MFPT. for conformation absolutely a We find coiled fast is,We find that a coiled DNA conformation is absolutely necessary for a fast MFPT.,14592
"such can be many Moreover, interesting e.g. intra-class applications, used in knowledge","Moreover, such intra-class knowledge can be used in many interesting applications, e.g.",14593
image retrieval completion. object and style-based style-based,style-based image retrieval and style-based object completion.,14594
several different sensory  spatial  information makes navigation of channels. use The mammalian system,  The mammalian spatial navigation system makes use of several different sensory information channels.,14595
is arise path analysis dynamic similar integration from noise errors in fluctuations. A performed when,A similar analysis is performed when errors in path integration arise from dynamic noise fluctuations.,14596
"forces them, their Cells stiffness can environment. to but of sense applied the also","Cells can sense forces applied to them, but also the stiffness of their environment.",14597
of TGFb large of sensor. uses complex the as latent model a Our example,Our model uses the example of large latent complex of TGFb as a sensor.,14598
"We dynamic behaviours investigate of comparing the steady-state model, the with them experiments. and","We investigate the dynamic and steady-state behaviours of the model, comparing them with experiments.",14599
optimize structure programs Current implementations of the order performance. in simplify to overall the,Current implementations simplify the structure of programs in order to optimize the overall performance.,14600
"particular, as aggregates. aggregates rewritten monotone In known are simpler forms into","In particular, aggregates are rewritten into simpler forms known as monotone aggregates.",14601
"function is of this no kind known. However,","However, no function of this kind is known.",14602
for semantics Some aggregates. similarities surprising results mainstream differences versus model highlight stable and,Some surprising results highlight similarities and differences versus mainstream stable model semantics for aggregates.,14603
"and the ""in scenarios. face regions) wild"" the with","with face regions) and the ""in the wild"" scenarios.",14604
performance hallmark cognitive is data processes major Inferring or of constructs psychometrics. from a,Inferring processes or constructs from performance data is a major hallmark of cognitive psychometrics.,14605
"if bodies or restricted \Luka are conjunctions. %, \Godel even to","%, even if bodies are restricted to \Luka or \Godel conjunctions.",14606
a these implementing also paper The prototype system translations. presents and evaluates,The paper also presents and evaluates a prototype system implementing these translations.,14607
tabling. planner is Picat's use implemented of by,Picat's planner is implemented by use of tabling.,14608
"domain the knowledge. is resulting shaped policy using Finally,","Finally, the resulting policy is shaped using domain knowledge.",14609
"(p,q)-Bezier study Bernstein the for degree degree and of reduction nature functions. We elevation","We study the nature of degree elevation and degree reduction for (p,q)-Bezier Bernstein functions.",14610
to The similar some new have q-Bezier curves curves. properties,The new curves have some properties similar to q-Bezier curves.,14611
"(p,q)-Bernstein for some discussed. are properties curves Furthermore, Bezier fundamental","Furthermore, some fundamental properties for (p,q)-Bernstein Bezier curves are discussed.",14612
real Experiments the ECMdd. are synthetic to sets data illustrate used of and performance in,Experiments in synthetic and real data sets are used to illustrate the performance of ECMdd.,14613
internal well in show data the results capture ECMdd structure. the The could uncertainty that,The results show that ECMdd could capture well the uncertainty in the internal data structure.,14614
"compared robust FCMdd. Moreover, is to more initializations it the with","Moreover, it is more robust to the initializations compared with FCMdd.",14615
applying networks. variational We Bayesian inference general stochastic a highlight when to pitfall,We highlight a pitfall when applying stochastic variational inference to general Bayesian networks.,14616
the elements. properties among interactions of ecosystem characterize impact Collectively indirect these,Collectively these properties characterize the impact of indirect interactions among ecosystem elements.,14617
"addition, expresses author formats, in i.e. the standard and In visual solutions his","In addition, the author expresses his solutions in standard and visual formats, i.e.",14618
probabilities terms of natural frequencies. and in,in terms of probabilities and natural frequencies.,14619
"First, gender introduces classification paper challenge and the for application research. this","First, this paper introduces the challenge and application for gender classification research.",14620
"described. development classification Then, and framework of the are gender","Then, the development and framework of gender classification are described.",14621
"the each discuss limitation mean strength In of highlight we and time, the method.","In mean time, we highlight the strength and discuss the limitation of each method.",14622
"work. discusses Finally, for the several applications also review future this promising","Finally, this review also discusses several promising applications for the future work.",14623
has research in graphics. of the It been one topics central computer,It has been one of the central research topics in computer graphics.,14624
"there single and both advantages one method perfect However, each and no has is disadvantages.","However, there is no single perfect method and each one has both advantages and disadvantages.",14625
available C++ implementation to in make use various A our ready framework is applications. to,A C++ implementation is available to make our framework ready to use in various applications.,14626
"using an matrix we setup. classify criteria two in $X$ any )$, random iterative",")$, we classify any random matrix $X$ using two criteria in an iterative setup.",14627
"speckle the benefits reduction In particular, by of inherits method methods. the diffusion-like","In particular, the method inherits the benefits of speckle reduction by diffusion-like methods.",14628
synthetic method. Results on performance PolSAR show the the real of and data,Results on synthetic and real PolSAR data show the performance of the method.,14629
motivated the assumption this from ideas causality. in of field can how We be show,We show how this assumption can be motivated from ideas in the field of causality.,14630
"for this guarantees provide method. we However, no","However, we provide no guarantees for this method.",14631
synthetic sets gene and on deletion We data results data set. a present,We present results on synthetic data sets and a gene deletion data set.,14632
"introduce framework frames for shift-invariant a signals. multi-scale, designing bi-frames adaptive, We for representing and","We introduce a framework for designing multi-scale, adaptive, shift-invariant frames and bi-frames for representing signals.",14633
in It frames improves wavelet multi-scale efficiency. coding basis classical such of as terms,It improves classical multi-scale basis such as wavelet frames in terms of coding efficiency.,14634
discussed. Connections also with deep networks convolutional are,Connections with deep convolutional networks are also discussed.,14635
"this for In representation. network learning face we a propose deep a convolution paper, robust","In this paper, we propose a deep convolution network for learning a robust face representation.",14636
dataset. LFW CASIA-WebFace is on trained dataset model The and on evaluated,The model is trained on CASIA-WebFace dataset and evaluated on LFW dataset.,14637
library. a In recognition presented this we have hand gesture paper,In this paper we have presented a hand gesture recognition library.,14638
"Various pointing etc. detecting cluster direction, count, orientation, finger functions cluster include","Various functions include detecting cluster count, cluster orientation, finger pointing direction, etc.",14639
"and activity. with human the is on model pattern automatic based distributed brain, The","The model is based on the human brain, with automatic and distributed pattern activity.",14640
suggested. carrying different for the Methods are processes out,Methods for carrying out the different processes are suggested.,14641
"gaze eye theory, a can provide tracking vision-based of good the estimate of In location.","In theory, vision-based tracking of the eye can provide a good estimate of gaze location.",14642
"We present a and reinforcement third one. algorithms, learning two devise","We present two reinforcement learning algorithms, and devise a third one.",14643
introduce and it variational Dirichlet apply allocation incremental inference We latent to (LDA).,We introduce incremental variational inference and apply it to latent Dirichlet allocation (LDA).,14644
sets. of LDA We benchmark performance the on data large incremental study,We study the performance of incremental LDA on large benchmark data sets.,14645
problems analysis data. data involve modern from inferences Many streaming,Many modern data analysis problems involve inferences from streaming data.,14646
"parameter high in space, inefficient problems. however, the Exploring this dimensional makes approach whole","Exploring the whole parameter space, however, makes this approach inefficient in high dimensional problems.",14647
"to inference First, in systems. of the we queuing an application present context","First, we present an application to inference in the context of queuing systems.",14648
tensors the factoring presence of in low-rank consider We slabs. outlying,We consider factoring low-rank tensors in the presence of outlying slabs.,14649
the sets on data Experiments proposed show other several methods. outperforms model that image state-of-the-art,Experiments on several image data sets show that the proposed model outperforms other state-of-the-art methods.,14650
in central Phylogenetic tool trees understanding evolution. a are,Phylogenetic trees are a central tool in understanding evolution.,14651
"sequence through from evolutionary inferred typically relationships and They are capture data, time.","They are typically inferred from sequence data, and capture evolutionary relationships through time.",14652
able sources (e.g. to It different trees to from essential data is be compare,It is essential to be able to compare trees from different data sources (e.g.,14653
genes and organisms) different same the inference several from methods.,several genes from the same organisms) and different inference methods.,14654
"a of for trees. labeled rooted, quantitative We comparison new metric robust, propose","We propose a new metric for robust, quantitative comparison of rooted, labeled trees.",14655
it summary This makes well-supported trees. select to possible,This makes it possible to select well-supported summary trees.,14656
approach Dengue phylogenies. on demonstrate fever our We,We demonstrate our approach on Dengue fever phylogenies.,14657
"train these expectation to As alternative, how here using propagation. efficiently we describe classifiers an","As an alternative, we describe here how to train these classifiers efficiently using expectation propagation.",14658
proposed instances. handling The data for with allows millions of datasets method,The proposed method allows for handling datasets with millions of data instances.,14659
that the indicate the competitive is experiments approach. Several with method described variational,Several experiments indicate that the method described is competitive with the variational approach.,14660
exhaustive on with several infeasible large dataset reasons. Working is search for,Working with exhaustive search on large dataset is infeasible for several reasons.,14661
of conduct first existing terms study degradations. of unified image-level in various the methods',conduct the first unified study of existing methods' in terms of various image-level degradations.,14662
"vital ""what hand, question other the the characteristics methods' performance?"" On object-level existing affect really","On the other hand, the vital question ""what object-level characteristics really affect existing methods' performance?""",14663
also localization various object analyse PASCAL classes. for accuracy methods' existing We latency and VOC,We also analyse existing methods' localization accuracy and latency for various PASCAL VOC object classes.,14664
paper method This describes to characters. handwritten offline recognize the,This paper describes the method to recognize offline handwritten characters.,14665
absence makes absent) task supervision and boxes challenging. bounding are The (labels the of,The absence of supervision (labels and bounding boxes are absent) makes the task challenging.,14666
analysis well. match the Numerical theoretical simulations,Numerical simulations match the theoretical analysis well.,14667
terms is poses This possible of adaptation. constraints in what on,This poses constraints on what is possible in terms of adaptation.,14668
appearance. representation of linear The is a into then incorporated metric learned,The learned metric is then incorporated into a linear representation of appearance.,14669
for presents two constraint algorithms performing satisfaction. new This work,This work presents two new algorithms for performing constraint satisfaction.,14670
operation. to anytime support is RDMaxWalkSat shown,RDMaxWalkSat is shown to support anytime operation.,14671
is structure for the understanding function. the Understanding of these activities oscillatory essential brain,Understanding the structure of these oscillatory activities is essential for understanding the brain function.,14672
oscillations results that Our chaotic are suggest by produced systems. deterministic neural,Our results suggest that neural oscillations are produced by deterministic chaotic systems.,14673
been recognition in topic research sequence Image-based long-standing a has computer vision.,Image-based sequence recognition has been a long-standing research topic in computer vision.,14674
the are over Significant baseline improvements classification implementation. shown,Significant classification improvements are shown over the baseline implementation.,14675
(CNNs) recent performance neural remarkable demonstrated history. nets in Convolutional have,Convolutional neural nets (CNNs) have demonstrated remarkable performance in recent history.,14676
tend Such to feed-forward bottom-up work unidirectional a in fashion. approaches,Such approaches tend to work in a unidirectional bottom-up feed-forward fashion.,14677
trained GPU-optimized to gradient allows This RGs with be descent.,This allows RGs to be trained with GPU-optimized gradient descent.,14678
"From CNNs RGs help probabilistic theoretical models. between a establish a perspective, and connection hierarchical","From a theoretical perspective, RGs help establish a connection between CNNs and hierarchical probabilistic models.",14679
results demonstrate state-of-the-art challenging benchmarks. on We,We demonstrate state-of-the-art results on challenging benchmarks.,14680
The phenomena origins of physical these discussed. are,The physical origins of these phenomena are discussed.,14681
"methods keypoint outliers, yield for Instead, with to state-of-the-art inference localization. robustness","Instead, inference methods with robustness to outliers, yield state-of-the-art for keypoint localization.",14682
"much performance. have can that recent models smaller works of shown levels achieve However, similar","However, recent works have shown that much smaller models can achieve similar levels of performance.",14683
"of NN the pruning address work, this a model. problem in parameters we trained In","In this work, we address the problem of pruning parameters in a trained NN model.",14684
"show similar them. to We way systematic redundant, propose are how remove a neurons and","We show how similar neurons are redundant, and propose a systematic way to remove them.",14685
"phase exposure contrast to TOF our camera cameras based on In profiles. modulation, enables general","In contrast to TOF cameras based on phase modulation, our camera enables general exposure profiles.",14686
approaches. flexibility in This requires added results computational novel and,This results in added flexibility and requires novel computational approaches.,14687
are present we state-of-the-art imaging. depth results These in,These results we present are state-of-the-art in depth imaging.,14688
The of flexibility generative our approach to easily us model. our allows enrich,The flexibility of our approach allows us to easily enrich our generative model.,14689
in integrated model no seamlessly new cost. is the additional The at computational system,The new model is seamlessly integrated in the system at no additional computational cost.,14690
question pulsed also optimal Our in addresses work of design exposure the TOF systems. important,Our work also addresses the important question of optimal exposure design in pulsed TOF systems.,14691
"features of is detectors, design considered. combining complexities, very complexity-aware different of cascaded The","The design of complexity-aware cascaded detectors, combining features of very different complexities, is considered.",14692
in vastly use This a detector. of features complexities enables single of the different,This enables the use of features of vastly different complexities in a single detector.,14693
"insurance, risk monitoring, environmental management anomaly and portfolio detection.","portfolio monitoring, insurance, environmental risk management and anomaly detection.",14694
towards up reduction technique existing This multivariate paves methods. way EVT the dimension scaling,This dimension reduction technique paves the way towards scaling up existing multivariate EVT methods.,14695
evidence results of the relevance approach. strong provide empirical of experimental Illustrative our,Illustrative experimental results provide strong empirical evidence of the relevance of our approach.,14696
"multiple, labels classes. Modeling dense relations within benefits temporal across from and","Modeling multiple, dense labels benefits from temporal relations within and across classes.",14697
"artificial Neural approach (DSNS). intelligence Synaptic introduce the 'Digital Substrate' new a We (AI) called,","We introduce a new artificial intelligence (AI) approach called, the 'Digital Synaptic Neural Substrate' (DSNS).",14698
objects attributes domains uses various (e.g. It from in selected,It uses selected attributes from objects in various domains (e.g.,14699
domain problem tested The was composition. of the chess in approach,The approach was tested in the domain of chess problem composition.,14700
also assessed experts were in domain. by the human They,They were also assessed by human experts in the domain.,14701
"a In source particular, chess quality (i.e. low","In particular, a low quality chess source (i.e.",14702
"compositions efficiently human published well. and as more experts), by","published compositions by human experts), and more efficiently as well.",14703
improvements cost and proposed. the enhance numerical computational bound to are stability Several,Several improvements to enhance numerical stability and bound the computational cost are proposed.,14704
"its each provided. the pseudo-code and corresponding are code analyzed Scilab For algorithm, in","For each analyzed algorithm, the pseudo-code and its corresponding code in Scilab are provided.",14705
special on symbolic Analysis data based Data is of - (SO). descriptions Symbolic objects,Symbolic Data Analysis is based on special descriptions of data - symbolic objects (SO).,14706
on clustering leaders based this method is result. The corresponding,The corresponding leaders clustering method is based on this result.,14707
solving methods compatible are same they Both are clustering optimization problem. the -,Both methods are compatible - they are solving the same clustering optimization problem.,14708
proposed applied methods different on data were The sets.,The proposed methods were applied on different data sets.,14709
"data presented. In paper, some clustering are results of of ESS the","In the paper, some results of clustering of ESS data are presented.",14710
analyze shapes Mathematical theory. using to and morphology describe set (MM) helps,Mathematical morphology (MM) helps to describe and analyze shapes using set theory.,14711
which binary treated effectively applied can MM are to sets. as be images,MM can be effectively applied to binary images which are treated as sets.,14712
used an can image defined as effective be in processing. morphological Basic tool operators,Basic morphological operators defined can be used as an effective tool in image processing.,14713
also and Morphological hypergraph. are operators on based developed graph,Morphological operators are also developed based on graph and hypergraph.,14714
have performance better and in applications processing. These image found operators,These operators have found better performance and applications in image processing.,14715
operation defined. hypergraph and morphological is erosion/dilation considered basic A is structure,A hypergraph structure is considered and basic morphological operation erosion/dilation is defined.,14716
operators the on opening/closing filtering also hypergraphs. Several and defined are new,Several new operators opening/closing and filtering are also defined on the hypergraphs.,14717
filters comparatively have performance found with on morphological filtering better based graph. Hypergraph based,Hypergraph based filtering have found comparatively better performance with morphological filters based on graph.,14718
paper binary images. hypergraph this ASF the In based effectiveness of we on evaluate,In this paper we evaluate the effectiveness of hypergraph based ASF on binary images.,14719
based ASF graph shows filters hypergraph ASF. have based Experimental that results outperformed,Experimental results shows that hypergraph based ASF filters have outperformed graph based ASF.,14720
"Knowing genome are model attempting to researchers this, currently structures. many accurately","Knowing this, many researchers are currently attempting to accurately model genome structures.",14721
"loci, and genome, scales fiber, chromosome, nucleosome, six full are nucleotide. These","These six scales are full genome, chromosome, loci, fiber, nucleosome, and nucleotide.",14722
already accuracy. Current in substantial achieved breakthroughs have prevailing approaches recognition,Current prevailing approaches have already achieved substantial breakthroughs in recognition accuracy.,14723
"if their usually dramatically performance are face samples misaligned. severely However, drops","However, their performance usually drops dramatically if face samples are severely misaligned.",14724
"of structure accelerate make we Moreover, use to derived the the solution. analytical block","Moreover, we make use of the block structure to accelerate the derived analytical solution.",14725
"extracellular properties However, is exact of the still uncertain. nature these","However, the exact nature of these extracellular properties is still uncertain.",14726
Probability discerning at are models statistical useless cause. and,Probability and statistical models are useless at discerning cause.,14727
"factor, ever hypothesis should again. or test, Bayes be No used","No hypothesis test, or Bayes factor, should ever be used again.",14728
not This over-certainty are when predictive worse used. methods made much and parametetric is,This over-certainty is made much worse when parametetric and not predictive methods are used.,14729
"industry in research and Big technology, has great society. science, data attention attracted","Big data research has attracted great attention in science, technology, industry and society.",14730
explores the data? answers is This paper What and following big questions:,This paper explores and answers the following questions: What is big data?,14731
"big are What representing, data? basic the and managing for analyzing methods","What are the basic methods for representing, managing and analyzing big data?",14732
big and between knowledge? is data relationship the What,What is the relationship between big data and knowledge?,14733
into we space? big from data knowledge find Can mapping a,Can we find a mapping from big data into knowledge space?,14734
is paradigm? the big science relationship data What and between,What is the relationship between big data and science paradigm?,14735
What of challenge the fundamental is computing? nature big data and,What is the nature and fundamental challenge of big data computing?,14736
of methodology multi-dimensional perspective data computing. is toward A big presented a,A multi-dimensional perspective is presented toward a methodology of big data computing.,14737
"however, have approaches almost empirically far, these investigated. all So only been","So far, however, almost all these approaches have only been empirically investigated.",14738
"requirements. always addition, was In motivation their on computational based","In addition, their motivation was always based on computational requirements.",14739
"For oracle local a inequality. this SVM, general derive we","For this local SVM, we derive a general oracle inequality.",14740
of evolution in Levy bacterial flights the space of model DNA. mutations time,Levy flights in the space of mutations model time evolution of bacterial DNA.,14741
conditions and This as covers background. dataset cluttered resolution hard imagery such low,This dataset covers hard conditions such as low resolution imagery and cluttered background.,14742
"are Convolutional Neural with marks. The Network results achieved approach, a similar rather compared achieving","The achieved results are compared with a Convolutional Neural Network approach, achieving rather similar marks.",14743
matcher. similarities combining by deep and are features k re-ranked from candidates COTS the The,The k candidates are re-ranked by combining similarities from deep features and the COTS matcher.,14744
"between this novel connection we two a discuss paper, In approaches. these","In this paper, we discuss a novel connection between these two approaches.",14745
and respond manage dynamic variation. policies to need to that systems to Ecological are them,Ecological systems are dynamic and policies to manage them need to respond to that variation.,14746
in of time) could fishing through Costs (here different policy changes forms. quotas adjustment take,Costs of policy adjustment (here changes in fishing quotas through time) could take different forms.,14747
for different costs show optimal of policies. have contrasting implications We forms policy how,We show how different forms of policy costs have contrasting implications for optimal policies.,14748
and of the PSS shown represent LHS to are SS the spectrum. True extremes,True SS and LHS are shown to represent the extremes of the PSS spectrum.,14749
derived along of with estimates The properties. asymptotic PSS is variance some,The variance of PSS estimates is derived along with some asymptotic properties.,14750
with and are designs discussed. use the Challenges PSS of their associated limitations,Challenges associated with the use of PSS designs and their limitations are discussed.,14751
numerical the the high-dimensional limitations and strengths examples Several method. highlight of,Several high-dimensional numerical examples highlight the strengths and limitations of the method.,14752
"approach we Secondly, theoretical results provide some to GP the our relating paradigm. standard","Secondly, we provide some theoretical results relating our approach to the standard GP paradigm.",14753
superimposed the markers images for system of The video pedestrians tracking. outputs also with,The system also outputs video images of the pedestrians with superimposed markers for tracking.,14754
is The the in complex object image a problem recognition processing.,The object recognition is a complex problem in the image processing.,14755
thinning using applications operators its paper hypergraph algorithms. briefly for and morphological This describes,This paper briefly describes morphological operators using hypergraph and its applications for thinning algorithms.,14756
an descriptors feed classifier. These to used are then SVM,These descriptors are then used to feed an SVM classifier.,14757
"stops that drift. summary, In growth show exponential we genetic","In summary, we show that exponential growth stops genetic drift.",14758
theoretical evolution with contrasts (e.g. non-selective of previous analyses result This,This result contrasts with previous theoretical analyses of non-selective evolution (e.g.,14759
"which populations how (fixate). genetic drift), over traits take spread investigated and eventually","genetic drift), which investigated how traits spread and eventually take over populations (fixate).",14760
"periodically particular shrinking models is expanding importance Consequently, populations. and it of of in life-cycles","Consequently, it is of particular importance in life-cycles models of periodically shrinking and expanding populations.",14761
The innovation. evolutionary an genes birth important new of motor of is,The birth of new genes is an important motor of evolutionary innovation.,14762
"selection that genes are indicating purifying they functional. show probably These signatures, significant","These genes show significant purifying selection signatures, indicating that they are probably functional.",14763
the is in pose automatic bottlenecks one face problem recognition. of The,The pose problem is one of the bottlenecks in automatic face recognition.,14764
approach experimental considerably recognition our could improve The the results that show performance.,The experimental results show that our approach could considerably improve the recognition performance.,14765
models in programming (PP) specification statistical Bayesian of code. flexible allows Probabilistic,Probabilistic programming (PP) allows flexible specification of Bayesian statistical models in code.,14766
We function targets performance. localization propose also both an that and loss efficient segmentation,We also propose an efficient loss function that targets both localization and segmentation performance.,14767
those the are related functions. The brain to of oscillations frequency,The frequency of those oscillations are related to the brain functions.,14768
life. key human in symbols) play a Words (phrases role or,Words (phrases or symbols) play a key role in human life.,14769
and understanding. problem fundamental symbol) or is (phrase for representation the representation knowledge Word,Word (phrase or symbol) representation is the fundamental problem for knowledge representation and understanding.,14770
usually A symbol) a represents a word of name (phrase or category.,A word (phrase or symbol) usually represents a name of a category.,14771
"is category complexity, proposed. to of In economy reduce the representation principle order representation","In order to reduce representation complexity, the economy principle of category representation is proposed.",14772
is of then A this relevant mesh way built cloud by from tetrahedralization.,A relevant mesh is then built from this cloud by way of tetrahedralization.,14773
computations. also The finite method element prove applications method other as can useful such for,The method can also prove useful for other applications such as finite element method computations.,14774
may influence distribution affect ecosystem Factors settlement livestock that degradation.,Factors that influence livestock settlement distribution may affect ecosystem degradation.,14775
availability settlements and livestock influence surface We hypothesize distribution. groundwater that,We hypothesize that surface and groundwater availability influence livestock settlements distribution.,14776
performance of the to are conducted method. proposed Extensive superior the show experiments,Extensive experiments are conducted to show the superior performance of the proposed method.,14777
"recollected complex, on timescales. Memories operating retained, and processes are stored, multiple coupled through","Memories are stored, retained, and recollected through complex, coupled processes operating on multiple timescales.",14778
"bidirectional. variables and interactions are fast Importantly, between the slow","Importantly, the interactions between fast and slow variables are bidirectional.",14779
are biological fingerprints dynamical Structural in unclear. networks evolutionary optimization and still of,Structural and dynamical fingerprints of evolutionary optimization in biological networks are still unclear.,14780
with Mixture three Applicability unsupervised real-world scenarios employing classification is Gaussian showcased Models.,Applicability is showcased with three real-world unsupervised classification scenarios employing Gaussian Mixture Models.,14781
The neural-symbolic algorithm with exploiting contrasted latter is side-information. a,The latter is contrasted with a neural-symbolic algorithm exploiting side-information.,14782
"analysis. has performed surveillance scene on reason, understanding For been and this research extensive","For this reason, extensive research has been performed on surveillance scene understanding and analysis.",14783
"groups scenes. have most considered single of adjacent or studies scenes, However,","However, most studies have considered single scenes, or groups of adjacent scenes.",14784
near-infrared spectroscopy pain investigated have based cerebral this hemodynamics (NIRS). In on we paper via,In this paper we have investigated pain based on cerebral hemodynamics via near-infrared spectroscopy (NIRS).,14785
"to segmentation. learning machine model-based ventricle applied successfully (LV) left been Recently, has","Recently, machine learning has been successfully applied to model-based left ventricle (LV) segmentation.",14786
techniques. Both supervised learning driven by are,Both are driven by supervised learning techniques.,14787
"and moving particular irregularly These, valid in not for objects. always, be however, may fast","These, however, may not be valid always, in particular for fast and irregularly moving objects.",14788
"meaning The this codon phenomenon, usage controversial. as is known biological of bias, still","The biological meaning of this phenomenon, known as codon usage bias, is still controversial.",14789
difficult. task The the makes human actions of complex recognition action of nature articulated and,The articulated and complex nature of human actions makes the task of action recognition difficult.,14790
models. either representations mainly Existing approaches body-part on articulated bag-of-feature or rely,Existing approaches mainly rely on either bag-of-feature representations or articulated body-part models.,14791
"the and relationship between the unexplored. substantially is the image segments action However, still","However, the relationship between the action and the image segments is still substantially unexplored.",14792
"To this a we VLAD. version end, of kernelized introduce","To this end, we introduce a kernelized version of VLAD.",14793
"adopted, GoogLeNet, Two three and learning recently proposed different CaffeNet are modalities. architectures, with","Two recently proposed architectures, CaffeNet and GoogLeNet, are adopted, with three different learning modalities.",14794
strongest as the unisensory enhancement expressed of Crossmodal response. commonly a is proportion,Crossmodal enhancement is commonly expressed as a proportion of the strongest unisensory response.,14795
summation the defined under unisensory probability combining responses is dependence. maximal Optimality by stochastic negative,Optimality is defined by probability summation combining the unisensory responses under maximal negative stochastic dependence.,14796
from set neurons. is single collected SC exemplified data This a with,This is exemplified with a data set collected from single SC neurons.,14797
index require any specific does new easy-to-compute distributional assumption. The not,The new easy-to-compute index does not require any specific distributional assumption.,14798
"potential has role been studied coding, in its and for decision-making. Metastability sensory memory","Metastability has been studied for its potential role in sensory coding, memory and decision-making.",14799
"Yet, about mechanisms genesis. little responsible for the is network very its known","Yet, very little is known about the network mechanisms responsible for its genesis.",14800
the in were data. predictions confirmed Both,Both predictions were confirmed in the data.,14801
We features. between the comparison fair CNN popular a provide also,We also provide a fair comparison between the popular CNN features.,14802
counts by from sources. The multiple method fusing information estimates,The method estimates counts by fusing information from multiple sources.,14803
"one At rely scale, for this on of estimation. cannot set we only features count","At this scale, we cannot rely on only one set of features for count estimation.",14804
using video Haar to detect we Firstly faces automatically frames. classifier dynamic human in,Firstly we using Haar classifier to detect human faces automatically in dynamic video frames.,14805
"Finally, achieve recognition. set we moment face to automatic the Hu using feature","Finally, we using Hu moment feature set to achieve the automatic face recognition.",14806
learning labeling is experiments prediction Our such both representation for in crucial and show space-time.,Our experiments show such representation learning is crucial for both labeling and prediction in space-time.,14807
of statisticians. more is result practice data for skills undergraduate The,The result is more practice of data skills for undergraduate statisticians.,14808
results pupil-size The the groups. difference normal population three a suggest in in,The results suggest a population difference in normal pupil-size in the three groups.,14809
We predictions obtain scenario. in to how this optimal study,We study how to obtain optimal predictions in this scenario.,14810
"provide contrast This which only in delayed tasks, is behavioral rewards. most time-sparse, with","This is in contrast with most behavioral tasks, which only provide time-sparse, delayed rewards.",14811
called A presented. minimalistic cognitive is architecture MANIC,A minimalistic cognitive architecture called MANIC is presented.,14812
"models, requires function machine. only three approximating MANIC one and state architecture The","The MANIC architecture requires only three function approximating models, and one state machine.",14813
"intelligence. of offers a approach objective artificial perspective, long-standing this an From alternate MANIC to","From this perspective, MANIC offers an alternate approach to a long-standing objective of artificial intelligence.",14814
architecture. analysis a of MANIC provides theoretical paper This the,This paper provides a theoretical analysis of the MANIC architecture.,14815
tasks content understanding. Image and two classification multimedia fundamental towards retrieval and representation are,Image representation and classification are two fundamental tasks towards multimedia content retrieval and understanding.,14816
shape (e.g. idea and texture that The information,The idea that shape and texture information (e.g.,14817
proposed (e.g. computing have A features gradients low-level number by of local been,A number of low-level features have been proposed by computing local gradients (e.g.,14818
"on have applications. achieved SIFT, HOG), LBP and successes great numerous multimedia and","SIFT, LBP and HOG), and have achieved great successes on numerous multimedia applications.",14819
of by color the contrast regions. distributions \emph{f}-divergence is two The between color computed measuring,The color contrast is computed by measuring \emph{f}-divergence between the color distributions of two regions.,14820
color descriptor enriches with and both image local representation information. contrast Our,Our descriptor enriches local image representation with both color and contrast information.,14821
can We that for (e.g. experimentally the strongly it descriptor compensate verified shape based,We verified experimentally that it can compensate strongly for the shape based descriptor (e.g.,14822
LCCD available. for will be the Codes,Codes for the LCCD will be available.,14823
of paper twofold. this scope The is,The scope of this paper is twofold.,14824
"properties set by a justifying and it of expands the one. new adding Firstly,","Firstly, it expands the set of properties by adding and justifying a new one.",14825
of adjusted version is so them. it An that proposed one fulfills index,An adjusted version of one index is proposed so that it fulfills them.,14826
"techniques, also two We rotational convolution introduce i.e.","We also introduce two rotational convolution techniques, i.e.",14827
encode adding without the extra parameters. two network invariance rotation into techniques These,These two techniques encode rotation invariance into the network without adding extra parameters.,14828
"for more RPC CNNs make and transformations. robust Additionally, FRPC rotation","Additionally, RPC and FRPC make CNNs more robust for rotation transformations.",14829
Gaussian image-set each model implicitly with distribution. State-of-the-art image-set typically techniques matching a,State-of-the-art image-set matching techniques typically implicitly model each image-set with a Gaussian distribution.,14830
"large TabletGaze of users, labeled tablet first dataset We collected unconstrained Rice the gaze dataset.","We collected the first large unconstrained gaze dataset of tablet users, labeled Rice TabletGaze dataset.",14831
"functional significant (perceptron we obtain using MMBANN. advantages, SOM) Besides, while models and which describe","Besides, we describe functional models (perceptron and SOM) which obtain significant advantages, while using MMBANN.",14832
miRNA on features relies precursors microRNAs characteristic from (miRNAs) models Discovery (pre-miRNAs). for of predictive,Discovery of microRNAs (miRNAs) relies on predictive models for characteristic features from miRNA precursors (pre-miRNAs).,14833
"by algorithms. (MCMC) Normally, Markov Monte is standard of means done this Carlo chain","Normally, this is done by means of standard Markov chain Monte Carlo (MCMC) algorithms.",14834
work scene In problem images. the of this address from RGB-D indoor understanding we,In this work we address the problem of indoor scene understanding from RGB-D images.,14835
"variables, systems large-scale spatial contain e.g. study exogenous We that","We study large-scale spatial systems that contain exogenous variables, e.g.",14836
predictors environmental spatial processes. are that in factors significant,environmental factors that are significant predictors in spatial processes.,14837
optimization We an procedure develop to hyperplanes. find also the desired,We also develop an optimization procedure to find the desired hyperplanes.,14838
a conjunction Machine used were Both Vector classifier. (SVM) techniques in with Support,Both techniques were used in conjunction with a Support Vector Machine (SVM) classifier.,14839
of acids constituting amino the Viral amongst by the constrained capsids interactions proteins. structurally are,Viral capsids are structurally constrained by interactions amongst the amino acids of the constituting proteins.,14840
We increased has. has epistasis significantly energy that found free but not,We found that free energy has not significantly increased but epistasis has.,14841
an for predicting Developing is structures. function accurate protein essential scoring successfully,Developing an accurate scoring function is essential for successfully predicting protein structures.,14842
expect beta-sheet We useful for proteins. to OPUS-Beta in potential be modeling,We expect OPUS-Beta potential to be useful in beta-sheet modeling for proteins.,14843
"variables making Furthermore, the are assumptions, not without strong identifiable. calibration","Furthermore, without making strong assumptions, the calibration variables are not identifiable.",14844
restricts the systems. real-time their This application in,This restricts their application in the real-time systems.,14845
dataset that it is The algorithms. competitive FDDB state-of-the-art test results show on with,The test results on FDDB dataset show that it is competitive with state-of-the-art algorithms.,14846
considerably our algorithms. GPU-based CPU-based all exceeds of speed the approach The detection and existing,The detection speed of our approach considerably exceeds all the existing CPU-based and GPU-based algorithms.,14847
the of achieved the cascade. asynchronous stages by computation is This in,This is achieved by the asynchronous computation of stages in the cascade.,14848
Protein-DNA for biological processes. are all crucial interactions,Protein-DNA interactions are crucial for all biological processes.,14849
Using theoretical we developed question detail. recently a framework this analyze in,Using a recently developed theoretical framework we analyze this question in detail.,14850
"two three have explicitly been Specifically, systems investigated. with and targets","Specifically, systems with two and three targets have been explicitly investigated.",14851
"proportional not However, always the acceleration the number of targets. is to","However, the acceleration is not always proportional to the number of targets.",14852
explanations of are Physical-chemical results presented. observed,Physical-chemical explanations of observed results are presented.,14853
Carlo Monte simulations fully Extensive calculations. theoretical our computer support,Extensive Monte Carlo computer simulations fully support our theoretical calculations.,14854
texture under still an open of is issue. The conditions color varying recognition lighting,The recognition of color texture under varying lighting conditions is still an open issue.,14855
"from preliminary points data. detected the are discrete Next, boundary A-Scans volume of the","Next, preliminary discrete boundary points are detected from the A-Scans of the volume data.",14856
"sensitive KL is it divergence is that known outliers. the However, to","However, it is known that the KL divergence is sensitive to outliers.",14857
"based these seems method For a reduction promising. on dimension developing supervised reasons, LSQMI","For these reasons, developing a supervised dimension reduction method based on LSQMI seems promising.",14858
demonstrate datasets our on algorithm. three Extensive experiments popular image multiview embedding the of effectiveness,Extensive experiments on three popular image datasets demonstrate the effectiveness of our multiview embedding algorithm.,14859
analysis. solution validated statistically The using been has statistical proposed by,The proposed solution has been statistically validated by using statistical analysis.,14860
as expressed contemporary scripts. Most of still AI the is hard-coded,Most of the contemporary AI is still expressed as hard-coded scripts.,14861
address we this this paper In issue.,In this paper we address this issue.,14862
of objects the extend concept inspired Behavior object-oriented and programming objects. smart are by,Behavior objects are inspired by object-oriented programming and extend the concept of smart objects.,14863
"we In propose method identification aliases a tackle to this avatar article, this problem.","In this article, we propose a method to tackle this avatar aliases identification problem.",14864
of optic nerve abnormal (e.g. Segmentation pathologically,Segmentation of pathologically abnormal optic nerve (e.g.,14865
unpredictable changes glioma) its poses due to additional nerve at challenges optic in locations. shape,optic nerve glioma) poses additional challenges due to changes in its shape at unpredictable locations.,14866
"comparisons, be other using to results methods scales. the presented may pairwise relevant Beyond subjective","Beyond pairwise comparisons, the results presented may be relevant to other methods using subjective scales.",14867
"scale, Keywords: comparisons, intelligence, assessment, inaccuracy, pairwise collective inconsistency. subjective","Keywords: pairwise comparisons, collective intelligence, scale, subjective assessment, inaccuracy, inconsistency.",14868
re-using datasets. consists object solution pre-trained in generic on detectors A scalable,A scalable solution consists in re-using object detectors pre-trained on generic datasets.,14869
with parameters. considers three present distribution  the paper The modified of  exponential extension,  The present paper considers modified extension of the exponential distribution with three parameters.,14870
"propose to sample Gibbs posterior from the we generate distribution. Therefore, method also sampling","Therefore, we also propose Gibbs sampling method to generate sample from the posterior distribution.",14871
illustrative set of has real data been for the considered A study. purpose,A real data set has been considered for illustrative purpose of the study.,14872
"has orientation the substrates on and biomolecules been However, single less arrangement of investigated.","However, the arrangement and orientation of single biomolecules on substrates has been less investigated.",14873
studies to Molecular positional used interpret distributions. were dynamics observed the,Molecular dynamics studies were used to interpret the observed positional distributions.,14874
cameras on by other experiments data produce The collected outperforming RGBD techniques. results,The experiments on data collected by RGBD cameras produce results outperforming other techniques.,14875
extracellular ongoing and spatially- time-varying ($V_e$). of neurons generates The voltage of field a activity,The ongoing activity of neurons generates a spatially- and time-varying field of extracellular voltage ($V_e$).,14876
"interactions"" these are but not find small that We ""ephaptic negligible.","We find that these ""ephaptic interactions"" are small but not negligible.",14877
These suggest MSO influence that results may ephaptic coupling function.,These results suggest that ephaptic coupling may influence MSO function.,14878
"manifolds, them? view Is by it or the preserving them achieved separating while collapsing","Is it achieved by collapsing the view manifolds, or separating them while preserving them?",14879
layer At view achieved? is invariance which,At which layer is view invariance achieved?,14880
recognition in effective be images. for to VGGNets still turned out object have,VGGNets have turned out to be effective for object recognition in still images.,14881
made available. performance the datasets Our are state-of-the-art public models trained achieve on and these,Our trained models achieve the state-of-the-art performance on these datasets and are made public available.,14882
dictionary. dictionaries a and build: dictionary feature are Two region-based voxel-wise A,Two dictionaries are build: A region-based feature dictionary and a voxel-wise dictionary.,14883
The dictionaries set level provide to the prior learned knowledge formulation.,The learned dictionaries provide prior knowledge to the level set formulation.,14884
scores challenge the by metric different results using computed organizers. are evaluated The,The results are evaluated using different metric scores computed by the challenge organizers.,14885
using is (DCNN) dataset. CASIA-WebFace the convolutional The deep trained network neural,The deep convolutional neural network (DCNN) is trained using the CASIA-WebFace dataset.,14886
provided. the are on IJB-A experiments dataset Extensive,Extensive experiments on the IJB-A dataset are provided.,14887
diagrams path learning consider problem structure bow-free of for We acyclic (BAPs). the,We consider the problem of structure learning for bow-free acyclic path diagrams (BAPs).,14888
present method this using a greedy algorithm. a We score-based problem first for search,We present a first method for this problem using a greedy score-based search algorithm.,14889
us causal infer to effects. This of lower allows bounds,This allows us to infer lower bounds of causal effects.,14890
using available R-package. simulated also datasets applications publicly to and We present our real,We also present applications to real and simulated datasets using our publicly available R-package.,14891
"analysis with feature is an Thus, desired. set highly integration of gene selection","Thus, an integration of gene set analysis with feature selection is highly desired.",14892
must taken that into is account. be The genes interplay basic among requirement,The basic requirement is that interplay among genes must be taken into account.,14893
regulated tightly The of is eukaryotes. dynamics in transcription gene,The dynamics of gene transcription is tightly regulated in eukaryotes.,14894
the transcriptional us may help better findings understand principles of regulation. design These,These findings may help us better understand the design principles of transcriptional regulation.,14895
"surface function, a The natural determined metric. a is which provides by potential stability","The surface is determined by a potential function, which provides a natural stability metric.",14896
"systems systems, quite gradient representation, amenable However, rare. to called are this","However, systems amenable to this representation, called gradient systems, are quite rare.",14897
spherical into angle force This a turn- coordinates. ing is then using converted,This force is then converted into a turn- ing angle using spherical coordinates.,14898
ligand with through environment is exponential The functions. modeled extracellular diffusion the linear or,The ligand diffusion through the extracellular environment is modeled with linear or exponential functions.,14899
cluster Siberian populations Paleo-Eskimos. related We to Kets established to that the of the belong,We established that the Kets belong to the cluster of Siberian populations related to Paleo-Eskimos.,14900
to varying optical flow solve a consider problem with method variational We the illumination.,We consider a variational method to solve the optical flow problem with varying illumination.,14901
of The parallelism solver MUMPS. level second massively uses the parallel,The second level of parallelism uses the massively parallel solver MUMPS.,14902
This procedure on the Gibbs built using relies algorithm particle a sampler. split-merge flexible,This algorithm relies on a flexible split-merge procedure built using the particle Gibbs sampler.,14903
important molecular interaction models class are networks. networks Boolean an of for computational,Boolean networks are an important class of computational models for molecular interaction networks.,14904
This networks. control of role molecular the work in Boolean studies of the canalization,This work studies the role of canalization in the control of Boolean molecular networks.,14905
seamlessly into reality the to art objects Augmented fuse ones. is real virtual,Augmented reality is the art to seamlessly fuse virtual objects into real ones.,14906
"effective, very choices. design involve many they user-defined Albeit","Albeit very effective, they involve many user-defined design choices.",14907
"network can learn? the "", ""what and",""", and ""what can the network learn?",14908
augmentation. using show renderings can when for overall improve that We performance image data,We show that overall performance can improve when using image renderings for data augmentation.,14909
many problem computer a low-level Image interesting long-standing vision applications. restoration in with is,Image restoration is a long-standing problem in low-level computer vision with many interesting applications.,14910
Reaction TNRD -- \textit{Trainable this Nonlinear approach We call Diffusion}.,We call this approach TNRD -- \textit{Trainable Nonlinear Reaction Diffusion}.,14911
"matter satisfactory resolve existing not a view, way. our do solutions in the In this","In our view, the existing solutions do not resolve this matter in a satisfactory way.",14912
the most fast transform. utilizes of and approach elegant efficient Fourier One the,One of the most elegant and efficient approach utilizes the fast Fourier transform.,14913
paper the In problem describe give this solution. satisfactory we a and,In this paper we describe the problem and give a satisfactory solution.,14914
probability a There of models is implementing third practicing. way and,There is a third way of implementing probability models and practicing.,14915
This questions in of to observables. is answer terms put,This is to answer questions put in terms of observables.,14916
eliminates testing and frequentist eliminates also hypothesis This factors and Bayes it estimation. parameter,This eliminates frequentist hypothesis testing and Bayes factors and it also eliminates parameter estimation.,14917
"stated verifiable fashion, publicly in as predictions. All are models","All models are stated in publicly verifiable fashion, as predictions.",14918
into put models is process any trust All a them. verification must before undergo,All models must undergo a verification process before any trust is put into them.,14919
convex hill climbing problems. is guarantied is to only method that for a L-BFGS converge,L-BFGS is a hill climbing method that is guarantied to converge only for convex problems.,14920
solving difficult with results L-BFGS. nice works by very obtain such problems Some,Some works obtain very nice results by solving such difficult problems with L-BFGS.,14921
"which In to techniques be regime, known algorithms sub-sampling effective. this utilize are","In this regime, algorithms which utilize sub-sampling techniques are known to be effective.",14922
algorithm us develop select which also to its allows near-optimal analysis parameters. theoretical We,We develop its theoretical analysis which also allows us to select near-optimal algorithm parameters.,14923
We how to problems. demonstrate results well-known our apply machine learning,We demonstrate how our results apply to well-known machine learning problems.,14924
"several various scenarios. we under our the on datasets of Lastly, performance evaluate algorithm","Lastly, we evaluate the performance of our algorithm on several datasets under various scenarios.",14925
miRNAs essential post-transcriptional various crucial in fate cell as regulators decision. serve,miRNAs serve as crucial post-transcriptional regulators in various essential cell fate decision.,14926
"However, the bistability not is mutual mRNA-miRNA contribution the to understood. of regulation fully","However, the contribution of the mRNA-miRNA mutual regulation to bistability is not fully understood.",14927
"stronger region a competing bistability mRNA by expanded (ceRNA). Second, is of the","Second, the region of bistability is expanded by a stronger competing mRNA (ceRNA).",14928
"importance in demonstrated regulation reciprocal cell the fate mRNA-miRNA decision. simple Thus, of we","Thus, we demonstrated the importance of simple mRNA-miRNA reciprocal regulation in cell fate decision.",14929
"paper, this Common series. Fuzzy In propose matching we Longest time Subsequence for","In this paper, we propose Fuzzy Longest Common Subsequence matching for time series.",14930
is semantic human regions analysis. into body Parsing crucial to human-centric,Parsing human body into semantic regions is crucial to human-centric analysis.,14931
"that a explores human paper, propose segment-based parsing this i.e. information, pipeline pose In we","In this paper, we propose a segment-based parsing pipeline that explores human pose information, i.e.",14932
"are though of proposed Graph. the Finally, the an inferred segments ensemble part And-Or best","Finally, the best ensemble of the proposed part segments are inferred though an And-Or Graph.",14933
data objective data major of Uncovering causal is in a analytics. relationships,Uncovering causal relationships in data is a major objective of data analytics.,14934
"with e.g. experiments, Causal normally designed are discovered relationships","Causal relationships are normally discovered with designed experiments, e.g.",14935
"many cases. or be conducted randomised which, expensive infeasible in to controlled are however trials,","randomised controlled trials, which, however are expensive or infeasible to be conducted in many cases.",14936
"develop In have tree causal paper, decision nodes a we this where interpretations. causal","In this paper, we develop a causal decision tree where nodes have causal interpretations.",14937
signals method data sets. finding practical for in causal The large is,The method is practical for finding causal signals in large data sets.,14938
machine computational in representations major advantages. learning ubiquitous Factored and to are lead,Factored representations are ubiquitous in machine learning and lead to major computational advantages.,14939
probabilistic large a representation. show of models a class compact have We Fourier that graphical,We show that a large class of probabilistic graphical models have a compact Fourier representation.,14940
approximating a up result opens probability of distribution. an theoretical This entirely new way,This theoretical result opens up an entirely new way of approximating a probability distribution.,14941
the demonstrate the this applying variable by significance elimination to We approach algorithm. of it,We demonstrate the significance of this approach by applying it to the variable elimination algorithm.,14942
"representation the traditional inference with significant obtain Compared other approximate bucket algorithms, we improvements. and","Compared with the traditional bucket representation and other approximate inference algorithms, we obtain significant improvements.",14943
distribution. model finger vein the using a user We Gaussian,We model the user finger vein using a Gaussian distribution.,14944
a results on state-of-the-art that SDUMLA-HMT perform algorithm benchmark. show our Experimental like,Experimental results show that our algorithm perform like a state-of-the-art on SDUMLA-HMT benchmark.,14945
also may useful results timescales. It relatively on deliver short,It may also deliver useful results on relatively short timescales.,14946
"become has Internet concept. popular a culture, development of the cuteness With","With the development of Internet culture, cuteness has become a popular concept.",14947
about what Many look factors person are making a curious cute. people,Many people are curious about what factors making a person look cute.,14948
"this rare However, answer there question. to research is interesting","However, there is rare research to answer this interesting question.",14949
prediction. effectiveness cuteness method the proposed Extensive evaluations the validate for of,Extensive evaluations validate the effectiveness of the proposed method for cuteness prediction.,14950
The applicable approach without is multi-class to any and binary proposed both modification. problems,The proposed approach is applicable to both binary and multi-class problems without any modification.,14951
investigations preliminary Our both existed. that suggest,Our preliminary investigations suggest that both existed.,14952
inform of animals modern-day examples discussions chirality. We our exhibit exterior that throughout with,We inform our discussions throughout with examples of modern-day animals that exhibit exterior chirality.,14953
of assumptions it these Here proven that is is neither correct.,Here it is proven that neither of these assumptions is correct.,14954
networks. encoded with of principles combinatorics inconsistent bounded of are They fundamental,They are inconsistent with fundamental principles of combinatorics of bounded encoded networks.,14955
MAP We the from also posterior learn ascent. solution by gradient the,We also learn the MAP solution from the posterior by gradient ascent.,14956
We classification temporal performance. constraints significantly that improve the show,We show that temporal constraints significantly improve the classification performance.,14957
"always reverse relationships the imply causal but The is associations, not is, true. problem","The problem is, causal relationships imply associations, but the reverse is not always true.",14958
However here. can see paradigms the between two the synergy we,However we can see the synergy between the two paradigms here.,14959
notion provide formalisms of Many some solution. these of,Many of these formalisms provide some notion of solution.,14960
"for a which the counterexamples. on provide formalization theorem we Unfortunately, relies","Unfortunately, the formalization relies on a theorem for which we provide counterexamples.",14961
are values. low a power sparsity The at by approximated law distributions closely,The distributions are closely approximated by a power law at low sparsity values.,14962
Model This a proposes Subdivision-Fusion (SFM) recognize to human novel actions. paper,This paper proposes a novel Subdivision-Fusion Model (SFM) to recognize human actions.,14963
"action distribution In feature tasks, is most leading overfitting. to common overlapping a problem recognition","In most action recognition tasks, overlapping feature distribution is a common problem leading to overfitting.",14964
"SFM, clustered. the subdivision are proposed in of stage In the category samples each","In the subdivision stage of the proposed SFM, samples in each category are clustered.",14965
"more multiple Then, grouped are into samples concentrated subcategories. such","Then, such samples are grouped into multiple more concentrated subcategories.",14966
for subcategories as overfitting and easier avoided. are the find to is Boundaries consequence,Boundaries for the subcategories are easier to find and as consequence overfitting is avoided.,14967
number to Two clusters methods of provided. determine the are,Two methods to determine the number of clusters are provided.,14968
popular has tested proposed The been thoroughly datasets. model with four,The proposed model has been thoroughly tested with four popular datasets.,14969
systems described utilizing are analogies. often Living informational,Living systems are often described utilizing informational analogies.,14970
completion an learn of problem image. to the missing The is tag tags of,The problem of tag completion is to learn the missing tags of an image.,14971
our the is of ICONIP. version This on paper preprint,This is the preprint version of our paper on ICONIP.,14972
of boundaries. important An detection the neuronal is reconstruction for computation,An important computation for reconstruction is the detection of neuronal boundaries.,14973
a gain in innovations. achieve Here three substantial we accuracy through,Here we achieve a substantial gain in accuracy through three innovations.,14974
deep We features. face based Deformable detection present algorithm and on Models a Part pyramidal,We present a face detection algorithm based on Deformable Part Models and deep pyramidal features.,14975
"the also we that Furthermore, regularizer of restoration images. can benefit show general natural this","Furthermore, we show that this regularizer can also benefit the restoration of general natural images.",14976
the aspects company. important one is a Warehouse of of,Warehouse is one of the important aspects of a company.,14977
"parameter generates backtracking First, it from solutions algorithm. a tree","First, it generates a solutions parameter tree from backtracking algorithm.",14978
for system. warehouse an method attractive proved This solution is be alternative layout to,This method is proved to be an attractive alternative solution for warehouse layout system.,14979
"these We approach into compare"", while account. taking an propose ""learns that to difficulties","We propose an approach that ""learns to compare"", while taking these difficulties into account.",14980
the paradigm. maximum trained with The likelihood network is,The network is trained with the maximum likelihood paradigm.,14981
presents in significantly terms convergence. a PISAA improved performance of,PISAA presents a significantly improved performance in terms of convergence.,14982
be parallel PISAA available. if environments can computing implemented in,PISAA can be implemented in parallel computing environments if available.,14983
reconstructed The variety imaging. x-ray to of than medical be is broader materials,The variety of materials to be reconstructed is broader than medical x-ray imaging.,14984
"to achieve algorithms require computations may results. in iterative quantitatively However, order accurate","However, iterative algorithms may require computations in order to achieve quantitatively accurate results.",14985
increase in approaches the convergence. are proposed to literature many speed of There,There are many approaches proposed in the literature to increase speed of convergence.,14986
"Initially, around surrogate coefficients approximation only. are created functions","Initially, surrogate functions are created around approximation coefficients only.",14987
"studies number air and of a have with dealt Recently, pollution. quality air","Recently, a number of studies have dealt with air quality and air pollution.",14988
"should in any software JavaScript, The SVG. run that browser web HTML, supports modern and","The software should run in any web browser that supports modern JavaScript, HTML, and SVG.",14989
release version. The is new every stable of updated with a manual,The manual is updated with every release of a new stable version.,14990
on Collective interactions. to cell cell-cell cues responses depend exogenous,Collective cell responses to exogenous cues depend on cell-cell interactions.,14991
"can enhanced stimuli. in and noisy result to these In sensitivity principle, weak","In principle, these can result in enhanced sensitivity to weak and noisy stimuli.",14992
as of problem We similarity regularized pose learning. hashing a learning,We pose hashing learning as a problem of regularized similarity learning.,14993
accepted our a marked several performance baselines approach increase. Comparing against demonstrates,Comparing our approach against several accepted baselines demonstrates a marked performance increase.,14994
based hierarchy model image is The partitions. a of first on,The first model is based on a hierarchy of image partitions.,14995
are integrated final saliency The a partitions different the obtained maps then for into map.,The maps obtained for the different partitions are then integrated into a final saliency map.,14996
saliency produce the We maps. that proposed show models high quality,We show that the proposed models produce high quality saliency maps.,14997
performance evaluation datasets. benchmark two achieve several state-of-the-art Objective that methods in the demonstrates,Objective evaluation demonstrates that the two methods achieve state-of-the-art performance in several benchmark datasets.,14998
given for models There tissue. are stochastic a risk numerous cancer for,There are numerous stochastic models for cancer risk for a given tissue.,14999
on Many two following the hypotheses. rely,Many rely on the following two hypotheses.,15000
each are other. nefarious not or of divisions independently Cell,Cell divisions are nefarious or not independently of each other.,15001
type The Boolean mathematical is considered that model of networks.,The mathematical model type considered is that of Boolean networks.,15002
practical This role data. play significant sequential in an important has time-series to implications representing,This has significant implications to play an important role in representing practical time-series sequential data.,15003
data eventually an and sequential provides retrieval. This efficient flexible,This eventually provides an efficient and flexible sequential data retrieval.,15004
to the in of each the HLA peptides The determined. affinity relevant was binding pair,The binding affinity of the peptides to the relevant HLA in each pair was determined.,15005
introduce novel the a the problem of photographer We identifying behind photograph.,We introduce the novel problem of identifying the photographer behind a photograph.,15006
a network deep new neural this convolutional task. also trained We for,We also trained a new deep convolutional neural network for this task.,15007
Our outperform features. features greatly high-level show results low-level that,Our results show that high-level features greatly outperform low-level features.,15008
applications method. We two demonstrate our also of,We also demonstrate two applications of our method.,15009
"Moreover, present filtering. outlier novel for we enhancements","Moreover, we present novel enhancements for outlier filtering.",15010
"the EpicFlow MPI-Sintel, significantly KITTI on outperform Middlebury. original and We","We significantly outperform the original EpicFlow on MPI-Sintel, KITTI and Middlebury.",15011
with model The carrying the basic emergent system is competition Lotka-Volterra capacities.,The basic model is the Lotka-Volterra competition system with emergent carrying capacities.,15012
invasion used for are of numerical The phenomena. latter study the,The latter are used for the numerical study of invasion phenomena.,15013
is the learning employed LAA thrombi solve Multiple-instance detection. method (MIL) to,Multiple-instance learning (MIL) method is employed to solve the LAA thrombi detection.,15014
importance. is an vital insight Enhancement human get of information to of content to vision,Enhancement of human vision to get an insight to information content is of vital importance.,15015
study handles the showed present efficiently Results method the of the noise amplification. proposed that,Results of the present study showed that the proposed method efficiently handles the noise amplification.,15016
"image. targeted it preserves Further, brightness the also by look natural of retaining","Further, it also preserves the brightness by retaining natural look of targeted image.",15017
wildlife protected establishment an for conservation. or tool become has The important areas of parks,The establishment of protected areas or parks has become an important tool for wildlife conservation.,15018
crop-raiding as effective hunting is of mesopredators Human only occurs it during less excursions.,Human hunting of mesopredators is less effective as it only occurs during crop-raiding excursions.,15019
"way, is segmentation achieved. In result this road a homogenous","In this way, a homogenous road segmentation result is achieved.",15020
"discontinuous road This also smooth road method centerlines. centerlines, the connects but extracts not only","This method not only extracts smooth road centerlines, but also connects the discontinuous road centerlines.",15021
"can As get network. we road a a centerline result, complete","As a result, we can get a complete road centerline network.",15022
simulation was to methods. The leave-one-out strategy different compare used,The leave-one-out strategy was used to compare different simulation methods.,15023
of same a set dominant with monomials class. or equilibrations the define Tropical equivalence branch,Tropical equilibrations with the same set of dominant monomials define a branch or equivalence class.,15024
the interesting branches Minimal reaction network. states are simplest they particularly describe as the of,Minimal branches are particularly interesting as they describe the simplest states of the reaction network.,15025
data. two applications The real methodological illustrated are to with novelties,The methodological novelties are illustrated with two applications to real data.,15026
complex Performing in of brains. typical a parallel tasks is more feature,Performing more tasks in parallel is a typical feature of complex brains.,15027
parallel Boolean networks. more of neuronal we investigate Here rules learning in,Here we investigate parallel learning of more Boolean rules in neuronal networks.,15028
"The is i.e. protein its function, sequence length a determined of by Background: largely","Background: The length of a protein sequence is largely determined by its function, i.e.",15029
size. functional each with optimal group associated an is,each functional group is associated with an optimal size.,15030
"comparative length revealed by that be factors. However, genomics additional proteins may affected","However, comparative genomics revealed that proteins length may be affected by additional factors.",15031
"genomes. and all in length As distributions have the different HHPs strikingly almost ORFans expected,","As expected, the HHPs and ORFans have strikingly different length distributions in almost all genomes.",15032
"genomes about thirty do rules. these obey not However,","However, about thirty genomes do not obey these rules.",15033
distributions also have genomes that showed We Mycoplasmataceae lengths. distinctive of ORFans,We also showed that Mycoplasmataceae genomes have distinctive distributions of ORFans lengths.,15034
possible We several of phenomenon. offer this biological explanations,We offer several possible biological explanations of this phenomenon.,15035
is secondary RNA compute that It follows conformational structure to entropy. state-of-the-art RNAentropy software,It follows that RNAentropy is state-of-the-art software to compute RNA secondary structure conformational entropy.,15036
interesting motivation. an weighted (DWD) classifier margin-based Distance with geometric is discrimination a,Distance weighted discrimination (DWD) is a margin-based classifier with an interesting geometric motivation.,15037
faults techniques identifying for on of rely Classical testing. the development acoustic,Classical techniques for identifying the development of faults rely on acoustic testing.,15038
different materials long. testing since carried been for destructive out has Non,Non destructive testing has been carried out for different materials since long.,15039
two each model representative sets are computed For parameters. with of solutions,For each model solutions are computed with two representative sets of parameters.,15040
phase plane and supports these model first bifurcation findings. analysis a the For,For the first model a bifurcation and phase plane analysis supports these findings.,15041
"an Bayesian algorithmic i.e. On abstract, nonlinear level, filtering,","On an abstract, algorithmic level, nonlinear Bayesian filtering, i.e.",15042
three show outperform on results methods. experimental state-of-the-art benchmark methods hashing the datasets The proposed,The experimental results on three benchmark datasets show the proposed methods outperform state-of-the-art hashing methods.,15043
is most of active biologically regularly consumed widely one substances. the Caffeine (CAF) and,Caffeine (CAF) is one of the most widely and regularly consumed biologically active substances.,15044
at to The concentrations target refers principal receptors. adenosine CAF physiologically important,The principal CAF target at physiologically important concentrations refers to adenosine receptors.,15045
adenosine. antagonist of a opinion CAF a that is It common competitive is,It is a common opinion that CAF is a competitive antagonist of adenosine.,15046
which SPF-CellTracker. suite We developed we software a call tracking,We developed a tracking software suite which we call SPF-CellTracker.,15047
the pose dot-product embeddings. image the score between The and function is then,The score function is then the dot-product between the image and pose embeddings.,15048
score are cost The and trained jointly using maximum-margin image-pose a function. function embedding,The image-pose embedding and score function are jointly trained using a maximum-margin cost function.,15049
as technique techniques other be compared to . best value Eigen two the with found,Eigen value technique found to be the best as compared with other two techniques .,15050
except small trees cases. understood well not gene anomalous some are for Such,Such anomalous gene trees are not well understood except for some small cases.,15051
how learning. discuss stimulate questions BrainPlay We,We discuss how BrainPlay questions stimulate learning.,15052
intelligence. that questions find We aspects BrainPlay many span of,We find that BrainPlay questions span many aspects of intelligence.,15053
"in many relationships been developed have methods detecting data. In recent years, observational for causal","In recent years, many methods have been developed for detecting causal relationships in observational data.",15054
sets. to large data the Some of tackle have potential them,Some of them have the potential to tackle large data sets.,15055
"i.e. these However, combined discover methods to cause, a fail","However, these methods fail to discover a combined cause, i.e.",15056
two causes. component cause multi-factor individually consisting variables are a of not which or more,a multi-factor cause consisting of two or more component variables which individually are not causes.,15057
in data mining causes sets. combined large,mining combined causes in large data sets.,15058
univocally cameras presence and of identifies system of the them. people detects The,The system of cameras detects the presence of people and univocally identifies them.,15059
our gradient the (APG) to efficiently. We utilize accelerated proximal solve method model,We utilize the accelerated proximal gradient (APG) method to solve our model efficiently.,15060
We annotation. intuition improve to this multilabel on build image,We build on this intuition to improve multilabel image annotation.,15061
path in the computation key-node. shortest This of a done temporal is neighborhood through,This is done through shortest path computation in a temporal neighborhood of the key-node.,15062
"momentum helps gradient. biodiversity steep tropical latitudinal to the perpetuate Thus, evolutionary","Thus, tropical evolutionary momentum helps to perpetuate the steep latitudinal biodiversity gradient.",15063
hyper-parameters process related are estimation of to Gaussian the presented. Examples,Examples related to the estimation of Gaussian process hyper-parameters are presented.,15064
cancer for Lung related the deaths. leading is cause cancer,Lung cancer is the leading cause for cancer related deaths.,15065
a imaging do to feature tool set. high-dimension this is radiomics: powerful A,A powerful tool to do this is radiomics: a high-dimension imaging feature set.,15066
"treatment and Finally, an implications. of effects antiretroviral we drug the study","Finally, we study the effects of an antiretroviral drug and treatment implications.",15067
complementary facial CNNs features multimodal set The from extracts data. of,The set of CNNs extracts complementary facial features from multimodal data.,15068
this method videos. unsupervised introduce to We representation from an generate,We introduce an unsupervised method to generate this representation from videos.,15069
model recognition The action multiple benchmarks. achieves proposed in state-of-the-art performance,The proposed model achieves state-of-the-art performance in multiple action recognition benchmarks.,15070
set the automatically tune possible approach nested it and to Our hyperparameters estimates. obtain makes,Our approach makes it possible to tune the hyperparameters automatically and obtain nested set estimates.,15071
also on presented. sets data Results actual are,Results on actual data sets are also presented.,15072
in navigation systems. autonomous a fundamental is  task Road  detection,  Road detection is a fundamental task in autonomous navigation systems.,15073
great using demonstrate contextual and benefit The computational from results efficiency. a blocks show their,The results show a great benefit from using contextual blocks and demonstrate their computational efficiency.,15074
of main contributions this are There two article.,There are two main contributions of this article.,15075
"used Second, various of be we that may various sensors data. list types for acquiring","Second, we list various types of sensors that may be used for acquiring various data.",15076
"for used classifying review Third, taxonomies different we motions.","Third, we review different taxonomies used for classifying motions.",15077
"involved we motion processes analysis. in review various Fourth,","Fourth, we review various processes involved in motion analysis.",15078
"how exhibit structured. surveys are we Fifth, different","Fifth, we exhibit how different surveys are structured.",15079
"provide we the computer general of aspect of subject Finally, the discussion vision. a from","Finally, we provide a general discussion of the subject from the aspect of computer vision.",15080
key manipulated Our a effective object generation contribution scheme. proposal yet is simple,Our key contribution is a simple yet effective manipulated object proposal generation scheme.,15081
in numerical assessment paper concerns filters. of Carlo This Monte particle error,This paper concerns numerical assessment of Monte Carlo error in particle filters.,15082
photosynthesis. their optimize to growth Chloroplasts regulate,Chloroplasts regulate their growth to optimize photosynthesis.,15083
more scatter Wild-type chloroplasts trend; chloroplasts highly this irregularly-shaped little scatter. exhibit around mutant exhibit,Wild-type chloroplasts exhibit little scatter around this trend; highly irregularly-shaped mutant chloroplasts exhibit more scatter.,15084
"depth robust individual Unfortunately, its preventing technique each from suffers sensing. limitations","Unfortunately, each technique suffers from its individual limitations preventing robust depth sensing.",15085
refer field imaging. as this combination to We depth,We refer to this combination as depth field imaging.,15086
"today's advertising rapidly huge, in growing is advertising market a world. Online","Online advertising is a huge, rapidly growing advertising market in today's world.",15087
advertising ads. of is image common form One using online,One common form of online advertising is using image ads.,15088
"there a fundamental However, more is layer.","However, there is a more fundamental layer.",15089
"to address recurrent In this neural introduce networks this we paper, (RNNs) issue.","In this paper, we introduce recurrent neural networks (RNNs) to address this issue.",15090
of treat We chain. the folded perturbed protein the strongly the as Rouse state,We treat the folded protein as the strongly perturbed state of the Rouse chain.,15091
properties. question addition the the whether intelligence show In we may fractal raise,In addition we raise the question whether the intelligence may show fractal properties.,15092
through finite carry analysis of the populations the this of to features Some key case.,Some of the key features of this analysis carry through to the finite populations case.,15093
possible this one to We answer offer question.,We offer one possible answer to this question.,15094
two amodal segmentation. datasets for We create semantic,We create two datasets for semantic amodal segmentation.,15095
classification rejection using supervised We new algorithm with for information. image multiscale contextual a introduce,We introduce a new supervised algorithm for image classification with rejection using multiscale contextual information.,15096
"The performance. with proposed local combines rejection, the classification information and multiscale improving contextual algorithm","The proposed algorithm combines local and multiscale contextual information with rejection, improving the classification performance.",15097
"for logistic multinomial a a adopt regression. probabilistic classification, we As model","As a probabilistic model for classification, we adopt a multinomial logistic regression.",15098
We the classification of of images H&E-stained tissues. the teratoma proposed method illustrate on,We illustrate the proposed method on the classification of images of H&E-stained teratoma tissues.,15099
provide of discussion by a the thorough Rodgers we study Here al. et conducted,Here we provide a thorough discussion of the study conducted by Rodgers et al.,15100
"access to easy In offers it addition, parallel computations.","In addition, it offers easy access to parallel computations.",15101
additive down approach We quantitative genetic as an write phenotype to use map. the the,We use the quantitative genetic approach to write down the phenotype as an additive map.,15102
then component integral phenotype. of construct the We for models projection each,We then construct integral projection models for each component of the phenotype.,15103
per projection can be shorter generation or formulated on time Population models steps. on,Population projection models can be formulated on per generation or on shorter time steps.,15104
"Yet, remains the large-scale networks. understanding control elusive of metabolic fluxes metabolic for efficient","Yet, understanding the efficient control of metabolic fluxes remains elusive for large-scale metabolic networks.",15105
find unicellular multicellular that control most extensive We less organisms than require organisms.,We find that most unicellular organisms require less extensive control than multicellular organisms.,15106
reactions in coli. strong transcriptional under identified Escherichia driver The regulation are,The identified driver reactions are under strong transcriptional regulation in Escherichia coli.,15107
we of detection. problem event paper efficient the video In this tackle,In this paper we tackle the problem of efficient video event detection.,15108
image microscopic in presented. counting a of cell this is An to application technique also,An application of this technique to cell counting in a microscopic image is also presented.,15109
exists. regular But this no not applicable grid is arbitrary to because approach surfaces,But this approach is not applicable to arbitrary surfaces because no regular grid exists.,15110
on paper nonstationary a will kernel approach data surfaces. for convolution describe This arbitrary,This paper will describe a nonstationary kernel convolution approach for data on arbitrary surfaces.,15111
"algorithm we as a probabilistic linear Using factorisation filter. model, matrix the derive a recursive","Using the probabilistic model, we derive a matrix factorisation algorithm as a recursive linear filter.",15112
comparisons image an given. and restoration are Demonstrations on task,Demonstrations and comparisons on an image restoration task are given.,15113
"for an we paper, RGB-D segmentation and propose unsupervised method indoor this analysis. In image","In this paper, we propose an unsupervised method for indoor RGB-D image segmentation and analysis.",15114
"manner. opens color an perspectives in unsupervised Moreover, it to geometry fuse interesting and","Moreover, it opens interesting perspectives to fuse color and geometry in an unsupervised manner.",15115
changes (fractal dynamics) among long-range consecutive correlations the strides AC anti-correlations. into,AC changes the long-range correlations among consecutive strides (fractal dynamics) into anti-correlations.,15116
floor. of (VC) with marks on step the lengths alignment is cueing the Visual,Visual cueing (VC) is the alignment of step lengths with marks on the floor.,15117
of walking the on effects not The been have VC fluctuation structure investigated. of,The effects of VC on the fluctuation structure of walking have not been investigated.,15118
"and conditions The no AC, VC. (NC), were cueing","The conditions were no cueing (NC), AC, and VC.",15119
"stepping the projected stones speed. with synchronized treadmill In VC, were","In VC, projected stepping stones were synchronized with the treadmill speed.",15120
fluctuation structure. the assessed analysis correlation Detrended,Detrended fluctuation analysis assessed the correlation structure.,15121
also The was of coefficient (CV) assessed. variation,The coefficient of variation (CV) was also assessed.,15122
"should induced gait by increased considered. However, be the VC variability","However, the increased gait variability induced by VC should be considered.",15123
for We the three wearable-camera testing collect of algorithm. proposed videos sets,We collect three sets of wearable-camera videos for testing the proposed algorithm.,15124
increasing interest vision from Proposals community. computer recent research technique is Object a receiving the,Object Proposals is a recent computer vision technique receiving increasing interest from the research community.,15125
Proposals field of techniques understanding in use text innovative. The is scene the Object,The use of Object Proposals techniques in the scene text understanding field is innovative.,15126
source publicly is available. The of made code method our,The source code of our method is made publicly available.,15127
"most denoising, are edge-preserving non-linear. signal filters, The smoothing for efficient e.g.,","The most efficient signal edge-preserving smoothing filters, e.g., for denoising, are non-linear.",15128
is filter dealt by freezing and restarting. careful non-linearity with The,The filter non-linearity is dealt with by careful freezing and restarting.,15129
segmentation an approach We for and recognition generative describe end-to-end the human of activities.,We describe an end-to-end generative approach for the segmentation and recognition of human activities.,15130
for data structured amount of models. sufficient when training available generative is,when sufficient amount of data is available for training structured generative models.,15131
called are tree These based networks. networks,These networks are called tree based networks.,15132
tree and are networks networks child-sibling based. all Reticulation-visible,Reticulation-visible networks and child-sibling networks are all tree based.,15133
"the action calculus. In theories paper, in investigate this situation bounded we","In this paper, we investigate bounded action theories in the situation calculus.",15134
classes bounded action We theories. discuss various of,We discuss various classes of bounded action theories.,15135
"form across quantification supports controlled Notably, situations. of variant a this","Notably, this variant supports a controlled form of quantification across situations.",15136
disadvantage Heterozygote potentially potent divergence. population a of driver is genetic,Heterozygote disadvantage is potentially a potent driver of population genetic divergence.,15137
components in task image central a multiple is computer into an Segmenting vision.,Segmenting an image into multiple components is a central task in computer vision.,15138
"components plausible knowledge is practical prior scenarios, available. about many In","In many practical scenarios, prior knowledge about plausible components is available.",15139
address with the semantic of images task We tuples. annotating,We address the task of annotating images with semantic tuples.,15140
each argument. is type output We best that for the representation of also specific conclude,We also conclude that the best output representation is specific for each type of argument.,15141
to paper This a the cross-domain problem. introduces new solve method recognition,This paper introduces a new method to solve the cross-domain recognition problem.,15142
"in source for we the labeled each of classes. samples construct subspaces given domain, Specifically,","Specifically, given labeled samples in source domain, we construct subspaces for each of the classes.",15143
the Comparison outperforms methods demonstrate datasets. results proposed both on comparison that method the,Comparison results demonstrate that the proposed method outperforms the comparison methods on both datasets.,15144
have devoted to pre-processing the Very efforts importance been the techniques. study of few,Very few efforts have been devoted to study the importance of the pre-processing techniques.,15145
"representation's mechanism physiological this unknown. However, details of remain many","However, many details of this representation's physiological mechanism remain unknown.",15146
the networks become primary learning recognition. have for neural Deep technique object,Deep neural networks have become the primary learning technique for object recognition.,15147
"makes the temporally which of non-trivial. deep application still are images, Videos, unlike networks coherent","Videos, unlike still images, are temporally coherent which makes the application of deep networks non-trivial.",15148
"object we recognition videos. Here, short how investigate in can aid motion","Here, we investigate how motion can aid object recognition in short videos.",15149
Long Short-Term (LSTM) approach is Our on networks. based deep Memory,Our approach is based on Long Short-Term Memory (LSTM) deep networks.,15150
"applications convolution. each we a gate previous LSTMs, of implement Unlike as","Unlike previous applications of LSTMs, we implement each gate as a convolution.",15151
computation size data analysis. sample Large brings the modern bottleneck for,Large sample size brings the computation bottleneck for modern data analysis.,15152
to is of this strategies handle Subsampling problem. one efficient,Subsampling is one of efficient strategies to handle this problem.,15153
"the we non-asymptotic estimator. consistency provide and result, asymptotic subsample normality of the Besides","Besides the non-asymptotic result, we provide asymptotic consistency and normality of the subsample estimator.",15154
subsampling are proposed our real data to and Synthetic study empirically used strategies.,Synthetic and real data are used to empirically study our proposed subsampling strategies.,15155
representations RL. approaches to have attempted for efficient Recent automatically discover,Recent approaches have attempted to automatically discover efficient representations for RL.,15156
a algorithm for space. structure examine greedy over searching We the,We examine a greedy algorithm for searching over the structure space.,15157
problems which improves positive stable definite This yields performance. regression dual,This yields dual stable positive definite problems which improves regression performance.,15158
"software, paintings. digital painting In organize layers","In digital painting software, layers organize paintings.",15159
"final painting. transmitted, not explicitly or are with layers published represented, However, digital the","However, layers are not explicitly represented, transmitted, or published with the final digital painting.",15160
technique We digital decompose into to painting propose a layers. a,We propose a technique to decompose a digital painting into layers.,15161
is decomposition Our painting's based geometry. the on RGB-space,Our decomposition is based on the painting's RGB-space geometry.,15162
suggest vertices pixels of paint colors. RGB-space in hull convex The the of,The vertices of the convex hull of pixels in RGB-space suggest paint colors.,15163
resulting We of for utility re-editing. decompositions the the demonstrate,We demonstrate the utility of the resulting decompositions for re-editing.,15164
"literature. compared work, with of number consider the classes previous this we relatively In larger","In this work, we consider relatively larger number of classes compared with the previous literature.",15165
"our accuracy system and To the the knowledge, best the achieves of highest speed.","To the best of our knowledge, the system achieves the highest accuracy and speed.",15166
our is model trained repository. and on dataset available The,The trained model and dataset is available on our repository.,15167
and paper runtime. for we derive estimates this In DFS average BFS,In this paper we derive estimates for average BFS and DFS runtime.,15168
basis more an for methods. search They may of the analysis advanced form also,They may also form the basis for an analysis of more advanced search methods.,15169
The treats tree search. and paper both graph search,The paper treats both tree search and graph search.,15170
empirical that verification come Experimental shows reality. to our close approximations analytical,Experimental verification shows that our analytical approximations come close to empirical reality.,15171
"by roles a challenge. remains functional played SNPs However, interpreting these the","However, interpreting the functional roles played by these SNPs remains a challenge.",15172
to use provides prototypes. an action the slow opportunity incremental feature analysis Applying,Applying incremental slow feature analysis provides an opportunity to use the action prototypes.,15173
path a essential rough The an of the in object signature is of paths. theory,The signature of a path is an essential object in the theory of rough paths.,15174
"recover data can signature the The standard stream statistics, representation of e.g.","The signature representation of the data stream can recover standard statistics, e.g.",15175
segmentation task. is object an under-explored important Instance-level yet,Instance-level object segmentation is an important yet under-explored task.,15176
"challenging. Nonetheless, proposals region accurate itself generating is quite","Nonetheless, generating accurate region proposals itself is quite challenging.",15177
ago a Fitness century by (IFT) proposed Theory half was Inclusive W.D.,Inclusive Fitness Theory (IFT) was proposed half a century ago by W.D.,15178
projections computation the Krylov eigendecomposition onto avoids suitable by costly subspaces. Polynomial of filtering,Polynomial filtering avoids costly computation of the eigendecomposition by projections onto suitable Krylov subspaces.,15179
"be e.g., Polynomial bilateral can filters. the filters based, guided and on","Polynomial filters can be based, e.g., on the bilateral and guided filters.",15180
"free. practical, is and method Our scalable model","Our method is practical, scalable and model free.",15181
E-T was The of instances hundreds tested in problems of particular several algorithm cases. and,The algorithm was tested in hundreds of instances of several E-T problems and particular cases.,15182
choose consist some to These pretreatment in based variables some and recoding on expertise. interactions,These pretreatment consist in recoding some variables and to choose some interactions based on expertise.,15183
"the dimension. of use selection one in high of variable Lasso, robust We methods","We use Lasso, one of the robust methods of variable selection in high dimension.",15184
avoid two used. is levels To cross-validation fitting a over,To avoid over fitting a two levels cross-validation is used.,15185
our We and examples justify theory. by matrix perturbation the intuitive method,We justify our method by intuitive examples and the matrix perturbation theory.,15186
generally the demonstrate state neuron exists separatrix space that of the models. We in,We demonstrate that the separatrix generally exists in the state space of neuron models.,15187
"accessible different evaluation settings: numbers. We discuss the and premise lemmas, ATPs,","We discuss the different evaluation settings: ATPs, accessible lemmas, and premise numbers.",15188
solving tools neural Convolutional becoming and visual object are standard for networks recognition tasks.,Convolutional neural networks are becoming standard tools for solving object recognition and visual tasks.,15189
"based most of implementation of trail-and-error. these complex on and are the However, models design","However, most of the design and implementation of these complex models are based on trail-and-error.",15190
"large with is kernels a classification networks considered. wide as general Specifically, with single-layer framework","Specifically, classification with wide single-layer networks with large kernels as a general framework is considered.",15191
learning itself. as is could Pool size as also be important procedure,Pool size is also could be as important as learning procedure itself.,15192
with Gaussian dimensional conditional include samplers This paper deals distributions. high that Gibbs,This paper deals with Gibbs samplers that include high dimensional conditional Gaussian distributions.,15193
"i.e. proved is to converge, algorithm The","The algorithm is proved to converge, i.e.",15194
are to the according drawn distribution. target the distributed asymptotically samples,the drawn samples are asymptotically distributed according to the target distribution.,15195
It methods. in some methods non-Gaussian semi-blind/unsupervised finds in direct as applications as well,It finds direct applications in semi-blind/unsupervised methods as well as in some non-Gaussian methods.,15196
an on for unsupervised illustration focused the methods. paper super-resolution The estimation provides,The paper provides an illustration focused on the unsupervised estimation for super-resolution methods.,15197
are amplitude of important oscillation. two and properties Frequency,Frequency and amplitude are two important properties of oscillation.,15198
network produces different Studied modes oscillation. two of,Studied network produces two different modes of oscillation.,15199
parameters/variables modes responsible of We identified oscillation. for the different key,We identified the key parameters/variables responsible for different modes of oscillation.,15200
modes flexible in the between appropriately parameters/variables. by network The choosing switching is required different,The network is flexible in switching between different modes by choosing appropriately the required parameters/variables.,15201
"Third, greyscale useful the using Un-sharp we in Masking method. image sharpen information resulting the","Third, we sharpen the useful information in the resulting greyscale image using Un-sharp Masking method.",15202
accuracy. can optical recognition significantly The and detection character improve approach text rate proposed,The proposed approach can significantly improve text detection rate and optical character recognition accuracy.,15203
"depend may in on mutant probability arises. structured the where However, fixation a this population,","However, in a structured population, this fixation probability may depend on where the mutant arises.",15204
epidemic only decisions information. algorithms Traditional detection using make local,Traditional epidemic detection algorithms make decisions using only local information.,15205
several explicitly fusion metapopulations. novel We a spatial from information that models propose approach,We propose a novel approach that explicitly models spatial information fusion from several metapopulations.,15206
regarding takes the considerations also announcement into cost-benefit method epidemic. Our account of,Our method also takes into account cost-benefit considerations regarding the announcement of epidemic.,15207
"adaptive, The alarms. false of strategy non-parametric detection probability detection balances delay vis-a-vis resulting optimally","The resulting adaptive, non-parametric detection strategy optimally balances detection delay vis-a-vis probability of false alarms.",15208
to proper predictions. model a crucial making is accurate time in characterize a Using series,Using a proper model to characterize a time series is crucial in making accurate predictions.,15209
demonstrating results performance are of approach. proposed Numerical provided the the,Numerical results are provided demonstrating the performance of the proposed approach.,15210
lot a recognition attention during last decades. drawn of has Fingerprint,Fingerprint recognition has drawn a lot of attention during last decades.,15211
and have used recognition past. fingerprint been algorithms for features the Different in,Different features and algorithms have been used for fingerprint recognition in the past.,15212
"representation scattering paper, In transform/network, for powerful used called recognition. image a is this","In this paper, a powerful image representation called scattering transform/network, is used for recognition.",15213
predefined network are Scattering transforms. its where network is wavelet filters and a architecture convolutional,Scattering network is a convolutional network where its architecture and filters are predefined wavelet transforms.,15214
"of After principal extraction (PCA). analysis is features, reduced component their applying dimensionality scattering by","After extraction of scattering features, their dimensionality is reduced by applying principal component analysis (PCA).",15215
"matching task. template used SVM for perform the end, is At recognition multi-class to the","At the end, multi-class SVM is used to perform template matching for the recognition task.",15216
"However, are optimal used in solutions on functions. dependent the fundamentally objective parameters the the","However, the optimal solutions are fundamentally dependent on the parameters used in the objective functions.",15217
range. maps their highly colour over provided perceptual Many vendors have by contrast uneven,Many colour maps provided by vendors have highly uneven perceptual contrast over their range.,15218
also maps may Colour features. the perceptual that false have appearance induce of discontinuities,Colour maps may also have perceptual discontinuities that induce the appearance of false features.,15219
"specific are maps rainbow and developed The cyclic colour linear, detail. for requirements diverging, in","The specific requirements for linear, diverging, rainbow and cyclic colour maps are developed in detail.",15220
images this To for colour maps support evaluating work presented. are two test,To support this work two test images for evaluating colour maps are presented.,15221
"construction of Finally, set of basis images colours the new presented. ternary for a are","Finally, a set of new basis colours for the construction of ternary images are presented.",15222
simple programs produce many can models variables. Even probabilistic infinitely with,Even simple probabilistic programs can produce models with infinitely many variables.,15223
classification. Existing their on convolutional shown success deep great have networks image (CNNs) neural,Existing deep convolutional neural networks (CNNs) have shown their great success on image classification.,15224
"such However, representation. explicit image important very generating are for dependencies","However, such dependencies are very important for generating explicit image representation.",15225
RNNs can be applied directly hardly General non-sequential on data.,General RNNs can hardly be directly applied on non-sequential data.,15226
"we proposed the Thus, hierarchical RNNs (HRNNs).","Thus, we proposed the hierarchical RNNs (HRNNs).",15227
"in respectively. it all argument an is accepted, one extensions, or is if eventually accepted","an argument is eventually accepted, if it is accepted in one or all extensions, respectively.",15228
"by presented. semantics are axiomatic ranking on semantics our counting Lastly, induced an the perspective","Lastly, an axiomatic perspective on the ranking semantics induced by our counting semantics are presented.",15229
Target Tracking Multiple is often tracking-by-detection within addressed the Online paradigm. (MTT),Online Multiple Target Tracking (MTT) is often addressed within the tracking-by-detection paradigm.,15230
"ambiguities or arise of errors. in Nevertheless, presence occlusions detection","Nevertheless, ambiguities arise in presence of occlusions or detection errors.",15231
"(e.g., ""Why questions hands?"") we test shake do The","The test questions (e.g., ""Why do we shake hands?"")",15232
a based used on The answering ConceptNet of version methods. spectral question,The question answering used a version of ConceptNet based on spectral methods.,15233
Large indicate areas potential subtests improvement. of among variations,Large variations among subtests indicate potential areas of improvement.,15234
sense. common strongly most is Comprehension subtest associated the with,Comprehension is the subtest most strongly associated with common sense.,15235
"on research AI. work continues Psychometric this Also, previous","Also, this work continues previous research on Psychometric AI.",15236
Parts an Expanded (e.g. Model for (EPM) recognizing We introduce attributes human,We introduce an Expanded Parts Model (EPM) for recognizing human attributes (e.g.,15237
"suit) actions short young, (e.g. wearing and hair,","young, short hair, wearing suit) and actions (e.g.",15238
consist is few to models contrast of (i.e. a in relatively which current This,This is in contrast to current models which consist of a relatively few (i.e.,15239
in image. an random it redundant ignores background and,it ignores redundant and random background in an image.,15240
attributes human and actions. our validate on We of three datasets recent method challenging,We validate our method on three recent challenging datasets of human attributes and actions.,15241
results convincing quantitative datasets. obtain qualitative on the three We and state-of-the-art,We obtain convincing qualitative and state-of-the-art quantitative results on the three datasets.,15242
"different faces, in etc.) persons major clothing, everyday Recognising challenges presents (occluded locations, photos","Recognising persons in everyday photos presents major challenges (occluded faces, different clothing, locations, etc.)",15243
"ones. existing challenging we limitations more the addition, propose benchmarks discuss of and In","In addition, we discuss the limitations of existing benchmarks and propose more challenging ones.",15244
results based a architecture. clustered We model on these spiking using a network explain,We explain these results using a spiking network model based on a clustered architecture.,15245
"on an the predicts the Moreover, model of dimensionality. bound existence upper","Moreover, the model predicts the existence of an upper bound on dimensionality.",15246
on and the trials. bounds empirical The depends of of duration estimation such number,The empirical estimation of such bounds depends on the number and duration of trials.,15247
The unknown. mechanism these observations common underlying remains,The common mechanism underlying these observations remains unknown.,15248
use of livers. then the otherwise We unobservable and metabolic model to properties quantify determine,We then use the model to determine and quantify otherwise unobservable metabolic properties of livers.,15249
of a Clustering parallel Linear Iterative We (SLIC) segmentation. Simple introduce the superpixel GPU implementation,We introduce a parallel GPU implementation of the Simple Linear Iterative Clustering (SLIC) superpixel segmentation.,15250
solving for useful Learning representations an unsupervised a instrument sparse learning feature is problem.,Learning sparse feature representations is a useful instrument for solving an unsupervised learning problem.,15251
"present n-MNIST. In datasets, handwritten paper, three this digit collectively called we labeled","In this paper, we present three labeled handwritten digit datasets, collectively called n-MNIST.",15252
already formalized New often proof concepts developments similar to involve assistant ones.,New proof assistant developments often involve concepts similar to already formalized ones.,15253
can combined advice. be internal external with further help Such,Such external help can be further combined with internal advice.,15254
dimensional high in are data. successful methods handling scale large based Vector quantization,Vector quantization based methods are successful in handling large scale high dimensional data.,15255
"computing perform methods datasets, distance excessive computations. these large However for","However for large datasets, these distance computing methods perform excessive computations.",15256
also cost. lower to Encoding We further the Forest(E-Forest) computation propose,We also propose Encoding Forest(E-Forest) to further lower the computation cost.,15257
quantization-based compatible is with E-Tree E-Forest methods. existing and various,E-Tree and E-Forest is compatible with various existing quantization-based methods.,15258
scale perform approximate Quantization large methods been to search tasks. introduced have nearest,Quantization methods have been introduced to perform large scale approximate nearest search tasks.,15259
quantization of one Vector Residual is (RVQ) the Quantization methods. effective,Residual Vector Quantization (RVQ) is one of the effective quantization methods.,15260
quantization by codebook multi-stage the scheme stage a to stage. RVQ error learning uses lower,RVQ uses a multi-stage codebook learning scheme to lower the quantization error stage by stage.,15261
performance quickly with The added gain diminishes stages.,The performance gain diminishes quickly with added stages.,15262
vector actually NP-hard. Encoding a RVQ with is,Encoding a vector with RVQ is actually NP-hard.,15263
a human presents wearing paper faces persons of This database spectacles. for,This paper presents a database of human faces for persons wearing spectacles.,15264
"to the valuable As be community. a can contribution this computer such, vision","As such, this can be a valuable contribution to the computer vision community.",15265
This note Jabbari al. presented by shows that the et Nooghabi results,This note shows that the results presented by Jabbari Nooghabi et al.,15266
"with analysis. proposed the framework support $(K,\epsilon)$-robustness We further a formal","We further support the proposed framework with a formal $(K,\epsilon)$-robustness analysis.",15267
periodically to monitored tracks safe transportation. ensure inspected and need Railroad to be,Railroad tracks need to be periodically inspected and monitored to ensure safe transportation.,15268
indispensable is Fine-tuned cellular replication nucleotide for pools the regulation DNA. of of faithful,Fine-tuned regulation of the cellular nucleotide pools is indispensable for faithful replication of DNA.,15269
DNA also damage information The recognition is repair genetic and by processes. safeguarded,The genetic information is also safeguarded by DNA damage recognition and repair processes.,15270
fast effective to encodings idea The allow generate to is approximation. distance,The idea is to generate effective encodings to allow fast distance approximation.,15271
than quantization is state-of-the-art. The error other lower,The quantization error is lower than other state-of-the-art.,15272
"search a efficient non-exhaustive HCLAE ANN-Search. (A-Tree), propose Aggregating-Tree we using perform to method Further,","Further, we propose Aggregating-Tree (A-Tree), a non-exhaustive search method using HCLAE to perform efficient ANN-Search.",15273
"state-of-the-art. magnitudes on ANN-Search to the achieves of compared speed-up tasks, A-Tree","A-Tree achieves magnitudes of speed-up on ANN-Search tasks, compared to the state-of-the-art.",15274
noise protein duplication of influences different how systematically genome components. the We investigate timing,We systematically investigate how the timing of genome duplication influences different protein noise components.,15275
"limitations. a this is these overcome paper, proposed new of version In EFIS to","In this paper, a new version of EFIS is proposed to overcome these limitations.",15276
work of focus image generation. on we In problem caption the this,In this work we focus on the problem of image caption generation.,15277
"Bayesian flexible Macau, factorization propose a data. and powerful for heterogeneous method We","We propose Macau, a powerful and flexible Bayesian factorization method for heterogeneous data.",15278
the Santos a of The is and modification algorithm Method. Proenka,The algorithm is a modification of the Santos and Proenka Method.,15279
"the the region. segment detected from We sclera, then eye out","We then segment out the sclera, from the detected eye region.",15280
"contour. segmented an sclera, the approximate segment we eyelid out From","From the segmented sclera, we segment out an approximate eyelid contour.",15281
corner candidates detector. are and Stephens obtained Eye corner Harris using,Eye corner candidates are obtained using Harris and Stephens corner detector.,15282
"eye the locate finally. the corners, post-pruning corner candidates a to of Eye We introduce","We introduce a post-pruning of the Eye corner candidates to locate the eye corners, finally.",15283
"as on database. well created been has The Yale, as algorithm databases tested our JAFFE","The algorithm has been tested on Yale, JAFFE databases as well as our created database.",15284
human We scheme facial evaluate landmarks and proposed localisation. the sheep both on,We evaluate the proposed scheme on both human and sheep facial landmarks localisation.,15285
"for of approach an characteristics this work, semantic-part we present representations. object analyzing In category","In this work, we present an approach for analyzing semantic-part characteristics of object category representations.",15286
"For experiments, for representation we sketch-based objects. recently use spatial our a category-epitome, proposed","For our experiments, we use category-epitome, a recently proposed sketch-based spatial representation for objects.",15287
convolutional a perform object How network (FCN) can detection? fully neural single on,How can a single fully convolutional neural network (FCN) perform on object detection?,15288
proposed We et. network spatial [Jaderberg the recently (SPN) transformer integrate,We integrate the recently proposed spatial transformer network (SPN) [Jaderberg et.,15289
digits classify to MNIST RNN-SPN sequences. in We cluttered use the,We use the RNN-SPN to classify digits in cluttered MNIST sequences.,15290
"a version SPN image. The skewed zoomed, the and outputs input rotated of","The SPN outputs a zoomed, rotated and skewed version of the input image.",15291
capacity. to Numerical artificial lighting extend tool have introduced the been simulation of codes,Numerical simulation codes of artificial lighting have been introduced to extend the tool capacity.,15292
for original these model calculations The semi-detailed (simplified). and used is,The model used for these calculations is original and semi-detailed (simplified).,15293
having reliability a of idea the of numerical precise This about simulations. results allowed the,This allowed having a precise idea about the reliability of the results of numerical simulations.,15294
non-smooth sparsity allow parameter We for in non-convex to penalties induce regularisation weights.,We allow for non-smooth non-convex regularisation penalties to induce sparsity in parameter weights.,15295
apply focus generally. We deep learning on logistic more although our methods multiclass,We focus on deep multiclass logistic learning although our methods apply more generally.,15296
"with we research. Finally, for future directions conclude","Finally, we conclude with directions for future research.",15297
"image-set To approach. this, querying an propose based localization we address","To address this, we propose an image-set querying based localization approach.",15298
proposed and shown effectiveness feasibility have Experiments approach. the of the,Experiments have shown the effectiveness and feasibility of the proposed approach.,15299
between learning features discriminative for Extracted multiple interactions within are framework recognizing people. a utilized,Extracted multiple features are utilized within a discriminative learning framework for recognizing interactions between people.,15300
and by investigate focal injury seizures acquired induced in epileptogenesis the head to rat.,to investigate focal seizures and acquired epileptogenesis induced by head injury in the rat.,15301
genome. move can around that DNA are elements different sequences Transposable to positions in the,Transposable elements are DNA sequences that can move around to different positions in the genome.,15302
"can an lead genome size. process, increase and cause During this mutations, to in they","During this process, they can cause mutations, and lead to an increase in genome size.",15303
"Despite genomic have no transposable biological function. large fraction, a representing elements clear","Despite representing a large genomic fraction, transposable elements have no clear biological function.",15304
tested is seven videos proposed on and with state-of-the-art method The compared several methods.,The proposed method is tested on several videos and compared with seven state-of-the-art methods.,15305
further robustness and proposed method. of support have the the effectiveness to results provided Our,Our results have provided further support to the effectiveness and robustness of the proposed method.,15306
on possible. based proposed the are results real-time indicate techniques that implementations Our,Our results indicate that real-time implementations based on the proposed techniques are possible.,15307
b) an How likes look works? ontology how it and,b) How an ontology look likes and how it works?,15308
it ontologies existing is consider reusing to not? required if Verify c) or,c) Verify if it is required to consider reusing existing ontologies or not?,15309
developed? the c) What is the of be complexity ontology to,c) What is the complexity of the ontology to be developed?,15310
of What design ontology are the d) development? and principles,d) What are the principles of ontology design and development?,15311
above. key This answers the all paper questions,This paper answers all the key questions above.,15312
"that framework. Here, we three our methods follow introduce","Here, we introduce three methods that follow our framework.",15313
arguments discuss explain theoretical implications. this to provide its practical result and We,We provide theoretical arguments to explain this result and discuss its practical implications.,15314
"heatmaps performance. we the Finally, unsupervised network investigate use of of assessment for neural","Finally, we investigate the use of heatmaps for unsupervised assessment of neural network performance.",15315
"main large-scale are the make Indexing approaches. To Estimation search and feasible, Subset Distance","To make large-scale search feasible, Distance Estimation and Subset Indexing are the main approaches.",15316
"relative their not However, has Subset performance been Indexing. for studied","However, their relative performance has not been studied for Subset Indexing.",15317
binary better strategies. investigate or n-ary works We retrieval under coding different whether,We investigate whether binary or n-ary coding works better under different retrieval strategies.,15318
new dictionary approach this a learning. present kernels paper we incorporating into of In,In this paper we present a new approach of incorporating kernels into dictionary learning.,15319
"automatically effective by an proposed method is detected training the Moreover, landmarks. disturbing","Moreover, an effective training method is proposed by disturbing the automatically detected landmarks.",15320
"for rare arise, nest-structured in dynamics a evolutionary example, can population migration. with These","These evolutionary dynamics can arise, for example, in a nest-structured population with rare migration.",15321
strength any selection. hold Our for of results,Our results hold for any strength of selection.,15322
intensive of is human the today. connectome research object an The,The human connectome is the object of an intensive research today.,15323
the of discovering directions question of is One the the field main edges.,One main question of the field is discovering the directions of the edges.,15324
"we $(v,u)$ from Then to edge $u$. $v$ direct this","Then we direct this $(v,u)$ edge from $v$ to $u$.",15325
Traditional characters. suffer character handwritten in Chinese similar for approaches recognition classifying,Traditional approaches for handwritten Chinese character recognition suffer in classifying similar characters.,15326
"field, medical are to diagnosis used the of increasingly images facilitate In diseases.","In the medical field, images are increasingly used to facilitate diagnosis of diseases.",15327
have patches based this solve been methods problem. to Some proposed,Some patches based methods have been proposed to solve this problem.,15328
the for departure this question This provides chapter. point,This question provides the departure point for this chapter.,15329
is games a variations. family Poker many card that of includes,Poker is a family of card games that includes many variations.,15330
We use two to training therefore propose sources. independent,We therefore propose to use two independent training sources.,15331
many in found Mixture uses areas. have models,Mixture models have found uses in many areas.,15332
"To list models. and latent unsupervised empirical trait a learning, few: class Bayes,","To list a few: unsupervised learning, empirical Bayes, latent class and trait models.",15333
considering carefully only scoring mechanism formulated cases visible. are partially is The faces where challenging,The scoring mechanism is carefully formulated considering challenging cases where faces are only partially visible.,15334
"speed. of despite that the DCN, achieve show our We runtime can practical use network","We show that despite the use of DCN, our network can achieve practical runtime speed.",15335
LBP cost The computational the property and their very features texture low. discriminative possess is,The LBP features possess the texture discriminative property and their computational cost is very low.,15336
compared. curves the and are results K-fold ROC obtained and proposed For approach are,For the proposed approach K-fold and ROC curves are obtained and results are compared.,15337
"these R package in are available on CRAN. implemented nonparametric Several of tests the spTest,","Several of these nonparametric tests are implemented in the R package spTest, available on CRAN.",15338
of Validation methods of importance. segmentation image critical is,Validation of image segmentation methods is of critical importance.,15339
is in image increasingly the as Probabilistic results. popular captures it segmentation uncertainty,Probabilistic image segmentation is increasingly popular as it captures uncertainty in the results.,15340
of segmentation robust measure These probabilistic accuracy. a accurate multi-region and provide,These provide a robust and accurate measure of multi-region probabilistic segmentation accuracy.,15341
"fully an graph, and connected both incorporating image Attribute-Graph is global characteristics. An local undirected","An Attribute-Graph is an undirected fully connected graph, incorporating both local and global image characteristics.",15342
problem demonstrate ranking. of applying the effectiveness the by We of them to image Attribute-Graphs,We demonstrate the effectiveness of Attribute-Graphs by applying them to the problem of image ranking.,15343
demonstrate its examples. We simulations and through applicability,We demonstrate its applicability through simulations and examples.,15344
to learn proposed methods been descriptors. data-dependent binary have Various,Various methods have been proposed to learn data-dependent binary descriptors.,15345
"properties multiple computed is the are Secondly, RMGD extracted. binary image from strings where","Secondly, the RMGD is computed from multiple image properties where binary strings are extracted.",15346
"and Prior statistical, studied iterative, techniques. has problem using representation work this sparse algebraic, low-rank","Prior work has studied this problem using algebraic, iterative, statistical, low-rank and sparse representation techniques.",15347
understanding for a is fundamental Scene environmental today's classification task in perception robotics.,Scene classification is a fundamental perception task for environmental understanding in today's robotics.,15348
a with segmentation. implemented of is This regularization semantic,This is implemented with a regularization of semantic segmentation.,15349
to connect the perceive world. language and vision Humans,Humans connect language and vision to perceive the world.,15350
similar for to build a connection computers? How,How to build a similar connection for computers?,15351
"Thus, other for vision is ISR helpful unclear is whether still tasks. it","Thus, it is still unclear whether ISR is helpful for other vision tasks.",15352
combines model of instead single multiple simultaneously individually. models framework ensemble a The ensemble using,The framework combines multiple ensemble models simultaneously instead of using a single ensemble model individually.,15353
benchmark challenging sequences. on proposed tracker We the image test,We test the proposed tracker on challenging benchmark image sequences.,15354
"boost use performance. regression resampled partial set Furthermore, we to on a least squares training","Furthermore, we use partial least squares regression on a resampled training set to boost performance.",15355
complex algorithm solution robust tracking object in is A scenes. for video proposed an,A robust algorithm solution is proposed for tracking an object in complex video scenes.,15356
efficacy the Experimental of validate radish approach. publicly the results available using proposed dataset the,Experimental results using the publicly available radish dataset validate the efficacy of the proposed approach.,15357
of include images. accessory These large clothing numbers and,These include large numbers of clothing and accessory images.,15358
survey synchronization theory noisy oscillators. of We in collections the,We survey the theory synchronization in collections of noisy oscillators.,15359
synchronization hydrodynamic flagellar applied framework to interactions. This by is,This framework is applied to flagellar synchronization by hydrodynamic interactions.,15360
"synchronization, break physical in for ways. We discuss flagellar different which symmetry different this mechanisms","We discuss different physical mechanisms for flagellar synchronization, which break this symmetry in different ways.",15361
diverse are abundant Chromatin remodeling machineries and eukaryotic in cells.,Chromatin remodeling machineries are abundant and diverse in eukaryotic cells.,15362
results us segmentation to benchmark enables on feature achieve state-of-the-art This datasets.,This feature enables us to achieve state-of-the-art segmentation results on benchmark datasets.,15363
chains if the These reactivated can considers algorithm necessary. extinct be,These extinct chains can be reactivated if the algorithm considers necessary.,15364
the scheme. benefits Numerical the simulations of shows novel,Numerical simulations shows the benefits of the novel scheme.,15365
in been in These full previous addressed not have generality work.,These have not been addressed in full generality in previous work.,15366
step in two images. or matching estimating from key the depth more Stereo is,Stereo matching is the key step in estimating depth from two or more images.,15367
"have performance. achieved Recently, some methods non-local been which proposed, state-of-the-art matching tree-based stereo","Recently, some tree-based non-local stereo matching methods have been proposed, which achieved state-of-the-art performance.",15368
MST. our because of partially is the overcome extreme the This DTs greediness,This is partially because our DTs overcome the extreme greediness of the MST.,15369
the Extensive on our demonstrate of experimental and benchmark effectiveness efficiency datasets some results framework.,Extensive experimental results on some benchmark datasets demonstrate the effectiveness and efficiency of our framework.,15370
give we Theory. on Bayesian the In outline an paper Decision this,In this paper we give an outline on the Bayesian Decision Theory.,15371
ARS high rejection rates. from with sampling independent generates via samples target acceptance the,ARS generates independent samples from the target via rejection sampling with high acceptance rates.,15372
"each sampling computational pdf updated. time demanding is becomes However, the it from proposal more","However, sampling from the proposal pdf becomes more computational demanding each time it is updated.",15373
"faster CARS than For is of a number large generating samples, desired ARS.","For generating a large number of desired samples, CARS is faster than ARS.",15374
a such properties filter. univariate of generalise Many filters median to,Many properties of univariate median filters generalise to such a filter.,15375
the demonstrate Numerical results. are the that experiments approximation validity presented of,Numerical experiments are presented that demonstrate the validity of the approximation results.,15376
"cells encode regulatory diverse surprisingly, to noise. Not buffer mechanisms","Not surprisingly, cells encode diverse regulatory mechanisms to buffer noise.",15377
feedforward the One such incoherent mechanism circuit. is,One such mechanism is the incoherent feedforward circuit.,15378
propose which the scalable stochastic gradient algorithm makes based method learning (w.r.t. a We,We propose a stochastic gradient based learning algorithm which makes the method scalable (w.r.t.,15379
"nonlinear. the examples), of training while number being","the number of training examples), while being nonlinear.",15380
trackers years. Tracking-by-detection recent are the most approaches successful object some in of,Tracking-by-detection approaches are some of the most successful object trackers in recent years.,15381
"solution. two algorithms Then, are proposed for the","Then, two algorithms are proposed for the solution.",15382
"search local TSP-D is it a generated, been After has solution improved through operators. then","After a TSP-D solution has been generated, it is then improved through local search operators.",15383
unmixing. normal model (NCM) has compositional used been in The hyperspectral extensively,The normal compositional model (NCM) has been extensively used in hyperspectral unmixing.,15384
"has their focused variability. previous However, of the and/or on of research estimation most endmembers","However, most of the previous research has focused on estimation of endmembers and/or their variability.",15385
"NCM. work Also, information little employed has spatial in","Also, little work has employed spatial information in NCM.",15386
be run. to to easy implement turns and The to out resulting algorithm efficient,The resulting algorithm turns out to be easy to implement and efficient to run.,15387
compared and state-of-the-art synthetic SCM current real We with on images. algorithms,We compared SCM with current state-of-the-art algorithms on synthetic and real images.,15388
the that accurate and endmembers SCM abundances. main in results more show The provide can,The results show that SCM can in the main provide more accurate endmembers and abundances.,15389
"a as endmember under uncertainty of prediction the Moreover, error can serve certain estimated conditions.","Moreover, the estimated uncertainty can serve as a prediction of endmember error under certain conditions.",15390
approach on dataset. evaluate our the We MNIST,We evaluate our approach on the MNIST dataset.,15391
several algorithms in immense possibilities of offer the cognitive Machine-learning applications. development,Machine-learning algorithms offer immense possibilities in the development of several cognitive applications.,15392
"In this War. Square game: paper, we present a board","In this paper, we present a board game: Square War.",15393
to Chinese War similar Go. board definition classic Square game is game the of The,The game definition of Square War is similar to the classic Chinese board game Go.,15394
Then game mathematical Square problem of War. we propose the a,Then we propose a mathematical problem of the game Square War.,15395
ding fundamental in cue Motion for and analysis a human activity scene understan- is videos.,Motion is a fundamental cue for scene analysis and human activity understan- ding in videos.,15396
can only be approach Each scenarios. applied limited on,Each approach can only be applied on limited scenarios.,15397
Gibbs learning. Monte often Markov Carlo a used Bayesian (MCMC) is sampling Chain method in,Gibbs sampling is a Markov Chain Monte Carlo (MCMC) method often used in Bayesian learning.,15398
"sequential We requirements. sampling, parallelism by asynchronous Gibbs achieves which study simply ignoring","We study asynchronous Gibbs sampling, which achieves parallelism by simply ignoring sequential requirements.",15399
to analysis. a Moving detection intelligent object is key video,Moving object detection is a key to intelligent video analysis.,15400
"the rich On prone objects to be texture hand, other without detected. moving not are","On the other hand, moving objects without rich texture are prone not to be detected.",15401
seek An algorithm to optimal minimization alternative proposed the solutions. is,An alternative minimization algorithm is proposed to seek the optimal solutions.,15402
a Sequential sampling framework. optimization the based global on Carlo (SMC) algorithm Monte We propose,We propose a global optimization algorithm based on the Sequential Monte Carlo (SMC) sampling framework.,15403
demonstrates result our The algorithm. appealing properties the of,The result demonstrates the appealing properties of our algorithm.,15404
"draw constant calculate posterior can normalization i.i.d. the Here, we tractably and","Here, we can tractably calculate the normalization constant and draw posterior i.i.d.",15405
"we the attain data, risk we Bayes simulations. simulated in how can With demonstrate","With simulated data, we demonstrate how we can attain the Bayes risk in simulations.",15406
of regions of different networks find molecular (NA) between Network alignment to aims similarities species.,Network alignment (NA) aims to find regions of similarities between molecular networks of different species.,15407
NA global (GNA). two exist There categories: or local (LNA),There exist two NA categories: local (LNA) or global (GNA).,15408
mapping. small finds a produces and conserved LNA many-to-many network node highly regions,LNA finds small highly conserved network regions and produces a many-to-many node mapping.,15409
conserved and finds regions mapping. one-to-one GNA node produces a large,GNA finds large conserved regions and produces a one-to-one node mapping.,15410
"to which LNA So, or GNA? one choose,","So, which one to choose, LNA or GNA?",15411
"NA systematic introduce first of answer evaluation To the we this, categories. two the","To answer this, we introduce the first systematic evaluation of the two NA categories.",15412
for evaluation efficient implements provide existing user-friendly and software alignment that new We the measures.,We provide user-friendly software for efficient alignment evaluation that implements the new and existing measures.,15413
and synthetic We on and evaluate real-world prominent biological networks. methods LNA GNA,We evaluate prominent LNA and GNA methods on synthetic and real-world biological networks.,15414
quality using interaction study levels. the on alignment of different confidence We and types effect,We study the effect on alignment quality of using different interaction types and confidence levels.,15415
that one other category of over context-dependent. NA the find is We the superiority,We find that the superiority of one NA category over the other is context-dependent.,15416
guidelines software method NA Our for development and evaluation. results future and provide,Our results and software provide guidelines for future NA method development and evaluation.,15417
robust general for approaches moving not captured cameras. dynamic scenes sparse with are These,These approaches are not robust for general dynamic scenes captured with sparse moving cameras.,15418
multiple improved demonstrates reconstruction. both state-of-the-art Comparison with accuracy in and approaches view dense segmentation,Comparison with state-of-the-art approaches demonstrates improved accuracy in both multiple view segmentation and dense reconstruction.,15419
structure proposed appearance. prior approach requirement also eliminates of for scene The and the knowledge,The proposed approach also eliminates the requirement for prior knowledge of scene structure and appearance.,15420
"perspective conceptual In the the paper, discriminative study CNN. this of generative we","In this conceptual paper, we study the generative perspective of the discriminative CNN.",15421
One in visual represented. is how of object crucial the problems tracking the is,One of the crucial problems in visual tracking is how the object is represented.,15422
are increasingly robust. to appearance-based Conventional trackers be in features complex using order more,Conventional appearance-based trackers are using increasingly more complex features in order to be robust.,15423
to from It background. also discriminative distinguish information foreground the incorporates,It also incorporates the discriminative information to distinguish foreground from background.,15424
"dictionary an DNBS bases problem the of finding over-complete NP-hard. is the from However,","However, the problem of finding the DNBS bases from an over-complete dictionary is NP-hard.",15425
through object SSD-based tracking applied proposed is representation to template DNBS The matching.,The proposed DNBS representation is applied to object tracking through SSD-based template matching.,15426
"human of interest topic, Recognition observational nowadays. low is a latency, under growing actions,","Recognition of human actions, under low observational latency, is a growing interest topic, nowadays.",15427
"Small as trees such approaches. parallel ensemble variations occur of MCTS, and in search","Small search trees occur in variations of MCTS, such as parallel and ensemble approaches.",15428
As becomes improved we an smaller trees performance. achieve the search,As the search trees becomes smaller we achieve an improved performance.,15429
of parallelism are large The results scale performance for MCTS. the important improving of,The results are important for improving the performance of large scale parallelism of MCTS.,15430
termed typically Such realisations diffusion bridges. are,Such realisations are typically termed diffusion bridges.,15431
"In proposed partial observation construct noisy a regime. recently adapt addition, and we a to","In addition, we adapt a recently proposed construct to a partial and noisy observation regime.",15432
Numerical the effects are carried of prevalence. disease residence to highlight out simulations times on,Numerical simulations are carried out to highlight the effects of residence times on disease prevalence.,15433
graph an random in analysis data. models tool the important Exponential are of statistical,Exponential random graph models are an important tool in the statistical analysis of data.,15434
real-world the illustrate analysis through of graphs. We methodology our,We illustrate our methodology through the analysis of real-world graphs.,15435
statistical Latent a models natural clustering. provide space and for framework graph visualizing,Latent space models provide a natural statistical framework for graph visualizing and clustering.,15436
an proposed method. part is regression essential step scaling The of novel the,The novel regression scaling step is an essential part of the proposed method.,15437
The described stochastic via evolution protein of is expression. a simple gene model,The protein evolution is described via a simple stochastic model of gene expression.,15438
the time. first-passage are formulas developed and of moments distribution for Exact analytical the,Exact analytical formulas are developed for the distribution and moments of the first-passage time.,15439
"discussed. are the of Lastly, this finding implications","Lastly, the implications of this finding are discussed.",15440
"them performance. it to Further, and improves combined complexity features with is higher when complementary","Further, it is complementary to higher complexity features and when combined with them improves performance.",15441
video. the is The in of assumes shown that the image background the algorithm majority,The algorithm assumes that the background image is shown in the majority of the video.,15442
"be cannot fulfilled. this However, constraint always","However, this constraint cannot always be fulfilled.",15443
be the derived. similarity From actions resulting appropriate can,From the resulting similarity appropriate actions can be derived.,15444
existing human advantage via perception We methods a demonstrate study. against its,We demonstrate its advantage against existing methods via a human perception study.,15445
also algorithm which is two-stage We robust model-mismatch recovery practical noise. to propose a and,We also propose a practical two-stage recovery algorithm which is robust to model-mismatch and noise.,15446
from convolutional DCNN. a The one layers uses first directly,The first one directly uses convolutional layers from a DCNN.,15447
as layer. the convolutional that on then pooling-guidance layer train We another convolutional of top,We then train another convolutional layer on top of that as the pooling-guidance convolutional layer.,15448
Scientists habit. using natural in these are to study increasingly their technologies creatures imaging-based,Scientists are increasingly using imaging-based technologies to study these creatures in their natural habit.,15449
"the human often below Methods accuracy is experts. exist, classification automated of for that but","Methods for automated classification exist, but the accuracy is often below that of human experts.",15450
for enhancement auxiliary car important transport systems. outdoor vision contrast is smart Image for,Image contrast enhancement for outdoor vision is important for smart car auxiliary transport systems.,15451
weather often by characterized visibility. The in conditions poor video frames are captured poor,The video frames captured in poor weather conditions are often characterized by poor visibility.,15452
wrong estimation. using threshold a cause may hard a Simply,Simply using a hard threshold may cause a wrong estimation.,15453
regions. images are in noise exposure dark good characterized enhanced and with little The,The enhanced images are characterized with little noise and good exposure in dark regions.,15454
The are enhanced processed images and edges also of textures significantly. the,The textures and edges of the processed images are also enhanced significantly.,15455
make and transformation FstCN operator novel possible. factorization We to in introduce permutation a,We introduce a novel transformation and permutation operator to make factorization in FstCN possible.,15456
Protein function its closely structure. sequence and and are related to dynamics,Protein function and dynamics are closely related to its sequence and structure.,15457
"classifier. propose this based To (MTF-SVM) fingerprint vector topological a machine support end, we molecular","To this end, we propose a molecular topological fingerprint based support vector machine (MTF-SVM) classifier.",15458
"of the MTF-SVM consider validate problems. To four present types approach, we","To validate the present MTF-SVM approach, we consider four types of problems.",15459
computational topology classification. present and as The protein independent study alternative for effective establishes an,The present study establishes computational topology as an independent and effective alternative for protein classification.,15460
"inappropriate devices. are mobile and systems for wearable these Consequently,","Consequently, these systems are inappropriate for mobile and wearable devices.",15461
hand also We using each pose frame classification. initialize at and detection finger,We also initialize hand pose at each frame using finger detection and classification.,15462
"death. in deletions) mutations, cell or results amplifications by","by mutations, amplifications or deletions) results in cell death.",15463
detected remaining instantiate used tracking. The as new objects are objects to information for,The remaining detected objects are used as information to instantiate new objects for tracking.,15464
events. parameters occlusion the how present adjust to we Finally,Finally we present how the parameters adjust to occlusion events.,15465
also of system possible adaptive used recovery. The is property for occlusion the,The adaptive property of the system is also used for possible occlusion recovery.,15466
"exponentially does not hyperbolically. is increase shown but population It if that, unchecked,","It is shown that, if unchecked, population does not increase exponentially but hyperbolically.",15467
transitions trajectories. were first The hyperbolic sustained between two,The first two transitions were between sustained hyperbolic trajectories.,15468
transition is current an to trajectory. unknown The,The current transition is to an unknown trajectory.,15469
The and oscillations high-gamma populations. model all low- visible exhibits in,The model exhibits low- and high-gamma oscillations visible in all populations.,15470
We derive error their forms. asymptotic functions and formulate new,We formulate new error functions and derive their asymptotic forms.,15471
in networks. two-layered Calculation of is the demonstrated error Bayesian functions,Calculation of the error functions is demonstrated in two-layered Bayesian networks.,15472
data this theory. contradiction These found are of examined to in be clear and closely,These data are closely examined and found to be in clear contradiction of this theory.,15473
Demographic best is data England. available Transition the contradicted Theory also for by,Demographic Transition Theory is also contradicted by the best available data for England.,15474
Other contradicting discussed. are examples of also evidence,Other examples of contradicting evidence are also discussed.,15475
by performance. We its the begin analyzing describing method and proposed,We begin by describing the proposed method and analyzing its performance.,15476
"theory network. a using validate Further, our we experimentally four-camera","Further, we experimentally validate our theory using a four-camera network.",15477
objects images. high paper detect to present for this resolution we In methods effective,In this paper we present effective methods to detect objects for high resolution images.,15478
boxes approach visual first features. based adjacent on to bounding The is predict,The first approach is to predict bounding boxes based on adjacent visual features.,15479
It is which variability characterise the normal to this seems state.,It is this variability which seems to characterise the normal state.,15480
standard for and modelling. analysis data purposes set We used a of experimental the,We used a standard experimental data set for the purposes of analysis and modelling.,15481
each separate of be will in these investigations of explored techniques Further publications.,Further investigations of each of these techniques will be explored in separate publications.,15482
indicate promise challenging this our solution experiments of the . settin Extensive proposed unsupervised in,Extensive experiments indicate the promise of our proposed solution in this challenging unsupervised settin .,15483
from mapping prediction onto of HA alone. tree This sequence antigenicity data the allows,This mapping onto the tree allows prediction of antigenicity from HA sequence data alone.,15484
is problems. a combinatorial oriented search programming paradigm Answer towards programming difficult set declarative,Answer set programming is a declarative programming paradigm oriented towards difficult combinatorial search problems.,15485
are programs perform Answer set task. this the that solvers,Answer set solvers are the programs that perform this task.,15486
"and this solvers DLV, general GnT then captures We how template illustrate Cmodels.","We then illustrate how this general template captures solvers DLV, GnT and Cmodels.",15487
method novel object presents (LLC). on This tracking Coding Linear paper based a approximated Locality-constrained,This paper presents a novel object tracking method based on approximated Locality-constrained Linear Coding (LLC).,15488
from in various these dictionaries of Weights are the learned LLC also framework. approximated similar,Weights of these various dictionaries are also learned from approximated LLC in the similar framework.,15489
"order In tracking method. detection occlusion a drifts, and alleviate we propose efficient simple to","In order to alleviate tracking drifts, we propose a simple and efficient occlusion detection method.",15490
incorporates relevant fusion ensemble increases our information accuracy. late contextual This and classification,This late fusion ensemble incorporates relevant contextual information and increases our classification accuracy.,15491
also results Additional distributions Normal bandits are provided. of on,Additional results on bandits of Normal distributions are also provided.,15492
"drive poorly to atherosclerosis dynamics its health, that significance public remain understood. the Despite","Despite its significance to public health, the dynamics that drive atherosclerosis remain poorly understood.",15493
existing new patch-based as special arise algorithm. of Many the cases algorithms,Many existing patch-based algorithms arise as special cases of the new algorithm.,15494
"way, estimated. To must transmission medium utilize the be this corresponding matrix its in","To utilize the medium in this way, its corresponding transmission matrix must be estimated.",15495
"field greatly science and fields information various emerged. The developed, have in applications has of","The field of information science has greatly developed, and applications in various fields have emerged.",15496
data contribute to applications. techniques and other processing may This framework theoretical,This theoretical framework may contribute to data processing techniques and other applications.,15497
and objects a geometric in very cartoon sharp have image surface The smooth edges. the,The geometric objects in the cartoon image have a very smooth surface and sharp edges.,15498
to heterogeneous ontology an DKP-AOM is designed tool merge ontologies. merging,DKP-AOM is an ontology merging tool designed to merge heterogeneous ontologies.,15499
"is DKP-AOM two with participating DKP-AOM_lite), performs and versions DKP-AOM (DKP-AOM coherence analysis.","DKP-AOM is participating with two versions (DKP-AOM and DKP-AOM_lite), DKP-AOM performs coherence analysis.",15500
"the we towards future some of development work discuss Finally, DKP-AOM.","Finally, we discuss some future work towards the development of DKP-AOM.",15501
habenula by study. axonal DTI tracts Hippocampus to medial my multisynaptic was linked in,Hippocampus was by multisynaptic axonal tracts linked to medial habenula in my DTI study.,15502
"and to amygdala hippocampus septum, to tracts BNST. These linked","These tracts linked hippocampus to septum, and amygdala to BNST.",15503
to the The response. is alternative with best the goal identify highest,The goal is to identify the best alternative with the highest response.,15504
each response the We of use to predict alternative. Bayesian regression logistic,We use Bayesian logistic regression to predict the response of each alternative.,15505
of datasets benchmark the the Experiments effectiveness UCI proposed method. demonstrate on,Experiments on benchmark UCI datasets demonstrate the effectiveness of the proposed method.,15506
WordNet hierarchy. shared Constraints the across labels are the through semantic,Constraints are shared across the labels through the semantic WordNet hierarchy.,15507
learning as two-stage problem. a this We pose,We pose this as a two-stage learning problem.,15508
compare to our state-of-the-art of et al. method approach the We Bell,We compare our method to the state-of-the-art approach of Bell et al.,15509
domain labeled new a data. target propose exploit sparsely to We unlabeled architecture and CNN,We propose a new CNN architecture to exploit unlabeled and sparsely labeled target domain data.,15510
"of (DCNNs) convolutional applications. recognition neural have range networks Meanwhile, revolutionized broad a object deep","Meanwhile, deep convolutional neural networks (DCNNs) have revolutionized a broad range of object recognition applications.",15511
"we apply DCNNs this to work, recognition. In logo","In this work, we apply DCNNs to logo recognition.",15512
element key models. augmentation Data high-dimensional training a in is,Data augmentation is a key element in training high-dimensional models.,15513
nets demonstrate training We improvements performance neural over significant manually-specified augmentation in schemes. deep,We demonstrate significant performance improvements in training deep neural nets over manually-specified augmentation schemes.,15514
datasets available code augmented Our are and online.,Our code and augmented datasets are available online.,15515
"invoked to data. to for biological account it be has various Nevertheless,","Nevertheless, it has to be invoked to account for various biological data.",15516
for The is method the emotions of individuals matrix analysis presented. basal innovational,The innovational matrix method of the basal emotions analysis for individuals is presented.,15517
matrix criteria The state the of been psycho-emotional have stability formulated.,The matrix criteria of the psycho-emotional state stability have been formulated.,15518
"understood. progression system is the between relationship tumor and still poorly nervous However, the","However, the relationship between tumor progression and the nervous system is still poorly understood.",15519
overcomes Biased Location network this novel limitation Our incorporating Convolutional layer. by a,Our network overcomes this limitation by incorporating a novel Location Biased Convolutional layer.,15520
"this paper, that face inaugurate promising recognition dogs. by In for we avenue pursuing","In this paper, we inaugurate that promising avenue by pursuing face recognition for dogs.",15521
display structural Some proteins patterns. recurrent natural,Some natural proteins display recurrent structural patterns.,15522
tracking requires trajectories. associations existing multi-person online online detection with correct the responses Robust of,Robust online multi-person tracking requires the correct associations of online detection responses with existing trajectories.,15523
challenging. high large-scale Annotating quality is datasets,Annotating large-scale high quality datasets is challenging.,15524
"environment. real-world diversity; not these reflect do datasets operating Typically, have the they limited","Typically, these datasets have limited diversity; they do not reflect the real-world operating environment.",15525
predator show a stabilizing as expected. cannibalism We influence provides in that the,We show that cannibalism in the predator provides a stabilizing influence as expected.,15526
and this distorts called is This phenomenon market. the asymmetry information,This phenomenon is called information asymmetry and this distorts the market.,15527
"intelligence, screening made the signaling With of artificial have been and advent easier.","With the advent of artificial intelligence, signaling and screening have been made easier.",15528
on impact of the of artificial studies the theory asymmetric This intelligence information. paper,This paper studies the impact of artificial intelligence on the theory of asymmetric information.,15529
"lampshade the realized surface. thin over the as tubes distributed holes Thus, are","Thus, the holes are realized as thin tubes distributed over the lampshade surface.",15530
probability chain Carlo asymptotically Markov sample (MCMC) complex Monte distributions. methods from,Markov chain Monte Carlo (MCMC) methods asymptotically sample from complex probability distributions.,15531
sampling methods. slice adaptive we other framework this can use and In,In this framework we can use slice sampling and other adaptive methods.,15532
"more chains, more often mix We obtain quickly. Markov which robust","We obtain more robust Markov chains, which often mix more quickly.",15533
"features captures it practical pre-tagging. models sharp Secondly, any on without","Secondly, it captures sharp features on practical models without any pre-tagging.",15534
state art out-perform benchmark methods two the of on datasets. the significantly Our,Our methods significantly out-perform the state of the art on two benchmark datasets.,15535
"cataract, cortical three and types cataract, nuclear cataract: of sub-capsular There cataract. senile are posterior/anterior","There are three types of senile cataract: cortical cataract, nuclear cataract, and posterior/anterior sub-capsular cataract.",15536
cataract. senile most cataract common is The cortical,The most common senile cataract is cortical cataract.,15537
"analyzed. understanding cataract are in cortical factors For the causing cataract, pathology the and cortical","For understanding cortical cataract, the pathology and the causing factors in cortical cataract are analyzed.",15538
crystallin The disrupted proteins. water will and release fiber,The disrupted fiber will release water and crystallin proteins.,15539
"become disrupt local then fibers successively. way, and the this In swollen","In this way, the local fibers become swollen and then disrupt successively.",15540
problem region. the localizing We in devices non-collaborative of large address a WiFi,We address the problem of localizing non-collaborative WiFi devices in a large region.,15541
"localizing motive WiFi main localize Our is e.g. devices, to by humans their","Our main motive is to localize humans by localizing their WiFi devices, e.g.",15542
operations natural during search-and-rescue a after disaster.,during search-and-rescue operations after a natural disaster.,15543
"is with very measurements. extremely to these it For make prudent reasons, few decision important","For these reasons, it is extremely important to make prudent decision with very few measurements.",15544
regression. Gaussian (GP) use We Bayesian on optimization the process approach based,We use the Bayesian optimization approach based on Gaussian process (GP) regression.,15545
This of can the image process help analyzing understanding better multiresolution and with registration.,This can help with analyzing and better understanding the process of multiresolution image registration.,15546
"technique global Lipschitz demonstrate we into a incorporate framework. this, optimization our multiresolution To","To demonstrate this, we incorporate our multiresolution technique into a Lipschitz global optimization framework.",15547
dense feature and framework proposed consists of a The detectors extractor important classes. three of,The proposed framework consists of a dense feature extractor and detectors of three important classes.,15548
"the shared features dense extracted, Once detectors. features with these been have are all","Once the dense features have been extracted, these features are shared with all detectors.",15549
achieves performance proposed the competitive approaches several The the-art state-of- benchmark on with framework datasets.,The proposed framework achieves the competitive performance with state-of- the-art approaches on several benchmark datasets.,15550
This to discriminative poorer power robustness. leads less and,This leads to less discriminative power and poorer robustness.,15551
"it higher challenging This to allows highly text in recall. a patterns, detect resulting","This allows it to detect highly challenging text patterns, resulting in a higher recall.",15552
data. and simulation to Our by are an theoretical genomic verified results studies application,Our theoretical results are verified by simulation studies and an application to genomic data.,15553
understanding. recognition is one of from still the for most images problems image Event important,Event recognition from still images is one of the most important problems for image understanding.,15554
for essential Left segmentation measuring left indices. ventricular is ventricular function,Left ventricular segmentation is essential for measuring left ventricular function indices.,15555
several images an contour. or of Segmentation guess of the one initial requires,Segmentation of one or several images requires an initial guess of the contour.,15556
of detection the The presented automatic first step includes here the algorithm of valve. mitral,The first step of the algorithm presented here includes automatic detection of the mitral valve.,15557
"the same detected is the frame. apex Next, in","Next, the apex is detected in the same frame.",15558
is the valve tracked The then cycle. cardiac throughout,The valve is then tracked throughout the cardiac cycle.,15559
an used resulting an active input contour The as is contour to algorithm.,The resulting contour is used as an input to an active contour algorithm.,15560
paper This multiclass the problem segmentation. adresses of interactive,This paper adresses the problem of interactive multiclass segmentation.,15561
 This a distinctive work representation using aims image  of categorization parts. for,  This work aims for image categorization using a representation of distinctive parts.,15562
Hough common The one line transform is of detection. for methods the most,The Hough transform is one of the most common methods for line detection.,15563
regular Hough paper In of we extension this the transform. a propose novel,In this paper we propose a novel extension of the regular Hough transform.,15564
and We rectangle detection. inspection in visual such benefits applications demonstrate quality as,We demonstrate benefits in applications such as visual quality inspection and rectangle detection.,15565
why is and of That set OOP. theory to close object-oriented this version is,That is why this version of set theory is object-oriented and close to OOP.,15566
the generating of address We problem features action for recognition. video,We address the problem of generating video features for action recognition.,15567
"human In technique presented. action is recognition video paper, a this from novel","In this paper, a novel human action recognition technique from video is presented.",15568
"Since training is independent. approach require proposed don't classify we view to the actions,","Since we don't require training to classify actions, the proposed approach is view independent.",15569
knowledge slow on fast Model networks of of relies variables. and biochemical reduction the,Model reduction of biochemical networks relies on the knowledge of slow and fast variables.,15570
predict effects Modern the of challenged biomedicine is to genetic variation.,Modern biomedicine is challenged to predict the effects of genetic variation.,15571
the United The of rise States. been in has use illicit on drugs,The use of illicit drugs has been on the rise in United States.,15572
idea been recently the has embodied EPFL camera. This as Gigavision,This idea has been recently embodied as the EPFL Gigavision camera.,15573
The and objects via based on their representation methods. is of properties model,The model is based on representation of objects via their properties and methods.,15574
"and addition, also In inheritance vague knowledge representation of within imprecise are discussed.","In addition, inheritance within representation of vague and imprecise knowledge are also discussed.",15575
facing perform user shelf. product a the We fine-grained is when recognition (b) class,(b) We perform fine-grained product class recognition when the user is facing a shelf.,15576
through active We improve (c) accuracy continuously learning. the recognition,(c) We continuously improve the recognition accuracy through active learning.,15577
Statistical descriptive widely statistics. are moments in used,Statistical moments are widely used in descriptive statistics.,15578
practice. important stable implementations efficient and numerically are Therefore in,Therefore efficient and numerically stable implementations are important in practice.,15579
is We present a to implement. simpler also version easier that,We present a simpler version that is also easier to implement.,15580
newly The system system compared background with based is a subtraction. on proposed,The newly proposed system is compared with a system based on background subtraction.,15581
performed means and accuracy. the time by is of execution Comparison,Comparison is performed by the means of execution time and accuracy.,15582
previous real data on experimental the statements. both synthetic and support sets Extensive results,Extensive experimental results on both synthetic and real data sets support the previous statements.,15583
"images. In we consider of this of unmixing a problem paper, the time hyperspectral series","In this paper, we consider the problem of unmixing a time series of hyperspectral images.",15584
dynamical linear a propose each based We time model on at processes instant. mixing,We propose a dynamical model based on linear mixing processes at each time instant.,15585
of the images. approach multitemporal hyperspectral on synthetic The demonstrated performance and proposed real is,The performance of the proposed approach is demonstrated on synthetic and real multitemporal hyperspectral images.,15586
"an scalable presented is results. offline and Moreover, to approach is compared online","Moreover, an online scalable approach is presented and is compared to offline results.",15587
"also conception discussed. within of exploiters is In addition, networks universal object-oriented dynamic","In addition, conception of universal exploiters within object-oriented dynamic networks is also discussed.",15588
calculated A before generation. precisely of be can quantity the classes these,A quantity of these classes can be precisely calculated before the generation.,15589
saving knowledge in set only classes basic of of base. the It allows,It allows saving of only basic set of classes in the knowledge base.,15590
"context, of analyze properties method. the this proposed the computational we and In statistical","In this context, we analyze the statistical and computational properties of the proposed method.",15591
validated Theoretical are analysis. thorough a complemented and by results experimental,Theoretical results are complemented and validated by a thorough experimental analysis.,15592
in presented. detectors framework principled for these is combining applications feature A also,A principled framework for combining feature detectors in these applications is also presented.,15593
metric online potential of applications. the the reveal for Timing results,Timing results reveal the potential of the metric for online applications.,15594
using machine Automatic vision track a inspection technology tool. effective has become inspection very,Automatic track inspection using machine vision technology has become a very effective inspection tool.,15595
the paradigms--perceptual a then We computing. provide overview three conceptual of these newest of,We then provide a conceptual overview of the newest of these three paradigms--perceptual computing.,15596
"MET. of using diversity However, from researchers it interpretations by not genetic producing deterred has","However, it has not deterred researchers from producing interpretations of genetic diversity by using MET.",15597
"To date, classified. identified to bacteriocins and close five hundred have been","To date, close to five hundred bacteriocins have been identified and classified.",15598
highly shown distributed species. discoveries and have among bacterial diverse widely Recent that are bacteriocins,Recent discoveries have shown that bacteriocins are highly diverse and widely distributed among bacterial species.,15599
feature satisfactorily No types of state-of-the-art transformations. image image works all detector under,No state-of-the-art image feature detector works satisfactorily under all types of image transformations.,15600
provide new feature The of into obtained detectors. behaviour results insights the,The results obtained provide new insights into the behaviour of feature detectors.,15601
"bandit optimization under model. consider problem stochastic a we of paper, feedback the this In","In this paper, we consider the problem of stochastic optimization under a bandit feedback model.",15602
the Solomonoff agent reinforcement answers and proposed induction question. this to AIXI learning are,Solomonoff induction and the reinforcement learning agent AIXI are proposed answers to this question.,15603
is illustrated applications. inpainting representation new with denoising This and,This new representation is illustrated with denoising and inpainting applications.,15604
in highlight use the color restoration results the of information. image of appropriate Therefore interest,Therefore results in image restoration highlight the interest of the appropriate use of color information.,15605
"refinement. a graph Laplacian model for saliency regularized Finally, we present regression nonlinear","Finally, we present a graph Laplacian regularized nonlinear regression model for saliency refinement.",15606
comparison our the demonstrate state-of-the-art Experimental results approach in approaches. of with effectiveness the,Experimental results demonstrate the effectiveness of our approach in comparison with the state-of-the-art approaches.,15607
we this modeling. general a In to Bayesian robust paper develop approach,In this paper we develop a general approach to robust Bayesian modeling.,15608
well the structures recent coffee with These of qualitative accord history the disease. rust,These qualitative structures accord well with the recent history of the coffee rust disease.,15609
Examples presented theory applications. are and to the illustrate,Examples are presented to illustrate the theory and applications.,15610
particular. towards These formation have supercomplex important implications in discoveries bacterial putative in OXPHOS new,These new discoveries have important implications towards putative supercomplex formation in bacterial OXPHOS in particular.,15611
We estimation high and density present for list- tree- structured methods data. dimensional binary/categorical,We present tree- and list- structured density estimation methods for high dimensional binary/categorical data.,15612
are histograms. density estimation dimensional analogies to variable bin high width models Our,Our density estimation models are high dimensional analogies to variable bin width histograms.,15613
"however, our higher easily can. visualized be in models whereas dimensions, cannot Histograms,","Histograms, however, cannot easily be visualized in higher dimensions, whereas our models can.",15614
"Our high-dimensional histograms. unlike models are sparse,","Our models are sparse, unlike high-dimensional histograms.",15615
main has of for to characteristics semantic actions. specifics embedding Our three accommodate the,Our semantic embedding has three main characteristics to accommodate for the specifics of actions.,15616
"objects. propose to a mechanism of descriptions First, multiple-word actions exploit and we","First, we propose a mechanism to exploit multiple-word descriptions of actions and objects.",15617
"Second, responsive action. automated selection incorporate of objects we the most the per","Second, we incorporate the automated selection of the most responsive objects per action.",15618
approach. Experiments of potential four action on the datasets demonstrate our,Experiments on four action datasets demonstrate the potential of our approach.,15619
severely of ADPKD cysts. the non-uniform of formation the to shape due alters kidneys,ADPKD severely alters the shape of the kidneys due to non-uniform formation of cysts.,15620
"challenging. segmentation consequence, very fully of such is a kidneys As automatic","As a consequence, fully automatic segmentation of such kidneys is very challenging.",15621
with We present forest random classifier. a a interaction based on minimal method user segmentation,We present a segmentation method with minimal user interaction based on a random forest classifier.,15622
"a way. faces render In in this paper, to aim aging we personalized automatically","In this paper, we aim to automatically render aging faces in a personalized way.",15623
"the learning taken into in factors process. dictionary are two consideration Moreover,","Moreover, two factors are taken into consideration in the dictionary learning process.",15624
"characteristics, dictionaries, personalized First, extra have may e.g. aging beyond subject each the facial","First, beyond the aging dictionaries, each subject may have extra personalized facial characteristics, e.g.",15625
"which process. mole, invariant are the in aging","mole, which are invariant in the aging process.",15626
"A generalized distributions, i.e. model for multivariate is developed spherical flexible","A flexible model is developed for multivariate generalized spherical distributions, i.e.",15627
star level are shaped. ones with that sets,ones with level sets that are star shaped.,15628
gensphere. implemented an R These techniques in package are,These techniques are implemented in an R package gensphere.,15629
GLASSES: optimisation and Look-Ahead We with Stochastic Expected-loss Search. Global Simulation through present,We present GLASSES: Global optimisation with Look-Ahead through Stochastic Simulation and Expected-loss Search.,15630
"novel the of GLASSES, the consideration of algorithm, Our permits into dozens future. evaluations","Our novel algorithm, GLASSES, permits the consideration of dozens of evaluations into the future.",15631
of order fractal paintings Jackson We degree dimension}. the of using \emph{Hausdorff-Besicovitch twenty-two Pollock's report,We report the degree of order of twenty-two Jackson Pollock's paintings using \emph{Hausdorff-Besicovitch fractal dimension}.,15632
identify a as fractal-order transition. this behavior We,We identify this behavior as a fractal-order transition.,15633
"characterize fractal \emph{Teri's spectrum, the We Find}. its called also by","We also characterize by its fractal spectrum, the called \emph{Teri's Find}.",15634
the is differential semantic The evaluated tests. of performance with help of (SD) methods the,The performance of the methods is evaluated with the help of semantic differential (SD) tests.,15635
viruses endocytosis a binding Influenza the enter via to after surface. cell,Influenza viruses enter a cell via endocytosis after binding to the surface.,15636
with this along studies conducted. other also policies involving are Simulation,Simulation studies involving this along with other policies are also conducted.,15637
high-dimensional new data. is method The to applicable,The new method is applicable to high-dimensional data.,15638
with mean inference approximation. solutions based on field variational Clustering are,Clustering solutions are based on variational inference with mean field approximation.,15639
and contains Our paragraph hierarchical a a framework generator. sentence generator,Our hierarchical framework contains a sentence generator and a paragraph generator.,15640
The generator short produces interval. describes that one simple a video short specific sentence sentence,The sentence generator produces one simple short sentence that describes a specific short video interval.,15641
exploits It both to selectively visual temporal- focus during spatial-attention on generation. mechanisms and elements,It exploits both temporal- and spatial-attention mechanisms to selectively focus on visual elements during generation.,15642
TACoS-MultiLevel. benchmark two on and datasets: YouTubeClips We evaluate approach our large-scale,We evaluate our approach on two large-scale benchmark datasets: YouTubeClips and TACoS-MultiLevel.,15643
that analyze to show present separations. also protein used be We can the domain mGNM,We show that the present mGNM can also be used to analyze protein domain separations.,15644
"showcase the of simulation our mANM ability motions. protein of we for collective Finally, the","Finally, we showcase the ability of our mANM for the simulation of protein collective motions.",15645
recognition face Such performance used can predict to be a a the of model system.,Such a model can be used to predict the performance of a face recognition system.,15646
(SfM) relies on feature Structure-from-motion tracking. largely,Structure-from-motion (SfM) largely relies on feature tracking.,15647
"for robustly contribute segment-based algorithm effective an SfM handling we large In addition, coarse-to-fine datasets.","In addition, we contribute an effective segment-based coarse-to-fine SfM algorithm for robustly handling large datasets.",15648
video of demonstrate on the the system. results data challenging effectiveness Experimental proposed,Experimental results on challenging video data demonstrate the effectiveness of the proposed system.,15649
representation of boosts This and performance the classification enriches less-represented classes. the,This boosts the classification performance and enriches the representation of less-represented classes.,15650
"evaluate and datasets, SIFTflow on two LMSun. our system large-scale We","We evaluate our system on two large-scale datasets, SIFTflow and LMSun.",15651
and SIFTflow the dataset on performance We results achieve state-of-the-art near-record on LMSun.,We achieve state-of-the-art performance on the SIFTflow dataset and near-record results on LMSun.,15652
Vehicle important ITS (Intelligent Traffic elements in one information is of the System). color,Vehicle color information is one of the important elements in ITS (Intelligent Traffic System).,15653
"using vehicle (CNN). method present convolutional this we color network In neural recognition a paper,","In this paper, we present a vehicle color recognition method using convolutional neural network (CNN).",15654
"test vehicle dataset method, To color by publicly we use provided Chen. our recognition","To test our method, we use publicly vehicle color recognition dataset provided by Chen.",15655
estimate in algorithm depth video scenes. an present dynamic We to,We present an algorithm to estimate depth in dynamic video scenes.,15656
into spatio-temporal We video regions. by a decomposing start,We start by decomposing a video into spatio-temporal regions.,15657
"maps. perform consistent to obtain smoothing we temporally Subsequently, temporal depth","Subsequently, we perform temporal smoothing to obtain temporally consistent depth maps.",15658
by the We abridged tables life use provided WHO.,We use the abridged life tables provided by WHO.,15659
the state more function theory have apply to We and estimates health also comparisons.,We also apply the health state function theory to have more estimates and comparisons.,15660
improved the in for results The recent cases. estimates majority of years the WHO suggest,The results suggest improved WHO estimates in recent years for the majority of the cases.,15661
now measured experimentally many mammalian been regions activity has in Coherent of cortex.,Coherent activity has now been measured experimentally in many regions of mammalian cortex.,15662
to explain limited This has ability neuronal responses. graded their,This has limited their ability to explain graded neuronal responses.,15663
called (SGSCs). these synfire synfire-gated circuits We chains,We called these circuits synfire-gated synfire chains (SGSCs).,15664
Markov field We pairwise in a learn (MRF) boundaries random occlusion framework.,We learn occlusion boundaries in a pairwise Markov random field (MRF) framework.,15665
"in remove temporal to inconsistencies temporally estimation. occlusion Then, we smooth boundaries boundary","Then, we temporally smooth boundaries to remove temporal inconsistencies in occlusion boundary estimation.",15666
"In attention efforts. critical take at chapter, we recent this look modeling a","In this chapter, we take a critical look at recent attention modeling efforts.",15667
to contours Active output results are MRF boost as seeds. used using,Active contours are used to boost results using MRF output as seeds.,15668
study Diploglossus is used. subject lizard millepunctatus the As,As study subject the Diploglossus millepunctatus lizard is used.,15669
image (e.g. between features relationship quality the model We,We model the relationship between image quality features (e.g.,15670
performance recognition density a function. measures probability and using,and recognition performance measures using a probability density function.,15671
of importance. determining someone For first impressions are of people,For people first impressions of someone are of determining importance.,15672
information. are further They hard through to alter,They are hard to alter through further information.,15673
This if the a question the reach begs same computer judgement. can,This begs the question if a computer can reach the same judgement.,15674
"introduce analysis this for complexes. acid work, FRI we protein-nucleic flexibility the In of","In this work, we introduce FRI for the flexibility analysis of protein-nucleic acid complexes.",15675
"paper, we finite to a dependency adjust on propose estimates framework samples. measure this In","In this paper, we propose a framework to adjust dependency measure estimates on finite samples.",15676
"recruitment. dependent survival we stochastic can of Moreover, density impacts adult dampen show","Moreover, we show density dependent adult survival can dampen impacts of stochastic recruitment.",15677
"Thus, and adult may variability food recruitment regulating by population compensatory simultaneously alter fluctuations survival.","Thus, food variability may alter population fluctuations by simultaneously regulating recruitment and compensatory adult survival.",15678
computational reviews recent of chapter This visual models attention.,This chapter reviews recent computational models of visual attention.,15679
outstanding proposed The benchmarks. state-of-the-art tracking compared algorithm performance methods with illustrates existing in,The proposed algorithm illustrates outstanding performance compared with state-of-the-art methods in existing tracking benchmarks.,15680
to subject control ensure order safety. quality Airbags passengers are strict to in,Airbags are subject to strict quality control in order to ensure passengers safety.,15681
aggregated developed using aggregation (e.g. approaches be features These local can for features local,These local features can be aggregated using aggregation approaches developed for local features (e.g.,15682
"providing powerful global thus descriptors. Fisher vectors), new","Fisher vectors), thus providing new powerful global descriptors.",15683
"risk e.g. is and overfitting when has This little bears few of efficient, method parameters,","This method is efficient, has few parameters, and bears little risk of overfitting when e.g.",15684
"considerably. benchmarks descriptor four on new the compact Overall, common state-of-the-art the improves global","Overall, the new compact global descriptor improves the state-of-the-art on four common benchmarks considerably.",15685
"regulatory of blockade Recently, (i.e. molecules the negative","Recently, the blockade of negative regulatory molecules (i.e.",15686
allows T model to enhance antitumor to evaluate us responses. order This cell in strategies,This model allows us to evaluate strategies in order to enhance antitumor T cell responses.,15687
the strategy of immunotherapy. a provides model effects therapeutic Our maximizing combination for rational cancer,Our model provides a rational combination strategy for maximizing the therapeutic effects of cancer immunotherapy.,15688
the these over persistence lifespan. modules stability concerns question of human and important the An,An important question concerns the persistence and stability of these modules over the human lifespan.,15689
exhibits age-dependent composition also scale-specific and Module changes.,Module composition also exhibits scale-specific and age-dependent changes.,15690
the most At are the associated regions fine network. flexible scales with mode default,At fine scales the most flexible regions are associated with the default mode network.,15691
"this answering questions. we the In problem of analogy study paper, visual","In this paper, we study the problem of answering visual analogy questions.",15692
argument. via algorithm a spectral It consistency the of the clustering standard implies perturbation,It implies the consistency of the spectral clustering algorithm via a standard perturbation argument.,15693
applying spectral A study denoising the of numerical necessity indicates a simple algorithms. before step,A simple numerical study indicates the necessity of a denoising step before applying spectral algorithms.,15694
when fragmented was the mycelial network further perturbed. RUs,RUs fragmented further when the mycelial network was perturbed.,15695
deficiencies problems GeoTutor in matter. provide the that can subject student address personalized practice,GeoTutor can provide personalized practice problems that address student deficiencies in the subject matter.,15696
rapid marked development by is of Adolescence executive function.,Adolescence is marked by rapid development of executive function.,15697
"unknown. and how Yet, adolescence is these contribute evolve to over development dynamics cognitive","Yet, how these dynamics evolve over adolescence and contribute to cognitive development is unknown.",15698
"Indeed, increasingly adulthood. toward flexibility performance is early for advantageous","Indeed, flexibility is increasingly advantageous for performance toward early adulthood.",15699
"can be complexity for especially time clustering However, their tasks. large rather","However, their time complexity can be rather large especially for clustering tasks.",15700
"bound scheme the original its clustering error the between Furthermore, we and randomization.","Furthermore, we bound the error between the original clustering scheme and its randomization.",15701
a propose work novel technique postprocessing In we for reduction. this compression-artifact,In this work we propose a novel postprocessing technique for compression-artifact reduction.,15702
for diverse transform method proposed on coding. rely The suitable techniques is compression that,The proposed method is suitable for diverse compression techniques that rely on transform coding.,15703
among selection. gametes imposes strong Competition for fertilization,Competition among gametes for fertilization imposes strong selection.,15704
Yet of functional existing fertilization dynamics nature can the alter ignore that models interactions. gamete,Yet existing fertilization models ignore dynamics that can alter the functional nature of gamete interactions.,15705
viable concentrations probabilities. potentially Such drastically processes allow affect egg fertilization to,Such processes potentially allow egg concentrations to drastically affect viable fertilization probabilities.,15706
(GGMs) are Gaussian Graphical structures. network popular Models tools for studying,Gaussian Graphical Models (GGMs) are popular tools for studying network structures.,15707
"estimation propose robust paper, of the Graphical sparse In this for Lasso GGMs. we Trimmed","In this paper, we propose the Trimmed Graphical Lasso for robust estimation of sparse GGMs.",15708
rigorous the of estimator We our a high-dimensional provide setting. statistical in analysis,We provide a rigorous statistical analysis of our estimator in the high-dimensional setting.,15709
"estimation lack approaches guarantees. sparse statistical robust for contrast, existing In GGMs","In contrast, existing approaches for robust sparse GGMs estimation lack statistical guarantees.",15710
evolve Can how these microscopic following we understand macroscopic processes? without the observables,Can we understand how the macroscopic observables evolve without following these microscopic processes?,15711
"scenes. in In work, the natural problem detection we pedestrian of consider this","In this work, we consider the problem of pedestrian detection in natural scenes.",15712
"exhibit with Intuitively, different instances different of may pedestrians dramatically spatial scales features.","Intuitively, instances of pedestrians with different spatial scales may exhibit dramatically different features.",15713
We divide-and-conquer this be alleviated that substantially the issue by argue philosophy. can,We argue that this issue can be substantially alleviated by the divide-and-conquer philosophy.,15714
disjoint which model scales The from ranges. pedestrians detect built-in with sub-networks multiple introduces,The model introduces multiple built-in sub-networks which detect pedestrians with scales from disjoint ranges.,15715
slight of Here treatment a present tempotron of graphical we a reformulation problem. the,Here we present a graphical treatment of a slight reformulation of the tempotron problem.,15716
gives of new algorithms. the phrasing learning into some gradient This based insights nature,This phrasing gives some new insights into the nature of gradient based learning algorithms.,15717
"can, A time. circumstances approximation provide certain approach an in based sampling polynomial under","A sampling based approach can, under certain circumstances provide an approximation in polynomial time.",15718
"networks problems, share hierarchical topological Other structure. related may some neural to","Other problems, related to hierarchical neural networks may share some topological structure.",15719
"the is suggest case. Here, this that we not","Here, we suggest that this is not the case.",15720
probability losses. portfolio in of We large the interested particularly are,We are particularly interested in the probability of large portfolio losses.,15721
is used capabilities. D* its for widely dynamic replanning,D* is widely used for its dynamic replanning capabilities.,15722
"for are and problems, A* yielding D*, suboptimal results. suited not well MOO","A* and D*, are not well suited for MOO problems, yielding suboptimal results.",15723
for presented designed is search this paper planning. MOO optimal path algorithm in The,The search algorithm presented in this paper is designed for optimal MOO path planning.,15724
"thus named Pareto into incorporates optimality and is D*-PO. The D*, algorithm","The algorithm incorporates Pareto optimality into D*, and is thus named D*-PO.",15725
solution calculating at Non-dominated the paths front Pareto step. by guaranteed search are each,Non-dominated solution paths are guaranteed by calculating the Pareto front at each search step.,15726
images Matching because appearance the cross-view challenging significantly viewpoints is and different. are,Matching cross-view images is challenging because the appearance and viewpoints are significantly different.,15727
"matching. used for semantically be can cross-view Consequently, labeled performing regions","Consequently, semantically labeled regions can be used for performing cross-view matching.",15728
area An images large results. and with experimental challenging promising shows evaluation query urban a,An experimental evaluation with challenging query images and a large urban area shows promising results.,15729
"it attempts, is predictable. that folding Underlying is supposed process this these","Underlying these attempts, it is supposed that this folding process is predictable.",15730
protein the folding dynamic of In topological evaluated. is this paper,In this paper the topological dynamic of protein folding is evaluated.,15731
structure prediction for both Consequences outlined. biology and are then,Consequences for both structure prediction and biology are then outlined.,15732
"spontaneous low analyzing However, and duration their very due challenging is MEs intensity. short to","However, analyzing spontaneous MEs is very challenging due to their short duration and low intensity.",15733
tasks: includes ME Automatic analysis ME recognition. ME and spotting two,Automatic ME analysis includes two tasks: ME spotting and ME recognition.,15734
"studies spotting, videos. spontaneous on rather than ME have posed focused previous For","For ME spotting, previous studies have focused on posed rather than spontaneous videos.",15735
"recognition, the low. ME of performance previous is For studies","For ME recognition, the performance of previous studies is low.",15736
free arbitrary on videos. This unseen training and is method works,This method is training free and works on arbitrary unseen videos.,15737
ring motivates to homomorphisms}. neural us This {\it define,This motivates us to define {\it neural ring homomorphisms}.,15738
"to that we find behave application, respect convexity. neural As an with nicely ring homomorphisms","As an application, we find that neural ring homomorphisms behave nicely with respect to convexity.",15739
which HQM's stochastic on are observation. formalize processes based quantum time-dependent,HQM's are based on quantum stochastic processes which formalize time-dependent observation.,15740
signal frequency-domain allow are They processing. autocovariance the calculation of which essential for functions,They allow the calculation of autocovariance functions which are essential for frequency-domain signal processing.,15741
(WIVs). weeks three after random-like became PS sparse and vitro patterns Initially in robust,Initially random-like and sparse PS patterns became robust after three weeks in vitro (WIVs).,15742
number short They characterized inter-paired and high spike occurrences of were (IPSIs). intervals by a,They were characterized by a high number of occurrences and short inter-paired spike intervals (IPSIs).,15743
"between within establishing PS connectivity the activity different participated in network. Moreover, developing sites functional","Moreover, PS activity participated in establishing functional connectivity between different sites within the developing network.",15744
spiking a be PS thus may considered predictor.,PS may thus be considered a spiking predictor.,15745
circular problem objects vision an Detection in applications. images of digital several is in important,Detection of circular objects in digital images is an important problem in several vision applications.,15746
algorithms the detection knowing rely signature advance. target on target Hyperspectral in desired,Hyperspectral target detection algorithms rely on knowing the desired target signature in advance.,15747
approach signatures learning target for data An is from training presented.,An approach for learning target signatures from training data is presented.,15748
real data. on shown are Results and simulated,Results are shown on simulated and real data.,15749
"videos. in to In we work, detect segment and water aim this","In this work, we aim to segment and detect water in videos.",15750
"The compared to texture specific less recognition. general is problem, discussed however,","The specific problem, however, is less discussed compared to general texture recognition.",15751
"we motion analyze water. Here, several properties of","Here, we analyze several motion properties of water.",15752
"investigate water spatial descriptors. temporal properties derive Second, we and corresponding and the local of","Second, we investigate the temporal and spatial properties of water and derive corresponding local descriptors.",15753
pipeline parts. this consists The three of of architecture,The pipeline of this architecture consists of three parts.,15754
been problem. difficult traditionally of regarded images still computational as objects has a Recognition in,Recognition of objects in still images has traditionally been regarded as a difficult computational problem.,15755
"estimation but computationally intensive, task for is motion animals. important a sighted Visual","Visual motion estimation is a computationally intensive, but important task for sighted animals.",15756
visual tasks biological of attempted sensing first one the field. the was neuromorphic in Replicating,Replicating biological visual sensing was one of the first tasks attempted in the neuromorphic field.,15757
on specifically paper focus we motion of In visual estimation. task the this,In this paper we focus specifically on the task of visual motion estimation.,15758
performs the by detecting measurement-to-target measurement automatically relevance. association It,It performs measurement-to-target association by automatically detecting the measurement relevance.,15759
"state-of-the-art techniques et to et due Khan A Laet comparison al. al., with","A comparison with state-of-the-art techniques due to Khan et al., Laet et al.",15760
emerging connectomics. variability of connectivity functional is field an in Investigating temporal,Investigating temporal variability of functional connectivity is an emerging field in connectomics.,15761
applying of extracted with scale-stability This varies the number k-means. clusters when,This scale-stability varies with the number of extracted clusters when applying k-means.,15762
evidence present plays first scale-invariance that in an We showing connectivity-state role important considerations.,We present first evidence showing that scale-invariance plays an important role in connectivity-state considerations.,15763
A repository provided. frequency-resolved our is of time-series data,A data repository of our frequency-resolved time-series is provided.,15764
anatomy section Serial specimen. method for Motivation: biological is microscopy reconstruction of established an detailed,Motivation: Serial section microscopy is an established method for detailed anatomy reconstruction of biological specimen.,15765
and section may that series in This be includes order. collected imaged unknown,This includes that section series may be collected and imaged in unknown order.,15766
Werman with four returned to later. decades synonymy schlegelii supraciliaris,Werman returned supraciliaris to synonymy with schlegelii four decades later.,15767
bound the previous initialize and search Most use methods to gyroscopes features. for,Most previous methods use gyroscopes to initialize and bound the search for features.,15768
the dimer the showed existing The envelope on cavity protein. of previous the structure research,The previous research showed the existing cavity on the dimer structure of the envelope protein.,15769
"this the prevented. can fusion process fashion, In be","In this fashion, the fusion process can be prevented.",15770
these character toxicity analised silico. were five in pharmacological of and The ligands,The pharmacological and toxicity character of these five ligands were analised in silico.,15771
The the was that PYRRP best result ligand. showed,The result showed that PYRRP was the best ligand.,15772
complex was Stability dynamics analyzed ligand-protein simulation. by molecular of,Stability of ligand-protein complex was analyzed by molecular dynamics simulation.,15773
depth-dependent segmentation of can the scene. This a cause,This can cause a depth-dependent segmentation of the scene.,15774
benchmarks. on videos than We thirty results more present different from,We present results on more than thirty videos from different benchmarks.,15775
essential for nutrients in stark of have processes many a Deprivation can cell. consequences,Deprivation of essential nutrients can have stark consequences for many processes in a cell.,15776
with charged amino missing tRNAs the acid.,tRNAs charged with the missing amino acid.,15777
"poorly known Onychophorans, worms, and animals. rare are or velvet","Onychophorans, or velvet worms, are poorly known and rare animals.",15778
with Scanning Microscopy; Specimens solorzanoi examined Electron sp. Peripatus were,Specimens were examined with Scanning Electron Microscopy; Peripatus solorzanoi sp.,15779
"be investigate separation. Furthermore, incremental we can and implementations real-time achieved if background/foreground algorithms for","Furthermore, we investigate if incremental algorithms and real-time implementations can be achieved for background/foreground separation.",15780
and The the the algorithm consistency of the estimators. assures stability,The algorithm assures the stability and the the consistency of estimators.,15781
technical the Li report al. et algorithms re-evaluates Y. proposed by This,This technical report re-evaluates the algorithms proposed by Y. Li et al.,15782
Meshlab. displayed be with can easily results analyzed The and,The results can easily be displayed and analyzed with Meshlab.,15783
proposed. both of advantages the We and shortcomings found methods,We found both advantages and shortcomings of the methods proposed.,15784
"wood results, histogram is an clustering image. Using SIFT the each calculated for keypoints","Using the clustering results, an SIFT keypoints histogram is calculated for each wood image.",15785
"sums. as computation rule operations of update such partition the avoids the Importantly, typical infeasible","Importantly, the update rule avoids typical infeasible operations such as the computation of partition sums.",15786
proof simulations discrete and a show for in action domains. concept We environment of,We show in simulations a proof of concept for discrete action and environment domains.,15787
A rule over label. of then posterior a application Bayes' simple probability the gives,A simple application of Bayes' rule then gives a posterior probability over the label.,15788
the a We dependent background spatially prior and foreground. use for,We use a spatially dependent prior for the background and foreground.,15789
the are habitat characteristics These Peripatopsidae. family southern with associated,These habitat characteristics are associated with the southern family Peripatopsidae.,15790
investment correlated parental conditions. environmental not and are ratio young general Sex with per,Sex ratio and parental investment per young are not correlated with general environmental conditions.,15791
more Peripatus off Larger (Peripatidae: Brazil) produce and heavier spring. females acacioi,Larger Peripatus acacioi females (Peripatidae: Brazil) produce more and heavier off spring.,15792
Then large assumptions by evaluate running we experiments. these number of a,Then we evaluate these assumptions by running a large number of experiments.,15793
become have few past networks very years. over the Deep popular,Deep networks have become very popular over the past few years.,15794
network. we In space a color introduce module this learns transformations paper within that a,In this paper we introduce a module that learns color space transformations within a network.,15795
of family a image It is transformations of (e.g. part broader,It is part of a broader family of image transformations (e.g.,15796
network populations. of an used with The arbitrary point-neuron can number be scheme,The scheme can be used with an arbitrary number of point-neuron network populations.,15797
common dimensional very field data in high are methods of reduction Dimensionality the analysis.,Dimensionality reduction methods are very common in the field of high dimensional data analysis.,15798
"are for expensive. Typically, dimensionality computationally algorithms reduction","Typically, algorithms for dimensionality reduction are computationally expensive.",15799
"of analysis of are for massive amounts their Therefore, applications the impractical. data","Therefore, their applications for the analysis of massive amounts of data are impractical.",15800
"computations example, due For to data computationally are prohibitive. repeated accumulated","For example, repeated computations due to accumulated data are computationally prohibitive.",15801
extension newly-arrived points. out-of-sample an We algorithm to an data which describe performs,We describe an algorithm which performs an out-of-sample extension to newly-arrived data points.,15802
We the the prove is bounded. proposed of error algorithm that,We prove that the error of the proposed algorithm is bounded.,15803
The similar have points a four stage. evaluated successional,The four evaluated points have a similar successional stage.,15804
major on on shown two fronts. has Recent developments subtraction work background,Recent work on background subtraction has shown developments on two major fronts.,15805
model from frequencies is also studied. identifiability of The parameters $k$-mer,The identifiability of model parameters from $k$-mer frequencies is also studied.,15806
Medicago In this we model the explore issue truncatula. legume this using wild paper,In this paper we explore this issue using the wild model legume Medicago truncatula.,15807
biomedicine breeding insight and can accelerate This research programs.,This insight can accelerate breeding and biomedicine research programs.,15808
segmentation (FCNs). by in state-of-the-art networks currently The represented semantic convolutional is fully,The state-of-the-art in semantic segmentation is currently represented by fully convolutional networks (FCNs).,15809
FCNs result object boundaries. around a As are produce segmentations to tend localized that poorly,As a result FCNs tend to produce segmentations that are poorly localized around object boundaries.,15810
and segment coherence used is to The information boundary to localization. object semantic improve enhance,The boundary information is used to enhance semantic segment coherence and to improve object localization.,15811
pairwise the boundaries energy. then We to define employ in our predicted potentials,We then employ the predicted boundaries to define pairwise potentials in our energy.,15812
performance image tasks. its ConvNets have classification Deep shown in good,Deep ConvNets have shown its good performance in image classification tasks.,15813
deep a remains However as recognition. it representation still action for video problem in,However it still remains as a problem in deep video representation for action recognition.,15814
pooled the video to encoded then The VLAD with representations. descriptors form local are,The pooled local descriptors are then encoded with VLAD to form the video representations.,15815
performance object recall. on paper of terms in focus In improving we this detection,In this paper we focus on improving object detection performance in terms of recall.,15816
by the is sampling exploration performed This image. in proposals object,This exploration is performed by sampling object proposals in the image.,15817
"we of propose In relations objects. method a between discover addition, to higher-order groups novel","In addition, we propose a novel method to discover higher-order relations between groups of objects.",15818
"for In a scheme content we paper, intra propose this prediction screen video. novel","In this paper, we propose a novel intra prediction scheme for screen content video.",15819
We SR techniques (i.e. benchmarks validate our on seven standard,We validate our seven techniques on standard SR benchmarks (i.e.,15820
detection. a hand introduce new We for pipeline and fingertip localization,We introduce a new pipeline for hand localization and fingertip detection.,15821
cities. an ozone task is of modern prediction level agencies important air quality The of,The ozone level prediction is an important task of air quality agencies of modern cities.,15822
computer inputs based with propose output. We a system three single and,We propose a computer based system with three inputs and single output.,15823
with This computationally machines. focus the model capable Turing modularized view the the binds with,This modularized view with the focus binds the model with the computationally capable Turing machines.,15824
"other related RAC mA+ relate contribute MADs. that we formalisms Finally, in to to","Finally, we relate mA+ to other related formalisms that contribute to RAC in MADs.",15825
"detect Moreover, to identification might be hard constraints statistically.","Moreover, identification constraints might be hard to detect statistically.",15826
prove on the bounds of We OA in an patterns model. properties theoretical,We prove theoretical bounds on the properties of patterns in an OA model.,15827
"degree differences explained. the were Then, proposed moving existing and rate between matching","Then, the differences between proposed moving rate and existing matching degree were explained.",15828
"proposed T-S model. method identification based rate the for moving is Next, on","Next, the identification method based on moving rate is proposed for T-S model.",15829
that significantly Test fuzzy identification. of proposed improves results show method effectiveness the the,Test results show that the proposed method significantly improves the effectiveness of fuzzy identification.,15830
"we neutrality theoretical derive the on Within NERM risks. bound a empirical generalization and framework,","Within the NERM framework, we derive a theoretical bound on empirical and generalization neutrality risks.",15831
problem. method This attractiveness the address a prediction to paper proposes leaning challenging deep facial,This paper proposes a deep leaning method to address the challenging facial attractiveness prediction problem.,15832
"unresolved been developed limitations proposed, remain. issues have and several certain and Although methods","Although several methods have been developed and proposed, certain limitations and unresolved issues remain.",15833
Different or measures have vectors. applied similarity these been dissimilarity to,Different similarity or dissimilarity measures have been applied to these vectors.,15834
"are their and advantages Specifically, disadvantages compared. mentioned and","Specifically, their advantages and disadvantages are mentioned and compared.",15835
adding the model capacity competition WTA neurons. information hidden of The increases the without,The WTA competition increases the information capacity of the model without adding hidden neurons.,15836
"Logo intellectual and recognition applications, for brand protection. particularly from has property detection many images","Logo detection from images has many applications, particularly for brand recognition and intellectual property protection.",15837
and describe a database. constructing ideas for large-scale the such We challenges,We describe the ideas and challenges for constructing such a large-scale database.,15838
neuroscientific research humans objects. demonstrates fixed size a Recent salient have for prior that,Recent neuroscientific research demonstrates that humans have a fixed size prior for salient objects.,15839
available source All pre-trained will at be and models GitHub. code,All source code and pre-trained models will be available at GitHub.,15840
The calculation only advantage the has that simulations is to over executed once. be,The advantage over simulations is that the calculation has to be executed only once.,15841
the matrix is Building automated algorithm. transition our by,Building the transition matrix is automated by our algorithm.,15842
on a paper perform analysis we In this image automated large-scale dataset. historical,In this paper we perform automated analysis on a large-scale historical image dataset.,15843
streaming a framework to selection. propose covariance perform We,We propose a framework to perform streaming covariance selection.,15844
the learnt in regularization an be online efficiently This parameter allows manner. for to,This allows for the regularization parameter to be efficiently learnt in an online manner.,15845
"assumptions, non-stochastic Under in mild are results to able convergence a setting. obtain we","Under mild assumptions, we are able to obtain convergence results in a non-stochastic setting.",15846
on work based Random a model In Field. Conditional Gaussian this new we present,In this work we present a new model based on Gaussian Conditional Random Field.,15847
"CK+ test on work two and We datasets, facial RU-FACS. our train and expression","We train and test our work on two facial expression datasets, CK+ and RU-FACS.",15848
outperform recognition. that the Experimental approach of state art proposed evaluation our shows expression,Experimental evaluation shows that our proposed approach outperform state of the art expression recognition.,15849
neural of systems. convolutional image networks (CNNs) the segmentation backbone semantic are state-of-art Deep,Deep convolutional neural networks (CNNs) are the backbone of state-of-art semantic image segmentation systems.,15850
the target This an semantic optimizing system task-specific segmentation quality. end-to-end produces trainable edges in,This produces task-specific edges in an end-to-end trainable system optimizing the target semantic segmentation quality.,15851
"variational the the linearly memory unlike resources in methods. dataset with required scale size, Nevertheless,","Nevertheless, the memory resources required scale linearly with the dataset size, unlike in variational methods.",15852
instances number This a is the large. of is very when severe limitation,This is a severe limitation when the number of instances is very large.,15853
areas an involves problem missing of image. the The inpainting reconstructing of,The problem of inpainting involves reconstructing the missing areas of an image.,15854
"removing old damaged as many Inpainting images. such applications, photographs or has reconstructing obfuscations from","Inpainting has many applications, such as reconstructing old damaged photographs or removing obfuscations from images.",15855
present this inpainting. for we directional algorithm paper In the diffusion,In this paper we present the directional diffusion algorithm for inpainting.,15856
diffusion by accurately. diffusion directional The algorithm edges regular reconstructing more algorithm on the improves,The directional diffusion algorithm improves on the regular diffusion algorithm by reconstructing edges more accurately.,15857
techniques. the Understanding ConvNets process internal commonly visualization of done is using,Understanding the internal process of ConvNets is commonly done using visualization techniques.,15858
"vital recognition, annotations semantic or for scene models Semantic object are training for segmentation understanding.","Semantic annotations are vital for training models for object recognition, semantic segmentation or scene understanding.",15859
developmental children's in of and drawings God(s) cross-cultural other agents. and supernatural We patterns detect,We detect developmental and cross-cultural patterns in children's drawings of God(s) and other supernatural agents.,15860
of for and experts These investigation. support the hypotheses findings questions further new raise the,These findings support the hypotheses of the experts and raise new questions for further investigation.,15861
depth We termed instance novel a object templateNet present recognition. deep architecture based for,We present a novel deep architecture termed templateNet for depth based object instance recognition.,15862
benefits parametrization template from layer. any this The the from additional network without,The network benefits from this without any additional parametrization from the template layer.,15863
updates the this in to weight efficiently an needed end-to-end derive network train manner. We,We derive the weight updates needed to efficiently train this network in an end-to-end manner.,15864
using benchmark for based We object datasets. available the two publicly depth recognition templateNet instance,We benchmark the templateNet for depth based object instance recognition using two publicly available datasets.,15865
"looking The and challenges pose distractors. large multiple clutter, present similar datasets of variations","The datasets present multiple challenges of clutter, large pose variations and similar looking distractors.",15866
of based alpha method minimization approximate the on is inference Black-box (BB-$\alpha$) new a $\alpha$-divergences.,Black-box alpha (BB-$\alpha$) is a new approximate inference method based on the minimization of $\alpha$-divergences.,15867
it gradient because datasets large to scales implemented be can using BB-$\alpha$ stochastic descent.,BB-$\alpha$ scales to large datasets because it can be implemented using stochastic gradient descent.,15868
easily gradients differentiation. obtained can be automatic These using,These gradients can be easily obtained using automatic differentiation.,15869
"approaches, recent video the three Compared representation makes contributions. this inference paper following to","Compared to recent video representation inference approaches, this paper makes the following three contributions.",15870
"operations attaining while Second, lessened non-linearity. are significantly computation more","Second, computation operations are significantly lessened while attaining more non-linearity.",15871
apply information new to method where temporal video the role. a We plays captioning crucial,We apply the new method to video captioning where temporal information plays a crucial role.,15872
benchmarks. outperforms our that method on the captioning video Experiments demonstrate state-of-the-art,Experiments demonstrate that our method outperforms the state-of-the-art on video captioning benchmarks.,15873
neural predictions functions. most provides fastest while the networks accurate metamodel yield the Kriging,Kriging provides the most accurate predictions while neural networks yield the fastest metamodel functions.,15874
computation All failures. strong metamodels three can detect conveniently,All three metamodels can conveniently detect strong computation failures.,15875
"most find we instability code Kriging tools. provides detection, that useful For the","For code instability detection, we find that Kriging provides the most useful tools.",15876
"time consuming. turns be out quite process the to Thus,","Thus, the process turns out to be quite time consuming.",15877
above to solution simple approach discussed. novel problem a and Here the is a,Here a novel approach and a simple solution to the above problem is discussed.,15878
"novel based classification propose this In convolutional architecture on ProNet we networks. neural paper, a","In this paper, we propose a novel classification architecture ProNet based on convolutional neural networks.",15879
We minimize formulated the to method proximal-subgradient use accelerated function. an cost,We use an accelerated proximal-subgradient method to minimize the formulated cost function.,15880
of its prove analyze and the convergence. the performance We proposed algorithm,We analyze the performance of the proposed algorithm and prove its convergence.,15881
methods the A novel heterogeneous is fusion approach detection object for proposed. of,A novel approach for the fusion of heterogeneous object detection methods is proposed.,15882
create spatially to Forest extends (SCRF) Random Spatially Forest Random labeling. Coherent coherent,Spatially Coherent Random Forest (SCRF) extends Random Forest to create spatially coherent labeling.,15883
in We can SCRF settings that be used believe as other well.,We believe that SCRF can be used in other settings as well.,15884
"which many of in are that remains from there However, far still humans. performance cases","However, there are still many cases in which performance remains far from that of humans.",15885
allows our network classify even deformations. under to intraclass adoption features samples precisely This,This features adoption allows our network to classify precisely intraclass samples even under deformations.,15886
"should long-term attractive campaigns effectively. To organized be managed these more deposit and overcome problems,","To overcome these problems, attractive long-term deposit campaigns should be organized and managed more effectively.",15887
applications. in search for vision hashing methods widely-used neighbor Supervised computer nearest are,Supervised hashing methods are widely-used for nearest neighbor search in computer vision applications.,15888
state-of-the-art batch-learners. hashing Most employ supervised approaches,Most state-of-the-art supervised hashing approaches employ batch-learners.,15889
"large batch-learning datasets. when be confronted inefficient can Unfortunately, training strategies with","Unfortunately, batch-learning strategies can be inefficient when confronted with large training datasets.",15890
"the with online algorithm, approach our an offers size. complexity is dataset linear it Since","Since it is an online algorithm, our approach offers linear complexity with the dataset size.",15891
"to we hash table reduce Thus, a framework also updates. propose","Thus, we also propose a framework to reduce hash table updates.",15892
networks. of of focus The learning Bayesian these paper this scalable is approximate,The focus of this paper is scalable approximate Bayesian learning of these networks.,15893
"pathways transduction disorders, variety cancer. lead signal can including complex to a Dysregulation of in","Dysregulation in signal transduction pathways can lead to a variety of complex disorders, including cancer.",15894
Each mutations. of components gain-of-function both loss-of-function under perturbed and model's the was,Each of the model's components was perturbed under both loss-of-function and gain-of-function mutations.,15895
"the of paying both However, and non-dispersal. entails costs dispersal this","However, this entails paying the costs of both dispersal and non-dispersal.",15896
approximate their to of distributions are methods broadly or posterior Importance moments. used Sampling some,Importance Sampling methods are broadly used to approximate posterior distributions or some of their moments.,15897
"single approach, In proposal weighted a distribution standard samples properly. from are drawn and its","In its standard approach, samples are drawn from a single proposal distribution and weighted properly.",15898
a formulate ZSR We problem. prediction binary as,We formulate ZSR as a binary prediction problem.,15899
Program for a puzzles. based novel We tion solving formula- (LP) Linear jigsaw propose,We propose a novel Linear Program (LP) based formula- tion for solving jigsaw puzzles.,15900
strategy. strength of approach assembly comes The main our LP from contribution and the,The main contribution and strength of our approach comes from the LP assembly strategy.,15901
feedback conditions. transmission primary the of for during is Such BSE epidemic means,Such feedback is the primary means of transmission for BSE during epidemic conditions.,15902
Aerial of many (MAVs) challenges. indoor Vehicles Micro Autonomous possesses navigation,Autonomous indoor navigation of Micro Aerial Vehicles (MAVs) possesses many challenges.,15903
main indoor in limited precision that has environments. is One GPS reason,One main reason is that GPS has limited precision in indoor environments.,15904
indoor system's experiments through diverse show real-time locations. our We performance in,We show our system's performance through real-time experiments in diverse indoor locations.,15905
"we network, To about use our understand trained more several visualization techniques.","To understand more about our trained network, we use several visualization techniques.",15906
of types. The is bee Spanish the authenticating validation out by pollen carried method,The validation of the method is carried out by authenticating Spanish bee pollen types.,15907
approach In take this paper parameters BSDF. the generating for to we a procedural wood,In this paper we take a procedural approach to generating parameters for the wood BSDF.,15908
data. series present and modeling probabilistic a We time to segmentation approach of convex,We present a convex approach to probabilistic segmentation and modeling of time series data.,15909
estimate converting intensity units of immunofluorescence factor An into given. numbers molecular the is,An estimate of the factor converting immunofluorescence intensity units into molecular numbers is given.,15910
on $D.$ model embryos. That transcription stochastic proposing binary gene for $melanogaster$ a suggests,That suggests proposing a stochastic binary model for gene transcription on $D.$ $melanogaster$ embryos.,15911
for of reproduce to data the transcripts. are number We average available able experimental,We are able to reproduce available experimental data for the average number of transcripts.,15912
be then to degraded environment naturally images. applied successfully The and/or low-light hardware network can,The network can then be successfully applied to naturally low-light environment and/or hardware degraded images.,15913
We demonstrate than vanilla performance our better has algorithms. experimentally that approach,We experimentally demonstrate that our approach has better performance than vanilla algorithms.,15914
a regression. novel coordinate facial We for neural network (CNN) present design landmark convolutional,We present a novel convolutional neural network (CNN) design for facial landmark coordinate regression.,15915
coding performance. has Sparse shown classification dictionary (DL) excellent with learning,Sparse coding with dictionary learning (DL) has shown excellent classification performance.,15916
crafted well-performing DL directly features. Many prevailing adopt methods current,Many current prevailing DL methods directly adopt well-performing crafted features.,15917
learned framework optimized. are dictionaries features and a and jointly propose where We,We propose a framework where features and dictionaries are jointly learned and optimized.,15918
constraints maximize graph and are imposed intra-class Discriminative inter-class simultaneously further separability. to compactness,Discriminative graph constraints are further imposed to simultaneously maximize intra-class compactness and inter-class separability.,15919
and theoretical present results handling of features. orthogonality appropriate redundant under of consistency We,We present theoretical results of consistency under orthogonality and appropriate handling of redundant features.,15920
defended benchmarks. practical examples is Every advantage by of means and,Every advantage is defended by means of practical examples and benchmarks.,15921
outperforms that the significantly compared Extensive baselines. approach our demonstrate experiments,Extensive experiments demonstrate that our approach significantly outperforms the compared baselines.,15922
information the contextual through Prior predicted maps. confidence methods post-processing leverage often,Prior methods often leverage the contextual information through post-processing predicted confidence maps.,15923
"representations. images, semantic studies on primarily focus ZSL Existing attribute-based still and","Existing ZSL studies focus primarily on still images, and attribute-based semantic representations.",15924
functional neuroimaging. tool Task-free emerged analyses a in connectivity have powerful as,Task-free connectivity analyses have emerged as a powerful tool in functional neuroimaging.,15925
"against data. this then tested measure, fluctuation human resting-state sensitivity signal We (SFS),","We then tested this measure, signal fluctuation sensitivity (SFS), against human resting-state data.",15926
discrimination sensitive specific problem and the biochemical consider species. between general of We,We consider the general problem of sensitive and specific discrimination between biochemical species.,15927
"the mathematically relating adaptation. first generic biochemical such of properties characterize it discrimination, to We","We characterize mathematically the generic properties of such discrimination, first relating it to biochemical adaptation.",15928
discriminability Convolutional enhance a We Neural to the propose Networks (CNN). of approach novel,We propose a novel approach to enhance the discriminability of Convolutional Neural Networks (CNN).,15929
opinion this concludes the author of on the The subject paper. personal,The personal opinion of the author on this subject concludes the paper.,15930
al termina dell'autore L'articolo riguardo. l'opinione con,L'articolo termina con l'opinione dell'autore al riguardo.,15931
the on present Selection. of We the Challenge ICON Algorithm results,We present the results of the ICON Challenge on Algorithm Selection.,15932
benchmarks. achieve edge effective modern detection proven top and for have on approaches Data-driven results,Data-driven approaches for edge detection have proven effective and achieve top results on modern benchmarks.,15933
"semantically used for which training. mark are annotators subsequently edges human Specifically, meaningful","Specifically, human annotators mark semantically meaningful edges which are subsequently used for training.",15934
"detect necessary actually form this to learn high-level accurately strong, of supervision edges? Is to","Is this form of strong, high-level supervision actually necessary to learn to accurately detect edges?",15935
smaller We subsequent gives results. that transfer superior on also show learning datasets,We also show that subsequent transfer learning on smaller datasets gives superior results.,15936
in visual important Rich recognition variety semantic problems. a relations of are,Rich semantic relations are important in a variety of visual recognition problems.,15937
Graphical are a for this standard models tool task.,Graphical models are a standard tool for this task.,15938
"BING++ can approaches, faster. with run Compared state-of-the-art significantly comparable achieve other but performance,","Compared with other state-of-the-art approaches, BING++ can achieve comparable performance, but run significantly faster.",15939
"This for which senescence. replicative however, undergo true somatic cells is human not assumption","This assumption however, is not true for human somatic cells which undergo replicative senescence.",15940
alignment available face Our outperforms datasets. publicly methods algorithm on previous significantly,Our algorithm significantly outperforms previous methods on publicly available face alignment datasets.,15941
"model two CNNs. Specifically, saliency our built is stacked upon","Specifically, our saliency model is built upon two stacked CNNs.",15942
"CNNs the the shot. one conduct collaboratively a testing computing For saliency in two image,","For a testing image, the two CNNs collaboratively conduct the saliency computing in one shot.",15943
interaction attention peripheral depend on and Visuospatial and processing. foveal control gaze the of,Visuospatial attention and gaze control depend on the interaction of foveal and peripheral processing.,15944
"filtering. size not Increasing central durations prolonged with fixation filtering, peripheral but filter with","Increasing filter size prolonged fixation durations with peripheral filtering, but not with central filtering.",15945
"low-pass high-pass with prolonged filtering, level not with but filter Increasing fixation filtering. durations","Increasing filter level prolonged fixation durations with low-pass filtering, but not with high-pass filtering.",15946
accuracy of on analysis complex MOT detection effects object shows performance. the Our system,Our analysis shows the complex effects of object detection accuracy on MOT system performance.,15947
our novel for performs task. This activation better,This novel activation performs better for our task.,15948
of we the outperform the each that need show retraining state-of-the-art for We without dataset.,We show that we outperform the state-of-the-art without the need of retraining for each dataset.,15949
which multispectral to a is image. technique aims Pansharpening improve of spatial resolution,Pansharpening is a technique which aims to improve spatial resolution of multispectral image.,15950
"processing state-of-art are implemented. several toolbox, image algorithms pan-sharpening MATLAB Using","Using MATLAB image processing toolbox, several state-of-art pan-sharpening algorithms are implemented.",15951
visually and of assessed pansharpened Quality are images quantitatively.,Quality of pansharpened images are assessed visually and quantitatively.,15952
"remarks. paper is the concluded Finally, with useful","Finally, the paper is concluded with useful remarks.",15953
"exploited convolution they As task for involving any such, be layers. can","As such, they can be exploited for any task involving convolution layers.",15954
Much research mathematical utilized this has modeling. of,Much of this research has utilized mathematical modeling.,15955
superior that Our public experiments our suggest datasets a has performance. two on method latest,Our experiments on two latest public datasets suggest that our method has a superior performance.,15956
difficult identify that also objectives evaluations. possible This capability it makes to require more decoupling,This decoupling capability also makes it possible to identify difficult objectives that require more evaluations.,15957
at aims from matching views. observed Person re-identification non-overlapping pedestrians camera,Person re-identification aims at matching pedestrians observed from non-overlapping camera views.,15958
two person Feature and significant are re-identification. descriptor in problems learning metric,Feature descriptor and metric learning are two significant problems in person re-identification.,15959
"well glasses, includes participants wearing also dataset as make-up. as contact lenses, The","The dataset also includes participants wearing glasses, contact lenses, as well as make-up.",15960
"viewed distinct However, they are approaches. as typically","However, they are typically viewed as distinct approaches.",15961
of a applications. of specimens requirement is range biological Background: major for Identification a,Background: Identification of biological specimens is a major requirement for a range of applications.,15962
(e.g. answer both help can This fundamental,This can help answer both fundamental (e.g.,15963
"diversity in phylogeny, ecological (e.g. analysis) practical questions and","in phylogeny, ecological diversity analysis) and practical questions (e.g.",15964
"changing a circuits, their dynamics. induces connectivity neural in changes Learning task thereby","Learning a task induces connectivity changes in neural circuits, thereby changing their dynamics.",15965
trained Recurrent elucidate To related task Networks. neural we Neural study dynamics,To elucidate task related neural dynamics we study trained Recurrent Neural Networks.,15966
"failure. training be of resulting or Stability assessed, success the can predicting ODE","Stability of the resulting ODE can be assessed, predicting training success or failure.",15967
"state proposed network in predicts theory the Finally, selectivity dependent frequency response.","Finally, the proposed theory predicts state dependent frequency selectivity in network response.",15968
mathematical framework propose selection populations. a natural finite We in for,We propose a mathematical framework for natural selection in finite populations.,15969
"extractedare Characteristic of and Ratio witch Area, parameters cells. Nucleus WBC white Granule blood","Characteristic parameters witch extractedare WBC Area, Nucleus and Granule Ratio of white blood cells.",15970
dataset full to publicly The at available the is community www.csvision.swansea.ac.uk/converse. made research,The full dataset is made publicly available to the research community at www.csvision.swansea.ac.uk/converse.,15971
"is best the assumption an However, this inverse of not convolution. approximation","However, this assumption is not the best approximation of an inverse convolution.",15972
"invariances Nevertheless, it categorical is texture represent to models these unclear still variations. how and","Nevertheless, it is still unclear how these models represent texture and invariances to categorical variations.",15973
"to honeycombed, more knitted. or an swirly, make image","to make an image more swirly, honeycombed, or knitted.",15974
"elucidation some functioning, and Implications etc.) (memory location neuropsychological the of mechanisms in dementia","Implications in the elucidation of some neuropsychological mechanisms (memory location and functioning, dementia etc.)",15975
"show method that (i) is general, We our i.e. learning","We show that our learning method is (i) general, i.e.",15976
"potential (ii) functions; arbitrary it and and architectures to scalable, CRF i.e. applies CNN","it applies to arbitrary CNN and CRF architectures and potential functions; (ii) scalable, i.e.",15977
low (iii) in memory on a has parallelizes it easy straightforwardly and footprint GPUs; implementation.,it has a low memory footprint and straightforwardly parallelizes on GPUs; (iii) easy in implementation.,15978
"simplifies approach hardware unified potential a CNN-CRF the optimization implementation. Additionally,","Additionally, the unified CNN-CRF optimization approach simplifies a potential hardware implementation.",15979
recently. actions studied Recognizing images is still from popularly,Recognizing actions from still images is popularly studied recently.,15980
"unsupervised parts in discover we of collections learning. set image via First, a","First, we discover a set of parts in image collections via unsupervised learning.",15981
two is datasets. be method effective benchmark to Our on shown,Our method is shown to be effective on two benchmark datasets.,15982
Polyvore). sharing datasets (Pinterest websites new collect from image and two We,We collect two new datasets from image sharing websites (Pinterest and Polyvore).,15983
images fashion attributes: We and represent based semantic data-driven on attributes attributes.,We represent fashion images based on attributes: semantic attributes and data-driven attributes.,15984
proposed the correlations the SPN capture attributes. of ranking can The high-order,The proposed ranking SPN can capture the high-order correlations of the attributes.,15985
We our on two collected datasets. of newly the method our show effectiveness,We show the effectiveness of our method on our two newly collected datasets.,15986
the propose new describe to qualitative geometric We networks. chemical dynamics a approach of reactions,We propose a new geometric approach to describe the qualitative dynamics of chemical reactions networks.,15987
a set show graph jumps. We geometrically of connectivity restricts computed possible the that,We show that a geometrically computed connectivity graph restricts the set of possible jumps.,15988
models dynamic for chain) changes. We provide also state (Markov finite machine such,We also provide finite state machine (Markov chain) models for such dynamic changes.,15989
"both aligned are videos With naturally above strategy, and spatially expression temporally.","With above strategy, expression videos are naturally aligned both spatially and temporally.",15990
"further discriminant the embedding. STM discriminative enhance expressionlet-based representation the processed power, is To with","To enhance the discriminative power, the expressionlet-based STM representation is further processed with discriminant embedding.",15991
"evaluated method and MMI, Oulu-CASIA, is Our on public FERA. databases, CK+, four expression","Our method is evaluated on four public expression databases, CK+, MMI, Oulu-CASIA, and FERA.",15992
"In cases, all large outperforms our state-of-the-art a margin. method known by the","In all cases, our method outperforms the known state-of-the-art by a large margin.",15993
and important face landmark an role in plays recognition Facial analysis localization applications.,Facial landmark localization plays an important role in face recognition and analysis applications.,15994
problem of in intensively years. has face past The been studied the alignment,The problem of face alignment has been intensively studied in the past years.,15995
"and face using re-retrained off-the-shelf models. synthetic data, and both on both real detection,","face detection, on both synthetic and real data, using both off-the-shelf and re-retrained models.",15996
"feasible. optimal"" it or be from can ""far Indeed, even not","Indeed, it can be ""far from optimal"" or even not feasible.",15997
"different the For completely to we possibility existing solution reasons, plan. the re-optimize hadn't or","For different reasons, we hadn't the possibility to completely re-optimize the existing solution or plan.",15998
lot computer attention attracted has a of Scene in vision. parsing,Scene parsing has attracted a lot of attention in computer vision.,15999
part a business in crucial IT of most today's infrastructure operations. is,IT infrastructure is a crucial part in most of today's business operations.,16000
"outages times short availability essential. High and and are reliability, response to","High availability and reliability, and short response times to outages are essential.",16001
Logic in Our Networks. Abduction Markov on based approach is,Our approach is based on Abduction in Markov Logic Networks.,16002
to titration application the of protein data. ISA-Tab paper This describes,This paper describes the application of ISA-Tab to protein titration data.,16003
"simultaneously arbitrary views, to of face handle alignment Learning e.g.","Learning to simultaneously handle face alignment of arbitrary views, e.g.",16004
"thought. profile frontal more and be appears views, than we challenging to","frontal and profile views, appears to be more challenging than we thought.",16005
inaccurate is the intuitive This performance but highly head pose to is solution estimation. susceptible,This solution is intuitive but the performance is highly susceptible to inaccurate head pose estimation.,16006
datasets State-of-the-art profile-view reported MultiPIE also frontaland performances on faces. AFW both containing are and,State-of-the-art performances are also reported on MultiPIE and AFW datasets containing both frontaland profile-view faces.,16007
"straightforwardly applications, the psi* particle implemented. In cannot unknown psi*-auxiliary is filter practical be so","In practical applications, psi* is unknown so the psi*-auxiliary particle filter cannot straightforwardly be implemented.",16008
Applications algorithm. include a Markov particle using Carlo parameter chain Monte estimation,Applications include parameter estimation using a particle Markov chain Monte Carlo algorithm.,16009
brain important networks. Modularity an for is attribute topological functional,Modularity is an important topological attribute for functional brain networks.,16010
"study a from single end, problem image. this the To estimation first we depth of","To this end, we first study the problem of depth estimation from a single image.",16011
and our the demonstrate indoor the both model We of scenarios. effectiveness outdoor in,We demonstrate the effectiveness of our model in both the indoor and outdoor scenarios.,16012
architecture based for answering deep (VQA). novel visual We a task learning propose question attention,We propose a novel attention based deep learning architecture for visual question answering task (VQA).,16013
(ABC-CNN) convolutional configurable network attention such neural to based an We learn attention. introduce question-guided,We introduce an attention based configurable convolutional neural network (ABC-CNN) to learn such question-guided attention.,16014
model ABC-CNN improvements datasets. state-of-the-art over significant on methods these achieves,ABC-CNN model achieves significant improvements over state-of-the-art methods on these datasets.,16015
Bond to can be systems. of biomolecular build graphs used thermodynamically-compliant models hierarchical,Bond graphs can be used to build thermodynamically-compliant hierarchical models of biomolecular systems.,16016
cascade example. pathway) as MAPK illustrative used The is an (Raf-MEK-ERK,The MAPK cascade (Raf-MEK-ERK pathway) is used as an illustrative example.,16017
support to to make action-recognition available We community new the dataset research. a,We make available to the community a new dataset to support action-recognition research.,16018
dataset This several from is ways. in key datasets different prior,This dataset is different from prior datasets in several key ways.,16019
poorly methods that this experiment perform recent state-of-the-art A on dataset. shows baseline,A baseline experiment shows that recent state-of-the-art methods perform poorly on this dataset.,16020
This will dataset research. to advances challenging this action-recognition in suggests be foster that a,This suggests that this will be a challenging dataset to foster advances in action-recognition research.,16021
transparent. version rather simple and is method In the basic its itself,In its basic version the method itself is rather simple and transparent.,16022
"admitted, But counter-examples it gets difficult. partially knowledge are background and more described when","But when background knowledge and partially described counter-examples are admitted, it gets more difficult.",16023
full of together object. point fused these Several a cloud maps give of the depth,Several of these depth maps fused together give a full point cloud of the object.,16024
in The mesh. turn surface cloud a can into point be transformed,The point cloud can in turn be transformed into a surface mesh.,16025
"those CNNs have enhanced attention, concepts in biology. to similar with of been found Recently,","Recently, CNNs have been enhanced with concepts of attention, similar to those found in biology.",16026
processing. on work spatial attention effective of focused Much on this serial has,Much of this work on attention has focused on effective serial spatial processing.,16027
In speaker channels. multiple find conversations over scenarios simultaneously recorded we recognition some,In some speaker recognition scenarios we find conversations recorded simultaneously over multiple channels.,16028
in of That NIST SRE the the case the is dataset. interviews,That is the case of the interviews in the NIST SRE dataset.,16029
"we for document, model. this In equations derive this the","In this document, we derive the equations for this model.",16030
"work, this a propose anti-spoofing we In on color method texture new face analysis. based","In this work, we propose a new face anti-spoofing method based on color texture analysis.",16031
"separately. the from feature extracted More each image are specifically, band histograms","More specifically, the feature histograms are extracted from each image band separately.",16032
"Most our depicts importantly, capabilities. promising generalization very approach evaluation showed the proposed that inter-database","Most importantly, our inter-database evaluation depicts that the proposed approach showed very promising generalization capabilities.",16033
semantic CNN image we segmentation. propose architecture for In this a paper,In this paper we propose a CNN architecture for semantic image segmentation.,16034
segmentation The addresses bilateral two arise general inception that issues CNN architectures. module with,The bilateral inception module addresses two issues that arise with general CNN segmentation architectures.,16035
for and prerequisite video-based robust eye-tracking settings. pupil Fast in is an essential real-world detection,Fast and robust pupil detection is an essential prerequisite for video-based eye-tracking in real-world settings.,16036
"novel filtered evaluation a edge on image. a propose of We algorithm based ElSe, ellipse","We propose ElSe, a novel algorithm based on ellipse evaluation of a filtered edge image.",16037
"architectures embedded We e.g. a at that resource-saving approach robust, integrated be in can aim","We aim at a robust, resource-saving approach that can be integrated in embedded architectures e.g.",16038
the of detection one studied in most vision the topics community. is Face computer,Face detection is one of the most studied topics in the computer vision community.,16039
by detection of face Much have the availability progresses datasets. been benchmark the made of,Much of the progresses have been made by the availability of face detection benchmark datasets.,16040
"and dataset bounding The occlusions, rich event face categories, boxes. poses, annotations, including contains","The dataset contains rich annotations, including occlusions, poses, event categories, and face bounding boxes.",16041
"detection. that FACE an training WIDER Furthermore, for source dataset face is effective we show","Furthermore, we show that WIDER FACE dataset is an effective training source for face detection.",16042
"further that investigated. to cases common we discuss Finally, worth be failure","Finally, we discuss common failure cases that worth to be further investigated.",16043
Dataset downloaded be can . at: mmlab.ie.cuhk.edu.hk/projects/WIDERFace,Dataset can be downloaded at: mmlab.ie.cuhk.edu.hk/projects/WIDERFace .,16044
"a data. Second, utilizing train model this","Second, train a model utilizing this data.",16045
and both benefits performance This scalability. in approach has,This approach has benefits in both performance and scalability.,16046
an to expanding for learning approach fine-grained compare We approach our datasets. active,We compare our approach to an active learning approach for expanding fine-grained datasets.,16047
crucially Handled recognition. objects are for egocentric ADL important,Handled objects are crucially important for egocentric ADL recognition.,16048
user's be can a body. occluded by They,They can be occluded by a user's body.,16049
"user's wrist. a described mount we camera As a on herein,","As described herein, we mount a camera on a user's wrist.",16050
Previous template. the warp approaches the appearance image input either or,Previous approaches warp either the input image or the appearance template.,16051
finally used bounding resulting The box re-ranking. image is for,The resulting bounding box is finally used for image re-ranking.,16052
curves. similar Bezier The to have curves properties new some,The new curves have some properties similar to Bezier curves.,16053
"are some Bezier Furthermore, and for properties curves surfaces fundamental Bernstein discussed.","Furthermore, some fundamental properties for Bernstein Bezier curves and surfaces are discussed.",16054
team is for critical the Tracking of sports. video-based ball analysis,Tracking the ball is critical for video-based analysis of team sports.,16055
on It material effects performance. has considerable,It has considerable effects on material performance.,16056
"there is objective quantitative indicator author's still no knowledge, for automated To the segregation.","To the author's knowledge, there is still no automated objective quantitative indicator for segregation.",16057
"segregation fill to In of task, analyzed. particles this is order full","In order to full fill this task, segregation of particles is analyzed.",16058
Edges digital particles the extracted are picture. the of from,Edges of the particles are extracted from the digital picture.,16059
"is the rectangles Then, small shape. whole of the splintered picture with particles same to","Then, the whole picture of particles is splintered to small rectangles with the same shape.",16060
in each rectangle index calculated. of Statistical edges the is,Statistical index of the edges in each rectangle is calculated.,16061
"the indexes to evaluated. rectangles the between corresponding segregation is Accordingly,","Accordingly, segregation between the indexes corresponding to the rectangles is evaluated.",16062
results evaluated The with subjective results. coincident show,The results show coincident with subjective evaluated results.,16063
"the Next, the regions k-dense in detected salient further are subgraphs image. enhance to","Next, k-dense subgraphs are detected to further enhance the salient regions in the image.",16064
structure measures. centrality local about Dense convey graph more than subgraphs simple information,Dense subgraphs convey more information about local graph structure than simple centrality measures.,16065
"benchmark evaluating image data the do sets. experiments proposed on extensive model, For we","For evaluating the proposed model, we do extensive experiments on benchmark image data sets.",16066
region algorithms comparable salient detection. performs method proposed to well-known in The,The proposed method performs comparable to well-known algorithms in salient region detection.,16067
attracted substantial has interest a The of communities. within subject variety from,The subject has attracted substantial interest from within a variety of communities.,16068
scenes. present an localizing for objects detection We in active model,We present an active detection model for localizing objects in scenes.,16069
"the words, In Bayesian universal. other is approach robust","In other words, the robust Bayesian approach is universal.",16070
is theory robust This exemplified by to relating analysis. Dempster-Shafer's result Bayesian evidence,This result is exemplified by relating Dempster-Shafer's evidence theory to robust Bayesian analysis.,16071
consider set We of cubic as patches. a video a being,We consider a video as being a set of cubic patches.,16072
have and for respect We global training models Gaussian two all patches descriptors. to local,We have two Gaussian models for all training patches respect to global and local descriptors.,16073
from different capture the properties descriptors video These aspects.,These descriptors capture the video properties from different aspects.,16074
"this sparse for image segmentation. is used In decomposition paper,","In this paper, sparse decomposition is used for image segmentation.",16075
based on the Fourier Fast Transform. of such is methods One,One of such methods is based on the Fast Fourier Transform.,16076
presented general mentioned where above more A limitation is is solution relaxed. the,A more general solution is presented where the above mentioned limitation is relaxed.,16077
"for density kernel also estimation Consequently, derivative selection supported. bandwidth is","Consequently, bandwidth selection for kernel density derivative estimation is also supported.",16078
numerical demonstrated practical usability solution by is comprehensive of new simulations. the The,The practical usability of the new solution is demonstrated by comprehensive numerical simulations.,16079
is Our similar PPCA. Bishop's to VB approach,Our approach is similar to Bishop's VB PPCA.,16080
separable saddle convex with functions. structure convex-concave We a problems non-strongly and consider point,We consider convex-concave saddle point problems with a separable structure and non-strongly convex functions.,16081
"and state-of-the-art empirically, these Theoretically than all better significantly methods performance we demonstrate applications. in","Theoretically and empirically, we demonstrate significantly better performance than state-of-the-art methods in all these applications.",16082
"require of image alignment they However, pairs. such manual","However, they require manual alignment of such image pairs.",16083
discrepancy the rely feature the between minimizing distributions domains. Fully unsupervised methods adaptation across on,Fully unsupervised adaptation methods rely on minimizing the discrepancy between the feature distributions across domains.,16084
data. extensive methods State-of-the-art boundary require learning detection training based,State-of-the-art learning based boundary detection methods require extensive training data.,16085
"networks, neural In (RF) depth. receptive convolutional with increases deep hierarchical field size","In convolutional deep neural networks, receptive field (RF) size increases with hierarchical depth.",16086
"the annotated datasets, contact For algorithm please first author. and implementation","For algorithm implementation and annotated datasets, please contact the first author.",16087
"and annotations. The are from human parts objects of objects, obtained pieces","The pieces are objects and parts of objects, obtained from human annotations.",16088
The protein-protein (PPI) interaction is and decision-making. processing crucial for information network cellular,The protein-protein interaction (PPI) network is crucial for cellular information processing and decision-making.,16089
"dense However, different. and are prediction classification structurally image","However, dense prediction and image classification are structurally different.",16090
resolution. without aggregate information systematically module to losing The multi-scale dilated contextual presented convolutions uses,The presented module uses dilated convolutions to systematically aggregate multi-scale contextual information without losing resolution.,16091
increases the semantic presented show state-of-the-art module systems. that segmentation context accuracy We of the,We show that the presented context module increases the accuracy of state-of-the-art semantic segmentation systems.,16092
cake too? eat max-pooled have we multi-layered it Can and our,Can we have our max-pooled multi-layered cake and eat it too?,16093
"to and Fast R-CNNs superior detection. in obtains performance it addition, object In Faster","In addition, it obtains superior performance to Fast and Faster R-CNNs in object detection.",16094
"single and we on geometry. clustering linkage its Specifically, study hierarchical (SLHC) focus","Specifically, we focus on single linkage hierarchical clustering (SLHC) and study its geometry.",16095
used Carlo (HMC) Hamiltonian in related algorithms and routinely computation. Monte Bayesian have become,Hamiltonian Monte Carlo (HMC) and related algorithms have become routinely used in Bayesian computation.,16096
states of the Hamiltonian achieved along trajectories This dynamics. {recycling intermediate by is simulated,This is achieved by {recycling the intermediate states along simulated trajectories of Hamiltonian dynamics.,16097
"Standard intermediate only points end discarding of wastefully the algorithms the use all states. trajectories,","Standard algorithms use only the end points of trajectories, wastefully discarding all the intermediate states.",16098
"sampler, the of to no-U-turn most applies algorithm variant popular arguably Our the straightforwardly HMC.","Our algorithm applies straightforwardly to the no-U-turn sampler, arguably the most popular variant of HMC.",16099
links. by problem optimization as segmentation derive an theoretically these We semantic parameterized the,We theoretically derive the semantic segmentation as an optimization problem parameterized by these links.,16100
several datasets results semantic demonstrate supreme our on representative Experimental performance for segmentation.,Experimental results on several representative datasets demonstrate our supreme performance for semantic segmentation.,16101
Spectral widely for of one used (SC) Clustering methods data the clustering. is most,Spectral Clustering (SC) is one of the most widely used methods for data clustering.,16102
and to appearance challenges key fine-grained variation classification. differences in Pose subtle are,Pose variation and subtle differences in appearance are key challenges to fine-grained classification.,16103
examined previous methods and have pose-normalized features. To to find how parts end extract this,To this end previous methods have examined how to find parts and extract pose-normalized features.,16104
"such data. collect considering However, prohibitive this when to manual becomes the labour quickly needed","However, this quickly becomes prohibitive when considering the manual labour needed to collect such data.",16105
strongest showed and on features. consistent influence similarity connection most Architectonic the,Architectonic similarity showed the strongest and most consistent influence on connection features.,16106
"connection distance By to similarity not and related features. systematically cortical contrast, thickness were","By contrast, cortical thickness similarity and distance were not systematically related to connection features.",16107
The of can useful applications. segmentation be very vision objects in computer transparent,The segmentation of transparent objects can be very useful in computer vision applications.,16108
optimization labeling the for pixel Graph-cut problem. applied is,Graph-cut optimization is applied for the pixel labeling problem.,16109
the from The segments background. transparent successfully that the proposed demonstrate results method objects,The results demonstrate that the proposed method successfully segments transparent objects from the background.,16110
"camera is It of and occlusion due illumination, challenging large pose, variations to view.","It is challenging due to large variations of pose, illumination, occlusion and camera view.",16111
"problem. re-identification, often the the However, in over-fitting deep from person suffer networks","However, in person re-identification, the deep networks often suffer from the over-fitting problem.",16112
person show methods of our on outperforms that re-identification. several benchmarks Experiments significantly state-of-the-art approach,Experiments show that our approach significantly outperforms state-of-the-art methods on several benchmarks of person re-identification.,16113
need State-of-the-art a relays speaker large that models on of amount recognition data. training,State-of-the-art speaker recognition relays on models that need a large amount of training data.,16114
data is SRE because successful This there available. sufficient like NIST are in tasks models,This models are successful in tasks like NIST SRE because there is sufficient data available.,16115
Bayes the variational variables. used We to estimate hidden,We used variational Bayes to estimate the hidden variables.,16116
"Here, the derive for model. this we equations","Here, we derive the equations for this model.",16117
super resolution the We methods. study of evaluating problem,We study the problem of evaluating super resolution methods.,16118
methodology of design. research is Validation in critical research,Validation of research methodology is critical in research design.,16119
Correlation before be observables experiments must or extensive established undertaking propose experimental mechanisms. between,Correlation between experimental observables must be established before undertaking extensive experiments or propose mechanisms.,16120
public impact development for direct Results are channeled health. on sweetener with,Results are channeled for sweetener development with direct impact on public health.,16121
likely observed in coupling step assay. the G-protein is,G-protein coupling step is likely observed in the assay.,16122
"spent been Thus, data. uncorrelated generating years have","Thus, years have been spent generating uncorrelated data.",16123
results. meaningful not uncorrelated from observables give does Data,Data from uncorrelated observables does not give meaningful results.,16124
This into projects. serves methodology researchers validate before to plunging term to article remind long,This article serves to remind researchers to validate methodology before plunging into long term projects.,16125
"the methodology, costly mistake test Ignoring validation a have in field. on been","Ignoring validation test on methodology, have been a costly mistake in the field.",16126
testing approach promising adaptive abilities. is and Computerized (CAT) testing to an interesting human,Computerized adaptive testing (CAT) is an interesting and promising approach to testing human abilities.,16127
model tested a use create to humans. In we Bayesian our research networks of,In our research we use Bayesian networks to create a model of tested humans.,16128
school collected tests data paper performed with grammar students. from We,We collected data from paper tests performed with grammar school students.,16129
the we experiments. of first provide for our In article data summary this used,In this article we first provide the summary of data used for our experiments.,16130
"compared We networks, we cross-validation. different and by Bayesian tested several which propose","We propose several different Bayesian networks, which we tested and compared by cross-validation.",16131
results the are paper. were obtained and Interesting in discussed,Interesting results were obtained and are discussed in the paper.,16132
analysis selection the on brought a The problem. view has model clearer,The analysis has brought a clearer view on the model selection problem.,16133
is the in part outlined concluding research the Future paper. of,Future research is outlined in the concluding part of the paper.,16134
learning. semantic of We deep address segmentation using problem the,We address the problem of semantic segmentation using deep learning.,16135
potentials We allow inference design field with to differentiable higher these order the algorithm. mean,We design these higher order potentials to allow inference with the differentiable mean field algorithm.,16136
"to The estimator function obtain without aim too many ""good"" evaluations. is","The aim is to obtain ""good"" estimator without too many function evaluations.",16137
This well-established coarse-grained describes model. protein chapter a CABS,This chapter describes a well-established CABS coarse-grained protein model.,16138
analysis of their numerical presents a the article extracted keypoints stability. This on impact,This article presents a numerical analysis of their impact on the extracted keypoints stability.,16139
to filter detections to shown are unstable Usual strategies out inefficient. be,Usual strategies to filter out unstable detections are shown to be inefficient.,16140
"hierarchical tracking. features learn paper, object for we propose an to In approach visual this","In this paper, we propose an approach to learn hierarchical features for visual object tracking.",16141
"robust from we diverse features auxiliary sequences. to patterns First, offline learn video motion","First, we offline learn features robust to diverse motion patterns from auxiliary video sequences.",16142
network. are a learned two-layer The hierarchical convolutional neural via features,The hierarchical features are learned via a two-layer convolutional neural network.,16143
methods. feature our algorithm integrate into learning three tracking We,We integrate our feature learning algorithm into three tracking methods.,16144
ways we of object paper for using context In explore detection. this two,In this paper we explore two ways of using context for object detection.,16145
Our proximal scenes are planar distal several first the method computes image. with which compatible,Our method first computes several distal scenes which are compatible with the proximal planar image.,16146
ground-truth We with the also on labeling. results images Berkeley of figure-ground dataset provided present,We also present results on images of the Berkeley dataset with provided figure-ground ground-truth labeling.,16147
"some recently,building architectures. promising Nevertheless, deep network the been on literature reported have results in","Nevertheless, some promising results have been reported in the literature recently,building on deep network architectures.",16148
experimental (SAT-based) demonstrating report scalability results approach. the of We our symbolic,We report experimental results demonstrating the scalability of our symbolic (SAT-based) approach.,16149
and video intrusive. systems tracking Image are based,Image and video based tracking systems are intrusive.,16150
indicated. to commonly Biopsy used is confirm radiologically diagnosis when cancer,Biopsy is commonly used to confirm cancer diagnosis when radiologically indicated.,16151
are for efficient features simplicity The of effective detection. pedestrian very and discrimination important and,The discrimination and simplicity of features are very important for effective and efficient pedestrian detection.,16152
"methods are efficiency. However, tradeoff between state-of-the-art and most unable achieve to accuracy good","However, most state-of-the-art methods are unable to achieve good tradeoff between accuracy and efficiency.",16153
SSF symmetrical the capture can shape. of pedestrian similarity,SSF can capture the symmetrical similarity of pedestrian shape.,16154
"characterization neighboring abilities. features to it's have difficult However, for above such","However, it's difficult for neighboring features to have such above characterization abilities.",16155
"propose Finally, pedestrian to non-neighboring combine we for and features neighboring detection. both","Finally, we propose to combine both non-neighboring and neighboring features for pedestrian detection.",16156
"require- is ment, often neglected. This however,","This require- ment, however, is often neglected.",16157
Automatically fascinating. been ever videos describing has,Automatically describing videos has ever been fascinating.,16158
game to the by domain limiting our address lawn We tennis. of this,We address this by limiting our domain to the game of lawn tennis.,16159
newly a data created We our on method video evaluate set. tennis,We evaluate our method on a newly created tennis video data set.,16160
is the of of cytoskeleton This bistability dynamics. behavior indicative,This behavior is indicative of bistability of the cytoskeleton dynamics.,16161
task. Sequential under constraint (SPM) pattern is challenging gap mining a,Sequential pattern mining (SPM) under gap constraint is a challenging task.,16162
"directly gap global be cannot support to constraint extended constraint. this However,","However, this global constraint cannot be directly extended to support gap constraint.",16163
the right pattern principle extensions. of relies GAP-SEQ on,GAP-SEQ relies on the principle of right pattern extensions.,16164
"important can be for. research point used to this we various problems Finally, that dataset","Finally, we point to various important research problems that this dataset can be used for.",16165
"ongoing focused efforts specific, have laboratory settings. training stereotyped on on in tasks Most decoders","Most ongoing efforts have focused on training decoders on specific, stereotyped tasks in laboratory settings.",16166
accuracy by comparing of annotations. to verify We the approach our manual,We verify the accuracy of our approach by comparing to manual annotations.,16167
new classical Our adaptive approach evolution. into this insight model of provides,Our approach provides new insight into this classical model of adaptive evolution.,16168
"alternative to choices bring image this for we paper, attention In restoration.","In this paper, we bring attention to alternative choices for image restoration.",16169
"the a propose differentiable performance function. error and of novel, compare losses, several We","We compare the performance of several losses, and propose a novel, differentiable error function.",16170
extensions decision Bayesian Influence diagrams of are theoretic networks.,Influence diagrams are decision theoretic extensions of Bayesian networks.,16171
decision to problems. are diverse applied They,They are applied to diverse decision problems.,16172
the a vehicle this of speed we to apply profile. In paper diagrams optimization influence,In this paper we apply influence diagrams to the optimization of a vehicle speed profile.,16173
well computed profiles pilots. speed test to achieved correspond lap by The time and those,The computed lap time and speed profiles correspond well to those achieved by test pilots.,16174
paper applications opens doors for influence of diagrams. This new,This paper opens doors for new applications of influence diagrams.,16175
features learning distance function. specific these by by a address challenges methods designing or Previous,Previous methods address these challenges by designing specific features or by learning a distance function.,16176
propose from hierarchical pairs. feature labeled invariant image learning We framework learns a representations that,We propose a hierarchical feature learning framework that learns invariant representations from labeled image pairs.,16177
"re-identification are small. the datasets of Critically, most person","Critically, most of the person re-identification datasets are small.",16178
on results datasets. stateof-the-art also the Our approach achieves these,Our approach also achieves the stateof-the-art results on these datasets.,16179
This systems trajectory thesis symbols. to that presents handwritten use bachelor's the multiple pen classify,This bachelor's thesis presents multiple systems that use the pen trajectory to classify handwritten symbols.,16180
Experiments various cases. challenging that show is in method of this performance superior,Experiments show that this method is of superior performance in various challenging cases.,16181
have Biological against immune evolved of a to wide mechanisms themselves defend organisms range pathogens.,Biological organisms have evolved a wide range of immune mechanisms to defend themselves against pathogens.,16182
abstract Symbolic Neutrosophic is (or use of Literal) the symbols Theory to referring (i.e.,Symbolic (or Literal) Neutrosophic Theory is referring to the use of abstract symbols (i.e.,16183
"Tj, F, their neutrosophics. or in indexed letters I, T, the refined Fl) letters Ik,","the letters T, I, F, or their refined indexed letters Tj, Ik, Fl) in neutrosophics.",16184
thesis-antithesis-synthesis triad the We the thesis-antithesis-neutrothesis-neutrosynthesis. extend neutrosophic to dialectical tetrad,We extend the dialectical triad thesis-antithesis-synthesis to the neutrosophic tetrad thesis-antithesis-neutrothesis-neutrosynthesis.,16185
properties. define actions We and three their neutrosophic,We define three neutrosophic actions and their properties.,16186
"neutrosophic given then respect prevalence order T,I,F to the a introduce operator. We on with","We then introduce the prevalence order on T,I,F with respect to a given neutrosophic operator.",16187
"and refinement A, And the entities of antiA. neutrosophic neutA,","And the refinement of neutrosophic entities A, neutA, and antiA.",16188
numbers. introduce neutrosophic quadruple refined the We quadruple (a+bT+cI+dF) neutrosophic numbers the and,We introduce the neutrosophic quadruple numbers (a+bT+cI+dF) and the refined neutrosophic quadruple numbers.,16189
of role The design. interaction the an inhibitory in drug plays of type important understanding,The understanding of the type of inhibitory interaction plays an important role in drug design.,16190
"this futile. would be side Without case, information, practically in the factorization the tensor","Without the side information, in this case, the tensor factorization would be practically futile.",16191
Our of approach. indicate the the proposed results effectiveness,Our results indicate the effectiveness of the proposed approach.,16192
"the model, in depth Using real-time. we this reconstruct map","Using this model, we reconstruct the depth map in real-time.",16193
representation the test label to then This collaborative sample. is used,This collaborative representation is then used to label the test sample.,16194
"on benchmark object the databases face, approach. our action Experiments and show of efficacy","Experiments on benchmark face, object and action databases show the efficacy of our approach.",16195
"propagates Activity the visual similar down hierarchy, visual associated experimental imagery. to results with","Activity propagates down the visual hierarchy, similar to experimental results associated with visual imagery.",16196
fMRI with connectivity. Cortico-cortical patterns interaction resting-state agree functional well,Cortico-cortical interaction patterns agree well with fMRI resting-state functional connectivity.,16197
challenging to remains Multi-object occlusions due outliers. of and tracking occurrence frequent,Multi-object tracking remains challenging due to frequent occurrence of occlusions and outliers.,16198
"this we propose an problem, optimization. to for sequential Approximation-Shrink Scheme order In handle","In order to handle this problem, we propose an Approximation-Shrink Scheme for sequential optimization.",16199
"the on states of window-wise a this self-organizing clustered targets are in framework, manner. Based","Based on this window-wise framework, the states of targets are clustered in a self-organizing manner.",16200
for for system collections can indexing useful video retrieval. organizing Our be and,Our system can be useful for organizing video collections for indexing and retrieval.,16201
through Plate deformation vary can that flexible a Spline It time. model Thin uses,It uses a flexible Thin Plate Spline deformation model that can vary through time.,16202
"a system on each step of dataset. fully new, annotated evaluate We carefully our","We carefully evaluate each step of our system on a new, fully annotated dataset.",16203
"On DTF outperform the behavior state-of-the-art discovery, descriptor. we Improved","On behavior discovery, we outperform the state-of-the-art Improved DTF descriptor.",16204
"Flow outperform alignment, the algorithm. On popular we SIFT spatial","On spatial alignment, we outperform the popular SIFT Flow algorithm.",16205
supporting effective learning are Deep transfer techniques learning. renowned for,Deep learning techniques are renowned for supporting effective transfer learning.,16206
"domain, to this multiple we work, In in source orthogonal classifiers. suggest the learn,","In this work, we suggest to learn, in the source domain, multiple orthogonal classifiers.",16207
"to the probabilities multiple are the classifiers softmax likely Interestingly, by be identical. produced","Interestingly, the softmax probabilities produced by the multiple classifiers are likely to be identical.",16208
variations variations. characterised fine-grained and image large classification The by is inter-class intra-class small problem,The fine-grained image classification problem is characterised by large intra-class variations and small inter-class variations.,16209
spreading this particularly in Invasions could manner have dramatic effects.,Invasions spreading in this manner could have particularly dramatic effects.,16210
"various suggested wave that authors Recently, demographic disrupts stochasticity have acceleration.","Recently, various authors have suggested that demographic stochasticity disrupts wave acceleration.",16211
on work and extend systems. paper we earlier In review recent our thermostated this,In this paper we review and extend our earlier recent work on thermostated systems.,16212
biological The implications. paper an of account with concludes the,The paper concludes with an account of the biological implications.,16213
Predictive it acquisition Search Constraints (PESC). and We develop one Entropy with such call function,We develop one such acquisition function and call it Predictive Entropy Search with Constraints (PESC).,16214
"solved function. a In these single paper, by this problems are objective two optimizing","In this paper, these two problems are solved by optimizing a single objective function.",16215
proposed The of support effectiveness approch. of experiments the the numerical results,The results of numerical experiments support the effectiveness of the proposed approch.,16216
"problem. it the delineation Unlike ones, standard specificities of takes the advantage of","Unlike standard ones, it takes advantage of the specificities of the delineation problem.",16217
widely data to for same partitions/clusterings used measures set. Adjusted chance compare of are the,Adjusted for chance measures are widely used to compare partitions/clusterings of the same data set.,16218
geometric the appearance are mixtures The utilising and distances. sub-trees learnt both of,The mixtures of the sub-trees are learnt utilising both geometric and appearance distances.,16219
the the method results occlusions the approaches. proposed robustness of to The over state-of-the-art show,The results show the robustness of the proposed method to occlusions over the state-of-the-art approaches.,16220
develop how efficient explore more CNNs. also to our analyses the field We lead can,We also explore how our analyses can lead the field to develop more efficient CNNs.,16221
"like defines ""kicking an action ball""? What","What defines an action like ""kicking ball""?",16222
"difficult be visual prototype. For may a real-world object categories, it establish to most unique","For most real-world visual object categories, it may be difficult to establish a unique prototype.",16223
state-of-the-art. and on Belga new establishes noteworthy a the logo especially dataset accuracy is Recognition,Recognition accuracy on the Belga logo dataset is especially noteworthy and establishes a new state-of-the-art.,16224
"problems, low-level vision typical WSNM to We subtraction. apply and denoising image e.g., background","We apply WSNM to typical low-level vision problems, e.g., image denoising and background subtraction.",16225
"the computed parameters. corresponding sets, and braingraphs we graph-theoretical have the numerous For computed","For the sets, we have computed the corresponding braingraphs and computed numerous graph-theoretical parameters.",16226
discussed. applications systems is in biology significance The for,The significance for applications in systems biology is discussed.,16227
"the dataset first created assembled this and have MegaFace end, MegaFace challenge. the we To","To this end, we have assembled the MegaFace dataset and created the first MegaFace challenge.",16228
algorithms. state of art results baseline of the and We report,We report results of state of the art and baseline algorithms.,16229
"face variation, etc. occlusion pose, rotation, Head illumination","Head rotation, face pose, illumination variation, occlusion etc.",16230
that increase applications. in of the the attributes complexity recognition of expressions practical spontaneous are,are the attributes that increase the complexity of recognition of spontaneous expressions in practical applications.,16231
database significantly recognition quality the expressions of used. the of depends Effective on,Effective recognition of expressions depends significantly on the quality of the database used.,16232
consist well-known databases Most expression facial expressions. of posed,Most well-known facial expression databases consist of posed expressions.,16233
more also response (CTLs). invite from Cytotoxic this time vigorous However increasing a would T-Lymphocytes,However increasing this time would also invite a more vigorous response from Cytotoxic T-Lymphocytes (CTLs).,16234
image de-noising classical and paper modern and fingerprint studies various This models. reviews,This paper studies and reviews various classical and modern fingerprint image de-noising models.,16235
in the is Information a element transmission information processing brain. for key,Information transmission is a key element for information processing in the brain.,16236
"programs and CPM_Neg_Side.f, CPM_Pos_Side.f find_molec_centr.f, we find_CP_coeffs.f, are: present tangent_sphere.f, FORTRAN here CPM_Zero_Side.f. The six","The six FORTRAN programs we present here are: find_molec_centr.f, tangent_sphere.f, find_CP_coeffs.f, CPM_Neg_Side.f, CPM_Pos_Side.f and CPM_Zero_Side.f.",16237
"the outside The protein of program second atoms the inside, on TS. and number finds","The second program finds the number of protein atoms inside, outside and on the TS.",16238
CP GC. to The tangent is TS the at,The CP is tangent to the TS at GC.,16239
"Specifically, first dictionary data a learn using sparsifying learning. spectral dictionary training we from","Specifically, we first learn a sparsifying dictionary from training spectral data using dictionary learning.",16240
over compressive The approaches. improvement method proposed sensing significant provides conventional the,The proposed method provides significant improvement over the conventional compressive sensing approaches.,16241
the matrix by performance improved matrix using The balancing. reconstruction further is reconditioning sensing,The reconstruction performance is further improved by reconditioning the sensing matrix using matrix balancing.,16242
Feature essential is for selection visual recognition. effective,Feature selection is essential for effective visual recognition.,16243
fully works both feature competitive method and classifier. Our as a mechanism a as selection,Our method works both as a feature selection mechanism and as a fully competitive classifier.,16244
"parts, Most human of multiple steps, or consist subtasks. behaviors","Most human behaviors consist of multiple parts, steps, or subtasks.",16245
"discriminative. colour distributions Moreover, be insufficiently alone can","Moreover, colour distributions alone can be insufficiently discriminative.",16246
mechanisms regarding the the many prominence Despite remain questions its photosymbiosis. unanswered there exact of,Despite its prominence there remain many unanswered questions regarding the exact mechanisms of the photosymbiosis.,16247
pipelines are assemble to intensive reads. such Complex computationally and required,Complex and computationally intensive pipelines are required to assemble such reads.,16248
in video. over features how frame-level time evolve a,how frame-level features evolve over time in a video.,16249
"ranking pooling a learn via a machines. we example, As specific function","As a specific example, we learn a pooling function via ranking machines.",16250
"action various action, generic for our benchmarks method recognition. on We fine-grained evaluate gesture and","We evaluate our method on various benchmarks for generic action, fine-grained action and gesture recognition.",16251
"used proposed to As the improve be localization. frequency can application, wavelets the an","As an application, the proposed wavelets can be used to improve the frequency localization.",16252
further is heuristics. Distinguishing background various by improved foreground from,Distinguishing foreground from background is further improved by various heuristics.,16253
positions trail. motion its a Then motion can emphasized object's be as each duplicating by,Then each object's motion can be emphasized by duplicating its positions as a motion trail.,16254
the clarify relationships. trails objects' spatial These,These trails clarify the objects' spatial relationships.,16255
learning Our in traditional deep advances the approaches aspects. model two,Our model advances the traditional deep learning approaches in two aspects.,16256
reported Subjects pressing keys. their by answers of one two,Subjects reported their answers by pressing one of two keys.,16257
"new propose based techniques. a pipeline for Deep We optical flow computation, Learning on","We propose a new pipeline for optical flow computation, based on Deep Learning techniques.",16258
the results We art present abstract. state the of in,We present state of the art results in the abstract.,16259
relaxations convex propose lifting. a on continuous for spatially We novel framework based functional,We propose a novel spatially continuous framework for convex relaxations based on functional lifting.,16260
be sublabel-accurate to Our a can solution method multilabel as interpreted problems.,Our method can be interpreted as a sublabel-accurate solution to multilabel problems.,16261
"a sense, relaxation. formulation tightest is the convex Moreover, possible our in local","Moreover, in a local sense, our formulation is the tightest possible convex relaxation.",16262
easy is allows efficient It an and implement primal-dual to GPUs. optimization on,It is easy to implement and allows an efficient primal-dual optimization on GPUs.,16263
of on We our vision several problems. approach effectiveness show the computer,We show the effectiveness of our approach on several computer vision problems.,16264
previous a are alleles given the conditional generation at received generation. upon The,The alleles received at a given generation are conditional upon the previous generation.,16265
computer data (Stryphnodendron and to numerical laws We simulations adstringens). real these to apply,We apply these laws to numerical computer simulations and to real data (Stryphnodendron adstringens).,16266
This genetic new implies mutations. diversity the in increase absence of in no,This implies no increase in genetic diversity in the absence of new mutations.,16267
"essential Now, evolution to mutations of obviously, the are species.","Now, obviously, mutations are essential to the evolution of species.",16268
that also mutual a We find of fixation. information allele measure is valid,We also find that mutual information is a valid measure of allele fixation.,16269
"in underwater Moreover, needed. other artificial media, is and lighting","Moreover, underwater and in other media, artificial lighting is needed.",16270
that descattering It move to optimally. also extends platforms,It also extends descattering to platforms that move optimally.,16271
"The optimization taken information from information criterion gain, theory. is","The optimization criterion is information gain, taken from information theory.",16272
setup. in We scaled-down this demonstrate principle a,We demonstrate this principle in a scaled-down setup.,16273
work the generation. future summarize map in also depth the We,We also summarize the future work in the depth map generation.,16274
these extract computations images Using tomographic the distribution.,Using these images tomographic computations extract the distribution.,16275
"off-axis scattering explicitly constrained, irradiance, for highly tomography must scattering. media In natural account and","In highly scattering media and constrained, natural irradiance, tomography must explicitly account for off-axis scattering.",16276
"a handles tomography model. monte-carlo We arbitrary using that orders scattering, formulate of","We formulate tomography that handles arbitrary orders of scattering, using a monte-carlo model.",16277
"Moreover, the is highly our model parallelizable formulation. in","Moreover, the model is highly parallelizable in our formulation.",16278
solve (RT) problems stability conditioning that from We stem in-situ. and transfer radiative modeling,We solve stability and conditioning problems that stem from radiative transfer (RT) modeling in-situ.,16279
"categories), active object and object categories. discover gestures, hands, hand","active hands, hand gestures, object categories), and discover object categories.",16280
"Moreover, observe finetuning Network that Convolutional we accuracy. consistently Neural improves","Moreover, we observe that finetuning Convolutional Neural Network consistently improves accuracy.",16281
"investigate from asynchronous, that possibility may independent oscillations firing. the We neural completely narrowband emerge","We investigate the possibility that narrowband oscillations may emerge from completely asynchronous, independent neural firing.",16282
"them understanding remains However, limited. our of","However, our understanding of them remains limited.",16283
pose and these framework effectiveness. demonstrate We its and generality a energy-minimization as regularized,We pose these as a regularized energy-minimization framework and demonstrate its generality and effectiveness.,16284
game. the relation standard information concurrent context in this We of investigate imperfect,We investigate this relation in the standard context of imperfect information concurrent game.,16285
"Fundamentally, pivot solve or we first. ""where"" they ""what"" around whether","Fundamentally, they pivot around whether we solve ""what"" or ""where"" first.",16286
drone constructs former the for first first. The while for route latter route constructs truck,The former constructs route for drone first while the latter constructs route for truck first.,16287
characteristics results with on obtained and sizes instances are different presented. many Numerical,Numerical results obtained on many instances with different sizes and characteristics are presented.,16288
also are choices algorithm Recommendations promising on provided.,Recommendations on promising algorithm choices are also provided.,16289
"and virtual of robotics. applications repair have Transformations fused kind of this scans, reality, in","Transformations of this kind have applications in virtual reality, repair of fused scans, and robotics.",16290
efficient tools of numerical for optimization We this present solution the problem.,We present numerical tools for the efficient solution of this optimization problem.,16291
We images for network. a method present using single neural objects detecting a in deep,We present a method for detecting objects in images using a single deep neural network.,16292
information is connectome. order Such fundamental human to thoroughly understand in the,Such information is fundamental in order to thoroughly understand the human connectome.,16293
to order objectives. Mixed address Neighborhood Selection in We the aforementioned three simultaneously propose,We propose Mixed Neighborhood Selection in order to simultaneously address the three aforementioned objectives.,16294
in advances revolutionized sequencing genomic next-generation have research. Recent,Recent advances in next-generation sequencing have revolutionized genomic research.,16295
from which present started. supervised approach the We here we also,We also present here the supervised approach from which we started.,16296
choose is a that from pool features. All positively the needs of to correlated algorithm,All that the algorithm needs is to choose from a pool of positively correlated features.,16297
"we small of a for set work, In this features issue. propose this","In this work, we propose a small set of features for this issue.",16298
architecture by learn using hierarchical models. data representations of deep aims It to,It aims to learn hierarchical representations of data by using deep architecture models.,16299
"a (e.g. city, smart lot of a data In","In a smart city, a lot of data (e.g.",16300
many captured automatically processed videos from sensors) analyzed. be distributed need to and,videos captured from many distributed sensors) need to be automatically processed and analyzed.,16301
is image of matching binary descriptors conventionally the using force performed feature Hamming Brute distance.,Brute force matching of binary image feature descriptors is conventionally performed using the Hamming distance.,16302
"McNemar's namely tests, were employed Variance) statistical test ANOVA Two for evaluation. and of (Analysis","Two statistical tests, namely ANOVA (Analysis of Variance) and McNemar's test were employed for evaluation.",16303
descriptors. and metrics Jackard-Needham some display better can Results that Dice show performance for,Results show that Jackard-Needham and Dice metrics can display better performance for some descriptors.,16304
"differences these significant. performance statistically Yet, be were not found to","Yet, these performance differences were not found to be statistically significant.",16305
a small. cell rather is molecules The subcellular structure involved or sometimes of in number,The number of molecules involved in a cell or subcellular structure is sometimes rather small.,16306
"main paper, this In propose tackle this problem. to we two contributions","In this paper, we propose two main contributions to tackle this problem.",16307
"state-of-the-art accuracy. competitive proposed among method very classification achieved has performance, i.e., Our","Our proposed method has achieved very competitive performance, i.e., among state-of-the-art classification accuracy.",16308
"environments, often important This but is problem streaming is intractable. in particularly","This problem is particularly important in streaming environments, but is often intractable.",16309
(KO). new call knock-out We this theoretical by procedure,We call this new procedure by theoretical knock-out (KO).,16310
"algorithmically. To and make devise two this we approaches: algebraically method,","To devise this method, we make two approaches: algebraically and algorithmically.",16311
"In called both state, vector asymptotic we vector. cases compute flux an a on","In both cases we compute a vector on an asymptotic state, called flux vector.",16312
"granularity among relationships classes. naturally the ultra-fine rich Fortunately, object brings","Fortunately, the ultra-fine granularity naturally brings rich relationships among object classes.",16313
the (BGL). rich bipartite-graph relationships approach This paper labels to proposes a novel exploit through,This paper proposes a novel approach to exploit the rich relationships through bipartite-graph labels (BGL).,16314
the thanks show efficient is also structure. inference it bipartite computationally to We that in,We also show that it is computationally efficient in inference thanks to the bipartite structure.,16315
vision computer that be Results techniques. obtained can realistic reconstructions show using sophisticated,Results show that realistic reconstructions can be obtained using sophisticated computer vision techniques.,16316
"network variability modularity. correlation was there clear positive to no Contrary and habitat between expectation,","Contrary to expectation, there was no clear positive correlation between habitat variability and network modularity.",16317
hypothesis that the suggests currently should reconsidered. be widely variability for study This accepted habitat,This study suggests that the currently widely accepted hypothesis for habitat variability should be reconsidered.,16318
component long the dynamics two-species in term the systems. three,the long term dynamics in the three two-species component systems.,16319
industrial synthetic for might This knowledge communities purposes. helpful constructing when be microbial,This knowledge might be helpful when constructing synthetic microbial communities for industrial purposes.,16320
equivalent is from other? the modelling mathematical either to perspective When,When is mathematical modelling from either perspective equivalent to the other?,16321
duality. is ensure also We provide combinatorial to sufficient characterisation a flux-concentration that,We also provide a combinatorial characterisation that is sufficient to ensure flux-concentration duality.,16322
a is the One the statistical of of estimation in distribution. core posterior models problems,One of the core problems in statistical models is the estimation of a posterior distribution.,16323
superior of Extensive behaviors of OPE new methods. our some and demonstrate learning experiments,Extensive experiments demonstrate some superior behaviors of OPE and of our new learning methods.,16324
and of in data tracking the study video. problem multi-target We association,We study the problem of multi-target tracking and data association in video.,16325
"extract representative via the most kinematic parameters the Furthermore, analysis. frequency primary we","Furthermore, we extract the most representative kinematic parameters via the primary frequency analysis.",16326
limit of would What the be capacity? memory,What would be the limit of memory capacity?,16327
contain multiple orientations material datasets corners two with properties. different and These in,These two datasets contain corners in multiple orientations and with different material properties.,16328
with pronounced We in most effects corners. as corners scenes multipath chose are,We chose scenes with corners as multipath effects are most pronounced in corners.,16329
"an by was order magnitude. variance Additionally, of the of lowered error","Additionally, variance of the error was lowered by an order of magnitude.",16330
Exemplar-based success semi-rigid the models of have on localizing objects. achieved parts great,Exemplar-based models have achieved great success on localizing the parts of semi-rigid objects.,16331
"as be objects However, yet highly explored. their is such humans articulated on to efficacy","However, their efficacy on highly articulated objects such as humans is yet to be explored.",16332
an the appears Adaptation of to different as of factors one animal resource. spending,Adaptation of an animal to different factors appears as the spending of one resource.,16333
adaptation. Adaptation hypothetical spent a extensive for energy is quantity,Adaptation energy is a hypothetical extensive quantity spent for adaptation.,16334
"one it as when a quantity, physical literally, i.e. debate takes term causes This much","This term causes much debate when one takes it literally, as a physical quantity, i.e.",16335
role physiological of key Natural evolution plays in mechanisms selection a of adaptation. the,Natural selection plays a key role in the evolution of mechanisms of physiological adaptation.,16336
materials Supplementary the include of in algorithms implementations Python.,Supplementary materials include implementations of the algorithms in Python.,16337
mode background the decomposition (cDMD) for of dynamic modeling. method compressed introduce We,We introduce the method of compressed dynamic mode decomposition (cDMD) for background modeling.,16338
"and are provided. of Results expressivity proposed approach complexity, on the the correctness, the the","Results on the correctness, the complexity, and the expressivity of the proposed approach are provided.",16339
considerable Recurrent classifying have Networks (RNNs) sequences. had Neural in success predicting and,Recurrent Neural Networks (RNNs) have had considerable success in classifying and predicting sequences.,16340
"connection the we between paper, formalisms. the In two theoretical investigate mathematical this","In this theoretical paper, we investigate the mathematical connection between the two formalisms.",16341
two the This highlights stochastic link the transform integral between processes.,This integral transform highlights the link between the two stochastic processes.,16342
"mixed jump-diffusion a well- kinetics this model is simulating paper, stochastic hybrid derived. for In","In this paper, a hybrid jump-diffusion model for simulating well- mixed stochastic kinetics is derived.",16343
as the acts SSA and Langevin bridge the between It chemical a equation. Gillespie,It acts as a bridge between the Gillespie SSA and the chemical Langevin equation.,16344
such are benefits illustrated examples. of a formalism The computational using,The benefits of such a formalism are illustrated using computational examples.,16345
train. difficult more are networks to Deeper neural,Deeper neural networks are more difficult to train.,16346
depth of importance visual of many for The central tasks. is representations recognition,The depth of representations is of central importance for many visual recognition tasks.,16347
"brain of is However, identification fMRI data human in difficult. notoriously its the","However, its identification in fMRI data of the human brain is notoriously difficult.",16348
"call ""Task-related this we Edge Density"" approach (TED). Hence,","Hence, we call this approach ""Task-related Edge Density"" (TED).",16349
"We source provide that reproduces also results. an implementation open our easy-to-use,","We also provide an easy-to-use, open source implementation that reproduces our results.",16350
of methods This applicability the cases. those limits more for realistic,This limits the applicability of those methods for more realistic cases.,16351
deep the neural a of prediction. for attribute We introduce network architecture task relative,We introduce a deep neural network architecture for the task of relative attribute prediction.,16352
to adopt network end-to-end loss in an ranking an whole We fashion. train the appropriate,We adopt an appropriate ranking loss to train the whole network in an end-to-end fashion.,16353
by is coefficients of enforced This corrupted the investigating paper samples. in assumption this,This assumption is enforced in this paper by investigating the coefficients of corrupted samples.,16354
"corrupted the zero. Concretely, we coefficients require of be samples","Concretely, we require the coefficients of corrupted samples be zero.",16355
actions. of of temporal simple Most human more of actions compositions complex consist,Most of human actions consist of complex temporal compositions of more simple actions.,16356
to model action. a sub-actions from complex enables capture hierarchical The,The hierarchical model enables to capture sub-actions from a complex action.,16357
the into changes. handling process in gives methods of capabilities insight This environmental,This process gives insight into the capabilities of methods in handling environmental changes.,16358
under environmental these We tested variations. methods three,We tested these three methods under environmental variations.,16359
point demonstrate the experiments under variations. view method illumination combination Our and of outperformance the,Our experiments demonstrate the outperformance of the combination method under illumination and view point variations.,16360
for model. multi-species developed using Bayesian of coalescent method the inference was A species phylogeny,A method was developed for Bayesian inference of species phylogeny using the multi-species coalescent model.,16361
new the simulation study statistical A properties examine of was method. the performed to,A simulation study was performed to examine the statistical properties of the new method.,16362
"has of numbers loci. species small The impact, some prior particularly trees on for","The prior on species trees has some impact, particularly for small numbers of loci.",16363
(for reanalyzed. rattlesnakes) An was dataset empirical,An empirical dataset (for rattlesnakes) was reanalyzed.,16364
multi-scale known accurate and are recognition. It is important visual for that representations well contextual,It is well known that contextual and multi-scale representations are important for accurate visual recognition.,16365
interest is outside neural recurrent Contextual information the integrated of region networks. using spatial,Contextual information outside the region of interest is integrated using spatial recurrent neural networks.,16366
"levels pooling at multiple information skip use abstraction. scales Inside, extract and to we of","Inside, we use skip pooling to extract information at multiple scales and levels of abstraction.",16367
Description used medical practice. logic routinely is CT) terminology based (DL) biomedical in (SNOMED,Description logic (DL) based biomedical terminology (SNOMED CT) is used routinely in medical practice.,16368
"However, inference diagnostic complexity. its terminology is using such by precluded","However, diagnostic inference using such terminology is precluded by its complexity.",16369
propose a inferential model simplifies Here we that these components.,Here we propose a model that simplifies these inferential components.,16370
and are for suggestions also Limitations these discussed. of improvements processes,Limitations of these processes and suggestions for improvements are also discussed.,16371
reference. mismatched with matched person a reference a one and,one person with a matched reference and a mismatched reference.,16372
"databases, other very promising outperforms our several state-of-the-art challenging and results On achieves approaches. approach","On several challenging databases, our approach achieves very promising results and outperforms other state-of-the-art approaches.",16373
criterion. are some to according matchings that user-specified optimal,matchings that are optimal according to some user-specified criterion.,16374
Practice is Logic paper consideration (TPLP). in of under Theory Programming This and,This paper is under consideration in Theory and Practice of Logic Programming (TPLP).,16375
classification. learning the images is of content-based involving image prevalent tasks One,One of the prevalent learning tasks involving images is content-based image classification.,16376
a the second as vocabulary construction. a adds phase filtering approach of The pre-processing,The second approach adds a filtering phase as a pre-processing of the vocabulary construction.,16377
detection is of Early leading in cancer factor survival rate. the lung,Early detection of lung cancer is the leading factor in survival rate.,16378
"increasing for an of rate. It accurate lung cancer, survival early the detection needs","It needs an accurate early detection of lung cancer, for increasing the survival rate.",16379
"accurate features For all features delete features. detection, need characterizes among efficient it redundancy and","For accurate detection, it need characterizes efficient features and delete redundancy features among all features.",16380
Feature among all informative selection features. of is problem the features selecting,Feature selection is the problem of selecting informative features among all features.,16381
and Youngand by Hong This of University indexed California in database collected Irvine repository. the,This database collected by Hong and Youngand indexed in the University of California Irvine repository.,16382
"the from data, include extracted the etc. contents and Experimental clinical data X-ray","Experimental contents include the extracted from the clinical data and X-ray data, etc.",16383
"study, our efficient for cancer. features of method In lung is new identify proposed","In our study, new method is proposed for identify efficient features of lung cancer.",16384
The Conclusions: of level proposed performance the has better method with highest accuracy.,Conclusions: The proposed method has better performance with the highest level of accuracy.,16385
"symptom model for an recommended identifying of efficient the Therefore, Disease. is proposed","Therefore, the proposed model is recommended for identifying an efficient symptom of Disease.",16386
our of examples. the set outline generalizations We of a illustrative implications via,We outline the implications of our generalizations via a set of illustrative examples.,16387
this re-identification. new propose architecture In for a person work we,In this work we propose a new architecture for person re-identification.,16388
"Multi-task instance-aware Network for segmentation. paper, this present we In semantic Cascades","In this paper, we present Multi-task Network Cascades for instance-aware semantic segmentation.",16389
"networks, differentiating three consists instances, objects. respectively estimating and of categorizing model Our masks,","Our model consists of three networks, respectively differentiating instances, estimating masks, and categorizing objects.",16390
"form networks designed a and cascaded share are structure, features. their to These convolutional","These networks form a cascaded structure, and are designed to share their convolutional features.",16391
"of end-to-end an algorithm this training nontrivial for develop structure. cascaded causal, the We","We develop an algorithm for the nontrivial end-to-end training of this causal, cascaded structure.",16392
accuracy instance-aware segmentation PASCAL demonstrate on VOC. We semantic state-of-the-art,We demonstrate state-of-the-art instance-aware semantic segmentation accuracy on PASCAL VOC.,16393
estimating problem paper spatially-varying of that This exhibit the the reflectance. shape objects addresses of,This paper addresses the problem of estimating the shape of objects that exhibit spatially-varying reflectance.,16394
discriminative variances classes. and between anchor intra-class These capture remain points,These anchor points capture intra-class variances and remain discriminative between classes.,16395
What before can is tumor detect a it? doing we,What is a tumor doing before we can detect it?,16396
The such still argumentation field. research of open an is confidence in issue,The issue of confidence in such argumentation is still an open research field.,16397
estimation complexity arguments. to quantitative an is confidence approach manage interesting of of Providing,Providing quantitative estimation of confidence is an interesting approach to manage complexity of arguments.,16398
"cell cells The (e.g. determines fit lower binary fitness, string with value","The binary string value determines cell fitness, with lower fit cells (e.g.",16399
Fitness defined birth-death are via prisoner's and matrix. the dilemma payoff rates,Fitness and birth-death rates are defined via the prisoner's dilemma payoff matrix.,16400
from sub-populations clonal expansions Dynamic cancer show of phylogenetic initial trees an malignant cell cell.,Dynamic phylogenetic trees show clonal expansions of cancer cell sub-populations from an initial malignant cell.,16401
the in We decision changes making attitude the discuss an fire ground. to at,We discuss the changes in an attitude to decision making at the fire ground.,16402
The changes shift. recent driven by are the technological,The changes are driven by the recent technological shift.,16403
"the theory descriptive human from decision question results of choices. Furthermore, rationality","Furthermore, results from descriptive decision theory question the rationality of human choices.",16404
need searching new making approaches creates and This the during emergency. decision for testing for,This creates the need for searching and testing new approaches for decision making during emergency.,16405
We that need. the propose framework this addresses,We propose the framework that addresses this need.,16406
"of human machine. for the performance allows and decisions, of also comparison The framework between","The framework also allows for comparison of the performance of decisions, between human and machine.",16407
(TPLP) Practice Logic and . Programming in Under Theory of consideration,Under consideration in Theory and Practice of Logic Programming (TPLP) .,16408
The com- a with set mapping is functions. algorithm of patible rich,The algorithm is com- patible with a rich set of mapping functions.,16409
Recurrent neural in many the of learning. deep networks developments forefront stand recent (RNNs) at,Recurrent neural networks (RNNs) stand at the forefront of many recent developments in deep learning.,16410
learning. tools variational deep our This extends in of arsenal,This extends our arsenal of variational tools in deep learning.,16411
are two autoregulation pressure flow. and Baroreflex mechanisms important regulating blood cerebral and (CA) (BR),Baroreflex (BR) and cerebral autoregulation (CA) are two important mechanisms regulating blood pressure and flow.,16412
"unknown. CA and functional However, humans the relationship between is in BR","However, the functional relationship between BR and CA in humans is unknown.",16413
"interaction data Consequently, interpretation. in knowledge is for important of networks biological types","Consequently, knowledge of interaction types in biological networks is important for data interpretation.",16414
generates precisely understanding these clear no what mechanisms of currently is behaviors. There,There is currently no clear understanding of precisely what mechanisms generates these behaviors.,16415
"methods continuous differential systems. time equations we Mathematically, use and dynamical ordinary from","Mathematically, we use methods from ordinary differential equations and continuous time dynamical systems.",16416
in multicellular to organisms. A transition evolutionary is issue unicellular fundamental from biology discussed the,A fundamental issue discussed in evolutionary biology is the transition from unicellular to multicellular organisms.,16417
in of The specialized the diversity of soma. number level determines type's cells,The level of type's diversity determines the number of cells specialized in soma.,16418
"in case. robust may even constraint, we convex energy receive Adding models","Adding energy constraint, we may receive robust models even in convex case.",16419
that with We methods. respect recently reported to approach proposed show is the very competitive,We show that the proposed approach is very competitive with respect to recently reported methods.,16420
higher promising accuracy. shows with classification results The for multiclass method,The method shows promising results for multiclass classification with higher accuracy.,16421
machine in learning. ranking Our the that is attention on focused problem is popular,Our attention is focused on the ranking problem that is popular in machine learning.,16422
"interactions, applicable methods intense need most not many seed-based cases. Second, and are in","Second, most seed-based methods need intense interactions, and are not applicable in many cases.",16423
conservation to of differential leads three of equations. coupled The mass a system,The conservation of mass leads to a coupled system of three differential equations.,16424
solutions All are model the steady-state categorized. of,All steady-state solutions of the model are categorized.,16425
each concentration Explicit expressions are for distribution. given stationary,Explicit expressions are given for each stationary concentration distribution.,16426
address approach by sought introducing the test-driven development ontology this of We authoring. to for,We sought to address this by introducing the approach of test-driven development for ontology authoring.,16427
We the their evaluated test two on performance. approaches,We evaluated the two test approaches on their performance.,16428
a detecting method even thus with JunctionSeq provides low-quality alternative powerful regulation for annotations. isoform,JunctionSeq thus provides a powerful method for detecting alternative isoform regulation even with low-quality annotations.,16429
as available R/Bioconductor an An of JunctionSeq implementation package. is,An implementation of JunctionSeq is available as an R/Bioconductor package.,16430
a countries very business different of challenging variety in this categories The makes problem. large,The large variety of business categories in different countries makes this a very challenging problem.,16431
"of manual this the is scale due to prohibitive Moreover, annotation problem.","Moreover, manual annotation is prohibitive due to the scale of this problem.",16432
"this the human achieves show evaluation, low-recall in In approach the accuracy settings. we","In the evaluation, we show this approach achieves human accuracy in the low-recall settings.",16433
evaluation provide business real world. of also the We end-to-end in discovery an,We also provide an end-to-end evaluation of business discovery in the real world.,16434
"variables. which variables or Causal random other random identifying ""cause"" control with deals inference","Causal inference deals with identifying which random variables ""cause"" or control other random variables.",16435
of framework differential provides ANM Our privacy variety guarantees variants. a for,Our framework provides differential privacy guarantees for a variety of ANM variants.,16436
"We techniques experiments, extensive easy our are to demonstrate that run implement. and and practical","We run extensive experiments, and demonstrate that our techniques are practical and easy to implement.",16437
"use, a Shannon entropy. entropy, also the besides We generalized","We use, besides the Shannon entropy, also a generalized entropy.",16438
"the or are on high-resolution embedding applied then sparse which locally linear models), manifold.","sparse or locally linear embedding models), which are then applied on the high-resolution manifold.",16439
structure component. methods popular single for SFM the of a recovery Incremental are basic most,Incremental SFM methods are most popular for the basic structure recovery of a single component.,16440
are robust but are They effective strictly nature. sequential and in,They are robust and effective but are strictly sequential in nature.,16441
"looking consider by them the systems cells, We making decision with mammalian dynamic as rhythms.","We consider the decision making by mammalian cells, looking them as dynamic systems with rhythms.",16442
"(i.e., or object. of the imaging The can be intrinsic to conditions) variation extrinsic source","The source of variation can be intrinsic or extrinsic (i.e., imaging conditions) to the object.",16443
of items the to images on the of illustrate focus proposed We effectiveness food approach.,We focus on images of food items to illustrate the effectiveness of the proposed approach.,16444
demonstrate model a dataset. MEG We on our,We demonstrate our model on a MEG dataset.,16445
"i.e. regularization, the sampling matrix, the of incoherence the image","the incoherence of the sampling matrix, the image regularization, i.e.",16446
"and prior, technique. sparsity optimization the the","the sparsity prior, and the optimization technique.",16447
"focused this In on we two the mainly last aspects. paper,","In this paper, we mainly focused on the last two aspects.",16448
on vivo simulated both method evaluated The data. and the performance is in of,The performance of the method is evaluated on both simulated and in vivo data.,16449
important learn value. what machine An is people use learning of to,An important use of machine learning is to learn what people value.,16450
shown? user a posts What should be photos or,What posts or photos should a user be shown?,16451
jobs rewarding? activities a person find Which or would,Which jobs or activities would a person find rewarding?,16452
"often from approximate deviate optimality. However, people","However, people often deviate from approximate optimality.",16453
"biases. preferences, and the about We explore models these beliefs, make inferences","We explore the inferences these models make about preferences, beliefs, and biases.",16454
The by ratings experts the unevenness is the entropy. in by measured,The unevenness in the ratings by the experts is measured by entropy.,16455
"case are effectiveness to of carried studies two proposed model. the out Finally, demonstrate","Finally, two case studies are carried out to demonstrate effectiveness of the proposed model.",16456
therefore concentration substances. are They odorous by measurements often of substituted,They are therefore often substituted by concentration measurements of odorous substances.,16457
as seven evaluated The are of mixtures. with their measurements well substances as methods olfactometric,The methods are evaluated with olfactometric measurements of seven substances as well as their mixtures.,16458
the conversion indicate EOC The SOI that values. and deliver results methods reliable,The results indicate that the SOI and EOC conversion methods deliver reliable values.,16459
"and provide diabetic early of hypertension. Retinal retinopathy, signs glaucoma, images","Retinal images provide early signs of diabetic retinopathy, glaucoma, and hypertension.",16460
on based microaneurysms investigated These or signs vessels. be smaller can,These signs can be investigated based on microaneurysms or smaller vessels.,16461
and interrupted; paths disconnected. may bifurcations may Vessel be be crossings also,Vessel paths may also be interrupted; crossings and bifurcations may be disconnected.,16462
"for single a camera. operates seats, passenger all It overhead using","It operates for all passenger seats, using a single overhead camera.",16463
descriptions study obtain discrete how concise multivariate of sequential data. to We,We study how to obtain concise descriptions of discrete multivariate sequential data.,16464
in show Experiments data. planted DITTO that discovers the patterns synthetic correctly,Experiments show that DITTO correctly discovers the patterns planted in synthetic data.,16465
objects pre-trained we frame detect in detectors propose to Hence using object every statically.,Hence we propose to detect objects using pre-trained object detectors in every frame statically.,16466
object Deep are Trained network as models detectors. used,Trained Deep network models are used as object detectors.,16467
"been strategies, on models. or Recent sophisticated more employing effective focused have seeking efforts optimization","Recent efforts have been focused on employing sophisticated optimization strategies, or seeking more effective models.",16468
perceptual units the performs features with and saliency. grouping individuates analysis local of This highest,This analysis performs grouping of local features and individuates perceptual units with the highest saliency.,16469
results transient spatial imaging using and prototype demonstrate We phase-sweep. a,We demonstrate a prototype and transient imaging results using spatial phase-sweep.,16470
one important topics the in is theory. Attribute of set reduction most rough,Attribute reduction is one of the most important topics in rough set theory.,16471
presented been problem. reduction attribute have solve reduction algorithms to the Heuristic attribute,Heuristic attribute reduction algorithms have been presented to solve the attribute reduction problem.,16472
validity can heuristic fitness reduction guarantee attribute of algorithms. monotonicity of functions The the,The monotonicity of fitness functions can guarantee the validity of heuristic attribute reduction algorithms.,16473
"there reduction are However, heuristic attribute on studies reducts. developing algorithms finding distribution for few","However, there are few studies on developing heuristic attribute reduction algorithms for finding distribution reducts.",16474
developed learning is corresponding the proposed framework. for supervised A dictionary algorithm also,A corresponding supervised dictionary learning algorithm is also developed for the proposed framework.,16475
algorithm. proposed digit recognition results on efficacy of the the verify datasets and robustness Experimental,Experimental results on digit datasets recognition verify the efficacy and robustness of the proposed algorithm.,16476
"""human-centric"" noisy exhibiting to as bias. these We refer human reporting annotations","We refer to these noisy ""human-centric"" annotations as exhibiting human reporting bias.",16477
"for annotations image we In visually learning classifiers. use paper, this correct noisy these","In this paper, we use these noisy annotations for learning visually correct image classifiers.",16478
"in reporting versus highly ""what's results are the saying."" Our ""what's for image"" interpretable worth","Our results are highly interpretable for reporting ""what's in the image"" versus ""what's worth saying.""",16479
"tasks. performance Recently, neural have vision convolutional impressive various demonstrated in networks (CNN) computer","Recently, convolutional neural networks (CNN) have demonstrated impressive performance in various computer vision tasks.",16480
"With mobile classify quantized CNN accurately within even devices second. can model, our one images","With our quantized CNN model, even mobile devices can accurately classify images within one second.",16481
Rectified activation convolutional units for deep state-of-the-art linear components important are networks.,Rectified linear activation units are important components for state-of-the-art deep convolutional networks.,16482
"parameters. SReLU four three linear piecewise of Specifically, formulated are consists functions, which by learnable","Specifically, SReLU consists of three piecewise linear functions, which are formulated by four learnable parameters.",16483
achieve high classification research accuracy. extracting Most focuses features current on to complex,Most current research focuses on extracting complex features to achieve high classification accuracy.,16484
the real simulated of both the predictive data. on and smartphone capability algorithm demonstrate We,We demonstrate the predictive capability of the algorithm on both simulated and real smartphone data.,16485
"paradigm recognition. Employing representation, visual the in shifted in mid-level particular,","Employing mid-level representation, in particular, shifted the paradigm in visual recognition.",16486
"In of task. the we image understanding, case detection/recognition on focus object","In the case of image understanding, we focus on object detection/recognition task.",16487
patches specifically We discriminative subcategory-aware the in employ a webly-supervised fashion.,We specifically employ the discriminative patches in a subcategory-aware webly-supervised fashion.,16488
"models the study employing We, for the bias. provided dataset undoing outcomes additionally, by subcategory-based","We, additionally, study the outcomes provided by employing the subcategory-based models for undoing dataset bias.",16489
and labels. Our maps semantically-meaningful to image model their corresponding subregions discovers,Our model discovers and maps semantically-meaningful image subregions to their corresponding labels.,16490
"were performed inferior was and caffe. results A similar on overfeat, to test","A similar test was performed on overfeat, and results were inferior to caffe.",16491
shown to be optimal. rate minimax of is Its convergence,Its rate of convergence is shown to be minimax optimal.,16492
"this popular is Tikhonov regression, ridge approach issue. address to a or regularization,","Tikhonov regularization, or ridge regression, is a popular approach to address this issue.",16493
MODWT We DWT method variable produces less than method. observe estimates the that the,We observe that the MODWT method produces less variable estimates than the DWT method.,16494
"object introduce an works CNNs algorithms. We without G-CNN, based technique detection proposal on which","We introduce G-CNN, an object detection technique based on CNNs which works without proposal algorithms.",16495
bounding grid boxes. a G-CNN multi-scale fixed of with starts,G-CNN starts with a multi-scale grid of fixed bounding boxes.,16496
towards train scale to move We and regressor elements iteratively. of the objects grid a,We train a regressor to move and scale elements of the grid towards objects iteratively.,16497
in are chamber employed automatically Convolutional to networks the MRI detect dataset. LV,Convolutional networks are employed to automatically detect the LV chamber in MRI dataset.,16498
utilized Stacked are autoencoders LV. shape to infer the of the,Stacked autoencoders are utilized to infer the shape of the LV.,16499
Excellent with the achieved. was agreement truth ground,Excellent agreement with the ground truth was achieved.,16500
How homeostasis an problem. size growing maintain fundamental exponentially important cells is,How exponentially growing cells maintain size homeostasis is an important fundamental problem.,16501
Deep Network We (DCNN). Convolutional based on semantic segmentation algorithm Neural a weakly-supervised propose novel,We propose a novel weakly-supervised semantic segmentation algorithm based on Deep Convolutional Neural Network (DCNN).,16502
issues. of networks training deep neural Effective from two suffers main,Effective training of deep neural networks suffers from two main issues.,16503
The the that models curvature. first spaces is of exhibit parameter these pathological,The first is that the parameter spaces of these models exhibit pathological curvature.,16504
by adaptive (SGD). Recent this using Descent methods address problem for Stochastic preconditioning Gradient,Recent methods address this problem by using adaptive preconditioning for Stochastic Gradient Descent (SGD).,16505
by methods of adapting the These convergence parameter to local improve geometry space.,These methods improve convergence by adapting to the local geometry of parameter space.,16506
"addressed overfitting, stopping. second is is by A typically early issue which","A second issue is overfitting, which is typically addressed by early stopping.",16507
"this that However, work problem. averaging demonstrated has mitigates recent model Bayesian","However, recent work has demonstrated that Bayesian model averaging mitigates this problem.",16508
Gradient The Stochastic by Dynamics using can (SGLD). posterior be Langevin sampled,The posterior can be sampled by using Stochastic Gradient Langevin Dynamics (SGLD).,16509
"the SGLD However, rapidly curvature renders methods default changing inefficient.","However, the rapidly changing curvature renders default SGLD methods inefficient.",16510
"preconditioners with we propose Here, SGLD. adaptive combining","Here, we propose combining adaptive preconditioners with SGLD.",16511
"In predictive idea, risk. of we convergence on support give and properties theoretical this asymptotic","In support of this idea, we give theoretical properties on asymptotic convergence and predictive risk.",16512
"and Nowadays, are to hospitals modern ubiquitous society. integral","Nowadays, hospitals are ubiquitous and integral to modern society.",16513
identifying discusses inference a to probabilistic approach variables level. at paper detection model-based a This,This paper discusses a model-based probabilistic inference approach to identifying variables at a detection level.,16514
determination structure a crystallography. in macromolecular by key Crystallization is step,Crystallization is a key step in macromolecular structure determination by crystallography.,16515
complete of Computational correctly requires data. formatted crystallization and analysis,Computational analysis of crystallization requires complete and correctly formatted data.,16516
discussed. needed steps The the this and to data for potential this are potential realize,The potential for this data and the steps needed to realize this potential are discussed.,16517
region proposals. accurate State-of-the-art on rely an object set systems of detection,State-of-the-art object detection systems rely on an accurate set of region proposals.,16518
Several architecture recent methods to promising object network neural a locations. use hypothesize,Several recent methods use a neural network architecture to hypothesize promising object locations.,16519
islets homeostasis. units Pancreatic in glucose involved are functional,Pancreatic islets are functional units involved in glucose homeostasis.,16520
of Our computational a principle islets. suggests design is synchronization study pancreatic that controllable,Our computational study suggests that controllable synchronization is a design principle of pancreatic islets.,16521
cognition. life for requirements minimal The autopoiesis and are,The minimal requirements for life are autopoiesis and cognition.,16522
propose models classes and We simulation. of with three autopoietic perform evolutionary cognition,We propose autopoietic models with cognition and perform three classes of evolutionary simulation.,16523
cells show cognitive adaptation evolution. The the the and,The cognitive cells show the adaptation and the evolution.,16524
interaction the The environment cells. the with also shows via of evolution the system,The environment also shows the evolution via the interaction with the system of cells.,16525
prototype This and the of the its is living system co-evolution environment. a of,This is a prototype of the co-evolution of the living system and its environment.,16526
recovered via based map standard The recovery Compressed technique. Sensing PD is,The PD map is recovered via standard Compressed Sensing based recovery technique.,16527
biopsy might preparation. during occur cracks) which,cracks) which might occur during biopsy preparation.,16528
the fast procedure Gibbs sampling. by easy-to-implement is The performed estimation using collapsed and,The estimation procedure is performed by using the fast and easy-to-implement collapsed Gibbs sampling.,16529
"the that apply We well. synthetic it performs method data, to showing proposed","We apply the proposed method to synthetic data, showing that it performs well.",16530
"problem. short, factorization matrix show k-means we clustering that is a In","In short, we show that k-means clustering is a matrix factorization problem.",16531
significant attention generated using Learning has Bayesian recently. in methods deep models,Learning in deep models using Bayesian methods has generated significant attention recently.,16532
"accurate, demonstrate the proposed more faster. is robust, We and converges scheme that","We demonstrate that the proposed scheme is more accurate, robust, and converges faster.",16533
demonstrated properties Bayesian These are learning. be desirable in deep to,These properties are demonstrated to be desirable in Bayesian deep learning.,16534
is Dicty. activation to first response the polarized gradients in cAMP Ras,Ras activation is the first polarized response to cAMP gradients in Dicty.,16535
is data outliers detect management. in meaningful It to traffic traffic for,It is meaningful to detect outliers in traffic data for traffic management.,16536
"a task for people large-scale to However, massive distinguish outliers. is this database from","However, this is a massive task for people from large-scale database to distinguish outliers.",16537
are normal distributed. assume data that We inlier,We assume that inlier data are normal distributed.,16538
performances. paper modeling address method of the would and This evaluate respective their each,This paper would address the modeling of each method and evaluate their respective performances.,16539
often making is on based Bayesian Decision networks.,Decision making is often based on Bayesian networks.,16540
blocks Bayesian building conditional are networks (CPTs). its tables for The probability,The building blocks for Bayesian networks are its conditional probability tables (CPTs).,16541
insufficient of approximations. representations knowledge these are Some,Some of these knowledge representations are insufficient approximations.,16542
observations Using to knowledge decisions. fusion and of predictive lead cause better effect,Using knowledge fusion of cause and effect observations lead to better predictive decisions.,16543
two probability The are of novel first space. mapping ways conditional to expectations the,The first two are novel ways of mapping conditional expectations to the probability space.,16544
"of explore Third, problem. this the learning for idea we user-guided","Third, we explore the idea of user-guided learning for this problem.",16545
study we stability. selection Here maintaining protein,Here we study selection maintaining protein stability.,16546
"negatively is Also, G$. to proportional the distribution of mean $\Delta the","Also, the mean of the distribution is negatively proportional to $\Delta G$.",16547
product of shapes We capacity. to modeling of introduce increase combinations also Kronecker different,We also introduce combinations of different shapes of Kronecker product to increase modeling capacity.,16548
"For and sparse perfect mean detection, namely, structures, consider set matching. combinatorial two we normal","For normal mean detection, we consider two combinatorial structures, namely, sparse set and perfect matching.",16549
"aggregators, preliminary as four as results report of we implemented discuss well We tests.","We discuss four implemented aggregators, as well as we report results of preliminary tests.",16550
This paper for specifies notation decision Markov a processes.,This paper specifies a notation for Markov decision processes.,16551
"limitation those scalability. the crucial of models,however, The is","The crucial limitation of those models,however, is the scalability.",16552
The stacking presents paper application attributes. non-linear Go an prediction for of ensembles of player,The paper presents an application of non-linear stacking ensembles for prediction of Go player attributes.,16553
depends in spread strongly on network. population's Epidemic systems the contact single-host,Epidemic spread in single-host systems strongly depends on the population's contact network.,16554
the Numeric validate to simulations were formulations. used analytic,Numeric simulations were used to validate the analytic formulations.,16555
"Hence, probabilities dynamics inter-specific networks. disease in infection multihost shapes asymmetry in","Hence, inter-specific asymmetry in infection probabilities shapes disease dynamics in multihost networks.",16556
multihost asymmetric Our and systems advances understanding with of non-biological rates. flow study systems,Our study advances understanding of multihost systems and non-biological systems with asymmetric flow rates.,16557
mutations identifying in could therapeutic targets. This and aid driver,This could aid in identifying driver mutations and therapeutic targets.,16558
"variables. target Using utilized the to evaluations different can machine algorithms, learning be relevant predict","Using machine learning algorithms, the evaluations can be utilized to predict different relevant target variables.",16559
strength playing of We predict and style the methodology the to (e.g. player this apply,We apply this methodology to predict the strength and playing style of the player (e.g.,16560
calls of connectivity these estimated populations Comparing functional in GGMs. subjects two for comparing,Comparing functional connectivity of subjects in two populations calls for comparing these estimated GGMs.,16561
in to similar structure. is known Our differences goal identify have to GGMs,Our goal is to identify differences in GGMs known to have similar structure.,16562
Sparse even statistical interpretable high-dimensional and low-sample in enable and guarantees penalties settings. models,Sparse penalties enable statistical guarantees and interpretable models even in high-dimensional and low-sample settings.,16563
"The approach discussed article, i.e. this in","The discussed approach in this article, i.e.",16564
"aware not are of comparative we supervoxel Indeed, single segmentation. a on study","Indeed, we are not aware of a single comparative study on supervoxel segmentation.",16565
"available of objective However, no these error any functions. bounds for modified are","However, no error bounds are available for any of these modified objective functions.",16566
"reasoning in Skolemization. of permit we theories, describe fine-grained formally cyclic variations limiting To","To permit reasoning in cyclic theories, we formally describe fine-grained variations of limiting Skolemization.",16567
"three top system A a with population predator population, i.e.","A three population system with a top predator population, i.e.",16568
categories. learning multi-task simultaneously to visual different share allows models attribute The among knowledge CNN,The multi-task learning allows CNN models to simultaneously share visual knowledge among different attribute categories.,16569
"other classifiers Furthermore, can their statistics classifiers shared to improve from performance. under-sampled leverage","Furthermore, under-sampled classifiers can leverage shared statistics from other classifiers to improve their performance.",16570
method of our effectiveness two popular the attribute datasets. on We show,We show the effectiveness of our method on two popular attribute datasets.,16571
"Euclidean only However, in to space. vector applies data SSC","However, SSC applies only to vector data in Euclidean space.",16572
verification. effective minor for the contain needed all algorithms filters object of Cascade detection,Cascade filters contain all of the minor algorithms needed for effective object detection verification.,16573
"can a mix deconvolution. This quantile between as technique be matching, described and","This technique can be described as a mix between quantile matching, and deconvolution.",16574
a validation. provide and empirical it theoretical an We for,We provide for it a theoretical and an empirical validation.,16575
"and states, timestamps. events model, actions canonical are with other In represented the","In the canonical model, states, actions and other events are represented with timestamps.",16576
"semantics This operational sound, incomplete. is but","This operational semantics is sound, but incomplete.",16577
won lottery his prize. the main a wife organized John state and,John organized a state lottery and his wife won the main prize.,16578
theory notion have not does events. of probability Traditional the even random,Traditional probability theory does not even have the notion of random events.,16579
"technique has The Lena, Elaine, on tested proposed been Baboon. and Pepper,","The proposed technique has been tested on Lena, Elaine, Pepper, and Baboon.",16580
model generative Machine is data. Restricted importan- t Boltzmann vectorial an modeling (RBM),Restricted Boltzmann Machine (RBM) is an importan- t generative model modeling vectorial data.,16581
"images, to While have vec- practice the torized. be data to an applying RBM in","While applying an RBM in practice to images, the data have to be vec- torized.",16582
got vectorization. results lost able and in data This has valu- high-dimensional spatial information in,This results in high-dimensional data and valu- able spatial information has got lost in vectorization.,16583
"a understanding still However, phenomenon is performance complete of the lacking. gain","However, a complete understanding of the performance gain phenomenon is still lacking.",16584
"from statistical perspective. study a learning gain performance paper, this In the we phenomenon","In this paper, we study the performance gain phenomenon from a statistical learning perspective.",16585
"BusPlus the hours Australia. aims service project in The public improving transit Canberra, off-peak at","The BusPlus project aims at improving the off-peak hours public transit service in Canberra, Australia.",16586
in process The processes convergent explicit implicit or creative and both involves either manner. divergent,The creative process involves both divergent and convergent processes in either implicit or explicit manner.,16587
experimentations robotic An is painter a example using of artificial by in platform. some described,An example of artificial painter is described in some experimentations by using a robotic platform.,16588
House We View Street variable Number on images. conduct containing number experiments highly dataset,We conduct experiments on Street View House Number dataset containing highly variable number images.,16589
the methods. superiority method results the demonstrate of HMM The traditional based over proposed,The results demonstrate the superiority of the proposed method over traditional HMM based methods.,16590
"for proposed. approach is novel coding data this In paper, nominal a","In this paper, a novel approach for coding nominal data is proposed.",16591
"a the rank is nominal complex of form For number assigned. given in a data,","For the given nominal data, a rank in a form of complex number is assigned.",16592
been used these based can approach classification. on knew The properties for,The approach based on these knew properties can been used for classification.,16593
chemicals cells cellular states the the these chondrocyte according to describes component different release. The,The cellular component describes the different chondrocyte states according to the chemicals these cells release.,16594
the The chemical of component change in models concentrations those chemicals.,The chemical component models the change in concentrations of those chemicals.,16595
"attention representations powerful of points. covariance set descriptors much received of Recently, as have","Recently, covariance descriptors have received much attention as powerful representations of set of points.",16596
"results yields promising approach that show recognition experimental pattern on Furthermore, we tasks. our","Furthermore, we show that our approach yields promising experimental results on pattern recognition tasks.",16597
public on method The several datasets. tested proposed is,The proposed method is tested on several public datasets.,16598
most state-of-the-art method outperforms experimental show the methods. The results,The experimental results show the method outperforms most state-of-the-art methods.,16599
represents factorization rank low commonly matrix (LRMF) The subspace utilized a learning strategy.,The low rank matrix factorization (LRMF) represents a commonly utilized subspace learning strategy.,16600
"tensor. calibrated three role similar trifocal to the In a assigned views, is","In three views, a similar role is assigned to the calibrated trifocal tensor.",16601
this paper. are Some in them described of,Some of them are described in this paper.,16602
"--- essential new notion a specifically, matrix. More we the trifocal define","More specifically, we define a new notion --- the trifocal essential matrix.",16603
"necessary Based we conditions three a tensor. these calibrated on propose characterizations, on trifocal","Based on these characterizations, we propose three necessary conditions on a calibrated trifocal tensor.",16604
rules. by consider excitatory and STDP inhibitory Hebbian synapses We driven of,We consider STDP of excitatory and inhibitory synapses driven by Hebbian rules.,16605
"envolves topology. all-to-all initial an Specifically, complex topology to a","Specifically, an initial all-to-all topology envolves to a complex topology.",16606
results demonstrate of gains approach the methods. existing performance image significant new restoration Experimental over,Experimental results demonstrate significant performance gains of the new approach over existing image restoration methods.,16607
FG-NET datasets. method applied PAL our on We and,We applied our method on FG-NET and PAL datasets.,16608
hand domain construct. rely on to These methods knowledge crafted,These hand crafted methods rely on domain knowledge to construct.,16609
to training would learn the and We based filters optimal features like examples. on,We would like to learn the optimal filters and features based on training examples.,16610
problems using Networks approaches. address the detection and Neural Deep (DNNs) classification vehicle We,We address the vehicle detection and classification problems using Deep Neural Networks (DNNs) approaches.,16611
"set, this by is each views. data different points In multiple data presented","In this data set, each data points is presented by multiple different views.",16612
"method we this this propose for In novel problem. paper, a","In this paper, we propose a novel method for this problem.",16613
two assumptions. method on This based is,This method is based on two assumptions.,16614
objective imposed also to terms function. in are Some regularization different the variables,Some regularization terms are also imposed to different variables in the objective function.,16615
novel desirable classified and investigated are by properties. and integration various Existing measures,Existing and novel integration measures are investigated and classified by various desirable properties.,16616
is a This on manifold. naturally Grassmann problem as described clustering,This is naturally described as a clustering problem on Grassmann manifold.,16617
in in applications many data method analysis computer new tasks. The vision has,The new method has many applications in data analysis in computer vision tasks.,16618
The models a have of data practical evaluated been proposed applications. analysis on number,The proposed models have been evaluated on a number of practical data analysis applications.,16619
been to benchmark developed this for have approaches Several different problem solve datasets.,Several approaches have been developed to solve this problem for different benchmark datasets.,16620
"novel present end, a facial expression this approach towards recognition. we To paper in this","To this end, in this paper we present a novel approach towards facial expression recognition.",16621
with domain deep encoding fuse patches. We features rich through discriminant knowledge facial,We fuse rich deep features with domain knowledge through encoding discriminant facial patches.,16622
We datasets; two experiments CK TFE. conduct on most popular of and benchmark the,We conduct experiments on two of the most popular benchmark datasets; CK and TFE.,16623
temporal untrimmed long We action localization videos. in address,We address temporal action localization in untrimmed long videos.,16624
Only network prediction. during are network and used the localization proposal the,Only the proposal network and the localization network are used during prediction.,16625
present formulation probabilistic graph We model for a the problem. clustering graphical,We present a probabilistic graphical model formulation for the graph clustering problem.,16626
problem approach works represented graphically Our any for instance.,Our approach works for any graphically represented problem instance.,16627
for is cluster demonstrated network social analysis. and image segmentation This,This is demonstrated for image segmentation and social network cluster analysis.,16628
relevant also for ansatz problems. combinatorial Our mathematical be other should,Our mathematical ansatz should be relevant also for other combinatorial problems.,16629
the description language as used the A of Essence' Savile Row. tool by,A description of the Essence' language as used by the tool Savile Row.,16630
"combine and we subset algorithm popular this In (Au the article, Probab. simulation Beck,","In this article, we combine the popular subset simulation algorithm (Au and Beck, Probab.",16631
simulation called resulting algorithm Bayesian is (BSS). subset The,The resulting algorithm is called Bayesian subset simulation (BSS).,16632
We model as stalled and recovered. the systems: forks two-state,We model the forks as two-state systems: stalled and recovered.,16633
(e.g. concepts mid-level semantic \emph{birthday by party}) described can multiple be,\emph{birthday party}) can be described by multiple mid-level semantic concepts (e.g.,16634
proposed results confirm the superiority experimental approach. the The of,The experimental results confirm the superiority of the proposed approach.,16635
a Animals wide psychophysical tasks. range probabilistic of in inference perform near-optimal,Animals perform near-optimal probabilistic inference in a wide range of psychophysical tasks.,16636
and work with task-dependent implemented networks using hand-crafted operations. has such computations Previous neural,Previous work has implemented such computations using neural networks with hand-crafted and task-dependent operations.,16637
sparsity-based novel population code. probabilistic develop trained a The networks,The trained networks develop a novel sparsity-based probabilistic population code.,16638
"connectivity, allowed us point-based the persistency of i.e. brain estimate Moreover, the to approach","Moreover, the point-based approach allowed us to estimate the persistency of brain connectivity, i.e.",16639
of patterns. trace duration resting-state connectivity or the memory intrinsic of the,the duration of the intrinsic trace or memory of resting-state connectivity patterns.,16640
the justified. are geometric of codewords mathematically learned then The properties,The geometric properties of the learned codewords are then mathematically justified.,16641
"hypothesis matched imposed which improves in through codewords, are consistency temporal constraints In addition, precision.","In addition, hypothesis constraints are imposed through temporal consistency in matched codewords, which improves precision.",16642
the little demonstrate Experiments with of runtime. precision/recall improved outperforming in art state loss,Experiments demonstrate improved precision/recall outperforming state of the art with little loss in runtime.,16643
"lower breakdown for more rapid, is level. the The trophic however,","The breakdown is more rapid, however, for the lower trophic level.",16644
between deterministic The degree species different can influence communities. then synchrony in the dynamics of,The deterministic dynamics can then influence the degree of synchrony between species in different communities.,16645
minimally Intuitively direction. this are damped fluctuations in,Intuitively fluctuations are minimally damped in this direction.,16646
"dictionary We a enable localized that develop design i.e. new geometric, block-based can","We develop a new localized block-based dictionary design that can enable geometric, i.e.",16647
"learning is performance and dictionary to method efficiency. a increase Further, incorporated","Further, a dictionary learning method is incorporated to increase performance and efficiency.",16648
minimization energy matching perform problem. as two-step to a propose stereo,propose to perform stereo matching as a two-step energy minimization problem.,16649
first a connected solve they fully For model. the step MRF,For the first step they solve a fully connected MRF model.,16650
mean The solved approximations. is by field model joint,The joint model is solved by mean field approximations.,16651
Averaging Dynamic Functions The called Weighted resulting functions Ordered are DYOWAs. ---,The resulting functions are called Dynamic Ordered Weighted Averaging Functions --- DYOWAs.,16652
to used described able The both are in here cases. be operators,The operators described here are able to be used in both cases.,16653
al. methodology the image reduction In terms provided by et Patermain we of apply,In terms of image reduction we apply the methodology provided by Patermain et al.,16654
can this memory data; memories The ability from brain reproduce critical recall. for partial is,The brain can reproduce memories from partial data; this ability is critical for memory recall.,16655
stored of memory. to the contain converges This which reliably model patterns kind,This kind of model reliably converges to stored patterns which contain the memory.,16656
models dynamics different plausible The memory of give account retrieval. of a sorts of our,The dynamics of our models give a plausible account of different sorts of memory retrieval.,16657
for and Annotating useful indexing tags is images. retrieving images with,Annotating images with tags is useful for indexing and retrieving images.,16658
"many data available annotation missing include inaccurate However, or annotations.","However, many available annotation data include missing or inaccurate annotations.",16659
utilize subspace clustering subspace the tag We for completion. via data property sparse of,We utilize the subspace property of data via sparse subspace clustering for tag completion.,16660
cues from individual image. are detections character derived an The bottom-up from,The bottom-up cues are derived from individual character detections from an image.,16661
"from prior, language statistics. i.e., obtained are interactions top-down These lexicon-based a cues","These interactions are top-down cues obtained from a lexicon-based prior, i.e., language statistics.",16662
such vertebral especially pathology challenging the compression is in of presence Segmentation as fractures.,Segmentation is especially challenging in the presence of pathology such as vertebral compression fractures.,16663
"movements pathways, functions, about brain neural information contain Eye abundant etc. cognitive","Eye movements contain abundant information about cognitive brain functions, neural pathways, etc.",16664
"eye the into In proposed data classified fixations and saccades. method, is movement","In the proposed method, eye movement data is classified into fixations and saccades.",16665
fusion is in classify approach layer. data the score to A adopted output the,A score fusion approach is adopted to classify the data in the output layer.,16666
as eye modality. strength indicate a pattern results The movement of the biometric,The results indicate the strength of eye movement pattern as a biometric modality.,16667
"a semantic describe However, contents. model to provide they rarely their","However, they rarely provide a semantic model to describe their contents.",16668
key the data are ingredients into models automatically to Such knowledge graphs. the publish,Such models are the key ingredients to automatically publish the data into knowledge graphs.,16669
work fields the of on attributes). (source the data of semantic focuses annotation Most related,Most of the related work focuses on semantic annotation of the data fields (source attributes).,16670
amounts present picturesque an large for from egocentric We data. highlights of identifying approach video,We present an approach for identifying picturesque highlights from large amounts of egocentric video data.,16671
"to properties we videos specifically Furthermore, leverage egocentric the highlight of our improve detection.","Furthermore, we specifically leverage the properties of egocentric videos to improve our highlight detection.",16672
photographic method and non-photo-realistic apply (artwork) the both tasks. to synthesis We,We apply the method to both photographic and non-photo-realistic (artwork) synthesis tasks.,16673
examples. the Fundamental of explored properties and through torque are illustrated,Fundamental properties of the torque are explored and illustrated through examples.,16674
extensions of multiple underway. are CFTV EMTV and and structures for The nonlinear extracting,The extensions of CFTV and EMTV for extracting multiple and nonlinear structures are underway.,16675
is this An version. included in addendum arXiv,An addendum is included in this arXiv version.,16676
d(y) show operator. to that integral V(x) is from compact mapping the a We,We show that the mapping from d(y) to V(x) is a compact integral operator.,16677
unbounded Its sense. ill-posed the which makes inverse problem inverse the in is mathematical,Its inverse is unbounded which makes the inverse problem ill-posed in the mathematical sense.,16678
latent We that individualizes predictions a of trajectories. hierarchical disease variable propose model,We propose a hierarchical latent variable model that individualizes predictions of disease trajectories.,16679
state-of-the-art compare accuracy. in We demonstrate significant approach against improvements our and predictive,We compare our approach against state-of-the-art and demonstrate significant improvements in predictive accuracy.,16680
"databases. two C-IQA with methods widely on IQA is Further, NR-IQA used state-of-the-art compared other","Further, C-IQA is compared with other state-of-the-art NR-IQA methods on two widely used IQA databases.",16681
for pipelines. object is detection state-of-the-art essential Object proposal current,Object proposal is essential for current state-of-the-art object detection pipelines.,16682
"results generally methods with producing However, localization existing fail satisfying the proposal accuracy. in","However, the existing proposal methods generally fail in producing results with satisfying localization accuracy.",16683
objects The quite however in are which common is worse even small practice. case for,The case is even worse for small objects which however are quite common in practice.,16684
"far many that still in remains which cases of from humans. performance However, are there","However, there are still many cases in which performance remains far from that of humans.",16685
benchmark also We a make new publicly available.,We also make a new benchmark publicly available.,16686
"compositional our we designed generative mechanism. To idea, models two implement with such particular a","To implement our idea, we designed two particular generative models with such a compositional mechanism.",16687
"an essential robust for is and pupil prerequisite eye-tracking. video-based Real-time, accurate, detection pervasive","Real-time, accurate, and robust pupil detection is an essential prerequisite for pervasive video-based eye-tracking.",16688
upon data Annotated request. made be sets available can,Annotated data sets can be made available upon request.,16689
"of which drawbacks control its has each own can these conditions. some method However,","However, each method has its own drawbacks which can control some of these conditions.",16690
"and different of eye will this categorized detection be In methods paper, explained.","In this paper, different methods of eye detection will be categorized and explained.",16691
"each method category, In and be will presented. of disadvantages each advantages the","In each category, the advantages and disadvantages of each method will be presented.",16692
on difficult show performance We face substantive improvement datasets. three thermal-visible,We show substantive performance improvement on three difficult thermal-visible face datasets.,16693
paper the in influence finetuning factors investigates detection. that for performance many This object,This paper investigates many factors that influence the performance in finetuning for object detection.,16694
for is sample of long-tailed distribution object numbers classes in a detection. There,There is a long-tailed distribution of sample numbers for classes in object detection.,16695
across classes. it more And to the uniform is make number better sample,And it is better to make the sample number more uniform across classes.,16696
be equally can considered multiple Generic tasks. detection important as object,Generic object detection can be considered as multiple equally important tasks.,16697
is each class task. a Detection of,Detection of each class is a task.,16698
appearance classes/tasks their in representation. These discriminative individuality visual have,These classes/tasks have their individuality in discriminative visual appearance representation.,16699
is scheme hierarchical feature proposed. learning A,A hierarchical feature learning scheme is proposed.,16700
of the T$_{\textrm{B}}$ were probability the disaggregated to functions density similar SMAP-T$_{\textrm{B}}$. The,The probability density functions of the disaggregated T$_{\textrm{B}}$ were similar to the SMAP-T$_{\textrm{B}}$.,16701
problem to based is solve on model. the proposed An effective optimization algorithm GPSSVR the,An effective algorithm is proposed to solve the optimization problem based on the GPSSVR model.,16702
show methods other outperform proposed methods. obviously experimental results The the that state-of-the-art,The experimental results show that the proposed methods obviously outperform other state-of-the-art methods.,16703
and thus engineers is importance for authorities. Visibility distance of road,Visibility distance is thus of importance for road engineers and authorities.,16704
a Terrestrial Mobile second involves Mapping of LIDAR the System. use approach The,The second approach involves the use of a Terrestrial LIDAR Mobile Mapping System.,16705
approaches developed SARI-VIZIR in were the These context the of project. PREDIT,These approaches were developed in the context of the SARI-VIZIR PREDIT project.,16706
"described, are compared. evaluated and Both approaches","Both approaches are described, evaluated and compared.",16707
Their pros following also are cons and to vehicle systems discussed. respect with,Their pros and cons with respect to vehicle following systems are also discussed.,16708
study. of The proposed the efficiency and a stability simulation algorithm in are illustrated,The efficiency and stability of the proposed algorithm are illustrated in a simulation study.,16709
any for complexity prohibitive their application. computational practical is Unfortunately,Unfortunately their computational complexity is prohibitive for any practical application.,16710
networks activity. Many variable neuronal exhibit spiking biological highly,Many biological neuronal networks exhibit highly variable spiking activity.,16711
networks offer this of model a Balanced variability. parsimonious,Balanced networks offer a parsimonious model of this variability.,16712
"our underscore phenotyping of importance the Together, longitudinal in neuroscience. cognitive results","Together, our results underscore the importance of longitudinal phenotyping in cognitive neuroscience.",16713
common is the to recognition One optimization visual nonlinear model approach problem. as,One common approach is to model the visual recognition as nonlinear optimization problem.,16714
this essential between distances the is units for solving problem. Measuring visual,Measuring the distances between visual units is essential for solving this problem.,16715
show behind the kernel. We theory each,We show the theory behind each kernel.,16716
provides framework to different This way explore kernels. convenient,This framework provides convenient way to explore different kernels.,16717
additive sparse model. We selection for of generalized present method a the variable,We present a method of variable selection for the sparse generalized additive model.,16718
the forward of It takes incremental stagewise regression. form,It takes the form of incremental forward stagewise regression.,16719
"we popular In machine competitive simulations, against the show learning approaches. method is new","In simulations, we show the new method is competitive against popular machine learning approaches.",16720
We some performance real also using its datasets. demonstrate,We also demonstrate its performance using some real datasets.,16721
"(e.g. Among we cancer drivers genes, the breast of several recover cis-associated known","Among the cis-associated genes, we recover several known drivers of breast cancer (e.g.",16722
be available made soon. will code Source,Source code will be made available soon.,16723
that We also of redundancy increases pathway. the show signalling fidelity,We also show that redundancy increases fidelity of the signalling pathway.,16724
False CNN second plate/non-plate the then positives by eliminated classifier. are,False positives are then eliminated by the second plate/non-plate CNN classifier.,16725
extracts effectively and both high cascade with plates The proposed recall license framework precision.,The proposed cascade framework extracts license plates effectively with both high recall and precision.,16726
"to license labelling} propose characters Last, recognize the {sequence as we problem. a","Last, we propose to recognize the license characters as a {sequence labelling} problem.",16727
of it that this is approach is free. main The advantage segmentation,The main advantage of this approach is that it is segmentation free.,16728
is The based the concept distinct validation on in optic disc. vessels structures,The validation concept is based on distinct vessels structures in the optic disc.,16729
the analyses are theoretic proposed model and given Some for algorithm.,Some theoretic analyses are given for the proposed model and algorithm.,16730
"ongoing it planners. favorably human compared with In experiments,","In ongoing experiments, it compared favorably with human planners.",16731
"approach processes Interleaved comprise attrition consumption and computational this tool. routing, the of planning, scheduling,","Interleaved planning, scheduling, routing, attrition and consumption processes comprise the computational approach of this tool.",16732
runtime We is practice in essentially linear that demonstrate $m\!\cdot\! in also the,We also demonstrate that in practice the runtime is essentially linear in $m\!\cdot\!,16733
it n$ analysis method and making shape shape for efficient an retrieval.,n$ making it an efficient method for shape analysis and shape retrieval.,16734
classification geometric according surfaces their the to of structure. fine,the classification of surfaces according to their geometric fine structure.,16735
type prevalence of occurrence and syndrome characteristics The each are symptom determined.,The prevalence and symptom occurrence characteristics of each syndrome type are determined.,16736
structural run-of-the-mill a areas in (SASA) biology. solvent Calculating is accessible surface calculation,Calculating solvent accessible surface areas (SASA) is a run-of-the-mill calculation in structural biology.,16737
"and stable library well-documented, is efficient. The","The library is well-documented, stable and efficient.",16738
"we an validate environment by empirically algorithm our psychology. inspired Finally, behavioral on","Finally, we empirically validate our algorithm on an environment inspired by behavioral psychology.",16739
main The of forest resources of forest modern regeneration management. of is the one objectives,The regeneration of forest resources is one of the main objectives of modern forest management.,16740
investigating identified techniques and methods Appropriate intensively. computer was technology used for regeneration were and,Appropriate methods and techniques were identified for investigating regeneration and computer technology was used intensively.,16741
thesis. Nine order the in were to achieve software the objectives stand-alone programs of produced,Nine stand-alone software programs were produced in order to achieve the objectives of the thesis.,16742
"a In be we parallelized. propose that discrete algorithm massively can optimization setting, novel the","In the discrete setting, we propose a novel optimization algorithm that can be massively parallelized.",16743
"reproduction of cost first, has a in diffusing faster In rate. the terms","In the first, diffusing faster has a cost in terms of reproduction rate.",16744
"distributed resources in uniformly second the case, space. In are not","In the second case, resources are not uniformly distributed in space.",16745
"are fluid third flow. In the a two transported species by case, the","In the third case, the two species are transported by a fluid flow.",16746
illustrate the Experimental one merits of on hyperspectral proposed image the dataset approach. results,Experimental results on one hyperspectral image dataset illustrate the merits of the proposed approach.,16747
"we the input descriptor to characteristics. paper, In this shape the according select propose to","In this paper, we propose to select the descriptor according to the input shape characteristics.",16748
an obtain to intermediate reconstructed The reconstruction the image. into final off-the-shelf fed is denoiser,The intermediate reconstruction is fed into an off-the-shelf denoiser to obtain the final reconstructed image.,16749
the no of function that biological genome DNA. rubbish fraction is The called has,The fraction of the genome that has no biological function is called rubbish DNA.,16750
alleviate decision planning Use can of challenges help complex intelligent aids of the operations.,Use of intelligent decision aids can help alleviate the challenges of planning complex operations.,16751
approach. a Dispersal is via modeled Lagrangian,Dispersal is modeled via a Lagrangian approach.,16752
are testing policies. both Our and methods in the study surveillance variations limitations country-level,Our study limitations are the country-level variations in both surveillance methods and testing policies.,16753
proportion B wide study Influenza over the period. variations displayed,Influenza B proportion displayed wide variations over the study period.,16754
examine whether other could additional factors. are studies there Future,Future studies could examine whether there are other additional factors.,16755
health potential in implications prioritizing public This measures. control study has,This study has potential implications in prioritizing public health control measures.,16756
"atrium. amounts contrast of small the are (CA) to used locate agent left only Commonly,","Commonly, only small amounts of contrast agent (CA) are used to locate the left atrium.",16757
images. onto are coordinates and world projected Building elevated in footprints,Building footprints are elevated in world coordinates and projected onto images.,16758
"projections procedure, critical. In accurate this are camera","In this procedure, accurate camera projections are critical.",16759
"affect sensors inherited which adversely position commonly camera results. exist, external However, errors from","However, camera position errors inherited from external sensors commonly exist, which adversely affect results.",16760
our real-world of the Experiments datasets method. show promise on,Experiments on real-world datasets show the promise of our method.,16761
a removal problem. Single challenging is ill-posed haze image,Single image haze removal is a challenging ill-posed problem.,16762
various methods get to constraints/priors dehazing plausible use solutions. Existing,Existing methods use various constraints/priors to get plausible dehazing solutions.,16763
"paper, called trainable DehazeNet, medium propose this transmission a system for end-to-end In we estimation.","In this paper, we propose a trainable end-to-end system called DehazeNet, for medium transmission estimation.",16764
between and in proposed of used methods. We components establish existing those connections the DehazeNet,We establish connections between components of the proposed DehazeNet and those used in existing methods.,16765
functions. models as heuristic are deployed learned new These,These learned models are deployed as new heuristic functions.,16766
problems. larger effective Consequently is more in it,Consequently it is more effective in larger problems.,16767
analysis statistical annotations. the accuracy of our a provide of We,We provide a statistical analysis of the accuracy of our annotations.,16768
rotation motion invariance insights network offer Our to design and nonspecific contributions estimation. on,Our contributions on network design and rotation invariance offer insights nonspecific to motion estimation.,16769
"Euclidean only methods data existing apply SRC to vector type in space. However, the","However, the existing SRC type methods apply only to vector data in Euclidean space.",16770
number function arbitrary oriented meshes. for measures generalized triangle The winding insideness,The generalized winding number function measures insideness for arbitrary oriented triangle meshes.,16771
"to Exploiting act operations meshes. binary such boolean I on this, generalize similarly","Exploiting this, I similarly generalize binary boolean operations to act on such meshes.",16772
"union, for etc. difference, The intersection, operations resulting","The resulting operations for union, intersection, difference, etc.",16773
key-point present convolutional networks We deep neural algorithm for descriptors (CNN). using extracting an,We present an algorithm for extracting key-point descriptors using deep convolutional neural networks (CNN).,16774
regression based face We a also present alignment using these on algorithm local descriptors.,We also present a face alignment algorithm based on regression using these local descriptors.,16775
input feature robust is operation This a to from seek images.,This operation is to seek a robust feature from input images.,16776
nonlinear included to approaches with their methods have Traditional respective modeling linear drawbacks. dynamical and,Traditional approaches to dynamical modeling have included linear and nonlinear methods with their respective drawbacks.,16777
human of brain networks learning. supports flexibility reconfiguration cognitive and Flexible,Flexible reconfiguration of human brain networks supports cognitive flexibility and learning.,16778
how perform Targeted measure can artists specific often employed studies user to tasks. well are,Targeted user studies are often employed to measure how well artists can perform specific tasks.,16779
"paper, we this models. workflows detailed In sculpting digital to produce investigate used","In this paper, we investigate digital sculpting workflows used to produce detailed models.",16780
"what to experiment In model. our artists freely can how choose and design,","In our experiment design, artists can choose freely what and how to model.",16781
recover trends We trends with goodness-of-fits validate and analyzes measures. sophisticated whole-workflow these statistical with,We recover whole-workflow trends with sophisticated statistical analyzes and validate these trends with goodness-of-fits measures.,16782
"rather not artists in work do coarse-to-fine bursts. Temporally, but","Temporally, artists do not work coarse-to-fine but rather in bursts.",16783
"artists periodicity. times area Spatio-temporally, return any the work apparent without on multiple to same","Spatio-temporally, artists return to work on the same area multiple times without any apparent periodicity.",16784
are regarding recognition discussed. Some object problems DCNNs open for then issues,Some open issues regarding DCNNs for object recognition problems are then discussed.,16785
what representing also feature the comprehensive really is the original analyze video. We for,We also analyze what the comprehensive feature really is for representing the original video.,16786
"generated from obtained by Subsequently, low-order comprehensive tensor decomposition. is tensor feature the the","Subsequently, the comprehensive feature is generated by the low-order tensor obtained from tensor decomposition.",16787
"feature. is computed video this the using Finally, fingerprint","Finally, the video fingerprint is computed using this feature.",16788
tests data. empirical conducted We CAT simulated on,We conducted simulated CAT tests on empirical data.,16789
and compared. model presented each for of are Results these tests separately,Results of these tests are presented for each model separately and compared.,16790
harmony also with economic in is study analysis historical of the the growth. This,This analysis is also in harmony with the study of the historical economic growth.,16791
"Malthusian of no range data, Within there analysable the was stagnation.","Within the range of analysable data, there was no Malthusian stagnation.",16792
"postulated never to from the growth, stagnation Theory Unified Growth by happened. Takeoffs","Takeoffs from stagnation to growth, postulated by the Unified Growth Theory never happened.",16793
were from there There Malthusian escapes no trap. no was the trap because,There were no escapes from the Malthusian trap because there was no trap.,16794
"takeoffs this claimed in timing happened. the Consequently, differential in theory never","Consequently, the claimed in this theory differential timing in takeoffs never happened.",16795
utility majority of the our method. in performance extensive of confirm the evaluations State-of-the-art these,State-of-the-art performance in the majority of these extensive evaluations confirm the utility of our method.,16796
"for these model purposes. or applied extensively been analyzed Continuous-variable have networks and ""rate""","Continuous-variable or ""rate"" model networks have been analyzed and applied extensively for these purposes.",16797
"connected advances, Despite remains training networks recurrently challenge. neural significant a spiking","Despite significant advances, training recurrently connected spiking neural networks remains a challenge.",16798
and of systems. are Capabilities visual significant of prediction inference components,Capabilities of inference and prediction are significant components of visual systems.,16799
"challenging this we and paper, task them: address prediction. path In an visual of important","In this paper, we address an important and challenging task of them: visual path prediction.",16800
"propose we estimation. algorithm on gradient based this paper, a novel In NGCA log-density","In this paper, we propose a novel NGCA algorithm based on log-density gradient estimation.",16801
different state-of-the-art tasks.  visual  has many network Convolutional performance neural in achieved (CNN),  Convolutional neural network (CNN) has achieved state-of-the-art performance in many different visual tasks.,16802
"among also domains. are CNN transferable different Moreover, features","Moreover, CNN features are also transferable among different domains.",16803
of a a network encode space? architecture can a stable with How dynamic map,How can a network with a dynamic architecture encode a stable map of space?,16804
parallel for model. We a new propose computationally sampler efficient the Gibbs,We propose a computationally efficient parallel Gibbs sampler for the new model.,16805
that from the conjectured retrieves regularities has brain stimuli. It statistical been repeatedly,It has been repeatedly conjectured that the brain retrieves statistical regularities from stimuli.,16806
approach allowing we statistical conjecture. address present new to this a Here,Here we present a new statistical approach allowing to address this conjecture.,16807
"PPA properties. number Moreover, a analytical interesting shows of","Moreover, PPA shows a number of interesting analytical properties.",16808
"which First, a is turn guarantees inverse. the map, of volume-preserving in PPA existence the","First, PPA is a volume-preserving map, which in turn guarantees the existence of the inverse.",16809
"be can in Second, such obtained an form. inverse closed","Second, such an inverse can be obtained in closed form.",16810
"dimensionality performance to evaluate sensible of units. in Moreover, it reduction (input-domain) allows the","Moreover, it allows to evaluate the performance of dimensionality reduction in sensible (input-domain) units.",16811
These demonstrated and illustrated experimentally. properties are theoretically,These properties are demonstrated theoretically and illustrated experimentally.,16812
A machine learning. mechanical realizes is that learning system a,A learning machine is a system that realizes mechanical learning.,16813
Vertebra produced into where segmentation leaked the errors the ribs. atlases,Vertebra atlases produced errors where the segmentation leaked into the ribs.,16814
"not placebo is understood these the completely can shape mechanisms it response. how However,","However, it is not completely understood how these mechanisms can shape the placebo response.",16815
challenging the signal involve task (PDF) estimation. function probability density multidimensional of processing problems Most,Most signal processing problems involve the challenging task of multidimensional probability density function (PDF) estimation.,16816
"convenience. unit for mean covariance at In zero a Gaussian particular, aim we","In particular, we aim at a zero mean unit covariance Gaussian for convenience.",16817
similar is formally algorithms. iterative (PP) RBIG classical Projection to Pursuit,RBIG is formally similar to classical iterative Projection Pursuit (PP) algorithms.,16818
"and analyzed. invertibility are RBIG and experimentally theoretically of The differentiability, convergence","The differentiability, invertibility and convergence of RBIG are theoretically and experimentally analyzed.",16819
feature of selection. This some form requires,This requires some form of feature selection.,16820
task source the becomes the we as and target between Improvements task. distance increase significant,Improvements becomes significant as we increase the distance between source task and the target task.,16821
The the gradients back-propagation efficiently using algorithm. can be pixel-wise computed,The pixel-wise gradients can be efficiently computed using the back-propagation algorithm.,16822
"saliency Our faster other deep based methods quite are much efficient, learning than methods. computationally","Our methods are quite computationally efficient, much faster than other deep learning based saliency methods.",16823
of assist detection (CADe) in fractures. and spine detection classification the Computer-aided could,Computer-aided detection (CADe) could assist in the detection and classification of spine fractures.,16824
"maps of posterior Then, elements are the edge computed.","Then, edge maps of the posterior elements are computed.",16825
split data cases. training We the testing and into randomly,We randomly split the data into training and testing cases.,16826
"dictionary, For low-rank enforce i.e. a constraint, shared we the","For the shared dictionary, we enforce a low-rank constraint, i.e.",16827
also its and be could to FDDL said extensions. applied algorithm The,The said algorithm could also be applied to FDDL and its extensions.,16828
"reduces is damped with For overall, synergistic the selection when oscillations. epistasis, probability fixation varied,","For synergistic epistasis, when selection is varied, the fixation probability reduces overall, with damped oscillations.",16829
developing In treated the an microbial pathogen. often identified world NS is without,In the developing world NS is often treated without an identified microbial pathogen.,16830
"approaches sensitive dense classic to However, alignment are illumination change.","However, classic dense alignment approaches are sensitive to illumination change.",16831
"k-means} algorithm chosen at random. the Traditionally, initialize with {\tt uniformly practitioners centers","Traditionally, practitioners initialize the {\tt k-means} algorithm with centers chosen uniformly at random.",16832
provides semi-supervised This analysis algorithm. clustering theoretical for linear roughly a justification,This analysis provides theoretical justification for a roughly linear semi-supervised clustering algorithm.,16833
Machines spatial learning framework sequential a Pose provide prediction for rich implicit models.,Pose Machines provide a sequential prediction framework for learning rich implicit spatial models.,16834
"In handling processing, the implies of patches. this image small image context","In the context of image processing, this implies handling small image patches.",16835
performance. The is new reconstruction Laplacian reminiscent the conventional to prior with but graph stronger,The new prior is reminiscent to the conventional graph Laplacian but with stronger reconstruction performance.,16836
with on experiments results inpainting. are The image validated,The results are validated with experiments on image inpainting.,16837
solve in the problem We steps. two,We solve the problem in two steps.,16838
in as extraction and component NER We pipeline model module. used cTAKES integrated Stanford relation,We integrated Stanford NER model as cTAKES pipeline component and used in relation extraction module.,16839
cut in is vessel by material the of method. the found boundary The the graph,The boundary of the material in the vessel is found by the graph cut method.,16840
material the of This the cut the to in image. boundary corresponds,This cut corresponds to the boundary of the material in the image.,16841
of linear We manifolds: subspaces. employ definite positive Riemannian symmetric and two categories matrices,We employ two categories of Riemannian manifolds: symmetric positive definite matrices and linear subspaces.,16842
under ideal FV highest obtains The accuracy conditions. the approach,The FV approach obtains the highest accuracy under ideal conditions.,16843
"translation and best FV scale moderate deals with changes. Moreover,","Moreover, FV best deals with moderate scale and translation changes.",16844
solve is A problem. invested methods developing great amount into of effort being to this,A great amount of effort is being invested into developing methods to solve this problem.,16845
"arise, the in are limits Several Which thus important alignment questions performance? including: fundamental multi-image","Several important questions thus arise, including: Which are the fundamental limits in multi-image alignment performance?",16846
more alignment? access Does images improve having the to,Does having access to more images improve the alignment?,16847
these is and target pharmaceutical by-products. One product unwanted two of are a substances,One of these substances is a target pharmaceutical product and two are unwanted by-products.,16848
comparing a a and process evaluating set arguments. is of Argumentation of,Argumentation is a process of evaluating and comparing a set of arguments.,16849
"which is takes broader perspective. study a no comparative However, there","However, there is no comparative study which takes a broader perspective.",16850
in we This what work. propose this is,This is what we propose in this work.,16851
proposed comparison general these semantics of provide a We with to properties. all the respect,We provide a general comparison of all these semantics with respect to the proposed properties.,16852
existing underline differences to behavior semantics. the allows between the of That,That allows to underline the differences of behavior between the existing semantics.,16853
unknown from actions and views problem. (novel) a unseen human Recognizing is challenging,Recognizing human actions from unknown and unseen (novel) views is a challenging problem.,16854
"scale incorporate Thus, action to efficiently classes. R-NKTM can new","Thus, R-NKTM can efficiently scale to incorporate new action classes.",16855
action that show benchmark Experiments existing our datasets state-of-the-art. cross-view on human method outperforms three,Experiments on three benchmark cross-view human action datasets show that our method outperforms existing state-of-the-art.,16856
method dubbed (EP). is Ensemble Projection The,The method is dubbed Ensemble Projection (EP).,16857
"the images. among individual also but of only not images, EP relationships captures characteristics the","EP captures not only the characteristics of individual images, but also the relationships among images.",16858
"yet computationally efficient, effective simple and and flexible. conceptually It is","It is conceptually simple and computationally efficient, yet effective and flexible.",16859
available of on code page. method the the is project The,The code of the method is available on the project page.,16860
Zou this et by paper build an In on al. approach we proposed,In this paper we build on an approach proposed by Zou et al.,16861
"computation, To speed et up Zou al.","To speed up computation, Zou et al.",16862
physical during changes activity. apply heart to method our detect We in rate,We apply our method to detect changes in heart rate during physical activity.,16863
"today analyzing the use for the tools data. experts visual Still,","Still, today the experts use visual tools for analyzing the data.",16864
converted representation the words Then to bag for are of clustering. SAX hierarchical symbols,Then the SAX symbols are converted to bag of words representation for hierarchical clustering.,16865
"chiller. proposed real the data life of technique to applied Moreover, is adsorption","Moreover, the proposed technique is applied to real life data of adsorption chiller.",16866
This paper the HA approach proposes modeling to a probabilistic algorithms. design of,This paper proposes a probabilistic modeling approach to the design of HA algorithms.,16867
are edges and common requiring cluttered problems segmentation Overlapping or regularization. weak colors additional,Overlapping colors and cluttered or weak edges are common segmentation problems requiring additional regularization.,16868
moves that submodular multi-shape show our constraints. We are a-expansion for,We show that a-expansion moves are submodular for our multi-shape constraints.,16869
"has its by ""hedgehog"" e.g. shape constrained Each field, normals vector surface some","Each ""hedgehog"" shape has its surface normals constrained by some vector field, e.g.",16870
of user transform for a scribbles. distance gradients,gradients of a distance transform for user scribbles.,16871
more allowed Wider normals cones gives relaxed hedgehog of shapes.,Wider cones of allowed normals gives more relaxed hedgehog shapes.,16872
clicks our defines approach all multi-star hedgehogs from come prior. single then If,If all hedgehogs come from single clicks then our approach defines multi-star prior.,16873
significantly Our has segmentation. more method than standard general applications one-star,Our general method has significantly more applications than standard one-star segmentation.,16874
"encoded of sequence depth Thus, a a reliably sequence into segment labels. can be","Thus, a depth sequence can be reliably encoded into a sequence of segment labels.",16875
various tasks. advances vision computer performances learning of challenging in have deep achieved remarkable Recent,Recent advances of deep learning have achieved remarkable performances in various challenging computer vision tasks.,16876
localization framework object Thus a for weakly is supervised introduced.,Thus a weakly supervised framework for object localization is introduced.,16877
simultaneously. optimizes classification and localization It jointly both networks,It jointly optimizes both classification and localization networks simultaneously.,16878
"best is how Therefore, understand profit the to important obtain to from existing it ConvNets.","Therefore, it is important to understand how to obtain the best profit from existing ConvNets.",16879
experiments popular ConvNets perform sensing datasets. three remote with We six using,We perform experiments with six popular ConvNets using three remote sensing datasets.,16880
and existing compare strategy descriptors each state-of-the-art with with also ConvNets in We baselines.,We also compare ConvNets in each strategy with existing descriptors and with state-of-the-art baselines.,16881
be strategy. that tuning best performing tends the to point Results fine,Results point that fine tuning tends to be the best performing strategy.,16882
three achieved state-of-the-art results also We used. the for datasets,We also achieved state-of-the-art results for the three datasets used.,16883
"process by reconstructing a involves followed from face classification. The primarily its image, occluded","The process primarily involves reconstructing a face from its occluded image, followed by classification.",16884
for the regions that methods the independent. assume pixel Typical occluded errors are reconstruction of,Typical methods for reconstruction assume that the pixel errors of the occluded regions are independent.,16885
and We hierarchical multi-action that a joint segmentation. approach performs classification propose recognition to,We propose a hierarchical approach to multi-action recognition that performs joint classification and segmentation.,16886
localization errors. and both Our results background-versus-foreground characterize,Our results characterize both localization and background-versus-foreground errors.,16887
"old says is picture adage a thousand well-known A {\em worth words!""} that ""A","A well-known old adage says that {\em ""A picture is worth a thousand words!""}",16888
"mean precisely, information images? in by more we do what But","But more precisely, what do we mean by information in images?",16889
it be machines? And can how by retrieved effectively,And how can it be retrieved effectively by machines?,16890
highlight briefly puzzling in We this column. these questions,We briefly highlight these puzzling questions in this column.,16891
to (CNNs) image be Neural require on Networks Convolutional trained corpora large tasks. classification,Convolutional Neural Networks (CNNs) require large image corpora to be trained on classification tasks.,16892
such to representations. encouraging CNN deal focused with variations Previous scale work attempting on scale-invariant,Previous work attempting to deal with such scale variations focused on encouraging scale-invariant CNN representations.,16893
"of representations contain images scale-variant images, incomplete as are However, because well. representations information scale-invariant","However, scale-invariant representations are incomplete representations of images, because images contain scale-variant information as well.",16894
This and representations. combined of development scale-variant scale-invariant the addresses paper,This paper addresses the combined development of scale-invariant and scale-variant representations.,16895
CNN single-scale show our that CNN. results The outperforms multi-scale,The results show that our multi-scale CNN outperforms single-scale CNN.,16896
Data tasks rare. more imbalance vision is classes one are many where common in or,Data imbalance is common in many vision tasks where one or more classes are rare.,16897
with small achieving for samples. latter prediction capability limited on dataset The is crucial accurate,The latter capability is crucial for achieving accurate prediction on small dataset with limited samples.,16898
the conducted user study A the also effectiveness of is metric. to validate,A user study is also conducted to validate the effectiveness of the metric.,16899
volumetric approximate priors. homogeneity used is CNN inference spatial to Alpha-expansion the to perform imposing,Alpha-expansion is used to perform approximate inference imposing spatial volumetric homogeneity to the CNN priors.,16900
"the avoid Non-deterministic Timed using we Concurrent Probabilistic extended Constraint calculus. this, To Ccfomi","To avoid this, we extended Ccfomi using the Probabilistic Non-deterministic Timed Concurrent Constraint calculus.",16901
"not interpreter execute ntcc of was capable a RT to However, there Ccfomi.","However, there was not a ntcc interpreter capable of RT to execute Ccfomi.",16902
interpreter We RT ntcc-- developed Ntccrt. Ccfomi capable on and executed Ntccrt for --a we,We developed Ntccrt --a RT capable interpreter for ntcc-- and we executed Ccfomi on Ntccrt.,16903
"Ntccrt to our we Ccfomi. plan to the to extend extension execute future, In","In the future, we plan to extend Ntccrt to execute our extension to Ccfomi.",16904
wayfinding model new for dynamics A proposed. is pedestrian,A new wayfinding model for pedestrian dynamics is proposed.,16905
representation every an and inaccuracies individual knowledge defines for implying uncertainties. The pedestrian model,The model defines for every pedestrian an individual knowledge representation implying inaccuracies and uncertainties.,16906
"introduced. In knowledge-driven strategies are search addition,","In addition, knowledge-driven search strategies are introduced.",16907
example is presented scenario. concept a fictive on The tested,The presented concept is tested on a fictive example scenario.,16908
"committed the this is this In class. case, to particular object","In this case, the object is committed to this particular class.",16909
"can uncontaminated tracker drifting alleviate this method, of to By utilize issues. virtue our information","By virtue of this method, our tracker can utilize uncontaminated information to alleviate drifting issues.",16910
non-active us allows Recently features/samples. part a of identify screening safe introduced to,Recently introduced safe screening allows us to identify a part of non-active features/samples.,16911
"consuming time however, segmentation, is and tedious. very Manual","Manual segmentation, however, is very time consuming and tedious.",16912
"WikiNav pairs Jeopardy! extend with the question-answer we Furthermore, from","Furthermore, we extend the WikiNav with question-answer pairs from Jeopardy!",16913
challenging model is a selection Performing between random Gibbs very fields task.,Performing model selection between Gibbs random fields is a very challenging task.,16914
an forms that best fits This a model issue pick to data. observed the central,This forms a central issue to pick the model that best fits an observed data.,16915
the version a new We of Criterion. Bayesian approximate introduce Information,We introduce a new approximate version of the Bayesian Information Criterion.,16916
the performances BLIC those for questions. of study We,We study the performances of BLIC for those questions.,16917
"this its palmprint representation. approach present for novel recognition and we paper, a In","In this paper, we present a novel approach for palmprint recognition and its representation.",16918
"technique binarization the algorithm is extract local lines, thresholding Niblack To adopted. palm","To extract the palm lines, local thresholding technique Niblack binarization algorithm is adopted.",16919
indirect also pollinator. by coexistence plant promoted facilitation We show the via how is,We also show how plant coexistence is promoted by indirect facilitation via the pollinator.,16920
has two consequences plant that conflicting coexistence. Simulations for show this,Simulations show that this has two conflicting consequences for plant coexistence.,16921
"optimal stable mutualism, evolutionarily strategy, Keywords: adaptation foraging, competition, . rate coexistence,","Keywords: mutualism, competition, optimal foraging, evolutionarily stable strategy, coexistence, adaptation rate .",16922
function. cataloguing the largest project Gene is resource The Ontology gene (GO) for,The Gene Ontology (GO) project is the largest resource for cataloguing gene function.,16923
"a primer we provide all users chapter, for GO. of concise this In the","In this chapter, we provide a concise primer for all users of the GO.",16924
"images. vectorizing we describe a clipart paper, In this for novel method","In this paper, we describe a novel method for vectorizing clipart images.",16925
potential This images on focuses sensing remote of registration. paper accuracy,This paper focuses on potential accuracy of remote sensing images registration.,16926
and properties. by texture defined CRLB local image This noise is,This CRLB is defined by local image texture and noise properties.,16927
"of and we common chapter, In annotations. ontology discuss misinterpretations the some the this","In this chapter, we discuss some common misinterpretations of the ontology and the annotations.",16928
analyses as such can biases gene We aggregate several confound discuss that analyses. enrichment also,We also discuss several biases that can confound aggregate analyses such as gene enrichment analyses.,16929
"best of suggest biases, pitfalls and For each we these practices. remedies and","For each of these pitfalls and biases, we suggest remedies and best practices.",16930
"In molecule the n the mRNA as sites. chain modeled of is RFM,","In the RFM, the mRNA molecule is modeled as chain of n sites.",16931
"is that case, first the we the show controllable, In i.e. system","In the first case, we show that the system is controllable, i.e.",16932
of discuss some We biological results. these of implications the,We discuss some of the biological implications of these results.,16933
solutions genetic to algorithm this a problem. to optimization We used have find,We have used a genetic algorithm to find solutions to this optimization problem.,16934
thoroughly. theoretical Both empirical and are methods of properties our studied,Both theoretical and empirical properties of our methods are studied thoroughly.,16935
extend two PS ways. framework the in We,We extend the PS framework in two ways.,16936
"color and and build using part First, we detectors robust images. depth both discriminative","First, we build robust and discriminative part detectors using both color and depth images.",16937
"(HDD). novel We present descriptor histogram images, depth of also differences depth a for called","We also present a novel descriptor for depth images, called histogram of depth differences (HDD).",16938
therefore process and is adaptive is any The problem formulation. expandable modular to whole,The whole process is modular expandable and is therefore adaptive to any problem formulation.,16939
for MoE be universal approximation models. can as viewed result theorem a Our,Our result can be viewed as a universal approximation theorem for MoE models.,16940
a long Repeated behavioral and in have sciences tradition the evolutionary biology. games,Repeated games have a long tradition in the behavioral sciences and evolutionary biology.,16941
"show exist strategies Furthermore, with randomly-alternating games moves. that for autocratic we even (iii)","Furthermore, we show that autocratic strategies exist even for (iii) games with randomly-alternating moves.",16942
"statistical test, We i.i.d. both derive for a","We derive a statistical test, both for i.i.d.",16943
"where procedure. a the wild estimate using we samples, quantiles null bootstrap distribution","samples, where we estimate the null distribution quantiles using a wild bootstrap procedure.",16944
the synthetic real on of datasets. the effectiveness approach illustrate both We and,We illustrate the effectiveness of the approach on both synthetic and real datasets.,16945
an function. internal which often utility maximize make decisions Humans,Humans often make decisions which maximize an internal utility function.,16946
"tend on However, change to how humans ""feel"". depending betting strategies their they","However, humans tend to change their betting strategies depending on how they ""feel"".",16947
"are decisions is, driven state. by That their emotional their","That is, their decisions are driven by their emotional state.",16948
"observations subspace paper, In Bayes from variational learning this is online an algorithm partial presented.","In this paper, an online variational Bayes subspace learning algorithm from partial observations is presented.",16949
"CNNs enable image highly training sufficient features layered learning data-driven, hierarchical data. from representative,","CNNs enable learning data-driven, highly representative, layered hierarchical image features from sufficient training data.",16950
and We first different CNN architectures. explore evaluate,We first explore and evaluate different CNN architectures.,16951
dataset of We spatial scale on context image then performance. evaluate influence the and,We then evaluate the influence of dataset scale and spatial image context on performance.,16952
under scheduling The problem constraints applicable. of resource widely is,The problem of scheduling under resource constraints is widely applicable.,16953
Such tightly temporal and problems constraints. continuous feature constraints coupled continuous resource,Such problems feature tightly coupled continuous resource constraints and continuous temporal constraints.,16954
"these each the proposed We describe applications, of involved solutions. and challenges","We describe each of these applications, the challenges involved and proposed solutions.",16955
"frames. video multiple one of each by fusion image reconstructed is For view, super-resolved","For each view, one super-resolved image is reconstructed by fusion of multiple video frames.",16956
evaluated a our video camera. mobile framework low-cost and fundus We for,We evaluated our framework for a mobile and low-cost video fundus camera.,16957
novel presented. is distributions triangular parametrization of A,A novel parametrization of triangular distributions is presented.,16958
"evaluations, shown be increase and to monotonically likelihood algorithm the both The is convergent. globally","The algorithm is shown to both monotonically increase the likelihood evaluations, and be globally convergent.",16959
general framework spatio-spectral A multi-spectral paper. introduced segmentation in images of for is this,A general framework of spatio-spectral segmentation for multi-spectral images is introduced in this paper.,16960
several steps. is decomposed into The present approach sequential,The present approach is decomposed into several sequential steps.,16961
"using dimensionality-reduction is analysis First, performed method. stage the a factor-correspondence","First, a dimensionality-reduction stage is performed using the factor-correspondence analysis method.",16962
"a of spectral image. a classification pre-segmentation spectral produces the Then,","Then, a spectral classification produces a spectral pre-segmentation of the image.",16963
is to A approach PC determine used non-intrusive coefficients. the,A non-intrusive approach is used to determine the PC coefficients.,16964
method. an be as iterative Newton It also viewed stochastic can,It can also be viewed as an iterative stochastic Newton method.,16965
samples. the in rMAP reduce to is presented approximate approach bias Metropolization An,An approximate Metropolization approach is presented to reduce the bias in rMAP samples.,16966
generation presented. multi-scale proposal is approach A greedy-based object,A multi-scale greedy-based object proposal generation approach is presented.,16967
"learning. processing, Sparse and representations applied signal to successfully vision been computer have machine","Sparse representations have been successfully applied to signal processing, computer vision and machine learning.",16968
"require when methods However, covariance region with computation. complex combined such often","However, such methods when combined with region covariance often require complex computation.",16969
Experiments on with the several demonstrate vision competitive tasks performance state-of-the-art methods.,Experiments on several vision tasks demonstrate competitive performance with the state-of-the-art methods.,16970
problem. this nonparametric a We class Bayesian address introduce to here models of,We introduce here a class of Bayesian nonparametric models to address this problem.,16971
out Swendsen-Wang original carried is computation Bayesian using an algorithm. generalized,Bayesian computation is carried out using an original generalized Swendsen-Wang algorithm.,16972
tasks. our on algorithm We several classification validate,We validate our algorithm on several classification tasks.,16973
the such results demonstrate that proposed manifolds. The method on also state-of-the-art methods outperforms experiment,The experiment results also demonstrate that the proposed method outperforms state-of-the-art methods on such manifolds.,16974
"visual handcrafted. used mostly are However, features the","However, the visual features used are mostly handcrafted.",16975
results methods show that both yields EndoNet tasks. Extensive to for other state-of-the-art comparisons experimental,Extensive experimental comparisons to other methods show that EndoNet yields state-of-the-art results for both tasks.,16976
promising also most of simulations. compare We the by ones numerical means,We also compare the most promising ones by means of numerical simulations.,16977
a computer face wild attributes Predicting images challenging problem. in vision the from is,Predicting attributes from face images in the wild is a challenging computer vision problem.,16978
"has using a problem, been face typical deep classification attribute As addressed learning. prediction","As a typical classification problem, face attribute prediction has been addressed using deep learning.",16979
"decades systems automatic study, accurate After detection widespread. now of recognition and are face and","After decades of study, automatic face detection and recognition systems are now accurate and widespread.",16980
we in Where race? this cat-and-mouse do stand,Where do we stand in this cat-and-mouse race?,16981
in everyone camera a live We society currently a in carries pocket. their where,We currently live in a society where everyone carries a camera in their pocket.,16982
multifaceted numerous spatial is unknown a variables. process regeneration Forest with,Forest regeneration is a spatial multifaceted process with numerous unknown variables.,16983
simulating diffusions. a residual-bridge new introduce We for conditioned approximately proposal,We introduce a new residual-bridge proposal for approximately simulating conditioned diffusions.,16984
in These illustrated study. gains simulation via efficiencies potential a are,These potential gains in efficiencies are illustrated via a simulation study.,16985
common The regularizer reduction. Laplacian in learning graph well-established and is dimensionality spectral semi-supervised,The common graph Laplacian regularizer is well-established in semi-supervised learning and spectral dimensionality reduction.,16986
"it high-dimensional manifolds. functions a However, first-order can degenerate to as lead in regularizer,","However, as a first-order regularizer, it can lead to degenerate functions in high-dimensional manifolds.",16987
by accompanied is the ATP of subunit rotation enzyme. Synthesis within,Synthesis of ATP is accompanied by subunit rotation within the enzyme.,16988
"can paper, investigate variability those this of be expressions how related. distinct we In","In this paper, we investigate how those distinct expressions of variability can be related.",16989
proposed to this methods make were reduction analysis Data easier.,Data reduction methods were proposed to make this analysis easier.,16990
of the the and samples outcome. covariates,samples of the covariates and the outcome.,16991
samples covariates the the of and outcome.,samples of the covariates and the outcome.,16992
problem. challenging an statistically Discovering databases significant from important is patterns,Discovering statistically significant patterns from databases is an important challenging problem.,16993
selective recently built is statistical on popularized called inference Our a approach inference. framework,Our approach is built on a recently popularized statistical inference framework called selective inference.,16994
We significant our patterns approach useful that demonstrate statistically is from finding databases. for,We demonstrate that our approach is useful for finding statistically significant patterns from databases.,16995
"to efficiency. of problems, as Manual variation due practical fission identification tracks such observer-observation has","Manual identification of fission tracks has practical problems, such as variation due to observer-observation efficiency.",16996
"is difficult processing. However, most in tasks of separation the non-trivial one of images image","However, separation of non-trivial images is one of the most difficult tasks in image processing.",16997
proposed available download. the to are Algorithms for corresponding method,Algorithms corresponding to the proposed method are available for download.,16998
that capable task. binary reveal all features not this evaluations Our are of,Our evaluations reveal that not all binary features are capable of this task.,16999
practice. evaluation in results our the of of feasibility preliminary suggest performance approach The our,The results of our preliminary performance evaluation suggest the feasibility of our approach in practice.,17000
a for network galactose galactose The is system responsible complex metabolism.,The galactose network is a complex system responsible for galactose metabolism.,17001
of the in demonstrated two is algorithm Application examples.,Application of the algorithm is demonstrated in two examples.,17002
of R these the efficient implementation package be An found in can minimaxdesign. algorithms,An efficient implementation of these algorithms can be found in the R package minimaxdesign.,17003
marker-less importance human-computer interaction. hand tracking in Real-time is increasing of,Real-time marker-less hand tracking is of increasing importance in human-computer interaction.,17004
publicly We quantitatively show these and available on qualitatively improvements both datasets.,We show these improvements both qualitatively and quantitatively on publicly available datasets.,17005
safe is inspired development SPP recent feature by screening. of method The,The SPP method is inspired by recent development of safe feature screening.,17006
Markerless human-computer hands interaction. enabler fingers of tracking for a promising is and,Markerless tracking of hands and fingers is a promising enabler for human-computer interaction.,17007
"decision into pixels the of the a parts classifies forest In detection hand. randomized step,","In the detection step, a randomized decision forest classifies pixels into parts of the hand.",17008
"or approach also arrangements. varying camera-to-scene static, The moving, supports","The approach also supports varying static, or moving, camera-to-scene arrangements.",17009
"locations. the characterness refines and Specifically, corresponding simultaneously network the predicts scores","Specifically, the network simultaneously predicts characterness scores and refines the corresponding locations.",17010
"definite to are even sensitive bounded using They kernels. when contaminated data, positive","They are sensitive to contaminated data, even when using bounded positive definite kernels.",17011
"propose we correlation kernel CCA). function canonical (classical kernel influence Second, analysis of classical","Second, we propose influence function of classical kernel canonical correlation analysis (classical kernel CCA).",17012
state-of-the-art over The the superior methods. methods the performance show robust,The robust methods show the superior performance over the state-of-the-art methods.,17013
is interplay mechanisms and addressed. efficiency evolutionary The between energy,The interplay between energy efficiency and evolutionary mechanisms is addressed.,17014
clear. gathering of an individuals is energetic The in advantage not efficient context immediate,The immediate advantage of gathering efficient individuals in an energetic context is not clear.,17015
"relatedness competition. increases this it also Although local among process increases individuals,","Although this process increases relatedness among individuals, it also increases local competition.",17016
structured well-mixed the dynamics both We in explore and populations.,We explore the dynamics in both well-mixed and structured populations.,17017
metabolic improve cells. of Periodic performance can enzyme activities the,Periodic enzyme activities can improve the metabolic performance of cells.,17018
behaviour different occur regions of types Both of same model. the may parameter,Both types of behaviour may occur different parameter regions of the same model.,17019
"boards other solved. several unsolved addition, In were hitherto","In addition, several other boards hitherto unsolved were solved.",17020
method aperture letter named radar (FAO). (SAR) registration Feature-Area Optimization This proposes synthetic image a,This letter proposes a synthetic aperture radar (SAR) image registration method named Feature-Area Optimization (FAO).,17021
"based the on are key these Then, three factors features. determined","Then, the three key factors are determined based on these features.",17022
"result. get factor-determined the can model optimization solving Finally, registration the","Finally, solving the factor-determined optimization model can get the registration result.",17023
"usually In photo vision, problem the computer geolocation approached retrieval image using is methods.","In computer vision, the photo geolocation problem is usually approached using image retrieval methods.",17024
methods Data-oriented super-large data. to extract been have from information statistical developed,Data-oriented statistical methods have been developed to extract information from super-large data.,17025
"algorithm, In this to computation introduce rounding scalable. parameters make we the","In this algorithm, we introduce rounding parameters to make the computation scalable.",17026
algorithms paper this function. we In reward the learning generalized reinforcement with proposed,In this paper we proposed reinforcement learning algorithms with the generalized reward function.,17027
"Real learning, Machine time Keywords Artificial : intelligence. strategy, learning, Reinforcement","Keywords : Reinforcement learning, Machine learning, Real time strategy, Artificial intelligence.",17028
neural presented. segmentation network a semantic is framework deconvolutional A with weakly-supervised tied,A weakly-supervised semantic segmentation framework with a tied deconvolutional neural network is presented.,17029
framework deconvolution and of operations. the consists deconvolution unpooling in layer Each,Each deconvolution layer in the framework consists of unpooling and deconvolution operations.,17030
inference positives pixel-level This stage. in in false reduction of a results,This results in reduction of false positives in a pixel-level inference stage.,17031
generating be used for class-specific stacked features are maps. to selectively Those activation,Those features are stacked to be selectively used for generating class-specific activation maps.,17032
multivariate proposed circular be shown Previously cases to this construction. special of distributions are,Previously proposed multivariate circular distributions are shown to be special cases of this construction.,17033
These tools leverage standard modelling can models (e.g.,These models can leverage standard modelling tools (e.g.,17034
and for functions covariance methods automatic determination). relevance,covariance functions and methods for automatic relevance determination).,17035
of of the each This process landmark extracts through algorithm at matching. template information momentum,This algorithm extracts information of momentum at each landmark through the process of template matching.,17036
representations use a Bayesian making collected We approach for on residual inference. momentum the,We use a Bayesian approach on the collected residual momentum representations for making inference.,17037
"are Eukaryota, databases and for viruses. Pre-trained Archaea, included Bacteria,","Pre-trained databases are included for Archaea, Bacteria, Eukaryota, and viruses.",17038
"computationally However, case. is worst solving general intractable the a in POMDP","However, solving a general POMDP is computationally intractable in the worst case.",17039
Bayesian algorithm We develop to learning POMDP-lite model-based reinforcement simple solve a models.,We develop a simple model-based Bayesian reinforcement learning algorithm to solve POMDP-lite models.,17040
conditions. is We suitable show further under the that near-Bayesian-optimal algorithm,We further show that the algorithm is near-Bayesian-optimal under suitable conditions.,17041
"just revealed polypeptides are are and single amino acids. considered, when constraints not Novel","Novel constraints are revealed when polypeptides are considered, and not just single amino acids.",17042
and formal several constraints. compute graph these to frameworks a characterize and We algorithm describe,We describe several formal frameworks and a graph algorithm to characterize and compute these constraints.,17043
rendering problem a is based in and fundamental computer graphics. vision Image,Image based rendering is a fundamental problem in computer vision and graphics.,17044
and demonstrate Our experimental efficacy accuracy. results the,Our experimental results demonstrate the efficacy and accuracy.,17045
"ideas As principles. few clear there overarching zoo a result, of is with a","As a result, there is a zoo of ideas with few clear overarching principles.",17046
surface with reconstruction and usually Point cloud for data contaminated is source outliers. noise,Point cloud source data for surface reconstruction is usually contaminated with noise and outliers.,17047
"based Then, is a scheme. to remove mean-shift thresholding outliers clustering technique used through","Then, mean-shift based clustering technique is used to remove outliers through a thresholding scheme.",17048
"robust approach, results The show and comparably, that efficient. is experimental this","The experimental results show that this approach, comparably, is robust and efficient.",17049
much optimism There around intelligence and (AI) both artificial today. pessimism is,There is both much optimism and pessimism around artificial intelligence (AI) today.,17050
within for rule a setting. answering query propose existential inconsistency-tolerant framework general We,We propose a general framework for inconsistency-tolerant query answering within existential rule setting.,17051
modifiers notions: inference and relies It two key on strategies.,It relies on two key notions: modifiers and inference strategies.,17052
inconsistency-tolerant a as inference composite is modifier an An plus seen semantics strategy.,An inconsistency-tolerant semantics is seen as a composite modifier plus an inference strategy.,17053
of We the view. point a compare from obtained semantics productivity,We compare the obtained semantics from a productivity point of view.,17054
the divide improve A employed is conquer strategy to and speed.,A divide and conquer strategy is employed to improve the speed.,17055
"divided into several on the planogram. image regions the input based Specifically, is","Specifically, the input image is divided into several regions based on the planogram.",17056
method. real Experimental results the efficacy the have proposed verified of data on,Experimental results on real data have verified the efficacy of the proposed method.,17057
Cross-modal quality the semantic across modalities. exploiting of improves coding correlations by different hash hashing,Cross-modal hashing improves the quality of hash coding by exploiting semantic correlations across different modalities.,17058
"correlated isoform to tailored first present quantification we the Here experiments. method RNA-seq DICEseq,","Here we present DICEseq, the first isoform quantification method tailored to correlated RNA-seq experiments.",17059
survey used for an pixel-level semantic over different segmentation. overview techniques This gives,This survey gives an overview over different techniques used for pixel-level semantic segmentation.,17060
of A is given. taxonomy algorithms segmentation,A taxonomy of segmentation algorithms is given.,17061
Residual use this Dipolar paper Coupling. Contact and Pseudo In we Shifts,In this paper we use Pseudo Contact Shifts and Residual Dipolar Coupling.,17062
in central We and distribution results. in single depth tendency reconstruct a discuss the the,We reconstruct a single central tendency in the distribution and discuss in depth the results.,17063
contamination characterized be mutual by learning models. can machine Many problems,Many machine learning problems can be characterized by mutual contamination models.,17064
"items have Indeed, no are examples there effect. duplicate where uncontrived","Indeed, there are uncontrived examples where duplicate items have no effect.",17065
to initialization. sensitive their are networks Filters in neural convolutional,Filters in convolutional neural networks are sensitive to their initialization.,17066
process affects and trial that is error a problem network. a designing the of This,This is a problem that affects the trial and error process of designing a network.,17067
"connectivity and pharmacogenetic Mendelian GWAS diseases, We phenotypes, pathways. observe extensive factors between and transcription","We observe extensive connectivity between transcription factors and Mendelian diseases, GWAS phenotypes, and pharmacogenetic pathways.",17068
"To of model, Motivation: quality protein assess a the i.e.","Motivation: To assess the quality of a protein model, i.e.",17069
compressive a the consider image using of reconstructing measurements from grid. multi-resolution an problem We,We consider the problem of reconstructing an image from compressive measurements using a multi-resolution grid.,17070
"Narrative craft, stories. is the intelligence respond to to and tell, affectively ability understand,","Narrative intelligence is the ability to craft, tell, understand, and respond affectively to stories.",17071
Simulating major arbitrary statistical distributions computing. program of a is research from probability samples,Simulating samples from arbitrary probability distributions is a major research program of statistical computing.,17072
framework. Many this in can analyzed existing methods be,Many existing methods can be analyzed in this framework.,17073
process Poisson a processes and for reviews Poisson Monte chapter methods. Carlo defines The model,The chapter reviews Poisson processes and defines a Poisson process model for Monte Carlo methods.,17074
of with a model Our model landscape dynamics. a approach stochastic spatially-explicit combines population,Our approach combines a stochastic landscape model with a spatially-explicit model of population dynamics.,17075
contrastive function. Finetuning Siamese loss using in performed a architecture is a,Finetuning is performed in a Siamese architecture using a contrastive loss function.,17076
"to state space. methods a were with models applied embedded HMM Previously, one-dimensional only","Previously, embedded HMM methods were only applied to models with a one-dimensional state space.",17077
data recognition We on video deep dimensional the consider task using learning. of emotion,We consider the task of dimensional emotion recognition on video data using deep learning.,17078
"easy: may making predicted influence However, forecasts is the such not features. factors many accurate","However, making such accurate forecasts is not easy: many factors may influence the predicted features.",17079
"real cases is assumption this in However, true. always not","However, in real cases this assumption is not always true.",17080
remaining cases. of this predicting work new running time for In the we methods present,In this work we present new methods for predicting the remaining time of running cases.,17081
two on real All methods these evaluated extensively case are studies.,All these methods are extensively evaluated on two real case studies.,17082
for also present Inception architectures both streamlined new non-residual networks. We residual several and,We also present several new streamlined architectures for both residual and non-residual Inception networks.,17083
"circuits types autocrine and Moreover, sensing quorum genetic of similar control cells. many","Moreover, similar types of genetic circuits control many autocrine and quorum sensing cells.",17084
We a few style power demonstrate on of the examples. space the suggested first,We first demonstrate the power of the suggested style space on a few examples.,17085
"analyses enrichment sense such GO be challenging. making of However, may","However, making sense of such GO enrichment analyses may be challenging.",17086
various of types objectively. displays thus compare We,We thus compare various types of displays objectively.,17087
proposed A conducted experiment perceptual the validate was to method.,A perceptual experiment was conducted to validate the proposed method.,17088
for velocities varying displays. nine test motion commercial We,We test varying motion velocities for nine commercial displays.,17089
method comparisons our for accurately characterizes various motion types. that display blur Our indicate,Our comparisons indicate that our method accurately characterizes motion blur for various display types.,17090
paper unconstrained in scenarios. problem identification the on This script of focuses,This paper focuses on the problem of script identification in unconstrained scenarios.,17091
is made proposed method of Source online. code available the,Source code of the proposed method is made available online.,17092
oral of very drug Complexity targets cancer designing difficult. renders,Complexity of oral cancer renders designing drug targets very difficult.,17093
to representation. contribution policy Our related second is,Our second contribution is related to policy representation.,17094
"multi-focus a handle environment. methods in To fusion situation, image are used such","To handle such a situation, image fusion methods are used in multi-focus environment.",17095
scene of problem script focuses images. text on in paper identification the This,This paper focuses on the problem of script identification in scene text images.,17096
with this identification state-of-the-art datasets. in results two public demonstrate procedure script Our learning experiments,Our experiments with this learning procedure demonstrate state-of-the-art results in two public script identification datasets.,17097
This centralising the rule. preferential by achieved connections through a attachment cross-modular is,This is achieved by centralising the cross-modular connections through a preferential attachment rule.,17098
other network dynamics benchmarks. model complex used widely models than as Our hierarchical hosts more,Our network model hosts more complex dynamics than other hierarchical models widely used as benchmarks.,17099
is with and others. shown The these two method be to methods new competitive,The new method is shown to be competitive with these two methods and others.,17100
This exceeds feature accuracies extraction pipelines. that earlier the of use manually engineered studies clearly,This clearly exceeds the accuracies of earlier studies that use manually engineered feature extraction pipelines.,17101
predominantly horizontal of (HGT). gene bacteria occurs transfer Adaptation via,Adaptation of bacteria occurs predominantly via horizontal gene transfer (HGT).,17102
"unit propose Thus, that form a transfer. of gene basic we horizontal SOCs","Thus, we propose that SOCs form a basic unit of horizontal gene transfer.",17103
pairwise of the threshold. trees constructs distance distances such FJ and on basis a,FJ constructs such trees on the basis of pairwise distances and a distance threshold.,17104
transmission sampled individuals a FJ-BIC sequences to involved known HIV from chain. applied was in,FJ-BIC was applied to HIV sequences sampled from individuals involved in a known transmission chain.,17105
all tree was transmission compatible almost The FJ-BIC be events. found to with,The FJ-BIC tree was found to be compatible with almost all transmission events.,17106
"In cases, results. promising both yields approach our","In both cases, our approach yields promising results.",17107
has Digitization image editing easier. of images made,Digitization of images has made image editing easier.,17108
general of The are economy pathogen the load of principles and minimization resources.,The general principles are minimization of the pathogen load and economy of resources.,17109
the describing constrain system. immune should parameters the They,They should constrain the parameters describing the immune system.,17110
a allows tissues. stability between comparison qualitative condition The,The stability condition allows a qualitative comparison between tissues.,17111
have improving reconstruction SIM at algorithm reported. been of Several attempts performance the aimed,Several attempts aimed at improving the performance of SIM reconstruction algorithm have been reported.,17112
of what its information subjected molecular can This fidelity. noise decrease to is transmission,This transmission of information is subjected to molecular noise what can decrease its fidelity.,17113
"drawing bounding-boxes is consuming. time very However, manually","However, manually drawing bounding-boxes is very time consuming.",17114
"variety problems of been a denoising, for deblurring, super-resolution. including has This demonstrated inpainting, and","This has been demonstrated for a variety of problems including denoising, inpainting, deblurring, and super-resolution.",17115
"formula the a algebraic using proposed measure. accomplished similarity we prediction Moreover,","Moreover, we accomplished a prediction formula using the proposed algebraic similarity measure.",17116
system. diagnosis recommender We for algorithm on new neutrosophic also a proposed based medical,We also proposed a new algorithm for medical diagnosis based on neutrosophic recommender system.,17117
of entire ligand. structure usually substructure the is the of It a,It is usually a substructure of the entire structure of the ligand.,17118
pocket induced a Sometimes is ligand by the itself.,Sometimes a pocket is induced by the ligand itself.,17119
The system respective its proteins. and sites regulator was in its studied binding FMN,The FMN system was studied and its binding sites in its respective regulator proteins.,17120
similarity. of been geometric not yet calculation the applied has in structural DCRR,DCRR has not yet been applied in the calculation of geometric structural similarity.,17121
Algorithms generate knowledge. content computer game design game that require,Algorithms that generate computer game content require game design knowledge.,17122
a as proof concept. Super Mario use We Bros. of,We use Super Mario Bros. as a proof of concept.,17123
in is tool an important Protein biology. structure structural representation,Protein structure representation is an important tool in structural biology.,17124
for is this interface a visualization the reduced aim to of Another representation. develop project,Another aim of this project is to develop a visualization interface for the reduced representation.,17125
"easier to than is comprehend DCRR AAR. Visually,","Visually, DCRR is easier to comprehend than AAR.",17126
degenerated time. lead performance query further at retrieval to This will,This will further lead to degenerated retrieval performance at query time.,17127
"TBRS*, TBRS, been implementation has of developed. recently the the A connectionist","A connectionist implementation of the TBRS, the TBRS*, has recently been developed.",17128
"algorithm is variety single-image a wavelet methods, including Our with SME of superresolution compared superresolution.","Our algorithm is compared with a variety of single-image superresolution methods, including wavelet SME superresolution.",17129
competitive results and demonstrate performance PSNR Our of SSIM. in terms numerical,Our numerical results demonstrate competitive performance in terms of PSNR and SSIM.,17130
errors. from This to statistical conflicts due optimally possible algorithm recovers,This algorithm optimally recovers from possible conflicts due to statistical errors.,17131
"Recently, detection single been image. there a for object in have doing major advances","Recently, there have been major advances for doing object detection in a single image.",17132
state-of-the-art object obtains method image single to our superior detection that We techniques. show results,We show that our method obtains superior results to state-of-the-art single image object detection techniques.,17133
tree are algorithms. list for competitors decision Rule algorithms,Rule list algorithms are competitors for decision tree algorithms.,17134
"association are they are built rules. associative classifiers, that in They pre-mined from","They are associative classifiers, in that they are built from pre-mined association rules.",17135
majority of comprises The a former significant proteins.,The former comprises a significant majority of proteins.,17136
"its a dimension. a axis Unlike along RSP an spheroprotein, possesses conspicuous longest","Unlike a spheroprotein, an RSP possesses a conspicuous axis along its longest dimension.",17137
do Rodrigues' formula the of form to matrix rotation. We a use,We use a matrix form of Rodrigues' formula to do the rotation.,17138
cylindrical to to transformation system. apply equations the then Cartesian coordinate We the,We then apply the Cartesian to cylindrical coordinate transformation equations to the system.,17139
The URL announced due publicly our will of web server course. in be,The URL of our web server will be announced publicly in due course.,17140
"GC. SP side the is of The side"" opposite ""exterior the","The ""exterior side"" of the SP is the side opposite GC.",17141
directly TSi the is is inversely while proportional. The LBS proportional to depth SPi,The SPi is directly proportional to LBS depth while the TSi is inversely proportional.,17142
both well Results reported agree SP TS from (ibid. data and methods the with,Results from both the SP and TS methods agree well with reported data (ibid.,17143
server two used may The be in web modes.,The web server may be used in two modes.,17144
"the the web cases, indices. SP In server outputs both and TS","In both cases, the web server outputs the SP and TS indices.",17145
due publicly be our announced will server course. in URL of web The,The URL of our web server will be announced publicly in due course.,17146
"this In estimation pose a for for SLR introduce domain. we dataset paper, human","In this paper, we introduce a dataset for human pose estimation for SLR domain.",17147
useful works. baseline these methods create a dataset from and future The results can for,The dataset and results from these methods can create a useful baseline for future works.,17148
"be no number there reproductive to benefits. when high, However, the is seems too","However, when the reproductive number is too high, there seems to be no benefits.",17149
Approximate a We learning propose formulation Component called Weighted (WARCA). metric Rank Analysis,We propose a metric learning formulation called Weighted Approximate Rank Component Analysis (WARCA).,17150
"Generally, CNN used proposals HOG+LUV generate and these then are to candidate the classifies proposals.","Generally, HOG+LUV are used to generate the candidate proposals and then CNN classifies these proposals.",17151
"room still its success, Despite there is improvement. for","Despite its success, there is still room for improvement.",17152
multi-layer of with a layer It integrates CNN image each channels. HOG+LUV firstly into,It firstly integrates HOG+LUV with each layer of CNN into a multi-layer image channels.,17153
"a image then AdaBoost the cascade on Based is channels, learned. multi-stage multi-layer","Based on the multi-layer image channels, a multi-stage cascade AdaBoost is then learned.",17154
combination recognition categories. achievable by various suggest is results of that experimental accurate a Our,Our experimental results suggest that accurate recognition is achievable by a combination of various categories.,17155
of of issue computation addressed. depth Here classical exact Tukey the the is data,Here the issue of exact computation of the classical Tukey data depth is addressed.,17156
similar presentation of complemented examples application. concepts to relationship and The is by,The presentation is complemented by relationship to similar concepts and examples of application.,17157
an important problem is applications. computer with vision Object variety a of localization,Object localization is an important computer vision problem with a variety of applications.,17158
to underlying are mechanisms that the give these observations? rise What,What are the underlying mechanisms that give rise to these observations?,17159
"of already algorithms e.g., common are recommendation areas, Personalization online in many retail.","Personalization of recommendation algorithms are already common in many areas, e.g., online retail.",17160
"profile routing third, during services. of is updated usage continuously And the the subsequently","And third, subsequently the profile is continuously updated during usage of the routing services.",17161
real-world dataset The FAVOUR this on that w.r.t. better-quality show experimental generates results recommendations,The experimental results on this real-world dataset show that FAVOUR generates better-quality recommendations w.r.t.,17162
classification of real-time the feasible. CNNs to detection makes application KDRP and,KDRP makes the application of CNNs to real-time detection and classification feasible.,17163
state sensory and processing Brain motor behavior. regulates for control adaptive,Brain state regulates sensory processing and motor control for adaptive behavior.,17164
enough. is of ability argue equal-size the that strategy not strong representation We,We argue that the representation ability of equal-size strategy is not strong enough.,17165
filters. consists two The of filter proposed subsequent subpatch,The proposed subpatch filter consists of two subsequent filters.,17166
network subpatch a networks called Cascaded (CSNet). These form new Network Subpatch,These subpatch networks form a new network called Cascaded Subpatch Network (CSNet).,17167
layer by The feature called is csconv layer. CSNet generated,The feature layer generated by CSNet is called csconv layer.,17168
and proposed on demonstrate the Experimental benchmark CSNet. four effectiveness of results the datasets compactness,Experimental results on four benchmark datasets demonstrate the effectiveness and compactness of the proposed CSNet.,17169
our We validate several optical using models. proposal flow variational,We validate our proposal using several optical flow variational models.,17170
Reducing challenge large the computational a remains in scale load of methods applications. these,Reducing the computational load of these methods remains a challenge in large scale applications.,17171
results illustrate the efficiency of proposed the Simulation method.,Simulation results illustrate the efficiency of the proposed method.,17172
real The approach a example. proposed illustrated simulated is data and using data,The proposed approach is illustrated using simulated data and a real data example.,17173
algorithm usefulness producing for Experiments accurate the more demonstrate the forecasts. of potential and thunderstorm,Experiments demonstrate the usefulness and potential of the algorithm for producing more accurate thunderstorm forecasts.,17174
problem from propose We manifold this learning. to of the tackle perspective,We propose to tackle this problem from the perspective of manifold learning.,17175
in of form most intestinal structures Colorectal is adenocarcinoma common the originating glandular cancer. colon,Colorectal adenocarcinoma originating in intestinal glandular structures is the most common form of colon cancer.,17176
of to automated morphology problem. the is glands a An the quantifies approach solution which,An automated approach which quantifies the morphology of glands is a solution to the problem.,17177
their is an oscillatory plants that during movement Nutation display development.,Nutation is an oscillatory movement that plants display during their development.,17178
our we experimental re-interpret example an As model. observations using previously measured,As an example we re-interpret previously measured experimental observations using our model.,17179
"scales neuroscience, many of dynamical the systems. At models mathematical in take appropriate form complex","At many scales in neuroscience, appropriate mathematical models take the form of complex dynamical systems.",17180
simulations. Protein We study folding dynamics coarse-grained L a using model molecular and,We study Protein L folding using a coarse-grained model and molecular dynamics simulations.,17181
transition. be folding is collapse to Protein concomitant in with The L the transition found,The collapse transition in Protein L is found to be concomitant with the folding transition.,17182
"we of overlap. small from In with paper, the address images scenario this self-localization","In this paper, we address the scenario of self-localization from images with small overlap.",17183
We group a of present form of novel evidence hunting.,We present evidence of a novel form of group hunting.,17184
"the the hunt. capturing model dynamics mathematical built we this, To of explore a","To explore this, we built a mathematical model capturing the dynamics of the hunt.",17185
"to more complex strategies. ""proto-cooperation"" group-hunting pre-cursor the Such be may","Such ""proto-cooperation"" may be the pre-cursor to more complex group-hunting strategies.",17186
"dynamics: absorption. and generation, the Three walking, govern actions particles'","Three actions govern the particles' dynamics: generation, walking, and absorption.",17187
particles for that generate Labeled edge compete rival new domination. against particles vertices,Labeled vertices generate new particles that compete against rival particles for edge domination.,17188
class. represent to Each subnetwork a set data connected a to tends form,Each set tends to form a connected subnetwork to represent a data class.,17189
"the uncovered relationships and vertices unfolding data the summarizes the relevant Consequently, between effect classes.","Consequently, the unfolding effect summarizes the relevant relationships between vertices and the uncovered data classes.",17190
distances is a phenomenon. long across neural Synchronization functionally important,Synchronization across long neural distances is a functionally important phenomenon.,17191
LiDAR scan utilizes LiDAR segment It line information to the data data. in the,It utilizes the scan line information in LiDAR data to segment the LiDAR data.,17192
"been have In literature, used, e.g. the algorithms many","In the literature, many algorithms have been used, e.g.",17193
"a automatic The with contours active combination and morphology. preprocessing, of segmentation achieved is","The automatic segmentation is achieved with a combination of preprocessing, active contours and morphology.",17194
that automatic possible. segmentation of The show spots results is,The results show that automatic segmentation of spots is possible.,17195
"to approaches and are based ICA. Indeed, seminal NGCA on PP","Indeed, seminal approaches to NGCA are based on PP and ICA.",17196
"and propose superiority. whitening-free paper, demonstrate experimentally In we its a this method LSNGCA","In this paper, we propose a whitening-free LSNGCA method and experimentally demonstrate its superiority.",17197
that performance. HC methods developed have utilize to Several structure the hierarchical improve the been,Several methods that utilize the hierarchical structure have been developed to improve the HC performance.,17198
We to data-driven filter hierarchy. an scalable expert-defined approach based modify propose rewiring a,We propose a scalable data-driven filter based rewiring approach to modify an expert-defined hierarchy.,17199
identification This and for proposes novel mammograms. view automatic paper algorithm MLO muscle pectoral a,This paper proposes a novel and automatic pectoral muscle identification algorithm for MLO view mammograms.,17200
suitable for scanned digital and both mammograms. It field full film is,It is suitable for both scanned film and full field digital mammograms.,17201
This a ImageJ. algorithm domain public using is demonstrated software,This algorithm is demonstrated using a public domain software ImageJ.,17202
crucial for are Standardized majority of the computer vision benchmarks applications.,Standardized benchmarks are crucial for the majority of computer vision applications.,17203
the benchmark. a paper MOTChallenge accompanies release new of This,This paper accompanies a new release of the MOTChallenge benchmark.,17204
"explanation However, the of lacked. the PCANet is","However, the explanation of the PCANet is lacked.",17205
depth. various The error is analyzed on of the PCANet parameters of rate in impact,The impact of various parameters on the error rate of PCANet is analyzed in depth.,17206
"analysis. further constructed matrix Laplacian Subsequently, for hypergraph is","Subsequently, hypergraph Laplacian matrix is constructed for further analysis.",17207
"are including learning then clustering semi-supervised algorithms, derived. hypergraph and unsupervised multi-class classification, New","New hypergraph learning algorithms, including unsupervised clustering and multi-class semi-supervised classification, are then derived.",17208
face experiments databases effectiveness method. of and Extensive the proposed on the demonstrate handwriting,Extensive experiments on face and handwriting databases demonstrate the effectiveness of the proposed method.,17209
This in inference black-box of the faster procedure. convergence results,This results in faster convergence of the black-box inference procedure.,17210
representation document features. study article the we on of image based problem visual In this,In this article we study the problem of document image representation based on visual features.,17211
in these features (i.e. We several tasks evaluate,We evaluate these features in several tasks (i.e.,17212
"and retrieval) different (e.g. in setups and clustering, classification,","classification, clustering, and retrieval) and in different setups (e.g.",17213
several domain and public datasets. in-house using transfer),domain transfer) using several public and in-house datasets.,17214
"neutrally-preferred a less) more previously preferred. became As (or result, faces","As a result, previously neutrally-preferred faces became more (or less) preferred.",17215
boosts up tasks It among synergy individual the exploits the which performances. their,It exploits the synergy among the tasks which boosts up their individual performances.,17216
"Our (HCP)"" the NIH-founded ""Human data an data Connectome release. Project is project, source public","Our data source is an NIH-founded project, the ""Human Connectome Project (HCP)"" public data release.",17217
population and a concepts play in role species The fundamental of biology.,The concepts of population and species play a fundamental role in biology.,17218
"we to species. of seek structure a fractal show First,","First, we seek to show a fractal structure of species.",17219
(PzDom) model. dominance with named model We zeta this the patch,We named this model the patch with zeta dominance (PzDom) model.,17220
on application We model. will of also developed focus an our,We will also focus on an application of our developed model.,17221
"for problem e.g. game systems exists, commercial merging This and known a level is","This is a known problem and commercial systems for game level merging exists, e.g.",17222
"more these PlasticSCM, ones. text-based are but robust only slightly than","PlasticSCM, but these are only slightly more robust than text-based ones.",17223
in biology. protein key current function is of a problem Systematic identification,Systematic identification of protein function is a key problem in current biology.,17224
allows NPPs. and discovery platform new mass NPs of The,The platform allows mass discovery of new NPs and NPPs.,17225
our classes. towards expanded protein approach challenging I other,I expanded our approach towards other challenging protein classes.,17226
ProFET (Protein a bioinformatics Engineering novel is This toolkit implemented called Toolkit). as Feature,This is implemented as a novel bioinformatics toolkit called ProFET (Protein Feature Engineering Toolkit).,17227
benchmark protein the on ProFET art of with many applied state performance. was datasets,ProFET was applied on many protein benchmark datasets with state of the art performance.,17228
"to it very solution closed-form is dimension, a compute. Importantly, efficient fixed has a and","Importantly, it has a fixed dimension, a closed-form solution and is very efficient to compute.",17229
"operate the methods the pixel instead of patch level. However, existing CNN-based at level","However, existing CNN-based methods operate at the patch level instead of the pixel level.",17230
"blurry, salient saliency especially the are objects. typically of boundary Resulting maps near","Resulting saliency maps are typically blurry, especially near the boundary of salient objects.",17231
map pixel-level image. The stream with directly an first produces saliency input from accuracy a,The first stream directly produces a saliency map with pixel-level accuracy from an input image.,17232
the Experimental our state demonstrate of model deep results art. significantly the that improves,Experimental results demonstrate that our deep model significantly improves the state of the art.,17233
problem from image this requires separating style Solving content.,Solving this problem requires separating image style from content.,17234
"limitations, general the main their of One data level remains quality. however, of low","One of their main limitations, however, remains the general low level of data quality.",17235
"address problem. to scalability Finally, suggest directions future possible this we","Finally, we suggest possible future directions to address this scalability problem.",17236
probabilities function. level for entropy the (GLCP) gray co-occurrence computing The are used,The gray level co-occurrence probabilities (GLCP) are used for computing the entropy function.,17237
experimental of indicate results high classification the accuracy. The degree,The experimental results indicate high degree of the classification accuracy.,17238
of confirm depth. the substantially the the proposed estimated that high-resolution approach Experiments quality improves,Experiments confirm that the proposed approach substantially improves the quality of the estimated high-resolution depth.,17239
accuracy method of validated data. by We simulated the the analyzing,We validated the accuracy of the method by analyzing simulated data.,17240
"key Examples the processing. interface, in a include player information immunological cell T","Examples include the T cell interface, a key player in immunological information processing.",17241
"investigate until a proceed. the reaction allowing the time empty, is sub-region to subsequent We","We investigate the time until the sub-region is empty, allowing a subsequent reaction to proceed.",17242
"hand-object propose for map. method interaction we only segmentation using depth hand Thus, a","Thus, we propose hand segmentation method for hand-object interaction using only a depth map.",17243
architecture proposed. system efficient Computationally is classification,Computationally efficient classification system architecture is proposed.,17244
multiplication algorithm apply to tensor-vector . linear It input signals utilizes fast operators upon,It utilizes fast tensor-vector multiplication algorithm to apply linear operators upon input signals .,17245
that Empirical remains versus prevention show subject a preference treatment studies for of debate.,Empirical studies show that preference for prevention versus treatment remains a subject of debate.,17246
prevention the for asymptotic an thus coverage. We expression obtain,We thus obtain an expression for the asymptotic prevention coverage.,17247
"folding, docking such in energy protein is Free predictive critical design. tasks calculation as and","Free energy calculation is critical in predictive tasks such as protein folding, docking and design.",17248
"calculation However, energy is practical these of applications. in change free expensive prohibitively rigorous","However, rigorous calculation of free energy change is prohibitively expensive in these practical applications.",17249
minimum widely to energy. The therefore utilized approximate free is potential energy,The minimum potential energy is therefore widely utilized to approximate free energy.,17250
"proxy. serve free potential may span Therefore, approximate as efficient energy alternative energy an","Therefore, potential energy span may serve as an alternative efficient approximate free energy proxy.",17251
et studied reanalyzed by then data al. We Jansen previously,We then reanalyzed data previously studied by Jansen et al.,17252
as of their splitting analysis. this take point input,take this splitting point as input of their analysis.,17253
considers variation color medium. produced model image The by the the,The model considers the image color variation produced by the medium.,17254
results proposed images obtained on underwater the foggy The method. of demonstrate capabilities and the,The obtained results on underwater and foggy images demonstrate the capabilities of the proposed method.,17255
"this with novel we optimization introduce a Patch Anchor an In paper, framework constraint.","In this paper, we introduce a novel optimization framework with an Anchor Patch constraint.",17256
reduce It containing given to non-rigidly objects. errors deformable supposed sequences overall is significantly long,It is supposed to significantly reduce overall errors given long sequences containing non-rigidly deformable objects.,17257
"dense algorithm, applied e.g. be can framework any Our tracking to","Our framework can be applied to any dense tracking algorithm, e.g.",17258
approach analysis also occlusions noise. We given quantitative our and provide of image synthetic,We also provide quantitative analysis of our approach given synthetic occlusions and image noise.,17259
and distributions evaluation Partition are probability of for important functions comparisons. model quantities,Partition functions of probability distributions are important quantities for model evaluation and comparisons.,17260
functions multimodal present a and method partition new to distributions. compute We complex of,We present a new method to compute partition functions of complex and multimodal distributions.,17261
"a this In pooling method. learning-based frame propose we paper, novel","In this paper, we propose a novel learning-based frame pooling method.",17262
"last studies intensive which, however, during remain despite decades. elusive of Physical origin two","Physical origin of which, however, remain elusive despite intensive studies during last two decades.",17263
activities. allosteric general protein strategy novel a of implicate These search for modulation findings,These findings implicate a general search strategy for novel allosteric modulation of protein activities.,17264
life the method also b estimation a direct of parameter tables. provide the We from,We also provide a direct estimation method of the parameter b from the life tables.,17265
matching on retrieval features Most are local methods shape state-of-the-art ranking. based and,Most state-of-the-art shape retrieval methods are based on local features matching and ranking.,17266
"To this decompose the overcome process decision two into we stages. challenge,","To overcome this challenge, we decompose the decision process into two stages.",17267
candidate from shape the set. Irrelevant removed are shapes,Irrelevant shapes are removed from the candidate shape set.,17268
"a second ranking that, stage. matching and After method the in (LMR) follows local-features-based","After that, a local-features-based matching and ranking (LMR) method follows in the second stage.",17269
TSR the is performance The demonstrated. of system robust retrieval,The robust retrieval performance of the TSR system is demonstrated.,17270
two mathematically We the methods are conclude that equivalent.,We conclude that the two methods are mathematically equivalent.,17271
"saliency detection. mechanism we an improved In for this paper, propose","In this paper, we propose an improved mechanism for saliency detection.",17272
"techniques existing involving hindered automation most to due feature human are Also, supervision. from extraction","Also, most existing feature extraction techniques are hindered from automation due to involving human supervision.",17273
saliency to labeling parts Object object are on based correctly. match initialized and parts relaxation,Object parts are initialized based on saliency and relaxation labeling to match object parts correctly.,17274
"non-rigid learned on criteria. then part based fitness, model and separation is A discrimination","A non-rigid part model is then learned based on fitness, separation and discrimination criteria.",17275
networks and at highly (CNNs) proven have style synthesis Convolutional neural transfer. effective image,Convolutional neural networks (CNNs) have proven highly effective at image synthesis and style transfer.,17276
meaningful control algorithm result content-aware outcome. The a offers is that the generative over,The result is a content-aware generative algorithm that offers meaningful control over the outcome.,17277
masterful include few and with paintings! into semantic doodles transfer turning style colors Applications,Applications include semantic style transfer and turning doodles with few colors into masterful paintings!,17278
"to segmentation the In information. with contextual semantic of improve this proffer we use work,","In this work, we proffer to improve semantic segmentation with the use of contextual information.",17279
"particular, in context we context `patch-patch' In CNNs. explore and `patch-background' deep","In particular, we explore `patch-patch' context and `patch-background' context in deep CNNs.",17280
"between patches. CNN-based we semantic potential correlations capture pairwise neighboring functions to formulate Specifically,","Specifically, we formulate CNN-based pairwise potential functions to capture semantic correlations between neighboring patches.",17281
We the comprehensive perform proposed method. of evaluation,We perform comprehensive evaluation of the proposed method.,17282
several hundred the It image into over-segmenting superpixels. by achieves this first,It achieves this by first over-segmenting the image into several hundred superpixels.,17283
data from and model real GDF-values for simulated to complexity-estimates models are compared of cross-validation.,GDF-values for models of simulated and real data are compared to model complexity-estimates from cross-validation.,17284
"dynamics for but mid-term to known. Neurofeedback not are last the effects short- can months,","Neurofeedback effects can last for months, but the short- to mid-term dynamics are not known.",17285
that Modeling in sessions. changes revealed successfully different induced results bidirectional behavioral DecNef,Modeling results revealed that DecNef successfully induced bidirectional behavioral changes in different sessions.,17286
about hypotheses the Neuroscientists very code. neural different formulate nature of,Neuroscientists formulate very different hypotheses about the nature of neural code.,17287
codes Information of model We encoded these for spike-trains. types binary which compare Sources,We compare these types of codes for binary Information Sources which model encoded spike-trains.,17288
of the to to vice probabilities It corresponds spike-state no-spike-state transitions from and the versa.,It corresponds to the probabilities of transitions from no-spike-state to the spike-state and vice versa.,17289
image in large-scale commonly retrieval. search a neighbour used hashing for is Similarity-preserving nearest method,Similarity-preserving hashing is a commonly used method for nearest neighbour search in large-scale image retrieval.,17290
This retrieval. suboptimal be multi-label setting for image may,This setting may be suboptimal for multi-label image retrieval.,17291
"propose representation. theoretic paper, a approach knowledge In set this we for","In this paper, we propose a set theoretic approach for knowledge representation.",17292
uses that first assumed primitive minimal We form a present knowledge constructs. and,We first present a primitive form that uses minimal assumed knowledge and constructs.,17293
"methods memory-consuming slow a and However, optimization process. requires their","However, their methods requires a slow and memory-consuming optimization process.",17294
computational a an stage. We here the burden propose learning that to moves approach alternative,We propose here an alternative approach that moves the computational burden to a learning stage.,17295
samples. systems motivates role collect datasets pivotal to The in verification signature researchers signature of,The pivotal role of datasets in signature verification systems motivates researchers to collect signature samples.,17296
culture-dependent datasets. for and of signature demands Distinct Persian signature richer offline characteristics,Distinct characteristics of Persian signature demands for richer and culture-dependent offline signature datasets.,17297
"more classes, samples, forgers. has with the other Compared more and datasets, UTSig more public","Compared with the other public datasets, UTSig has more samples, more classes, and more forgers.",17298
setups evaluate We and UTSig. four different propose test for and training,We propose and evaluate four different training and test setups for UTSig.,17299
chain of The the demonstrated ergodicity adaptive theoretically. is,The ergodicity of the adaptive chain is demonstrated theoretically.,17300
via The performance real examples. studied and is data simulations,The performance is studied via simulations and real data examples.,17301
the in The models game. relationships elements between represent probabilistic,The models represent probabilistic relationships between elements in the game.,17302
resulting reconstruction overcome complex to of exploited scenes. Temporal is in improved ambiguities coherence visual,Temporal coherence is exploited to overcome visual ambiguities resulting in improved reconstruction of complex scenes.,17303
activation implicated have long neuronal brain functions. of variety a been of Sequences in,Sequences of neuronal activation have long been implicated in a variety of brain functions.,17304
for introduces a new This tool bias paper neuronal the understanding matrix. sequences:,This paper introduces a new tool for understanding neuronal sequences: the bias matrix.,17305
"hippocampus, sequence-detection events data Using we and from interest rat describe techniques. of also","Using data from rat hippocampus, we describe events of interest and also sequence-detection techniques.",17306
"when sub-problems appealing observed interdependent. performance are seldom can the be However,","However, appealing performance can be seldom observed when the sub-problems are interdependent.",17307
"original global guaranteed. is Meanwhile, convergence to the optimum problem) (of still the the","Meanwhile, the convergence to the global optimum (of the original problem) is still guaranteed.",17308
is empirically non-separable sets The DAC of two on effectiveness demonstrated high-dimensional problems. of,The effectiveness of DAC is demonstrated empirically on two sets of non-separable high-dimensional problems.,17309
"In reality, flow depending optical class. on image varies across an object","In reality, optical flow varies across an image depending on object class.",17310
are wild fish populations. marine of factors Abiotic fundamental of the dynamics drivers,Abiotic factors are fundamental drivers of the dynamics of wild marine fish populations.,17311
of sparse propose different methods. detection also framework with codes combining our We,We also propose our framework of combining sparse codes with different detection methods.,17312
"and we image that, candidate shape, gender, can face After key determine face points parts.","After that, we can determine image candidate gender, face shape, key points and face parts.",17313
"military, model can police, healthcare, used in be technology and defense, sectors. The","The model can be used in military, police, defense, healthcare, and technology sectors.",17314
"nationality face image. Finally, the Computer distinguish from can","Finally, Computer can distinguish nationality from the face image.",17315
object recognition on object preprocessing performing a step. build top proposals Current systems as,Current top performing object recognition systems build on object proposals as a preprocessing step.,17316
edge an We algorithm clear without handling devise for objects grouping boundaries.,We devise an edge grouping algorithm for handling objects without clear boundaries.,17317
both of and question. experimental We this explorations theoretical present,We present both theoretical and experimental explorations of this question.,17318
regularization show how can results aid approximation. Our this also,Our results also show how regularization can aid this approximation.,17319
the and solvers also model. proposed efficient a minimize to We provide energy framework,We also provide a framework and efficient solvers to minimize the proposed energy model.,17320
[Yang multivariate graphical al. models Previous et,Previous multivariate graphical models [Yang et al.,17321
"dependencies. real-world datasets, have many in However, variables clearly positive","However, in many real-world datasets, variables clearly have positive dependencies.",17322
and top-down connecting selection the computation. form and a fusion The bottom-up feedback,The selection and fusion form a feedback connecting the top-down and bottom-up computation.,17323
approach. to experiments demonstrate We our the of provide effectiveness numerical,We provide numerical experiments to demonstrate the effectiveness of our approach.,17324
object detection task state-of-the-art in works. proposals benefit recent greatly Object,Object proposals greatly benefit object detection task in recent state-of-the-art works.,17325
encountered Manifold-valued are tasks. in datasets many computer widely vision,Manifold-valued datasets are widely encountered in many computer vision tasks.,17326
"efficient we an Riemannian (CCM-EPGA). for PGA paper, In propose manifolds curvature exact this constant","In this paper, we propose an efficient exact PGA for constant curvature Riemannian manifolds (CCM-EPGA).",17327
capture of indoor both to a shape reflectance scene. an We device portable present and,We present a portable device to capture both shape and reflectance of an indoor scene.,17328
device of real-world on Effectiveness scenes. demonstrated by results our is quality,Effectiveness of our device is demonstrated by quality results on real-world scenes.,17329
"external timestep deformation, large extreme force stable It under is or size.","It is stable under extreme deformation, external force or large timestep size.",17330
user constraints we readily And in Euclidean handle various can space.,And we can readily handle various user constraints in Euclidean space.,17331
performed system specification. condition scene illumination and by the evaluation considering camera Pattern is,Pattern evaluation is performed by considering the scene illumination condition and camera system specification.,17332
accurate It can be estimation and nonparametric for densities resampling. of copula further used,It can be further used for accurate nonparametric estimation of copula densities and resampling.,17333
is simulations. the performance methods illustrated The of by,The performance of the methods is illustrated by simulations.,17334
can number axioms large provers. overwhelm theorem of irrelevant resolution-based easily A,A large number of irrelevant axioms can easily overwhelm resolution-based theorem provers.,17335
"ordering reasoning this we of in for optimization paper, heuristics systems. such describe two In","In this paper, we describe two ordering heuristics for optimization of reasoning in such systems.",17336
that methods time. in inference to magnitude lead Results an these order show reduction of,Results show that these methods lead to an order of magnitude reduction in inference time.,17337
"primarily used codes accelerating search. Conventionally, are for image hash compact","Conventionally, compact hash codes are primarily used for accelerating image search.",17338
is binary first kind work classifiers to Our its of represent codes. using,Our work is first of its kind to represent classifiers using binary codes.,17339
"classification image we formulate an multi-class optimization Specifically, binary variables. as problem over","Specifically, we formulate multi-class image classification as an optimization problem over binary variables.",17340
binary The codes. classifiers optimization over proceeds hash alternatively and the image,The optimization alternatively proceeds over the binary classifiers and image hash codes.,17341
benchmarks. image conducted are several on evaluations Comprehensive representative,Comprehensive evaluations are conducted on several representative image benchmarks.,17342
"reduced model complexities The consistently of experiments accuracies. deployment, sacrifice observe and of training without","The experiments consistently observe reduced complexities of model training and deployment, without sacrifice of accuracies.",17343
"Finally, fitted primitives. candidate are geometric to the ellipses","Finally, candidate ellipses are fitted to the geometric primitives.",17344
"deviation is during reading. rate of blink higher variability standard Moreover, the","Moreover, standard deviation of the blink rate variability is higher during reading.",17345
"detection. applications have neural convolutional computer on other under-performed pedestrian vision, to Compared in networks","Compared to other applications in computer vision, convolutional neural networks have under-performed on pedestrian detection.",17346
empirically implementation disclose We important and details. identify,We empirically identify and disclose important implementation details.,17347
We a also KITTI the state-of-the-art to on dataset. comparable result achieve the approaches,We also achieve a comparable result to the state-of-the-art approaches on the KITTI dataset.,17348
with in Recognizing unsolved text is many a natural task problems. images challenging,Recognizing text in natural images is a challenging task with many unsolved problems.,17349
"detection we In multi-object using method paper, depth a and propose cameras. this tracking","In this paper, we propose a multi-object detection and tracking method using depth cameras.",17350
object detection. Depth are noisy and in very maps obscure,Depth maps are very noisy and obscure in object detection.,17351
some not in enough. NNQ accurate is scenarios,NNQ in some scenarios is not accurate enough.,17352
"connects keypoints frames, local matching frame. pairs TRGMC of each by of densely","TRGMC densely connects pairs of frames, by matching local keypoints of each frame.",17353
has TRGMC a wide scenarios. in of results superior Experimental range that performance demonstrate,Experimental results demonstrate that TRGMC has superior performance in a wide range of scenarios.,17354
can Tick-Borne be transmission. (NS) via non-systemic diseases transmitted,Tick-Borne diseases can be transmitted via non-systemic (NS) transmission.,17355
"systemic peculiar, require does not host. infection it as This any transmission the of is","This transmission is peculiar, as it does not require any systemic infection of the host.",17356
We vanishing a the rare prove relative property. possesses that error asymptotic estimator,We prove that the estimator possesses a rare vanishing relative error asymptotic property.,17357
to of due Reducing the candidates scale is possible number important the the problem. of,Reducing the number of possible candidates is important due to the scale of the problem.,17358
problem Interactive and has selection many very important is object applications. a research,Interactive object selection is a very important research problem and has many applications.,17359
foreground and background require estimate algorithms the user Previous distributions. interactions to substantial,Previous algorithms require substantial user interactions to estimate the foreground and background distributions.,17360
is Our beneficial. integrating into calculation indicate that saliency experiments depth video,Our experiments indicate that integrating depth into video saliency calculation is beneficial.,17361
"data are the segmented clustering the spectral matrix. to by applying Second, affinity","Second, the data are segmented by applying spectral clustering to the affinity matrix.",17362
segmentation consistent Our results. more yields method,Our method yields more consistent segmentation results.,17363
on Experimental of results effectiveness benchmarks method. the our demonstrate,Experimental results on benchmarks demonstrate the effectiveness of our method.,17364
a test with We method phantoms methods. and our different variety reconstruction of,We test our method with a variety of phantoms and different reconstruction methods.,17365
"localization. learnt, approach to segmentation a the refine to detector Once the is resort we","Once the detector is learnt, we resort to a segmentation approach to refine the localization.",17366
"state-of-the-art despite approach its methods. show our that simplicity, We outperforms","We show that despite its simplicity, our approach outperforms state-of-the-art methods.",17367
demonstrate compared the the algorithm state-of-the-arts. our to of effectiveness Extensive experiments,Extensive experiments demonstrate the effectiveness of our algorithm compared to the state-of-the-arts.,17368
"power, that baseline We it outperforms previous showing by consistently predictive approaches. demonstrate EgoNet's","We demonstrate EgoNet's predictive power, by showing that it consistently outperforms previous baseline approaches.",17369
transitions? of anatomical facilitate network such does How connections brain's the help,How does the brain's network of anatomical connections help facilitate such transitions?,17370
difficult? contribute transition easy this to transition features and network Which one of making another,Which features of this network contribute to making one transition easy and another transition difficult?,17371
"these we control address theory. using questions Here, network","Here, we address these questions using network control theory.",17372
brain made The contributions input by to signals allow assess the different us regions.,The input signals allow us to assess the contributions made by different brain regions.,17373
Networks segmentation neural State-of-the-art (FCNs). established by semantic Fully are of Convolutional results,State-of-the-art results of semantic segmentation are established by Fully Convolutional neural Networks (FCNs).,17374
"the of spatially recurrent work, (i.e. use In this layers advocate we","In this work, we advocate the use of spatially recurrent layers (i.e.",17375
global representations. and which contexts capture lead directly improved layers) feature ReNet to,ReNet layers) which directly capture global contexts and lead to improved feature representations.,17376
"with (H-ReNet). ReNet Hybrid Furthermore, a ReNet novel and integrate deep FCNs, we develop layers","Furthermore, we integrate ReNet layers with FCNs, and develop a novel Hybrid deep ReNet (H-ReNet).",17377
approximations convolutional efficient two XNOR-Networks. propose networks: standard to neural Binary-Weight-Networks and We,We propose two efficient approximations to standard convolutional neural networks: Binary-Weight-Networks and XNOR-Networks.,17378
"both input XNOR-Networks, to convolutional the and the are layers binary. In filters","In XNOR-Networks, both the filters and the input to convolutional layers are binary.",17379
binary approximate operations. using primarily convolutions XNOR-Networks,XNOR-Networks approximate convolutions using primarily binary operations.,17380
state-of-the-art of CPUs running (rather possibility the than offer XNOR-Nets in GPUs) on real-time. networks,XNOR-Nets offer the possibility of running state-of-the-art networks on CPUs (rather than GPUs) in real-time.,17381
"on Our tasks. networks work accurate, simple, efficient, challenging and visual binary are","Our binary networks are simple, accurate, efficient, and work on challenging visual tasks.",17382
ImageNet We task. evaluate on our approach classification the,We evaluate our approach on the ImageNet classification task.,17383
of for (ML) likelihood regression is computation The maximum the considered. estimator models heteroscedastic,The computation of the maximum likelihood (ML) estimator for heteroscedastic regression models is considered.,17384
models time processes. stochastic representations serve Linear autoregressive as basic discrete of,Linear autoregressive models serve as basic representations of discrete time stochastic processes.,17385
of the mathematical of evolutionary We analysis perform dynamics a policing.,We perform a mathematical analysis of the evolutionary dynamics of policing.,17386
derive invasion policing the and of conditions for precise alleles. We stability,We derive precise conditions for the invasion and stability of policing alleles.,17387
calculate when evolutionarily are strategies determine evolutionarily singular stable. they also and We,We also calculate evolutionarily singular strategies and determine when they are evolutionarily stable.,17388
"As optimization additional result, an a possible. is","As a result, an additional optimization is possible.",17389
using consumer-level mobile such device phone. a as a,using a mobile consumer-level device such as a phone.,17390
a achieve autoencoder. analogous this to network the in by training an We manner,We achieve this by training the network in a manner analogous to an autoencoder.,17391
"Internet. part teaching statistics should include actually the from an However, data of important retrieving","However, an important part of teaching statistics should include actually retrieving data from the Internet.",17392
performs KPCA missing and presence in Our approach of noise. robust the data,Our approach performs robust KPCA in the presence of missing data and noise.,17393
on the missing standard state-of-the-art We predicting entries demonstrate results in flow oil dataset.,We demonstrate state-of-the-art results on predicting missing entries in the standard oil flow dataset.,17394
challenging We of of on the Camera RGB demonstrate Relocalization. our approach efficacy the problem,We demonstrate the efficacy of our approach on the challenging problem of RGB Camera Relocalization.,17395
able the relocalization results. method camera experiments to reveal is proposed state-of-the-art achieve Our that,Our experiments reveal that the proposed method is able to achieve state-of-the-art camera relocalization results.,17396
signatures needed data. from estimate to labels set mixed are level) target,set level) labels are needed to estimate target signatures from mixed data.,17397
the regions analyst by labels for eFUMI. highlighted The as imprecise the serve,The regions highlighted by the analyst serve as the imprecise labels for eFUMI.,17398
"iteratively analyst want labels. to refine Often, an may imprecise their","Often, an analyst may want to iteratively refine their imprecise labels.",17399
shown sets. on hyperspectral real data are detection Results target sub-pixel,Results are shown on real hyperspectral sub-pixel target detection data sets.,17400
illustrated detection is on images. several digital algorithm This,This detection algorithm is illustrated on several digital images.,17401
about wall? desk How a a pushing against,How about pushing a desk against a wall?,17402
by large-scale Training dataset requires model movements external caused of object forces. our a,Training our model requires a large-scale dataset of object movements caused by external forces.,17403
to are receiver other leading Results a methods. summarized (ROC) operator compared and as curve,Results are summarized as a receiver operator curve (ROC) and compared to other leading methods.,17404
"behavioral and are with practice. psychological Overall, the congruent activation of aims dissociable patterns each","Overall, dissociable activation patterns are congruent with the psychological and behavioral aims of each practice.",17405
balance modeling problems population can that Such need applied. classical be numerical technique so non-conventional,Such problems need non-conventional population balance modeling so that classical numerical technique can be applied.,17406
developed validated method the solutions. analytical A recently numerical numerical and is against efficient,A recently developed efficient numerical method is validated against the analytical and numerical solutions.,17407
latent family structure network models. variables augments model neural graphical Our with in observation,Our model family augments graphical structure in latent variables with neural network observation models.,17408
phenotyping. mouse example models framework several We illustrate this with to application an behavioral and,We illustrate this framework with several example models and an application to mouse behavioral phenotyping.,17409
"labeled of millions training such networks, needed are To large images. train with very sets","To train such networks, very large training sets are needed with millions of labeled images.",17410
a of invariances into discriminative general classifier. a for We these introduce types incorporating framework,We introduce a general framework for incorporating these types of invariances into a discriminative classifier.,17411
as from The process. slice Levy a drawn imagines a being data framework of,The framework imagines data as being drawn from a slice of a Levy process.,17412
applies FLG MLG subgraphs kernel such to The kernels recursively.,The MLG kernel applies such FLG kernels to subgraphs recursively.,17413
of large with a a multi-neighborhood algorithm local consider We search neighborhoods. possible number,We consider a multi-neighborhood local search algorithm with a large number of possible neighborhoods.,17414
"potentially Moreover, similar applied can in this contexts. problem-independent method is be and","Moreover, this method is problem-independent and potentially can be applied in similar contexts.",17415
"but related In approaches, other in not contrast are the to corresponding weights layers shared.","In contrast to other approaches, the weights in corresponding layers are related but not shared.",17416
similarity compared and the are discussed. The and of methods difference two,The similarity and difference of the two methods are compared and discussed.,17417
the start We of frame-interpolation problem from that insight for solves implicitly the correspondences. inter-frame,We start from the insight that the problem of frame-interpolation implicitly solves for inter-frame correspondences.,17418
matching present image simply therefore The model learns watching videos. by,The present model therefore learns image matching by simply watching videos.,17419
Current by and images end-to-end sampling uniform learning generate from frameworks noise generative use distribution.,Current generative frameworks use end-to-end learning and generate images by sampling from uniform noise distribution.,17420
"first and joint independently, are merged GANs The via learning. trained two together then","The two GANs are first trained independently, and then merged together via joint learning.",17421
uncertainty This elliptic equation. paper for quantification considers an nonlocal,This paper considers uncertainty quantification for an elliptic nonlocal equation.,17422
cannot posteriors elliptic As be are equation solved approximate nonlocal the constructed.,As the elliptic nonlocal equation cannot be solved approximate posteriors are constructed.,17423
"most pairwise based geometric relations between graph parts. local However, consider trackers","However, most graph based trackers consider pairwise geometric relations between local parts.",17424
an A in of candidates. results large in search a usually such number index,A search in such an index usually results in a large number of candidates.,17425
a aims review in paper this field. of progress to provide recent the This,This paper aims to provide a review of the recent progress in this field.,17426
"users servers simpler, These the process. exclude from modeling make but largely","These servers make modeling simpler, but largely exclude users from the process.",17427
We Modeling (IMPi) for Pipeline present modeling. Interactive an homology,We present an Interactive Modeling Pipeline (IMPi) for homology modeling.,17428
pre-trained convolutional features neural Frame-level network a are by firstly deep extract (CNN).,Frame-level features are firstly extract by a pre-trained deep convolutional neural network (CNN).,17429
"and are DFT features. domain interpolated features Then, transferred time into","Then, time domain features are transferred and interpolated into DFT features.",17430
fused for and emotion and CNN DFT features classification. further encoded are,CNN and DFT features are further encoded and fused for emotion classification.,17431
"can Shadows them be challenging. artifacts very and in photographs, create unwanted removing often","Shadows often create unwanted artifacts in photographs, and removing them can be very challenging.",17432
refined handle on confidence image the based result unique then The reconstruction is to patterns.,The result is then refined based on the reconstruction confidence to handle unique image patterns.,17433
results Many show removal the improvement. effectiveness our of comparisons shadow are and,Many shadow removal results and comparisons are show the effectiveness of our improvement.,17434
discriminative supervised dictionary literature. been Many methods in the effective developed learning have,Many effective supervised discriminative dictionary learning methods have been developed in the literature.,17435
"are in not applications, labels many always accurate Yet, feasible.","Yet, in many applications, accurate labels are not always feasible.",17436
are WEMI and data. on Results discussed presented measured,Results are presented and discussed on measured WEMI data.,17437
image technical algorithm. report improved describes mosaicking This an,This technical report describes an improved image mosaicking algorithm.,17438
"research video Recently, classification efforts. intensive attracts","Recently, video classification attracts intensive research efforts.",17439
extract by deep features pre-trained are convolutional neural firstly (CNN). Frame-level a network,Frame-level features are firstly extract by a pre-trained deep convolutional neural network (CNN).,17440
"Then, features interpolated features. time are into DFT domain transformed and","Then, time domain features are transformed and interpolated into DFT features.",17441
We and test video emotion recognition. classification for method action our,We test our method for video emotion classification and action recognition.,17442
"a develop Utilizing notion self-similarity, local we local of for model fractal first the images.","Utilizing the notion of local self-similarity, we first develop a local fractal model for images.",17443
The a proposed with method CNN-based image us interpretation provides model. filtering geometric novel of,The proposed filtering method provides us with a novel geometric interpretation of CNN-based image model.,17444
approximate but biased. is The gradient computationally-cheap,The approximate gradient is computationally-cheap but biased.,17445
questions. the provides open algorithm Whether are asymptotically and estimate still why consistent an CD,Whether and why the CD algorithm provides an asymptotically consistent estimate are still open questions.,17446
"are memory intensive. and computation typically very algorithms However, ConvNet","However, ConvNet algorithms are typically very computation and memory intensive.",17447
on proposed. integer linear aided (ALIP) disaggregation based Load programming is,Load disaggregation based on aided linear integer programming (ALIP) is proposed.,17448
"IP-based aid improved. these of With performance disaggregation significantly is the enhancements, the of","With the aid of these enhancements, the performance of IP-based disaggregation is significantly improved.",17449
parts localize is features. discriminative extract to to pose-invariant problem this A to key address,A key to address this problem is to localize discriminative parts to extract pose-invariant features.,17450
"to ground-truth be However, can annotations expensive part acquire.","However, ground-truth part annotations can be expensive to acquire.",17451
"it many Moreover, for is hard classes. fine-grained parts define to","Moreover, it is hard to define parts for many fine-grained classes.",17452
"for to context ""how contents? its better surrounding a local proposal inferring exploit and","and ""how to exploit local context surrounding a proposal for better inferring its contents?",17453
information incorporates and CNN into effectively region-based local (e.g. AC-CNN contextual global the,AC-CNN effectively incorporates global and local contextual information into the region-based CNN (e.g.,17454
better and provides model performance. detection RCNN) detection Fast object,Fast RCNN) detection model and provides better object detection performance.,17455
"violate rules, parking detection example, For traffic enforcement, vehicles of that etc.","For example, detection of vehicles that violate traffic rules, parking enforcement, etc.",17456
log-scale. Taylor used the expansions improve series are on accuracy Simple to,Simple Taylor series expansions are used to improve accuracy on the log-scale.,17457
of functions package part the statmod the CRAN IGD from are The repository. available,The IGD functions are part of the statmod package available from the CRAN repository.,17458
"this In replace MLP shallow we MLP. shallow to propose sparse with paper, dense","In this paper, we propose to replace dense shallow MLP with sparse shallow MLP.",17459
"benchmark development attribute for recognition However, datasets the remains suitable of lagged behind.","However, the development of suitable benchmark datasets for attribute recognition remains lagged behind.",17460
challenge fine-grained recognition developing algorithms. poses Such attribute heterogeneous quality big high a collection on,Such heterogeneous collection poses a big challenge on developing high quality fine-grained attribute recognition algorithms.,17461
novel introduces convolutional estimation. of the a for work architecture pose This task network human,This work introduces a novel convolutional network architecture for the task of human pose estimation.,17462
the MPII benchmarks FLIC results recent State-of-the-art methods. on and achieved all are outcompeting,State-of-the-art results are achieved on the FLIC and MPII benchmarks outcompeting all recent methods.,17463
responses a understanding Written provide reasoning wealth topic. student on a in data can of,Written responses can provide a wealth of data in understanding student reasoning on a topic.,17464
inter-rater We between is there the and scoring. substantial find reliability that ML human,We find that there is substantial inter-rater reliability between the human and ML scoring.,17465
"particularly image. labels facial, Attribute recognition, many extracts each for","Attribute recognition, particularly facial, extracts many labels for each image.",17466
images in tasks. building between a the similarity various patches in block Measuring fundamental is,Measuring the similarity between patches in images is a fundamental building block in various tasks.,17467
"patches our patch assumption using the sampled, is large sufficiently that database (e.g. Under","Under the assumption that our patch database is sufficiently sampled, using large patches (e.g.",17468
This of small the applications. in patch-size favored choice most explains,This explains the favored choice of small patch-size in most applications.,17469
work approach. we offer such In this an,In this work we offer such an approach.,17470
"increase (e.g. of dimensions minor a Therefore, with the","Therefore, with a minor increase of the dimensions (e.g.",17471
based surrounding. The computed patch self-similarity a additional behavior on the are of descriptors,The additional descriptors are computed based on a self-similarity behavior of the patch surrounding.,17472
a present using single image. We surface estimation normal (NIR) infrared near,We present surface normal estimation using a single near infrared (NIR) image.,17473
are light on focusing fine-scale surface We with uncalibrated source. an geometry captured,We are focusing on fine-scale surface geometry captured with an uncalibrated light source.,17474
doing have so methods studied. for been Several,Several methods for doing so have been studied.,17475
"systems for are embedded high such as too smartphones. their However, costs computational","However, their computational costs are too high for embedded systems such as smartphones.",17476
"yet simple, stabilizer. a algorithm, surprisingly effective propose the called We stitching","We propose a simple, yet surprisingly effective algorithm, called the stitching stabilizer.",17477
stitches It a frames with computational only cost. minimal together two,It stitches only two frames together with a minimal computational cost.,17478
"To apply we clearly hyperlapse. to effect, show the it","To clearly show the effect, we apply it to hyperlapse.",17479
or RGB modality sequences explored depth recognition extensively recently. on been action has Single,Single modality action recognition on RGB or depth sequences has been extensively explored recently.,17480
"uniformly behaves various under resolutions. energy mesh this As such,","As such, this energy behaves uniformly under various mesh resolutions.",17481
literature. of in of our exist the the elements All already approach,All of the elements of our approach already exist in the literature.,17482
This drawing paper undirected graphs. and multi-surface techniques general studies force-directed algorithms for the,This paper studies the force-directed algorithms and multi-surface techniques for drawing general undirected graphs.,17483
"new for multi-surface Next, algorithms. we clustering approach method on based fuzzy a introduce","Next, we introduce a new method for multi-surface approach based on fuzzy clustering algorithms.",17484
image fully We a automatic colorization system. develop,We develop a fully automatic image colorization system.,17485
"approach advances both in networks, low-level and leverages Our exploiting deep representations. recent semantic","Our approach leverages recent advances in deep networks, exploiting both low-level and semantic representations.",17486
"and we tasks, colorization partially outperform fully methods. On existing automatic both","On both fully and partially automatic colorization tasks, we outperform existing methods.",17487
for explore self-supervised colorization visual a We also vehicle as learning. representation,We also explore colorization as a vehicle for self-supervised visual representation learning.,17488
analysis segmentation. tasks microarray challenging image is of spot most in the One,One of the most challenging tasks in microarray image analysis is spot segmentation.,17489
have leaps. capabilities recognition extraordinary made recently Face,Face recognition capabilities have recently made extraordinary leaps.,17490
"Rather harvesting synthesize and more manually faces, we than them. simply labeling","Rather than manually harvesting and labeling more faces, we simply synthesize them.",17491
system for recognition and deep learning models. face method our multiple using pose-aware We introduce,We introduce our method and system for face recognition using multiple pose-aware deep learning models.,17492
segmentation and Semantic annotations. require labelling are two costly instance that tasks particularly,Semantic labelling and instance segmentation are two tasks that require particularly costly annotations.,17493
"of this a we propose Gaussian generalization called Morphable In SSMs, Models (GPMMs). Process paper,","In this paper, we propose a generalization of SSMs, called Gaussian Process Morphable Models (GPMMs).",17494
as resulting can an model seen The be of continuous a SSM. analogon,The resulting model can be seen as a continuous analogon of an SSM.,17495
"models. different combine Gaussian possible processes to make it Furthermore,","Furthermore, Gaussian processes make it possible to combine different models.",17496
or a algorithm for a a GPMM to fitting surface We image. simple introduce,We introduce a simple algorithm for fitting a GPMM to a surface or image.,17497
"whose GPMM. approach, This non-rigid properties defined results a regularization in registration by are a","This results in a non-rigid registration approach, whose regularization properties are defined by a GPMM.",17498
"and falls accuracy of challenging remains it short However, still and classification efficiency.","However, it still remains challenging and falls short of classification accuracy and efficiency.",17499
"leverages computing. MCNN also it is naturally computationally efficient, as GPU","MCNN is also computationally efficient, as it naturally leverages GPU computing.",17500
"makes problem the is which low, difficult. patent very The of rate litigations consequently","The rate of patent litigations is very low, which consequently makes the problem difficult.",17501
for capture perspective. modified time-to-litigation to likelihood initial predicting model the is a further The,The initial model for predicting the likelihood is further modified to capture a time-to-litigation perspective.,17502
have End-to-end of methods vision. in many achieved learning areas computer results impressive,End-to-end learning methods have achieved impressive results in many areas of computer vision.,17503
is This as shift the effect. known domain,This is known as the domain shift effect.,17504
network parameters. focus adaptation retraining Recently methods proposed the on,Recently proposed adaptation methods focus on retraining the network parameters.,17505
"shift pop we the layer, Surprisingly, that up. domain already at first effects very find","Surprisingly, we find that already at the very first layer, domain shift effects pop up.",17506
improves performance datasets. network of the This benchmark various on deep the,This improves the performance of the deep network on various benchmark datasets.,17507
"neural network Recently, recognition. deep image face shown in performance promising has","Recently, deep neural network has shown promising performance in face image recognition.",17508
demonstrate Experiments video the on the network. of datasets effectiveness two proposed face public,Experiments on two public face video datasets demonstrate the effectiveness of the proposed network.,17509
exact computation of This enables and average velocity run-length.,This enables exact computation of average velocity and run-length.,17510
for easily method use. a general is provide package implementable and we Our software,Our method is easily implementable and we provide a software package for general use.,17511
an objectively measure of performance offensive one NFL? the lineman individual the in does How,How does one objectively measure the performance of an individual offensive lineman in the NFL?,17512
"sets, training and set respectively. each First, data data is split testing into","First, each data set is split into training and testing data sets, respectively.",17513
"using training model estimated parameters and are linear data nonlinear the classification set. Then,","Then, linear and nonlinear classification model parameters are estimated using the training data set.",17514
maximize reduction performed to is classification Data then accuracy. dimensionality,Data dimensionality reduction is then performed to maximize classification accuracy.,17515
stereo using expansion present matching cuts. We accurate local an method moves based graph on,We present an accurate stereo matching method using local expansion moves based on graph cuts.,17516
defined alpha-expansions presented small regions. as The for grid are expansion local many moves,The local expansion moves are presented as many alpha-expansions defined for small grid regions.,17517
localization ways: moves local and propagation. The by traditional spatial two expansion moves extend expansion,The local expansion moves extend traditional expansion moves by two ways: localization and spatial propagation.,17518
"the different localization, alpha-labels local we locations By of candidate according alpha-expansions. use to","By localization, we use different candidate alpha-labels according to the locations of local alpha-expansions.",17519
with frame. attention a mixture visual each Gaussians We of at model,We model visual attention with a mixture of Gaussians at each frame.,17520
of is each This distribution express for saliency pixel. probability the to used,This distribution is used to express the probability of saliency for each pixel.,17521
"regularization In techniques circumvent cases so, bias. the to we doing various normal-vs-diseased large adopt","In doing so, we adopt various regularization techniques to circumvent the large normal-vs-diseased cases bias.",17522
video vision common Non-rigid task. computer is a interpolation,Non-rigid video interpolation is a common computer vision task.,17523
semidefinite quadratic matrix (QML) (SDP) as standard problem. We a learning programming formulate,We formulate quadratic matrix learning (QML) as a standard semidefinite programming (SDP) problem.,17524
"as such in filters, result Biologically inspired features, very dimensional Gabor high measurement.","Biologically inspired features, such as Gabor filters, result in very high dimensional measurement.",17525
Does dimensionality advantages the space efficiency? computational feature beyond the afford reducing of,Does reducing the dimensionality of the feature space afford advantages beyond computational efficiency?,17526
reduction approaches dimensionality action unit detection? to some Do (DR) yield improved,Do some approaches to dimensionality reduction (DR) yield improved action unit detection?,17527
tracked using and aligned (AAM). Facial models features appearance were active,Facial features were tracked and aligned using active appearance models (AAM).,17528
SIFT local Gabor and facial were from features regions. extracted,SIFT and Gabor features were extracted from local facial regions.,17529
"as a no-DR further For included condition well. comparison, was control","For further comparison, a no-DR control condition was included as well.",17530
Gabor with PCA previous results were features research. with for comparable Baseline,Baseline results for PCA with Gabor features were comparable with previous research.,17531
"to exceptions, AU no-DR. DR With notable some improved relative detection","With some notable exceptions, DR improved AU detection relative to no-DR.",17532
approaches proved Locality problems. embedding to vulnerable \emph{out-of-sample},Locality embedding approaches proved vulnerable to \emph{out-of-sample} problems.,17533
the better to SIFT than Gradient-based features. filter-based AU Gabor detection lead,Gradient-based SIFT lead to better AU detection than the filter-based Gabor features.,17534
"the approaches. differences area were under found linear few DR other and For curve, between","For area under the curve, few differences were found between linear and other DR approaches.",17535
"both among varied of results units. metrics, action pattern the For","For both metrics, the pattern of results varied among action units.",17536
LDA approaches; least most efficient. were and was the efficient PCA KDA the,PCA and LDA were the most efficient approaches; KDA was the least efficient.,17537
"been shiftable to restricted range Previous kernels. based approaches Gaussian have, however, approximation on","Previous approaches based on shiftable approximation have, however, been restricted to Gaussian range kernels.",17538
form the decomposition. recursive QR be using can coefficients of The computed a efficiently,The coefficients can be efficiently computed using a recursive form of the QR decomposition.,17539
We accuracy results speed the present algorithm. and of proposed simulation the to demonstrate,We present simulation results to demonstrate the speed and accuracy of the proposed algorithm.,17540
This allows us given to to to a required $N$ accuracy. obtain estimate the order,This allows us to to estimate the order $N$ required to obtain a given accuracy.,17541
the is subject. order the The partial to index authors' approach,The partial order index is the authors' approach to the subject.,17542
of the measures. Conducted considered properties experiments the show,Conducted experiments show the properties of the considered measures.,17543
"of are strong conclusions, presented. weak sides In variant and the each","In conclusions, the strong and weak sides of each variant are presented.",17544
simple detects by and mode effective model seeking algorithm. a MSH intuitively instances,MSH intuitively detects model instances by a simple and effective mode seeking algorithm.,17545
novel method semi-supervised on recognition. This object proposes paper a,This paper proposes a novel semi-supervised method on object recognition.,17546
rationality of is analyzed theory. the The in strategy also,The rationality of the strategy is also analyzed in theory.,17547
signal problem to of blind wavelet domain. The is separation reassigned the,The problem of blind signal separation is reassigned to the wavelet domain.,17548
"domain. pre-whitened is image the the Initially, observed space in","Initially, the observed image is pre-whitened in the space domain.",17549
highly depends immune cells. collection a B human The diverse system antibody-making of on,The human immune system depends on a highly diverse collection of antibody-making B cells.,17550
receptors resulting throughput and in The sequenced for research can be high diagnostics.,The resulting receptors can be sequenced in high throughput for research and diagnostics.,17551
to to sampled cluster came same according events. from sequences which the rearrangement,to cluster sampled sequences according to which came from the same rearrangement events.,17552
"inference."" this problem clustering call ""clonal We family","We call this clustering problem ""clonal family inference.""",17553
convincing over experimental state-of-the-art Our results deblurring existing improvements methods. extensive shown have,Our extensive experimental results have shown convincing improvements over existing state-of-the-art deblurring methods.,17554
This accurate inherent tracking to and enables standard alignment. reduces model-to-depth drift drastically,This enables accurate tracking and drastically reduces drift inherent to standard model-to-depth alignment.,17555
Our motion results features. tracking geometric robust that for and fast even lack scenes demonstrate,Our results demonstrate robust tracking even for fast motion and scenes that lack geometric features.,17556
propose We automatic produces colorizations. realistic fully vibrant and that a approach,We propose a fully automatic approach that produces vibrant and realistic colorizations.,17557
results in several benchmarks. learning This state-of-the-art feature approach on performance,This approach results in state-of-the-art performance on several feature learning benchmarks.,17558
"We apply our to segmentation. method a scheme, multi-region motion provide and","We provide a multi-region scheme, and apply our method to motion segmentation.",17559
estimation show effect investigate the of parameters important. is realism on We performance and rendering,We investigate the effect of rendering parameters on estimation performance and show realism is important.,17560
We the validated qualitative promising convergence have and results. quantitative and demonstrated empirically,We have empirically validated the convergence and demonstrated promising quantitative and qualitative results.,17561
higher Category are those previous significantly quality of discovered. work in labels than,Category labels of significantly higher quality than those in previous work are discovered.,17562
"a detection propose we this Transform-based novel In paper, Hough object approach.","In this paper, we propose a novel Hough Transform-based object detection approach.",17563
an variant (PLS). is Partial Squares efficient BPLS of Least,BPLS is an efficient variant of Partial Least Squares (PLS).,17564
"multi-scale handle Second, a scheme. to efficiently we propose scale voting novel object changes,","Second, to efficiently handle object scale changes, we propose a novel multi-scale voting scheme.",17565
"images obtained simultaneously. In this object can scheme, Hough corresponding be to scales multiple multiple","In this scheme, multiple Hough images corresponding to multiple object scales can be obtained simultaneously.",17566
"various a problem, difficult re-rendering notoriously to is This applications. key yet","This is a notoriously difficult problem, yet key to various re-rendering applications.",17567
"solely that illumination trained maps), synthetic spherical data. high-resolution on is","high-resolution spherical illumination maps), that is solely trained on synthetic data.",17568
Here not MCI. shown out but diagnosed subject that Disease selects Alzheimer's it TUT is,Here it is shown that subject TUT selects out diagnosed Alzheimer's Disease but not MCI.,17569
detection consider multivariate of We in problem time-series. the change-point,We consider the problem of change-point detection in multivariate time-series.,17570
and arising from neuroscience. synthetic illustrate both on experimental our We data results biology and,We illustrate our results on both synthetic and experimental data arising from biology and neuroscience.,17571
In truth nonrigidly we paper real-world this present dense a ground dataset scenes. deforming of,In this paper we present a dense ground truth dataset of nonrigidly deforming real-world scenes.,17572
tracked positions metrics. This formation comparison allows and automatically of the RGB for with error,This allows for comparison with automatically tracked RGB positions and the formation of error metrics.,17573
containing sequences deforming datasets synthetic Most nonrigidly previous on based data. are,Most previous datasets containing nonrigidly deforming sequences are based on synthetic data.,17574
"Next, used the to feature sets. classification distinguish ranking feature optimal are strategies","Next, feature ranking strategies are used to distinguish the optimal classification feature sets.",17575
"However, threshold fix issue. this open an is how still to","However, how to fix this threshold is still an open issue.",17576
"we the interested paper, inferring are In video. this fluents in from of vehicles","In this paper, we are interested in inferring the fluents of vehicles from video.",17577
This And-Or car learns paper to a model represent fluents. hierarchical spatial-temporal,This paper learns a spatial-temporal And-Or hierarchical model to represent car fluents.,17578
this latent of learning the SVM is framework. model formulated structural The under,The learning of this model is formulated under the latent structural SVM framework.,17579
entertainments from wide to Face synthesis enforcements. law applications sketch ranging digital has,Face sketch synthesis has wide applications ranging from digital entertainments to law enforcements.,17580
available implementation paper at are and online code in www.ihitworld.com/WNN/IQA_Sketch.zip. this Data,Data and implementation code in this paper are available online at www.ihitworld.com/WNN/IQA_Sketch.zip.,17581
chooses set a source then from set subspaces. Our method restricted of ordered this,Our method then chooses a restricted set from this ordered set of source subspaces.,17582
work of a first This kind is vision-based estimation. friction its,This work is a first of its kind vision-based friction estimation.,17583
for instantaneously. measurements BRDF reflectance develop that a capture disks We novel partial representation,We develop a novel representation for reflectance disks that capture partial BRDF measurements instantaneously.,17584
"alignment object for instance-level Amazon model Challenge, Picking surface mesh correspondence). the and","instance-level object model alignment for the Amazon Picking Challenge, and mesh surface correspondence).",17585
"camera. We of behind the infer the ""invisible egocentric to propose pose"" a person","We propose to infer the ""invisible pose"" of a person behind the egocentric camera.",17586
scheme novel the pose energy to introduce sequence. further a We minimization infer,We further introduce a novel energy minimization scheme to infer the pose sequence.,17587
tested energy method We our provided on by data several usage utilities. U.S. power,We tested our method on energy usage data provided by several U.S. power utilities.,17588
to the then recover image used details. De-convolutional layers are,De-convolutional layers are then used to recover the image details.,17589
Experimental network previously reported achieves all that results better performance our methods. than state-of-the-art show,Experimental results show that our network achieves better performance than all previously reported state-of-the-art methods.,17590
Experiments both conducted semi-real and synthetic are data. for,Experiments are conducted for both synthetic and semi-real data.,17591
essential We configurations. types also models independence with independence/conditional different four of generative define,We also define four essential types of generative models with different independence/conditional independence configurations.,17592
in concerned direction theorems. this Results representation adapting are mainly with,Results in this direction are mainly concerned with adapting representation theorems.,17593
"fragment-driven also leads novel to belief research questions. However, change","However, fragment-driven belief change also leads to novel research questions.",17594
problem a nearest to Finding influential many subspace applications. the and is fundamental,Finding the nearest subspace is a fundamental problem and influential to many applications.,17595
"multiple the existing by methods. Intuitively, represents points a the subspace proposed unlike method","Intuitively, the proposed method represents a subspace by multiple points unlike the existing methods.",17596
a problem ANSS This large-scale tractable. makes,This makes a large-scale ANSS problem tractable.,17597
associate differencing tracks. Most aim into to detections from approaches frame,Most approaches aim to associate detections from frame differencing into tracks.,17598
"missing cause stopped However, in tracking and or discontinuities. result vehicles detections further long-term slow","However, slow or stopped vehicles result in long-term missing detections and further cause tracking discontinuities.",17599
to It takes the targets. context constraints to drifting nearby advantage of avoid,It takes the advantage of context constraints to avoid drifting to nearby targets.,17600
efficiently dense sparse We hypotheses generate to solutions keep tractable. from and flow,We generate hypotheses from sparse and dense flow efficiently to keep solutions tractable.,17601
from merged detections We them. handle additional hypotheses by generating explicitly,We explicitly handle merged detections by generating additional hypotheses from them.,17602
area significant shows evaluation wide improvement Our sequences imagery over on aerial state-of-the-art methods.,Our evaluation on wide area aerial imagery sequences shows significant improvement over state-of-the-art methods.,17603
movement address real-world a problem recognition then classification multi-label our problem. We crowd as,We then address our real-world crowd movement recognition problem as a multi-label classification problem.,17604
work. dataset This for analysis or crowd training a set benchmark as serve can,This dataset can serve as a training set or benchmark for crowd analysis work.,17605
"its of approach, generative A nature. drawback however, is the","A drawback of the approach, however, is its generative nature.",17606
"infinite take use may Functions Conditional Ordinal To such that values. represent problems, we","To represent such problems, we use Ordinal Conditional Functions that may take infinite values.",17607
attacks. of to notion of use enthymemes We and uncertainty this illustrate arguments,We use enthymemes to illustrate this notion of uncertainty of arguments and attacks.,17608
person by generic specific improve model detection models. a active learning further speaker We for,We further improve a generic model for active speaker detection by learning person specific models.,17609
use of training clean lack the temporal continuity of overcomes data. The,The use of temporal continuity overcomes the lack of clean training data.,17610
hierarchy of We so-called amount classes the specifying of required verification introduce a information.,We introduce a hierarchy of so-called verification classes specifying the required amount of information.,17611
class. well-known a semantics that We standard through show exactly certain verifiable are such,We show that well-known standard semantics are exactly verifiable through a certain such class.,17612
"probabilistic simple, to introduced. intuitive inferences is approach the of assessment A","A simple, intuitive approach to the assessment of probabilistic inferences is introduced.",17613
The the are information probability translated Shannon metrics to domain.,The Shannon information metrics are translated to the probability domain.,17614
"for often keypoint On obtain hand, other labels can of the we bags. pairs","On the other hand, we can often obtain labels for pairs of keypoint bags.",17615
"average, matching than non-matching pairs. pairs On corresponding should contain keypoints more","On average, matching pairs should contain more corresponding keypoints than non-matching pairs.",17616
learned and their Features in system. jointly learning an end-to-end are relationships,Features and their relationships are jointly learned in an end-to-end learning system.,17617
framework substantially. learning proposed improves feature The,The proposed framework improves feature learning substantially.,17618
released public. The is the to code,The code is released to the public.,17619
of an object? is What vector representation good a,What is a good vector representation of an object?,17620
this usefulness and Extensive experimental embedding. demonstrates versatility analysis the of,Extensive experimental analysis demonstrates the usefulness and versatility of this embedding.,17621
object-level requires segmentation and data. information both low-level Object pixel,Object segmentation requires both object-level information and low-level pixel data.,17622
object is resulting generating of bottom-up/top-down efficiently masks. capable architecture The high-fidelity,The resulting bottom-up/top-down architecture is capable of efficiently generating high-fidelity object masks.,17623
"Similarly leverages at approach features our to net. skip layers connections, all of the","Similarly to skip connections, our approach leverages features at all layers of the net.",17624
"connections, at does attempt each predictions output Unlike approach our skip not layer. independent to","Unlike skip connections, our approach does not attempt to output independent predictions at each layer.",17625
"effective. is approach The fast, simple, and","The approach is simple, fast, and effective.",17626
"are FCNs proposing this In that instance-level of paper, develop capable we candidates. segment","In this paper, we develop FCNs that are capable of proposing instance-level segment candidates.",17627
results of present on competitive proposal VOC COCO. PASCAL We both MS and segment instance,We present competitive results of instance segment proposal on both PASCAL VOC and MS COCO.,17628
"(HQI). we Iteration off-policy In paper, Hierarchical an HRL Q-value introduce algorithm: this","In this paper, we introduce an off-policy HRL algorithm: Hierarchical Q-value Iteration (HQI).",17629
tabular prove formally MDP. algorithm for We first the the of convergence,We first formally prove the convergence of the algorithm for tabular MDP.,17630
properties. several social with networks unique videos Micro-videos are media on six-second popular,Micro-videos are six-second videos popular on social media networks with several unique properties.,17631
Oxford portable Nanopore by Motivation: sequencing the device is MinION The device. first,Motivation: The MinION device by Oxford Nanopore is the first portable sequencing device.,17632
This may applicability sequencing for applications. various of enhance further genome MinION clinical and advance,This advance may further enhance applicability of MinION for genome sequencing and various clinical applications.,17633
is algorithm features. Deep high-level learning in while learning machine classification extracting learning power,Deep learning is power learning machine learning algorithm in classification while extracting high-level features.,17634
This more to enables expand systems. approach also predict to us methodology complicated our,This approach also enables us to expand our methodology to predict more complicated systems.,17635
representation we problem learning study of In the without image human this paper annotation.,In this paper we study the problem of image representation learning without human annotation.,17636
"we CNN. context-free the a network tasks the To introduce siamese-ennead compatibility across maintain (CFN),","To maintain the compatibility across tasks we introduce the context-free network (CFN), a siamese-ennead CNN.",17637
content. that show the relevant semantically evaluations experimental capture learned Our features,Our experimental evaluations show that the learned features capture semantically relevant content.,17638
"on extraction two significantly other Our results achieves promising competitors. and skeleton method outperforms datasets,","Our method achieves promising results on two skeleton extraction datasets, and significantly outperforms other competitors.",17639
text pipeline a detection. introduce for scene top-down new We,We introduce a new top-down pipeline for scene text detection.,17640
convolutional We rectangle ing fusions. in-network and multiple develop- by convolutions customize network,We customize convolutional network by develop- ing rectangle convolutions and multiple in-network fusions.,17641
efficiently. This it handle enables to and text multi-scale multi-shape,This enables it to handle multi-shape and multi-scale text efficiently.,17642
"networks). curved Man-made exhibit usually descriptive curve objects features (i.e.,","Man-made objects usually exhibit descriptive curved features (i.e., curve networks).",17643
geometric network conveys The its structure. of and an high-level object curve topological,The curve network of an object conveys its high-level geometric and topological structure.,17644
curve point feature present a from unstructured We extracting data. framework cloud networks for,We present a framework for extracting feature curve networks from unstructured point cloud data.,17645
initial of segments curved fitting a highly generates first set Our regions. framework curved,Our framework first generates a set of initial curved segments fitting highly curved regions.,17646
both curved and structural regularities. to segments We data optimize fitting respect then these,We then optimize these curved segments to respect both data fitting and structural regularities.,17647
The by compatibility incorporating bilinear model variables. proposed state-of-the-art augments method latent the,The proposed method augments the state-of-the-art bilinear compatibility model by incorporating latent variables.,17648
the lot of during attention drawn Palmprint recent recognition years. a has,Palmprint recognition has drawn a lot of attention during the recent years.,17649
"called representation, powerful is a this paper, recognition. used scattering network/transform, image In for palmprint","In this paper, a powerful image representation, called scattering network/transform, is used for palmprint recognition.",17650
where network predefined convolutional wavelet its is Scattering filters architecture transforms. network and a are,Scattering network is a convolutional network where its architecture and filters are predefined wavelet transforms.,17651
classifiers different classifier. minimum-distance SVM for Two multi-class used are and recognition:,Two different classifiers are used for recognition: multi-class SVM and minimum-distance classifier.,17652
"that In ABA+ exhibit. semantics we desirable establish properties particular,","In particular, we establish desirable properties that ABA+ semantics exhibit.",17653
of numbers up remains Scaling category recognition to visual challenging. large classes,Scaling up visual category recognition to large numbers of classes remains challenging.,17654
zero-short in state-of-the-art results significantly retrieval. on improve and consistently recognition Our and the,Our results consistently and significantly improve on the state-of-the-art in zero-short recognition and retrieval.,17655
iteration Pearl. discussion on postulates according adaptations and of centers The Darwiche the to classical,The discussion centers on adaptations of the classical iteration postulates according to Darwiche and Pearl.,17656
computer around result final strongly of contours tasks. disturb Local perturbations the vision,Local perturbations around contours strongly disturb the final result of computer vision tasks.,17657
introduce process. a common information to the is priori in estimation It,It is common to introduce a priori information in the estimation process.,17658
be deformable a model such model. as the can Improvement via achieved snake,Improvement can be achieved via a deformable model such as the snake model.,17659
filtered data the noisiest out. guarantees that procedure be will This,This procedure guarantees that the noisiest data will be filtered out.,17660
with We output. use mixture partially-latent tu propose of a linear regressions,We propose tu use a mixture of linear regressions with partially-latent output.,17661
The generating known a a Markov function. with chain of resulting is process branching that,The resulting Markov chain is that of a branching process with a known generating function.,17662
"time modeling Finally, we mechanisms to propose of this perception. framework underlying a application possible","Finally, we propose a possible application of this modeling framework to mechanisms underlying time perception.",17663
"are an with In object concerned interest this article, tracking of in we video stream.","In this article, we are concerned with tracking an object of interest in video stream.",17664
framework. within develop a the Bayesian We modeling algorithm,We develop the algorithm within a Bayesian modeling framework.,17665
in fusing proposed The observations. information for (BMA) strategy averaging multi-clue is Bayesian the model,The Bayesian model averaging (BMA) strategy is proposed for fusing multi-clue information in the observations.,17666
of proposed allowed features are object Any number in to framework. involved be the,Any number of object features are allowed to be involved in the proposed framework.,17667
the performed by particle state The inference employing filter methods. is,The state inference is performed by employing the particle filter methods.,17668
role a in ventral object The recognition. human temporal (VTC) cortex plays critical,The human ventral temporal cortex (VTC) plays a critical role in object recognition.,17669
"to participants associate sessions, Over learned (e.g. features multiple semantic training","Over multiple training sessions, participants learned to associate semantic features (e.g.",17670
"associations made and contextual wood, of (e.g. spatial floats)","made of wood, floats) and spatial contextual associations (e.g.",17671
"that second these In to effects real-life a we translated validated experiment, contextual objects.","In a second experiment, we validated that these contextual effects translated to real-life objects.",17672
evaluate group such A then graphs. collectively people of can,A group of people can then collectively evaluate such graphs.,17673
models formal of These help to real-world bring closer results application. argumentation,These results help bring formal models of argumentation closer to real-world application.,17674
recognition a is Color applications. image step segmentation in many and pattern computer crucial vision,Color image segmentation is a crucial step in many computer vision and pattern recognition applications.,17675
"e.g. The concentrations, metabolite of enzyme complex and interplay","The complex interplay of enzyme and metabolite concentrations, e.g.",17676
"positive and to high have these rates. sensitive false However, thresholds are methods","However, these methods are sensitive to thresholds and have high false positive rates.",17677
a field sensors Imaging evolving. Image and is that is continuously,Imaging and Image sensors is a field that is continuously evolving.,17678
day. into new market coming every products There are the,There are new products coming into the market every day.,17679
conditions Some require to met be these both simultaneously.,Some require both these conditions to be met simultaneously.,17680
analog signal a work solution we propose this processing problem. as In this to,In this work we propose analog signal processing as a solution to this problem.,17681
analog stereo merits correspondence processing to of the show are order In algorithms two implemented.,In order to show the merits of analog processing two stereo correspondence algorithms are implemented.,17682
applied learning. cross-domain knowledge have compression transfer cross-architecture and been distillation and for successfully Model,Model compression and knowledge distillation have been successfully applied for cross-architecture and cross-domain transfer learning.,17683
"However, in is examples key the are that across requirement training domains. a correspondence","However, a key requirement is that training examples are in correspondence across the domains.",17684
"three we in Systematically, facts study CNN transfer.","Systematically, we study three facts in CNN transfer.",17685
solving LCA (NP-complete) in shown potentials great time has also polynomial non-deterministic The problems.,The LCA has also shown great potentials in solving non-deterministic polynomial time (NP-complete) problems.,17686
future as studies. for can a systematic be basis categorization used This,This systematic categorization can be used as a basis for future studies.,17687
settings. We streaming batch both and handle the variants develop to,We develop variants to handle both the batch and streaming settings.,17688
activity during Neural process. Ketamine-Medetomidine induction recorded anesthetic a was,Neural activity was recorded during a Ketamine-Medetomidine anesthetic induction process.,17689
in were estimated Networks serially of seconds. time intervals five,Networks were serially estimated in time intervals of five seconds.,17690
architecture. changes of on These a networks the reveal occurrence transition the,These changes reveal the occurrence of a transition on the networks architecture.,17691
to state. highly observed specific It brain was dynamic shifted a that the also and,It was also observed that the brain shifted to a highly specific and dynamic state.,17692
"what motivations properties, is future There and considerable uncertainty will have. capabilities AGIs about","There is considerable uncertainty about what properties, capabilities and motivations future AGIs will have.",17693
"may In security accidents defects. some plausible and risks scenarios, pose from AGIs arising","In some plausible scenarios, AGIs may pose security risks arising from accidents and defects.",17694
"containers, mechanisms, identify AGI for available need We and that requirements to addressed. weaknesses be","We identify requirements for AGI containers, available mechanisms, and weaknesses that need to be addressed.",17695
mining pattern generally flexibility. lack Traditional from a algorithms suffer of,Traditional pattern mining algorithms generally suffer from a lack of flexibility.,17696
paper This depth water estimation introduces from drops.,This paper introduces depth estimation from water drops.,17697
commonly adherent drops. a any water total in observed band dark generates reflection This,This total reflection generates a dark band commonly observed in any adherent water drops.,17698
"using depth can compute stereo. we Subsequently,","Subsequently, we can compute depth using stereo.",17699
"we addition estimation, In refocusing. can depth apply also to image","In addition to depth estimation, we can also apply image refocusing.",17700
evaluation our show a Experiments the on of and quantitative effectiveness method. proposed real images,Experiments on real images and a quantitative evaluation show the effectiveness of our proposed method.,17701
"have drops knowledge, estimate To used to never been our depth. adherent best before water","To our best knowledge, never before have adherent water drops been used to estimate depth.",17702
representations. descriptors our tensor The higher-order outer-products of form these kernel,The higher-order outer-products of these kernel descriptors form our tensor representations.,17703
train from an to used are then kernels Tensors SVM. formed these,Tensors formed from these kernels are then used to train an SVM.,17704
"In conceived, fabricated. then shapes first design, and traditional are","In traditional design, shapes are first conceived, and then fabricated.",17705
Extremes Anomaly Detection. a special role play in,Extremes play a special role in Anomaly Detection.,17706
Illustrative of relevance strong results empirical provide the experimental our evidence of approach.,Illustrative experimental results provide strong empirical evidence of the relevance of our approach.,17707
evaluate graph. be argumentation There to way can than such one consistent more an,There can be more than one consistent way to evaluate such an argumentation graph.,17708
preferences to ought This individual be. is have because the agents what outcome about their,This is because agents have their individual preferences about what the outcome ought to be.,17709
relationship the social into choice theory. investigation motivate argumentation between results further and Our,Our results motivate further investigation into the relationship between social choice and argumentation theory.,17710
sphere's perpendicular axis optical camera's the to lies surface. The,The camera's optical axis lies perpendicular to the sphere's surface.,17711
"the is this rotation camera's parameters. case, three by pose minimally represented In","In this case, the camera's pose is minimally represented by three rotation parameters.",17712
"models a toolbox. used, neuron fitting are published point leveraging integrate-and-fire generalized recently Here,","Here, generalized integrate-and-fire point neuron models are used, leveraging a recently published fitting toolbox.",17713
point-neuron applicable it other The also synapse makes for of modularity models. and approach the,The modularity of the approach makes it applicable also for other point-neuron and synapse models.,17714
"is this, to of to In the instantaneous. all motion detect nearly ability addition","In addition to all of this, the ability to detect motion is nearly instantaneous.",17715
Such on information the provides intra-attribute important cues variations.,Such information provides important cues on the intra-attribute variations.,17716
state-of-the-art. with experiments The over our of the model effectiveness demonstrate improvement significant,The experiments demonstrate the effectiveness of our model with significant improvement over state-of-the-art.,17717
Labeling global (GAL) Assisted and exploits The features proposed local system Global-attributes both attributes.,The proposed Global-attributes Assisted Labeling (GAL) system exploits both local features and global attributes.,17718
"lines, They ground etc. include vanishing lines, lines, sky","They include sky lines, ground lines, vanishing lines, etc.",17719
utilized paper. quality of assessment method Reference) in the is (Full FR this image,FR (Full Reference) method of the image quality assessment is utilized in this paper.,17720
object understanding content image critical segmentation Semantic is and localization. to,Semantic segmentation is critical to image content understanding and object localization.,17721
Recent development fully-convolutional pixel-level neural network has in labeling. enabled accurate (FCN),Recent development in fully-convolutional neural network (FCN) has enabled accurate pixel-level labeling.,17722
in re-identification critical surveillance is applications. Person,Person re-identification is critical in surveillance applications.,17723
paper In to feature this this a dynamics-based overcome limitation. propose we,In this paper we propose a dynamics-based feature to overcome this limitation.,17724
"datasets. merits new the the ""appearance-impaired"" we introduce illustrate of features three proposed To","To illustrate the merits of the proposed features we introduce three new ""appearance-impaired"" datasets.",17725
achieves and performance simple the on a scheme state-of-the-art Such benchmarks. several effective yet,Such a simple and yet effective scheme achieves the state-of-the-art performance on several benchmarks.,17726
"more on state-of-the-art images shows outperforms capability methods. Surprisingly, and noisy superior it","Surprisingly, it shows more superior capability on noisy images and outperforms state-of-the-art methods.",17727
"extraction a to major feature In role. order information, plays obtain these","In order to obtain these information, feature extraction plays a major role.",17728
made An all obtaining has attempt been (i.e. in,An attempt has been made in obtaining all (i.e.,17729
"go. and test age gender, range, expression information in a single a from ethnicity) image","gender, age range, expression and ethnicity) information from a test image in a single go.",17730
We introduce first the method radiometric decomposition scene handles intricacies. for those that,We introduce the first method for radiometric scene decomposition that handles those intricacies.,17731
back-propagation whole are criterion. networks using The the updated the parameters by of,The parameters of the whole networks are updated by using the back-propagation criterion.,17732
an problem. the to algorithm resulting We efficiently also solve develop optimization,We also develop an optimization algorithm to efficiently solve the resulting problem.,17733
to complexity the techniques hierarchical two of We topologies. probe introduce,We introduce two techniques to probe the hierarchical complexity of topologies.,17734
"measure a Firstly, model. hierarchical Hierarchy Weighted secondly, complexity; metric new to (WCH) a Complex","Firstly, a new metric to measure hierarchical complexity; secondly, a Weighted Complex Hierarchy (WCH) model.",17735
"Further, networks to EEG at has phase-lag complexity peak performance. the model equivalent","Further, the model has equivalent complexity to EEG phase-lag networks at peak performance.",17736
model of hierarchical rigorous a complexity. Our provide and metric characterisation,Our metric and model provide a rigorous characterisation of hierarchical complexity.,17737
MNA aligned aim (node Existing over total maximize similarity nodes to methods all conservation).,Existing MNA methods aim to maximize total similarity over all aligned nodes (node conservation).,17738
"called achieve we can novel a that present approach multiMAGNA++ Thus, MNA this.","Thus, we present a novel MNA approach called multiMAGNA++ that can achieve this.",17739
"multiMAGNA++ to effectively. and data scales network be well parallelized larger can That is,","That is, multiMAGNA++ scales well to larger network data and can be parallelized effectively.",17740
collect diverse sufficiently lives? boring how So samples do but many representing our we,So how do we collect sufficiently many diverse but boring samples representing our lives?,17741
propose We such a novel approach in collect Homes to data. Hollywood,We propose a novel Hollywood in Homes approach to collect such data.,17742
"deep enormously detection of the from in benefited large-scale especially has datasets, learning. context Object","Object detection has benefited enormously from large-scale datasets, especially in the context of deep learning.",17743
reproducibility. models make We network and dataset our aggregated multi-domain available for,We make available our network models and aggregated multi-domain dataset for reproducibility.,17744
several proposed multiple interaction video on datasets. model pedestrian evaluate public our We,We evaluate our proposed model on several public multiple pedestrian interaction video datasets.,17745
"to biological transcription, sciences, transcription. RNA Till term, DNA mainly now, in refers the to","Till now, in biological sciences, the term, transcription, mainly refers to DNA to RNA transcription.",17746
are Crucially types movement both cortical driven same from the site.,Crucially both movement types are driven from the same cortical site.,17747
novel retrieval. propose a image for approach instance-level We,We propose a novel approach for instance-level image retrieval.,17748
training using that the our to approach. key of We is clean success data show,We show that using clean training data is key to the success of our approach.,17749
architecture pass. global representation a The a image proposed forward in single produces,The proposed architecture produces a global image representation in a single forward pass.,17750
approach global on based significantly outperforms descriptors previous standard datasets. approaches Our on,Our approach significantly outperforms previous approaches based on global descriptors on standard datasets.,17751
costly most works verification. and local prior indexing based descriptor on surpasses spatial even It,It even surpasses most prior works based on costly local descriptor indexing and spatial verification.,17752
traffic instance-level using participants the We estimate propose to segmentation.,We propose to estimate the traffic participants using instance-level segmentation.,17753
flow of estimation This core our pipeline. a is element,This is a core element of our flow estimation pipeline.,17754
than longer NA condense also predicts shorter model readily that ones. The will more fragments,The model also predicts that longer NA fragments will condense more readily than shorter ones.,17755
approach addressed via this \emph{hybrid} learning are and a termed deep representations in Such paper.,Such representations are termed \emph{hybrid} and addressed via a deep learning approach in this paper.,17756
"In recognition. past, recognized been attributes separately have however, the material preceding category","In the past, however, material attributes have been recognized separately preceding category recognition.",17757
hold material recognition? attribute and for Does same the true category,Does the same hold true for material attribute and category recognition?,17758
"interest to representations gained have image increasing Meanwhile, hierarchical describe content.","Meanwhile, hierarchical representations have gained increasing interest to describe image content.",17759
"However, unordered the features. equipped nodes are with are trees and numerical the induced","However, the induced trees are unordered and the nodes are equipped with numerical features.",17760
study the problem videos. We analysis of facial in,We study the problem of facial analysis in videos.,17761
"a automatically (eg. mined, of discriminative as sequence sub-events","as a sequence of automatically mined, discriminative sub-events (eg.",17762
"for smile, lower brow for and phase raise pain). and cheek onset offset","onset and offset phase for smile, brow lower and cheek raise for pain).",17763
"state-of-the-art complimentary In these we on with combination datasets. results features, report","In combination with complimentary features, we report state-of-the-art results on these datasets.",17764
is of existence detection investigated. anomalous Nonparametric of structure a network over an,Nonparametric detection of existence of an anomalous structure over a network is investigated.,17765
network is Detection studied. interval line anomalous of over a first an,Detection of an anomalous interval over a line network is first studied.,17766
results to developed. Generalization other further networks the of is,Generalization of the results to other networks is further developed.,17767
to Numerical are provided of demonstrate proposed performance the the results tests.,Numerical results are provided to demonstrate the performance of the proposed tests.,17768
"inexpensive stores. products, is This of handling e.g., usual retailers large grocery for varieties","This is usual for retailers handling large varieties of inexpensive products, e.g., grocery stores.",17769
is to and base throughout Knowledge unstructured store data the structured way web. the,Knowledge base is the way to store structured and unstructured data throughout the web.,17770
can significantly. the semantic automatically to Therefore web society ontology contribute learning,Therefore learning ontology automatically can contribute to the semantic web society significantly.,17771
based property is algorithm on which approach data-driven Our of the instance. each is,Our algorithm is data-driven approach which is based on the property of each instance.,17772
by class its is instances. Ontology of elaborated property frequent being generalizing,Ontology class is being elaborated by generalizing frequent property of its instances.,17773
"class information, instance each matching By can its find that developed most using relatively class.","By using that developed class information, each instance can find its most relatively matching class.",17774
"two basic achieve these we from fully-automated steps, evolution ontology incomplete repeating base. knowledge By","By repeating these two steps, we achieve fully-automated ontology evolution from incomplete basic knowledge base.",17775
"sees. we In video, a camera see wearable what the wearer camera","In a wearable camera video, we see what the camera wearer sees.",17776
approach array outperforms wide Our a existing of methods.,Our approach outperforms a wide array of existing methods.,17777
the go of top-ranked most winnings to the entries.,most of the winnings go to the top-ranked entries.,17778
find We practice. approach well that in our performs,We find that our approach performs well in practice.,17779
"expression, illumination, as such has itself occlusion pose, etc. limitations recognition Face","Face recognition itself has limitations such as illumination, pose, expression, occlusion etc.",17780
"periocular has spectacles, as biometrics head hair limitations. expression and Also, angle, its","Also, periocular biometrics has spectacles, head angle, hair and expression as its limitations.",17781
limitations. biometrics Unimodal surmount these cannot all,Unimodal biometrics cannot surmount all these limitations.,17782
is of cancer pathogenesis in poorly human The understood. still,The pathogenesis of cancer in human is still poorly understood.,17783
data to opportunities those great Deciphering challenges biologists. and computational poses,Deciphering those data poses great opportunities and challenges to computational biologists.,17784
routinely from Carlo sampling used Hamiltonian for (HMC) become posterior Monte has distributions.,Hamiltonian Monte Carlo (HMC) has become routinely used for sampling from posterior distributions.,17785
a establish behind and motivate theoretical properties. GTHMC We characterize its theory RMHMC to geometric,We establish a geometric theory behind RMHMC to motivate GTHMC and characterize its theoretical properties.,17786
training through accomplished model is Expectation-Maximization an The method. entire,The entire model training is accomplished through an Expectation-Maximization method.,17787
applications. all seamlessly these blended are of processed image The the for regions,The processed regions of the image are blended seamlessly for all these applications.,17788
problem. study sparse squares We the (S-NNLS) least non-negative,We study the sparse non-negative least squares (S-NNLS) problem.,17789
"we approaches and the the Finally, of data. performance study proposed real-world on synthetic","Finally, we study the performance of the proposed approaches on synthetic and real-world data.",17790
metagenomic projects different biodiversity scale between environmental Large conditions. knowledge extract aim to,Large scale metagenomic projects aim to extract biodiversity knowledge between different environmental conditions.,17791
comparing for limitations. important communities methods face Current microbial,Current methods for comparing microbial communities face important limitations.,17792
"novo a de comparative development method, the These Simka. metagenomic new of limitations called motivated","These limitations motivated the development of a new de novo metagenomic comparative method, called Simka.",17793
a We neural superpixel-based semantic multi-view propose segmentation. network convolutional for novel image,We propose a novel superpixel-based multi-view convolutional neural network for semantic image segmentation.,17794
Migration detected as a as week. be soon can in,Migration can be detected as soon as in a week.,17795
"week is of It statistical for purpose. faster, setting be the a could","It could be faster, the setting of a week is for statistical purpose.",17796
model sample regression. few with quantitative done exact would and research More be research also,More exact quantitative research would also be done with few sample research and model regression.,17797
"a image. location estimate been As such, its have methods proposed from single many to","As such, many methods have been proposed to estimate its location from a single image.",17798
along a road. work is this Our study systematic,Our work is a systematic study along this road.,17799
uses long error correction present only. reads method an that We,We present an error correction method that uses long reads only.,17800
view. their to videos person Egocentric have ability characterised first by the are,Egocentric videos are characterised by their ability to have the first person view.,17801
egocentric action the is wearer videos important an of from problem. Recognizing,Recognizing action of the wearer from egocentric videos is an important problem.,17802
representation We feature derived first from trajectories. of the a actions person novel propose,We propose a novel representation of the first person actions derived from feature trajectories.,17803
the (CST). for network We threshold a Threshold named analysis new Cluster-Span unbiased propose,We propose a new unbiased threshold for network analysis named the Cluster-Span Threshold (CST).,17804
"does perspective. I systems dynamical As take a Boutilier,","As does Boutilier, I take a dynamical systems perspective.",17805
the hour. the is of need Image recognition,Image recognition is the need of the hour.,17806
large The world. direct of applications real has the in work number a,The work has a large number of direct applications in the real world.,17807
validate results were the MATLAB The proposed tool simulated encouraging using algorithm. are and,The results were simulated using MATLAB tool are encouraging and validate the proposed algorithm.,17808
"Calculation, Detection, . Terms: Shape Area Edge Detection, Object Index Recognition","Index Terms: Edge Detection, Area Calculation, Shape Detection, Object Recognition .",17809
"Wild We (FIW). date, present largest in the the kinship to recognition Families dataset","We present the largest kinship recognition dataset to date, Families in the Wild (FIW).",17810
"image-based several include two family We recognition. benchmarks for verification tasks, kinship and","We include several benchmarks for two image-based tasks, kinship verification and family recognition.",17811
"baselines. and For learning features several visual metric methods this, incorporate we as","For this, we incorporate several visual features and metric learning methods as baselines.",17812
object COCO object challenges The dataset new recent detection several presents for detection.,The recent COCO object detection dataset presents several new challenges for object detection.,17813
"We refer network. modified a classifier to our ""MultiPath"" as","We refer to our modified classifier as a ""MultiPath"" network.",17814
by to Cell neuroengineering. neuroscience advance and pivotal assemblies optogenetics manipulation is,Cell assemblies manipulation by optogenetics is pivotal to advance neuroscience and neuroengineering.,17815
"interpret. to former of downstream, straightforward direct activation and targets from The are arise","The former arise from direct activation of targets downstream, and are straightforward to interpret.",17816
an computer Scene applications. role in text recognition plays many vision important,Scene text recognition plays an important role in many computer vision applications.,17817
"paper, In this propose algorithm. recognition CNN we a Chinese text based","In this paper, we propose a CNN based Chinese text recognition algorithm.",17818
Chinese datasets. The proposed is evaluated text text with two algorithm recognition Chinese,The proposed Chinese text recognition algorithm is evaluated with two Chinese text datasets.,17819
The compared to the baseline methods. accuracy algorithm recognize better achieves,The algorithm achieves better recognize accuracy compared to the baseline methods.,17820
"paper parsing, of i.e. the addresses geometric scene This problem","This paper addresses the problem of geometric scene parsing, i.e.",17821
"and sky, relations interaction vertical determining and the ground plane) (e.g.","sky, ground and vertical plane) and determining the interaction relations (e.g.",17822
"siding regions. main affinity) supporting, and between layering,","layering, supporting, siding and affinity) between main regions.",17823
Neural computer (CNNs) in performance achieve Convolutional Networks state-of-the-art vision many tasks.,Convolutional Neural Networks (CNNs) achieve state-of-the-art performance in many computer vision tasks.,17824
step in of probabilistic The fundamental estimation normalizing a model comparison. constants is,The estimation of normalizing constants is a fundamental step in probabilistic model comparison.,17825
Annealed further over A to advantage is ARM Sampling tune. easier Importance that is,A further advantage over Annealed Importance Sampling is that ARM is easier to tune.,17826
of operation. Lewis Imaging use as the revision version We a,We use a version of Lewis Imaging as the revision operation.,17827
of The is proved. correctness the approach,The correctness of the approach is proved.,17828
"some applications. under belief of restricted, rather consideration has but The the are expressivity bases","The expressivity of the belief bases under consideration are rather restricted, but has some applications.",17829
addresses segmentation their shortcomings. reviews in manuscript recent the cross-country scene achievements This and,This manuscript reviews the recent achievements in cross-country scene segmentation and addresses their shortcomings.,17830
related a classification of features. describes to then texture high dimensional problem It,It then describes a problem related to classification of high dimensional texture features.,17831
"resolving learning machine algorithms aimed compares Finally, at three it this problem.","Finally, it compares three machine learning algorithms aimed at resolving this problem.",17832
"flow Optical an ""data optional and minimizing estimated a by is typically regularizer. cost""","Optical flow is typically estimated by minimizing a ""data cost"" and an optional regularizer.",17833
the to clearly most hand-object with deal is The multiple having case task challenging tracks.,The hand-object case is clearly the most challenging task having to deal with multiple tracks.,17834
this are dataset. Extensive presented on results benchmarking,Extensive benchmarking results are presented on this dataset.,17835
distribution image methods posterior the estimate of transformation. Probabilistic registration,Probabilistic image registration methods estimate the posterior distribution of transformation.,17836
"never this justified. convention acceptance, wide Despite has the been","Despite the wide acceptance, this convention has never been justified.",17837
the . of fields examples ensemble by potential demonstrate pilot We,We demonstrate the potential of ensemble fields by pilot examples .,17838
regular take structures uniform that Objects dimensions. or are,Objects or structures that are regular take uniform dimensions.,17839
not or are models regular. processes all But,But not all processes or models are regular.,17840
OTU binning. clustering also referred is to as,OTU clustering is also referred to as binning.,17841
performance COCACOLA The aspects. on relies superior of two,The superior performance of COCACOLA relies on two aspects.,17842
"clustering and regularization. sparsity by COCACOLA clustering soft advantage importantly, hard More takes of both","More importantly, COCACOLA takes advantage of both hard clustering and soft clustering by sparsity regularization.",17843
"addition, binning In accuracy. seamlessly knowledge facilitate customized to framework COCACOLA the embraces","In addition, the COCACOLA framework seamlessly embraces customized knowledge to facilitate binning accuracy.",17844
"and faster MetaBAT. CONCOCT than scalable is MaxBin COCACOLA and ,GroopM,","COCACOLA is scalable and faster than CONCOCT ,GroopM, MaxBin and MetaBAT.",17845
to hierarchical structured from of follows the finger The regression topology wrist tree hand tips.,The hierarchical regression follows the tree structured topology of hand from wrist to finger tips.,17846
to a surface also approximation efficient estimate new We introduce normals.,We also introduce a new efficient approximation to estimate surface normals.,17847
networks that theoretical communities. are abstractions Ecological represent ecological,Ecological networks are theoretical abstractions that represent ecological communities.,17848
"large we evaluate multiple interactions of Here, networks. spatiotemporal consumer-resource variability of marine the","Here, we evaluate the spatiotemporal variability of multiple consumer-resource interactions of large marine networks.",17849
produced Reads of repertoires to moderate cover MiSeq are size. able by,Reads produced by MiSeq are able to cover repertoires of moderate size.,17850
"of from machines to throughput increases year. sequencing the year At time, same","At the same time, throughput of sequencing machines increases from year to year.",17851
of repertoires ultra of diversity. scanning their and deep will adaptive enable This immune analysis,This will enable ultra deep scanning of adaptive immune repertoires and analysis of their diversity.,17852
efficient data highly repertoire construction requires Such accurate and tools. both,Such data requires both efficient and highly accurate repertoire construction tools.,17853
"of immunoproteogenomics and analysis. accurate for presented a tool repertoire antibody construction IgRepertoireConstructor,","presented IgRepertoireConstructor, a tool for accurate construction of antibody repertoire and immunoproteogenomics analysis.",17854
IgReC experiments Results that field. the computational in state-of-the-art improves show of,Results of computational experiments show that IgReC improves state-of-the-art in the field.,17855
source available Availability: running an and on platforms. IgReC Linux open is program freely,Availability: IgReC is an open source and freely available program running on Linux platforms.,17856
at code is yana-safonova.github.io/ig_repertoire_constructor. available The GitHub: source,The source code is available at GitHub: yana-safonova.github.io/ig_repertoire_constructor.,17857
substructure mixture in identifying effectiveness We in data. real CNV demonstrate tumor their further,We further demonstrate their effectiveness in identifying mixture substructure in real tumor CNV data.,17858
"However, MSTs functional network the that for analyses outperforms of detection differences. find threshold we","However, we find that the threshold analyses outperforms MSTs for detection of functional network differences.",17859
"fail find any MSTs significant Particularly, differences. to","Particularly, MSTs fail to find any significant differences.",17860
object significantly two past The has for state-of-the-art improved the detection been years. over performance,The state-of-the-art performance for object detection has been significantly improved over the past two years.,17861
information of utilized. investigated fully contextual are not Temporal and videos and,Temporal and contextual information of videos are not fully investigated and utilized.,17862
traditional extend methods in In direction. algebraic work we this this specification,In this work we extend traditional algebraic specification methods in this direction.,17863
from such that interface integration. we Here adequate an emerges AI for assume Neural-Symbolic,Here we assume that such interface for AI emerges from an adequate Neural-Symbolic integration.,17864
on We extraction knowledge from specified method this of tested the structures.,We tested this method on the extraction of knowledge from specified structures.,17865
"gap, framework deep propose we search. person for To learning new close the a","To close the gap, we propose a new deep learning framework for person search.",17866
"our approach, and validate benchmark a for annotate dataset we search. large-scale collect person To","To validate our approach, we collect and annotate a large-scale benchmark dataset for person search.",17867
"analysing captured of majority spectrum. in visible on focus images Nevertheless, the methods","Nevertheless, the majority of methods focus on analysing images captured in visible spectrum.",17868
salient purpose For this we employing proposed features.,For this purpose we proposed employing salient features.,17869
"scale, view and variations to are robust Salient angle. in features brightness","Salient features are robust to variations in scale, brightness and view angle.",17870
SURF two comparison of a also We implementations. provide,We also provide a comparison of two SURF implementations.,17871
extracted of SURF terrain The features types. from are different images,The SURF features are extracted from images of different terrain types.,17872
estimate terrain we feature class For every function. a membership,For every feature we estimate a terrain class membership function.,17873
nearest obtained or either perceptron means by multi-layer the membership values of neighbours. are The,The membership values are obtained by means of either the multi-layer perceptron or nearest neighbours.,17874
"interest. Advanced Assist broad have vision-based gained Driver Recently, Systems","Recently, vision-based Advanced Driver Assist Systems have gained broad interest.",17875
"online offline. self-supervised FCN the of instead of facilitates Additionally, training training our","Additionally, our self-supervised training facilitates online training of the FCN instead of offline.",17876
"improves, ratio will as worse overall gating This only segmentation get possible speed-up.","This ratio will only get worse as segmentation improves, gating overall possible speed-up.",17877
We introduce meaningful algorithm metrics evaluate size. our on further novel of to connectomes,We further introduce novel metrics to evaluate our algorithm on connectomes of meaningful size.,17878
is the hardware-oriented in of approximation. development steps model One accelerator important most,One of the most important steps in accelerator development is hardware-oriented model approximation.,17879
point. using arithmetic condense fixed point can models instead of and by representation floating Ristretto,Ristretto can condense models by using fixed point arithmetic and representation instead of floating point.,17880
"the fixed fine-tunes resulting Ristretto Moreover, point network.","Moreover, Ristretto fine-tunes the resulting fixed point network.",17881
optimization a optical approach estimation. present flow to We global,We present a global optimization approach to optical flow estimation.,17882
model and this paper segmentation pixel image is In clustering hierarchical a for developed.,In this paper a hierarchical model for pixel clustering and image segmentation is developed.,17883
image model hierarchically an structured. In is the,In the model an image is hierarchically structured.,17884
The of is by clustering clustering optimal hierarchical simulating none-hierarchical studied. pixel,The simulating of none-hierarchical optimal pixel clustering by hierarchical clustering is studied.,17885
surreptitious societal video is important recordings protection Privacy an challenge. from,Privacy protection from surreptitious video recordings is an important societal challenge.,17886
"However, prevents monitoring. real-time usage its complexity the video in computation","However, the computation complexity prevents its usage in real-time video monitoring.",17887
We an present for recognition. end-to-end model attention-based handwriting,We present an attention-based model for end-to-end handwriting recognition.,17888
the does input segmentation require not Our system of paragraph. any,Our system does not require any segmentation of the input paragraph.,17889
"main multi-dimensional and a difference covert overt LSTM The the is implemented as network. attention,","The main difference is the covert and overt attention, implemented as a multi-dimensional LSTM network.",17890
carried the Database. on well-known We out experiments IAM,We carried out experiments on the well-known IAM Database.,17891
"analogue In based a on new quantum is this introduced. post blossom calculus of paper,","In this paper, a new analogue of blossom based on post quantum calculus is introduced.",17892
fundamental in detection image Object is problem a understanding.,Object detection is a fundamental problem in image understanding.,17893
and versions. framework One popular solution is the R-CNN its fast,One popular solution is the R-CNN framework and its fast versions.,17894
"Therefore, parameters optimal nearly shape. more it corresponding contains tailored the to","Therefore, it contains more nearly optimal parameters tailored to the corresponding shape.",17895
voice algorithm propose activity an detection for on signal. based We audio-visual the,We propose an algorithm for voice activity detection based on the audio-visual signal.,17896
voice outperforms that activity Simulation proposed detection fusion the approaches. and competing results show algorithm,Simulation results show that the proposed algorithm outperforms competing fusion and voice activity detection approaches.,17897
object like tasks detection. be estimated can Counts from of vision also outputs other,Counts can also be estimated from outputs of other vision tasks like object detection.,17898
"study object Subsequently, used detection. counting can be how improve to we","Subsequently, we study how counting can be used to improve object detection.",17899
and questions VQA in datasets. the COCO-QA,questions in the VQA and COCO-QA datasets.,17900
are provided development. of Examples illustrate our theoretical to the utility,Examples are provided to illustrate the utility of our theoretical development.,17901
"Instead, deformation predefined basic (e.g. achieved is a rules set by of","Instead, deformation is achieved by a set of predefined basic rules (e.g.",17902
cannot patch). sub-patch across the right left pass the,the left sub-patch cannot pass across the right patch).,17903
that Recent deep show studies achieve two learning can approaches performance these impressive on tasks.,Recent studies show that deep learning approaches can achieve impressive performance on these two tasks.,17904
multi-target present online neural on to recurrent approach a based novel networks (RNNs). tracking We,We present a novel approach to online multi-target tracking based on recurrent neural networks (RNNs).,17905
Most methods models tuning parameters. that involve of previous complex tedious require,Most previous methods involve complex models that require tedious tuning of parameters.,17906
"approach propose for multi-target we learning the online for an time, first Here, end-to-end tracking.","Here, we propose for the first time, an end-to-end learning approach for online multi-target tracking.",17907
principled points solution of in the addresses a all above way. Our,Our solution addresses all of the above points in a principled way.,17908
"the (F,B) matte selected the pairs. on depends of The quality","The quality of the matte depends on the selected (F,B) pairs.",17909
"Here, frames. from is by samples the dictionary formed multiple","Here, the dictionary is formed by samples from multiple frames.",17910
We the the constraint demonstrate non-convex using of benefits experimentally.,We demonstrate the benefits of using the non-convex constraint experimentally.,17911
"objects, are directly treated general-purpose thin reconstructs our Moreover, usually specialized algorithms. with formulation which","Moreover, our general-purpose formulation directly reconstructs thin objects, which are usually treated with specialized algorithms.",17912
"empirical traits, different validation facial multiple of on We e.g. face datasets show image","We show empirical validation on multiple face image datasets of different facial traits, e.g.",17913
codes final classification for We for combine retrieval. with the global barcodes Radon neural,We combine neural codes for global classification with Radon barcodes for the final retrieval.,17914
We based image matching examine regions of (ROI) image search retrieval. on also after interest,We also examine image search based on regions of interest (ROI) matching after image retrieval.,17915
our approach show Experimental works. results superior is many to that published,Experimental results show that our approach is superior to many published works.,17916
"for this we In CNNs detection. propose paper, object subcategory-aware","In this paper, we propose subcategory-aware CNNs for object detection.",17917
a health Zika epidemic The ongoing major poses public global emergency. (ZIKV) virus,The ongoing Zika virus (ZIKV) epidemic poses a major global public health emergency.,17918
infection ZIKV and control not How spread affects transmission sexual well-understood. the of is,How sexual transmission affects the spread and control of ZIKV infection is not well-understood.,17919
"to artistic realistic noise brown Such can directly decode paintings. texture, or to network photos","Such network can directly decode brown noise to realistic texture, or photos to artistic paintings.",17920
"to neural obtain training, recent comparable adversarial quality With we texture methods. synthesis","With adversarial training, we obtain quality comparable to recent neural texture synthesis methods.",17921
"this We synthesis, and stylization. to apply transfer, idea texture style video","We apply this idea to texture synthesis, style transfer, and video stylization.",17922
"approach in a detec- tion natural for images. paper, propose novel we text this In","In this paper, we propose a novel approach for text detec- tion in natural images.",17923
"are estimated by line hypotheses combining character the and Then, text salient map components.","Then, text line hypotheses are estimated by combining the salient map and character components.",17924
"framework general fonts. is and ori- languages text The in entations, handling for multiple","The framework is general for handling text in multiple ori- entations, languages and fonts.",17925
Many Web querying Knowledge (KGs). Graphs efficient applications require large of,Many Web applications require efficient querying of large Knowledge Graphs (KGs).,17926
encoded into terms data locality. semantically grouped ontological to Infrequent classes increase are and,Infrequent terms are semantically grouped into ontological classes and encoded to increase data locality.,17927
"noise in Frames of for used processing basis non-orthogonal traditionally signal removal. are functions, sets","Frames are sets of non-orthogonal basis functions, traditionally used in signal processing for noise removal.",17928
"true other comparable It of k-mers, recall has error the to correction methods. high","It has high recall of the true k-mers, comparable to other error correction methods.",17929
relations Their VID thoroughly in are studied contributions task evaluated. the and and,Their relations and contributions in the VID task are thoroughly studied and evaluated.,17930
focused is systems. a and of AI on subfield studying Computational creative creativity developing,Computational creativity is a subfield of AI focused on developing and studying creative systems.,17931
"this notion. we paper General extends traditional Creativity, introduce that which In","In this paper we introduce General Creativity, which extends that traditional notion.",17932
rely approaches in assumptions matrix different structure the on on hand. Different,Different approaches rely on different structure assumptions on the matrix in hand.,17933
compute of We an a variational algorithm pseudo-posterior. the approximation to propose,We propose an algorithm to compute a variational approximation of the pseudo-posterior.,17934
performance this PAC-Bayesian approximation of bounds. learning the We also variational study through,We also study the performance of this variational approximation through PAC-Bayesian learning bounds.,17935
"the However, a loss. with approximation variational to discuss deal also logistic we","However, we also discuss a variational approximation to deal with the logistic loss.",17936
popular has become representation. image Fisher a vector (FV),Fisher vector (FV) has become a popular image representation.,17937
"in practice. is decorrelation supported local a As usually assumption result, the not","As a result, the local decorrelation assumption is usually not supported in practice.",17938
CFV discriminative power. leads to thus The improved,The CFV thus leads to improved discriminative power.,17939
"used generate Radon images. to has Recently, tagging for been barcodes medical transformation","Recently, Radon transformation has been used to generate barcodes for tagging medical images.",17940
"proposed such Smoothed To new we Dirichlet a Process (sHDP). overcome shortcomings, Hierarchical","To overcome such shortcomings, we proposed a new Smoothed Hierarchical Dirichlet Process (sHDP).",17941
elaborate algorithm and the We present inference solutions. on its,We present the inference algorithm and elaborate on its solutions.,17942
shown has of desirable Our using model. keywords the the effect experiment NIPS,Our experiment using NIPS keywords has shown the desirable effect of the model.,17943
carefully end. considered design factors this are to A few,A few design factors are carefully considered to this end.,17944
"we (ii) online method bootstrapping an propose for training. Second,","(ii) Second, we propose an online bootstrapping method for training.",17945
online We accuracy. critically for important is achieving bootstrapping demonstrate good that,We demonstrate that online bootstrapping is critically important for achieving good accuracy.,17946
"an additive noise. $X_i$, random points design superposed with","random design points $X_i$, superposed with an additive noise.",17947
the can and very be is The of unknown design distribution general. points,The distribution of the design points is unknown and can be very general.,17948
improves results This on completes obtained in settings. or related previous,This improves on or completes previous results obtained in related settings.,17949
main-effect results present several on designs. In the factorial we of this simulations work,In this work we present the results of several simulations on main-effect factorial designs.,17950
"decades, performed For research more image on retrieval (CBIR). has been content-based two than","For more than two decades, research has been performed on content-based image retrieval (CBIR).",17951
is further and memory approach fast low requirement. The has very,The approach further is very fast and has low memory requirement.,17952
differ There receptor NMDA different conductance. single-channel their are subtypes in which,There are different NMDA receptor subtypes which differ in their single-channel conductance.,17953
low-conductance Both to high-conductance as cell were contain found as well receptors. types NMDA,Both cell types were found to contain high-conductance as well as low-conductance NMDA receptors.,17954
art. the more current is Even problematic irreproducibility of,Even more problematic is the irreproducibility of current art.,17955
reproducibility initiate to our community also a We improve . dialog in,We also initiate a dialog to improve reproducibility in our community .,17956
"interaction, To learning this framework DARI, an called i.e. their explore work proposes end-to-end","To explore their interaction, this work proposes an end-to-end learning framework called DARI, i.e.",17957
deep objective convolutional by We networks. architecture neural of a solve this building,We solve this objective by building a deep architecture of convolutional neural networks.,17958
still vision. an classification computer images in is task in important Action,Action classification in still images is an important task in computer vision.,17959
be and information time of scale context difficult labeling up. Manually to would consuming,Manually labeling of context information would be time consuming and difficult to scale up.,17960
Paris visit does typical a like? to look What,What does a typical visit to Paris look like?,17961
Tower? then first Louvre photos the of Do the people take Eiffel and,Do people first take photos of the Louvre and then the Eiffel Tower?,17962
"we frameworks? temporal a event visually ""Paris Can model like using current Vacation""","Can we visually model a temporal event like ""Paris Vacation"" using current frameworks?",17963
"of anatomy-aware we medical measurement scheme segmentation this for In paper, an propose accuracy images.","In this paper, we propose an anatomy-aware measurement scheme for segmentation accuracy of medical images.",17964
"proposed However, extended easily for saliency the be space-time model prediction. can","However, the proposed model can be easily extended for space-time saliency prediction.",17965
other The over the the approaches. clearly performance illustrate results proposed superior of model,The results clearly illustrate the superior performance of the proposed model over other approaches.,17966
"networks have deep Recently, tasks. performance neural machine vision obtained many on state-of-the-art","Recently, deep neural networks have obtained state-of-the-art performance on many machine vision tasks.",17967
new degradation. like and gradient networks neural introduced convolutional deep Very problems vanishing,Very deep convolutional neural networks introduced new problems like vanishing gradient and degradation.,17968
Residual The recent are these Networks. Highway towards and contributions solving problems successful,The recent successful contributions towards solving these problems are Residual and Highway Networks.,17969
specific are from corresponding features also extracted domain Geometric regions.,Geometric features are also extracted from corresponding domain specific regions.,17970
human Typical characteristic last spatio-temporal structure. several seconds exhibit and actions,Typical human actions last several seconds and exhibit characteristic spatio-temporal structure.,17971
with Recent methods networks. and representations capture attempt this neural convolutional to structure action learn,Recent methods attempt to capture this structure and learn action representations with convolutional neural networks.,17972
we video long-term networks using with (LTC). learn work temporal representations convolutions this neural In,In this work we learn video representations using neural networks with long-term temporal convolutions (LTC).,17973
action increased the improve recognition. models with of temporal demonstrate LTC-CNN We that accuracy extents,We demonstrate that LTC-CNN models with increased temporal extents improve the accuracy of action recognition.,17974
of structuredness the studied chloroplast a Siberian genome We larch. in,We studied the structuredness in a chloroplast genome of Siberian larch.,17975
A the found. those seven-cluster has previously of in structure reported been distribution fragments,A seven-cluster structure in the distribution of those fragments reported previously has been found.,17976
feed-forward inspired by visual models framework for processing. is The cortical,The framework is inspired by feed-forward cortical models for visual processing.,17977
higher accuracy state-of-the-art based model other The to achieve compared proposed event processing classification methods.,The proposed model achieve higher classification accuracy compared to other state-of-the-art event based processing methods.,17978
"what many However, that's do. not in images, people real","However, in many real images, that's not what people do.",17979
detangle in propose multi-person approach people We to images. an,We propose an approach to detangle people in multi-person images.,17980
region the a task formulate We problem. assembly as,We formulate the task as a region assembly problem.,17981
"previously-published of elitist-particle In we process. advancements propose the this selection two article,","In this article, we propose two advancements of the previously-published elitist-particle selection process.",17982
the datasets. multiple capture from can model regularities Our,Our model can capture the regularities from multiple datasets.,17983
"cannot different. or be games, question-answering) similar easily board compared, however","board games, question-answering) cannot be easily compared, however similar or different.",17984
"recognition triangle independently. features accuracy are analyzed and The using from point, line","The recognition accuracy using features from point, line and triangle are analyzed independently.",17985
distortion-free have predicting in the models images. been saliency and designed evaluated for saliency Existing,Existing saliency models have been designed and evaluated for predicting the saliency in distortion-free images.,17986
"In with an complexity. in increase found model performance decrease addition, image is also to","In addition, model performance is also found to decrease with an increase in image complexity.",17987
according lymphangiogenesis to different these models two describing two Here we PDE present hypotheses.,Here we present two PDE models describing lymphangiogenesis according to these two different hypotheses.,17988
the method. Aggregation We (SCDA) Selective propose Descriptor Convolutional,We propose the Selective Convolutional Descriptor Aggregation (SCDA) method.,17989
"or box SCDA annotation. bounding using no unsupervised, image is label","SCDA is unsupervised, using no image label or bounding box annotation.",17990
confirm datasets the image Experiments on for SCDA fine-grained retrieval. of fine-grained effectiveness six,Experiments on six fine-grained datasets confirm the effectiveness of SCDA for fine-grained image retrieval.,17991
disciplines. Statistics of most the valuable one of is,Statistics is one of the most valuable of disciplines.,17992
language Statistics of acceptable science. is only proof the in,Statistics is the only acceptable language of proof in science.,17993
"simple! statistics, be presentation must complex, the In material is precisely the because","In statistics, precisely because the material is complex, the presentation must be simple!",17994
just to article do This endeavors that.,This article endeavors to do just that.,17995
"define systems, a we call intelligent Reflexivity. of We which property","We define a property of intelligent systems, which we call Reflexivity.",17996
"beings, one element of consciousness, is it an human deliberation. and In aspect of","In human beings, it is one aspect of consciousness, and an element of deliberation.",17997
"processes non symbolic, e.g. be These symbolic or may","These processes may be symbolic, or non symbolic e.g.",17998
"In this this article, is the explain why case. we","In this article, we explain why this is the case.",17999
"issues, empirical relevant The several evidence. associated with discusses paper","The paper discusses several associated issues, with relevant empirical evidence.",18000
of There potential interventions on the invasion. is ongoing an therapeutic against glioma debate vaso-modulatory,There is an ongoing debate on the therapeutic potential of vaso-modulatory interventions against glioma invasion.,18001
normalisation. functional deterioration vasculature-targeting therapies involve Prominent and tumour-associated vessel blood,Prominent vasculature-targeting therapies involve functional tumour-associated blood vessel deterioration and normalisation.,18002
"therapy chemo-, stress-induced alleviating to vaso-occlusion, immuno- and is efficacy. via improve radiation","via alleviating stress-induced vaso-occlusion, is to improve chemo-, immuno- and radiation therapy efficacy.",18003
varying of Vaso-modulatory degree the modelled vaso-occlusion. by interventions are,Vaso-modulatory interventions are modelled by varying the degree of vaso-occlusion.,18004
state-of-the-art proposed benchmark Our system public achieves classification accuracy on dataset.,Our proposed system achieves state-of-the-art classification accuracy on public benchmark dataset.,18005
quality in measures importance Learned gain improvement increasing stereo removal and for outlier vision. confidence,Learned confidence measures gain increasing importance for outlier removal and quality improvement in stereo vision.,18006
I to medical this EBSR In propose note some imaging. of applications,In this note I propose some applications of EBSR to medical imaging.,18007
expensive approach object to annotations. quo bounding status The requires detectors box training,The status quo approach to training object detectors requires expensive bounding box annotations.,18008
then boxes. those videos object match We corresponding retrieve to tracked regions the and,We then match those regions to videos and retrieve the corresponding tracked object boxes.,18009
state-space methods propose unknown with two general new We smoothing parameters. Bayesian models for,We propose two new Bayesian smoothing methods for general state-space models with unknown parameters.,18010
of neural We glacialis) networks. right whales recognizing using individual convolutional the studied (Eubalaena feasibility,We studied the feasibility of recognizing individual right whales (Eubalaena glacialis) using convolutional neural networks.,18011
"Here, performance the we describe the of models.","Here, we describe the performance of the models.",18012
of We case the images. single inpainting depth consider,We consider the case of inpainting single depth images.,18013
"previous image frames, corresponding Without next color images, quite depth is challenging. inpainting or","Without corresponding color images, previous or next frames, depth image inpainting is quite challenging.",18014
to may by sparse the gradients inspire penalize non-zero gradient A shallow us regularization. observation,A shallow observation may inspire us to penalize the non-zero gradients by sparse gradient regularization.,18015
We gradient regularization regularization. sparse low with gradient our compare proposed,We compare our proposed low gradient regularization with sparse gradient regularization.,18016
effectiveness of results show experimental approach. proposed The the our,The experimental results show the effectiveness of our proposed approach.,18017
present experimental that effectiveness approach. of results this We the demonstrate,We present experimental results that demonstrate the effectiveness of this approach.,18018
visual dense object This problem: establishing correspondence tackles one different instances. such across paper,This paper tackles one such problem: establishing dense visual correspondence across different object instances.,18019
"CAD are no time, models At required. test","At test time, no CAD models are required.",18020
"for used. finding widely corresponding points Instead, methods are","Instead, methods for finding corresponding points are widely used.",18021
similarity stereo measure suggests suitable for paper a matching This images.,This paper suggests a stereo matching similarity measure suitable for images.,18022
state-of-the-art and using images to validate We method compare real-world the it methods. proposed,We validate the proposed method using real-world images and compare it to state-of-the-art methods.,18023
"illustrated. This the as voxel results using processing directly, in as classification, against intensities improved","This processing results in improved classification, as against using the voxel intensities directly, as illustrated.",18024
"streaming skeleton of this from online study action detection data. we In paper, the problem","In this paper, we study the problem of online action detection from streaming skeleton data.",18025
"our large evaluate annotations. we dataset To a build with proposed streaming video model,","To evaluate our proposed model, we build a large streaming video dataset with annotations.",18026
 tree-structured architecture scale classification. for  present network We large image a,  We present a tree-structured network architecture for large scale image classification.,18027
of all layers the The over convolutional optimized trunk network classes. contains,The trunk of the network contains convolutional layers optimized over all classes.,18028
(CNN). existing be network from neural can structure any convolutional The built proposed,The proposed structure can be built from any existing convolutional neural network (CNN).,18029
has a traditionally major decentralized solving. for Privacy problem motivation been,Privacy has traditionally been a major motivation for decentralized problem solving.,18030
fundamental of problems. families paradigm to is various used a approach programming Constraint,Constraint programming is a fundamental paradigm used to approach various families of problems.,18031
"traditional planning Therefore, problem. viewed privacy is with a requirements as a DisCSP","Therefore, a traditional DisCSP with privacy requirements is viewed as a planning problem.",18032
communication inference. to available local and agents The actions are:,The actions available to agents are: communication and local inference.,18033
technology. UAV is to nowadays in in emerging Application due the advancements indoor of environment,Application of UAV in indoor environment is emerging nowadays due to the advancements in technology.,18034
"terms The facial color, in attributes facial people etc.) (ethnicity, hair of look hair, way","The way people look in terms of facial attributes (ethnicity, hair color, facial hair, etc.)",18035
"accessories wear the clothes hoodies, or they etc.) (sunglasses, and hat,","and the clothes or accessories they wear (sunglasses, hat, hoodies, etc.)",18036
"weather geo-location is highly dependent on condition, and respectively.","is highly dependent on geo-location and weather condition, respectively.",18037
"manually fine-tuned annotated Finally, samples. is network with the","Finally, the network is fine-tuned with manually annotated samples.",18038
trained by a network a scratch. outperforms Our large from method margin,Our method outperforms by a large margin a network trained from scratch.,18039
analysis. task document in recognition spotting important Word an is historical,Word spotting is an important recognition task in historical document analysis.,18040
cases developed evaluated word methods segmentations. assuming and are In most perfect,In most cases methods are developed and evaluated assuming perfect word segmentations.,18041
of methods. the spotting allow word experiments for estimate end-to-end an The of performance,The experiments allow for an estimate of the end-to-end performance of word spotting methods.,18042
"projections. data Euclidean random into a compressed are lower-dimensional Initially, using high-dimensional space","Initially, high-dimensional data are compressed into a lower-dimensional Euclidean space using random projections.",18043
in dataset. an a works well simulations and The supermarket application to in scanner estimator,The estimator works well in simulations and in an application to a supermarket scanner dataset.,18044
major distributed Privacy motivation has a problem been optimization. for,Privacy has been a major motivation for distributed problem optimization.,18045
we content-based In paper image this study retrieval. problem the of,In this paper we study the problem of content-based image retrieval.,18046
problem. is programming minimization quadratic as solved This a problem,This minimization problem is solved as a quadratic programming problem.,18047
is a for problem four This challenging very reasons.,This is a very challenging problem for four reasons.,18048
"Second, data. variability there is negative in a large","Second, there is a large variability in negative data.",18049
"large Finally, world exists. within-class variability data, in real","Finally, in real world data, large within-class variability exists.",18050
"to only some This before, but extent. addressed problem has been","This problem has been addressed before, but only to some extent.",18051
to threefold. action Our detection are online contributions,Our contributions to online action detection are threefold.,18052
protocol fair comparison. an for introduce evaluation We,We introduce an evaluation protocol for fair comparison.,18053
spectrum single neutral $\theta/f$ the well extension the SFS. is immediate of This site known,This spectrum is the immediate extension of the well known single site $\theta/f$ neutral SFS.,18054
local output CNN-Ensemble. pixels of are The beliefs by,The local beliefs of pixels are output by CNN-Ensemble.,18055
"challenging state-of-the-art we any the and SiftFlow on without benchmarks. Even Barcelona achieve results post-processing,","Even without any post-processing, we achieve state-of-the-art results on the challenging SiftFlow and Barcelona benchmarks.",18056
"in disparity times with Thus, strategies. for deal great the we solving can the","Thus, we can deal with the great disparity in solving times for the strategies.",18057
"filters, very indicate other results that surpasses high Experimental the filter levels. especially noise at","Experimental results indicate that the filter surpasses other filters, especially at very high noise levels.",18058
localizing recognizing has video actions understanding. ranges importance human crucial of for and wide Automatically,Automatically recognizing and localizing wide ranges of human actions has crucial importance for video understanding.,18059
proposing We several conclude future and challenges. for THUMOS directions improvements by,We conclude by proposing several directions and improvements for future THUMOS challenges.,18060
"distributions cases, distribution). In most satisfactorily these a law (Pareto described by are power","In most cases, these distributions are satisfactorily described by a power law (Pareto distribution).",18061
a function coordinates logarithmic power straight is line. In,In logarithmic coordinates power function is a straight line.,18062
there was during evolution a differential that angiosperms. species This of extinction the of suggested,This suggested that there was a differential extinction of species during the evolution of angiosperms.,18063
"a less. result, and bigger getting small become genera As large genera became","As a result, large genera became getting bigger and small genera become less.",18064
law. distribution power to species The of in according frequencies a varied the genera,The species frequencies of distribution in the genera varied according to a power law.,18065
platform safety industrial labeled multi-Bernoulli filter presents This for track-before-detect mobile tailored applications. a paper,This paper presents a track-before-detect labeled multi-Bernoulli filter tailored for industrial mobile platform safety applications.,18066
brain volume of average the that of The larger females. human males than is the,The average human brain volume of the males is larger than that of the females.,18067
model differentiable is image parametric Our formation w.r.t.,Our parametric image formation model is differentiable w.r.t.,18068
of challenging ability very blur. demonstrate object experiments cope to cases its Our motion with,Our experiments demonstrate its ability to cope with very challenging cases of object motion blur.,18069
Deep success Convolutional evinced Neural tasks. have various for image immense recognition (CNNs) recently Networks,Deep Convolutional Neural Networks (CNNs) have recently evinced immense success for various image recognition tasks.,18070
"two operations viz. architecture refinement, use We for","We use two operations for architecture refinement, viz.",18071
demonstrate of operations. two empirically necessity the We the,We empirically demonstrate the necessity of the two operations.,18072
substantiate usefulness the Our of method. proposed results the,Our results substantiate the usefulness of the proposed method.,18073
score derived ensemble is Our final from output the of this novelty of classifiers.,Our final novelty score is derived from the output of this ensemble of classifiers.,18074
the model corresponding inaccurate. indicates is misclassification A that saliency,A misclassification indicates that the corresponding saliency model is inaccurate.,18075
"images saliency be Hence, need to updated. which the for the selects classifier models","Hence, the classifier selects images for which the saliency models need to be updated.",18076
be (GNA). (LNA) NA global or can either local,NA can be either local (LNA) or global (GNA).,18077
for a method images. new we in text natural detection In introduce this paper,In this paper we introduce a new method for text detection in natural images.,18078
"We survey generative and hybrid methods, will discriminative methods methods.","We will survey discriminative methods, generative methods and hybrid methods.",18079
agenda aggregation the rules. judgment computation on discuss separability implications of of the We briefly,We briefly discuss the implications of agenda separability on the computation of judgment aggregation rules.,18080
a discriminative learning Network present Convolutional using method Neural shallow We filters a (CNN). for,We present a method for learning discriminative filters using a shallow Convolutional Neural Network (CNN).,18081
well-suited used can extract filters be features for to These classification. invariant image rotation,These filters can be used to extract rotation invariant features well-suited for image classification.,18082
We obtain results to the state-of-the-art. comparable,We obtain results comparable to the state-of-the-art.,18083
to a down-weight then classifier non-action learn and use it irrelevant video segments. We,We then learn a non-action classifier and use it to down-weight irrelevant video segments.,18084
so building problems. far Iterative studied has for been consensus voting only,Iterative consensus building has so far only been studied for voting problems.,18085
"these novel small traits. paths evolutionary However, towards different populations and followed large","However, small and large populations followed different evolutionary paths towards these novel traits.",18086
combination of of optimal becoming an Selection variables crucial. is thus,Selection of an optimal combination of variables is thus becoming crucial.,18087
number number of models possible in the increases The candidate exponentially variables. of,The number of possible models increases exponentially in the number of candidate variables.,18088
"terms of space Moreover, models probabilities. in local model the posterior numerous extrema of has","Moreover, the space of models has numerous local extrema in terms of posterior model probabilities.",18089
The algorithm likelihoods based can be marginal within that efficiently on each model. calculated is,The algorithm is based on that marginal likelihoods can be efficiently calculated within each model.,18090
of marginal or either expressions that applied. likelihoods recommended are precise It exact approximations is,It is recommended that either exact expressions or precise approximations of marginal likelihoods are applied.,18091
convolutional the networks. depth problem estimation addressed works utilizing neural have single-view Recent by,Recent works have addressed the single-view depth estimation problem by utilizing convolutional neural networks.,18092
"transformation-invariant. other neural convolutional we formulate be the hand, networks in On features to","On the other hand, we formulate features in convolutional neural networks to be transformation-invariant.",18093
"new by was proposed the of al. Recently, Ghitany et Lindley distribution a generalization","Recently, a new generalization of the Lindley distribution was proposed by Ghitany et al.",18094
was Sharma proposed tion et generaliza- by Another al.,Another generaliza- tion was proposed by Sharma et al.,18095
"we paper, method. this a Dynamics interpreted that Projective In as can quasi-Newton show be","In this paper, we show that Projective Dynamics can be interpreted as a quasi-Newton method.",18096
The optimization. quasi-Newton ideas numerical leverage interpretation allows from also to us,The quasi-Newton interpretation also allows us to leverage ideas from numerical optimization.,18097
"software implement, reduced method implying development costs. is Our to also easier","Our method is also easier to implement, implying reduced software development costs.",18098
"refinement poses by alignment limb Finally, schemes. the refined are and","Finally, the poses are refined by limb alignment and refinement schemes.",18099
an has computer essential been Finding in features suitable vision. problem,Finding suitable features has been an essential problem in computer vision.,18100
"requiring to features, units, hidden robust our are rotations. compact less learns which Overall, method","Overall, requiring less hidden units, our method learns compact features, which are robust to rotations.",18101
enables weighted-least time variation squares linear and -total solvers for problems. This,This enables linear time solvers for weighted-least squares and -total variation problems.,18102
"even and Our quickly, comparable its solver local converges is to state-of-the-art approaches. EPS runtime","Our solver converges quickly, and its runtime is even comparable to state-of-the-art local EPS approaches.",18103
non-convex propose prior iteratively a We using family term. re-weighted also algorithms a of fast,We also propose a family of fast iteratively re-weighted algorithms using a non-convex prior term.,18104
"definite we $\mathcal{P}_n$. matrices, extend the symmetric of Moreover, positive to this manifold framework","Moreover, we extend this framework to the manifold of symmetric positive definite matrices, $\mathcal{P}_n$.",18105
"this Further, we proved {\it is sparse that a the KL-center combination atoms}. of dictionary","Further, {\it we proved that this KL-center is a sparse combination of the dictionary atoms}.",18106
This an using gradient descent cost algorithm. is acceleterated minimized function,This cost function is minimized using an acceleterated gradient descent algorithm.,18107
proposed of results the show extensive An experimental our framework. set effectiveness of,An extensive set of experimental results show the effectiveness of our proposed framework.,18108
classification problems experiments in variety We Computer Vision several of applications. present involving a,We present several experiments involving a variety of classification problems in Computer Vision applications.,18109
a directed graph. as formulate deep acyclic CNN We a,We formulate a deep CNN as a directed acyclic graph.,18110
our are and of generally favorable. results on a the method set We CNNs evaluated,We evaluated our method on a set of CNNs and the results are generally favorable.,18111
video. Multi-person role analysis surveillance of critical plays tracking in a the,Multi-person tracking plays a critical role in the analysis of surveillance video.,18112
"However, focus work existing shorter-term most (e.g. on","However, most existing work focus on shorter-term (e.g.",18113
"for very a (e.g. tracking algorithm long-term we Therefore, multi-person propose","Therefore, we propose a multi-person tracking algorithm for very long-term (e.g.",18114
our tracking on video output. further summarization performed based We experiments,We further performed video summarization experiments based on our tracking output.,18115
Shoham's a database logic intentions We temporal for on based perspective. and introduce beliefs,We introduce a logic for temporal beliefs and intentions based on Shoham's database perspective.,18116
from weak beliefs separate beliefs. We strong,We separate strong beliefs from weak beliefs.,18117
and intentions. on coherence strong formalize conditions We beliefs,We formalize coherence conditions on strong beliefs and intentions.,18118
revision intentions. strong and provide of the We AGM-style beliefs postulates for,We provide AGM-style postulates for the revision of strong beliefs and intentions.,18119
an is issue image of mining. Object hot categorization a,Object categorization is a hot issue of an image mining.,18120
between is the of important an semantic information knowledge objects of Contextual image. one,Contextual information between objects is one of the important semantic knowledge of an image.,18121
and the proposed this Experimental reliability validity show of method. results,Experimental results show the validity and reliability of this proposed method.,18122
other. knowledge Information transformable and into each are,Information and knowledge are transformable into each other.,18123
with functions propose image learning hash for search. kernels incremental for strategy large-scale an We,We propose an incremental strategy for learning hash functions with kernels for large-scale image search.,18124
"codes become a binary space of SVM. multi-class the stage, the second In input","In the second stage, binary codes become the input space of a multi-class SVM.",18125
"and representation novel is sparse a proposed. learning dictionary In paper, (SS-DLSR) this semi-supervised","In this paper, a novel semi-supervised dictionary learning and sparse representation (SS-DLSR) is proposed.",18126
This is independence performed criterion Hilbert-Schmidt (HSIC). using maximization,This maximization is performed using Hilbert-Schmidt independence criterion (HSIC).,18127
makes the the for fast. algorithm This proposed very solution,This makes the solution for the proposed algorithm very fast.,18128
engine game. implemented an for analyzing We the,We implemented an engine for analyzing the game.,18129
first theoretically solving the the These are steps tablebases towards game.,These tablebases are the first steps towards theoretically solving the game.,18130
"trees on decision we Finally, the tablebases. based endgame constructed","Finally, we constructed decision trees based on the endgame tablebases.",18131
note we this report on some interesting In patterns.,In this note we report on some interesting patterns.,18132
actions localization videos. spatio-temporal strive of We for in,We strive for spatio-temporal localization of actions in videos.,18133
"in error is video cumbersome, tedious, boxes and prone. Annotating action","Annotating action boxes in video is cumbersome, tedious, and error prone.",18134
a procedure We with functions flexible by search for kernels. hash propose large-scale image,We propose a flexible procedure for large-scale image search by hash functions with kernels.,18135
learning re-identification are Feature two metric and components representation critical models. person in,Feature representation and metric learning are two critical components in person re-identification models.,18136
"by CNN FFN, features handcrafted the constrained In makes propagation features. back","In FFN, back propagation makes CNN features constrained by the handcrafted features.",18137
which divergent under We consider cumulative the evolution cultural can conditions occur.,We consider the conditions under which divergent cumulative cultural evolution can occur.,18138
hypothesize necessary. that We are two conditions,We hypothesize that two conditions are necessary.,18139
stored in agent. that cultural First separately and the are information genetic,First that genetic and cultural information are stored separately in the agent.,18140
information between generations. Second cultural of agents horizontally must be transferred different,Second cultural information must be transferred horizontally between agents of different generations.,18141
"maps action the and detection. apply estimated we proposal actionness generation on Moreover, action","Moreover, we apply the estimated actionness maps on action proposal generation and action detection.",18142
state-of-the-art tasks the of substantially. advance Our actionness these performance current maps,Our actionness maps advance the current state-of-the-art performance of these tasks substantially.,18143
"retrained model objects. be a the different and to As initialized for need result,","As a result, the model need to be initialized and retrained for different objects.",18144
specific be Thus any to it confined object. not will,Thus it will not be confined to any specific object.,18145
community. in detection computer area and newly is rapidly Co-saliency growing vision research emerging a,Co-saliency detection is a newly emerging and rapidly growing research area in computer vision community.,18146
learners towards Social were philopatry. more natal inclined,Social learners were more inclined towards natal philopatry.,18147
reproductive a learners including labor. evolved also of Social of eusociality division culture,Social learners also evolved a culture of eusociality including reproductive division of labor.,18148
is motion an undergoing TFG for periodic prove also object periodic. We that,We prove that TFG for an object undergoing periodic motion is also periodic.,18149
method improve gradient a performance. PGRD's variance-reducing We also adopt to,We also adopt a variance-reducing gradient method to improve PGRD's performance.,18150
We computable. particular is agent standard determination show whether of any not meets that an,We show that determination of whether an agent meets any particular standard is not computable.,18151
anti-cancer therapies. to linked is NTDhXPB also,NTDhXPB is also linked to anti-cancer therapies.,18152
TB target be for therapeutics. potential These could,These could be potential target for TB therapeutics.,18153
effectiveness Experiments of show algorithm. the the pipeline proposed and our,Experiments show the effectiveness of our pipeline and the proposed algorithm.,18154
deep architecture The video excellent exhibited action recognition. on based two-stream performance,The deep two-stream architecture exhibited excellent performance on video based action recognition.,18155
"transfer, transfer initialization combination. introduce strategies and this, supervision we their three for Specifically,","Specifically, we introduce three strategies for this, initialization transfer, supervision transfer and their combination.",18156
without a function. consider products firm periods over that demand We $T$ sells knowing the,We consider a firm that sells products over $T$ periods without knowing the demand function.,18157
which performance. dispersion forced introduced and They integrate algorithms price GILS achieve asymptotically optimal with,They introduced algorithms which integrate forced price dispersion with GILS and achieve asymptotically optimal performance.,18158
"this a we In this environment. dynamic pricing in data-rich paper, problem consider","In this paper, we consider this dynamic pricing problem in a data-rich environment.",18159
this in that prove $\log(T)$. order asymptotically regret setting optimal of GILS achieves We,We prove that in this setting GILS achieves asymptotically optimal regret of order $\log(T)$.,18160
extensive We via and real our synthetic sets. numerical validate results on simulations data,We validate our results via extensive numerical simulations on synthetic and real data sets.,18161
"to most independent. have date, works be considered In attributes to been","In most works to date, attributes have been considered to be independent.",18162
"the However, be to this know we case. not","However, we know this not to be the case.",18163
"very Many related, are heavy wearing strongly and makeup such attributes lipstick. as","Many attributes are very strongly related, such as heavy makeup and wearing lipstick.",18164
might be training data further possible improvement suggests Our time. with more,Our data suggests further improvement might be possible with more training time.,18165
"fingerprint addition, feature from is a In image fingerprint explored. also minutia-based a","In addition, a minutia-based fingerprint feature from a fingerprint image is also explored.",18166
involves Sparse a GEP solving optimization non-convex problem.,Sparse GEP involves solving a non-convex optimization problem.,18167
"this computational In two-stage the solve a GEP. paper, we framework to propose sparse","In this paper, we propose a two-stage computational framework to solve the sparse GEP.",18168
"a At we the of first sparse GEP. solve relaxation stage, the convex","At the first stage, we solve a convex relaxation of the sparse GEP.",18169
studies are numerical theoretical Thorough the results. to provided validate,Thorough numerical studies are provided to validate the theoretical results.,18170
scenes semantic visual object between for focuses This work recognition. the on relations objects and,This work focuses on the semantic relations between scenes and objects for visual object recognition.,18171
"Neural network perception mechanism are oscillations consciousness. for and a fundamental cognition,","Neural network oscillations are a fundamental mechanism for cognition, perception and consciousness.",18172
"brain perturbations of pathophysiology in an of disorders. activity important role the play Consequently, network","Consequently, perturbations of network activity play an important role in the pathophysiology of brain disorders.",18173
such methodology clustering found in can be financial assets. Applications of the of,Applications of such methodology can be found in the clustering of financial assets.,18174
"experiments found A tutorial, research for www.datagrapple.com/Tech. implementation can at and be reproducible","A tutorial, experiments and implementation for reproducible research can be found at www.datagrapple.com/Tech.",18175
"applicable egocentric show it both conventional that is Furthermore, to and we images.","Furthermore, we show that it is applicable to both conventional and egocentric images.",18176
the video-based video-based re-id re-id show methods. proposed the outperforms experiments framework that Our state-of-the-art,Our experiments show that the proposed video-based re-id framework outperforms the state-of-the-art video-based re-id methods.,18177
"this problem. In we image super-resolution consider work, the (SR)","In this work, we consider the image super-resolution (SR) problem.",18178
"algorithm analysis, developed. is our To an denoising verify image iterative","To verify our analysis, an iterative image denoising algorithm is developed.",18179
then moments equations. recurrent a time exit of satisfy differential-difference system of the The,The moments of the exit time then satisfy a system of recurrent differential-difference equations.,18180
The of the also deviation standard is eradication investigated. time,The standard deviation of the eradication time is also investigated.,18181
this a was Doing single-handed imagination. video sequence for beyond,Doing this for a video sequence single-handed was beyond imagination.,18182
baselines simpler both method that the qualitatively and outperforms We quantitatively. clearly show proposed,We show that the proposed method clearly outperforms simpler baselines both qualitatively and quantitatively.,18183
validate our of method. Experimental effectiveness results presented the,Experimental results validate the effectiveness of our presented method.,18184
fundamental images correspondences dense Establishing is a many between in applications. task multiple,Establishing dense correspondences between multiple images is a fundamental task in many applications.,18185
of The is by computational applying fast dense edge-aware redundancy descriptors reduced filtering. dramatically,The computational redundancy of dense descriptors is dramatically reduced by applying fast edge-aware filtering.,18186
proposed qualitatively the method's We quantitatively. demonstrate effectiveness and both,We demonstrate the proposed method's effectiveness both qualitatively and quantitatively.,18187
"particular, and presented. the beauty feeling are the of pleasure definitions In of","In particular, the definitions of pleasure and the feeling of beauty are presented.",18188
"""good"", are ""hot"" analyzed. ""wish"", The meanings ""wait"", of","The meanings of ""wish"", ""wait"", ""good"", ""hot"" are analyzed.",18189
always from Understanding a fundamental of humans computer has photographs vision. goal been,Understanding humans from photographs has always been a fundamental goal of computer vision.,18190
have enlargement of with been transient associated mechanisms identified experimentally. Many molecular spine the,Many of the molecular mechanisms associated with transient spine enlargement have been identified experimentally.,18191
the able of dynamics transient is capture volume. to observed spine Our experimentally model,Our model is able to capture the experimentally observed dynamics of transient spine volume.,18192
"actin provide robustness volume we a show Furthermore, remodeling that to spine events dynamics.","Furthermore, we show that actin remodeling events provide a robustness to spine volume dynamics.",18193
variational hybrid gradient by solved (PDHG) algorithm. primal-dual is the problem The,The variational problem is solved by the primal-dual hybrid gradient (PDHG) algorithm.,18194
disorganized loss. absence Their movements associated with is division cell chromosome and mitotic during chromosome,Their absence during cell division is associated with disorganized mitotic chromosome movements and chromosome loss.,18195
first. in predicted are video Salient frames patches,Salient patches in video frames are predicted first.,18196
maps the them. fixation Then are built predicted upon visual,Then the predicted visual fixation maps are built upon them.,18197
toolkit. basis of design implemented Caffe CaffeNet We the deep architecture the with on,We design the deep architecture on the basis of CaffeNet implemented with Caffe toolkit.,18198
on two The conducted experiments datasets. available publicly are,The experiments are conducted on two publicly available datasets.,18199
"murky in problem. very However, challenging a water this is","However, in murky water this is a very challenging problem.",18200
"learning important. statistical and Consequently, different machine models are data to tailored encodings","Consequently, statistical and machine learning models tailored to different data encodings are important.",18201
and to achieve difficult a normally adjustment artifact-free problem This is quality. very manual requires,This is a very difficult problem and normally requires manual adjustment to achieve artifact-free quality.,18202
"a colorization aims the previous Unlike fully-automatic at this high-quality method. paper methods,","Unlike the previous methods, this paper aims at a high-quality fully-automatic colorization method.",18203
"To proposed. post-processing a is artifact-free quality, step ensure filtering bilateral based joint","To ensure artifact-free quality, a joint bilateral filtering based post-processing step is proposed.",18204
image We clustering to image global adaptive further incorporate an information. the technique develop,We further develop an adaptive image clustering technique to incorporate the global image information.,18205
have Fourier central role convolution development. theory and such in analysis,Fourier analysis and convolution theory have central role in such development.,18206
"one-dimensional Fourier q-Gaussian some theoretical elements we and review behind transform. the its Firstly,","Firstly, we review some theoretical elements behind the one-dimensional q-Gaussian and its Fourier transform.",18207
order actual in further such station. Both visits the cases require to of meet demands,Both cases require further visits in order to meet the actual demands of such station.,18208
to heuristic We based the (ILS) an problem. solve local iterated search propose,We propose an iterated local search (ILS) based heuristic to solve the problem.,18209
the action. motion primitives of assembled of an The representation improved vector is,The assembled vector of motion primitives is an improved representation of the action.,18210
on method four different datasets. demonstrate our We,We demonstrate our method on four different datasets.,18211
formability in consider work. We this coalition abstract-argumentation-theoretic,We consider abstract-argumentation-theoretic coalition formability in this work.,18212
"any general, However, can presented method and the wavelet is it be on applied transform.","However, the presented method is general, and it can be applied on any wavelet transform.",18213
exhaustive is Our experiments reasoning high-end cards. on by supported graphics,Our reasoning is supported by exhaustive experiments on high-end graphics cards.,18214
reviews the paper classification state-of-the comprehensively This AV in methods. art,This paper comprehensively reviews the state-of-the art in AV classification methods.,18215
"new improve is a To proposed. Bi- Pattern-based (LBP) on methods, previous nary method Local","To improve on previous methods, a new Local Bi- nary Pattern-based method (LBP) is proposed.",18216
"this detection for hyperspectral a data. model we (HSI) propose paper, imaging In burnscar","In this paper, we propose a burnscar detection model for hyperspectral imaging (HSI) data.",18217
official the are available public conducted website. on dataset NASA MODIS Experiments at,Experiments are conducted on the public MODIS dataset available at NASA official website.,18218
is known technique the scheduling a jobs short to performance. the long Separating from improve,Separating the short jobs from the long is a known technique to improve scheduling performance.,18219
available unique with dataset annotations. High-Resolution Flight is publicly Our Moth made (HRMF),Our unique High-Resolution Moth Flight (HRMF) dataset is made publicly available with annotations.,18220
"not for errors. standard a EM does covariance produce parameter matrix Unfortunately,","Unfortunately, EM does not produce a parameter covariance matrix for standard errors.",18221
made Complete source code available. freely is,Complete source code is made freely available.,18222
employ vast cameras of The a mechanism. consumer-grade majority rolling shutter modern,The vast majority of modern consumer-grade cameras employ a rolling shutter mechanism.,18223
"significance, to inadequate its date. However, received has despite attention it","However, despite its significance, it has received inadequate attention to date.",18224
requires rate methods and have diverse which steps Solving convergence optimization error. minimization,Solving minimization steps requires optimization methods which have diverse convergence rate and error.,18225
results parameters used method. Lucas-Kanade could for experimental tuning Our as be,Our experimental results could be used as tuning parameters for Lucas-Kanade method.,18226
"to regularization with functions. provide complementary Dropout DisturbLabel well Meanwhile, cooperates","Meanwhile, DisturbLabel cooperates well with Dropout to provide complementary regularization functions.",18227
demonstrate recognition recognition popular datasets. image several Experiments results on competitive,Experiments demonstrate competitive recognition results on several popular image recognition datasets.,18228
chemoattractant be in have shown can concentrations. to bacteria variations small sensitive (CA) Experiments that,Experiments have shown that bacteria can be sensitive to small variations in chemoattractant (CA) concentrations.,18229
"as Once sources such point bacterial effect"". parameterized, the ""volcano behavior model our recapitulates around","Once parameterized, our model recapitulates bacterial behavior around point sources such as the ""volcano effect"".",18230
of proposed. is algorithms for variables A fuzzy described new distributional c-means by set data,A new set of fuzzy c-means algorithms for data described by distributional variables is proposed.,18231
product-to-one weights computed in the process relevance clustering The introducing constraints. are,The relevance weights are computed in the clustering process introducing product-to-one constraints.,18232
We artificial proposed data. tested have real-world on the and algorithms,We have tested the proposed algorithms on artificial and real-world data.,18233
of symmetries these advantages The broken evolutionary unknown. remain,The evolutionary advantages of these broken symmetries remain unknown.,18234
as can be interpreted that neuron spatial-weighted activeness indicates Visualization responses. neuron,Visualization indicates that neuron activeness can be interpreted as spatial-weighted neuron responses.,18235
image state-of-the-art achieve range datasets. performance a of We wide classification on,We achieve state-of-the-art classification performance on a wide range of image datasets.,18236
unifying framework for a This multi-atlas segmentation. provides,This provides a unifying framework for multi-atlas segmentation.,18237
"studied promising in visual widely has shown tracking, which performance. has representation Sparse been tracking","Sparse representation has been widely studied in visual tracking, which has shown promising tracking performance.",18238
qualitative that trackers. state-of-the-art outperforms quantitative and algorithm previous our results significantly demonstrate Both,Both qualitative and quantitative results demonstrate that our algorithm significantly outperforms previous state-of-the-art trackers.,18239
"first paper, this new formally. define task this we In","In this paper, we first define this new task formally.",18240
Experimental well results real as on synthetic MVMDS. datasets effectiveness as demonstrate the of,Experimental results on synthetic as well as real datasets demonstrate the effectiveness of MVMDS.,18241
get. hard usually The results right are to,The right results are usually hard to get.,18242
"these improve the can areas, matting. effects flags are On certain provided, it of if","On these areas, if certain flags are provided, it can improve the effects of matting.",18243
"segmentation, in entity resolution. community discovery Applications this and image include domain","Applications in this domain include image segmentation, community discovery and entity resolution.",18244
pseudo-code are algorithms at this of so whole end detailed document. the developed in The,The whole so developed algorithms are detailed in pseudo-code at the end of this document.,18245
a detection multiantenna moving for proposes spatio-temporal paper in decomposition targets the SAR. This of,This paper proposes a spatio-temporal decomposition for the detection of moving targets in multiantenna SAR.,18246
and detection due target to smearing more is masking target challenging Moving clutter. by,Moving target detection is more challenging due to target smearing and masking by clutter.,18247
results to real- upcoming These be used will our time inform experiments.,These results will be used to inform our upcoming real- time experiments.,18248
field. a possesses estimation significant Automatic and in healthcare medical intensity pain position,Automatic pain intensity estimation possesses a significant position in healthcare and medical field.,18249
continuous-valued then recurrent output the to pain the network intensity. carefully of design architecture We,We then carefully design the architecture of the recurrent network to output continuous-valued pain intensity.,18250
suppress the promote others. and some emergence Structure populations can in it of in cooperation,Structure can promote the emergence of cooperation in some populations and suppress it in others.,18251
"randomly previous fixation mutants. on Our work of extend probabilities single, placed results","Our results extend previous work on fixation probabilities of single, randomly placed mutants.",18252
never for for that (BD) is birth-death We any cooperation favored find configuration updating.,We find that for any configuration cooperation is never favored for birth-death (BD) updating.,18253
robots complex on decision-making perception Robust essential operating in environment domains. for is,Robust environment perception is essential for decision-making on robots operating in complex domains.,18254
"fast may and in always, be not however, These, objects. for particular valid moving irregularly","These, however, may not be valid always, in particular for fast and irregularly moving objects.",18255
the the on method efficiency of demonstrate studies. We two and inference case usefulness,We demonstrate the usefulness and efficiency of the inference method on two case studies.,18256
of are order also increases higher moments The accuracy considered. when two than,The accuracy increases when also moments of order higher than two are considered.,18257
fourfold. The of this paper contribution is,The contribution of this paper is fourfold.,18258
recognition of cameracaptured for third document The a method contribution is the novel images.,The third contribution is a novel method for the recognition of cameracaptured document images.,18259
"to In we a robust paper, solution propose this problem. this","In this paper, we propose a robust solution to this problem.",18260
"estimation, separation. performs optical and proposed both layer image The flow method","The proposed method performs both optical flow estimation, and image layer separation.",18261
limited a The set of fixations. refines on truth the ground step second pre-trained model,The second step refines the pre-trained model on a limited set of ground truth fixations.,18262
results in of promising recent ADMM Plug-and-Play a empirical demonstrated number papers. has,Plug-and-Play ADMM has demonstrated promising empirical results in a number of recent papers.,18263
"In Plug-and-Play convergence. we a fixed algorithm with paper, point this ADMM provable propose","In this paper, we propose a Plug-and-Play ADMM algorithm with provable fixed point convergence.",18264
fast restoration super-resolution two on also image We implementations single-photon problems present imaging. for and,We also present fast implementations for two image restoration problems on super-resolution and single-photon imaging.,18265
discover that all permanent dipoles. charges have and net surface electric negative We,We discover that all have net negative surface charges and permanent electric dipoles.,18266
is vision. an Image and in inpaiting processing task image important,Image inpaiting is an important task in image processing and vision.,18267
science datasets a in large is challenge Detecting extreme in research. major events climate,Detecting extreme events in large datasets is a major challenge in climate science research.,18268
"dataset. methods same different on vastly multiple competing the Often, results produce","Often, multiple competing methods produce vastly different results on the same dataset.",18269
at computational cost. is extra done This no,This is done at no extra computational cost.,18270
effects. provide interventions source causal learning information most the Controlled direct of for,Controlled interventions provide the most direct source of information for learning causal effects.,18271
"can be and time-consuming. However, interventions expensive","However, interventions can be expensive and time-consuming.",18272
"not Observational available. a is data, the by is where treatment known mechanism, sometimes controlled","Observational data, where the treatment is not controlled by a known mechanism, is sometimes available.",18273
"observational allows data for dose-response strong estimation some curves. Under assumptions, of the","Under some strong assumptions, observational data allows for the estimation of dose-response curves.",18274
representations We the for task employ image-caption ranking. of these,We employ these representations for the task of image-caption ranking.,18275
captions and interpret allows of images from a to wide model This variety perspectives. the,This allows the model to interpret images and captions from a wide variety of perspectives.,18276
consistency captions images between and incorporating reasoning We significantly performance. find improves and about that,We find that incorporating and reasoning about consistency between images and captions significantly improves performance.,18277
Using is behaviour such data critical maintenance. safety and efficient to for anomalous detect,Using such data to detect anomalous behaviour is critical for safety and efficient maintenance.,18278
introduction modeling analysis. image the gradual techniques hierarchical This in document of showing is last,This last document is showing the gradual introduction of hierarchical modeling techniques in image analysis.,18279
axonal frontal dynamic that of lobe. behavior We the hypothesize development copies this the fiber,We hypothesize that this dynamic behavior copies the axonal fiber development of the frontal lobe.,18280
was de by Vienne It observed recently et al.,It was recently observed by de Vienne et al.,18281
We patch-based address this problem. framework a to describe,We describe a patch-based framework to address this problem.,18282
requires resulting object boxes. only approach The bounding,The resulting approach only requires object bounding boxes.,18283
represented a meshes. new volumes for performing Boolean operations method triangle We present on as,We present a new method for performing Boolean operations on volumes represented as triangle meshes.,18284
edge-preserving smoothing. range kernel filter a uses The a perform spatial along kernel with to,The filter uses a spatial kernel along with a range kernel to perform edge-preserving smoothing.,18285
"the kernels where In Gaussian the this bilateral we filter are both Gaussian. consider paper,","In this paper, we consider the Gaussian bilateral filter where both the kernels are Gaussian.",18286
"it well-known In in slow implementation fact, the is that practice. is direct","In fact, it is well-known that the direct implementation is slow in practice.",18287
"endows that properties system the underlying However, remains the response with mechanism these elusive.","However, the underlying mechanism that endows the system with these response properties remains elusive.",18288
beliefs of of requires information to plausibility the be sources the the merged. Merging,Merging beliefs requires the plausibility of the sources of the information to be merged.,18289
"Yet, may no or available. previous examples be revisions","Yet, no examples or previous revisions may be available.",18290
article first scenario indeed The proves this a possible. that of theorem is such,The first theorem of this article proves that such a scenario is indeed possible.,18291
are reliability merging. various Other and under results definition obtained of,Other results are obtained under various definition of reliability and merging.,18292
The described the first to tailored distribution a loss fits by function. errors a,The first fits to the errors a distribution described by a tailored loss function.,18293
for spatial that continuity. describing errors will We show characterization is effective this with joint,We will show that this joint characterization is effective for describing errors with spatial continuity.,18294
pixel-wise a estimation We depth formulate as task. classification,We formulate depth estimation as a pixel-wise classification task.,18295
performance. our on and indoor evaluate We state-of-the-art approach datasets achieve and both outdoor,We evaluate our approach on both indoor and outdoor datasets and achieve state-of-the-art performance.,18296
filter The slow direct of non-linear in this computation practice. is,The direct computation of this non-linear filter is slow in practice.,18297
maps. that Our approach experiments improves the show disparity significantly the of proposed accuracy,Our experiments show that the proposed approach significantly improves the accuracy of disparity maps.,18298
issues regarding verification Some are problems DCNNs face open discussed. for then,Some open issues regarding DCNNs for face verification problems are then discussed.,18299
cannot prevalent techniques. in these Uncertainties captured by be an image,Uncertainties prevalent in an image cannot be captured by these techniques.,18300
"other are noise, artifacts. outliers corrupted and when deteriorates The by images further the performance","The performance further deteriorates when the images are corrupted by noise, outliers and other artifacts.",18301
Malevolent (MAI). Artificial This paper a some Intelligence creation guidelines the of general provides for,This paper provides some general guidelines for the creation of a Malevolent Artificial Intelligence (MAI).,18302
"The the ability of non-stationary outliers, i.e. proposed advantage is key to handle method our","The key advantage of our proposed method is the ability to handle non-stationary outliers, i.e.",18303
un pasos. de proceso este Actualmente en se resuelve dos problema,Actualmente este problema se resuelve en un proceso de dos pasos.,18304
"capturando pr\'acticos, con costosos. dispositivos datos intrusivos y Primero poco","Primero capturando datos con dispositivos poco pr\'acticos, intrusivos y costosos.",18305
automatizar futuros el de calculo esas variables. la de En proceso es idea todo trabajos,En trabajos futuros la idea es automatizar todo el proceso de calculo de esas variables.,18306
"benefit e.g. Deep from networks usually neural pre-training, unsupervised","Deep neural networks usually benefit from unsupervised pre-training, e.g.",18307
"classifier for However, supervised discrimination. needs further the good methods fine-tuning","However, the classifier further needs supervised fine-tuning methods for good discrimination.",18308
also promising face models. a could be of recognition training fast for This direction,This could also be a promising direction for fast training of face recognition models.,18309
Matrix Component Analysis (PCA) investigated Non-negative approaches. Principal candidate Factorization and (NMF) clustering as are,Principal Component Analysis (PCA) and Non-negative Matrix Factorization (NMF) are investigated as candidate clustering approaches.,18310
reading/memorization session. and a a with by are resting preceded session The followed presented subjects,The subjects are presented with a reading/memorization session preceded and followed by a resting session.,18311
during signals recorded these were sessions. EEG,EEG signals were recorded during these sessions.,18312
"was the in variability. no difference However, significant there deviation standard blink in rate","However, there was no significant difference in standard deviation in the blink rate variability.",18313
"tree a As the game of may paradigm properties unexploited. this leave result, some","As a result, this paradigm may leave some properties of the game tree unexploited.",18314
of to basis use not a is of size. the However point large P-splines part,However part of the point of P-splines is not to use a large basis size.,18315
images body orientations different for in captured contain each person. datasets multiple Both,Both datasets contain multiple images captured in different body orientations for each person.,18316
"we property Particularly, model: investigate ideas under two fusion. the derivation primary and property CNN","Particularly, we investigate two primary ideas under the CNN model: property derivation and property fusion.",18317
framework So a of learning is two-stage fusion consisting derivation and constructed. property,So a two-stage learning framework consisting of property derivation and fusion is constructed.,18318
either or the Here can their pairs the color/depth derived provided from (e.g. properties be,Here the properties can be derived either from the provided color/depth or their pairs (e.g.,18319
geometry the contour this adopted in paper).,the geometry contour adopted in this paper).,18320
"proposed dataset, method the evaluate challenging and on experimentally performance. state-of-the-art We the achieved have","We experimentally evaluate the proposed method on the challenging dataset, and have achieved state-of-the-art performance.",18321
is progress a expansion. undergoing experimental period rapid and of Neuroscience,Neuroscience is undergoing a period of rapid experimental progress and expansion.,18322
surface Polarization at polarizer from Shape photos rotations. different using (SfP) normals estimates captured,Shape from Polarization (SfP) estimates surface normals using photos captured at different polarizer rotations.,18323
"model assumes specularly. or Fundamentally, either light that SfP diffusely reflected the is","Fundamentally, the SfP model assumes that light is reflected either diffusely or specularly.",18324
three of from saliency visual We issue the address perspectives.,We address the issue of visual saliency from three perspectives.,18325
"consider saliency detection analysis frequency we problem. domain First, a as","First, we consider saliency detection as a frequency domain analysis problem.",18326
"this the employing of non-saliency}. concept Second, {\it by achieve we","Second, we achieve this by employing the concept of {\it non-saliency}.",18327
"of size. different Third, we of the regions consider detection simultaneously salient","Third, we simultaneously consider the detection of salient regions of different size.",18328
Fourier performs Transform analysis Hypercomplex in frequency A domain. the the,A Hypercomplex Fourier Transform performs the analysis in the frequency domain.,18329
"databases, demonstrate Using that human predict proposed we fixation experimentally available model the can data.","Using available databases, we demonstrate experimentally that the proposed model can predict human fixation data.",18330
anatomical on motion. details compromise It and between relies visibility to a robustness,It relies on a compromise between anatomical details visibility and robustness to motion.,18331
improvement for computing. the of Another facilities provision parallel major is,Another major improvement is the provision of facilities for parallel computing.,18332
has using islands the master-slave been approach implemented the Parallelisation and both model. evolution,Parallelisation has been implemented using both the master-slave approach and the islands evolution model.,18333
the finds challenging aspect these its The simplicity of algorithms classification stage. in most,The simplicity of these algorithms finds its most challenging aspect in the classification stage.,18334
the MAPE kernel regression illustrated simulated of data. is The behavior on,The behavior of the MAPE kernel regression is illustrated on simulated data.,18335
tracker two propose a at that We models detail. the of levels object,We propose a tracker that models the object at two levels of detail.,18336
dynamic territory. is emergent analysis properties Formal networks of of uncharted structural the largely,Formal analysis of the emergent structural properties of dynamic networks is largely uncharted territory.,18337
"for calculate thresholds node dynamics, networks the growth FRS. For the with of we","For networks with node dynamics, we calculate thresholds for the growth of the FRS.",18338
size The explored of population simulation are approximation. finite effects via and,The effects of finite population size are explored via simulation and approximation.,18339
using theoretical validate tests Numerical and simulated findings. datasets our real,Numerical tests using simulated and real datasets validate our theoretical findings.,18340
is sets. Data segmentation data used handle large to,Data segmentation is used to handle large data sets.,18341
"favoriting We purchase through listing popularity clicks, search define and activity. first","We first define listing popularity through search clicks, favoriting and purchase activity.",18342
"pixel-level listing images quality listed the from quality features. as we infer of Next, information","Next, we infer listing quality from the pixel-level information of listed images as quality features.",18343
We to then models our compare findings prediction. popularity text-only for,We then compare our findings to text-only models for popularity prediction.,18344
in action approach video techniques. recognition presents sparse an coding category-based work This to using,This work presents an approach to category-based action recognition in video using sparse coding techniques.,18345
be categories. among discriminate resulting to demonstrates ITRA The highly action to effective descriptor,The resulting ITRA descriptor demonstrates to be highly effective to discriminate among action categories.,18346
affordances functional regions aspect or is Localizing understanding. scene objects important of an of,Localizing functional regions of objects or affordances is an important aspect of scene understanding.,18347
"can digraph. as viewed people the Then, the in flow integer constructed be an flow","Then, the people flow can be viewed as an integer flow in the constructed digraph.",18348
will it? question is: they The to want use,The question is: will they want to use it?,18349
technology. driving Autonomous evolving is a rapidly,Autonomous driving is a rapidly evolving technology.,18350
"proof of hypothesis, a This strong AI i.e. the is","This is a proof of the strong AI hypothesis, i.e.",18351
systems achievable are artifacts. Pattern recognition technical,Pattern recognition systems are achievable technical artifacts.,18352
introspective art. proof links philosophical of theory this consciousness with technical respected The,The proof links this respected introspective philosophical theory of consciousness with technical art.,18353
to paradigms are possible and AI several Similarities applications discussed.,Similarities and possible applications to several AI paradigms are discussed.,18354
that fronts. end contribute we To on three,To that end we contribute on three fronts.,18355
and two pose is two estimation single-person multi-person Evaluation benchmarks. done on,Evaluation is done on two single-person and two multi-person pose estimation benchmarks.,18356
blind has separation the years of been processing investigated. image recent In,In recent years the processing of blind image separation has been investigated.,18357
Conclusion without The automatic any user fully interaction. proposed method is,Conclusion The proposed method is fully automatic without any user interaction.,18358
and clinical main independent linked consciousness-related features. FC-traits Three to identified were,Three main independent FC-traits were identified and linked to consciousness-related clinical features.,18359
reflects second sensory-motor The FC-trait patterns. disconnection and of the connectivity visual the,The second FC-trait reflects the disconnection of the visual and sensory-motor connectivity patterns.,18360
starting on point. segmentation rely as video Graph-based superpixels methods,Graph-based video segmentation methods rely on superpixels as starting point.,18361
question arguing by the address to paper aims This following. this,This paper aims to address this question by arguing the following.,18362
of operations The action mutational simplification under provides performance selection natural and pressure. these joint,The joint action of these mutational operations and natural selection provides simplification under performance pressure.,18363
accident from local but not network-level Novelty evolution. arises genetic a gradual from thus,Novelty thus arises not from a local genetic accident but from gradual network-level evolution.,18364
bounding same they These since the image. highly originate boxes from are correlated,These bounding boxes are highly correlated since they originate from the same image.,18365
How intelligent can for design good goals agents? we arbitrarily,How can we design good goals for arbitrarily intelligent agents?,18366
(RL) a approach. is natural Reinforcement learning,Reinforcement learning (RL) is a natural approach.,18367
learning (VRL). In we called suggest this paper an RL to value reinforcement alternative,In this paper we suggest an alternative to RL called value reinforcement learning (VRL).,18368
"use agents learn VRL, a the reward function. utility to signal In","In VRL, agents use the reward signal to learn a utility function.",18369
effective tools are object visual recognition. distances very for Manifold,Manifold distances are very effective tools for visual object recognition.,18370
fuzzy the advances in Important made quantification been have field.,Important advances have been made in the fuzzy quantification field.,18371
on statistical matrix problem simple models. is very based A introduced completion interpretation of,A very simple interpretation of matrix completion problem is introduced based on statistical models.,18372
model and makes electrophysiological mechanism Purkinje The the predictions. learning accurately simulates cell testable,The model accurately simulates the Purkinje cell learning mechanism and makes testable electrophysiological predictions.,18373
among very makes challenging (ReID) multiple person cameras task. a Re-Identification This,This makes person Re-Identification (ReID) among multiple cameras a very challenging task.,18374
"a training. framework three-stage involves this Specifically,","Specifically, this framework involves a three-stage training.",18375
"namely predicted attributes} The exhibit attributes, different across \emph{deep datasets. generalization superior ability","The predicted attributes, namely \emph{deep attributes} exhibit superior generalization ability across different datasets.",18376
that first of We review how describes clustering behaviour. user we trajectories obtained,We first review how we obtained clustering of trajectories that describes user behaviour.,18377
"predict Secondly, we trajectories destination. final the to characteristics use from these clusters inside","Secondly, we use characteristics from trajectories inside these clusters to predict the final destination.",18378
"and orientation. depends critically both both the by particle cell on shape processes, engulfment In","In both processes, engulfment by the cell depends critically on both particle shape and orientation.",18379
of treewidth complexity the the reduces Bayesian inferences. a of greatly Bounding,Bounding the treewidth of a Bayesian greatly reduces the complexity of inferences.,18380
"a in classification the we this pipeline proposed images. image-level novel for In hyperspectral paper,","In this paper, we proposed a novel pipeline for image-level classification in the hyperspectral images.",18381
explored images. feature hyperspectral potential also descriptors the We in of the traditional,We also explored the potential of traditional feature descriptors in the hyperspectral images.,18382
"micro-fluidic applications. tracking and ecological, many is common Particle biophysical, in","Particle tracking is common in many biophysical, ecological, and micro-fluidic applications.",18383
symmetry of advantage the accuracy. of for the spatial algorithm The particle allowing subpixel takes,The algorithm takes advantage of the spatial symmetry of the particle allowing for subpixel accuracy.,18384
with experimental both compare proposed We algorithm synthetic and methods the four different using datasets.,We compare the proposed algorithm with four different methods using both synthetic and experimental datasets.,18385
"of addressing outlier issue extensions, robustness. the also discuss We","We also discuss extensions, addressing the issue of outlier robustness.",18386
with CNNs are limited. representations feature is training when hard Learning data powerful,Learning powerful feature representations with CNNs is hard when training data are limited.,18387
multiscale classical explicitly CNNs. analysis This connects image with approach general,This approach explicitly connects classical multiscale image analysis with general CNNs.,18388
genetic gene flow. under Genome-wide mechanisms divergence patterns of reveal adaptation of,Genome-wide patterns of genetic divergence reveal mechanisms of adaptation under gene flow.,18389
narrow genomic mostly divergence concentrated Empirical data is that regions. show in,Empirical data show that divergence is mostly concentrated in narrow genomic regions.,18390
jointly the Our metrics. learns constrained proposed (CNNs) temporally method and convolutional networks neural,Our proposed method jointly learns the convolutional neural networks (CNNs) and temporally constrained metrics.,18391
"the By method, tracklet challenging can proposed even in association employing accomplished situations. be","By employing the proposed method, tracklet association can be accomplished even in challenging situations.",18392
to used chromosomes. examine aegypti the new conservation We of A. assembly the,We used the new assembly to examine the conservation of A. aegypti chromosomes.,18393
Ae. between strongly that Our synteny is suggest results conserved,Our results suggest that synteny is strongly conserved between Ae.,18394
The data complete to available statistical public. the are and freely per-model dataset,The complete dataset and per-model statistical data are freely available to the public.,18395
similar. different Convolutional the Often from appear learned Networks (CNNs) datasets by Neural filters,Often the filters learned by Convolutional Neural Networks (CNNs) from different datasets appear similar.,18396
layers. in prominent few This the first is,This is prominent in the first few layers.,18397
gained have potential more goals. a as Binary descriptors vehicle to achieve these recently attention,Binary descriptors have recently gained more attention as a potential vehicle to achieve these goals.,18398
texture-based is also a information. to transform transform extract Gabor powerful,Gabor transform is also a powerful transform to extract texture-based information.,18399
"novel a introduces as Barcodes Gabor for framework paper the This (GBCs), annotation. image","This paper introduces Gabor Barcodes (GBCs), as a novel framework for the image annotation.",18400
as modelled extraction reciprocal for (HRC). develops Bayesian hidden chains paper This algorithms targets track,This paper develops Bayesian track extraction algorithms for targets modelled as hidden reciprocal chains (HRC).,18401
"learn heartbeat individual. to approach this of for ""concept"" goal an a personalized The is","The goal of this approach is to learn a personalized heartbeat ""concept"" for an individual.",18402
in and are two jointly connected end-to-end an The parts trained fashion.,The two parts are connected and jointly trained in an end-to-end fashion.,18403
activity. observed unrealistic neuronal modes of The is two in presence view of,The presence of two modes is unrealistic in view of observed neuronal activity.,18404
mode neurobiological high-activity The a grounds. is on unrealistic of prediction,The prediction of a high-activity mode is unrealistic on neurobiological grounds.,18405
asymmetric a bimodality inhibition. avoids This model thanks to minimal,This model avoids bimodality thanks to a minimal asymmetric inhibition.,18406
statistical approach likelihood of by Estimation penalized a has based advantage provided directly the inference.,Estimation by a penalized likelihood based approach has the advantage of directly provided statistical inference.,18407
the natural fluorescence of interest widespread an image. There materials estimating properties in in is,There is widespread interest in estimating the fluorescence properties of natural materials in an image.,18408
provide solver and our prior methods. We the implementation a for software method of,We provide a software implementation of the solver for our method and prior methods.,18409
both data. method simulations and reliability of evaluate We using accuracy the and the experimental,We evaluate the accuracy and reliability of the method using both simulations and experimental data.,18410
algorithms. lower approach earlier compared Our produces errors to,Our approach produces lower errors compared to earlier algorithms.,18411
upon structure shift. ele- (RNATs) thermometers RNA secondary are cis-regulatory that Motivation: temperature ments change,Motivation: RNA thermometers (RNATs) are cis-regulatory ele- ments that change secondary structure upon temperature shift.,18412
software The in C++ OR-Tools search and written engine. is uses CP,The software is written in C++ and uses OR-Tools CP search engine.,18413
have signals. to agents study avoid aversive a We framework where,We study a framework where agents have to avoid aversive signals.,18414
implement. to and parameter-free easy is A-learning,A-learning is parameter-free and easy to implement.,18415
with correction color compared least-square. delivers fidelity linear improved the Our homography application color,Our homography color correction application delivers improved color fidelity compared with the linear least-square.,18416
trained Neural to then Term Long-Short Memory-based A classify series. is time each Recurrent Network,A Long-Short Term Memory-based Recurrent Neural Network is then trained to classify each time series.,18417
"a We differentiation. treatment ABtree, fast effect method, tree-based for computationally present","We present a computationally fast tree-based method, ABtree, for treatment effect differentiation.",18418
Local Resource-bounded computation; Binary inference; Generic algebras; Anytime Approximation; join algorithms; computation; Valuation Keywords: trees.,Keywords: Approximation; Anytime algorithms; Resource-bounded computation; Generic inference; Valuation algebras; Local computation; Binary join trees.,18419
localization sparse action spatio-temporal We supervision. approach human an for using spatial introduce,We introduce an approach for spatio-temporal human action localization using sparse spatial supervision.,18420
a of can as time pipeline algorithms. implemented be It polynomial,It can be implemented as a pipeline of polynomial time algorithms.,18421
"practice. are in easily However, these conditions violated","However, these conditions are easily violated in practice.",18422
confirm of on approach. challenging a variety the images of efficiency Experiments the,Experiments on a variety of challenging images confirm the efficiency of the approach.,18423
"(e.g. include, as a subclass, proposals These discretized Langevin diffusion","These proposals include, as a subclass, discretized Langevin diffusion (e.g.",18424
enabled dynamics of a DNA proviral and origin analysis of detailed ART. during This the,This enabled a detailed analysis of the dynamics and origin of proviral DNA during ART.,18425
are real-world both supported studies Our data. simulated and examples by with experiments,Our studies are supported by experiments with both simulated examples and real-world data.,18426
consumer-grade additional without webcams can tracking for hardware. Even any low-cost gaze used be,Even low-cost consumer-grade webcams can be used for gaze tracking without any additional hardware.,18427
proposed centre iris is localization. two-stage algorithm for A,A two-stage algorithm is proposed for iris centre localization.,18428
of the proposed eye. method characteristics geometrical uses The,The proposed method uses geometrical characteristics of the eye.,18429
Spectral-spatial classification. explored image hyperspectral in increasingly been processing has remote sensing,Spectral-spatial processing has been increasingly explored in remote sensing hyperspectral image classification.,18430
"the non-linear of remains unresolved. case structures However,","However, the case of non-linear structures remains unresolved.",18431
"images images one conditions, in suffer low from captures low-light often When visibility. the","When one captures images in low-light conditions, the images often suffer from low visibility.",18432
"can Having the enhancement illumination map, accordingly. well-constructed the achieved be","Having the well-constructed illumination map, the enhancement can be achieved accordingly.",18433
biometric human are passive technologies for recognition. body Gait measurement static and important,Gait and static body measurement are important biometric technologies for passive human recognition.,18434
is feature the argue based completely on works gait previous that recognition Many performance limited.,Many previous works argue that recognition performance based completely on the gait feature is limited.,18435
remains for The this reason unclear. performance limited,The reason for this limited performance remains unclear.,18436
"games available. in some However, forward is readily not such model","However, in some games such forward model is not readily available.",18437
these data. how also combat can show learned from We be replay models,We also show how these combat models can be learned from replay data.,18438
domain. our use StarCraft We as application,We use StarCraft as our application domain.,18439
statistics rate of Heterogeneity coding. neural known to on severe is firing have consequences,Heterogeneity of firing rate statistics is known to have severe consequences on neural coding.,18440
use to We architecture. show also can effective neural how our theory predict be,We also show how our theory can be use to predict effective neural architecture.,18441
Estimation tasks. human-computer direction eye is various gaze useful in interaction,Estimation eye gaze direction is useful in various human-computer interaction tasks.,18442
point regarding of Knowledge valuable information can attention. give direction of gaze users,Knowledge of gaze direction can give valuable information regarding users point of attention.,18443
geometric A using relations obtained landmarks. is eye region and facial rough,A rough eye region is obtained using geometric relations and facial landmarks.,18444
phase. less complexity computational testing the very in of The the is algorithm,The computational complexity of the algorithm is very less in the testing phase.,18445
"development, patterns arrangement the gene determine tissue types. spatial During expression cell of of","During tissue development, patterns of gene expression determine the spatial arrangement of cell types.",18446
"process. secreted signaling many gradients morphogens - cases, of In guide molecules - this","In many cases, gradients of secreted signaling molecules - morphogens - guide this process.",18447
steady the these state mechanisms. Previous emphasize output of analyses,Previous analyses emphasize the steady state output of these mechanisms.,18448
of explain This observed experimentally dynamics could pattern formation.,This could explain experimentally observed dynamics of pattern formation.,18449
"be to constant. are Often, be the to in parameters such estimated models assumed","Often, in such models the parameters to be estimated are assumed to be constant.",18450
to This equivalent extended algorithm Kalman to out turns the filter. be,This algorithm turns out to be equivalent to the extended Kalman filter.,18451
"momentum. the time, is same At imagery gaining aerial","At the same time, aerial imagery is gaining momentum.",18452
objects. recognizing way introduce a as of context complementary Here we,Here we introduce context as a complementary way of recognizing objects.,18453
are in final later The combined layers of two the processing.,The two are later combined in the final layers of processing.,18454
many usually performance. detection to satisfactory labeled methods Supervised require obtain images contour training,Supervised contour detection methods usually require many labeled training images to obtain satisfactory performance.,18455
"However, might data large be unavailable labor of set extremely intensive. annotated or a","However, a large set of annotated data might be unavailable or extremely labor intensive.",18456
have task captioning increasing Recent in in image advances led to captioning task. video interests,Recent advances in image captioning task have led to increasing interests in video captioning task.,18457
Our results theoretical by numerical are simulations. supported,Our theoretical results are supported by numerical simulations.,18458
language to neural limitations i.e. by models scratch; from training propose We these overcome,We propose to overcome these limitations by training neural language models from scratch; i.e.,18459
and characters. consuming pre-training only and words without,without pre-training and only consuming words and characters.,18460
train of to content with end-to-end and images. the align models category-specific fine-grained proposed Our,Our proposed models train end-to-end to align with the fine-grained and category-specific content of images.,18461
the be gap DNA The mechanism. in repair filled will dsDNA through,The gap in the dsDNA will be filled through DNA repair mechanism.,18462
"of cycle. could cutout-repair-cutout rounds Occasionally, multiple multiple transcripts ssDNA through of copies be produced","Occasionally, multiple copies of ssDNA transcripts could be produced through multiple rounds of cutout-repair-cutout cycle.",18463
computationally real-time method Our is offers and cheap proposed description.,Our proposed method is computationally cheap and offers real-time description.,18464
the using superpixels. investigates image paper segmentation of problem This,This paper investigates the problem of image segmentation using superpixels.,18465
enhance to superpixel's approaches of descriptors. discriminative the ability two the covariance We propose,We propose two approaches to enhance the discriminative ability of the superpixel's covariance descriptors.,18466
from transformations works showed can invariance Several recent to complex be learned shape examples. that,Several recent works showed that invariance to complex shape transformations can be learned from examples.,18467
different far. large number detectors has proposed of A been so feature,A large number of different feature detectors has been proposed so far.,18468
been far. feature proposed large has number different %A so detectors of,%A large number of different feature detectors has been proposed so far.,18469
"general-purpose, biologically-plausible present a of neural new Here implementation approximate we inference.","Here we present a new general-purpose, biologically-plausible neural implementation of approximate inference.",18470
"a multivariate graphical PPCs multiple By together represent connecting probability distributions. model, as we probabilistic","By connecting multiple PPCs together as a probabilistic graphical model, we represent multivariate probability distributions.",18471
"subtle it Unfortunately, a LBP implausible. feature neurally renders of","Unfortunately, a subtle feature of LBP renders it neurally implausible.",18472
An results is and presented. given system promising are the design of overview,An overview of the system design is given and promising results are presented.,18473
that categorisation introduced. video fine-grained designed is pedestrians A of study is to dataset,A video dataset that is designed to study fine-grained categorisation of pedestrians is introduced.,18474
"from moving ""in-the-wild"" vehicle. Pedestrians recorded a were","Pedestrians were recorded ""in-the-wild"" from a moving vehicle.",18475
variational allows using to expectation-maximization of a algorithm. path a It models find,It allows to find a path of models using a variational expectation-maximization algorithm.,18476
over is maximized The path. marginal likelihood this then exact,The exact marginal likelihood is then maximized over this path.,18477
and illustrated data is synthetic real sets. on approach This,This approach is illustrated on real and synthetic data sets.,18478
other properties We the compare also of models. six,We also compare other properties of the six models.,18479
"reveals In the addition, population-wise probability. parameters mutation between the difference of the correlation","In addition, correlation between the parameters reveals the population-wise difference of the mutation probability.",18480
against non-independence support hypothesis These mutational mutation. random results the,These results support the mutational non-independence hypothesis against random mutation.,18481
undirected data. from We consider graphical of models discovery observational structure,We consider structure discovery of undirected graphical models from observational data.,18482
precision likelihood a matrix. estimating methods the on of maximum penalized rely Popular,Popular methods rely on estimating a penalized maximum likelihood of the precision matrix.,18483
"In our problems to experiments, detection we tracking. and algorithm pedestrian of apply the the","In the experiments, we apply our algorithm to the problems of pedestrian detection and tracking.",18484
We how select and show to platforms algorithm-parameter adaptively combinations.,We show how to adaptively select platforms and algorithm-parameter combinations.,18485
analysis of of efficiency increasing is the method process. the Multiresolution segmentation a,Multiresolution analysis is a method of increasing the efficiency of the segmentation process.,18486
selected is criteria. The starting resolution with selection clear generally no for arbitrarily segmentation,The starting resolution for segmentation is generally selected arbitrarily with no clear selection criteria.,18487
be an for thus method selection for automated beneficial. An would image resolution input,An automated method for resolution selection for an input image would thus be beneficial.,18488
Cluster to score various was to develop taxonomies. analysis similarity the space applied,Cluster analysis was applied to the similarity score space to develop various taxonomies.,18489
"subject, and illumination supports by of existence the condition. gender, Evidence race, clustering","Evidence supports the existence of clustering by race, gender, subject, and illumination condition.",18490
observed an The ancestry. Amerindian has lineage DNA,The observed DNA lineage has an Amerindian ancestry.,18491
conclusion elite idea support the current do not marriages. This interethnic the Spanish avoided that,This conclusion do not support the current idea that the Spanish elite avoided interethnic marriages.,18492
yield visual powerful hierarchies that models of Convolutional features. are networks,Convolutional networks are powerful visual models that yield hierarchies of features.,18493
from is inspired in cell the approach the representation Our visual double-opponent cortex. human,Our approach is inspired from the double-opponent cell representation in the human visual cortex.,18494
"In traditional image elaborate applications we deblurring restoration experiments, on denoising. inpainting, of and","In experiments, we elaborate on traditional image restoration applications of inpainting, deblurring and denoising.",18495
computational has The a is low algorithm efficient cost. memory and also,The algorithm is also memory efficient and has a low computational cost.,18496
person-specific to require facial system does and Our capture is not able training. expressions any,Our system is able to capture facial expressions and does not require any person-specific training.,18497
is in vision. segmentation Superpixel ubiquitous becoming computer,Superpixel segmentation is becoming ubiquitous in computer vision.,18498
"Unfortunately, there generate scales accurately no all that is of in can superpixels real-time. method","Unfortunately, there is no method that can generate all scales of superpixels accurately in real-time.",18499
"convolutional object networks accurate We detection. present efficient region-based, for fully and","We present region-based, fully convolutional networks for accurate and efficient object detection.",18500
fine-grained to in find represent recognition local key how A and discriminative challenge is regions.,A key challenge in fine-grained recognition is how to find and represent discriminative local regions.,18501
Contour skeleton two recognition. are and representations complementary shape for,Contour and skeleton are two complementary representations for shape recognition.,18502
contour. with information a The correspondence a shape associate provides way to skeletal straightforward,The correspondence provides a straightforward way to associate skeletal information with a shape contour.,18503
"More descriptor. a propose we specifically, new shape","More specifically, we propose a new shape descriptor.",18504
ancestry. corresponds lineage to American The Native MtDNA a observed,The MtDNA lineage observed corresponds to a Native American ancestry.,18505
this in miscegenation Rica. Examples Costa early one that began confirm like,Examples like this one confirm that miscegenation began early in Costa Rica.,18506
"So, themselves actually colonial the mestizos. ""Spaniards"" considered were era, those late biologically who in","So, those who considered themselves ""Spaniards"" in the late colonial era, were actually biologically mestizos.",18507
This to containing in has collections mining image applications related matching content. handwritten and,This has applications related to matching and mining in image collections containing handwritten content.,18508
method segmentation does accurate on depend of and not an words Our lines.,Our method does not depend on an accurate segmentation of words and lines.,18509
"convolutional To neural a we two based descriptor. (CNN) feature network match propose word images,","To match two word images, we propose a convolutional neural network (CNN) based feature descriptor.",18510
state-of-the-art this handwritten on spotting. surpasses word of representation the Performance,Performance of this representation surpasses the state-of-the-art on handwritten word spotting.,18511
scores the quality state to of Observed art. results show the comparable,Observed results show quality scores comparable to the state of the art.,18512
the cases. particular phenomenological models presented reachs model microscopic This as,This microscopic model reachs the phenomenological models presented as particular cases.,18513
kind field model of That is mean means model. Verhulst a,That means Verhulst model is a kind of mean field model.,18514
four matrilineal from is A presented genealogy Da. generations,A four generations matrilineal genealogy is presented from Da.,18515
"recent This consistent expected, European with is -as lineage. the documented","This is consistent -as expected, with the recent documented European lineage.",18516
the These parameter and from differences predictions same different data. yield may estimates,These differences may yield different predictions and parameter estimates from the same data.,18517
images. fine-grained object address resolution low categories the We of in distinguishing difficult problem,We address the difficult problem of distinguishing fine-grained object categories in low resolution images.,18518
perceive and that humans hierarchically. Psychological anatomical visual have concepts shown studies,Psychological and anatomical studies have shown that humans perceive visual concepts hierarchically.,18519
"ranking, image We retargeting, features and the aesthetics deep apply on retrieval. learned","We apply the learned deep features on image retargeting, aesthetics ranking, and retrieval.",18520
approach. demonstrate subjective our experimental and results of the objective Both competitiveness thoroughly,Both subjective and objective experimental results thoroughly demonstrate the competitiveness of our approach.,18521
segmentation are debate of subject image Quality algorithms of evaluation still and research.,Quality evaluation of image segmentation algorithms are still subject of debate and research.,18522
"metric could there algorithm applied generic that be any is no to reliably. Currently,","Currently, there is no generic metric that could be applied to any algorithm reliably.",18523
are exchanges markets donors. swap organized but willing Kidney patients incompatible where,Kidney exchanges are organized markets where patients swap willing but incompatible donors.,18524
"The indeed, that, small attributes suffice. show of experiments numbers","The experiments show that, indeed, small numbers of attributes suffice.",18525
"(i) approach a First, to instance semantic propose simple we effective segmentation. yet","(i) First, we propose a simple yet effective approach to semantic instance segmentation.",18526
also (iv) segmentation. for achieve state-of-the-art results We instance-level,(iv) We also achieve state-of-the-art results for instance-level segmentation.,18527
representation a dimensional problem. SIFT) for to retrieval image useful the higher,SIFT) to a higher dimensional representation useful for the image retrieval problem.,18528
This patterns. requires fate model about our of extrapolate boundaries motion the learn to and,This requires our model to learn about the fate of boundaries and extrapolate motion patterns.,18529
for boundary spatio-temporal challenging show time this scenario. extrapolation the in first We,We show for the first time spatio-temporal boundary extrapolation in this challenging scenario.,18530
(E-pulses). of pulses the vortices We all in with phases pinned parallel scan field electric,We scan the phases of all pinned vortices in parallel with electric field pulses (E-pulses).,18531
a termination specify that pacing guarantees vortex. parameters one on condition We of,We specify a condition on pacing parameters that guarantees termination of one vortex.,18532
a similar pinned) (not found free a that vortex. also terminates We mechanism,We found that a similar mechanism terminates also a free (not pinned) vortex.,18533
"extract of aspects data. In applications, is desirable to it only relevant the many","In many applications, it is desirable to extract only the relevant aspects of data.",18534
"computationally method however, the data is non-gaussian. IB when Unfortunately and/or high-dimensional are demanding","Unfortunately however, the IB method is computationally demanding when data are high-dimensional and/or non-gaussian.",18535
"learning nets. approach, a novel present deeply-fused we paper, this In deep","In this paper, we present a novel deep learning approach, deeply-fused nets.",18536
resulting The deeply benefits. enjoys fused several net,The resulting deeply fused net enjoys several benefits.,18537
as Quickest derived problem method of The Change Detection. is a,The method is derived as a problem of Quickest Change Detection.,18538
"compare by network approaches their proteins Then, can PSNs. comparing","Then, network approaches can compare proteins by comparing their PSNs.",18539
interesting. may approach PSN patterns biochemically be that finds Our,Our approach finds PSN patterns that may be biochemically interesting.,18540
an of interactions. social part integral dilemmas Social are,Social dilemmas are an integral part of social interactions.,18541
microbial tractable combined became experimentally recently dilemmas. in processes social Only eco-evolutionary,Only recently combined eco-evolutionary processes became experimentally tractable in microbial social dilemmas.,18542
"maintained and densities often at cooperation ecological particular, is In through supported high fluctuations.","In particular, cooperation is often supported and maintained at high densities through ecological fluctuations.",18543
methods Most on current rely fully semantic convolutional networks segmentation (FCNs).,Most current semantic segmentation methods rely on fully convolutional networks (FCNs).,18544
the localization to predictions This poor around with leads boundaries.,This leads to predictions with poor localization around the boundaries.,18545
to attempted this CRFs issue address predictions post-processing with has by Prior MRFs. or work,Prior work has attempted to address this issue by post-processing predictions with CRFs or MRFs.,18546
"these recent an methods into MRFs framework. integrated overcome FCN problems, CRFs or To","To overcome these problems, recent methods integrated CRFs or MRFs into an FCN framework.",18547
pixelwise Our jointly and the affinity of proposed semantic optimizes objectives RWN segmentation.,Our proposed RWN jointly optimizes the objectives of pixelwise affinity and semantic segmentation.,18548
using multiplication. RWN matrix is Our and standard convolution implemented,Our RWN is implemented using standard convolution and matrix multiplication.,18549
phenotypic to the maintain complexity. increases ability genomic significantly alone Random already mating and,Random mating alone already significantly increases the ability to maintain genomic and phenotypic complexity.,18550
provides capacity for complexity. additional selection Sexual this,Sexual selection provides additional capacity for this complexity.,18551
a be trained scenario alone. using object annotation weakly-supervised detectors image-level to In need,In a weakly-supervised scenario object detectors need to be trained using image-level annotation alone.,18552
select a the dictionary size. can appropriate simultaneously method The and learn sparse proposed dictionary,The proposed method can simultaneously learn a sparse dictionary and select the appropriate dictionary size.,18553
fluoroquinolones bacterial Antibiotics quinolones DNA as pathogens through ultimately damage. kill such the and,Antibiotics such as the quinolones and fluoroquinolones kill bacterial pathogens ultimately through DNA damage.,18554
their discuss of the Here development and action. these of method antibiotics we,Here we discuss the development of these antibiotics and their method of action.,18555
are context in the structures. dynamics studied game of different often population Evolutionary,Evolutionary game dynamics are often studied in the context of different population structures.,18556
multicellular structure by life population inspired simple is that forms. new propose we Here a,Here we propose a new population structure that is inspired by simple multicellular life forms.,18557
"In cells but after our reproduction. can together model, reproduce stay","In our model, cells reproduce but can stay together after reproduction.",18558
"size, again. n, complexes of single reach before cells producing certain a They","They reach complexes of a certain size, n, before producing single cells again.",18559
proportional payoff. cells rate reproductive The is of to their,The reproductive rate of cells is proportional to their payoff.,18560
examples As study and we dilemma specific prisoner's hawk-dove games. the,As specific examples we study the prisoner's dilemma and hawk-dove games.,18561
within model advances work on multicellularity Our by allowing interactions theoretical for frequency-dependent groups.,Our model advances theoretical work on multicellularity by allowing for frequency-dependent interactions within groups.,18562
introduce macro-architecture strategy on We neural for based network a self-similarity. design,We introduce a design strategy for neural network macro-architecture based on self-similarity.,18563
regularization fixed-depth extraction allows subnetworks. Such high-performance of,Such regularization allows extraction of high-performance fixed-depth subnetworks.,18564
"other. however, with each These contradict two properties,","These two properties, however, contradict with each other.",18565
the or of generators Random component metaheuristic in key algorithms. structure engines are stochastic a,Random generators or stochastic engines are a key component in the structure of metaheuristic algorithms.,18566
distributions. are and the non-Gaussian The Weibull distributions Gamma two,The two non-Gaussian distributions are the Weibull and Gamma distributions.,18567
practical The implications are results in nature simulations. demonstrated of and the,The nature and practical implications of the results are demonstrated in simulations.,18568
"state-of-the-art evaluate First, problem. search on this we instance methods","First, we evaluate state-of-the-art instance search methods on this problem.",18569
works buildings for that what on shoes. generality We its loses observe,We observe that what works for buildings loses its generality on shoes.,18570
"of re-identification special Third, as we a search. treat instance generic case person","Third, we treat person re-identification as a special case of generic instance search.",18571
"with we popular performance same state-of-the-art method. the On VIPeR the reach dataset,","On the popular VIPeR dataset, we reach state-of-the-art performance with the same method.",18572
"extend to category. restriction we specifically search method the to Fourth, objects our known without","Fourth, we extend our method to search objects without restriction to the specifically known category.",18573
Lung for number cancer globally. the of highest  accounts deaths cancer ,  Lung cancer accounts for the highest number of cancer deaths globally.,18574
sampling the random a stable subspaces using Our scheme. method starts of by computation novel,Our method starts by the computation of stable subspaces using a novel random sampling scheme.,18575
residual many performance appealing challenging tasks. on computer shown have vision Deep recently networks,Deep residual networks have recently shown appealing performance on many challenging computer vision tasks.,18576
stochastic are by descent. optimized networks gradient The projected,The networks are optimized by projected stochastic gradient descent.,18577
"five language, memory, perception, describes intelligence. cognitive Fuster attention, and functions:","Fuster describes five cognitive functions: perception, memory, attention, language, and intelligence.",18578
as in cognitive is between and well. A cognition given human comparison memory radar,A comparison between memory in human cognition and cognitive radar is given as well.,18579
each We in video. large a proposals action of number obtain first,We first obtain a large number of action proposals in each video.,18580
the that We this preserves action ranking high demonstrate quality proposals.,We demonstrate that this ranking preserves the high quality action proposals.,18581
each for video proposals same action. Several of the generated such are,Several such proposals are generated for each video of the same action.,18582
video. each The output the of most method action proposals representative is from our,The output of our method is the most action representative proposals from each video.,18583
method annotate instances can of same a video. multiple Our action also the in,Our method can also annotate multiple instances of the same action in a video.,18584
borne The susceptible vectors criss-cross dynamics. disease through is to hosts by,The disease is borne by vectors to susceptible hosts through criss-cross dynamics.,18585
equations differential compartments. to Partial formulated interaction describe the are the Methods: model of,Methods: Partial differential equations are formulated to describe the interaction of the model compartments.,18586
are be proven analyzed of well-posed. The partial the differential model equations and Results: to,Results: The partial differential equations of the model are analyzed and proven to be well-posed.,18587
one of descriptor Dominant is them. The color,The Dominant color descriptor is one of them.,18588
can extract the method dominant pixel proposed successfully Results that values. color show,Results show that the proposed method can successfully extract dominant color pixel values.,18589
that on effective selection the Previous selection studies higher-level lower-level multilevel emerges from assumed reproduction.,Previous studies on multilevel selection assumed that the effective higher-level selection emerges from lower-level reproduction.,18590
"leads although This selection among only to reproduce. individuals groups,","This leads to selection among groups, although only individuals reproduce.",18591
substantially ratio for When accounting critical lowered. the this is benefit-to-cost,When accounting for this the critical benefit-to-cost ratio is substantially lowered.,18592
technology. recognition novel is Palm a vein biometric identification,Palm vein recognition is a novel biometric identification technology.,18593
a multispectral We image. palm method on test our,We test our method on a multispectral palm image.,18594
robustifying We of structured problem consider the high-dimensional estimation.,We consider the problem of robustifying high-dimensional structured estimation.,18595
which often involve in corruption. applications real-world key Robust data are outliers techniques and,Robust techniques are key in real-world applications which often involve outliers and data corruption.,18596
"transfer paper, problem with multi-instance In learning the we this data. invest domain","In this paper, we invest the domain transfer learning problem with multi-instance data.",18597
cannot But used to domain. be target the it directly,But it cannot be directly used to the target domain.,18598
transfer a adaptive The multi-instance function linear dictionary. a domain is function based,The adaptive function is a linear function based a domain transfer multi-instance dictionary.,18599
fitness simple dynamic approximated a The Markov is population's by stochastic well long-term chain model.,The population's long-term fitness dynamic is well approximated by a simple stochastic Markov chain model.,18600
species consistent times fossil typical years. persistence of persistence times several Long-term are with million,Long-term persistence times are consistent with typical fossil species persistence times of several million years.,18601
"over risk long later population time. Should the constant extinction enough, survive becomes","Should the population survive long enough, extinction risk later becomes constant over time.",18602
"context coarse-graining the propose equilibrium of data-driven, formulation mechanics. in statistical a We","We propose a data-driven, coarse-graining formulation in the context of equilibrium statistical mechanics.",18603
This specific filters sub-events to learn activity. latent temporal allows attention each to our,This allows our temporal attention filters to learn latent sub-events specific to each activity.,18604
choice The is of for passenger understanding public route management. transit critical,The understanding of passenger route choice is critical for public transit management.,18605
Automated Fare wide deployment systems new opens a Collection(AFC) The of up opportunity.,The wide deployment of Automated Fare Collection(AFC) systems opens up a new opportunity.,18606
"well situations. don't methods work in situations, specific complicated existing for they While some work","While existing methods work well in some specific situations, they don't work for complicated situations.",18607
measured inputs when provide results model. passenger path with building the The us useful choice,The measured results provide us with useful inputs when building the passenger path choice model.,18608
of Stochastic biodemographic Process longitudinal analysis data. in Model many applications has,Stochastic Process Model has many applications in analysis of longitudinal biodemographic data.,18609
contain physiological known as covariates). variables data various (sometimes Such,Such data contain various physiological variables (sometimes known as covariates).,18610
potentially of participants. genetic all can for or contain information part also It available a,It also can potentially contain genetic information available for all or a part of participants.,18611
in video important play an Key annotation. role frames,Key frames play an important role in video annotation.,18612
paper entropy novel for In using proposed. value is this a key-frame approach extraction,In this paper a novel approach for key-frame extraction using entropy value is proposed.,18613
selected eliminates entropy feature. It using redundant from also frames key-frames value as local,It also eliminates redundant frames from selected key-frames using entropy value as local feature.,18614
clips has been approach presented. Evaluation of the several video on,Evaluation of the approach on several video clips has been presented.,18615
show automatically that is helping annotators algorithm video key-frames. in identify Results the successful,Results show that the algorithm is successful in helping annotators automatically identify video key-frames.,18616
without shape explicit requiring localizes model. parts accurately Our body an method,Our method accurately localizes body parts without requiring an explicit shape model.,18617
multiple body by in evidence The recovered views are from real-time. then locations combining joint,The body joint locations are then recovered by combining evidence from multiple views in real-time.,18618
solution Standard no report that return inaccurate may exists. solvers double-precision solutions or,Standard double-precision solvers may return inaccurate solutions or report that no solution exists.,18619
"not labeled for some categories. In may available problems, all data recognition object be of","In some of object recognition problems, labeled data may not be available for all categories.",18620
protocol incorporates extraction in metric advances the recent learning. The both most experimental feature and,The experimental protocol incorporates the most recent advances in both feature extraction and metric learning.,18621
community The publicly results evaluation will and use. available for made be codebase,The evaluation codebase and results will be made publicly available for community use.,18622
composition among technique those on The methods which rely. is indispensable often triangle professional photographers,The triangle technique is among those indispensable composition methods on which professional photographers often rely.,18623
"carefully positioning a examine of For human photo. the portraits, in subjects we","For portraits, we carefully examine the positioning of human subjects in a photo.",18624
highly that line modeling categories. is advantageous for We show both analysis composition in,We show that line analysis is highly advantageous for modeling composition in both categories.,18625
"we In to approach quantile-based propose this problems. RBDO a paper, solve","In this paper, we propose a quantile-based approach to solve RBDO problems.",18626
"design. degree In quantile the level formulation, conservatism this the controls of the of","In this formulation, the quantile level controls the degree of conservatism of the design.",18627
such optimized that multi-layer not spaces modeling. suggests feature Our imperative are finding texture for,Our finding suggests that such optimized multi-layer feature spaces are not imperative for texture modeling.,18628
regularization. that important An our is structure-promoting it easily handle design can benefit of,An important benefit of our design is that it can easily handle structure-promoting regularization.,18629
"will databases multiple the test In paper, be networks. to this used","In this paper, multiple databases will be used to test the networks.",18630
majority interactions cellular The understanding crucial processes. of protein-protein (PPIs) are the for,The protein-protein interactions (PPIs) are crucial for understanding the majority of cellular processes.,18631
the complexes of relevant unknown. biologically structure many For of protein-protein cases remain,For many of biologically relevant cases the structure of protein-protein complexes remain unknown.,18632
"Therefore pipelines. have computational molecular part become of docking, valuable drug a including techniques, discovery","Therefore computational techniques, including molecular docking, have become a valuable part of drug discovery pipelines.",18633
"serious none widely Unfortunately, the tools limitations. docking is free from protein-protein used of","Unfortunately, none of the widely used protein-protein docking tools is free from serious limitations.",18634
"the or limited. protein flexibility neglected docking completely simulations is either very in Typically,","Typically, in docking simulations the protein flexibility is either completely neglected or very limited.",18635
mostly of number the of arise enormous freedom degrees limitations protein-protein of from Such systems.,Such limitations arise mostly from the enormous number of degrees of freedom of protein-protein systems.,18636
procedure. two-step The proposed is on method based a,The proposed method is based on a two-step procedure.,18637
"are deal with these analysis To massive sequence elaborated employed. pipelines data,","To deal with these massive data, elaborated sequence analysis pipelines are employed.",18638
mapping We alignment accuracy. propose read that significantly approach read improves a dynamic,We propose a dynamic read mapping approach that significantly improves read alignment accuracy.,18639
is capable OCOCO input fashion. caller the reads in first process the consensus to online,OCOCO is the first consensus caller capable to process input reads in the online fashion.,18640
theory analysis. the distribution approach Our coexistence modern between species large-scale and bridges data gap,Our approach bridges the gap between modern coexistence theory and large-scale species distribution data analysis.,18641
"of we problem In the prediction intents. the of new address paper, human this","In this paper, we address the new problem of the prediction of human intents.",18642
trained the models Are valid in reality?,Are the trained models valid in reality?,18643
bias? to the ways What performance are possible reduce,What are possible ways to reduce the performance bias?,18644
This stimulation. the important of applications growing adjusted will dynamically potential given be neural,This will be important given the growing potential applications of dynamically adjusted neural stimulation.,18645
objects of recognizing from input. the address We video task,We address the task of recognizing objects from video input.,18646
"important compared unexplored, This relatively with object problem recognition. image-based is","This important problem is relatively unexplored, compared with image-based object recognition.",18647
"contributions. we end, following To make this the","To this end, we make the following contributions.",18648
"comprehensive we object for two video-based datasets First, introduce recognition.","First, we introduce two comprehensive datasets for video-based object recognition.",18649
"(LBSVM), video-based propose we object Latent Second, Bi-constraint for recognition. framework maximum-margin a SVM","Second, we propose Latent Bi-constraint SVM (LBSVM), a maximum-margin framework for video-based object recognition.",18650
substitutions two least rearrangements. single-base DNA Levy have random large and flights components: at These,These Levy flights have at least two components: random single-base substitutions and large DNA rearrangements.,18651
costly. time-consuming is need design these to and pipelines optimize and The,The need to design and optimize these pipelines is time-consuming and costly.,18652
problem. difficult a docking Protein-peptide molecular is modeling,Protein-peptide molecular docking is a difficult modeling problem.,18653
"have statistics RG models responses. predictions However, joint DSP on the and spiking contradicting of","However, DSP and RG models have contradicting predictions on the joint statistics of spiking responses.",18654
modelling that to efficient potential-level stochasticity correlations. argue We membrane provides strategy model of an,We argue that membrane potential-level modelling of stochasticity provides an efficient strategy to model correlations.,18655
"our methods. programming with to Finally, we that show genetic results competitive regard are","Finally, we show that our results are competitive with regard to genetic programming methods.",18656
data-driven possibility using also We to the neural proposals. explore improve networks of,We also explore the possibility of using neural networks to improve data-driven proposals.,18657
tasks. Neural exhibit of performance Networks extraordinary (CNNs) variety Convoulutional learning a on machine,Convoulutional Neural Networks (CNNs) exhibit extraordinary performance on a variety of machine learning tasks.,18658
"their and poorly understood. quite behavior mathematical properties However, are","However, their mathematical properties and behavior are quite poorly understood.",18659
been recognition and face long problem Human has recognition. a pattern in computer standing vision,Human face recognition has been a long standing problem in computer vision and pattern recognition.,18660
the of experiments demonstrate methods. our power over Simulation superior meta-analysis method,Simulation experiments demonstrate the superior power of our method over meta-analysis methods.,18661
high-dimensional on be diagnosis clinical Predictive models a condition. images for can of brain used,Predictive models can be used on high-dimensional brain images for diagnosis of a clinical condition.,18662
any generated desired sequence maintains the globally precision. convergence optimal The speed towards,The generated sequence maintains the optimal convergence speed towards any globally desired precision.,18663
"proximal classical over gradient methods. an is improvement We rate, a convergence which smoothing provide","We provide a convergence rate, which is an improvement over classical proximal gradient smoothing methods.",18664
a study relevance relation particular We in paper. this image-word,We study a particular image-word relevance relation in this paper.,18665
model. arrive We a tagging at versatile quite,We arrive at a quite versatile tagging model.,18666
"set in test the fast time given w.r.t.\ It constant a size. training runs image,","It runs fast given a test image, in constant time w.r.t.\ the training set size.",18667
brain atypical Is shift SFN activity? in this identifiable,Is this SFN shift identifiable in atypical brain activity?,18668
Pareto the hub of signature networks. The a Degree on Distribution scale-free (D) (P(D)~D^-\b{eta}) is,The Pareto Distribution (P(D)~D^-\b{eta}) on the hub Degree (D) is a signature of scale-free networks.,18669
"Degree During atypical subjects. resting-state, and we a assess across exponents large of range neurotypical","During resting-state, we assess Degree exponents across a large range of neurotypical and atypical subjects.",18670
the structure. We of brain complexity network theory provide theory graph to a predictive use,We use graph complexity theory to provide a predictive theory of the brain network structure.,18671
"are limited. sensors often depth resolution increasingly While remains popular, becoming spatial their","While depth sensors are becoming increasingly popular, their spatial resolution often remains limited.",18672
solution emerged as a this Depth to super-resolution therefore problem.,Depth super-resolution therefore emerged as a solution to this problem.,18673
"how employed low-resolution learning Furthermore, in can we be maps depth show strategy. training our","Furthermore, we show how low-resolution training depth maps can be employed in our learning strategy.",18674
"on As runs in real-time images videos. or a it result,","As a result, it runs in real-time on images or videos.",18675
categorization hundreds occurs few process scale. of millisecond and is very This time fast in,This categorization process is very fast and occurs in few hundreds of millisecond time scale.,18676
were this categories immobile visual stimuli and analysed of in mobile. Two study:,Two categories of visual stimuli were analysed in this study: immobile and mobile.,18677
and was (intra-condition) similarity categories assessed Networks category within (inter-condition). between each,Networks similarity was assessed within each category (intra-condition) and between categories (inter-condition).,18678
low the similarity similarity between showed each categories. Results two and within high category,Results showed high similarity within each category and low similarity between the two categories.,18679
"multiple segment Second, atrous robustly pyramid pooling spatial (ASPP) to at we propose scales. objects","Second, we propose atrous spatial pyramid pooling (ASPP) to robustly segment objects at multiple scales.",18680
code available publicly of online. made is All our,All of our code is made publicly available online.,18681
based maximum likelihood estimation efficiently on implemented Kalman principles Fortran. The filtering in is and,The estimation is based on maximum likelihood principles and Kalman filtering efficiently implemented in Fortran.,18682
capabilities of three to provided examples the ctsmr. Finally demonstrate are,Finally three examples are provided to demonstrate the capabilities of ctsmr.,18683
"daughter cells. division partitioned occur, randomly events whenever size among is Moreover,","Moreover, whenever division events occur, size is randomly partitioned among daughter cells.",18684
"partitioning size-based division mechanisms. errors In mean the contrast, with increasing size cell decreases for","In contrast, the mean cell size decreases with increasing partitioning errors for size-based division mechanisms.",18685
"CCA kernel contaminated data. not to robust addition, In classical is","In addition, classical kernel CCA is not robust to contaminated data.",18686
"covariance, discuss robust the operators. element, kernel cross-covariance We, and therefore, mean kernel robust first","We, therefore, first discuss robust kernel mean element, the robust kernel covariance, and cross-covariance operators.",18687
"the methods synthesized Finally, genetic and to set. imaging data data proposed we investigate","Finally, we investigate the proposed methods to synthesized data and imaging genetic data set.",18688
"introduce a we few Finally, emerging topics.","Finally, we introduce a few emerging topics.",18689
R. study the presented medical Stupp by Based on al. et,Based on the medical study presented by R. Stupp et al.,18690
competitive method in proposed The very benchmarks. state-of-the-art the two performance obtains,The proposed method obtains very competitive state-of-the-art performance in the two benchmarks.,18691
said and between order It been complexity disorder. has that lies,It has been said that complexity lies between order and disorder.,18692
LH slow velocities high outcomes. to We that show lead,We show that slow velocities lead to high LH outcomes.,18693
a conditional system of We dealing argumentation the with obligations. detachment present formal general for,We present a general formal argumentation system for dealing with the detachment of conditional obligations.,18694
of document for accurate crucial images recognition. scanning sheets text and paper Digitally unwrapping is,Digitally unwrapping images of paper sheets is crucial for accurate document scanning and text recognition.,18695
"We on including present letters book folded examples shopping results and receipts. several pages,","We present results on several examples including book pages, folded letters and shopping receipts.",18696
important step tasks. is segmentation an most in visual Image,Image segmentation is an important step in most visual tasks.,18697
"to need outliers. Therefore, identified we errors examine transcription of for","Therefore, we need to examine for transcription errors of identified outliers.",18698
"stem-and-leaf examine display. outliers, the To use we the","To examine the outliers, we use the stem-and-leaf display.",18699
"implementation this downloadable saliency we the report, SALICON of In publicly model. present our technical","In this technical report, we present our publicly downloadable implementation of the SALICON saliency model.",18700
"downloadable a model. issue, created we implementation the of address SALICON this publicly To have","To address this issue, we have created a publicly downloadable implementation of the SALICON model.",18701
and data. spatial sensors high of Remote sensing hyperspectral large spectral collect dimensional volumes,Remote sensing hyperspectral sensors collect large volumes of high dimensional spectral and spatial data.,18702
that proposed state-of-the-art method to subspace Experimental well identification compares shows results recent methods. our,Experimental results shows that our proposed method compares well to recent state-of-the-art subspace identification methods.,18703
is predominant customer employing predict artificial NTL. direction to The intelligence research whether a causes,The predominant research direction is employing artificial intelligence to predict whether a customer causes NTL.,18704
of calculus one QSTR notion formalisms. of The prominent most a is qualitative the,The notion of a qualitative calculus is one of the most prominent QSTR formalisms.,18705
"to properties. Moreover, classification we according provide algebraic of calculi a their","Moreover, we provide a classification of calculi according to their algebraic properties.",18706
latent a truncated are out assumed marginal and model. to induce integrated variables The,The truncated variables are assumed latent and integrated out to induce a marginal model.,18707
"blending have blending image little been video Unlike algorithms studied. algorithms,","Unlike image blending algorithms, video blending algorithms have been little studied.",18708
quality provided and both these visual objectively also by The algorithms evaluated subjectively. is,The visual quality provided by these algorithms is also evaluated both objectively and subjectively.,18709
available. publicly algorithm are The implementations and benchmark video,The video benchmark and algorithm implementations are publicly available.,18710
implements tasks the C++. package computationally intensive The in,The package implements the computationally intensive tasks in C++.,18711
large scene photos. consecutive leads complex structure Such at transitions gaps between content to,Such complex structure leads to large content gaps at scene transitions between consecutive photos.,18712
outperforms BMRNN three Experiments model the methods. show that standard the state-of-the-art datasets on storytelling,Experiments on three standard storytelling datasets show that the BMRNN model outperforms the state-of-the-art methods.,18713
of networks. successes deep neural recent to function the Activation crucial is,Activation function is crucial to the recent successes of deep neural networks.,18714
"is to important train addition, weight In networks. deep very initialization very","In addition, weight initialization is very important to train very deep networks.",18715
and to paper range. extends the wider the complements theory it current This,This paper complements the current theory and extends it to the wider range.,18716
asthma. and To oral between assess the relationship health Aim:,Aim: To assess the relationship between oral health and asthma.,18717
"proportion other significant. BOP plaque and hand, the index sites, not statistically of On were","On the other hand, proportion of BOP sites, and plaque index were not statistically significant.",18718
to Participants likely Conclusion: periodontitis severe with have less asthma. were,Conclusion: Participants with severe periodontitis were less likely to have asthma.,18719
outcome. when inverse an found using of association Stronger asthma was evidence as medication,Stronger evidence of an inverse association was found when using asthma medication as outcome.,18720
. asthma hypothesis Keywords: periodontal disease; periodontitis; medication; hygiene asthma;,Keywords: asthma; periodontal disease; asthma medication; periodontitis; hygiene hypothesis .,18721
best determines by the using model sequential Bayes' strategies estimators. The,The model determines the best strategies by using sequential Bayes' estimators.,18722
problem; solution environment. is to a universal can any learn RL it the computable AIXI,AIXI is a universal solution to the RL problem; it can learn any computable environment.,18723
constraint. when reaches a state It the and if satisfying network ceases the stable,It ceases if and when the network reaches a stable state satisfying the constraint.,18724
As in models (SQR) the square graphical work al. with et recent root [Inouye,As in the recent work with square root graphical (SQR) models [Inouye et al.,18725
for situations. in finish We methods the parameter by space various discussing reducing,We finish by discussing methods for reducing the parameter space in various situations.,18726
Thorough theory. real synthetic and on our back datasets experiments up both,Thorough experiments on both synthetic and real datasets back up our theory.,18727
systems. useful expect We to insight design be the of in this image captioning scalable,We expect this insight to be useful in the design of scalable image captioning systems.,18728
the This Single as Super-Resolution known generally is problem. Image (SISR),This is generally known as the Single Image Super-Resolution (SISR) problem.,18729
"fetal placenta. blood oxygen in from pregnancy, diffuses the maternal During trees villous to through","During pregnancy, oxygen diffuses from maternal to fetal blood through villous trees in the placenta.",18730
"Doppler monitoring ultrasound. for proposed been auscultation and have estimation, including Several FHR techniques","Several monitoring techniques have been proposed for FHR estimation, including auscultation and Doppler ultrasound.",18731
represented more than to object. by enables be each class one mechanism This,This mechanism enables each class to be represented by more than one object.,18732
exponentially the parent number in of increases of parameters with included number CPTs variables. The,The number of parameters included in CPTs increases exponentially with the number of parent variables.,18733
the The of common solution canonical the is so-called most application gates.,The most common solution is the application of the so-called canonical gates.,18734
BNOR with dealing capable network. both of and the epistemic is of aleatory uncertainty,BNOR is capable of dealing with both aleatory and epistemic uncertainty of the network.,18735
"NOR. is BNOR when Specially, degrades epistemic no there uncertainty, into","Specially, when there is no epistemic uncertainty, BNOR degrades into NOR.",18736
are (FS) brain inhibitory Fast-spiking by connections. self-innervated in the interneurons autaptic powerful GABAergic,Fast-spiking (FS) interneurons in the brain are self-innervated by powerful inhibitory GABAergic autaptic connections.,18737
"response the such inhibition we firing investigate By regulates how interneurons. computational modelling, autaptic of","By computational modelling, we investigate how autaptic inhibition regulates the firing response of such interneurons.",18738
"modifications higher However, deliver that straightforward could more there accuracy? are","However, are there more straightforward modifications that could deliver higher accuracy?",18739
resolution that image We increasing input (i.e. show,We show that increasing input image resolution (i.e.,18740
provide where layers CNN of earlier/shallower than layers. also accuracy later/deeper situations higher find We,We also find situations where earlier/shallower layers of CNN provide higher accuracy than later/deeper layers.,18741
that further accuracy. and We yield images upsampled show models shallow competitive,We further show that shallow models and upsampled images yield competitive accuracy.,18742
Deep to recognition with methods. state-of-the-art tasks networks advance image the of various neural continue,Deep neural networks continue to advance the state-of-the-art of image recognition tasks with various methods.,18743
"of remain applications limited. to methods multimodality these However,","However, applications of these methods to multimodality remain limited.",18744
models explored alternative our are Various on study. multimodality based by introduced,Various alternative models introduced by multimodality are explored based on our study.,18745
Extensive can experiments that show on state-of-the-art achieve these benchmarks results. our model,Extensive experiments on these benchmarks show that our model can achieve state-of-the-art results.,18746
in spike-time irregularity modulating is self-feedback understood. The still importance poorly autaptic transmission of,The importance of self-feedback autaptic transmission in modulating spike-time irregularity is still poorly understood.,18747
of most Clustering important in statistics. one and learning machine the is unsupervised problems,Clustering is one of the most important unsupervised problems in machine learning and statistics.,18748
"weakly achieves whereas exact i.e. K-SVD is consistent, recovery,","achieves exact recovery, whereas K-SVD is weakly consistent, i.e.",18749
present applications season inpainting and We to also transfer.,We also present applications to inpainting and season transfer.,18750
essentially consistency temporal under is and matching method some DTW point-to-point a boundary constraints.,DTW is essentially a point-to-point matching method under some boundary and temporal consistency constraints.,18751
"Although global not obtains locally necessarily a it achieve optimal matchings. does solution, DTW sensible","Although DTW obtains a global optimal solution, it does not necessarily achieve locally sensible matchings.",18752
"structures points may two local be dissimilar matched with entirely by DTW. Concretely, temporal","Concretely, two temporal points with entirely dissimilar local structures may be matched by DTW.",18753
"and provided our reproduction, dataset To facilitate new the implementation also are online.","To facilitate reproduction, the new dataset and our implementation are also provided online.",18754
luxury the to bane accurate Is a of recognition? data,Is the luxury of data a bane to accurate recognition?,18755
important assessing the dynamic linear behavior stochastic of systems. response are functions for Frequency (FRFs),Frequency response functions (FRFs) are important for assessing the behavior of stochastic linear dynamic systems.,18756
"evaluations For a even time-consuming their single large systems, are for simulation.","For large systems, their evaluations are time-consuming even for a single simulation.",18757
"is In feasible. crude Monte-Carlo such quantification cases, uncertainty simulation by not","In such cases, uncertainty quantification by crude Monte-Carlo simulation is not feasible.",18758
for PCEs strategy lower-order frequency. results in each This,This strategy results in lower-order PCEs for each frequency.,18759
reduce number the Principal of then is analysis to outputs. employed random component,Principal component analysis is then employed to reduce the number of random outputs.,18760
image multilevel benefit well It segmentation accepted utilizing can cues. that is from,It is well accepted that image segmentation can benefit from utilizing multilevel cues.,18761
phase. without expensive pre-training This codes the forces produce network the to sparse,This forces the network to produce sparse codes without the expensive pre-training phase.,18762
the constraints enables global on foreground observed function. imposing This mass probability,This enables imposing global constraints on the observed foreground probability mass function.,18763
localization benchmark detection four evaluations on Our performance. demonstrate and datasets superior,Our evaluations on four benchmark datasets demonstrate superior detection and localization performance.,18764
vision for Haze and photography computational applications. important is computer removal,Haze removal is important for computational photography and computer vision applications.,18765
"we algorithm images. Accordingly, for propose dehazing hazy efficient an nighttime","Accordingly, we propose an efficient dehazing algorithm for nighttime hazy images.",18766
steps. The three proposed algorithm sequential includes,The proposed algorithm includes three sequential steps.,18767
the Maximum typically (EM) parameters carried is estimation likelihood Expectation-Maximization of via the out algorithm.,Maximum likelihood estimation of the parameters is typically carried out via the Expectation-Maximization (EM) algorithm.,18768
large datasets. This can quite high-dimensional for time-consuming become and/or,This can become quite time-consuming for large and/or high-dimensional datasets.,18769
"changes small simple easy and to requiring approach code. is only Our to standard implement,","Our approach is simple and easy to implement, requiring only small changes to standard code.",18770
with proposed mechanism especially The combined progressive attention. well hard when attention works,The proposed progressive attention mechanism works well especially when combined with hard attention.,18771
DNNs the to These probe music. used neural representations of were subsequently,These DNNs were subsequently used to probe the neural representations of music.,18772
"In we problem of estimation. age this the paper, apparent address","In this paper, we address the problem of apparent age estimation.",18773
are of The produce final these the models learning combined outputs then deep to estimation.,The outputs of these deep learning models are then combined to produce the final estimation.,18774
affective In we cognitive paper (NEUromodulating ARchitecture). a neurobiologically-inspired new architecture: this NEUCOGAR COGnitive present,In this paper we present a new neurobiologically-inspired affective cognitive architecture: NEUCOGAR (NEUromodulating COGnitive ARchitecture).,18775
successfully Our localization eye tasks inclduing to different approach and detection. two car applied is,Our approach is successfully applied to two different tasks inclduing eye localization and car detection.,18776
results the analysis of provided are and experimental the algorithm. to proposed demonstrate Theoretical robustness,Theoretical analysis and experimental results are provided to demonstrate the robustness of the proposed algorithm.,18777
short-term tasks promising Visual disease Alzheimer's memory a for are binding early marker (AD).,Visual short-term memory binding tasks are a promising early marker for Alzheimer's disease (AD).,18778
feature sets resultant two then the form to are of histogram. concatenated features The,The two sets of features are then concatenated to form the resultant feature histogram.,18779
the Experiments and databases. SMIC were on II CASME performed,Experiments were performed on the CASME II and SMIC databases.,18780
in entrain hypothesized that neural production speech rhythmic regions. oscillations acts motor brain We,We hypothesized that rhythmic motor acts entrain neural oscillations in speech production brain regions.,18781
"of times were these the Importantly, observed synchronized speech we that with rhythm production. spiking","Importantly, we observed that these spiking times were synchronized with the rhythm of speech production.",18782
"and algorithm. LAND to EM provide corresponding mixture further the models, extend We the","We further extend the LAND to mixture models, and provide the corresponding EM algorithm.",18783
"is aging databases, various approach face using evaluated The proposed i.e.","The proposed approach is evaluated using various face aging databases, i.e.",18784
three a one. The very these difficult combination this inverse problem makes properties of,The combination of these three properties makes this inverse problem a very difficult one.,18785
"features into Rather than image on relying spectral purely content, play. come appearance-based","Rather than relying purely on spectral content, appearance-based image features come into play.",18786
Potsdam for sets. Vaihingen dense accuracy yields labelling ISPRS state-of-the-art The data benchmark the and,The dense labelling yields state-of-the-art accuracy for the ISPRS Vaihingen and Potsdam benchmark data sets.,18787
"experiments. Finally, important design the of we interpretation decoding caveats and time-series outline in","Finally, we outline important caveats in the design and interpretation of time-series decoding experiments.",18788
generalize fail way. the to machine same Existing in approaches learning,Existing machine learning approaches fail to generalize in the same way.,18789
can chaotic. even new be These attractors periodic or,These new attractors can be periodic or even chaotic.,18790
learning. robopsychological questions theoretical of approach This raises fundamental some machine,This robopsychological approach raises some fundamental theoretical questions of machine learning.,18791
Our discussion is of constrained these machines. questions to Turing,Our discussion of these questions is constrained to Turing machines.,18792
(aka describe given algorithm algorithms. Alan Turing Machine) an Turing had to the,Alan Turing had given an algorithm (aka the Turing Machine) to describe algorithms.,18793
"to investigate write algorithms paper, we the algorithms. In present","In the present paper, we investigate algorithms to write algorithms.",18794
of We demonstrate our fine forensic models potential arts arts. and in finally applications,We finally demonstrate potential applications of our models in fine arts and forensic arts.,18795
(CNNs) of require of can very amounts labeled convolutional Supervised networks learning data. neural large,Supervised learning of convolutional neural networks (CNNs) can require very large amounts of labeled data.,18796
Labeling training time examples or can of costly. and thousands be millions consuming extremely,Labeling thousands or millions of training examples can be extremely time consuming and costly.,18797
to data. is addressing unlabeled from create direction this One features towards problem,One direction towards addressing this problem is to create features from unlabeled data.,18798
to shown is resistant also to It be overfitting.,It is also shown to be resistant to overfitting.,18799
methods. method with alternative our that performance show competitive existing results provides The compared,The results show that our method provides competitive performance compared with existing alternative methods.,18800
"subclass Simplified Boardgames games. formalize arbitrary describes language, board which We a of","We formalize Simplified Boardgames language, which describes a subclass of arbitrary board games.",18801
Ground-based earth's for whole cameras opened monitoring new the have sky atmosphere. opportunities up,Ground-based whole sky cameras have opened up new opportunities for monitoring the earth's atmosphere.,18802
"image available analysis. have learning become to with the However, machine techniques powerful aid","However, powerful machine learning techniques have become available to aid with the image analysis.",18803
boundaries to consistent consistent that We key component assignment. argue temporally a region are temporally,We argue that temporally consistent boundaries are a key component to temporally consistent region assignment.,18804
information voxels. The the mutual point-wise of spatio-temporal based method (PMI) is proposed on,The proposed method is based on the point-wise mutual information (PMI) of spatio-temporal voxels.,18805
provides simplicity This the to problems algorithm it some also along. but to leads,This provides simplicity to the algorithm but it also leads to some problems along.,18806
out-sample this uses remove StackMC overfitting. / techniques to in-sample,StackMC uses in-sample / out-sample techniques to remove this overfitting.,18807
extend work. we this Here earlier substantially,Here we substantially extend this earlier work.,18808
new role astrocytes Retinal polarizing play sprouts to in VEGFA. a by crucial appear secreting,Retinal astrocytes appear to play a crucial role in polarizing new sprouts by secreting VEGFA.,18809
"extracellular. The description: three intracellular, model intercellular, incorporates levels of and","The model incorporates three levels of description: intracellular, intercellular, and extracellular.",18810
can organisms by How sensory by get of hard-coding number neurons? the,How can organisms get by by hard-coding the number of sensory neurons?,18811
We question theory. approach this rate-distortion using,We approach this question using rate-distortion theory.,18812
from users active of have changed content. consumers data mere multimedia to producers Common,Common users have changed from mere consumers to active producers of multimedia data content.,18813
expressions are way through interact humans important socially. which Facial an,Facial expressions are an important way through which humans interact socially.,18814
present the of also important influential methods. most datasets and We bench-marking the,We also present the important datasets and the bench-marking of most influential methods.,18815
"We research. discussion lines trends, a general and with questions of about future conclude important","We conclude with a general discussion about trends, important questions and future lines of research.",18816
a regions of given the automatic set semantic Co-segmentation of common is the images. extraction,Co-segmentation is the automatic extraction of the common semantic regions given a set of images.,18817
paper the the This process prediction quantification regression. considers in Gaussian of performance,This paper considers the quantification of the prediction performance in Gaussian process regression.,18818
that of systematic omission underestimation the errors. prediction show We leads this to a,We show that this omission leads to a systematic underestimation of the prediction errors.,18819
"unary it ill-posed are becomes attributes However, the when (heavily) corrupted.","However, it becomes ill-posed when the unary attributes are (heavily) corrupted.",18820
in mombf. the are article All methods proposed library R in implemented this,All methods proposed in this article are implemented in the R library mombf.,18821
"detailed Neuroscience studying neural dynamics circuits. the has implementation computation, focused and codes, on of","Neuroscience has focused on the detailed implementation of computation, studying neural codes, dynamics and circuits.",18822
think of the in about these brain ideas. Here terms we,Here we think about the brain in terms of these ideas.,18823
could refine to neuroscience test by which and these seek hypotheses. We suggest directions,We suggest directions by which neuroscience could seek to refine and test these hypotheses.,18824
(a.k.a. captured cameras by Sky/cloud ground-based images,Sky/cloud images captured by ground-based cameras (a.k.a.,18825
that many experiment method methods. state-of-the-art outperforms the results show proposed clustering The,The experiment results show that the proposed method outperforms many state-of-the-art clustering methods.,18826
procedure loss is compute each to in step. performed This,This procedure is performed to compute loss in each step.,18827
shared the make using single-step question a prediction for new model. a We,We make a single-step prediction for a new question using the shared model.,18828
We extend depths. the arbitrary to epistemic gossip problem,We extend the gossip problem to arbitrary epistemic depths.,18829
domains separate We on closely identify three related anthropo-centered levels. working functional,We identify three closely related anthropo-centered domains working on separate functional levels.,18830
"The this three layers and identified MindComputing, PhysioComputing, named MetaComputing. in model been have","The three layers identified in this model have been named PhysioComputing, MindComputing, and MetaComputing.",18831
"recent dynamics. have there simulations of spontaneous In been computational neural years, many","In recent years, there have been many computational simulations of spontaneous neural dynamics.",18832
"transformation perform correspondence SuperBIG space. decorrelated to guides the the refined color Finally, in a","Finally, the refined correspondence guides SuperBIG to perform the transformation in a decorrelated color space.",18833
most sets cores seven proteins snRNPs. different of the of Various form Sm,Various sets of seven different Sm proteins form the cores of most snRNPs.,18834
distinct Each has radiological-pathological parameters meaning. these of,Each of these parameters has distinct radiological-pathological meaning.,18835
"technique. local method, for propose level set the improving We a window, novel segmentation adaptive","We propose a novel method, the adaptive local window, for improving level set segmentation technique.",18836
co-occurrence considered statistics and local gray several are level calculating by Global matrices.,Global and local statistics are considered by calculating several gray level co-occurrence matrices.,18837
"models. analyzed images energy Moreover, different three we using","Moreover, we analyzed images using three different energy models.",18838
the parsimony. to introduce of are implemented also Thirteen model variants,Thirteen variants of the model are also implemented to introduce parsimony.,18839
"artificial data to and the exemplify package, To are applications the use presented. real of","To exemplify the use of the package, applications to artificial and real data are presented.",18840
and (KB) inherent limited knowledge of commonsense The their bases has usefulness. inflexibility incompleteness,The inherent inflexibility and incompleteness of commonsense knowledge bases (KB) has limited their usefulness.,18841
"ways, for a other's KB DSVS In and a these weaknesses. make can each up","In these ways, a KB and a DSVS can make up for each other's weaknesses.",18842
on results R-CNN demonstrated impressive has recently The Faster detection benchmarks. various object,The Faster R-CNN has recently demonstrated impressive results on various object detection benchmarks.,18843
analysis MATLAB of a in this as used Application is of study. tool,Application of MATLAB is used as a tool of analysis in this study.,18844
warped images. trained end-to-end fashion using an are in Our MS-COCO networks,Our networks are trained in an end-to-end fashion using warped MS-COCO images.,18845
need the feature local stages. without transformation works for estimation Our separate and detection approach,Our approach works without the need for separate local feature detection and transformation estimation stages.,18846
"three geometric involves transformation, This technique main steps: overlapping. and segmentation,","This technique involves three main steps: segmentation, geometric transformation, and overlapping.",18847
"present an overlapping. asymmetry analysis-based technique to alternative paper, In this we","In this paper, we present an alternative technique to asymmetry analysis-based overlapping.",18848
used of for been structures. different microscopy has Electronic morphology materials evaluation,Electronic microscopy has been used for morphology evaluation of different materials structures.,18849
"affected microscopy be may results factors. However, several by","However, microscopy results may be affected by several factors.",18850
Image to methods improve quality can be processing used the correct results. of and these,Image processing methods can be used to correct and improve the quality of these results.,18851
in application An order nanoparticles presented locate rubber membranes. natural is in to gold,An application is presented in order to locate gold nanoparticles in natural rubber membranes.,18852
this using demonstrated. Validation and vivo been of has ex experiments phantom system,Validation of this system using phantom and ex vivo experiments has been demonstrated.,18853
in its efficiency multimedia facilitating large-scale shown effectiveness Hashing has applications. and,Hashing has shown its efficiency and effectiveness in facilitating large-scale multimedia applications.,18854
"we independent project i.e. data Specifically, labels","Specifically, we project independent data labels i.e.",18855
"alternating algorithm the solve develop ZSH efficient model. to Besides, an we","Besides, we develop an efficient alternating algorithm to solve the ZSH model.",18856
organization modular demonstrated Several studies character of brain the activities. of have,Several studies have demonstrated the modular character of organization of brain activities.,18857
time in seconds. five serially Networks intervals were estimated of,Networks were serially estimated in time intervals of five seconds.,18858
First-order for compilation inference. knowledge proven techniques efficient lifted have,First-order knowledge compilation techniques have proven efficient for lifted inference.,18859
target their used as structures Early methods data circuit.,Early methods used data structures as their target circuit.,18860
attracted mechanisms captioning to its powerful have performance. considerable image due interest in Attention,Attention mechanisms have attracted considerable interest in image captioning due to its powerful performance.,18861
on extensive We dataset. perform the MS-COCO experiments,We perform extensive experiments on the MS-COCO dataset.,18862
"current gradient with QuickProp standard We compare empirically is method. descent, the which","We compare QuickProp empirically with gradient descent, which is the current standard method.",18863
"function, objective novel coefficient. Dice training, optimise introduce We a during based that we on","We introduce a novel objective function, that we optimise during training, based on Dice coefficient.",18864
task is Multiple applications. in (MHT) many computer tracking fundamental a vision human,Multiple human tracking (MHT) is a fundamental task in many computer vision applications.,18865
the be in to indices compute surrogate negligible Then can time. the sensitivity used,Then the surrogate can be used to compute the sensitivity indices in negligible time.,18866
analytically indices the coefficients. total from Sobol') can computed PCE be,total Sobol') indices can be computed analytically from the PCE coefficients.,18867
"we synthetic segmentation with images, truth the accuracy. on good the Furthermore, ground recover","Furthermore, on the synthetic images, we recover the ground truth segmentation with good accuracy.",18868
"theory potential formulate steps questions. w.r.t.\ Finally, and open future experimentation, discuss and we","Finally, we discuss potential future steps w.r.t.\ theory and experimentation, and formulate open questions.",18869
evaluate benchmarks. our recognition performance multiple action the using method We of,We evaluate the performance of our method using multiple action recognition benchmarks.,18870
methods. action consistently state-of-the-art outperforms recognition model and baselines Our,Our model consistently outperforms baselines and state-of-the-art action recognition methods.,18871
on (i.e. image approach our evaluate AttractioNet We datasets extensively several,We extensively evaluate our AttractioNet approach on several image datasets (i.e.,18872
maximum model. for the likelihood principle derive criterion a the using learning we Then,Then using the maximum likelihood principle we derive a criterion for learning the model.,18873
problem tackled approach. is using a optimization The majorization-minimization resulting,The resulting optimization problem is tackled using a majorization-minimization approach.,18874
"theoretical these superficial approximations similarities, surprisingly properties behave practice. in and Despite differently have different","Despite superficial similarities, these approximations have surprisingly different theoretical properties and behave differently in practice.",18875
"recognition. method this propose character paper, novel scene we robust In a for","In this paper, we propose a novel method for robust scene character recognition.",18876
"cortices. events frontal incongruent frontal events temporally asynchronous Semantically right recruited while left recruited regions,","Semantically incongruent events recruited left frontal regions, while temporally asynchronous events recruited right frontal cortices.",18877
theorems state operators. representation We for some these,We state some representation theorems for these operators.,18878
hallmark human This across generalize ability of to visual is a complex intelligence. transformations,This ability to generalize across complex transformations is a hallmark of human visual intelligence.,18879
the might hypothesis invariant that test Here we action gap. this fill discrimination,Here we test the hypothesis that invariant action discrimination might fill this gap.,18880
"odontocete non-human mammals, both PRLS been confirmed has cetaceans. Prolonged only in two","Prolonged PRLS has been confirmed in only two non-human mammals, both odontocete cetaceans.",18881
"self-search first-person we tailored videos. risks, developed Motivated by to a these technique benefits and","Motivated by these benefits and risks, we developed a self-search technique tailored to first-person videos.",18882
improves self-search Our and well-known recognizers. detectors face performance approach over several significantly,Our approach significantly improves self-search performance over several well-known face detectors and recognizers.,18883
cross-immunity produce period. striking in incidence and the increases can Asymmetric,Asymmetric cross-immunity can produce striking increases in the incidence and period.,18884
a by autoencoder. knowledge deep illustrated by a concept generative specific model is provided The,The concept is illustrated by a specific knowledge model provided by a deep generative autoencoder.,18885
of neural data. sets large Effective labeled convolutional on are trained networks,Effective convolutional neural networks are trained on large sets of labeled data.,18886
"and datasets task. labeled time-consuming very large However, creating a is costly","However, creating large labeled datasets is a very costly and time-consuming task.",18887
"consider semi-supervised In neural paper, learning the problem with we convolutional this networks. of","In this paper, we consider the problem of semi-supervised learning with convolutional neural networks.",18888
We the evaluate on proposed datasets. method benchmark several,We evaluate the proposed method on several benchmark datasets.,18889
"in raises question, whether optimisation holds problems This the general. for this","This raises the question, whether this holds for optimisation problems in general.",18890
"have deep based and performance. several remarkable learning been achieved re-identification approaches proposed person Recently,","Recently, several deep learning based person re-identification approaches have been proposed and achieved remarkable performance.",18891
"many research scientific is tracking to From applications, eye important commercial tool domains. across an","From scientific research to commercial applications, eye tracking is an important tool across many domains.",18892
"has tracking become a range yet its eye technology. applications, pervasive to of Despite","Despite its range of applications, eye tracking has yet to become a pervasive technology.",18893
different has fields. many This applied to successfully been algorithm,This algorithm has been successfully applied to many different fields.,18894
clustering parts: consists and PSF algorithm prediction. of major The two,The PSF algorithm consists of two major parts: clustering and prediction.,18895
of part selection the clusters. of number The optimum clustering includes,The clustering part includes selection of the optimum number of clusters.,18896
data clusters. such time It series reference with to labels,It labels time series data with reference to such clusters.,18897
of implement to the PSF The functions various consists PSF algorithm. package,The PSF package consists of various functions to implement the PSF algorithm.,18898
automates functions function also prediction a results. contains optimized all to other It which obtain,It also contains a function which automates all other functions to obtain optimized prediction results.,18899
package syntax. paper describes the PSF all in their with This the functions,This paper describes all the functions in the PSF package with their syntax.,18900
usage. provides It also example a of simple,It also provides a simple example of usage.,18901
survival agents. is strategy problems one fundamental the biological a (policy) of Obtaining of,Obtaining a survival strategy (policy) is one of the fundamental problems of biological agents.,18902
"size, bounds our groups of bounds. are than known less When equal all are conservative","When all groups are of equal size, our bounds are less conservative than known bounds.",18903
"our even different have where situations to sizes. Moreover, where groups results apply the","Moreover, our results apply even to situations where where the groups have different sizes.",18904
"method achieved also evaluation decoders' used is tune a hyper-parameters. cross-validation, via to Such","Such evaluation is achieved via cross-validation, a method also used to tune decoders' hyper-parameters.",18905
review decoding This paper procedures in cross-validation a neuroimaging. for on is,This paper is a review on cross-validation procedures for decoding in neuroimaging.,18906
overview didactic It includes the theoretical relevant a considerations. of,It includes a didactic overview of the relevant theoretical considerations.,18907
cross-validation Nested parameters tune can bias. while avoiding decoders' circularity,Nested cross-validation can tune decoders' parameters while avoiding circularity bias.,18908
algorithms. food this The deep learning-based is recognition innovation image key technique the paper in,The key technique innovation in this paper is the deep learning-based food image recognition algorithms.,18909
framework for to and training makes our and testing. efficient effective both be This,This makes our framework to be efficient and effective for both training and testing.,18910
unsupervised adaptation be can in manner. implemented an effectively process and The efficiently,The adaptation process can be efficiently and effectively implemented in an unsupervised manner.,18911
exploding becomes deep a issue. such the extremely vanishing architectures problem gradient or key In,In such extremely deep architectures the vanishing or exploding gradient problem becomes a key issue.,18912
refer as to residual this We memory convolutional a network.,We refer to this as a convolutional residual memory network.,18913
"GRTwIND both issues. is If valid, solves universal perception","If universal perception is valid, GRTwIND solves both issues.",18914
to with its discussion perception. of primacy conclude perceptual and universal relationship a We,We conclude with a discussion of perceptual primacy and its relationship to universal perception.,18915
(Gatys artistic an the note neural al. et algorithm to presents transfer This extension style,This note presents an extension to the neural artistic style transfer algorithm (Gatys et al.,18916
image. algorithm to The an the given have style of another image original transforms,The original algorithm transforms an image to have the style of another given image.,18917
"can painting. photograph style be example, have a of famous For a to transformed the","For example, a photograph can be transformed to have the style of a famous painting.",18918
describe style colors. simple while for methods We linear preserving transferring,We describe simple linear methods for transferring style while preserving colors.,18919
limit. to the 'true model this perfectly that in We show OF' the adapts,We show that this model adapts perfectly to the 'true OF' in the limit.,18920
geometric small described interpretation. with of is perfect OF straightforward number This parameters a by,This perfect OF is described by a small number of parameters with straightforward geometric interpretation.,18921
"across implicitly However, assumes homogeneous variance entire space. the the method explanatory-variable","However, the method implicitly assumes homogeneous variance across the entire explanatory-variable space.",18922
the when behaves with It faced data. unknown is heteroscedastic how algorithm,It is unknown how the algorithm behaves when faced with heteroscedastic data.,18923
promising Our insights large scale video in analysis. provide experiments,Our experiments provide promising insights in large scale video analysis.,18924
"the the performances an presented of in evaluation feasibility shows Finally, approach. experimental setting","Finally, performances evaluation in an experimental setting shows the feasibility of the presented approach.",18925
Agents scenarios must real-world to of intelligence general deployed environmental conditions. adapt in ever-changing,Agents of general intelligence deployed in real-world scenarios must adapt to ever-changing environmental conditions.,18926
sharing learning architecture proposed leading enables both capacity. The simplicity and filter increased to,The proposed architecture enables both simplicity and filter sharing leading to increased learning capacity.,18927
other The demonstrated is to by to existing proposed also converting network networks. tasks generalize,The proposed network is also demonstrated to generalize to other tasks by converting existing networks.,18928
Concept Trees presented. grow from are sources concepts the of consistent sequences can semi-structured when,Concept Trees can grow from the semi-structured sources when consistent sequences of concepts are presented.,18929
be enriched structure can dynamic restricted the and by complimented more context. This then,This restricted structure can then be complimented and enriched by the more dynamic context.,18930
tests A architecture. have number of the the demonstrated soundness of,A number of tests have demonstrated the soundness of the architecture.,18931
"are of query language types two described. Then,","Then, two types of query language are described.",18932
showed that interpretable. A ensemble synthetic approximated a complicated was as tree experiment reasonably,A synthetic experiment showed that a complicated tree ensemble was approximated reasonably as interpretable.,18933
"a assumes Moreover, toxin a produces byproduct. that strain model the the as inefficient","Moreover, the model assumes that the inefficient strain produces a toxin as a byproduct.",18934
strength. strains both affects toxin rate This with of the growth different,This toxin affects the growth rate of both strains with different strength.,18935
anomaly features. different detection investigated these Two approaches are using,Two different anomaly detection approaches are investigated using these features.,18936
parallelized thousands are enabling molecules sequence simultaneously. methods to highly of to millions NGS,NGS methods are highly parallelized enabling to sequence thousands to millions of molecules simultaneously.,18937
are for performed. be analysis to Specific algorithms written task data specific,Specific data analysis algorithms are written for specific task to be performed.,18938
"algorithms analysing in group, NGS tool data. as The the act a in","The algorithms in group, act as a tool in analysing the NGS data.",18939
present In Semi-Aggregated model. (SAMDP) we MDP work the this,In this work we present the Semi-Aggregated MDP (SAMDP) model.,18940
describe spatial exhibiting A temporal and both best hierarchies. suited model policies to,A model best suited to describe policies exhibiting both spatial and temporal hierarchies.,18941
decoding interest of in primary neuroimaging many approaches studies. the interpretability is of brain Improving,Improving the interpretability of brain decoding approaches is of primary interest in many neuroimaging studies.,18942
"paper, present of interpretability In this we simple brain decoding models. linear a for definition","In this paper, we present a simple definition for interpretability of linear brain decoding models.",18943
"benchmarked proposed challenging i.e. approach detection on recent is face two databases, The","The proposed approach is benchmarked on two recent challenging face detection databases, i.e.",18944
"require answer, factual which sense, example. excludes or It knowledge to common basic for questions","It excludes questions which require common sense, or basic factual knowledge to answer, for example.",18945
"FVQA, supports, Here dataset we and introduce deeper VQA which a much reasoning. requires,","Here we introduce FVQA, a VQA dataset which requires, and supports, much deeper reasoning.",18946
which FVQA answer. contains questions information to external only require,FVQA only contains questions which require external information to answer.,18947
"triplet, fact a supporting as represented as structural such is <Cat,CapableOf,ClimbingTrees>. The","The supporting fact is represented as a structural triplet, such as <Cat,CapableOf,ClimbingTrees>.",18948
"Mondrian Laplace kernel, kernel. the random feature We the introduce to a fast approximation","We introduce the Mondrian kernel, a fast random feature approximation to the Laplace kernel.",18949
methods relationship and This kernel provides link the a random forests. between new insight into,This link provides a new insight into the relationship between kernel methods and random forests.,18950
success have great based deep on achieved approaches Recent handwriting  learning recognition. ,  Recent deep learning based approaches have achieved great success on handwriting recognition.,18951
widely characters most writing the Chinese the are in adopted among world. systems,Chinese characters are among the most widely adopted writing systems in the world.,18952
mainly on focused handwritten research Chinese recognizing characters. Previous has,Previous research has mainly focused on recognizing handwritten Chinese characters.,18953
of connectedness Examples the include neural vascular structures under constraints. reconstruction or,Examples include the reconstruction of neural or vascular structures under connectedness constraints.,18954
"this multiple presented. paper, characterization, MI-ACE, for MI-SMF target are instance and In two methods","In this paper, two methods for multiple instance target characterization, MI-SMF and MI-ACE, are presented.",18955
data. signature discriminative training a and imprecisely-labeled target mixed MI-SMF estimate and from MI-ACE,MI-SMF and MI-ACE estimate a discriminative target signature from imprecisely-labeled and mixed training data.,18956
these The the has been geometry used algorithms. cones analyze to of inference,The geometry of these cones has been used to analyze the inference algorithms.,18957
demonstrate proposed Four approach. are the of provided to numerical examples validity the,Four numerical examples are provided to demonstrate the validity of the proposed approach.,18958
the bound expression and proceed to when is the TF promoter. Transcription gene,Transcription and gene expression proceed when the TF is bound to the promoter.,18959
codominant and heterozygotes additive Cis-site therefore phenotypes. expression have,Cis-site heterozygotes have additive expression and therefore codominant phenotypes.,18960
on infinite-dimensional often sampling Bayesian distributions posterior inverse spaces. problems involve function,Bayesian inverse problems often involve sampling posterior distributions on infinite-dimensional function spaces.,18961
"of convergence with has methods class mesh-independent MCMC Recently, a times emerged. new","Recently, a new class of MCMC methods with mesh-independent convergence times has emerged.",18962
"work, subspace infinite-dimensional methods approaches. finite-dimensional on a this with mesh-independent geometric combine In we","In this work, we combine geometric methods on a finite-dimensional subspace with mesh-independent infinite-dimensional approaches.",18963
volumetric from learns for a that volumetric introduces segmentation paper sparsely network This images. annotated,This paper introduces a network for volumetric segmentation that learns from sparsely annotated volumetric images.,18964
"segments set, this network new images. on Trained volumetric densely data the","Trained on this data set, the network densely segments new volumetric images.",18965
from proposed the et previous Ronneberger u-net al. architecture The network extends,The proposed network extends the previous u-net architecture from Ronneberger et al.,18966
on-the-fly augmentation implementation data training. efficient for deformations performs The during elastic,The implementation performs on-the-fly elastic deformations for efficient data augmentation during training.,18967
"no trained from is is i.e., required. It scratch, end-to-end pre-trained network","It is trained end-to-end from scratch, i.e., no pre-trained network is required.",18968
"build critical to efficient is embedding and an images texts. of it Thus,","Thus, it is critical to build an efficient embedding of images and texts.",18969
DualNet an performance. boosts further of Building the ensemble,Building an ensemble of DualNet further boosts the performance.,18970
"two framework, practical to regions. also the this general base we propose Besides create ways","Besides this general framework, we also propose two practical ways to create the base regions.",18971
of of number in a case the consider probabilistic classifiers large training classes. We,We consider training probabilistic classifiers in the case of a large number of classes.,18972
is The too assumed perform over large classes. number classes of all to normalisation exact,The number of classes is assumed too large to perform exact normalisation over all classes.,18973
directly approximates a approach for this simple the To consider account we that likelihood.,To account for this we consider a simple approach that directly approximates the likelihood.,18974
"ranking a relate Furthermore, approach we to simple objective. this","Furthermore, we relate this approach to a simple ranking objective.",18975
auto-context using decision boosted train sequence trees features. We of a,We train a sequence of boosted decision trees using auto-context features.,18976
with softmax DeepWriter takes patches as input local loss. handwritten trained is classification and,DeepWriter takes local handwritten patches as input and is trained with softmax classification loss.,18977
In we image problem this on focus the of for compression. colorization work,In this work we focus on the problem of colorization for image compression.,18978
poses challenges. colorization for several But compression,But colorization for compression poses several challenges.,18979
"possible exhibit Second, objects multiple many world in colors. real the","Second, many objects in the real world exhibit multiple possible colors.",18980
account breast grading. nuclear when into size take cancer Pathologists performing,Pathologists take into account nuclear size when performing breast cancer grading.,18981
"been mean value. In nuclear shown the have has (MNA) independent to prognostic area addition,","In addition, the mean nuclear area (MNA) has been shown to have independent prognostic value.",18982
approach performing segmentation. is nuclei size straightforward measuring nuclear The by to,The straightforward approach to measuring nuclear size is by performing nuclei segmentation.,18983
activities recognizing task from to kinematics. We of apply surgical neural recurrent robot networks the,We apply recurrent neural networks to the task of recognizing surgical activities from robot kinematics.,18984
"we the to first apply this knowledge, task. our networks recurrent To are to neural","To our knowledge, we are the first to apply recurrent neural networks to this task.",18985
an crowd. An IHC the crowdsourcing is quantification undefined approach of to alternative images,An alternative approach is crowdsourcing the quantification of IHC images to an undefined crowd.,18986
"to valid Instead, fitting model usually they poses. step separate hand a rely generate","Instead, they usually rely a separate model fitting step to generate valid hand poses.",18987
inconvenient post sub-optimal. and Such a processing is,Such a post processing is inconvenient and sub-optimal.,18988
is public performance. challenging Our datasets on approach and verified state-of-the-art achieves,Our approach is verified on challenging public datasets and achieves state-of-the-art performance.,18989
"noted the are, It that so-called, nested. was labelings","It was noted that the labelings are, so-called, nested.",18990
views than two a We PAC-Bayesian multiview under study learning the framework. more with two-level,We study a two-level multiview learning with more than two views under the PAC-Bayesian framework.,18991
"redundant complex, is language ambiguous. sometimes and highly However,","However, language is highly complex, redundant and sometimes ambiguous.",18992
Many semantic different captions same express the concept. may,Many different captions may express the same semantic concept.,18993
four This provides contributions to problem. paper this,This paper provides four contributions to this problem.,18994
"this semi-automatic task. general a to benchmarks procedure Second, we describe for create","Second, we describe a general semi-automatic procedure to create benchmarks for this task.",18995
user-friendly application for an of environment The is upper-limb this interactive amputees. set-up,The application of this set-up is an interactive user-friendly environment for upper-limb amputees.,18996
distorted. to The tend observed get images,The observed images tend to get distorted.,18997
"paper, display we a calibration to the propose In fix this situation. method novel","In this paper, we propose a novel display calibration method to fix the situation.",18998
"mostly frequency accuracy patterns in our robustness, For the and analyzes method domain. high","For high accuracy and robustness, our method analyzes the patterns mostly in frequency domain.",18999
and Image problem computational a vision in is segmentation medical imaging. fundamental,Image segmentation is a fundamental problem in computational vision and medical imaging.,19000
to conditions guide algorithm design. derived are Then stabilization,Then stabilization conditions are derived to guide algorithm design.,19001
"of proposed effectiveness method are robustness and the Finally, demonstrated.","Finally, the effectiveness and robustness of proposed method are demonstrated.",19002
"annotated For a and new compiled evaluation, we dataset.","For evaluation, we compiled and annotated a new dataset.",19003
structure stabilize Of diverse interests are and features the networks. primary which underlying ecological,Of primary interests are the underlying structure and features which stabilize diverse ecological networks.,19004
and of context the provided. numerical interpreted in HIV/CTL within-host are Results are simulations evolution,Results are interpreted in the context of within-host HIV/CTL evolution and numerical simulations are provided.,19005
human fixation. are understand to maps and used attention Saliency visual,Saliency maps are used to understand human attention and visual fixation.,19006
"approximated for methodology The Criteria"". named stands is Information developed ""Minimizing MIC which","The developed methodology is named MIC which stands for ""Minimizing approximated Information Criteria"".",19007
"a result, a superior MIC performance As estimation. in is sparse with computationally fast","As a result, MIC is computationally fast with a superior performance in sparse estimation.",19008
"reparameterization yields Furthermore, circumventing in the additional terms advantage inference. post-selection of tactic an","Furthermore, the reparameterization tactic yields an additional advantage in terms of circumventing post-selection inference.",19009
PBC introduced its method and implementation the and The data. are illustrated R MIC with,The MIC method and its R implementation are introduced and illustrated with the PBC data.,19010
generative propose images. for of network learning (CoGAN) a joint distribution We coupled multi-domain adversarial,We propose coupled generative adversarial network (CoGAN) for learning a joint distribution of multi-domain images.,19011
from just can drawn with samples the a distributions. It distribution learn marginal joint,It can learn a joint distribution with just samples drawn from the marginal distributions.,19012
it For learns any each distribution images. task successfully tuple without of the corresponding joint,For each task it successfully learns the joint distribution without any tuple of corresponding images.,19013
image transformation. its We applications to and also adaptation demonstrate domain,We also demonstrate its applications to domain adaptation and image transformation.,19014
"Neural a decade, the last tremendous in Networks surge performance. saw Convolutional Over (CNN)","Over the last decade, Convolutional Neural Networks (CNN) saw a tremendous surge in performance.",19015
"challenging a to be a understanding has network learned still task. However, what proves","However, understanding what a network has learned still proves to be a challenging task.",19016
is proposed Carlo algorithm chain Markov based approach. on Monte The pseudo-marginal,The proposed algorithm is based on pseudo-marginal Markov chain Monte Carlo approach.,19017
in estimation an human-computer pose plays hand Articulated important interaction. role,Articulated hand pose estimation plays an important role in human-computer interaction.,19018
Experiments method a outperforms show that challenging dataset. state-of-the-art proposed the largely on,Experiments show that the proposed method largely outperforms state-of-the-art on a challenging dataset.,19019
"generalization Moreover, good cross-dataset method. demonstrates of a experiment the the ability proposed also","Moreover, a cross-dataset experiment also demonstrates the good generalization ability of the proposed method.",19020
over a outcomes. set associated of some finite group with Each is distributions,Each group is associated with some distributions over a finite set of outcomes.,19021
"faster. using segmentation DNLS much converges First,","First, segmentation using DNLS converges much faster.",19022
"DNLS regular function evolution. the throughout its Second, remains level set","Second, the DNLS level set function remains regular throughout its evolution.",19023
method. of the experimental proposed show the The results potential,The experimental results show the potential of the proposed method.,19024
medical Accurate important image yet analysis. automatic segmentation organ challenging problem for an is,Accurate automatic organ segmentation is an important yet challenging problem for medical image analysis.,19025
is abdominal pancreas anatomical organ high The variability. very an with,The pancreas is an abdominal organ with very high anatomical variability.,19026
class preserving method pancreas boundary segmentation. labels for pixel-wise generates Our,Our method generates boundary preserving pixel-wise class labels for pancreas segmentation.,19027
"this way is enhancing novel, in A practical images and of paper. fast introduced","A novel, fast and practical way of enhancing images is introduced in this paper.",19028
"very with acquired often low this the resolution. reason, are For data","For this reason, the data are often acquired with very low resolution.",19029
"data solution. provides resolution, To enhance interesting software interpolation DTI an","To enhance DTI data resolution, interpolation provides an interesting software solution.",19030
"methods. Chain Markov we Monte For inference, posterior use Carlo","For posterior inference, we use Markov Chain Monte Carlo methods.",19031
perform toy data. real in We experiments and,We perform experiments in toy and real data.,19032
"compared in other the of when literature, Results GWP protocols. validation different methods in outperform","Results of GWP outperform other methods in the literature, when compared in different validation protocols.",19033
results the The proposed the method. show experimental of potential,The experimental results show the potential of the proposed method.,19034
as problem. this a autoencoder conditional solution a variational to We propose,We propose a conditional variational autoencoder as a solution to this problem.,19035
"many has It but demanding. computationally is properties, nice","It has many nice properties, but is computationally demanding.",19036
"tests their and the $C$-sample one- location implemented. illustrate multivariate are corresponding use, To","To illustrate their use, the corresponding multivariate one- and $C$-sample location tests are implemented.",19037
"online the learning, tutoring popularity increasing of increased regaining With systems attention. are intelligent","With the increasing popularity of online learning, intelligent tutoring systems are regaining increased attention.",19038
by multi-armed of The bandit SBTS is algorithms. inspired class the,The SBTS is inspired by the class of multi-armed bandit algorithms.,19039
algorithm basic with practical is scenario The from a a evaluated programming course. java,The algorithm is evaluated with a practical scenario from a basic java programming course.,19040
"this relevant From scalar possible quantification, is (i.e. calculate measures to it","From this quantification, it is possible to calculate relevant scalar measures (i.e.",19041
of diagnosis anisotropy diffusivity) neurological and fractional in diseases. clinical mean employed,fractional anisotropy and mean diffusivity) employed in clinical diagnosis of neurological diseases.,19042
size structures. tissue smaller than is This much voxel,This voxel size is much smaller than tissue structures.,19043
"inaccurate. several dMRI derived procedures be from Therefore, may clinical","Therefore, several clinical procedures derived from dMRI may be inaccurate.",19044
resolution to enhance been Interpolation a in space. used dMRI of tensorial has,Interpolation has been used to enhance resolution of dMRI in a tensorial space.,19045
rank. that interpolates any generalizes show fields TDP HOT accurately obtained the to Results and,Results obtained show that TDP interpolates accurately the HOT fields and generalizes to any rank.,19046
We spatial axioms. and of default spatial and formalise frame reasoning concepts implement,We formalise and implement concepts of default spatial reasoning and spatial frame axioms.,19047
This paper for is TPLP. under consideration in publication,This paper is under consideration for publication in TPLP.,19048
model proposed area. The municipal makers helps to alternative preferable their for a decision choose,The proposed model helps decision makers to choose a preferable alternative for their municipal area.,19049
fact. confirms model proposed A sensitivity analysis by this our,A sensitivity analysis by our proposed model confirms this fact.,19050
standpoint will problem I physics. the discuss of this from,I will discuss this problem from the standpoint of physics.,19051
established. the switched is identifiable are proposed model Conditions which under,Conditions under which the proposed switched model is identifiable are established.,19052
is proximal-gradient An resulting problem. optimization first-order solve efficient developed algorithm to the,An efficient first-order proximal-gradient algorithm is developed to solve the resulting optimization problem.,19053
their on of the by given AHP analyzes relative preferences criteria Rough experts. based importance,Rough AHP analyzes the relative importance of criteria based on their preferences given by experts.,19054
the based criteria the alternative sites on weights. evaluates MABAC Rough,Rough MABAC evaluates the alternative sites based on the criteria weights.,19055
between explore its study the variants of and both We transferability datasets. model,We explore variants of the model and study its transferability between both datasets.,19056
perspective. a from This considers paper Verified AI methods formal,This paper considers Verified AI from a formal methods perspective.,19057
this framework. and provide all to solution necessary derivation We,We provide all necessary derivation and solution to this framework.,19058
"formally resemblance, not Despite been has their stated. relationship the","Despite their resemblance, the relationship has not been formally stated.",19059
"a them this we the describe precise In relationship paper, perspective. semantic logical between from","In this paper, we describe the precise semantic relationship between them from a logical perspective.",19060
"control relationship fitting Also, to of by provides binding DNA. a cAMP best to","Also, provides a relationship of best fitting to cAMP control by binding to DNA.",19061
cancer. inverse prediction cAMP level of and allows between relationship neurodegeneration the The an,The cAMP level allows the prediction of an inverse relationship between neurodegeneration and cancer.,19062
Learning critical vision. in computer representations is invariant task a,Learning invariant representations is a critical task in computer vision.,19063
"filters during we Instead, Divergence the they algorithm. gradient rotate Contrastive computed the when are","Instead, we rotate the gradient filters when they are computed during the Contrastive Divergence algorithm.",19064
structured consider We parameters. linear models multi-response high-dimensional learning with,We consider learning high-dimensional multi-response linear models with structured parameters.,19065
"their However, enormous complexity. interpretability the is to limited due critically","However, their interpretability is critically limited due to the enormous complexity.",19066
"Specifically, problem. model selection formalize tree as simplification of the ensembles a we","Specifically, we formalize the simplification of tree ensembles as a model selection problem.",19067
region power. could and genomic SNPs testing impair functional non-functional the Combining in arbitrary,Combining functional and non-functional SNPs in arbitrary genomic region could impair the testing power.,19068
"region next are of the refined. the Focusing step, boundaries In zoomed","In the next Focusing step, boundaries of the zoomed region are refined.",19069
"Furthermore, real-world synthetically-generated combining videos. the by background increase we behaviors with realism can","Furthermore, we can increase the realism by combining synthetically-generated behaviors with real-world background videos.",19070
WWW. would on the be LCrowdV released,LCrowdV would be released on the WWW.,19071
image contents of eliminating while capture The layers the convolutional abstraction corruptions.,The convolutional layers capture the abstraction of image contents while eliminating corruptions.,19072
maps capability details. layers to the feature the upsample recover have the Deconvolutional and image,Deconvolutional layers have the capability to upsample the feature maps and recover the image details.,19073
"the into significant amplitudes. would sustained In with translate flare-ups such ODE context, oscillations","In the ODE context, such flare-ups would translate into sustained oscillations with significant amplitudes.",19074
"chain RFM, the modeled is as $n$ a In of the molecule mRNA sites.","In the RFM, the mRNA molecule is modeled as a chain of $n$ sites.",19075
regulation the down optimization cases special rigorously problem. We of also two analyze,We also rigorously analyze two special cases of the down regulation optimization problem.,19076
"the along molecule. this the slowest necessarily transition is rate However, mRNA rate not","However, this rate is not necessarily the slowest transition rate along the mRNA molecule.",19077
the We some implications these biological discuss of results. of,We discuss some of the biological implications of these results.,19078
under term idea This more 'self-configuration'. often general is the captured,This more general idea is often captured under the term 'self-configuration'.,19079
the on jumps. sampling is virtual algorithm based The idea of,The algorithm is based on the idea of sampling virtual jumps.,19080
small establish end To drift a geometric a we towards this condition set.,To this end we establish a geometric drift condition towards a small set.,19081
be can as image. seen of a factorized epitome an An representation,An epitome can be seen as a factorized representation of an image.,19082
base pairs co-located layer learns of linear The between approach epitome patches. and mappings second,The second approach learns linear mappings between pairs of co-located base layer and epitome patches.,19083
for Multiple convolutional extract attention generated to features canvases are attention.,Multiple attention canvases are generated to extract convolutional features for attention.,19084
employed attentiveness unit An and learn to LSTM attention of recurrent canvases. is discrimination the,An LSTM recurrent unit is employed to learn the attentiveness and discrimination of attention canvases.,19085
Thus does in temporal long-term reef variability variability. among-site than more composition contributes variability to,Thus among-site variability contributes more to long-term variability in reef composition than does temporal variability.,19086
(DCM) under trained Dynamic the Causal on generative classifier Modeling generated The datasets model. is,The classifier is trained on datasets generated under the Dynamic Causal Modeling (DCM) generative model.,19087
"is statistical series, based between time upon The directionality two dependencies node e.g. inferred the","The directionality is inferred based upon statistical dependencies between the two node time series, e.g.",19088
online are samplers developed for Batch Gibbs and inference.,Batch and online Gibbs samplers are developed for inference.,19089
measure paper making. abnormality for new decision a introduces The,The paper introduces a new abnormality measure for decision making.,19090
and data. real synthetic on is proposed both evaluated The method,The proposed method is evaluated on both synthetic and real data.,19091
"Moreover, the dataset baseline on recently-released, first large-scale COCO-Text. the we result report","Moreover, we report the first baseline result on the recently-released, large-scale dataset COCO-Text.",19092
posterior is Approximate for framework for models. likelihood-free inference (ABC) Bayesian performing a Computation simulation,Approximate Bayesian Computation (ABC) is a framework for performing likelihood-free posterior inference for simulation models.,19093
all we avoid nearly derivations. exploiting By model-specific libraries can automatic then differentiation,By then exploiting automatic differentiation libraries we can avoid nearly all model-specific derivations.,19094
compare problems on demonstrate SVI three and performance We algorithms. existing to,We demonstrate performance on three problems and compare to existing SVI algorithms.,19095
the efficiency and results of demonstrate our Our correctness algorithm.,Our results demonstrate the correctness and efficiency of our algorithm.,19096
"introduce two novel uncertainty. additionally multi-class for we settings, criteria For","For multi-class settings, we additionally introduce two novel criteria for uncertainty.",19097
sequential is for computationally or massive It intractable data.,It is computationally intractable for massive or sequential data.,19098
"incrementally It observation. updating allows sequential the model data, process a to new by","It allows to process sequential data, incrementally updating the model by a new observation.",19099
is The abnormal in applied behaviour detection to sequences. video model,The model is applied to abnormal behaviour detection in video sequences.,19100
is decision A making. new proposed abnormality measure for,A new abnormality measure is proposed for decision making.,19101
in into challenge problems. stated be divided main the can The   two title,  The challenge stated in the title can be divided into two main problems.,19102
The problem to interfaces. is the users mimic that way with reliably user first interact,The first problem is to reliably mimic the way that users interact with user interfaces.,19103
"is agent, build problem to instructible second an i.e. The","The second problem is to build an instructible agent, i.e.",19104
expressed as be commands. previously execute can natural taught to unseen tasks language one that,one that can be taught to execute tasks expressed as previously unseen natural language commands.,19105
"to This proposes a the system a paper we second solution problem, Helpa. call","This paper proposes a solution to the second problem, a system we call Helpa.",19106
programming. not Teaching Helpa does involve any,Teaching Helpa does not involve any programming.,19107
on knowledge. not pre-existing any does Helpa rely domain-specific,Helpa does not rely on any pre-existing domain-specific knowledge.,19108
"respectively extended to -Probability, etc. Neutrosophic -Statistics -Measure, were Over-/Under-/Off-Logic,","were extended to respectively Neutrosophic Over-/Under-/Off-Logic, -Measure, -Probability, -Statistics etc.",19109
"convolutional Recently, power expressive to the made exploit attempts of have been several networks. deep","Recently, several attempts have been made to exploit the expressive power of deep convolutional networks.",19110
"not did Zygophyllales. of However, topologies the on placement these agree","However, these topologies did not agree on the placement of Zygophyllales.",19111
datasets. simple and makes efficient solution large implement MBFA-ZSL The close-form on run to to,The close-form solution makes MBFA-ZSL simple to implement and efficient to run on large datasets.,19112
This has genetics in has fueled research. human increase an,This has has fueled an increase in human genetics research.,19113
construction and This copper robust tape of observation a circuit soldering. traces enables using,This observation enables construction of a robust circuit traces using copper tape and soldering.,19114
actual through our We the demonstrate approach effectiveness construction of examples. various,We demonstrate the effectiveness of our approach through various actual construction examples.,19115
in processing. image Segmentation an important problem remains,Segmentation remains an important problem in image processing.,19116
Both results theoretical and demonstrations provided. on are images natural,Both theoretical results and demonstrations on natural images are provided.,19117
"a given context. manner in information ""situated"" done a This relevant by including is only","This is done in a ""situated"" manner by only including information relevant a given context.",19118
"fragile can However, generalizable. visual easily be detection methods and not","However, visual detection methods can be fragile and not easily generalizable.",19119
"a detection robust this neural deep on we present In convolutional based algorithm networks. paper,","In this paper, we present a robust detection algorithm based on deep convolutional neural networks.",19120
generate fixations. scene perception our eyes During complex sequences of,During scene perception our eyes generate complex sequences of fixations.,19121
for inhibitory viable planning dynamical We tagging that of mechanism a scanpaths. conclude is,We conclude that inhibitory tagging is a viable mechanism for dynamical planning of scanpaths.,19122
the of complexity depends on number this algorithm the quadratically observations. However of,However the complexity of this algorithm depends quadratically on the number of observations.,19123
leaves our algorithm We distribution ergodic. the prove posterior that,We prove that our algorithm leaves the posterior distribution ergodic.,19124
"inference show time. in reasonable is possible we that a Moreover,","Moreover, we show that inference is possible in a reasonable time.",19125
existing the contextual symmetry paper of symmetry. extends of introducing notion definitions by novel Our,Our paper extends existing definitions of symmetry by introducing the novel notion of contextual symmetry.,19126
in attention of recently. gained The finding overlapping networks problem has much communities,The problem of finding overlapping communities in networks has gained much attention recently.,19127
accuracy and both real-world on its We datasets. demonstrate simulated,We demonstrate its accuracy on both simulated and real-world datasets.,19128
"inconsistent serious and, can intermittent therefore, consequences. poor assessment long-term induce The cause treatment and","The intermittent and inconsistent assessment can induce poor treatment and, therefore, cause serious long-term consequences.",19129
analysis We give a of approach. risk the,We give a risk analysis of the approach.,19130
"Significant similarity. neural exist, and contribute response correlations to the","Significant neural correlations exist, and contribute to the response similarity.",19131
"selection. components, extraction two DDRL contains More main and i.e., feature specifically,","More specifically, DDRL contains two main components, i.e., feature extraction and selection.",19132
parameters Both millions carefully are designed to algorithms tuning. avoid of,Both algorithms are carefully designed to avoid millions of parameters tuning.,19133
This more big compact to a of leads classification image data. for solution,This leads to a more compact solution for image classification of big data.,19134
is proposed friendly approach be with We designed the note parallel computing. to that,We note that the proposed approach is designed to be friendly with parallel computing.,19135
easy to It to different distributed deployed generic resources. computing be is and,It is generic and easy to be deployed to different distributed computing resources.,19136
in models lines with complex few of a data represent Probabilistic programming intermingled languages code.,Probabilistic programming languages represent complex data with intermingled models in a few lines of code.,19137
"rather probabilistic the a is of given, program constructing is straightforward. structure model When","When the structure of model is given, constructing a probabilistic program is rather straightforward.",19138
"parameters have learn the marginal focus been main probabilities. compute model to and best Thus,","Thus, main focus have been to learn the best model parameters and compute marginal probabilities.",19139
description report programming derives such accurately. a descriptive covariance We that efficiently structure probabilistic,We report that such descriptive covariance structure efficiently derives a probabilistic programming description accurately.,19140
processing human Network brain. efficient key is modularity a feature in for information the,Network modularity is a key feature for efficient information processing in the human brain.,19141
neuroimaging resolution. This requires time sufficient techniques with,This requires neuroimaging techniques with sufficient time resolution.,19142
picture identify We task. naming functional during networks,We identify functional networks during picture naming task.,19143
"(tools, Two (scrambled) of and animals) were objects. visual presented, meaningful categories stimuli meaningless","Two categories of visual stimuli were presented, meaningful (tools, animals) and meaningless (scrambled) objects.",19144
cases. activated in ventral of Networks both visual were the pathway,Networks of the ventral visual pathway were activated in both cases.,19145
strong meaningless occipitotemporal appeared object for object. connectivity but meaningful However functional a for not,However a strong occipitotemporal functional connectivity appeared for meaningful object but not for meaningless object.,19146
depth Estimating RGB is inherently image an single and a ambiguous problem. from ill-posed,Estimating depth from a single RGB image is an ill-posed and inherently ambiguous problem.,19147
makes tools for argumentation. valuable the abstract dialectical This frameworks,This makes the abstract dialectical frameworks valuable tools for argumentation.,19148
form unknown. prove when We both this theorems are of parameters model and known,We prove theorems of this form both when model parameters are known and unknown.,19149
commonly Our proposed datasets. system state-of-the-art used the two benchmark performance offers on,Our proposed system offers the state-of-the-art performance on two commonly used benchmark datasets.,19150
agent article an for agent presents autonomous environments. in stochastic an architecture controlling This,This article presents an agent architecture for controlling an autonomous agent in stochastic environments.,19151
the approach feasible. experiments indicate of is The results simulation the that,The results of the simulation experiments indicate that the approach is feasible.,19152
recognition Previous part-based and detection separate approaches attribute in part perform steps. recognition attribute,Previous part-based attribute recognition approaches perform part detection and attribute recognition in separate steps.,19153
are recognition for not The therefore sub-optimal. be parts optimized could attribute and,The parts are not optimized for attribute recognition and therefore could be sub-optimal.,19154
overcome approach end-to-end deep an the to learning present We limitation.,We present an end-to-end deep learning approach to overcome the limitation.,19155
setting. jointly a learnt recognition key Both attribute estimation multi-task are and in point,Both key point estimation and attribute recognition are learnt jointly in a multi-task setting.,19156
verify the efficacy approach. on two Extensive of end-to-end datasets experiments proposed,Extensive experiments on two datasets verify the efficacy of proposed end-to-end approach.,19157
"i.e. components images, crucial of face approach, two this In","In this approach, two crucial components of face images, i.e.",19158
"approach by The face using in datasets, is proposed various applications challenging i.e. evaluated","The proposed approach is evaluated in various applications by using challenging face datasets, i.e.",19159
the system. This alert health public warrants the preparedness of,This alert warrants the preparedness of the public health system.,19160
One processing phase. this major the challenges of the image of technique is,One of the major challenges of this technique is the image processing phase.,19161
"The into and segmentation, four be mainly organized isolation, processing gridding, quantification. can spot steps:","The processing can be organized mainly into four steps: gridding, spot isolation, segmentation, and quantification.",19162
that automatic gridding four spot compares perform and This paper and describes isolation. techniques,This paper describes and compares four techniques that perform automatic gridding and spot isolation.,19163
into that methods optimization-based frameworks. their We survey focus information our incorporate on prior,We focus our survey on optimization-based methods that incorporate prior information into their frameworks.,19164
makes parameters which of Most codependent HTM-based efficient solutions are challenging. creating,Most of parameters are codependent which makes creating efficient HTM-based solutions challenging.,19165
the performance of on knowledge impact their profound requires and of the settings system. It,It requires profound knowledge of the settings and their impact on the performance of system.,19166
input efficient patterns. may given HTM designing be for This as models considered for recipes,This may be considered as recipes for designing efficient HTM models for given input patterns.,19167
and Rotation recognition. been interest invariants moment of processing great have pattern in image,Rotation moment invariants have been of great interest in image processing and pattern recognition.,19168
"also the data. not on discuss only theoretically experiments the We but real prove invariance,","We not only theoretically prove the invariance, but also discuss the experiments on real data.",19169
during human interest. the topic is rest of great brain a networks Exploring,Exploring the human brain networks during rest is a topic of great interest.,19170
conducted brain the been have structural studies to Several and study intrinsic previously networks. functional,Several structural and functional studies have previously been conducted to study the intrinsic brain networks.,19171
Our considered functional sets findings as hubs. of confirmed existence of regions the brain,Our findings confirmed the existence of sets of brain regions considered as functional hubs.,19172
"and isthmus the regions In orbitofrontal integration. particular, levels cingulate the reveal high of","In particular, the isthmus cingulate and the orbitofrontal regions reveal high levels of integration.",19173
web for is server novel a SMISS protein function prediction.,SMISS is a novel web server for protein function prediction.,19174
different for different usage. selected can predictors Three be,Three different predictors can be selected for different usage.,19175
server. are used Perl PHP programming the in and two language primary,PHP and Perl are two primary programming language used in the server.,19176
plugins for website. are No our needed,No plugins are needed for our website.,19177
studies point generating thought. Lesion to self-generated consciously regions to critical and experiencing,Lesion studies point to regions critical to generating and consciously experiencing self-generated thought.,19178
active We for method instances of describe situations. performing in objects localization of a visual,We describe a method for performing active localization of objects in instances of visual situations.,19179
"in four a dimensional is method polytopes A proposed. visualize euclidian space to $(x,y,z,w)$","A method to visualize polytopes in a four dimensional euclidian space $(x,y,z,w)$ is proposed.",19180
the world coordinates. curve placed a on parabola in multiple three-dimensional are The slices,The multiple slices are placed on a parabola curve in the three-dimensional world coordinates.,19181
a The window oval view an in slices appearance. form,The slices in a view window form an oval appearance.,19182
shapes the a change rotation to polytope. synchronously is slices All applied their when,All slices synchronously change their shapes when a rotation is applied to the polytope.,19183
"proposed two framework consists mechanisms, of The i.e. ICP","The proposed ICP framework consists of two mechanisms, i.e.",19184
ICP). DICP SICP and (Dynamic ICP) (Static,SICP (Static ICP) and DICP (Dynamic ICP).,19185
"MADM comparison existing is a described. two with methods Finally, other","Finally, a comparison with two other existing MADM methods is described.",19186
path the domains is In on conditioned the many taken. cells navigational of traversability,In many navigational domains the traversability of cells is conditioned on the path taken.,19187
addressing play needs. sites Question an in (CQA) important Community-based Answering information role health,Community-based Question Answering (CQA) sites play an important role in addressing health information needs.,19188
"significant remain posted a unanswered. questions of However, number","However, a significant number of posted questions remain unanswered.",19189
Our answers information identify from algorithm resolved to proposed candidate retrieval techniques QA. uses,Our proposed algorithm uses information retrieval techniques to identify candidate answers from resolved QA.,19190
on assessed Yahoo! a corpus We this from curated approach,We assessed this approach on a curated corpus from Yahoo!,19191
compared string Answers and a rule-based against similarity baseline.,Answers and compared against a rule-based string similarity baseline.,19192
of reasonably given data that considerably rate noisy. is A the high is obtained accuracy,A reasonably high rate of accuracy is obtained given that the data is considerably noisy.,19193
Feature applicable MATLAB widely (FSLib) Selection for Selection Feature library (FS). is a Library,Feature Selection Library (FSLib) is a widely applicable MATLAB library for Feature Selection (FS).,19194
"consists filtering, automatic thresholding, algorithm processing. main of The data four shifting, and proposed steps:","The proposed algorithm consists of four main steps: automatic thresholding, shifting, filtering, and data processing.",19195
bands. poor that is also used improve quality images faint have DNA to Enhancement,Enhancement is also used to improve poor quality images that have faint DNA bands.,19196
for nonparametric a time-varying networks. prior propose Bayesian We,We propose a Bayesian nonparametric prior for time-varying networks.,19197
is The the time-varying described sociabilities gamma evolution by process. tractable of a,The evolution of the sociabilities is described by a tractable time-varying gamma process.,19198
priors The and apply suitable challenge lies them in effectively. how to thus define,The challenge thus lies in how to define suitable priors and apply them effectively.,19199
transformation The variable can priors. to other also non-Gaussian be extended,The variable transformation can also be extended to other non-Gaussian priors.,19200
both environmental fields. in Optimization and arise control with many variables variables problems,Optimization problems with both control variables and environmental variables arise in many fields.,19201
han- This such paper personalized problems. to a framework introduces dle optimization of,This paper introduces a framework of personalized optimization to han- dle such problems.,19202
models optimization complex computer is discussed. The personalized implementation of for,The implementation of personalized optimization for complex computer models is discussed.,19203
show examples our effectiveness Numerical of the algorithms.,Numerical examples show the effectiveness of our algorithms.,19204
"presented for is paper, (i.e. method this superimposition a In","In this paper, a method is presented for superimposition (i.e.",19205
non-linear linear estimators followed optimisation. involves Each by stage,Each stage involves linear estimators followed by non-linear optimisation.,19206
of computation. for image fast The model warping invertible also is,The model of image warping is also invertible for fast computation.,19207
"results. better two methods, ours state-of-the-art gives to Compared","Compared to two state-of-the-art methods, ours gives better results.",19208
Monte formulation other admits chain algorithms many special This Carlo as cases. Markov,This formulation admits many other Markov chain Monte Carlo algorithms as special cases.,19209
propose to a based We novel face approach recognition. template,We propose a novel approach to template based face recognition.,19210
translate environmental networks responses. use Cells biochemical into information intracellular to,Cells use biochemical networks to translate environmental information into intracellular responses.,19211
information in and demonstrate mechanisms dynamic of limitations results These the storage cells.,These results demonstrate the mechanisms and limitations of dynamic information storage in cells.,19212
architectures. state-of-the-art Such CNN is data train necessary to,Such data is necessary to train state-of-the-art CNN architectures.,19213
viable is ours. large with such an Such only training sets as approach,Such an approach is viable only with large training sets such as ours.,19214
generalize CNNs well This demonstrates trained that to images. images real on artificial,This demonstrates that CNNs trained on artificial images generalize well to real images.,19215
equivalence This kernel Bayesian of provides understanding new a inference.,This equivalence provides a new understanding of kernel Bayesian inference.,19216
settings. and work problem in provides open solves this theoretical consistency regression Our analysis,Our theoretical work solves this open problem and provides consistency analysis in regression settings.,19217
the image to due super-resolution an is task extremely Depth challenging in information sub-sampling. loss,Depth image super-resolution is an extremely challenging task due to the information loss in sub-sampling.,19218
convolutional applied been Deep widely neural color image network to super-resolution. have,Deep convolutional neural network have been widely applied to color image super-resolution.,19219
"been super-resolution. has success to depth surprisingly, not this matched Quite","Quite surprisingly, this success has not been matched to depth super-resolution.",19220
images. and the depth This mainly is between inherent color difference to due,This is mainly due to the inherent difference between color and depth images.,19221
in complexity complexity from of of part neural stems anatomy. underlying The the the dynamics,The complexity of neural dynamics stems in part from the complexity of the underlying anatomy.,19222
injury. with a vulnerability theoretical versatility comes to This,This theoretical versatility comes with a vulnerability to injury.,19223
some treated the framework using We algebras. of genetic that cannot consider be examples also,We also consider some examples that cannot be treated using the framework of genetic algebras.,19224
"with from along individual static features the approach exploits advantageously frames our motion. Second,","Second, along with the static features from individual frames our approach advantageously exploits motion.",19225
summarizes This of advances review these dynamical computational processes. modeling in recent,This review summarizes recent advances in computational modeling of these dynamical processes.,19226
"guidelines ensure needed practical a to more suitable posterior. However, working numerical are","However, more practical guidelines are needed to ensure a suitable working numerical posterior.",19227
The data. examples with synthetic method using illustrated some is,The method is illustrated with some examples using synthetic data.,19228
relies Bayes a The classification Naive on classifier. process,The classification process relies on a Naive Bayes classifier.,19229
"thesis This DNA is studying the in trajectories understood configuration aimed at as space. mutations,","This thesis is aimed at studying mutations, understood as trajectories in the DNA configuration space.",19230
mutations flights of Levy proposed. model evolutive is terms of An in,An evolutive model of mutations in terms of Levy flights is proposed.,19231
and the cells. discuss bacterial the analogy the mutator found cancerous qualitative phenotype between We,We discuss the qualitative analogy found between the bacterial mutator phenotype and the cancerous cells.,19232
as radiation of is role analyzed. of source mutations The,The role of radiation as source of mutations is analyzed.,19233
decay in We breathing. Radon's in focus the case the of on lungs,We focus on the case of Radon's decay in the lungs in breathing.,19234
challenges in one important most It cyber-security. technical policy and the is of,It is one of the most important technical and policy challenges in cyber-security.,19235
"contributions. novel from the makes VideoLSTM three Starting LSTM, soft-Attention","Starting from the soft-Attention LSTM, VideoLSTM makes three novel contributions.",19236
the LSTM hardwire exploit soft-Attention correlation in the spatial architecture. convolutions we To,To exploit the spatial correlation we hardwire convolutions in the soft-Attention LSTM architecture.,19237
comparisons on classification datasets action and localization and Experiments challenging for our claims. support,Experiments and comparisons on challenging datasets for action classification and localization support our claims.,19238
"widespread in {\gamma}-crystallins eye causes disease a aggregation cataract. aging, Amorphous the of lens of","Amorphous aggregation of {\gamma}-crystallins in the eye lens causes a widespread disease of aging, cataract.",19239
"and different in oscillate the genotypes abundances. hosts Usually, parasites of their","Usually, the different genotypes of hosts and parasites oscillate in their abundances.",19240
their internal stability. allows solutions general fixed analytical for framework This for points and,This general framework allows for analytical solutions for internal fixed points and their stability.,19241
"has more chaotic apparently For than dynamics two reported. been species,","For more than two species, apparently chaotic dynamics has been reported.",19242
"from representations are feature cross the learning transfer enhanced. effectively By domains, the leveraging","By leveraging the transfer learning from cross domains, the feature representations are effectively enhanced.",19243
are by The improved refinement. iterative results further the,The results are further improved by the iterative refinement.,19244
activity states of How those constrain brain neuronal circuitry transitions between large-scale states? and does,How does large-scale brain circuitry constrain states of neuronal activity and transitions between those states?,19245
a has recently is instance task that called important. This become increasingly segmentation,This is a task called instance segmentation that has recently become increasingly important.,19246
"are with and work dealing there Recently problems, pose illumination respectively. many","Recently there are many work dealing with pose and illumination problems, respectively.",19247
same the However at encountered and time. variation the lighting always pose both will be,However both the lighting and pose variation will always be encountered at the same time.,19248
taken different images diverse. quite for views global is in structure the Normally,Normally the global structure for images taken in different views is quite diverse.,19249
"towards central in However, intelligence. machine avoids progress general-purpose learning of role the artificial prediction","However, the central role of prediction in machine learning avoids progress towards general-purpose artificial intelligence.",19250
"environments. how under understanding unseen these provide Causal structures systems behave about changing,","Causal structures provide understanding about how these systems behave under changing, unseen environments.",19251
three in inference ways. thesis This art causal of different advances the,This thesis advances the art of causal inference in three different ways.,19252
"we neural convolutional algorithms. structures causal network our using discover in Third, features","Third, we discover causal structures in convolutional neural network features using our algorithms.",19253
"performance DCNN incurs of However, excellent and requirement. large memory computational the high a complexity","However, the excellent performance of a DCNN incurs high computational complexity and large memory requirement.",19254
From one Shape a is The of Shading field. vision computer,The Shape From Shading is one of a computer vision field.,19255
in ambiguity local this concave). can be of field expressed The the (convex difficulty /,The difficulty of this field can be expressed in the local ambiguity (convex / concave).,19256
local have a the and J.Shi method View) Q.Zhu solve to proposed (Global ambiguity.,J.Shi and Q.Zhu have proposed a method (Global View) to solve the local ambiguity.,19257
relationship between the method singular the the theory on and points. based graph This,This method based on the graph theory and the relationship between the singular points.,19258
"same sequences mostly ending exploring camera motion, starting the at position. All contain and","All sequences contain mostly exploring camera motion, starting and ending at the same position.",19259
"In calibrated. photometrically existing datasets, to all sequences contrast are","In contrast to existing datasets, all sequences are photometrically calibrated.",19260
"the and represent model, vertices hyperedges denote In hypergraph points data hypotheses. model","In the hypergraph model, vertices represent data points and hyperedges denote model hypotheses.",19261
"algorithm fitting. a partition to robust addition, sub-hypergraphs for we model develop detect In hypergraph","In addition, we develop a robust hypergraph partition algorithm to detect sub-hypergraphs for model fitting.",19262
An is the variables select efficient interaction to the and algorithm estimate importance network. developed,An efficient algorithm is developed to select the importance variables and estimate the interaction network.,19263
real obtained examples. data in Excellent is and simulated performance,Excellent performance is obtained in simulated and real data examples.,19264
category. to Local resolution belongs Our approach,Our approach belongs to Local resolution category.,19265
"gray is The unknown the from the level, be TILT can the determined SLANT.","The TILT can be determined from the gray level, the unknown is the SLANT.",19266
process). test of step third The images the is (on-line application,The third step is the application of test images (on-line process).,19267
parameters These geometrical defines the reveal stage formulas key infection. viral cytoplasmic how of,These formulas reveal how key geometrical parameters defines the cytoplasmic stage of viral infection.,19268
"oxygenation, potentially function imaging information perfusion surgery. provide during can Multispectral and about tissue (MSI)","Multispectral imaging (MSI) can provide information about tissue oxygenation, perfusion and potentially function during surgery.",19269
"efficiency is with even closed multiple providing Our from, computational cameras. method","Our method is closed from, providing computational efficiency even with multiple cameras.",19270
analysis Colombia. for an We policy recidivism illustrate with of options reducing in ex-combatant,We illustrate with an analysis of policy options for reducing ex-combatant recidivism in Colombia.,19271
"this compression. a work, In present comparison two techniques of image we between","In this work, we present a comparison between two techniques of image compression.",19272
"the mentioned to blocks. cases, is Karhunen-Loeve both applied in transform Later,","Later, in both cases, the Karhunen-Loeve transform is applied to mentioned blocks.",19273
Our of performance proposed experiments the our promising model. validate,Our experiments validate the promising performance of our proposed model.,19274
"generalization paper, present In of exploits Hamiltonian we this HMC dynamics. a which \textit{non-canonical}","In this paper, we present a generalization of HMC which exploits \textit{non-canonical} Hamiltonian dynamics.",19275
We new LPCS efficiently. present perform also to a approach straightforward,We also present a new straightforward approach to perform LPCS efficiently.,19276
Further extension considered. to binary problems tackle is also classification,Further extension to tackle binary classification problems is also considered.,19277
our at experiments preference Psychophysical perceptual the filling-in reveal spot. in horizontal blind,Psychophysical experiments reveal our horizontal preference in perceptual filling-in at the blind spot.,19278
"hand, filling-in. of preference in is case vertical other in On exhibited the tolerance the","On the other hand, vertical preference is exhibited in the case of tolerance in filling-in.",19279
perception? anisotropy What causes in our this,What causes this anisotropy in our perception?,19280
filling-in. filling-in vertical horizontal preference preference tolerance of and in in,horizontal preference in filling-in and vertical preference in tolerance of filling-in.,19281
optimally solved polynomial Sequence time. alignment be in can,Sequence alignment can be optimally solved in polynomial time.,19282
"compare we several functions Finally, using topological SANA. objective","Finally, we compare several topological objective functions using SANA.",19283
a appearance. image on deformations image method based We predict patch-wise present to,We present a method to predict image deformations based on patch-wise image appearance.,19284
areas show that ambiguous uncertainty We deformations. highlights of deformation,We show that deformation uncertainty highlights areas of ambiguous deformations.,19285
The stimulus-based novel onset a two paper (vmoBCI). of visual presents interfaces brain-computer motion study,The paper presents a study of two novel visual motion onset stimulus-based brain-computer interfaces (vmoBCI).,19286
are afferent screen computer Two settings center patterns. motion to compared with and efferent a,Two settings are compared with afferent and efferent to a computer screen center motion patterns.,19287
a formulation. novel propose odometry visual We direct sparse,We propose a novel direct sparse visual odometry formulation.,19288
several of hours We method on three video. datasets our evaluate comprising different thoroughly,We thoroughly evaluate our method on three different datasets comprising several hours of video.,19289
vision deep based paper target learning a for approach introduces based This novel single tracking.,This paper introduces a novel deep learning based approach for vision based single target tracking.,19290
an We correspondence and higher significantly accuracy. recovery producing a guaranteeing alternative technique propose bijective,We propose an alternative recovery technique guaranteeing a bijective correspondence and producing significantly higher accuracy.,19291
moment. favorably use methods most results to the Despeckling at compare of in the,Despeckling results compare favorably to the most of methods in use at the moment.,19292
"(e.g., persons, description The are actions). typically objects, heterogeneous in the concepts","The concepts in the description are typically heterogeneous (e.g., objects, persons, actions).",19293
on can be also location concepts these constraints from Certain the inferred description.,Certain location constraints on these concepts can also be inferred from the description.,19294
"proposed using approximation. posterior for inference variational formulation the mean-field we Finally, develop","Finally, we develop posterior inference for the proposed formulation using mean-field variational approximation.",19295
yet applied by learned retrieval. have Representations to GANs not been,Representations learned by GANs have not yet been applied to retrieval.,19296
"multi-branch background jitters. model noise method, our proposed to In suppress a from is","In our method, a multi-branch model is proposed to suppress noise from background jitters.",19297
using by show the that our The model convolutional achieves state-of-art results features. only,The results show that our model achieves the state-of-art by only using convolutional features.,19298
misspecification. outliers and are against Resulting estimators generally robust,Resulting estimators are generally robust against outliers and misspecification.,19299
are properties Convergence treated. two algorithms of the,Convergence properties of the two algorithms are treated.,19300
classifier engine. an a to use We the learning build game emotion approach as deep,We use a deep learning approach to build an emotion classifier as the game engine.,19301
subgraph main maximum-weight The activity idea is a problem. as connected detection pose to,The main idea is to pose activity detection as a maximum-weight connected subgraph problem.,19302
several Semantic can recognition useful object be  visual parts for tasks. ,  Semantic object parts can be useful for several visual recognition tasks.,19303
"addressed Neural Networks have Lately, tasks Convolutional using achieving (CNN), these been results. outstanding","Lately, these tasks have been addressed using Convolutional Neural Networks (CNN), achieving outstanding results.",19304
in representation. learn CNNs parts work this study internal whether semantic we their In,In this work we study whether CNNs learn semantic parts in their internal representation.,19305
"depths, emergence explore different layers, and this levels. for We network supervision","We explore this emergence for different layers, network depths, and supervision levels.",19306
"semantics. power connections discriminative several we Moreover, explore and between","Moreover, we explore several connections between discriminative power and semantics.",19307
"such analysis. for used mixture cluster are models, as models, probabilistic Hierarchical","Hierarchical probabilistic models, such as mixture models, are used for cluster analysis.",19308
and These types latent. of have models two variables: observable,These models have two types of variables: observable and latent.,19309
site procedure observed for evolutionary A from frequency described matrices rate data. estimating is,A procedure is described for estimating evolutionary rate matrices from observed site frequency data.,19310
to is not matrix be the reversible. particular In rate assumed,In particular the rate matrix is not assumed to be reversible.,19311
"Structure-from-Motion camera distributed for We (SfM). a the introduce model, model novel","We introduce the distributed camera model, a novel model for Structure-from-Motion (SfM).",19312
J. and and B. Given Bush L.,Given and B. L. Bush and J.,19313
model neural deconvolution. approach a for network multi-frame present We blind,We present a neural network model approach for multi-frame blind deconvolution.,19314
increases rate. production pipelining This the protein,This pipelining increases the protein production rate.,19315
natural question protein rate. and maximizes production density important ribosomal the is A what,A natural and important question is what ribosomal density maximizes the protein production rate.,19316
whose fish pelagic are close Tuna large are to panmixia. populations,Tuna are large pelagic fish whose populations are close to panmixia.,19317
be can film parts. and (plasma) separated background foreground Thin to blood easily (cell) images,Thin blood film images can easily be separated to foreground (cell) and background (plasma) parts.,19318
of takes the move each of lots analysis Unfortunately times.,Unfortunately the analysis of each move takes lots of times.,19319
"At any architecture network testing the time, inference or change overhead. DSD doesn't incur","At testing time, DSD doesn't change the network architecture or incur any inference overhead.",19320
Our (negative) this robust proposed to uses minimize regret model method (inaccurate) w.r.t. the directly,Our proposed robust method uses this (inaccurate) model to directly minimize the (negative) regret w.r.t.,19321
approximate NP-hard show and our We formulation is an propose that algorithm.,We show that our formulation is NP-hard and propose an approximate algorithm.,19322
"their be precomputation parameters algorithm fixed flexibility. offline, However, requires certain limiting","However, precomputation requires certain algorithm parameters be fixed offline, limiting their flexibility.",19323
representing structures values. and brain their several for stimuli has both work Prior identified critical,Prior work has identified several brain structures critical for representing both stimuli and their values.,19324
"which define have functions, for gradient-like piecewise quantities general Hence, we new, applicability. wide constant","Hence, we define new, general gradient-like quantities for piecewise constant functions, which have wide applicability.",19325
operator feature regions. extraction extract useful features from to the target The aims,The feature extraction operator aims to extract useful features from the target regions.,19326
"sites After safety and pose a concerns. of earthquake, multitude health an disaster","After an earthquake, disaster sites pose a multitude of health and safety concerns.",19327
be shows This of with proof constructed surfaces can minimally textured modified that cystoscopes. concept,This proof of concept shows that textured surfaces can be constructed with minimally modified cystoscopes.,19328
textures with on superimposed phantoms and known dimensions. are and shape human Pig bladder,Pig and human bladder textures are superimposed on phantoms with known shape and dimensions.,19329
the correctness proposed illustrate Example simulation and results algorithms. of the efficiency,Example results illustrate the correctness and efficiency of the proposed simulation algorithms.,19330
an part environment is surveillance multi-camera in re-identification systems. modern a of Person important,Person re-identification in a multi-camera environment is an important part of modern surveillance systems.,19331
from re-identification. pixels) for spectral skin signatures,spectral signatures from skin pixels) for re-identification.,19332
configuration studies or contrast spatial In (e.g. to assume mostly existing that face pre-aligned ignore,In contrast to existing studies that mostly ignore or assume pre-aligned face spatial configuration (e.g.,19333
proposes network learning This and algorithms for and supervised hashing. models unsupervised binary work deep,This work proposes deep network models and learning algorithms for unsupervised and supervised binary hashing.,19334
Our output the design binary novel codes. hidden directly to constrains layer network one,Our novel network design constrains one hidden layer to directly output the binary codes.,19335
"our in Furthermore, we include objective property similarity function. preserving","Furthermore, we include similarity preserving property in our objective function.",19336
"resulting and is balance binary, independence, difficult Our with to optimization solve. these constraints","Our resulting optimization with these binary, independence, and balance constraints is difficult to solve.",19337
propose careful optimization and to relaxation. attack alternating with We it,We propose to attack it with alternating optimization and careful relaxation.,19338
"different most However, such cross-layer ignore when complexities. existing works classifying heterogeneities samples of","However, most existing works ignore such cross-layer heterogeneities when classifying samples of different complexities.",19339
"on we top of multiple this Towards layers. target, introduce classifiers multiple","Towards this target, we introduce multiple classifiers on top of multiple layers.",19340
"Allowing conveniently be training, our into networks. state-of-the-art embedded can method end-to-end deep for","Allowing for end-to-end training, our method can be conveniently embedded into state-of-the-art deep networks.",19341
our significantly that performance. the We boosts show architecture proposed,We show that our proposed architecture significantly boosts the performance.,19342
unmixing hyperspectral associated and models mixture two novel paper This algorithms. presents,This paper presents two novel hyperspectral mixture models and associated unmixing algorithms.,19343
order generalizes The NL into models by account model taking terms. bilinear higher interaction,The NL model generalizes bilinear models by taking into account higher order interaction terms.,19344
General scarce. laws are parasitology in ecological,General laws in ecological parasitology are scarce.,19345
of years. significant in received has signals attention Spare representation recent,Spare representation of signals has received significant attention in recent years.,19346
handle to data. We nonlinearly kernelized also its develop variant and separable present,We also develop and present its kernelized variant to handle nonlinearly separable data.,19347
developed algorithm. of this angular an also preserving locality variant We,We also developed an angular locality preserving variant of this algorithm.,19348
alternate relating quantifying is for to duration. an logic activity approach Fuzzy uncertainty,Fuzzy logic is an alternate approach for quantifying uncertainty relating to activity duration.,19349
fuzzy The the problem the application to is of maximum. be shown,The problem is shown to be the application of the fuzzy maximum.,19350
and geometry Our information. material categories goal recognize images to using is,Our goal is to recognize material categories using images and geometry information.,19351
"available. applications, geometry In many management, information is construction as coarse such","In many applications, such as construction management, coarse geometry information is available.",19352
elements maintained counter-regulating systems. by living equilibrium Dynamic is in,Dynamic equilibrium is maintained by counter-regulating elements in living systems.,19353
{\alpha} glucagon and and pancreatic homeostasis. The glucose for produce {\beta} insulin cells,The pancreatic {\alpha} and {\beta} cells produce glucagon and insulin for glucose homeostasis.,19354
"effort although some families amplifiers was known, few to required of prove it. A are","A few families of amplifiers are known, although some effort was required to prove it.",19355
also are Monte Some larger experiments graphs. methods numerical using performed for Carlo,Some numerical experiments using Monte Carlo methods are also performed for larger graphs.,19356
common reality a (e.g. Such multi-body is in scenario non-rigid,Such a multi-body non-rigid scenario is common in reality (e.g.,19357
alternating method (ADMM). optimization with multipliers We resultant direction the solve of the,We solve the resultant optimization with the alternating direction method of multipliers (ADMM).,19358
forested shaded sites over Vireos coffee sites sun coffee selected sites. preferentially and,Vireos preferentially selected forested sites and shaded coffee sites over sun coffee sites.,19359
to into interactions potential conservation. for the occupancy Incorporating has models monitoring species improve,Incorporating species interactions into occupancy models has the potential to improve monitoring for conservation.,19360
called has increasingly task This that recently a segmentation become important. instance is,This is a task called instance segmentation that has recently become increasingly important.,19361
"we In easily parsable by a detail, and concrete syntax applications. effectively present","In detail, we present a concrete syntax easily and effectively parsable by applications.",19362
"such. the software be can deployed as Alternatively,","Alternatively, the software can be deployed as such.",19363
"results for are of Further, deployment local and both benchmarks web presented.","Further, results of benchmarks for both local and web deployment are presented.",19364
the experiments. behavioral two of We results report,We report the results of two behavioral experiments.,19365
"received and positive their negative The decisions, incorrect correct participants for respectively. rewards and","The participants received positive and negative rewards for their correct and incorrect decisions, respectively.",19366
"of first experiment, were trials In two and easy. randomly: hard the intermixed types","In the first experiment, two types of trials were intermixed randomly: hard and easy.",19367
small was decision for a trials. The threshold optimal hard to adopt strategy,The optimal strategy was to adopt a small decision threshold for hard trials.,19368
participants strategy. this that did learn results the showed several The not of simple,The results showed that several of the participants did not learn this simple strategy.,19369
the proposed effectiveness justify of techniques. the experiments Systematic,Systematic experiments justify the effectiveness of the proposed techniques.,19370
"have focus this long a methods machine of regard, In been attention. learning","In this regard, machine learning methods have long been a focus of attention.",19371
methods the learning also have underline with intrinsic We methods. deep learning differences other machine,We also underline the intrinsic differences deep learning methods have with other machine learning methods.,19372
for models significant saliency in images still years. popularity gained have Computational recent,Computational saliency models for still images have gained significant popularity in recent years.,19373
"annotated sparsely to frames. video Additionally, datasets tend have","Additionally, video datasets tend to have sparsely annotated frames.",19374
"for a First, neural train network a pseudo-labeler, that domain-adapted convolutional we is, object detection.","First, we train a pseudo-labeler, that is, a domain-adapted convolutional neural network for object detection.",19375
"valuable even demonstrate that neighboring labels. can Our absent frames information, experiments provide","Our experiments demonstrate that neighboring frames can provide valuable information, even absent labels.",19376
video on approach densely verify datasets. annotated our We two,We verify our approach on two densely annotated video datasets.,19377
"competitive slightly, We interactions, likely. is affect coexistence when even show can stable mutations that","We show that when mutations can affect competitive interactions, even slightly, stable coexistence is likely.",19378
evolutionary our the populations. the in clonal We discuss of of large results dynamics context,We discuss our results in the context of the evolutionary dynamics of large clonal populations.,19379
"paper two-step for This unsupervised) proposes in two (supervised, inferencing codes hashing. binary approaches","This paper proposes two approaches for inferencing binary codes in two-step (supervised, unsupervised) hashing.",19380
unsupervised formulation first hashing. both and an for supervised introduce We unified,We first introduce an unified formulation for both supervised and unsupervised hashing.,19381
"as bit Binary Quadratic learning of we (BQP). the Problem one Then, cast a","Then, we cast the learning of one bit as a Binary Quadratic Problem (BQP).",19382
propose BQP. to approaches two We solve,We propose two approaches to solve BQP.,19383
neural problems. networks Deep achieved success of practical a range wide remarkable have in,Deep neural networks have achieved remarkable success in a wide range of practical problems.,19384
"low-level Random In transition paper, we Forests patterns. this expression to Conditional use capture","In this paper, we use Conditional Random Forests to capture low-level expression transition patterns.",19385
or pairs point of feature evaluated upon images. variations) are movements texture,feature point movements or texture variations) are evaluated upon pairs of images.,19386
"conditioned also can head FER. PCRF Moreover, pose be dynamic estimation multi-view for on collections","Moreover, PCRF collections can also be conditioned on head pose estimation for multi-view dynamic FER.",19387
temporal cost per-pixel information. comes The the resolution the at of losing high intensity familiar,The high temporal resolution comes at the cost of losing the familiar per-pixel intensity information.,19388
use Current the pixel on to Convolutional representations. ModelNet submissions Challenge (CNNs) leading Neural Networks,Current leading submissions to the ModelNet Challenge use Convolutional Neural Networks (CNNs) on pixel representations.,19389
"CNN architectures. this, new Volumetric do To (V-CNN) we introduce","To do this, we introduce new Volumetric CNN (V-CNN) architectures.",19390
Uncertain have streams data applications. in been widely many generated Web,Uncertain data streams have been widely generated in many Web applications.,19391
streams anomaly uncertainty far detection sensor from more in streams data data challenging. makes The,The uncertainty in data streams makes anomaly detection from sensor data streams far more challenging.,19392
the corrected stream. use We in data anomaly classification methods detection for,We use classification methods for anomaly detection in the corrected data stream.,19393
morphology. strongly of their with are coupled neurons Functional properties,Functional properties of neurons are strongly coupled with their morphology.,19394
neuronal morphological in spines. of characteristics Changes activity alter dendritic,Changes in neuronal activity alter morphological characteristics of dendritic spines.,19395
neuroscientists of understand help the underlying relationships. dendritic can Shape spines analysis,Shape analysis of dendritic spines can help neuroscientists understand the underlying relationships.,19396
"perspective. we In a paper, to address aim issues clustering this presenting by these","In this paper, we aim to address these issues by presenting a clustering perspective.",19397
This argument of observation the types. supports intermediate shape,This observation supports the argument of intermediate shape types.,19398
studying indispensable for learning inequalities generalization Concentration the are capacity models. tools of,Concentration inequalities are indispensable tools for studying the generalization capacity of learning models.,19399
"Hoeffding's the inequalities giving independent of distribution. used, McDiarmid's data are commonly and bounds","Hoeffding's and McDiarmid's inequalities are commonly used, giving bounds independent of the data distribution.",19400
"bounds hierarchy-bounded unbounded obtain for generalization loss we and functions. Furthermore,","Furthermore, we obtain generalization bounds for unbounded and hierarchy-bounded loss functions.",19401
Finally the extensions we applications of to potential discuss our learning theory.,Finally we discuss the potential applications of our extensions to learning theory.,19402
"terms of precision distinctiveness, tracking and is evaluation tracking Our speed. in performed","Our evaluation is performed in terms of distinctiveness, tracking precision and tracking speed.",19403
effort annotate data. requires to This considerable human,This requires considerable human effort to annotate data.,19404
"is complexity. modern favorable Considering its hashing the for low datasets, of scale","Considering the scale of modern datasets, hashing is favorable for its low complexity.",19405
the associated a perform scaling finite-size exponents. We and we scaling obtain,We perform a finite-size scaling and we obtain the associated scaling exponents.,19406
of direction methods images. one palmprint of usually only the the Conventional most dominant capture,Conventional methods usually capture the only one of the most dominant direction of palmprint images.,19407
rhythms. brain They especially have useful in understanding proven,They have proven especially useful in understanding brain rhythms.,19408
"prior starts of appearances. captures superpixel from segmentation, which effectively feature SDF information","SDF starts from superpixel segmentation, which effectively captures prior information of feature appearances.",19409
The fitting the feature reduce beneficial for appearances computational to methods. deterministic are complexity,The feature appearances are beneficial to reduce the computational complexity for deterministic fitting methods.,19410
demonstrate basis. per-pixel effective is on results a highly algorithm The the,The results demonstrate the algorithm is highly effective on a per-pixel basis.,19411
response with binary is multilevel Fitting challenging. models cross-classified,Fitting cross-classified multilevel models with binary response is challenging.,19412
"images Natural under are many including shape, generated illumination factors, pose, etc.","Natural images are generated under many factors, including shape, pose, illumination etc.",19413
to LEPs well be categorical combined describe expressions can facial (AUs). as Action Units as,LEPs can be combined to describe categorical facial expressions as well as Action Units (AUs).,19414
"provided an Furthermore, autoencoder network. confidence can weighted be by scores LEPs by","Furthermore, LEPs can be weighted by confidence scores provided by an autoencoder network.",19415
variations detection. that to be face remain real-word a pose challenge confronts Large,Large pose variations remain to be a challenge that confronts real-word face detection.,19416
automatically face/non-face This to the scale of selects learning patterns. the supervised transformations differentiate best,This supervised learning of the transformations automatically selects the best scale to differentiate face/non-face patterns.,19417
information independent neural processes deep that propose along two dual-stream pathways. network a We,We propose a dual-stream deep neural network that processes information along two independent pathways.,19418
algorithms. many for vision computer Illumination effects cause problems,Illumination effects cause problems for many computer vision algorithms.,19419
present illumination-invariant system robust generation. a interactive image user-friendly for We,We present a user-friendly interactive system for robust illumination-invariant image generation.,19420
by illumination-invariant input. is the The of derivation image guided user,The derivation of illumination-invariant image is guided by the user input.,19421
"results present our Finally, some method. we evaluation of","Finally, we present some evaluation results of our method.",19422
classify textual connect and by visual to the methods modalities. their mechanism We,We classify methods by their mechanism to connect the visual and textual modalities.,19423
knowledge structured with also discuss modular bases. and We interface that architectures memory-augmented,We also discuss memory-augmented and modular architectures that interface with structured knowledge bases.,19424
are candidate confounding These terms. interaction the,These are the candidate confounding interaction terms.,19425
data. are issues Unfortunately and analysing longitudinal there many complexities when observational underlying,Unfortunately there are many complexities and underlying issues when analysing longitudinal observational data.,19426
be considered can algorithm a of clustering. Our extension as an spectral,Our algorithm can be a considered as an extension of spectral clustering.,19427
"Systems interpretable Rule-Based so-called provide potential the have i.e. Fuzzy Classification to (FRBCSs) classifiers,","Fuzzy Rule-Based Classification Systems (FRBCSs) have the potential to provide so-called interpretable classifiers, i.e.",19428
resulting an excessive in without rules. of number,without resulting in an excessive number of rules.,19429
"can step some in dimensional measurements, the high model), situations (informative prove However, inefficient. resampling","However, in some situations (informative measurements, high dimensional model), the resampling step can prove inefficient.",19430
budget. which validate the take results our simulations computational account Finally carefully into via we,Finally we validate our results via simulations which carefully take into account the computational budget.,19431
listed new experiments Further are proposed. predictions are and,Further predictions are listed and new experiments are proposed.,19432
"Finally, is of interpretation results the presented. an above","Finally, an interpretation of the above results is presented.",19433
The optical of atmospheric media survey discussing scattering formation. image begins models and with the,The survey begins with discussing the optical models of atmospheric scattering media and image formation.,19434
"knowledge, techniques. first our this evaluate benchmark to metrics the is numerical using To dehazing","To our knowledge, this is the first benchmark using numerical metrics to evaluate dehazing techniques.",19435
aid cardiac in diagnostic Echocardiography diseases. an plays in important part,Echocardiography plays an important part in diagnostic aid in cardiac diseases.,19436
efficiency. It method both outperforms in the classification demonstrates accuracy and computational our that baselines,It demonstrates that our method outperforms the baselines in both classification accuracy and computational efficiency.,19437
propose new a on based homography. color We color method correction,We propose a new color correction method based on color homography.,19438
color that homography solving calibration. problem more the demonstrate leads to Experiments accurate,Experiments demonstrate that solving the color homography problem leads to more accurate calibration.,19439
"isolation. recognition trajectories years, hand considering gesture been mostly addressed in hand has Over the","Over the years, hand gesture recognition has been mostly addressed considering hand trajectories in isolation.",19440
"sign languages, a defined are on However, region). hand particular context most (body gestures in","However, in most sign languages, hand gestures are defined on a particular context (body region).",19441
dataset in YouTube verify dataset MultiPIE static face Celebrities(YTC). RRNN We and video face,We verify RRNN in static face dataset MultiPIE and face video dataset YouTube Celebrities(YTC).,19442
method. experimental of demonstrate the The the proposed results comprehensive RRNN effectiveness,The comprehensive experimental results demonstrate the effectiveness of the proposed RRNN method.,19443
"an this representation efficient we model In for work, propose classification. image","In this work, we propose an efficient image representation model for classification.",19444
consider We neuron process. integrate-and-fire with stimulated Poisson a single Leaky,We consider a single Leaky integrate-and-fire neuron stimulated with Poisson process.,19445
"Our image consensus, taking towards consistency. all label account deforms approach volumes and into similarities","Our approach deforms all volumes towards consensus, taking into account image similarities and label consistency.",19446
incorporate Our can pipeline classifier any similarity metric. and,Our pipeline can incorporate any classifier and similarity metric.,19447
the convergence process the theory population risk to risk. implies uniform of empirical empirical Classical,Classical empirical process theory implies uniform convergence of the empirical risk to the population risk.,19448
to regression non-convex apply We in very classification binary robust and result this high-dimension.,We apply this result to non-convex binary classification and robust regression in very high-dimension.,19449
across recognize images. can many different People beyond modalities scenes natural,People can recognize scenes across many different modalities beyond natural images.,19450
"across In this cross-modal learn to investigate transfer representations scene that how paper, modalities. we","In this paper, we investigate how to learn cross-modal scene representations that transfer across modalities.",19451
"this a study introduce cross-modal we To scene dataset. new problem,","To study this problem, we introduce a new cross-modal scene dataset.",19452
representation our representations across Our scene suggest experiments modalities that help retrieval. for transfer can,Our experiments suggest that our scene representation can help transfer representations across modalities for retrieval.,19453
for types two of features assignment We our utilize procedure.,We utilize two types of features for our assignment procedure.,19454
between pair Pairwise relationship of visual features the viewers. of encode a content the,Pairwise features encode the relationship between the visual content of a pair of viewers.,19455
Cells are the system immune with anti-inflammatory opposing signals. pro- and confronted of,Cells of the immune system are confronted with opposing pro- and anti-inflammatory signals.,19456
detection and MS-CNN sub-network. of The a consists a sub-network proposal,The MS-CNN consists of a proposal sub-network and a detection sub-network.,19457
These scale-specific object detector. to detectors complementary strong are combined produce a multi-scale,These complementary scale-specific detectors are combined to produce a strong multi-scale object detector.,19458
"unified optimizing learned by a network loss. multi-task end-to-end, The is","The unified network is learned end-to-end, by optimizing a multi-task loss.",19459
architecture performs coarse-to-fine with additional an warping Our of intensity deep individual pixels. correction,Our deep architecture performs coarse-to-fine warping with an additional intensity correction of individual pixels.,19460
"in its and The evaluated robustness i.e. is capabilities, proposed to various approach demonstrate applications","The proposed approach is evaluated in various applications to demonstrate its robustness and capabilities, i.e.",19461
"and Face FG-NET. Parts Wild in (LFPW), Helen the Labeled","Labeled Face Parts in the Wild (LFPW), Helen and FG-NET.",19462
from increases labor. pregnancy showed significantly that node of Results the to each strength,Results showed that the strength of each node significantly increases from pregnancy to labor.,19463
"learn continuous world We meta-architecture raw, a can novel video. from dynamics network that present","We present a novel network meta-architecture that can learn world dynamics from raw, continuous video.",19464
built implementation We where perceptrons. demonstrate components architecture are of multi-layer from this an the,We demonstrate an implementation of this architecture where the components are built from multi-layer perceptrons.,19465
tracking par exceeding or performance Results state-of-the-art show algorithms. with on,Results show performance on par with or exceeding state-of-the-art tracking algorithms.,19466
either trained regimes. unsupervised-then-briefly-supervised fully or The supervised tracker in can be,The tracker can be trained in either fully supervised or unsupervised-then-briefly-supervised regimes.,19467
Person an and challenging problem vision. in is computer re-identification open,Person re-identification is an open and challenging problem in computer vision.,19468
Convolutional visual playing Neural state-of-the-art roles (CNNs) important are Deep recognition. Networks in,Deep Convolutional Neural Networks (CNNs) are playing important roles in state-of-the-art visual recognition.,19469
show Experimental classification. GNPP gain results accuracy produces in image that consistent significant and,Experimental results show that GNPP produces significant and consistent accuracy gain in image classification.,19470
"ontologies. of of we focus on SHIQ In paper, descriptions generating this particularly based individuals","In this paper, we particularly focus on generating descriptions of individuals of SHIQ based ontologies.",19471
"representation. performed image classification finally, probabilistic on And image such is","And finally, image classification is performed on such probabilistic image representation.",19472
is cancer. of the one Melanoma dangerous of forms most,Melanoma is one of the most dangerous forms of cancer.,19473
"Finally, for demonstrate utilize neurons semantically can relevant regions. we we finding highly-activating that","Finally, we demonstrate that we can utilize highly-activating neurons for finding semantically relevant regions.",19474
"not test for The required the information is called data, ""privileged prior"". kinematic-layout thus information","The kinematic-layout information is not required for the test data, thus called ""privileged information prior"".",19475
"we querying for work, system search. this content-based large-scale video propose In fast video a","In this work, we propose a fast content-based video querying system for large-scale video search.",19476
two distinguished from with similar system contributions. works is The proposed major,The proposed system is distinguished from similar works with two major contributions.,19477
and hashing mechanisms. repeated First efficient content usage joint contribution of representation is superiority of,First contribution is superiority of joint usage of repeated content representation and efficient hashing mechanisms.,19478
"proposed (PGS), procedure, training. purpose back-propagation A network is suppression special peak for gradient","A special purpose back-propagation procedure, peak gradient suppression (PGS), is proposed for network training.",19479
of There perhaps that which language is emotions. suggest numerous music are studies the truly,There are numerous studies which suggest that perhaps music is truly the language of emotions.,19480
induce pictures were participants. the context to used in IAPS The emotional,The IAPS pictures were used to induce the emotional context in participants.,19481
and focuses below. on urban echelons of specifically This at program battalion battles,This program focuses specifically on urban battles at echelons of battalion and below.,19482
"his faster, more he able should achieve efficiently. to be safer, goals operational and","he should be able to achieve his operational goals safer, faster, and more efficiently.",19483
problem alone of the No can complex face solve recognition. classifier single,No single classifier can alone solve the complex problem of face recognition.,19484
their enhances Researchers recognition usually base found that classifiers combining some rate.,Researchers found that combining some base classifiers usually enhances their recognition rate.,19485
are the system. base of weaknesses The reflected classifiers resulting the combined on,The weaknesses of the base classifiers are reflected on the resulting combined system.,19486
"for is combining proposed. unstable, In classifiers this system performance work, low a","In this work, a system for combining unstable, low performance classifiers is proposed.",19487
shows and of parameters. recognition high system The a reduced rate stability remarkable using number,The system shows remarkable stability and high recognition rate using a reduced number of parameters.,19488
has addressed as a Detecting general topic pedestrian beyond been detection. special object arguably,Detecting pedestrian has been arguably addressed as a special topic beyond general object detection.,19489
gap. work Our to technological fill the aims above,Our work aims to fill the above technological gap.,19490
"entailment information in enact textual documents, conflicting we medical technology. To identify","To identify conflicting information in medical documents, we enact textual entailment technology.",19491
"encapsulate ontologies. we To rely on medical knowledge, existing","To encapsulate existing medical knowledge, we rely on ontologies.",19492
"disagreement, design To a we multi-agent framework. this mediation within harmonize protocols","To harmonize this disagreement, we design mediation protocols within a multi-agent framework.",19493
also our report details present in baseline setup results. experiment We promising and,We also present in details our experiment setup and report promising baseline results.,19494
Our problems in computer vision. benchmark task to could of largest lead one the classification,Our benchmark task could lead to one of the largest classification problems in computer vision.,19495
Ulyanov method It we stylization fast introduced the in revisit paper et. this,It this paper we revisit the fast stylization method introduced in Ulyanov et.,19496
method generation. resulting to train can architectures high-performance real-time image be The for used,The resulting method can be used to train high-performance architectures for real-time image generation.,19497
"of working use The mammalian phases and is metaphor we the ""day"" ""night"" life. the","The working metaphor we use is the ""day"" and the ""night"" phases of mammalian life.",19498
"of have long-term developmental retardation infant disabilities. and risk a mortality, higher Preemies","Preemies have a higher risk of infant mortality, developmental retardation and long-term disabilities.",19499
"Predicting experienced the most even is birth preterm clinicians. for difficult,","Predicting preterm birth is difficult, even for the most experienced clinicians.",19500
databases hospital take a of exploiting approach by normal We operations. different,We take a different approach by exploiting databases of normal hospital operations.,19501
"a more can efficiently. solved be purely propose contrast, which we continuous formulation In","In contrast, we propose a purely continuous formulation which can be solved more efficiently.",19502
"mutation-selection-drift steady limit, adopts eventually balance. by characterized the In state a population this","In this limit, the population eventually adopts a steady state characterized by mutation-selection-drift balance.",19503
"we epistatic, extension develop to arbitrary, possibly of fitness formula this an Here landscapes.","Here we develop an extension of this formula to arbitrary, possibly epistatic, fitness landscapes.",19504
Gauss-Gamma is dense general too but in solutions. the provides fastest The Bayesian algorithm,The Bayesian Gauss-Gamma is in general the fastest algorithm but provides too dense solutions.,19505
but is sparse Gaussian/inverse-Gamma fast in maximum also The very general solutions. provides likelihood very,The maximum likelihood Gaussian/inverse-Gamma is also very fast but provides in general very sparse solutions.,19506
of the very characteristic consuming. systems time CAD new makes a This design,This characteristic makes the design of new CAD systems a very time consuming.,19507
published on risk. is structure derived Model literature ASI from,Model structure is derived from published literature on ASI risk.,19508
"hierarchical based of representation, abstraction model and method the two-level on knowledge e.g. The is","The method is based on the hierarchical two-level model of abstraction and knowledge representation, e.g.",19509
feedback between interfaces both are inter-level described. planners and is examined and Interaction loops,Interaction between both planners is examined and inter-level interfaces and feedback loops are described.,19510
"construct video first auxiliary this from To sequences. discriminative set training a we end,","To this end, we first construct a discriminative training set from auxiliary video sequences.",19511
on network then is trained this classification deep set. offline A training neural,A deep classification neural network is then trained offline on this training set.,19512
about noising. of elucidating the further concept This manuscript is,This manuscript is about further elucidating the concept of noising.,19513
to learning techniques Deep contributed this significantly.,Deep learning techniques contributed to this significantly.,19514
"the same method computationally time, At still the is efficient.","At the same time, the method still is computationally efficient.",19515
exchange. breath % on per gas based,% based on per breath gas exchange.,19516
to propose attribute-based In an we alleviate this strategy problem. paper this,In this paper we propose an attribute-based strategy to alleviate this problem.,19517
This passed the infer is content. generative missing the then model to encoding through,This encoding is then passed through the generative model to infer the missing content.,19518
viewpoints epipolar very very between cameras difficult. Computing with geometry different often is the,Computing the epipolar geometry between cameras with very different viewpoints is often very difficult.,19519
includes the These fail moving when methods multiple objects. scene,These methods fail when the scene includes multiple moving objects.,19520
"is paper, optimal presented. problem this statistically the (PnP) solution to Perspective-n-Point a In","In this paper, a statistically optimal solution to the Perspective-n-Point (PnP) problem is presented.",19521
"features paper, retrieval. In for image representations local we and image this review","In this paper, we review local features and image representations for image retrieval.",19522
"image addition, In briefly learning-based retrieval are reviewed. for approaches deep recent","In addition, recent deep learning-based approaches for image retrieval are briefly reviewed.",19523
"the of Salient problem Object Subitizing, i.e. We study","We study the problem of Salient Object Subitizing, i.e.",19524
in holistic the image and number the using objects predicting cues. an of existence salient,predicting the existence and the number of salient objects in an image using holistic cues.,19525
"methods recent advantage. to maintain this for image fail (SR) However, single super-resolution","However, recent methods for single image super-resolution (SR) fail to maintain this advantage.",19526
"work, for this propose a fine-grained we In generalization novel scene recognition. domain approach","In this work, we propose a novel domain generalization approach for fine-grained scene recognition.",19527
then into transform probabilities for occurrences such We each scene image. scene,We then transform such occurrences into scene probabilities for each scene image.,19528
"the dataset in settings. To scene new cross-domain proposed a fine-grained evaluate we method, propose","To evaluate the proposed method, we propose a new fine-grained scene dataset in cross-domain settings.",19529
Our competing paradigms. main combines method advantages the two of the,Our method combines the advantages of the two main competing paradigms.,19530
We enable region-based to semantic for approaches training modern to how end-to-end segmentation. modify show,We show how to modify modern region-based approaches to enable end-to-end training for semantic segmentation.,19531
layer Region-of-Interest This layer. a achieved a is pooling region-to-pixel differentiable differentiable via free-form and,This is achieved via a differentiable region-to-pixel layer and a differentiable free-form Region-of-Interest pooling layer.,19532
In single present super-resolution. we map a novel for this work approach depth,In this work we present a novel approach for single depth map super-resolution.,19533
"to with recover method depth variational high-resolution a maps. deep networks, convolutional accurate","deep convolutional networks, with a variational method to recover accurate high-resolution depth maps.",19534
"is To training of networks, a corpus train large ground-truth required. data deep with accurate","To train deep networks, a large corpus of training data with accurate ground-truth is required.",19535
"processing. unpleasing, further is visually can This but only hamper not","This is not only visually unpleasing, but can hamper further processing.",19536
perfect scale the demonstrate invariance We measure in the case. satisfies the continuous property,We demonstrate the measure satisfies the perfect scale invariance property in the continuous case.,19537
"keypoint description. algorithms for discrete detection In setting, the we and blob derive","In the discrete setting, we derive algorithms for blob detection and keypoint description.",19538
they and architectures precomputation. but allow only accurate feature Classification slow partial Hybrid are since,Classification and Hybrid architectures are accurate but slow since they allow only partial feature precomputation.,19539
of correspondence the azimuth) even more making street unknown is views difficult.,azimuth) of the street views is unknown making correspondence even more difficult.,19540
explicit improves supervision prediction that orientation accuracy. also out turns location It,It turns out that explicit orientation supervision also improves location prediction accuracy.,19541
similarity the plaques in makes detection very soft This of intensity isolation and difficult.,This similarity in intensity makes the isolation and detection of soft plaques very difficult.,19542
aims for minimal develop This vulnerable with work dependency. segmentation to user plaques framework of,This work aims to develop framework for segmentation of vulnerable plaques with minimal user dependency.,19543
unsupervised work deals This with deblurring. image,This work deals with unsupervised image deblurring.,19544
as method a The process. we defined propose is two-step,The method we propose is defined as a two-step process.,19545
"reduction original we for an domain. technique wavelet in First, use noise","First, we use an original technique for noise reduction in wavelet domain.",19546
"work, between different this a we compression. techniques In image present of comparison","In this work, we present a comparison between different techniques of image compression.",19547
"organized is are image a to divided First, the scan. certain according which in blocks","First, the image is divided in blocks which are organized according to a certain scan.",19548
"applied, combined several or compression Later, techniques are alone.","Later, several compression techniques are applied, combined or alone.",19549
"are: etc. wavelets Karhunen-Loeve techniques (Haar's Such basis), Transform,","Such techniques are: wavelets (Haar's basis), Karhunen-Loeve Transform, etc.",19550
"a noisy, depth network estimate map. low-resolution a from high-resolution joint input noise-free, computes The","The joint network computes a noise-free, high-resolution estimate from a noisy, low-resolution input depth map.",19551
"Additionally, image the network. to intensity the is in a guide high-resolution used reconstruction","Additionally, a high-resolution intensity image is used to guide the reconstruction in the network.",19552
large deep of dataset such supervision. requires for The training a a network,The training of such a deep network requires a large dataset for supervision.,19553
"color and based with maps we physically corresponding high-quality generate a depth renderer. images Therefore,","Therefore, we generate high-quality depth maps and corresponding color images with a physically based renderer.",19554
we evaluation method In benchmarks. our exhaustive the on outperforms multiple an state-of-the-art that show,In an exhaustive evaluation we show that our method outperforms the state-of-the-art on multiple benchmarks.,19555
"In video data. a this approach human present paper, pose data-driven in for we tracking","In this paper, we present a data-driven approach for human pose tracking in video data.",19556
an parameter thermal after increase latter cycling. exhibit does indeed The,The latter parameter does indeed exhibit an increase after thermal cycling.,19557
on demonstrate clearly our state-of-the-art. benchmark results proposed Experimental the approach several that outperforms datasets,Experimental results on several benchmark datasets demonstrate that our proposed approach clearly outperforms the state-of-the-art.,19558
"settings. for global low-dimensional this in algorithm propose minimizer a we paper, general the In","In this paper, we propose a general algorithm for the global minimizer in low-dimensional settings.",19559
"real-time For of and its accuracy performance this reason, is significance. great","For this reason, its accuracy and real-time performance is of great significance.",19560
on these gene expression. of processes based are Many,Many of these processes are based on gene expression.,19561
"and be such can they fluctuations"", we ""timing Here, consider controlled. we ask how","Here, we consider such ""timing fluctuations"", and we ask how they can be controlled.",19562
Influence problems. of graphical decision compact a diagrams provide representation,Influence diagrams provide a compact graphical representation of decision problems.,19563
"probabilistic often a and they uncertainties utility values. However, both quantification on rely of full","However, often they rely on a full quantification of both probabilistic uncertainties and utility values.",19564
utilities correspond expected families approach Within of this polynomials. to,Within this approach expected utilities correspond to families of polynomials.,19565
diagrams characterize transformations as manipulations the We then of many of of standard polynomials. influence,We then characterize many of the standard manipulations of influence diagrams as transformations of polynomials.,19566
"adopt layers. we Third, sizes but filter smaller mapping more","Third, we adopt smaller filter sizes but more mapping layers.",19567
the updating from stems The schedule. stochasticity,The stochasticity stems from the updating schedule.,19568
deterministic dynamics latter former while a stochastic dynamics. produces The the a,The former produces a deterministic dynamics while the latter a stochastic dynamics.,19569
more A each setting for stochastic parameters considers general node. propensity updating,A more general stochastic setting considers propensity parameters for updating each node.,19570
paper This for estimating presents the propensity a SDDS. method for parameters,This paper presents a method for estimating the propensity parameters for SDDS.,19571
"Then of propensity parameters estimated. the the a genetic algorithm, with are use","Then with the use of a genetic algorithm, the propensity parameters are estimated.",19572
a interactions of variety (e.g. images wide Visual pairs in objects relationships of capture between,Visual relationships capture a wide variety of interactions between pairs of objects in images (e.g.,19573
"pushing bicycle"" riding ""man and bicycle""). ""man","""man riding bicycle"" and ""man pushing bicycle"").",19574
"their infrequent, (e.g. are Though relationships objects most","Though most relationships are infrequent, their objects (e.g.",19575
"""riding"" independently frequently. and occur more ""pushing"")","""riding"" and ""pushing"") independently occur more frequently.",19576
model examples. relationships of predict to few scale of Our from can thousands types a,Our model can scale to predict thousands of types of relationships from a few examples.,19577
"we Additionally, relationships objects the the as in image. localize boxes in bounding predicted the","Additionally, we localize the objects in the predicted relationships as bounding boxes in the image.",19578
image demonstrate content further understanding We based retrieval. can improve relationships that,We further demonstrate that understanding relationships can improve content based image retrieval.,19579
Network forms critical architecture on a function. constraint neuronal,Network architecture forms a critical constraint on neuronal function.,19580
Coefficients new Projection (POAC). The Approximation algorithm called is Onto,The new algorithm is called Projection Onto Approximation Coefficients (POAC).,19581
"the with a rate. Besides, coefficients, bigger compression elimination the we of obtain details subbands","Besides, with the elimination of the details subbands coefficients, we obtain a bigger compression rate.",19582
"this In new fundamentally to approach propose Datalog a evaluation. paper, we","In this paper, we propose a fundamentally new approach to Datalog evaluation.",19583
"linear purely another, tilde_Eq(DB). into translate then but matrix equations Eq(DB) We","We then translate Eq(DB) into another, but purely linear matrix equations tilde_Eq(DB).",19584
DB. solution to computing by is specified least the computing Hence tilde_Eq(DB) equivalent M_DB of,Hence computing the least solution of tilde_Eq(DB) is equivalent to computing M_DB specified by DB.,19585
programs. compute Datalog of that least Herbrand models experiments two the linear conducted We,We conducted two experiments that compute the least Herbrand models of linear Datalog programs.,19586
Theory in To of Programming Logic Practice appear (TPLP). and,To appear in Theory and Practice of Logic Programming (TPLP).,19587
the past area in been research active few an has Face decades. recognition,Face recognition has been an active research area in the past few decades.,19588
"from problem paper, we address this this perspectives. the two In following","In this paper, we address this problem from the following two perspectives.",19589
large on scale used have methods image for Hashing database. widely similarity efficient retrieval been,Hashing methods have been widely used for efficient similarity retrieval on large scale image database.,19590
imagery of wavelet Synthetic has Aperture (SAR) to despeckling improved application performance. transforms Radar The,The application of wavelet transforms to Synthetic Aperture Radar (SAR) imagery has improved despeckling performance.,19591
according estimated An to variance. initial the is value threshold noise,An initial threshold value is estimated according to the noise variance.,19592
The is from image by the thresholded logarithmic filtered then reconstruction obtained coefficients.,The filtered logarithmic image is then obtained by reconstruction from the thresholded coefficients.,19593
"time, to decomposition. of level exclusively process and single This applied the a first is","This process is applied a single time, and exclusively to the first level of decomposition.",19594
exponential function gives image the reconstructed image. this The final of filtered,The exponential function of this reconstructed image gives the final filtered image.,19595
label relies these The methods accuracy. on performance of,The performance of these methods relies on label accuracy.,19596
approaches object other for SwiDeN that problem. recognition the outperforms show results Experimental depiction-invariant,Experimental results show that SwiDeN outperforms other approaches for the depiction-invariant object recognition problem.,19597
"the between cortex neurons. shaping synaptic The with plastic, experience ongoing of connectivity interactions is","The synaptic connectivity of cortex is plastic, with experience shaping the ongoing interactions between neurons.",19598
"related the to of people. cases, these diseases are many consumption In food","In many cases, these diseases are related to the food consumption of people.",19599
"of In particular, focuses dishes. recognition the our and on work Mediterranean food","In particular, our work focuses on the recognition of Mediterranean food and dishes.",19600
automatic of a diet development basis for serve recognition. dataset will as This the,This dataset will serve as a basis for the development of automatic diet recognition.,19601
divides It landmarks elaborate and principal subset. into subset given,It divides given landmarks into principal subset and elaborate subset.,19602
Next subsets weight is weights. the gradually until of two equivalent subset have principal decreased,Next the weight of principal subset is gradually decreased until two subsets have equivalent weights.,19603
performance results Theoretical simulation. the are of by new confirmed algorithm superior and analysis,Theoretical analysis results and superior performance of the new algorithm are confirmed by simulation.,19604
technique propose objects a that referring We context understand to between expressions. integrates,We propose a technique that integrates context between objects to understand referring expressions.,19605
MIL max-margin utilize training for based objective the LSTM. We functions,We utilize max-margin based MIL objective functions for training the LSTM.,19606
text-to-region The our validated usefulness further in the association of method task. is,The usefulness of our method is further validated in the text-to-region association task.,19607
of background video More problems very detection in and object advanced challenging. subtraction are streams,More advanced problems of background subtraction and object detection in video streams are very challenging.,19608
implemented software split in Computation the package the has score SplitSup. of been,Computation of the split score has been implemented in the software package SplitSup.,19609
we these of knowledge domain to use distributions? assist the estimation Can,Can we use domain knowledge to assist the estimation of these distributions?,19610
"space Specifically, descriptor. effective the we to with semantic project semantic an images","Specifically, we project images to the semantic space with an effective semantic descriptor.",19611
methods on We results benchmark four achieve than better state-of-the-art multi-label learning datasets.,We achieve better results than state-of-the-art multi-label learning methods on four benchmark datasets.,19612
"paper, the tackle problem of of images. RGB-D indoor this we In segmentation semantic","In this paper, we tackle the problem of RGB-D semantic segmentation of indoor images.",19613
transformation networks. a deconvolutional networks convolutional propose novel and feature network We to the bridge,We propose a novel feature transformation network to bridge the convolutional networks and deconvolutional networks.,19614
"we the visual visible specific With one that features, in capture are patterns only modality.","With specific features, we capture the visual patterns that are only visible in one modality.",19615
amount present unique The data challenges. in of massive these computational studies produced,The massive amount of data produced in these studies present unique computational challenges.,19616
"In we this the algorithmic make concrete above demixing paper, for progress some problem.","In this paper, we make some concrete algorithmic progress for the above demixing problem.",19617
In relying paper CNN-based this we present architecture. an system a downsample-then-upsample on,In this paper we present a CNN-based system relying on an downsample-then-upsample architecture.,19618
"of frame-to-frame a basis. other frames video), instead a in","other frames in a video), instead of a frame-to-frame basis.",19619
enhancement for color grading color transfer re-application. applications demonstrate and video two We,We demonstrate two applications for color transfer enhancement and video color grading re-application.,19620
simple is important future color model for also This color algorithm transfer design. transfer of,This simple model of color transfer is also important for future color transfer algorithm design.,19621
are action decode Facial facial expressions. (AUs) units human to essential,Facial action units (AUs) are essential to decode human facial expressions.,19622
on have detectors and features AU of training Researchers classifiers. a variety focused with,Researchers have focused on training AU detectors with a variety of features and classifiers.,19623
"and correlation. These modeling, temporal representation, AU are spatial","These are spatial representation, temporal modeling, and AU correlation.",19624
"report CNN over we significant datasets, On standard feature-based state-of-the-art. a and multi-label improvement both","On both datasets, we report significant improvement over a standard multi-label CNN and feature-based state-of-the-art.",19625
mining its play techniques data applications. important Clustering an and related role in,Clustering techniques play an important role in data mining and its related applications.,19626
"is and algorithm low-complexity a paper, robust proposed. this clustering In","In this paper, a robust and low-complexity clustering algorithm is proposed.",19627
independent. shape It processing that data manner centroid capable is a is and of in,It is capable of processing data in a manner that is shape and centroid independent.,19628
synthetic reported. data was real test algorithm are tested and and on The results,The algorithm was tested on real and synthetic data and test results are reported.,19629
"however, undesirable. large are CNNs, in needed number The parameters of and often","The number of parameters needed in CNNs, however, are often large and undesirable.",19630
"various methods CNN developed once it prune Consequently, is have been to a trained.","Consequently, various methods have been developed to prune a CNN once it is trained.",19631
"Nevertheless, offer benefits. CNNs resulting the limited","Nevertheless, the resulting CNNs offer limited benefits.",19632
to a and present speed economy realize CNNs. while size improvement simultaneously We pruning method,We present a method to realize simultaneously size economy and speed improvement while pruning CNNs.,19633
easily precisely could and be later. The gaps ssDNA repaired,The ssDNA gaps could be easily and precisely repaired later.,19634
in released some role progression. of cycle may ssDNA the fragments play regulation cell The,The released ssDNA fragments may play some role in the regulation of cell cycle progression.,19635
widely adopted to scheme a ground labels. sourcing collect become Crowd has truth,Crowd sourcing has become a widely adopted scheme to collect ground truth labels.,19636
"is very it can labels problem that be noisy. a well-known However, these","However, it is a well-known problem that these labels can be very noisy.",19637
their of boundary. as Shapes closed are seen functions real,Shapes are seen as real functions of their closed boundary.,19638
methods a Experiments results. localized the to validate theoretical with comparison and noise,Experiments with noise and a comparison to localized methods validate the theoretical results.,19639
convolutional visual images. great have still Deep achieved in recognition for success networks,Deep convolutional networks have achieved great success for visual recognition in still images.,19640
"recognition traditional action not in advantage the videos, methods evident. However, so is over for","However, for action recognition in videos, the advantage over traditional methods is not so evident.",19641
"novel Our for network video-based (TSN), segment first contribution recognition. framework a is action temporal","Our first contribution is temporal segment network (TSN), a novel framework for video-based action recognition.",19642
idea temporal is based the long-range of modeling. which on structure,which is based on the idea of long-range temporal structure modeling.,19643
modular. expected are be to Brain networks,Brain networks are expected to be modular.,19644
environments is Real dynamic navigation in time crucial highly problem. outdoor an,Real time outdoor navigation in highly dynamic environments is an crucial problem.,19645
to The time scale on literature outdoor real don't environments. up static SLAM dynamic recent,The recent literature on real time static SLAM don't scale up to dynamic outdoor environments.,19646
exploit poses. the of the the moving for accuracy improving We objects relationship between,We exploit the relationship between moving objects for improving the accuracy of the poses.,19647
incorporating a the as graph constraints. We all poses formulate the factor,We formulate the poses as a factor graph incorporating all the constraints.,19648
incremental problem solution a time. full exact in We optimization by achieve nonlinear solving real,We achieve exact incremental solution by solving a full nonlinear optimization problem in real time.,19649
"heuristic this method In research, a character present for we recognition.","In this research, we present a heuristic method for character recognition.",19650
are of contains original important of sketch that image. This most the image representatives pixels,This sketch contains the most important pixels of image that are representatives of original image.,19651
depth end-to-end of for field computational We proof-of-concept imaging. (EDOF) present extended system a,We present a proof-of-concept end-to-end system for computational extended depth of field (EDOF) imaging.,19652
three stages. framework is composed of Our,Our framework is composed of three stages.,19653
relatively analysis X-ray of the field immature review We for imagery. automated image cargo,We review the relatively immature field of automated image analysis for X-ray cargo imagery.,19654
into image preprocessing field the split We understanding. of and the image classical pipeline,We split the field into the classical pipeline of image preprocessing and image understanding.,19655
Image (TIP); includes: material segmentation. and Threat and improvement; Projection image quality discrimination Preprocessing manipulation;,Preprocessing includes: image manipulation; quality improvement; Threat Image Projection (TIP); and material discrimination and segmentation.,19656
Automated Image Verification and Detection understanding includes: Threat (ATD); Automated (ACV). Contents,Image understanding includes: Automated Threat Detection (ATD); and Automated Contents Verification (ACV).,19657
multiple much unifying saliency saliency on Saliency attracted has attention maps models. integration from,Saliency integration has attracted much attention on unifying saliency maps from multiple saliency models.,19658
"the Mixture Models. mathematically Bernoulli of Kernels Then, application to formalize Fisher we","Then, we mathematically formalize the application of Fisher Kernels to Bernoulli Mixture Models.",19659
algorithm We present physically-based of smoke. the to novel a animation control,We present a novel algorithm to control the physically-based animation of smoke.,19660
"more over practice, order than In observe improvement an prior methods. magnitude of we","In practice, we observe more than an order of magnitude improvement over prior methods.",19661
"LDS. sparsely In two to are encode an solutions particular, developed","In particular, two solutions are developed to sparsely encode an LDS.",19662
"For concept, introducing by novel namely time-series the two-fold learning, dictionary LDSs. a encode we","For dictionary learning, we encode time-series by introducing a novel concept, namely the two-fold LDSs.",19663
planning. heuristics for investigate learning domain-specific We,We investigate learning heuristics for domain-specific planning.,19664
framed a ordinary an heuristic as learning regression problem. work Prior,Prior work framed learning a heuristic as an ordinary regression problem.,19665
algorithm introduces novel for paper a real-time landmark facial This tracking.,This paper introduces a novel real-time algorithm for facial landmark tracking.,19666
"detection, to has both tracking challenges and Compared additional opportunities.","Compared to detection, tracking has both additional challenges and opportunities.",19667
"evaluate performance. learning importance we and incremental in the of iCCR show state-of-the-art Finally, achieving","Finally, we evaluate iCCR and show the importance of incremental learning in achieving state-of-the-art performance.",19668
"very Here, the of simple control strategies. we performance investigate","Here, we investigate the performance of very simple control strategies.",19669
with demonstrate a of pests. pest-pathogen-crop our We on rate a tunable immigration model method,We demonstrate our method on a pest-pathogen-crop model with a tunable immigration rate of pests.,19670
"high-quality A method user-centric is interactive, fast, robust and presented. shadow removal for","A user-centric method for fast, interactive, robust and high-quality shadow removal is presented.",19671
probability by measures functions induced with constraints. set study We,We study probability measures induced by set functions with constraints.,19672
of on based reduction the filters are this speckle Most for property.,Most of the filters for speckle reduction are based on this property.,19673
and providing humans. is High-resolution rich of brain animals imaging activity functional measurements in increasingly,High-resolution functional imaging is providing increasingly rich measurements of brain activity in animals and humans.,19674
array averaging recordings). sparse sampling in fMRI in local voxels or,local averaging in fMRI voxels or sparse sampling in array recordings).,19675
The resulting (BCM-MM) produces measurements. simulated model generative,The resulting generative model (BCM-MM) produces simulated measurements.,19676
less is beyond NP search Tackling using problems straightforward. ASP,Tackling search problems beyond NP using ASP is less straightforward.,19677
techniques easily The such of application that to are non-experts. encodings leads to inaccessible,The application of such techniques easily leads to encodings that are inaccessible to non-experts.,19678
"the QBFs other On modeling hand, the ASP perspective. over has from advantages","On the other hand, ASP has advantages over QBFs from the modeling perspective.",19679
of levels on Recursive PH. construction programming high enable incarnations of arbitrarily logic the this,Recursive incarnations of this construction enable logic programming on arbitrarily high levels of the PH.,19680
We paradigm. develop our implementation proof-of-concept for a new,We develop a proof-of-concept implementation for our new paradigm.,19681
under paper is acceptance This for TPLP. consideration in,This paper is under consideration for acceptance in TPLP.,19682
"contexts used to knowledge. be background extension, the structure In can further this","In this extension, contexts can be used to further structure the background knowledge.",19683
demonstrate the scalability algorithms both applying gain in learning tasks. We by various to,We demonstrate the gain in scalability by applying both algorithms to various learning tasks.,19684
This consideration paper for in acceptance is TPLP. under,This paper is under consideration for acceptance in TPLP.,19685
Flegontov approach al. et by Our recently adopted been has,Our approach has been recently adopted by Flegontov et al.,19686
at focusing which information aims performed. The be regions seeds selecting from will propagation,The focusing aims at selecting seeds regions from which information propagation will be performed.,19687
the promising An on mammograms of approach application . proposed results shows,An application of the proposed approach on mammograms shows promising results .,19688
"geographical toolbox an visualizing open-source data. We geoplotlib, python introduce for","We introduce geoplotlib, an open-source python toolbox for visualizing geographical data.",19689
"functionalities geoplotlib and cases. use design, We describe","We describe geoplotlib design, functionalities and use cases.",19690
Neural networks classification system. Convolutional tremendous importance nowadays any of for image are,Convolutional Neural networks nowadays are of tremendous importance for any image classification system.,19691
forks research accuracy. Some layers after the found increases that intermediate auxiliary adding,Some research found that adding auxiliary forks after intermediate layers increases the accuracy.,19692
which shoud just have Specifying recently. the fork layer addressed intermediate,Specifying which intermediate layer shoud have the fork just addressed recently.,19693
supervision deep as This with networks convolutional known technique neural (CNDS).,This technique known as convolutional neural networks with deep supervision (CNDS).,19694
"we study the to connections CNDS of network. paper, effect In adding residual this","In this paper, we study the effect of adding residual connections to CNDS network.",19695
of only. CNDS results accuracy Our increasing show over experiments using,Our experiments results show increasing of accuracy over using CNDS only.,19696
which in terms of implemented achieves recognition performance have a promising system very We accuracy.,We have implemented a system which achieves very promising performance in terms of recognition accuracy.,19697
OpenCL. system parts selected Therefore of to we have using the decided accelerate,Therefore we have decided to accelerate selected parts of the system using OpenCL.,19698
the paper for explanation following proposes results. This these,This paper proposes the following explanation for these results.,19699
then humans. mice applied model This and is to,This model is then applied to mice and humans.,19700
"are interests in there sequences. image inferring facial Recently, from increasing mirco-expression","Recently, there are increasing interests in inferring mirco-expression from facial image sequences.",19701
Recent usually pattern used binary spatiotemporal for micro-expression works analysis. local,Recent works usually used spatiotemporal local binary pattern for micro-expression analysis.,19702
binary local spatiotemporal seriously on the of limit micro-expression The problems application above-mentioned pattern recognition.,The above-mentioned problems seriously limit the application of spatiotemporal local binary pattern on micro-expression recognition.,19703
"micro-expressions. Firstly, preserving develop improved of an integral shape projection for we attribute the","Firstly, we develop an improved integral projection for preserving the shape attribute of micro-expressions.",19704
"spatiotemporal extract incorporating Specifically, into the attributes we novel features texture shape spatiotemporal features.","Specifically, we extract the novel spatiotemporal features incorporating shape attributes into spatiotemporal texture features.",19705
Intensive conducted experiments on availably are databases. published three micro-expression,Intensive experiments are conducted on three availably published micro-expression databases.,19706
compare state-of-the-art method our We the with algorithms.,We compare our method with the state-of-the-art algorithms.,19707
for promising micro-expression that achieves demonstrate results Experimental our performance recognition. proposed method,Experimental results demonstrate that our proposed method achieves promising performance for micro-expression recognition.,19708
procedure. of We prove the proposed formal properties,We prove formal properties of the proposed procedure.,19709
Recent by large in has progress been models vision trained computer high-capacity on driven datasets.,Recent progress in computer vision has been driven by high-capacity models trained on large datasets.,19710
"outliers, enhance robustness is reconstruction error. to estimator imposed the the on To Cauchy","To enhance robustness to outliers, the Cauchy estimator is imposed on the reconstruction error.",19711
for extensive present big data practices practitioners. with along empirical We best information,We present extensive empirical information along with best practices for big data practitioners.,19712
"a To start dataset. first we create (CIFE) candid-images-for-facial-expression with,","To start with, we first create a candid-images-for-facial-expression (CIFE) dataset.",19713
"paper, the EM the some basic In of FIM. ideas review and this we","In this paper, we review some basic ideas of the EM and the FIM.",19714
"theoretical of and analysis efficiency show the method. our Finally, examples to we numerical conduct","Finally, we conduct theoretical analysis and numerical examples to show the efficiency of our method.",19715
"matrix, . algorithm, perturbation EM words: stochastic approximation Simultaneous Fisher Carlo, Key Monte information","Key words: Fisher information matrix, EM algorithm, Monte Carlo, Simultaneous perturbation stochastic approximation .",19716
deeply significant detection was networks. made by face trained convolutional performance Recently in improvement possible,Recently significant performance improvement in face detection was made possible by deeply trained convolutional networks.,19717
"training is novel approach this face described. for detector a In report, state-of-the-art","In this report, a novel approach for training state-of-the-art face detector is described.",19718
widely surveillance Recently is increasing. of video systems use the,Recently the use of video surveillance systems is widely increasing.,19719
it Therefore proper these becomes of guarantee significant working systems. the to,Therefore it becomes significant to guarantee the proper working of these systems.,19720
vision. Camera processing techniques computer Detection based uses Tampering several image on and,Camera Tampering Detection uses several techniques based on image processing and computer vision.,19721
under has been checking entailment polynomial that It rational has instance proven complexity. a,It has been proven that instance checking under rational entailment has a polynomial complexity.,19722
"model this semantics. strengthen entailment, paper in we minimal rational a consider To","To strengthen rational entailment, in this paper we consider a minimal model semantics.",19723
Theory of Practice in is Programming. acceptance for paper consideration The under and Logic,The paper is under consideration for acceptance in Theory and Practice of Logic Programming.,19724
"relevant is reference not Unfortunately, images providing straightforward.","Unfortunately, providing relevant reference images is not straightforward.",19725
"of sources to two there difficulty. are previous formulations, Compared","Compared to previous formulations, there are two sources of difficulty.",19726
action video study of problem human the analysis classification We for recognition. facial and,We study the problem of video classification for facial analysis and human action recognition.,19727
"for and running offset phase ""highjump""). onset jumping and ""smile"", for","onset and offset phase for ""smile"", running and jumping for ""highjump"").",19728
algorithms for learning combines to learners obtain (biased) Boosting classification prediction. weak and effective,Boosting combines weak (biased) learners to obtain effective learning algorithms for classification and prediction.,19729
We synthetic of examples with applications numerical these some data. real on illustrate and,We illustrate some of these applications with numerical examples on synthetic and real data.,19730
"early image In (CBIR) content-based the retrieval with global was studied days, features.","In the early days, content-based image retrieval (CBIR) was studied with global features.",19731
"Two broad categories, SIFT-based and CNN-based are methods, presented.","Two broad categories, SIFT-based and CNN-based methods, are presented.",19732
video natural retrieval objective queries. based on language Our is,Our objective is video retrieval based on natural language queries.,19733
"Finally, can be the proposed show to model we how applied generation. description","Finally, we show how the proposed model can be applied to description generation.",19734
"one such carefully must algorithms. tune However, design, choose and","However, one must carefully design, choose and tune such algorithms.",19735
"directly is the our learn to process Instead, itself. goal iterative","Instead, our goal is to directly learn the iterative process itself.",19736
"we our a manually such task. from for labeled train Consequently, enhancement network data","Consequently, we train such a network from manually labeled data for our enhancement task.",19737
highlight et al. We work done by Gusti,We highlight work done by Gusti et al.,19738
on centralized conflict We resolution plans. focus independently computed for,We focus on centralized conflict resolution for independently computed plans.,19739
"methods used many image surprisingly, fractional-order in based been Not processing have field.","Not surprisingly, many fractional-order based methods have been used in image processing field.",19740
"and jointly features, localization, ranker. Unlike attribute's our learns network previous the methods,","Unlike previous methods, our network jointly learns the attribute's features, localization, and ranker.",19741
framework and faster significantly also speeds methods. than much end-to-end processing is up previous Our,Our end-to-end framework also significantly speeds up processing and is much faster than previous methods.,19742
in are big models used networks Bayesian analytics. often data graphical probabilistic,Bayesian networks are probabilistic graphical models often used in big data analytics.,19743
"In approach introduce this a for exact new learning. we structure paper,","In this paper, we introduce a new approach for exact structure learning.",19744
regulate MicroRNAs expression (miRNAs) small RNAs that post-transcriptionally. are non-coding gene,MicroRNAs (miRNAs) are small non-coding RNAs that regulate gene expression post-transcriptionally.,19745
"the camera pose. to algorithm we In work, estimate a this propose algebraic novel","In this work, we propose a novel algebraic algorithm to estimate the camera pose.",19746
mismatched approach is included. correspondences algebraic line to handle An also,An algebraic approach to handle mismatched line correspondences is also included.,19747
"compression introduces compression artifacts, Lossy ringing blurring. and blocking artifacts, particularly effects complex","Lossy compression introduces complex compression artifacts, particularly blocking artifacts, ringing effects and blurring.",19748
Freehand sequential sketching an process. is inherently,Freehand sketching is an inherently sequential process.,19749
"quantization. as particular state-of-the-art previous residual cases, formulation product or such This as encompasses, methods","This formulation encompasses, as particular cases, previous state-of-the-art methods such as product or residual quantization.",19750
recognizing learning task e-learning online Automatically important improving the an process. the for is activities,Automatically recognizing the e-learning activities is an important task for improving the online learning process.,19751
"are such and observation processes. For the crucial for models, inference sequences of training","For such models, the sequences of observation are crucial for training and inference processes.",19752
content important segmentation analysis based detection and tasks. for many prerequisite image an Text is,Text detection and segmentation is an important prerequisite for many content based image analysis tasks.,19753
segmentation for also extraction paper and proposes in The various text inputs. challenges multilingual,The paper also proposes various challenges in text extraction and segmentation for multilingual inputs.,19754
novel performance two proposed of with method ideas. The improves the CNN,The proposed method improves the performance of CNN with two novel ideas.,19755
consistent temporally object extracts detector. off-the-shelf The approach on an based tubes,The approach extracts temporally consistent object tubes based on an off-the-shelf detector.,19756
"we final with the combine this cues. For segmentation, motion video information","For the final video segmentation, we combine this information with motion cues.",19757
their to stimulus ecologically intensities. receptor dynamic neurons range match Sensory relevant,Sensory receptor neurons match their dynamic range to ecologically relevant stimulus intensities.,19758
poorly is How tuning is this achieved understood in receptors. most,How this tuning is achieved is poorly understood in most receptors.,19759
years. attracted the in has analysis Visual attentions many fashion recent,Visual fashion analysis has attracted many attentions in the recent years.,19760
regions either by clothing Previous or work boxes bounding represented joints. human,Previous work represented clothing regions by either bounding boxes or human joints.,19761
of gradually landmark the stages improve accuracies These predictions.,These stages gradually improve the accuracies of landmark predictions.,19762
mapping RoR substitutes residual optimizing residual for mapping. mapping residual original optimizing of,RoR substitutes optimizing residual mapping of residual mapping for optimizing original residual mapping.,19763
questions multiple style fill-in-the-blank focuses answering choice on Madlibs from paper This the dataset. Visual,This paper focuses on answering fill-in-the-blank style multiple choice questions from the Visual Madlibs dataset.,19764
"catalytically only nucleotide are $\beta$ subunits subunit similar Each pockets, in yet binds active.","Each subunit binds nucleotide in similar pockets, yet only $\beta$ subunits are catalytically active.",19765
We axes differences. their symmetry internal interesting and observe re-examine,We re-examine their internal symmetry axes and observe interesting differences.,19766
model using this three approaches. We $Methods$: validate,$Methods$: We validate this model using three approaches.,19767
noise. approach GVZM neurological The managing successful paradigm represents and for this new a understanding,The GVZM approach represents a successful new paradigm for understanding and managing this neurological noise.,19768
"signed particular, discriminative learns patch we groups. that train (DeepPattern) CNN In newly de- a","In particular, we train a newly de- signed CNN (DeepPattern) that learns discriminative patch groups.",19769
two aspects innovative to There this. are,There are two innovative aspects to this.,19770
in pay nal we hand one to information fashion. attention origi- On the contextual an,On the one hand we pay attention to contextual information in an origi- nal fashion.,19771
Suenderhauf  Recent al. et  work by,  Recent work by Suenderhauf et al.,19772
then landmarks relative consistency between the views. positions of enforces Matching of descriptors,Matching descriptors then enforces consistency of the relative positions of landmarks between views.,19773
This impact on significant has a performance.,This has a significant impact on performance.,19774
extraction. recognition vital to accurate A feature expression precursor is,A vital precursor to accurate expression recognition is feature extraction.,19775
"In we sensing SenTion: this paper, A for expressions. facial propose framework","In this paper, we propose SenTion: A framework for sensing facial expressions.",19776
a mesh such typical for The is a of process vertices. input,The typical input for such a process is a mesh of vertices.,19777
average curves the is to circle locally ability demonstrated. by The approximate,The ability to locally approximate curves by the circle average is demonstrated.,19778
"An nonlinear, exact $\mathrm{PPD}$s. integro-partial non-closed, is equations of system differential the derived for","An exact non-closed, nonlinear, system of integro-partial differential equations is derived for the $\mathrm{PPD}$s.",19779
"customary, leads procedure As closing is field a a limit. mean to","As is customary, a closing procedure leads to a mean field limit.",19780
bone are such in nature. Porous trabecular widely seen as structures,Porous structures such as trabecular bone are widely seen in nature.,19781
lightweight. exhibit superior properties being mechanical structures whilst These,These structures exhibit superior mechanical properties whilst being lightweight.,19782
method voxel-wise extends optimization. upon and Our builds topology,Our method builds upon and extends voxel-wise topology optimization.,19783
includes Our and solution method error. optimization nonlinear re-projection close-form minimizing a calibration a by,Our calibration method includes a close-form solution and a nonlinear optimization by minimizing re-projection error.,19784
"when estimating methods for learning were In explore paper, deep made. objects this we","In this paper, we explore deep learning methods for estimating when objects were made.",19785
or large-scale data Direct organization applications retrieval. include,Direct applications include large-scale data organization or retrieval.,19786
to normalization (STDN) depth estimated depth enforce temporal introduce our consistency in spatio-temporal We sequences.,We introduce spatio-temporal depth normalization (STDN) to enforce temporal consistency in our estimated depth sequences.,19787
recognition components two action the improve These significantly performance.,These two components significantly improve the action recognition performance.,19788
"not using recent evaluated outcome However, consistent have systematically measures. techniques been","However, recent techniques have not been systematically evaluated using consistent outcome measures.",19789
among methods. trade-offs delineate results Our two clear,Our results delineate two clear trade-offs among methods.,19790
"structural a that function network results maturation executive delineate process Together, youth. in supports of","Together, results delineate a process of structural network maturation that supports executive function in youth.",19791
"In handwritten generation such this data domain. we a exploit for in work, framework","In this work, we exploit such a framework for data generation in handwritten domain.",19792
render data source and fonts using incorporate schemes. augmentation We open data synthetic,We render synthetic data using open source fonts and incorporate data augmentation schemes.,19793
"intuitively. it extract what images from unclear still However, the filters is","However, it is still unclear what the filters extract from images intuitively.",19794
reveal We extracts texture. that a filter every specific,We reveal that every filter extracts a specific texture.,19795
layer and intricate from more The structures. contains higher texture colours more,The texture from higher layer contains more colours and more intricate structures.,19796
be of demonstrate style these texture could primitives. We combination a that also images of,We also demonstrate that style of images could be a combination of these texture primitives.,19797
purposefully. maps of to Two and randomly feature methods energy are reallocate distribution proposed,Two methods are proposed to reallocate energy distribution of feature maps randomly and purposefully.,19798
"code Then, we generate the styles. modified inverse images of diverse and","Then, we inverse the modified code and generate images of diverse styles.",19799
predicts model that long-term could plasticity. short- and both pathways affect both profoundly Our,Our model predicts that both pathways could profoundly affect both short- and long-term plasticity.,19800
"implications models. on in simple, deterministic from derived classical relies ecology However, paradigm this","However, this classical paradigm in ecology relies on implications derived from simple, deterministic models.",19801
for heterogeneous learning. across hashing work assumes hash function modalities Existing relationship on cross-modal,Existing work on cross-modal hashing assumes heterogeneous relationship across modalities for hash function learning.,19802
surveillance of detection. motion work the in video This field is including,This work is in the field of video surveillance including motion detection.,19803
"have rectangle). (circle, detected geometric that The objects are and ellipse, some square, objects shape","The detected objects are objects that have some geometric shape (circle, ellipse, square, and rectangle).",19804
"probability distributions experiment. paper the Luria-Delbr\""uck This modeling calculates","This paper calculates probability distributions modeling the Luria-Delbr\""uck experiment.",19805
a Haldane's function distribution to for due Ycart. includes generating probability This,This includes a generating function for Haldane's probability distribution due to Ycart.,19806
reduction storage expedited represents visual and data search. Similarity-based technique crucial hashing image image for,Similarity-based image hashing represents crucial technique for visual data storage reduction and expedited image search.,19807
hash codes not The discrete gradient-based optimization. them for nature makes amenable of,The discrete nature of hash codes makes them not amenable for gradient-based optimization.,19808
conducted Comprehensive are on several evaluations quantitative image benchmarks. widely-used,Comprehensive quantitative evaluations are conducted on several widely-used image benchmarks.,19809
"algorithm all state-of-the-art all outperforms hashing proposed significant margins. deep by our On benchmarks, competitors","On all benchmarks, our proposed deep hashing algorithm outperforms all state-of-the-art competitors by significant margins.",19810
require distances) themselves algorithms that Wasserstein compute. numerical to,Wasserstein distances) that themselves require numerical algorithms to compute.,19811
"this WLMs. stochastic based paper, introduce simulated we a annealing In approach solving on for","In this paper, we introduce a stochastic approach based on simulated annealing for solving WLMs.",19812
stability has for advantages of warm and numerical new starts. the readiness Our approach,Our new approach has the advantages of numerical stability and readiness for warm starts.,19813
methods are biological complex nonlinear to and discover medical used in data. Machine learning relationships,Machine learning methods are used to discover complex nonlinear relationships in biological and medical data.,19814
"with for features. models computationally of are learning unfeasible data However, sophisticated millions","However, sophisticated learning models are computationally unfeasible for data with millions of features.",19815
cloud computing implemented can platforms. be commodity Our algorithm on,Our algorithm can be implemented on commodity cloud computing platforms.,19816
compare segmentation mechanisms semantic three The selected algorithms. three using selection algorithm different experiments,The experiments compare three different algorithm selection mechanisms using three selected semantic segmentation algorithms.,19817
selection the of the algorithm algorithm accuracy. theoretical limits analyze with to respect We selection,We analyze the theoretical limits of algorithm selection with respect to the algorithm selection accuracy.,19818
"accuracy. disk their without number parameters, ambiguity, size overall reduces degrading It of on and","It reduces ambiguity, number of parameters, and their size on disk without degrading overall accuracy.",19819
minimization} paper (SMF). matrix factorization \emph{volume considers (VolMin)-based This structured,This paper considers \emph{volume minimization} (VolMin)-based structured matrix factorization (SMF).,19820
practical on both and VolMin. This theoretical focuses of aspects paper,This paper focuses on both theoretical and practical aspects of VolMin.,19821
of Occlusion depth problems in one the is most estimation. challenging,Occlusion is one of the most challenging problems in depth estimation.,19822
map. Then depth regularize anti-occlusion energy an to is function built,Then an anti-occlusion energy function is built to regularize depth map.,19823
and composed system nodes is of vessels. The primo vascular,The primo vascular system is composed of nodes and vessels.,19824
recent Multimedia event receiving years. increasing has detection been attention in,Multimedia event detection has been receiving increasing attention in recent years.,19825
function methods region region-based a scoring Most to merging. use determine segmentation,Most region-based segmentation methods use a scoring function to determine region merging.,19826
this that approach We a semi-supervised propose demand. reduces,We propose a semi-supervised approach that reduces this demand.,19827
object Discriminative (DCF) Filters tracking. performance visual demonstrated have for Correlation excellent,Discriminative Correlation Filters (DCF) have demonstrated excellent performance for visual object tracking.,19828
"formulation the However, is restricted significantly its to potential. feature single-resolution limiting maps, DCF underlying","However, the underlying DCF formulation is restricted to single-resolution feature maps, significantly limiting its potential.",19829
We our in point demonstrate learning also the formulation tracking effectiveness extensive experiments. of feature,We also demonstrate the effectiveness of our learning formulation in extensive feature point tracking experiments.,19830
"By at on for once, inter-object occlusions. jointly operating we account scene the full","By operating jointly on the full scene at once, we account for inter-object occlusions.",19831
KDE is the produced an to effective confidence by also means outliers. detect The,The confidence produced by the KDE is also an effective means to detect outliers.,19832
accuracy is The animation related for realistic critical and applications.,The accuracy is critical for realistic animation and related applications.,19833
"sufficient using semi-automatic is method. Also, proposed (SDR) new dimension of algorithm ABC a reduction","Also, a new semi-automatic algorithm of ABC is proposed using sufficient dimension reduction (SDR) method.",19834
the literature. primarily surfaced has frequentist SDR in,SDR has primarily surfaced in the frequentist literature.,19835
demonstrated that this has have But ABC well. with we paper connections in as it,But we have demonstrated in this paper that it has connections with ABC as well.,19836
and learned. fusion the pixel-level And CNN jointly CNN are,And the pixel-level CNN and fusion CNN are jointly learned.,19837
"and image in Our on based model encoder--decoder popular video captioning systems. pipeline, the is","Our model is based on the encoder--decoder pipeline, popular in image and video captioning systems.",19838
we separate in train diverse sub-domains. two input Using models features these specializing,Using these diverse features we train models specializing in two separate input sub-domains.,19839
"based table. ranked second the metrics automatic evaluation were we Additionally, in","Additionally, we were ranked second in the automatic evaluation metrics based table.",19840
presents of set methods a lossy This based on image networks. full-resolution paper compression neural,This paper presents a set of full-resolution lossy image compression methods based on neural networks.,19841
"reconstruction also additive a introduce and framework. new architectures We ""one-shot"" study versus scaled-additive","We also study ""one-shot"" versus additive reconstruction architectures and introduce a new scaled-additive framework.",19842
distributed spatially regions. usually network brain involves The epileptic,The epileptic network usually involves spatially distributed brain regions.,19843
"The showed higher cMEM wMNE solution performance and than inverse dSPM, sLORETA.","The wMNE inverse solution showed higher performance than dSPM, cMEM and sLORETA.",19844
model captioning. a attention propose then split novel for image We,We then propose a novel split attention model for image captioning.,19845
competing experiments distinguish often theories. run to Scientists,Scientists often run experiments to distinguish competing theories.,19846
our experiments validation optimal. strong that designed find empirical were We indeed automatically,We find strong empirical validation that our automatically designed experiments were indeed optimal.,19847
of conclude We research. for a future discussing number questions by interesting,We conclude by discussing a number of interesting questions for future research.,19848
"hundred However, (e.g. of that volumes fluid thousand consist particles several larger","However, larger fluid volumes that consist of several hundred thousand particles (e.g.",19849
oceans) a the power. many of and simulation iterations lot of require for simulation,for the simulation of oceans) require many iterations and a lot of simulation power.,19850
This end-to-end cutout. automatic an catergory-specific amounts pipeline to object for,This amounts to an automatic end-to-end pipeline for catergory-specific object cutout.,19851
a for face video-based alignment. encoder-decoder network We propose real-time model recurrent novel,We propose a novel recurrent encoder-decoder network model for real-time video-based face alignment.,19852
whether processes. model all machines A can fundamental Turing is reasoning question,A fundamental question is whether Turing machines can model all reasoning processes.,19853
"is experiments, hypothesis about introduced. which A creative based systems, on computer is","A hypothesis about creative systems, which is based on computer experiments, is introduced.",19854
"We deep CNN namely, have popular the VGG-Face, evaluated models, three AlexNet, GoogLeNet. and","We have evaluated three popular deep CNN models, namely, the AlexNet, VGG-Face, and GoogLeNet.",19855
"offer on comprehensive topics these in overview we two this Wi-Fi. paper, an In","In this paper, we offer an comprehensive overview on these two topics in Wi-Fi.",19856
"From the we and feasibility proposed the conducted, experiments demonstrate methods. of efficacy the","From the experiments conducted, we demonstrate the feasibility and efficacy of the proposed methods.",19857
"norm. methods single-step function to For be non-linear appear however, the approximation,","For non-linear function approximation, however, single-step methods appear to be the norm.",19858
"to very are of value particular, divergence susceptible they estimates. In","In particular, they are very susceptible to divergence of value estimates.",19859
"paper, this reason we the behind In identify this.","In this paper, we identify the reason behind this.",19860
"captured shading cues exploiting (IR) from an infrared camera. Kinect, by","Kinect, by exploiting shading cues captured from an infrared (IR) camera.",19861
Our approach mesh model on directly for geometry the operates refinement.,Our approach directly operates on the mesh model for geometry refinement.,19862
through challenging several The effectiveness of real-world approach is demonstrated examples. our,The effectiveness of our approach is demonstrated through several challenging real-world examples.,19863
"labeling sure that make receive labels. to is similar order regions neighboring Afterward, in smoothed","Afterward, labeling is smoothed in order to make sure that neighboring regions receive similar labels.",19864
"and they expensive to ignore and regions due among dependencies training inference. expressive non-local However,","However, they ignore expressive and non-local dependencies among regions due to expensive training and inference.",19865
We approach obtained promising datasets three on our and assessed results.,We assessed our approach on three datasets and obtained promising results.,19866
"proposed PSD further on the the we algorithm an optimize derive To function, objective manifold.","To optimize the proposed objective function, we further derive an algorithm on the PSD manifold.",19867
"Therefore, the of convolutional recurrent deeply-supervised the advantages networks nets. DSRCNN and combines neural","Therefore, the DSRCNN combines the advantages of recurrent convolutional neural networks and deeply-supervised nets.",19868
registering approach for propose point We a clouds. systematic cross-source,We propose a systematic approach for registering cross-source point clouds.,19869
to The and micro robustness of the of the structures. method is attributed extraction macro,The robustness of the method is attributed to the extraction of macro and micro structures.,19870
great method also proposed shows ability datasets. same-source in The,The proposed method also shows great ability in same-source datasets.,19871
class important Planning of decision the theory. plays role in broad an,Planning plays an important role in the broad class of decision theory.,19872
"model, e.g. generative RL, can one in assume Generally a","Generally in RL, one can assume a generative model, e.g.",19873
"types assume of behavior, on the generative environment agent models, can e.g. Based various","Based on environment behavior, the agent can assume various types of generative models, e.g.",19874
More (POMDP) Markov Decision Partially for Process complex models Observable can this drawback. compensate like,More complex models like Partially Observable Markov Decision Process (POMDP) can compensate for this drawback.,19875
infinite-dimensional handling paradigm is for important an (FDA) Functional modern data. data analysis,Functional data analysis (FDA) is an important modern paradigm for handling infinite-dimensional data.,19876
a algorithm approach mixed-effects (MLMM) linear intensive mixture models for The estimation. requires computationally of,The mixture of linear mixed-effects models (MLMM) approach requires a computationally intensive algorithm for estimation.,19877
of model novel (GMM) a characterization model-based Gaussian the clustering provide problem. We mixture,We provide a novel Gaussian mixture model (GMM) characterization of the model-based clustering problem.,19878
approach considerations discussed. for are the Theoretical GMM,Theoretical considerations for the GMM approach are discussed.,19879
using homogeneity this examine such work analysis. clustering an approach datasets we to In,In this work we examine an approach to clustering such datasets using homogeneity analysis.,19880
Homogeneity determines a of the data. representation analysis euclidean,Homogeneity analysis determines a euclidean representation of the data.,19881
represent vertex programming computer each We as shader a program. genetic to use,We use genetic programming to represent each vertex shader as a computer program.,19882
We and the discuss project's current challenges. status development,We discuss the project's current status and development challenges.,19883
"platform popular to the public photos. become has share Nowadays, for a social media","Nowadays, social media has become a popular platform for the public to share photos.",19884
proposed aesthetic-oriented into our integrate We further models. classification category,We further integrate category classification into our proposed aesthetic-oriented models.,19885
a (FACD). create Aesthetic called We Dataset dataset new Filter Comparison,We create a new dataset called Filter Aesthetic Comparison Dataset (FACD).,19886
images is dataset first user preference the labels. and containing filtered It,It is the first dataset containing filtered images and user preference labels.,19887
This largely of development limits classification algorithms. the scene,This largely limits the development of scene classification algorithms.,19888
Dataset Aerial for (AID): describes dataset Image This paper the scene a aerial classification. large-scale,This paper describes the Aerial Image Dataset (AID): a large-scale dataset for aerial scene classification.,19889
"aerial more than we ten AID, annotate creating and collect For thousands scene images.","For creating AID, we collect and annotate more than ten thousands aerial scene images.",19890
It problem using formulated optimization methods. is solved as standard continuous usually an and,It is usually formulated as an optimization problem and solved using standard continuous methods.,19891
"sparse. to be the is corruption assumed Meanwhile,","Meanwhile, the corruption is assumed to be sparse.",19892
the iteratively to derive problem. (PIRA) algorithm proximal reweighted solve a We,We derive a proximal iteratively reweighted algorithm (PIRA) to solve the problem.,19893
of approach. results results on benchmarks encouraging show Experimental our,Experimental results on benchmarks show encouraging results of our approach.,19894
work. contributions least are two at in main There our,There are at least two main contributions in our work.,19895
"programming convex Specifically, using a efficient solve we the Frank-Wolfe (QP) quadratic algorithm. relaxation","Specifically, we solve a convex quadratic programming (QP) relaxation using the efficient Frank-Wolfe algorithm.",19896
the human attributes. from problem pictures generating of on focuses paper specific This face,This paper focuses on the problem of generating human face pictures from specific attributes.,19897
"loss convolutional is With ""VGG-Face"", deep convolutional maps. the on the the defined feature network","With the deep convolutional network ""VGG-Face"", the loss is defined on the convolutional feature maps.",19898
the solve We apply to decent optimization this algorithm problem. then gradient,We then apply the gradient decent algorithm to solve this optimization problem.,19899
the our results and face method generation. of for attribute driven effectiveness The validate identity-preserving,The results validate the effectiveness of our method for attribute driven and identity-preserving face generation.,19900
evaluate benchmark and achieve We approach datasets three state-of-the-art on performance on our each.,We evaluate our approach on three benchmark datasets and achieve state-of-the-art performance on each.,19901
"best the In is previous method. faster significantly approach than our addition,","In addition, our approach is significantly faster than the previous best method.",19902
"In unconstrained of problem video paper, address in clips. the searching we this proposals action","In this paper, we address the problem of searching action proposals in unconstrained video clips.",19903
"designed searched further refine paths. to tracking-by-detection is action Finally, a the scheme","Finally, a tracking-by-detection scheme is designed to further refine the searched action paths.",19904
transmitted by directly contact. sexual only through be It (mosquito) but not also vectors can,It can be transmitted not only by (mosquito) vectors but also directly through sexual contact.,19905
"in Africa. especially endemic in Malaria tropical remains areas,","Malaria remains endemic in tropical areas, especially in Africa.",19906
well Climatic recordedthroughout factors as as were study. household characteristics the,Climatic factors as well as household characteristics were recordedthroughout the study.,19907
article is Supplementary for this available material online.,Supplementary material for this article is available online.,19908
inductive rule sets propositional the extends definitions: logic semantics. with under PC(ID) well-founded,PC(ID) extends propositional logic with inductive definitions: rule sets under the well-founded semantics.,19909
"notion Recently, relevance was of a language. for introduced this","Recently, a notion of relevance was introduced for this language.",19910
"changes During solver state have effect search, literals. an the relevance the on of","During search, the solver state changes have an effect on the relevance of literals.",19911
most tracking Robust is the accurate vision visual computer challenging one and of problems.,Robust and accurate visual tracking is one of the most challenging computer vision problems.,19912
(SRDCF) tracking. Filters for Discriminative Spatially propose We Correlation Regularized,We propose Spatially Regularized Discriminative Correlation Filters (SRDCF) for tracking.,19913
datasets. state-of-the-art approach Our on four results all achieves,Our approach achieves state-of-the-art results on all four datasets.,19914
video involving effect also good features a segmenting on temporal has We investigate how data.,We investigate how involving temporal features also has a good effect on segmenting video data.,19915
and near-infrared image the details. enhancement Image (V) using visible useful usually (NIR) enhances,Image enhancement using the visible (V) and near-infrared (NIR) usually enhances useful image details.,19916
"The enhanced evaluated images quantitative of evaluation. instead perception, by feature are observers","The enhanced images are evaluated by observers perception, instead of quantitative feature evaluation.",19917
the area. Determining cell mutation is one of such a rate,Determining the mutation rate of a cell is one such area.,19918
"for have convolutional Recently, networks optical proven (convnets) predicting useful flow.","Recently, convolutional networks (convnets) have proven useful for predicting optical flow.",19919
Together for on form losses flow. based the groundtruth these measure a proxy losses,Together these losses form a proxy measure for losses based on the groundtruth flow.,19920
schemes of scalar A wide devised for color have variety to mapping been data color.,A wide variety of color schemes have been devised for mapping scalar data to color.,19921
the color value scale. a use to index data Some,Some use the data value to index a color scale.,19922
"colors usually to assign blended Others to overlap. materials areas handle materials, disjoint where different,","Others assign colors to different, usually blended disjoint materials, to handle areas where materials overlap.",19923
precision assessed is quantitatively method using recall the and accuracy metrics. The proposed,The proposed method is assessed quantitatively using the precision and recall accuracy metrics.,19924
"Clothing as stimuli style, expressed such are color, and usually texture. with fashions visual","Clothing fashions are usually expressed with visual stimuli such as style, color, and texture.",19925
society next The sciences. for and grand challenges the science brain in are,The next grand challenges for society and science are in the brain sciences.,19926
"medicine. We on and challenges, spanning physiology, three eventually settled anatomy,","We eventually settled on three challenges, spanning anatomy, physiology, and medicine.",19927
computational three novel all challenges Addressing infrastructure. requires,Addressing all three challenges requires novel computational infrastructure.,19928
states requires and observations. between model The linear relationship a,The model requires a linear relationship between states and observations.,19929
"specified. highly assumptions, for be given we can models non-linear mild demonstrate that Further, $p(\text{state}|\text{observation})$","Further, we demonstrate that given mild assumptions, highly non-linear models for $p(\text{state}|\text{observation})$ can be specified.",19930
video a shot detection. approach boundary novel for paper This presents,This paper presents a novel approach for video shot boundary detection.,19931
The based and proposed merge is concept. approach split on,The proposed approach is based on split and merge concept.,19932
proposed also a validated method experimentally video. cricket is on The,The proposed method is experimentally also validated on a cricket video.,19933
Exhaustive the HSMM. performance simulations revealed in superior models comparison have proposed with of,Exhaustive simulations have revealed superior performance of the proposed models in comparison with HSMM.,19934
"Next, budget. we total the bits assigned layer to bit each given select the adaptively","Next, we adaptively select the bits assigned to each layer given the total bit budget.",19935
"the that, the network After quantized retrain fine-tune centroids. to we","After that, we retrain the network to fine-tune the quantized centroids.",19936
Two units theta and excitable different systems studied: neurons. Adler are,Two different excitable systems are studied: Adler units and theta neurons.,19937
"observed. are as behaviors, dimensional as Compatible solutions, higher well chaotic","Compatible behaviors, as well as higher dimensional chaotic solutions, are observed.",19938
study simulations further equations. validate numerical We the to,We study numerical simulations to further validate the equations.,19939
system. gesture The in surveillance understanding from recognition simple complicated range applications behaviour to,The applications range from simple gesture recognition to complicated behaviour understanding in surveillance system.,19940
to development motion recognition. and to techniques leads in This related human major representation the,This leads to major development in the techniques related to human motion representation and recognition.,19941
bird's A the view in for eye is domain the in new paper. presented researchers,A bird's eye view for new researchers in the domain is presented in the paper.,19942
R-package available make We in \verb+cstab+. the our method,We make our method available in the R-package \verb+cstab+.,19943
"Continuous L-DOPA, fast jejunal absorption. as administration pumps, of of have the issue addressed such","Continuous administration of L-DOPA, such as jejunal pumps, have addressed the issue of fast absorption.",19944
"we function evaluation Second, propose specifically targeting $\psi_{p,\sigma}$ criterion. new Maximin a the","Second, we propose a new evaluation function $\psi_{p,\sigma}$ specifically targeting the Maximin criterion.",19945
both in consumption expensive memory is runtime. Volume-based usually of reconstruction terms and,Volume-based reconstruction is usually expensive both in terms of memory consumption and runtime.,19946
"produce huge a overhead. geometric computational volumetric sparse structures, Especially representations for","Especially for sparse geometric structures, volumetric representations produce a huge computational overhead.",19947
algorithm. matri a DLT combined the the xestimated cornerstone is method projection by A of,A cornerstone of the method is a combined projection matri xestimated by the DLT algorithm.,19948
Multiplicity improve proposed of the the is the to of accuracy method. exploited estimates,Multiplicity of the estimates is exploited to improve the accuracy of the proposed method.,19949
modeled Samples being generally from as from are (non-linear) intimate a drawn manifold. mixtures smooth,Samples from intimate (non-linear) mixtures are generally modeled as being drawn from a smooth manifold.,19950
algorithm embedding improved and for clustering of mixture-manifolds. propose reconstruction-based We novel a,We propose a novel reconstruction-based algorithm for improved clustering and embedding of mixture-manifolds.,19951
program this call a We epistemic in logic program language an (ELP).,We call a program in this language an epistemic logic program (ELP).,19952
solver. occurrences) known any M and and K ELP faster than other,K and M occurrences) and faster than any other known ELP solver.,19953
"computational while expense, storage benefits The manageable. include data making more reduction,","The benefits include storage reduction, computational expense, while making data more manageable.",19954
a presented from recorded on algorithm trajectories hand-tracking was tested The system.,The presented algorithm was tested on trajectories recorded from a hand-tracking system.,19955
the is ranking method. simplest average One the methods of metalearning,One of the simplest metalearning methods is the average ranking method.,19956
average construct are ranking. to ranks the The used,The ranks are used to construct the average ranking.,19957
time. fewer computation thus and simply could conduct We save tests,We could simply conduct fewer tests and save thus computation time.,19958
"death's or nematodes, In of alter time. physiological environmental scaling perturbations","In nematodes, environmental or physiological perturbations alter death's scaling of time.",19959
"curvature of cancer, alter In time. death's human perturbations genetic","In human cancer, genetic perturbations alter death's curvature of time.",19960
and death's scale Those the changes follow geometry. of in contours curvature invariant constraining,Those changes in scale and curvature follow the constraining contours of death's invariant geometry.,19961
hands egocentric masks accurate propose We very for in a extracting of views. method,We propose a method for extracting very accurate masks of hands in egocentric views.,19962
The two confirmed illustrative of satisfying examples. algorithm performance the new is by,The satisfying performance of the new algorithm is confirmed by two illustrative examples.,19963
"is boundary. explicit this In framework, there no decision","In this framework, there is no explicit decision boundary.",19964
"our mid-size the for the dataset knowledge, is light-field first of To this images. best","To the best of our knowledge, this is the first mid-size dataset for light-field images.",19965
learning method a endoscopic we for this propose paper In transfer of images.,In this paper we propose a method for transfer learning of endoscopic images.,19966
The fastest Saddle proposed. the among is detector,The Saddle detector is among the fastest proposed.,19967
the in are NN. trained other with parameters These along parameters,These parameters are trained along with other parameters in the NN.,19968
Networks Neural multiple tasks. performance of in (NN) AAFs classification improved have,AAFs have improved performance of Neural Networks (NN) in multiple classification tasks.,19969
"for tasks. NNs and we this apply In AAFs feedforward propose paper, regression on","In this paper, we propose and apply AAFs on feedforward NNs for regression tasks.",19970
"However, overfitting. lead AAFs may using existing to","However, using existing AAFs may lead to overfitting.",19971
avoid regularizing parameters. with SAAFs simply overfitting the NNs by can,NNs with SAAFs can avoid overfitting by simply regularizing the parameters.,19972
"NNs overfitting. parameters regularizing avoids Thus, in SAAFs with the","Thus, regularizing the parameters in NNs with SAAFs avoids overfitting.",19973
We and on multiple achieved state-of-the-art with results SAAFs regression datasets. NNs evaluated empirically,We empirically evaluated NNs with SAAFs and achieved state-of-the-art results on multiple regression datasets.,19974
 increasingly alignment.  varying for popular Spatially are image warps,  Spatially varying warps are increasingly popular for image alignment.,19975
"a varying matches. number of warps sufficient requires spatially However, estimating feature","However, estimating spatially varying warps requires a sufficient number of feature matches.",19976
This of correlation filter promote to the robustness study the intends learning.,This study intends to promote the robustness of the correlation filter learning.,19977
"are three a implemented. trackers As real-time result,","As a result, three real-time trackers are implemented.",19978
a robust useful filter tracking. in a for visual designing correlation reference criterion is This,This is a useful reference criterion in designing a robust correlation filter for visual tracking.,19979
The based to of two methods. the is existing address framework proposed limits aiming CNN,The proposed framework is aiming to address two limits of the existing CNN based methods.,19980
"depth be RGB-D applied method detection Furthermore, to saliency generally can by refinement. our","Furthermore, our method can be generally applied to RGB-D saliency detection by depth refinement.",19981
Symmetry breaking has be for proven (SAT). been technique solving efficient preprocessing to satisfiability an,Symmetry breaking has been proven to be an efficient preprocessing technique for satisfiability solving (SAT).,19982
"the breaker compare symmetry our current ASP. sbass, tool with We for state-of-the-art","We compare our tool with sbass, the current state-of-the-art symmetry breaker for ASP.",19983
"problem. the for this framework a we estimation deep paper, illumination In learning adopt","In this paper, we adopt a deep learning framework for the illumination estimation problem.",19984
variation problems sensing. effectiveness proved in solving for compressive has inverse its Total,Total variation has proved its effectiveness in solving inverse problems for compressive sensing.,19985
examine provide scientific ways powerful to Multidimensional techniques questions. imaging kinds of various,Multidimensional imaging techniques provide powerful ways to examine various kinds of scientific questions.,19986
of the objects Tracking useful monitoring surveillance is for analysis. in the and videos motion,Tracking of motion objects in the surveillance videos is useful for the monitoring and analysis.,19987
deteriorate shadows The surveillance will performance system detected of the moving when as are objects.,The performance of the surveillance system will deteriorate when shadows are detected as moving objects.,19988
"the stages. benefits usually Therefore, shadow elimination and detection next","Therefore, shadow detection and elimination usually benefits the next stages.",19989
"and of a is shadows detection issue, this overcome for proposed. elimination method To","To overcome this issue, a method for detection and elimination of shadows is proposed.",19990
we from been plethora perspective. first person captured a of have the videos with provided,we have been provided with a plethora of videos captured from the first person perspective.,19991
"relationship yet explored two thoroughly. However, to between be the the has","However, the relationship between the two has yet to be explored thoroughly.",19992
"In this this we approaching effort, two attempt to questions. explore relationship by","In this effort, we attempt to explore this relationship by approaching two questions.",19993
"top-view we content second, And identify in can egocentric the the video? of viewers the","And second, can we identify the egocentric viewers in the content of the top-view video?",19994
"the surveillance videos? In other we find can cameramen the words, in","In other words, can we find the cameramen in the surveillance videos?",19995
videos These time-synchronous. more can challenging the problems are become when not,These problems can become more challenging when the videos are not time-synchronous.,19996
"the top-view a observers localize video. spatiotemporally egocentric As result, the we in","As a result, we spatiotemporally localize the egocentric observers in the top-view video.",19997
study we al. introduced case O'Shaughnessy considered models et three a mathematical by As,As a case study we considered three mathematical models introduced by O'Shaughnessy et al.,19998
yet structure is replication a mechanism such of not The on understood.,The mechanism of replication on such a structure is not yet understood.,19999
"DNA the With replication. involved picture this in in initiated mind, we characterization have bigger","With this bigger picture in mind, we have initiated the characterization involved in DNA replication.",20000
applications. a variety widely used of Fingerprint is in,Fingerprint is widely used in a variety of applications.,20001
Security the data. privacy of fingerprint to to have measures be taken protect,Security measures have to be taken to protect the privacy of fingerprint data.,20002
of using an and proposed protecting biometrics. effective biometrics as Cancelable mechanism is,Cancelable biometrics is proposed as an effective mechanism of using and protecting biometrics.,20003
"number user one is generator. each random with generated synthetic given Specifically, minutia template","Specifically, each user is given one synthetic minutia template generated with random number generator.",20004
set k-nearest verification arbitrary constructed by of the neighbors. The template is an combining,The verification template is constructed by combining an arbitrary set of the k-nearest neighbors.,20005
"three on prove scheme, of the To validity testing the is out carried databases.","To prove the validity of the scheme, testing is carried out on three databases.",20006
the cancelable The the requirements results of that show constructed templates biometrics. satisfy,The results show that the constructed templates satisfy the requirements of cancelable biometrics.,20007
been Explanations the century. in previous introduced have,Explanations have been introduced in the previous century.,20008
no is Their search questioned. interest in the space reducing longer,Their interest in reducing the search space is no longer questioned.,20009
"still Yet, their CSP a implementation challenge. into solver is efficient","Yet, their efficient implementation into CSP solver is still a challenge.",20010
for dynamic backtracking enables computation selection an explanations efficient intelligent This al- of gorithms.,This dynamic selection enables an efficient computation of explanations for intelligent backtracking al- gorithms.,20011
"as this a we end, joint-to-person multi-person association pose consider estimation problem. To","To this end, we consider multi-person pose estimation as a joint-to-person association problem.",20012
a computer problems. Linking vision sources basic building data two is numerous block in,Linking two data sources is a basic building block in numerous computer vision problems.,20013
"detect a assist data we events. processing, mitosis based to learning automatically To method developed","To assist data processing, we developed a learning based method to automatically detect mitosis events.",20014
proposals top a on perform exploit these of CNN detection. then object to We,We then exploit a CNN on top of these proposals to perform object detection.,20015
spatial-temporal is based the STANNCR on proposed action The component and information. incorporates,The proposed STANNCR is based on action component and incorporates the spatial-temporal information.,20016
approach extensively datasets. The three evaluated on is proposed public,The proposed approach is extensively evaluated on three public datasets.,20017
the action STANNCR demonstrate effectiveness of recognition. results for experimental The,The experimental results demonstrate the effectiveness of STANNCR for action recognition.,20018
new been ideas made some to has to image applied mathematicians effort analysis. An show,An effort has been made to show mathematicians some new ideas applied to image analysis.,20019
operation. one only with constructed are They,They are constructed with only one operation.,20020
in and operations. of tilings Allowance are types Constraint used the Two,Two tilings are used in the Constraint and Allowance types of operations.,20021
concavity New of dale. a described: type,New type of concavity described: a dale.,20022
"to operations All realize are clock-less. parallel, possible","All operations are parallel, possible to realize clock-less.",20023
They gray as are separate objects. treated,They are treated as separate gray objects.,20024
their multiple relations are them among descendants. and There,There are multiple relations among them and their descendants.,20025
"proposed. concavities that, is Via topological of size","Via that, topological size of concavities is proposed.",20026
"""strokes"" concavities. like are defined Notions through","Notions like ""strokes"" are defined through concavities.",20027
effects levelized objects are gray on shown. Unusual,Unusual effects on levelized gray objects are shown.,20028
of living organisms proposed. analysis image hypothesis is A,A hypothesis of living organisms image analysis is proposed.,20029
by robust Bayesian formulated multi-object a framework. presents paper (MCMOT) multi-class a tracking This filtering,This paper presents a robust multi-class multi-object tracking (MCMOT) formulated by a Bayesian filtering framework.,20030
tracking encouraging techniques comparison to very shows The video results. state-of-the-art,The comparison to state-of-the-art video tracking techniques shows very encouraging results.,20031
"can developing how discipline ""data science."" ontologies the of be explore vis-a-vis We computational impactful","We explore how computational ontologies can be impactful vis-a-vis the developing discipline of ""data science.""",20032
"is, would implemented as data rules management within That engine. theories analytics a be","That is, management theories would be implemented as rules within a data analytics engine.",20033
work of This advantages models. exploits the both,This work exploits the advantages of both models.,20034
employed two matrix method is the A to integrate completion models.,A matrix completion method is employed to integrate the two models.,20035
"appearance target tracking the as is a formulated of Specifically, estimation. problem","Specifically, the tracking is formulated as a problem of target appearance estimation.",20036
"their Consequently, in limited. reasoning is multimedia application","Consequently, their application in multimedia reasoning is limited.",20037
reality for information new possibilities developing offers navigation surgical and augmented applications. This,This information offers new possibilities for developing surgical navigation and augmented reality applications.,20038
suggested is process a sampler random in discrete-time The environment. stochastic,The suggested sampler is a discrete-time stochastic process in random environment.,20039
"object videos. learn we In from of models time-lapse paper, computational this transformations","In this paper, we learn computational models of object transformations from time-lapse videos.",20040
"to impaired loss processing information whether a information of transfer. local leads Therefore, we asked","Therefore, we asked whether impaired local information processing leads to a loss of information transfer.",20041
decrease entropy in The PFC. in was stronger source,The decrease in source entropy was stronger in PFC.,20042
"compromised. As multi-view the is consequence, data for a clustering performance","As a consequence, the clustering performance for multi-view data is compromised.",20043
manifest classical-like in We that universes. singular is only incompatibility the discover,We discover that the incompatibility is only manifest in singular classical-like universes.,20044
dependence graphical models Gaussian are widely among represent variables. used random to conditional,Gaussian graphical models are widely used to represent conditional dependence among random variables.,20045
mouse technique plausible obtain and to biologically results. data We genomics apply our,We apply our technique to mouse genomics data and obtain biologically plausible results.,20046
"total than divergences Unsurprisingly, bounds non-symmetric give the sharper Hellinger. or variation","Unsurprisingly, the non-symmetric divergences give sharper bounds than total variation or Hellinger.",20047
multi-gate metal oxide flexible Flexible substrates. graphene on neuron were hybrid fabricated transistors oxide/graphene,Flexible metal oxide/graphene oxide hybrid multi-gate neuron transistors were fabricated on flexible graphene substrates.,20048
was detector lobula model visual for giant motion emulating neuron investigated. system proof-of-principle A,A proof-of-principle visual system model for emulating lobula giant motion detector neuron was investigated.,20049
for interest cognitive results great neuromorphic are flexible of Our systems.,Our results are of great interest for flexible neuromorphic cognitive systems.,20050
objects of occluded is a Detection computer vision partially challenging problem.,Detection of partially occluded objects is a challenging computer vision problem.,20051
will also as State be a analysed metric. forecasts and observation performance,State and observation forecasts will also be analysed as a performance metric.,20052
Attention a present tracking novel We visual technique modeling. based Shared on attention,We present a novel visual attention tracking technique based on Shared Attention modeling.,20053
present in We color measurements OLED from an noises model for empirical displays.,We present an empirical model for noises in color measurements from OLED displays.,20054
XYZ is model presented empirical values. The depends measured the simple only on and,The presented empirical model is simple and depends only on the measured XYZ values.,20055
machine analysis pipelines. learning a data in is and coding block core many building Sparse,Sparse coding is a core building block in many data analysis and machine learning pipelines.,20056
this study for In we the paper such reasons acceleration.,In this paper we study the reasons for such acceleration.,20057
are problems several of Many composed interacting components. real-world,Many real-world problems are composed of several interacting components.,20058
"this With four ways. article, contribute in we","With this article, we contribute in four ways.",20059
"contribute portfolio. to algorithms this Finally, which study most we","Finally, we study which algorithms contribute most to this portfolio.",20060
recovery. of dynamics the importance interplay is starvation and reproduction Of between and the particular,Of particular importance is the interplay between reproduction and the dynamics of starvation and recovery.,20061
of a driver This selective mechanism framework rule. provides a for Cope's principled,This framework provides a principled mechanism for a selective driver of Cope's rule.,20062
space-time Our inputs. calculates functions grid-based of a the approach signed-distance using deformation dense,Our approach calculates a dense space-time deformation using grid-based signed-distance functions of the inputs.,20063
We solve. flow five-dimensional employ optical a,We employ a five-dimensional optical flow solve.,20064
"in is the computed, arbitrary match created can between very be efficiently. variants Once","Once the match is computed, arbitrary in between variants can be created very efficiently.",20065
"we alignment concatenate deformations, long-range a propose technique. novel multiple To","To concatenate multiple long-range deformations, we propose a novel alignment technique.",20066
"swirling smoke clouds, simulations. As and a interpolate splashing liquid can result, we","As a result, we can interpolate swirling smoke clouds, and splashing liquid simulations.",20067
unsupervised learning similarities discovering powerful in paradigm a an for Exemplar is visual manner.,Exemplar learning is a powerful paradigm for discovering visual similarities in an unsupervised manner.,20068
are cliques. compact samples into batches are and Conflicting different similar distributed relations over grouped,Conflicting relations are distributed over different batches and similar samples are grouped into compact cliques.,20069
framed categorization tasks. Learning exemplar of clique sequence a as is similarities,Learning exemplar similarities is framed as a sequence of clique categorization tasks.,20070
analysis and The shown performance competitive proposed on posture has unsupervised classification. approach object detailed,The proposed unsupervised approach has shown competitive performance on detailed posture analysis and object classification.,20071
"which for matching, probing, i.e. a pattern in is metric, double-sided useful Aspl\""und consists 's","Aspl\""und 's metric, which is useful for pattern matching, consists in a double-sided probing, i.e.",20072
function a the sub-graph and probed jointly. the over-graph are of,the over-graph and the sub-graph of a function are probed jointly.,20073
"we log-likelihood penalize this function. the problem, To deal with","To deal with this problem, we penalize the log-likelihood function.",20074
log-likelihood inference function. based on We also hypothesis testing penalized consider the,We also consider hypothesis testing inference based on the penalized log-likelihood function.,20075
"likelihood signed and score consider Wald tests. ratio, In ratio, likelihood we particular,","In particular, we consider likelihood ratio, signed likelihood ratio, score and Wald tests.",20076
BS to laws. between use distinguish nonnested hypothesis test We bimodal a two,We use a nonnested hypothesis test to distinguish between two bimodal BS laws.,20077
We some corrections analytical to tests. derive,We derive analytical corrections to some tests.,20078
discussed. applications presented and and are results empirical simulation Carlo Monte,Monte Carlo simulation results and empirical applications are presented and discussed.,20079
the interactions developed phenomena. theorists natural complex have characterize to Network methods in,Network theorists have developed methods to characterize the complex interactions in natural phenomena.,20080
is PDDL enables planning mixed dynamics. extension of discrete-continuous an that PDDL+ with domains modelling,PDDL+ is an extension of PDDL that enables modelling planning domains with mixed discrete-continuous dynamics.,20081
models provide encoding into We PDDL+ an of problems. CASP,We provide an encoding of PDDL+ models into CASP problems.,20082
histogram extracted threshold first for region based of The a on regions. predefined is interest,The region of interest is first extracted based on a predefined threshold for histogram regions.,20083
as is This tradeoffs grows. of number reduced tolerance the,This tolerance is reduced as the number of tradeoffs grows.,20084
"supervision. For unannotated tracks discover videos, can without content be used any to","For unannotated videos, tracks can be used to discover content without any supervision.",20085
annotations and consuming expensive to are obtain. time Pixel-level,Pixel-level annotations are expensive and time consuming to obtain.,20086
"Hence, segmentation. could image weak tags a in impact significant semantic have only using supervision","Hence, weak supervision using only image tags could have a significant impact in semantic segmentation.",20087
"Recently, using CNN-based fine-tune proposed image pre-trained to have methods tags. networks","Recently, CNN-based methods have proposed to fine-tune pre-trained networks using image tags.",20088
"accuracy. localization poor information, additional to Without this leads","Without additional information, this leads to poor localization accuracy.",20089
"use This alleviated was by priors foreground/background however, to making masks. problem, generate objectness of","This problem, however, was alleviated by making use of objectness priors to generate foreground/background masks.",20090
"pixel-level these annotations/bounding boxes, require Unfortunately object either yield priors training inaccurate boundaries. or still","Unfortunately these priors either require training pixel-level annotations/bounding boxes, or still yield inaccurate object boundaries.",20091
"we of introduce inexpensive weak accuracy an form boost. Furthermore, supervision a additional yielding new","Furthermore, we introduce a new form of inexpensive weak supervision yielding an additional accuracy boost.",20092
systems. dynamics globally cortical distinct to by perturbations are local Such neural integrated,Such perturbations to local cortical dynamics are globally integrated by distinct neural systems.,20093
"brain computational Here, we activity. localized briefly studies addressing affect perturbations review and how empirical","Here, we briefly review empirical and computational studies addressing how localized perturbations affect brain activity.",20094
"To aspects. main contributions we three end, from the following this make","To this end, we make three main contributions from the following aspects.",20095
as can storage Chesson's understood This of a phenomenon consequence be effect. direct,This phenomenon can be understood as a direct consequence of Chesson's storage effect.,20096
"In reduced. the influence incomplete negative of label problem way, this be can the","In this way, the negative influence of the label incomplete problem can be reduced.",20097
residual learning aesthetics of hierarchical A automatic representation. enable deep very to utilized network is,A very deep residual network is utilized to enable automatic learning of hierarchical aesthetics representation.,20098
"is an learning deep framework two these ideas, Integrating established. end-to-end","Integrating these two ideas, an end-to-end deep learning framework is established.",20099
structured the presents SPCA-TV) This paper (denoted optimization sparse its and framework resolution. PCA,This paper presents the structured sparse PCA (denoted SPCA-TV) optimization framework and its resolution.,20100
sets. of on and SPCA-TV three different the versatility We demonstrate efficiency data,We demonstrate the efficiency and versatility of SPCA-TV on three different data sets.,20101
segmentation achieved (DL) have perfor- Deep learning approaches state-of-the-art mance.,Deep learning (DL) approaches have achieved state-of-the-art segmentation perfor- mance.,20102
problems. be a optimization wide hence of to used PESMOC range can solve,PESMOC can hence be used to solve a wide range of optimization problems.,20103
of problems. expensive These black-box a plethora in constraints optimization arise,These constraints arise in a plethora of expensive black-box optimization problems.,20104
addresses the point multiple paper of sets. This problem registering,This paper addresses the problem of registering multiple point sets.,20105
"sets an contrast, footing. formulation proposed the all point the on In equal treats","In contrast, the proposed formulation treats all the point sets on an equal footing.",20106
stereo videos? Are we ready to segment consumer,Are we ready to segment consumer stereo videos?,20107
"data unexplored. the is segmentation still such largely of However,","However, the segmentation of such data is still largely unexplored.",20108
segmentation combines and scheme. image existing efficient This an in techniques video,This combines existing image and video segmentation techniques in an efficient scheme.,20109
"to should desirable able interface However, a natural user's understand mental be user status automatically.","However, a desirable natural user interface should be able to understand user's mental status automatically.",20110
"for presented. model DAIA, engagement paper, this is novel multi-modal detection, a In","In this paper, a novel multi-modal model for engagement detection, DAIA, is presented.",20111
"State a Finite this is Transducer For designed. purpose, (FST)","For this purpose, a Finite State Transducer (FST) is designed.",20112
employed state smoothly transition several is to combination the of FST boolean expressions. using make,FST is employed to make the state transition smoothly using combination of several boolean expressions.,20113
Results segment show robustly. FST can hand user gestures also more,Results also show FST can segment user hand gestures more robustly.,20114
sampling (ABC) in functions. intractable popular applications likelihood is method Bayesian involving Computation a Approximate,Approximate Bayesian Computation (ABC) is a popular sampling method in applications involving intractable likelihood functions.,20115
Monte Experiments are both done ABC rejection sequential and methods. Carlo simple with ABC,Experiments are done with both simple rejection ABC and sequential Monte Carlo ABC methods.,20116
to detection The many scenes crowded has deal challenges. of behaviours abnormal in with,The detection of abnormal behaviours in crowded scenes has to deal with many challenges.,20117
and of anomalies efficient in videos. This method detection localization an presents for paper,This paper presents an efficient method for detection and localization of anomalies in videos.,20118
"architecture addresses and cascaded tasks, This two main detection. outlier FCN-based feature representation","This FCN-based architecture addresses two main tasks, feature representation and cascaded outlier detection.",20119
"for and experimental datasets We validation, in qualitative these tasks. both four present public quantitative,","We present experimental validation, both qualitative and quantitative, in four public datasets for these tasks.",20120
"OpenCL to a addition paper, this writing open-source developed we functioning raycaster. In","In addition to writing this paper, we developed a functioning open-source OpenCL raycaster.",20121
"""volume and the paper This of with ray introduction topic provides OpenCL"". an overview casting","This paper provides an introduction and overview of the topic ""volume ray casting with OpenCL"".",20122
"we .NET OpenCL raycaster describe using how built Then, we a and C#. rudimentary","Then, we describe how we built a rudimentary raycaster using OpenCL and .NET C#.",20123
"we sequences case are ordinal temporal bags instances. consider Moreover, where of the","Moreover, we consider the case where bags are temporal sequences of ordinal instances.",20124
"we (MI-DORF). Multi-Instance model Dynamic Ordinal propose this, Fields Random the To novel","To model this, we propose the novel Multi-Instance Dynamic Ordinal Random Fields (MI-DORF).",20125
"bag treat inside states. instance-labels ordinal as this the model, we latent In","In this model, we treat instance-labels inside the bag as latent ordinal states.",20126
computer-aided present in optical colonoscopy We a detection images. for polyps algorithm,We present a computer-aided detection algorithm for polyps in optical colonoscopy images.,20127
cancer. are to precursors the colon Polyps,Polyps are the precursors to colon cancer.,20128
attractor Local structure. modular by minima specific characterized states patterns represent model this of in,Local minima in this model represent attractor states characterized by specific patterns of modular structure.,20129
"ROIs to single subcortical attention, tend community. and form a Visual, sensorimotor, functional","Visual, attention, sensorimotor, and subcortical ROIs tend to form a single functional community.",20130
"a stimuli. complex interactions, is and involving signaling Bone biochemical mechanical cell-cell process remodeling","Bone remodeling is a complex process involving cell-cell interactions, biochemical signaling and mechanical stimuli.",20131
"emergence clusters. scenarios, revealed three-dimensional of sponge-like bone the coexistence simulations In","In coexistence scenarios, three-dimensional simulations revealed the emergence of sponge-like bone clusters.",20132
"that do this However, well is not known it cell cultures law. obey","However, it is well known that cell cultures do not obey this law.",20133
design the systems implications important with relevance. downscaled results in-vitro physiological have The for of,The results have important implications for the design of downscaled in-vitro systems with physiological relevance.,20134
for The of of proposed global logo characteristics classification. system classification use images makes,The proposed classification system makes use of global characteristics of logo images for classification.,20135
"logo global Color, a shape texture, of wholly characteristics images. describe the of and logo","Color, texture, and shape of a logo wholly describe the global characteristics of logo images.",20136
The these of combinations classification. are used characteristics various for,The various combinations of these characteristics are used for classification.,20137
(K-NN) Neighbour is for classification. The classifier K-Nearest used,The K-Nearest Neighbour (K-NN) classifier is used for classification.,20138
are results promising fusion for most experimental of results the features. obtained The that show,The experimental results show that the most promising results are obtained for fusion of features.,20139
image Deep is current for the learning classification. bet,Deep learning is the current bet for image classification.,20140
of context. imaging amounts for its usage annotated Its in limits huge data medical greed,Its greed for huge amounts of annotated data limits its usage in medical imaging context.,20141
appears solution. scenario In transfer a prominent as learning this,In this scenario transfer learning appears as a prominent solution.,20142
learning results. transfer this may influence classification to schemes clarify we report how In aim,In this report we aim to clarify how transfer learning schemes may influence classification results.,20143
"findings future Although drive some our open, researches. may remain issues","Although some issues remain open, our findings may drive future researches.",20144
Gene Motivation: biological of processes. and are functions various importance regulatory fundamental to interactions,Motivation: Gene regulatory interactions are of fundamental importance to various biological functions and processes.,20145
present structures. algorithm a network scalable We genome-wide Result: regulatory identifying for novel,Result: We present a novel scalable algorithm for identifying genome-wide regulatory network structures.,20146
article in The in Availability is MATLAB$^\circledR$. and Implementation: algorithm this described implemented,Availability and Implementation: The algorithm described in this article is implemented in MATLAB$^\circledR$.,20147
information: online. Contact: available hongyu.miao@uth.tmc.edu Supplementary jliu@cs.rochester.edu;  Supplementary data are ,Contact: jliu@cs.rochester.edu; hongyu.miao@uth.tmc.edu   Supplementary information: Supplementary data are available online.,20148
realistic not worlds is virtual creating easy. But,But creating realistic virtual worlds is not easy.,20149
(CP) uses solver. base currently Constraint as MiniSAT(ID) IDP Programming its knowledge The backend system,The IDP knowledge base system currently uses MiniSAT(ID) as its backend Constraint Programming (CP) solver.,20150
as similar a solver few systems backend. (MIP) Programming used Integer have A Mixed,A few similar systems have used a Mixed Integer Programming (MIP) solver as backend.,20151
"far so when However, solver little known is preferable. MIP the about is","However, so far little is known about when the MIP solver is preferable.",20152
"model rigged a fully then The embedded watertight skeleton, mesh, of and skinning weights. consists","The fully rigged model then consists of a watertight mesh, embedded skeleton, and skinning weights.",20153
deep for novel paper prediction. architecture This a presents saliency,This paper presents a novel deep architecture for saliency prediction.,20154
an representation. important multi-instance Dictionary in plays data role,Dictionary plays an important role in multi-instance data representation.,20155
to histograms. bags instances maps It of,It maps bags of instances to histograms.,20156
mapped a are histograms These to dictionary. using multi-instance bags,These bags are mapped to histograms using a multi-instance dictionary.,20157
"on design condition, loss. Base a hinge this we","Base on this condition, we design a hinge loss.",20158
differential solving is ordinary equation the then segmentation distance. realised by an geodesic with The,The segmentation is then realised by solving an ordinary differential equation with the geodesic distance.,20159
compared The with layer retinal GDM segmented the results manually of are boundaries/surfaces.,The results of the GDM are compared with manually segmented retinal layer boundaries/surfaces.,20160
popular in Action topic still computer in research has been vision. classification images a,Action classification in still images has been a popular research topic in computer vision.,20161
"and categories pre-defined fixed. the Besides, in action such are datasets are vocabularies","Besides, the action categories in such datasets are pre-defined and vocabularies are fixed.",20162
for built. learning the task descriptions of A dataset is from new actions,A new dataset for the task of learning actions from descriptions is built.,20163
several that results baseline significantly. methods our Experimental outperforms show method,Experimental results show that our method outperforms several baseline methods significantly.,20164
We Lagrange onto map problem this problem minimizing functional multipliers. dual the of of a,We map this problem onto the dual problem of minimizing a functional of Lagrange multipliers.,20165
identify phonemes Visual to from aims recognition the continuous speech speech. sequence of,Visual speech recognition aims to identify the sequence of phonemes from continuous speech.,20166
method matching We environments. in novel unconstrained a for template propose,We propose a novel method for template matching in unconstrained environments.,20167
i.e. computational Common for movement automated eye - methods detection,Common computational methods for automated eye movement detection - i.e.,20168
learning decision Reinforcement as typically Markov specified processes. are tasks,Reinforcement learning tasks are typically specified as Markov decision processes.,20169
"examples, the of a formalism. series demonstrate generality Through and this we of utility","Through a series of examples, we demonstrate the generality and utility of this formalism.",20170
This proposed versatility the framework. demonstrate again of and effectiveness,This again demonstrate the effectiveness and versatility of proposed framework.,20171
meeting that network this promise in offers neuroscience challenge. posit We,We posit that network neuroscience offers promise in meeting this challenge.,20172
We then analyses. further mechanisms molecular the incorporate into,We then incorporate further molecular mechanisms into the analyses.,20173
"work, classify based images. this Triangle we propose In flower to a approach","In this work, we propose a Triangle based approach to classify flower images.",20174
"merging Initially, segmented whorl segmentation. using flowers are based region","Initially, flowers are segmented using whorl based region merging segmentation.",20175
the method. a flower flower using pruning of from obtained is Skeleton a segmented skeleton,Skeleton of a flower is obtained from the segmented flower using a skeleton pruning method.,20176
the Delaunay junction is from points the obtained endpoints and detected triangulation The on skeleton.,The Delaunay triangulation is obtained from the endpoints and junction points detected on the skeleton.,20177
suitable explored of classification. classifier been has A for purpose the,A suitable classifier has been explored for the purpose of classification.,20178
"edge-preserving for approach new this image deconvolution. In efficient propose a work, we","In this work, we propose a new approach for efficient edge-preserving image deconvolution.",20179
of - on based algorithm guided Our is novel image filter. filter a type explicit,Our algorithm is based on a novel type of explicit image filter - guided filter.,20180
for We algorithm propose motion dense of smoke. estimation an,We propose an algorithm for dense motion estimation of smoke.,20181
upgraded flow. to This sparse a is dense flow,This sparse flow is upgraded to a dense flow.,20182
photometric Our and requires geometric approach methods calibration. geometric hybrid of and a no is,Our approach is a hybrid of geometric and photometric methods and requires no geometric calibration.,20183
a poses. capturing having of photometric in single overcome We by multiple viewpoint problem the,We overcome the problem of having a single photometric viewpoint by capturing in multiple poses.,20184
renderings provide any resulting view. in photorealistic relightable models The,The resulting relightable models provide photorealistic renderings in any view.,20185
translucent key of perception scattering is to our materials. Sub-surface,Sub-surface scattering is key to our perception of translucent materials.,20186
based a model angular tabulating of Photon on is Our Diffusion. approach compressed Beam,Our approach is based on tabulating a compressed angular model of Photon Beam Diffusion.,20187
"model. generate, to explain and how We our sample evaluate","We explain how to generate, evaluate and sample our model.",20188
of modeling artificial machine intelligence. learning of foundations the Probabilistic modern and is one,Probabilistic modeling is one of the foundations of modern machine learning and artificial intelligence.,20189
modeling It context-specific capable of independence. therefore is,It is therefore capable of modeling context-specific independence.,20190
dependency first-order a grammar. We parameterize non-projective LDFM a using,We parameterize a LDFM using a first-order non-projective dependency grammar.,20191
Our results models. competitive that LDFMs probabilistic with show existing are experimental,Our experimental results show that LDFMs are competitive with existing probabilistic models.,20192
"update, target around the multiple we different positions At and examples (at use each scales).","At each update, we use multiple examples around the target (at different positions and scales).",20193
CC inter-CC application. the existing models of We for degradation clear show,We show the clear degradation of existing CC models for inter-CC application.,20194
rain We removing for introduce network an from DerainNet deep architecture streaks called a image.,We introduce a deep network architecture called DerainNet for removing rain streaks from an image.,20195
"improve we the the to augment Moreover, CNN with image visual enhancement results. framework","Moreover, we augment the CNN framework with image enhancement to improve the visual results.",20196
aggressive of types cancer. is Melanoma amongst most,Melanoma is amongst most aggressive types of cancer.,20197
"detected it in is its stages. highly if early curable However,","However, it is highly curable if detected in its early stages.",20198
for of lesions Prescreening importance. moles suspicious is of great malignancy and,Prescreening of suspicious moles and lesions for malignancy is of great importance.,20199
into lesion and as segmentation two normal skin. image regions of an,segmentation of an image into two regions as lesion and normal skin.,20200
mask The processing output post segmentation by some refined is operations.,The output segmentation mask is refined by some post processing operations.,20201
"the this work, in animal of classification In for images. a propose method we","In this work, we propose a method for the classification of animal in images.",20202
Probabilistic here neural network classification. are K-nearest neighbors considered and for,Probabilistic neural network and K-nearest neighbors are considered here for classification.,20203
success face Recently super-resolution. sparse in representation gained has image great,Recently sparse representation has gained great success in face image super-resolution.,20204
"In super-resolution AdaBoost. on we paper, new weighted-patch this a based propose method","In this paper, we propose a new weighted-patch super-resolution method based on AdaBoost.",20205
by recovery of reliance theories stroke are a after limited on studies. Current language small,Current theories of language recovery after stroke are limited by a reliance on small studies.,20206
mathematical networks. utilized networks Boolean gene for as model regulatory are,Boolean networks are utilized as mathematical model for gene regulatory networks.,20207
solve Boolean provides This an control paper the in algorithm to problem networks.,This paper provides an algorithm to solve the control problem in Boolean networks.,20208
continuous topic. interest the The competitors in growing demonstrates of a increase,The continuous increase of competitors demonstrates a growing interest in the topic.,20209
using paper Neural is Networks. Convolutional human estimation on pose This,This paper is on human pose estimation using Convolutional Neural Networks.,20210
"detection-followed-by-regression CNN this To a end, propose cascade. we","To this end, we propose a detection-followed-by-regression CNN cascade.",20211
over advocates standard the representation learning current flow optical model. This Eulerian Lagrangian work motion,This work advocates Eulerian motion representation learning over the current standard Lagrangian optical flow model.,20212
provide direction a each phase-based motivate For and task we approach. the possible,For each task we motivate the phase-based direction and provide a possible approach.,20213
findings. lesion-symptom corroborated these tracking analyses fiber Multivariate mapping and,Multivariate lesion-symptom mapping and fiber tracking analyses corroborated these findings.,20214
particular coefficients. spatially-varying for includes a issue We consider that model this,We consider this issue for a particular model that includes spatially-varying coefficients.,20215
face application carried out. also A related real perception is to,A real application related to face perception is also carried out.,20216
"confounding of inflates coefficient) variance fixed estimates. spatial effect the (regression Moreover,","Moreover, spatial confounding inflates the variance of fixed effect (regression coefficient) estimates.",20217
and up setting Standard speeds the computations MCMC well mix algorithms iteration. reduced-dimensional per,Standard MCMC algorithms mix well and the reduced-dimensional setting speeds up computations per iteration.,20218
gastrointestinal (VCE) capsule tract. widely endoscopy used the (GI) for Video is visualizing nowadays,Video capsule endoscopy (VCE) is used widely nowadays for visualizing the gastrointestinal (GI) tract.,20219
"automatic degrees have detection success. algorithms Recently, of with various proposed polyp been","Recently, automatic polyp detection algorithms have been proposed with various degrees of success.",20220
"use model circuits. we FORCE create biologically motivated two training Finally, to","Finally, we use FORCE training to create two biologically motivated model circuits.",20221
inspired successfully is and by singing. songbird One the zebra-finch reproduces,One is inspired by the zebra-finch and successfully reproduces songbird singing.,20222
"propose support-agnostic In sparse a this novel denoising algorithm paper, using we collaborative image reconstruction.","In this paper, we propose a novel image denoising algorithm using collaborative support-agnostic sparse reconstruction.",20223
observed first into image is patches. divided An,An observed image is first divided into patches.,20224
structured for to processing. are together patches be grouped collaborative Similarly utilized,Similarly structured patches are grouped together to be utilized for collaborative processing.,20225
"proposed In schemes, same support patches assumed to collaborative are taps. the share the similar","In the proposed collaborative schemes, similar patches are assumed to share the same support taps.",20226
"estimation, good patch provides support the image enhancing of very quality restoration. hence This","This provides very good patch support estimation, hence enhancing the quality of image restoration.",20227
Image of vision and learning. in one machine classification main research problems computer is the,Image classification is one of the main research problems in computer vision and machine learning.,20228
noise a camera). in low-quality surveillance sensor,sensor noise in a low-quality surveillance camera).,20229
relational models. on provide survey a We,We provide a survey on relational models.,20230
account global Relational {domains by taking into the complete dependencies models describe networked in data}.,Relational models describe complete networked {domains by taking into account global dependencies in the data}.,20231
can predictions if lead to accurate learning more models approaches. non-relational to compared Relational machine,Relational models can lead to more accurate predictions if compared to non-relational machine learning approaches.,20232
"wild paper, in for learning alignment to conditions. exploit person-specific incremental In we this propose","In this paper, we propose to exploit incremental learning for person-specific alignment in wild conditions.",20233
to are the image method. employed video and valid datasets proposed Both,Both image and video datasets are employed to valid the proposed method.,20234
arise innovations so How do complex quickly? evolutionary,How do complex evolutionary innovations arise so quickly?,20235
time past a long ago. events great of happened The the,The great events of the past happened a long time ago.,20236
study a provides Cancer innovation. evolutionary model to,Cancer provides a model to study evolutionary innovation.,20237
many an A must become evolve aggressive traits tumor novel to cancer.,A tumor must evolve many novel traits to become an aggressive cancer.,20238
"evolutionary hierarchy recursive In general, processes. systems of form a","In general, evolutionary systems form a hierarchy of recursive processes.",20239
"generated, spread recursive and innovations are which Those determine transmitted. at the processes rates","Those recursive processes determine the rates at which innovations are generated, spread and transmitted.",20240
recursive I evolutionary to the processes abrupt innovation. relate,I relate the recursive processes to abrupt evolutionary innovation.,20241
yet high-throughput challenging computational is Genome reads from assembly sequencing a the (HTS) fundamental problem.,Genome assembly from the high-throughput sequencing (HTS) reads is a fundamental yet challenging computational problem.,20242
challenge the by uncertainty repetitive caused widespread elements. the intrinsic An is,An intrinsic challenge is the uncertainty caused by the widespread repetitive elements.,20243
of iterations. mainly of two types consists It,It mainly consists of two types of iterations.,20244
structural layouts proposed possible split the at is test sites. to variation A statistical,A statistical test is proposed to split the layouts at possible structural variation sites.,20245
"iterations scaffolding/contig-extension, and of merging. mapping, includes contig type The second","The second type of iterations includes mapping, scaffolding/contig-extension, and contig merging.",20246
"(OLC) method, the Instead the de graph take paradigm. Bruijn we of overlap-layout-consensus","Instead of the de Bruijn graph method, we take the overlap-layout-consensus (OLC) paradigm.",20247
"selected calculations. by Throughout the the is probabilistic criterion mapping assembling,","Throughout the assembling, the mapping criterion is selected by probabilistic calculations.",20248
These assemblers. to innovations be de can used the complementary novo existing,These innovations can be used complementary to the existing de novo assemblers.,20249
and is parameterless metric a MUG training. does need not,MUG is a parameterless metric and does not need training.,20250
"NR-IQAs, and to other is cropping. independent MUG Unlike block size","Unlike other NR-IQAs, MUG is independent to block size and cropping.",20251
also more is stable introduced. index MUG+ A called,A more stable index called MUG+ is also introduced.,20252
knowledge encoded a locigal knowledge on applications base (KB). in rely Many AI,Many AI applications rely on knowledge encoded in a locigal knowledge base (KB).,20253
space help search for restrict KB the answers given to The repairs. gradually,The given answers help to gradually restrict the search space for KB repairs.,20254
interactive query rely pool-based Existing debuggers often computation. on for a strategy,Existing interactive debuggers often rely on a pool-based strategy for query computation.,20255
criteria query heuristic efficient These are search methods. to devising key the,These criteria are the key to devising efficient heuristic query search methods.,20256
"active different relations learning we study Further, measures. between","Further, we study different relations between active learning measures.",20257
neural a based semantic for is segmentation. model network Our trained on deep,Our model is based on a deep neural network trained for semantic segmentation.,20258
"initial, about us to an semantic This reason instances segmentation. from allows category-level","This allows us to reason about instances from an initial, category-level semantic segmentation.",20259
shown image applications. recognition vision computer Machine-learning have or outstanding for performance algorithms classification,Machine-learning algorithms have shown outstanding image recognition or classification performance for computer vision applications.,20260
information human action is recognition. Learning representation motion of the to crucial spatial-temporal,Learning the spatial-temporal representation of motion information is crucial to human action recognition.,20261
"region of automatic initialization is by method a For applied growing segmentation, seed points.","For segmentation, a region growing method is applied by automatic initialization of seed points.",20262
measured are segmentation The performance the with appreciable. different results and is known well measures,The segmentation performance is measured with different well known measures and the results are appreciable.,20263
"areas texture color represented the and features. are Subsequently, lesion extracted by","Subsequently, the extracted lesion areas are represented by color and texture features.",20264
It the supply. increases reduces blood oxygen skull the and and inside pressure,It increases the pressure inside the skull and reduces blood and oxygen supply.,20265
"some imposing done i.e. can be covariance constraints This matrices, on the","This can be done imposing some constraints on the covariance matrices, i.e.",20266
the structure incorporating information priori a components. on mixture of covariance the by,by incorporating a priori information on the covariance structure of the mixture components.,20267
we Puzzling something do failure designed. are implies how organisms about that understand not,Puzzling failure implies something that we do not understand about how organisms are designed.,20268
of Solving processes hidden the reveals design. the puzzle,Solving the puzzle reveals the hidden processes of design.,20269
"via path style-transfer the There generalization towards of task, handling algorithms. alternative exists an texture-synthesis","There exists an alternative path towards handling the style-transfer task, via generalization of texture-synthesis algorithms.",20270
The system them performance evaluate different of to the allowed two setups experiments (i.e.,The experiments allowed them to evaluate the performance of two different system setups (i.e.,20271
Results tests all SVM to baseline setup. were compared of the,Results of all the tests were compared to SVM baseline setup.,20272
interactive video. Our iteratively cluster able and of images classes to approach is,Our interactive approach is able to iteratively cluster classes of images and video.,20273
an autonomous also has into system for been real-time It control. incorporated vehicle driving,It has also been incorporated into an autonomous driving system for real-time vehicle control.,20274
for is The source the code available online. algorithm,The source code for the algorithm is available online.,20275
theme import Structure an motion in from is vision. computer,Structure from motion is an import theme in computer vision.,20276
"research Finally, in are non-rigid future discussed. structure topics from some motion","Finally, some future research topics in non-rigid structure from motion are discussed.",20277
use of the proportion constraints enforces The values. Dirichlet positivity and sum-to-one distribution on the,The use of the Dirichlet distribution enforces positivity and sum-to-one constraints on the proportion values.,20278
semantic propose part-based a to of models classes Google from technique object Images. We train,We propose a technique to train semantic part-based models of object classes from Google Images.,20279
manual this A key of requires part is location annotations. no that approach it benefit,A key benefit of this approach is that it requires no manual part location annotations.,20280
"is a completely is for undesirable of reasons. it unsupervised which Firstly, number","Firstly, it is completely unsupervised which is undesirable for a number of reasons.",20281
"a most work Secondly, relational is to limited single graph.","Secondly, most work is limited to a single relational graph.",20282
"ii) alternativeness. i) to iii) and diversity enforce ticular explore In par- supervision we sparsity,","In par- ticular we explore supervision to enforce i) sparsity, ii) diversity and iii) alternativeness.",20283
to We how for this lift then work show multi-relational graphs.,We then show how to lift this work for multi-relational graphs.,20284
"thickness the print In the slice time. increases reducing largely addition,","In addition, reducing the slice thickness largely increases the print time.",20285
geometry the sloped of reproduces increasing better print time. the without surfaces It,It better reproduces the geometry of sloped surfaces without increasing the print time.,20286
"remains reliable. height remain and small the Therefore, changes plastic deposition","Therefore, the height changes remain small and plastic deposition remains reliable.",20287
between The units. critically antisymmetric to is network due balanced short-ranged connections,The network is critically balanced due to short-ranged antisymmetric connections between units.,20288
"from face introduce a a we image. single approach Here, reconstructing learning-based a three-dimensional for","Here, we introduce a learning-based approach for reconstructing a three-dimensional face from a single image.",20289
characteristic accurate localization face rely on methods points. of key recovery Recent,Recent face recovery methods rely on accurate localization of key characteristic points.,20290
on FRI performance and The of of rigidity formulations index depends index. flexibility its,The performance of FRI depends on its formulations of rigidity index and flexibility index.,20291
"formulations. work, and alternative flexibility this introduce we rigidity In","In this work, we introduce alternative rigidity and flexibility formulations.",20292
"gFRI based visualization surface demonstrated. Finally, molecular generation are flexibility and","Finally, gFRI based molecular surface generation and flexibility visualization are demonstrated.",20293
Matching image a Deep high-quality popular for method (DM) quasi-dense matching. is,Deep Matching (DM) is a popular high-quality method for quasi-dense image matching.,20294
useful detection. representation object object are skeletons and Object for,Object skeletons are useful for object representation and object detection.,20295
"outperforms other two extraction achieves Our promising results method competitors. on datasets, significantly and skeleton","Our method achieves promising results on two skeleton extraction datasets, and significantly outperforms other competitors.",20296
"graph and diffusion one-class cut, maps classification.","graph cut, diffusion maps and one-class classification.",20297
"complexity to favorable/comparable state-of-the-art Furthermore, many of computational the proposed techniques. is the method","Furthermore, the computational complexity of the proposed method is favorable/comparable to many state-of-the-art techniques.",20298
"spatial a transformations. translation useful just However, of is of one myriad","However, translation is just one of a myriad of useful spatial transformations.",20299
considering spatial other efficiency be invariances? when the same Can attained,Can the same efficiency be attained when considering other spatial invariances?,20300
"Such generalized but a been the high convolutions in have computational considered at cost. past,","Such generalized convolutions have been considered in the past, but at a high computational cost.",20301
all output model keeping as realistic edits automatically the as possible. The adjusts,The model automatically adjusts the output keeping all edits as realistic as possible.,20302
algorithm color. task the realistic and We shape of evaluate on our manipulation of photo,We evaluate our algorithm on the task of realistic photo manipulation of shape and color.,20303
integrates HNSolve HA a dReal of by network. tightly abstraction with discrete solving the,HNSolve tightly integrates with dReal by solving a discrete abstraction of the HA network.,20304
occurring patterns frequently (or the spectral represent the vectors components). Anchor,Anchor vectors represent the frequently occurring patterns (or the spectral components).,20305
model. explained necessity rectification the is of RECOS The using,The necessity of rectification is explained using the RECOS model.,20306
"(CNN), Keywords: (ReLU), Rectified Unit Network Dataset. Convolutional Neural Activation, Model, Linear Nonlinear MNIST RECOS","Keywords: Convolutional Neural Network (CNN), Nonlinear Activation, RECOS Model, Rectified Linear Unit (ReLU), MNIST Dataset.",20307
face highly face to recognition. accurate Robust is imperative representation,Robust face representation is imperative to highly accurate face recognition.,20308
based is license. VIPLFaceNet under open-source SDK on released BSD An C++,An open-source C++ SDK based on VIPLFaceNet is released under BSD license.,20309
and point both start for VIPLFaceNet a recognition industrial academic applications. face state-of-the-art provides,VIPLFaceNet provides a state-of-the-art start point for both academic and industrial face recognition applications.,20310
to involve that intracellular components. of bind/unbind diffusion Many pathways messengers the signaling cell,Many cell signaling pathways involve the diffusion of messengers that bind/unbind to intracellular components.,20311
of a range approach to also Our applicable articulated is objects. wide,Our approach is also applicable to a wide range of articulated objects.,20312
for design GPU it makes processing. suited LATCH The of well,The design of LATCH makes it well suited for GPU processing.,20313
"speeds. together, leads Taken descriptor matching breakneck extraction this and to","Taken together, this leads to breakneck descriptor extraction and matching speeds.",20314
"are In gender two and estimation face-based tasks. learning classification human typical age biometrics,","In human face-based biometrics, gender classification and age estimation are two typical learning tasks.",20315
for stimulation powerful disorder a Neural treatment. neural technique represents,Neural stimulation represents a powerful technique for neural disorder treatment.,20316
"step for to this a structured methods prediction. inference Overall, developing Bayesian automated fundamental is","Overall, this is a fundamental step to developing automated inference methods for Bayesian structured prediction.",20317
sufficient condition disease-free asymptotic global stability approved. equilibrium is of of The,The sufficient condition of global asymptotic stability of disease-free equilibrium is approved.,20318
simulations results. are to analytical Numerical presented illustrate the,Numerical simulations are presented to illustrate the analytical results.,20319
all goal items identify The of is means by testing. to group,The goal is to identify all items by means of group testing.,20320
generalized problem. group testing the This is,This is the generalized group testing problem.,20321
"the paper, Sterrett this $S$). In (procedure we investigate procedure","In this paper, we investigate the Sterrett procedure (procedure $S$).",20322
that concept classifier experiments for rank-based this leading systems. a is indicate method Additional,Additional experiments indicate that this method is a leading concept for rank-based classifier systems.,20323
combined local distributions These are produce surface a normal single map. consistent then to,These local distributions are then combined to produce a single consistent surface normal map.,20324
based fully cascade convolutional networks a three-stage (FCNs). propose on neural We structure,We propose a three-stage cascade structure based on fully convolutional neural networks (FCNs).,20325
map score generated FCN each stage. is A after,A score map is generated after each FCN stage.,20326
to fed face next regions of selected the stage. Probable and are,Probable regions of face are selected and fed to the next stage.,20327
objects supervision to in using only. image-level images We localize aim,We aim to localize objects in images using image-level supervision only.,20328
neurone-to-neurone consider the complex the networks. of dynamical analysis mathematical implications We of,We consider the implications of the mathematical analysis of neurone-to-neurone dynamical complex networks.,20329
of build In the we consciousness. turn problem on address these ideas to hard,In turn we build on these ideas to address the hard problem of consciousness.,20330
"First, the on strong classification conditions performance. influence viewing a have can","First, the viewing conditions can have a strong influence on classification performance.",20331
studies Empirical concepts. of feasibility proposed and the the demonstrate effectiveness,Empirical studies demonstrate the effectiveness and feasibility of the proposed concepts.,20332
the is HMC particular hard In to well. mass tune of matrix,In particular the mass matrix of HMC is hard to tune well.,20333
We modifications explaining model propose to that these come data. closer to the,We propose modifications to the model that come closer to explaining these data.,20334
implications results. We these discuss the of,We discuss the implications of these results.,20335
"model, pairwise Our on wavelet products complex-valued convolutions, simplifies the of edge-based methods. previous built","Our model, built on the products of complex-valued wavelet convolutions, simplifies previous edge-based pairwise methods.",20336
model was algorithm. Caputo new Differential The sense by using Transformation Fractional solved,The new model was solved using Fractional Differential Transformation by Caputo sense algorithm.,20337
are of Human a sequence poses. actions comprised of,Human actions are comprised of a sequence of poses.,20338
makes and of videos dense human source humans This a of rich poses.,This makes videos of humans a rich and dense source of human poses.,20339
data supervision. is train to allowing automatically for task without human The generated us this,The data for this task is automatically generated allowing us to train without human supervision.,20340
scene computed correspondence formulation. the a fields refine in We flow variational finally,We finally refine the computed correspondence fields in a variational scene flow formulation.,20341
succeed. guaranteed to If unknown well-separated these algorithm is are this subspaces,If these unknown subspaces are well-separated this algorithm is guaranteed to succeed.,20342
the upon The well algorithm assumption on same that spread. the rests points subspace are,The algorithm rests upon the assumption that points on the same subspace are well spread.,20343
violated happens this investigated. yet question what if not The been has condition is,The question what happens if this condition is violated has not yet been investigated.,20344
"this particular In distributions analyzed. same the will work, be on of effect the subspace","In this work, the effect of particular distributions on the same subspace will be analyzed.",20345
"Guided-Radon-of-Gabor Gabor-of-Radon-Image two techniques: We Barcodes and propose different (GRIBCs), Barcodes (GRGBCs).","We propose two different techniques: Gabor-of-Radon-Image Barcodes (GRIBCs), and Guided-Radon-of-Gabor Barcodes (GRGBCs).",20346
has promising emerged as digital binary a to tag content-based codes technology. images retrieval Using,Using content-based binary codes to tag digital images has emerged as a promising retrieval technology.,20347
"as for a have image been (RBCs) descriptor search. binary Radon introduced new barcodes Recently,","Recently, Radon barcodes (RBCs) have been introduced as a new binary descriptor for image search.",20348
simple thresholding suggested for has A binarization. been local,A simple local thresholding has been suggested for binarization.,20349
"""autoencoded barcodes"". put idea In the Radon paper, this forward of we","In this paper, we put forward the idea of ""autoencoded Radon barcodes"".",20350
mini-batch descent employed stochastic gradient the approach for the We training.,We employed the mini-batch stochastic gradient descent approach for the training.,20351
"usually multiple of real-world composite images However, contain complex, mixtures distortions.","However, real-world images usually contain complex, composite mixtures of multiple distortions.",20352
kinematic object motion is variables. parameterized function defined The on the appropriately,The kinematic function is defined on the appropriately parameterized object motion variables.,20353
with in This estimators yields variance. a reduction large approach biased potentially,This approach yields biased estimators with a potentially large reduction in variance.,20354
default as We ASPIC+. using Brewka's argumentation prioritised logic express (PDL),We express Brewka's prioritised default logic (PDL) as argumentation using ASPIC+.,20355
"performance However, always computational improved. are be to of the methods expected","However, the performance of computational methods are always expected to be improved.",20356
"recognition. audiovisual personality Here, we for trait deep apparent multimodal an network residual develop","Here, we develop an audiovisual deep residual network for multimodal apparent personality trait recognition.",20357
Operationalization the applications has Wusterian terminology revived IT for of approach.,Operationalization of terminology for IT applications has revived the Wusterian approach.,20358
conceptual The more taking seat specialised lexicography. back prevails after to once dimension,The conceptual dimension once more prevails after taking back seat to specialised lexicography.,20359
demonstrated This of by terminology. emergence ontology is in the,This is demonstrated by the emergence of ontology in terminology.,20360
and framework agents interact Our actions to take with enables objects.,Our framework enables agents to take actions and interact with objects.,20361
"of number we can a efficiently. collect training samples Hence, huge","Hence, we can collect a huge number of training samples efficiently.",20362
expensive prohibitively extensive cloudscapes to optimisations. time in are Volumetric without real render,Volumetric cloudscapes are prohibitively expensive to render in real time without extensive optimisations.,20363
the implementation. remove from we noise resulting temporal a image apply anti-aliasing To,To remove noise from the resulting image we apply a temporal anti-aliasing implementation.,20364
learning is learning. planning scaling Temporal abstraction and to in key up reinforcement,Temporal abstraction is key to scaling up learning and planning in reinforcement learning.,20365
to combined required intelligently tasks. for classification are be which,which are required to be intelligently combined for classification tasks.,20366
three We experiments datasets perform exhaustive four modalities. different spanning on,We perform exhaustive experiments on three different datasets spanning four modalities.,20367
"PK Sheep?"" ""Do Dream Electric Androids asked of Dick once","PK Dick once asked ""Do Androids Dream of Electric Sheep?""",20368
Can live do? as and they humans change,Can they live and change as humans do?,20369
"AI, personalities. creates personality a Extreme engine, NPC psychology-based adaptive","Extreme AI, a psychology-based personality engine, creates adaptive NPC personalities.",20370
guard acts spoofing attacks. Liveliness a as detection against safe,Liveliness detection acts as a safe guard against spoofing attacks.,20371
classification decreases drastically all approaches conditions. The results in light these variable of,The classification results of all these approaches decreases drastically in variable light conditions.,20372
processes. variables internal that the global are multiple microscopic These are outcome of,These are global variables that are the outcome of multiple internal microscopic processes.,20373
roles growth components the of and We explore division. different controlling possible in cell,We explore possible roles of the different components in controlling cell growth and division.,20374
recently. have and Transport-based attention for analysis received techniques increased data signal,Transport-based techniques for signal and data analysis have received increased attention recently.,20375
in framework demonstrate scenarios. results challenging the our of effectiveness Experimental,Experimental results demonstrate the effectiveness of our framework in challenging scenarios.,20376
"System conceptual discussion Human this a is paper, (VL-HVS) In Visual model discussed. Light-Based Visible","In this discussion paper, a Visible Light-Based Human Visual System (VL-HVS) conceptual model is discussed.",20377
"data, amount has  video-monitored surgeries, recorded increased. of surgical The extremely during ","  The amount of surgical data, recorded during video-monitored surgeries, has extremely increased.",20378
"function t, as time for consciousness, constructed. C, of based measurement A formula a is","A measurement based formula for consciousness, C, as a function of time t, is constructed.",20379
"dependent P(t), them From a probability a time identity. function, by extracted is mathematical using","From them a time dependent probability function, P(t), is extracted by using a mathematical identity.",20380
as classification we the task. formulate this a In paper segmentation challenge,In this paper we formulate the segmentation challenge as a classification task.,20381
efficacy algorithms with of Comparison model. our other major high substantiates the,Comparison with other major algorithms substantiates the high efficacy of our model.,20382
independently. estimator each presented robust The processes structure in this paper,The robust estimator presented in this paper processes each structure independently.,20383
elemental only subsets for specify to sampling. has user The of the number random,The user has to specify only the number of elemental subsets for random sampling.,20384
"algorithm also described detail. has Like estimators, are limitations which this in any robust","Like any robust estimators, this algorithm also has limitations which are described in detail.",20385
are presented Several aspect real algorithm. the and examples of illustrate every to synthetic,Several synthetic and real examples are presented to illustrate every aspect of the algorithm.,20386
"people. Interpersonal defines between two relation friendliness, association, and e.g., more dominance, or warm, the","Interpersonal relation defines the association, e.g., warm, friendliness, and dominance, between two or more people.",20387
on benchmarks. expression facial demonstrate existing state-of-the-art we With results network recognition the,With the network we demonstrate state-of-the-art results on existing facial expression recognition benchmarks.,20388
"a However, suffers training from lengthy it process.","However, it suffers from a lengthy training process.",20389
nature due the space of vast the features and to Adaboost. exhaustive This is search,This is due to the vast features space and the exhaustive search nature of Adaboost.,20390
performance demonstrated recent methods in Tracking-by-detection competitive have years.,Tracking-by-detection methods have demonstrated competitive performance in recent years.,20391
"In the model quality set. approaches, relies the training on of tracking these heavily the","In these approaches, the tracking model heavily relies on the quality of the training set.",20392
samples. dynamically manages approach training by Our the estimating of the set quality the,Our approach dynamically manages the training set by estimating the quality of the samples.,20393
"Finally, on method all three our state-of-the-art datasets. results achieves","Finally, our method achieves state-of-the-art results on all three datasets.",20394
dominated are approaches by in years. recent methods patch-based State-of-the-art clearly non-local,State-of-the-art approaches are clearly dominated by patch-based non-local methods in recent years.,20395
variational several for this task. We variants feasible exploit,We exploit several feasible variational variants for this task.,20396
"low problem SNR. in encounters of strategy cases this However,","However, this strategy encounters problem in cases of low SNR.",20397
in directly prior domain. model the the transform FoE retrain We,We retrain the FoE prior model directly in the transform domain.,20398
the literature. in heterogeneity single-species existence phenotypic of in well-established is bacterial biofilms published The,The existence of phenotypic heterogeneity in single-species bacterial biofilms is well-established in the published literature.,20399
"population modeling the from biofilms of However, viewpoint dynamics interactions, in social of the i.e.","However, the modeling of population dynamics in biofilms from the viewpoint of social interactions, i.e.",20400
"shifted a for revised Then, by the validation boundaries our were inferred algorithm bit boundaries.","Then, the inferred boundaries were revised by our validation algorithm for a bit shifted boundaries.",20401
applied It IoT our the in to work is environment. enough proves be that feasible,It proves that our work is feasible enough to be applied in the IoT environment.,20402
molecule before reaching codon). stop drop ribosomes mRNA a off may the,ribosomes may drop off the mRNA molecule before reaching a stop codon).,20403
show Our not this true. that is results,Our results show that this is not true.,20404
are species either diploid. Evolving or haploid,Evolving species are either haploid or diploid.,20405
rates. pleiotropically A and attack locus single determines nutritional prey quality predator,A single locus pleiotropically determines prey nutritional quality and predator attack rates.,20406
"does rates, coexistence, synergistic mediate antagonistic not. affecting these can pleiotropy By pleiotropy while","By affecting these rates, synergistic pleiotropy can mediate coexistence, while antagonistic pleiotropy does not.",20407
interactions idiotype. The clones of between links represent nodes almost complementary possible between,The links between nodes represent possible interactions between clones of almost complementary idiotype.,20408
of several we model or nodes. presence the self To permanently one occupy,To model the presence of self we permanently occupy one or several nodes.,20409
of results be autoimmune helpful understand The could to network aspects and disorders. probabilistic,The results could be helpful to understand network and probabilistic aspects of autoimmune disorders.,20410
"import reach the to an plays role. goal, color different Among features proposed this","Among the different features proposed to reach this goal, color plays an import role.",20411
uncertain for vague structure information. algebraic elegant triangle Bilattice-based and reasoning an with provides,Bilattice-based triangle provides an elegant algebraic structure for reasoning with vague and uncertain information.,20412
the over sometimes intuitive. triangle not by intervals is ordering induced Moreover the bilattice-based the,Moreover the ordering induced over the intervals by the bilattice-based triangle is not sometimes intuitive.,20413
"to Firstly, converted procedure. are Z-numbers a using standard numbers fuzzy","Firstly, Z-numbers are converted to fuzzy numbers using a standard procedure.",20414
of approach. the feasibility obtained the Results show,Results obtained show the feasibility of the approach.,20415
model simple SIR implicit derived vector An and used. dynamics was with Methods:,Methods: An simple SIR model with implicit vector dynamics was derived and used.,20416
a constraint side-information. propagation of success exploiting such to been be proved has The,The constraint propagation has been proved to be a success of exploiting such side-information.,20417
"have of propagation In multi-view proposed. been some methods years, constraint recent","In recent years, some methods of multi-view constraint propagation have been proposed.",20418
"problem views of fusing remains reasonably the unaddressed. However, different","However, the problem of reasonably fusing different views remains unaddressed.",20419
calculate factors This all algorithm of avoid PLSR suboptimal globally the solutions. can to,This algorithm can calculate all the factors of PLSR globally to avoid suboptimal solutions.,20420
"exploration, quick and sampled are efficient according Metropolis-Hastings For a to algorithm. these segmentations the","For a quick and efficient exploration, these segmentations are sampled according to the Metropolis-Hastings algorithm.",20421
of dataset. We on strategy MRI the segmentation a potential our prostate demonstrate,We demonstrate the potential of our strategy on a prostate segmentation MRI dataset.,20422
"Secondly, scale-space that individuality. encode we fin fingerprinting selective show successfully combinatorial, can","Secondly, we show that combinatorial, scale-space selective fingerprinting can successfully encode fin individuality.",20423
"labels applications, obtaining machine is learning labeled many and more is expensive. scarce data In","In many machine learning applications, labeled data is scarce and obtaining more labels is expensive.",20424
"constraints derived laws are knowledge, from from domain These of e.g., known physics. prior","These constraints are derived from prior domain knowledge, e.g., from known laws of physics.",20425
vision the on this We effectiveness approach of tasks. real world demonstrate computer and simulated,We demonstrate the effectiveness of this approach on real world and simulated computer vision tasks.,20426
"debuggers compliance proposed. the of been To requirements, KBs (non-interactive) given KB guarantee with have","To guarantee the compliance of KBs with given requirements, (non-interactive) KB debuggers have been proposed.",20427
Previous work of visual features samples. material on relies on hand-engineered classification strongly,Previous work on material classification relies strongly on hand-engineered features of visual samples.,20428
multi-residual number the residual functions residual network The of the proposed increases blocks. in,The proposed multi-residual network increases the number of residual functions in the residual blocks.,20429
"improves accuracy. significantly wider, that which architecture generates rather than deeper, are models Our","Our architecture generates models that are wider, rather than deeper, which significantly improves accuracy.",20430
problem visual challenging tracking. of estimation a a is object Accurate target scale research in,Accurate scale estimation of a target is a challenging research problem in visual object tracking.,20431
estimate employ Most an exhaustive target size. the search state-of-the-art methods to scale,Most state-of-the-art methods employ an exhaustive scale search to estimate the target size.,20432
search encountered expensive exhaustive The strategy with and is computationally large struggles when variations. scale,The exhaustive search strategy is computationally expensive and struggles when encountered with large scale variations.,20433
of This in paper tracking-by-detection investigates accurate the problem robust a framework. and scale estimation,This paper investigates the problem of accurate and robust scale estimation in a tracking-by-detection framework.,20434
"of approach. to computational strategies Additionally, our reduce investigate the cost we","Additionally, we investigate strategies to reduce the computational cost of our approach.",20435
"This us network and use both to accurate a which enables generic fast. is architecture,","This enables us to use a generic network architecture, which is both accurate and fast.",20436
method the is of Our faster considerably many than methods. state-of-the-art,Our method is considerably faster than many of the state-of-the-art methods.,20437
mechanism a Visual in scene key human is perception. grouping,Visual grouping is a key mechanism in human scene perception.,20438
two study to I cooperation the two-population game. mechanisms based in on punishment snowdrift promote,I study two mechanisms based on punishment to promote cooperation in the two-population snowdrift game.,20439
with defectors consists given to The exogenous letting a mechanism frequency. punish second cooperators of,The second mechanism consists of letting cooperators to punish defectors with a given exogenous frequency.,20440
Results an agent-based non-linear differential analytically also using and obtained simulation. through equations were,Results were obtained analytically through non-linear differential equations and also using an agent-based simulation.,20441
"codes is trained the the labels data. supervised, When training are hashing on using","When hashing is supervised, the codes are trained using labels on the training data.",20442
upper-bound give baselines the a These and supervised on a of lower- scheme. hashing performance,These baselines give a lower- and upper-bound on the performance of a supervised hashing scheme.,20443
of understanding findings in These timing. improved event regulation an facilitate noise,These findings facilitate an improved understanding of noise regulation in event timing.,20444
which and are illusions specific orientations in misperceived. a angles are subfamily Geometric,Geometric illusions are a specific subfamily in which orientations and angles are misperceived.,20445
lung standard for a Computed imaging and modality detecting cancer. tomography is assessing,Computed tomography imaging is a standard modality for detecting and assessing lung cancer.,20446
an motion. suit-based to Alternative exoskeleton or inertial use systems measurement units capture several,Alternative suit-based systems use several inertial measurement units or an exoskeleton to capture motion.,20447
"performance. enhancing block scheme, this to diagonal the form, computational matrix transition reduces In","In this scheme, the transition matrix reduces to block diagonal form, enhancing computational performance.",20448
"considerable methods drawn based the have attention. graph Recently,","Recently, the graph based methods have drawn considerable attention.",20449
offline. ($k$NN) construct neighbor graph methods $k$-nearest a These,These methods construct a $k$-nearest neighbor ($k$NN) graph offline.,20450
and the advantages hierarchical nearest-neighbor-graph methods. of nicely Efanna based based combines methods structure,Efanna nicely combines the advantages of hierarchical structure based methods and nearest-neighbor-graph based methods.,20451
released on is based on this research Github. library EFANNA A,A library EFANNA based on this research is released on Github.,20452
"training carefully this, is algorithm Based on a designed. two-stage","Based on this, a two-stage training algorithm is carefully designed.",20453
Visualization model our shows the captures semantics. trained high-level expression method that with improved,Visualization shows that the model trained with our method captures improved high-level expression semantics.,20454
provided A assay concentration-inhibition chemical. ToxCast each relationships aromatase for,A ToxCast aromatase assay provided concentration-inhibition relationships for each chemical.,20455
led individual such chemicals to Exposures effects. to never,Exposures to individual chemicals never led to such effects.,20456
thousands Arguably papers are of image several dedicated denoising. to,Arguably several thousands of papers are dedicated to image denoising.,20457
fields cascade include reaction (CSF) trainable Successful approaches and shrinkage nonlinear diffusion (TNRD). of,Successful approaches include cascade of shrinkage fields (CSF) and trainable nonlinear reaction diffusion (TNRD).,20458
"paper overcome this problem, we In order this introduce to strategy. in multi-scale a","In order to overcome this problem, in this paper we introduce a multi-scale strategy.",20459
"qualitatively quite both and verify model accurate, which often is quantitatively. we Our","Our model is often quite accurate, which we verify both qualitatively and quantitatively.",20460
"and the between deconvolution layer? is our relationship proposed Firstly, the layer What","Firstly, What is the relationship between our proposed layer and the deconvolution layer?",20461
"convolutions why a And are space low-resolution (LR) choice? better in secondly,","And secondly, why are convolutions in low-resolution (LR) space a better choice?",20462
degrade plant polysaccharides. efficiently cell-wall are machineries which Cellulosomes complex multi-enzyme,Cellulosomes are complex multi-enzyme machineries which efficiently degrade plant cell-wall polysaccharides.,20463
The proteins domains cellulosome multiple disordered intrinsically are together tethered by the regions. of often,The multiple domains of the cellulosome proteins are often tethered together by intrinsically disordered regions.,20464
not these The of and properties functions understood. disordered are well linkers,The properties and functions of these disordered linkers are not well understood.,20465
the Heated target in parallel run convergence and the chains posterior to distribution. accelerate are,Heated chains are run in parallel and accelerate the convergence to the target posterior distribution.,20466
implementing addressed by Identifiability algorithms. switching are label issues,Identifiability issues are addressed by implementing label switching algorithms.,20467
these different explore from two whether we mechanisms. fundamentally patterns Here arise,Here we explore whether these two patterns arise from fundamentally different mechanisms.,20468
same true using or is images. The corrupted for less also clearer,The same is also true for using clearer or less corrupted images.,20469
due is to The two improvement innovations.,The improvement is due to two innovations.,20470
affects Sepsis year. worldwide each people of millions,Sepsis affects millions of people worldwide each year.,20471
"rate has immediate survivors. requires readmission and high Sepsis, for treatment a","Sepsis, requires immediate treatment and has a high readmission rate for survivors.",20472
of the is also expensive conditions treat. It to most one,It is also one of the most expensive conditions to treat.,20473
most leads scalability training in poor This to practical of scenarios. DNN,This leads to poor scalability of DNN training in most practical scenarios.,20474
"reconstructions noisy the incomplete. However, and are generally","However, the reconstructions are noisy and generally incomplete.",20475
"input a the inverse TTS. and model between More intensity importantly, relationship demonstrates linear biophysical","More importantly, the biophysical model demonstrates a linear relationship between input intensity and inverse TTS.",20476
The practical dedicated segmentation. documents application administrative to is,The practical application is dedicated to administrative documents segmentation.,20477
as then set a be can bigger batch. process automatic performed on the An much,An automatic process can then be performed on the much bigger set as a batch.,20478
show results a processing Our experiments in very very interesting with time. competitive,Our experiments show very interesting results in with a very competitive processing time.,20479
(FISTA). shrinkage solved is fast problem using The algorithm inverse thresholding iterative,The inverse problem is solved using fast iterative shrinkage thresholding algorithm (FISTA).,20480
results effectiveness proposed algorithm. Experimental of the show,Experimental results show the effectiveness of proposed algorithm.,20481
"the this plane vision. ground framework with In of detection we propose stereo a paper,","In this paper, we propose a framework of the ground plane detection with stereo vision.",20482
The semantic define among one of the fashion-product fashion-attributes to images. similarity ingredients are,The fashion-attributes are one of the ingredients to define semantic similarity among fashion-product images.,20483
by can different be features. that image described know We one,We know that one image can be described by different features.,20484
ranking of list improve Multi-feature be ranking can fusion utilized the query. to,Multi-feature fusion ranking can be utilized to improve the ranking list of query.,20485
"and analyze this re-ranking In graph first aspect. we manifold structure paper, fusion from multi-feature","In this paper, we first analyze graph structure and multi-feature fusion re-ranking from manifold aspect.",20486
results outperform than that TTNG our methods. MFR other and state-of-the-art experimental show The,The experimental results show that our TTNG and MFR outperform than other state-of-the-art methods.,20487
and ranking in retrieval. neglected Incompatibility is image descriptor of image always,Incompatibility of image descriptor and ranking is always neglected in image retrieval.,20488
"PUD we realize image and Thus, to retrieval. use manifold ranking","Thus, we use manifold ranking and PUD to realize image retrieval.",20489
"criterion presented, rejection/incorporation. is based specificity ligand on ratio A of new that the is","A new criterion of ligand specificity is presented, that is based on the ratio rejection/incorporation.",20490
herbivory that costs insect suffers This producers. from crop millions from,This crop suffers from insect herbivory that costs millions from producers.,20491
real-world from experimented a Brazil. soybean leaves with We in crop collected,We experimented with real-world leaves collected from a soybean crop in Brazil.,20492
publicly justify images thermal model experimental and infrared assumptions. The proposals available our studies on,The experimental studies on publicly available thermal infrared images justify our proposals and model assumptions.,20493
data Our and publicly will sets codes available. be,Our codes and data sets will be publicly available.,20494
"reducing With day, dimensionality the the every scale of growing data (a.k.a.","With the scale of data growing every day, reducing the dimensionality (a.k.a.",20495
sketching) paramount importance. task emerged of a as high-dimensional of has data,sketching) of high-dimensional data has emerged as a task of paramount importance.,20496
modeling convolutional to paper evaluated. application the of uncertainty neural is networks In this,In this paper the application of uncertainty modeling to convolutional neural networks is evaluated.,20497
method is novel based information the A introduced. network's predictions for uncertainty on adjusting,A novel method for adjusting the network's predictions based on uncertainty information is introduced.,20498
pessimistic in to scores. the or allows its optimistic be either network This prediction,This allows the network to be either optimistic or pessimistic in its prediction scores.,20499
"allowing evaluation implementation details the methodological for presented. a are fast aspects, Besides","Besides the methodological aspects, implementation details allowing for a fast evaluation are presented.",20500
"presented strongly introduced a architecture multilabel network approach. Furthermore, benefits is that from the","Furthermore, a multilabel network architecture is introduced that strongly benefits from the presented approach.",20501
Variational inference lies the state-of-the-art at core of many algorithms.,Variational inference lies at the core of many state-of-the-art algorithms.,20502
"we HMC geared some algorithm extensions to convergence. faster Additionally, towards the introduce","Additionally, we introduce some extensions to the HMC algorithm geared towards faster convergence.",20503
performance these by of results. The theoretical are in improvements reflected modifications our advantages experimental,The theoretical advantages of these modifications are reflected by performance improvements in our experimental results.,20504
to accurate single-view use measurement. metrology compute We,We use single-view metrology to compute accurate measurement.,20505
thus can be subproblems closed-forms and All solved have the efficiently.,All the subproblems have closed-forms and can thus be solved efficiently.,20506
viable simulation code a option. real Testing in is,Testing real code in simulation is a viable option.,20507
for results highlight test agents of The generation. using BDI advantages the,The results highlight the advantages of using BDI agents for test generation.,20508
also posterior by approximate the accelerating using probability. Fisher of We vector the propose value,We also propose accelerating the Fisher vector by using the approximate value of posterior probability.,20509
presented solution to 'long tail' by A DM. is the problem described,A solution is presented to the 'long tail' problem described by DM.,20510
The human memory incredibly of is basis neural complex.,The neural basis of human memory is incredibly complex.,20511
"if texture. lacks Processing distinct, gracefully non-repeating and overlapping, footage degrade display the","Processing and display degrade gracefully if the footage lacks distinct, overlapping, non-repeating texture.",20512
"(x,y) odometry a visual algorithm translations pairs. produces proposed Our over image for distribution","Our proposed visual odometry algorithm produces a distribution over (x,y) translations for image pairs.",20513
motions. Robustness with by scenes camera known is training on obtained synthetic,Robustness is obtained by training on synthetic scenes with known camera motions.,20514
"joint model In work, network architectures. language-visual three this neural different study we","In this work, we study three different joint language-visual neural network model architectures.",20515
of expression result genes. many from These traits,These traits result from expression of many genes.,20516
"phenotype increases robustness mutations. with Moreover, respect learning this to","Moreover, this learning increases phenotype robustness with respect to mutations.",20517
"However, risk procedure since additional learning an leads acceleration can errors. this produce to","However, this acceleration leads to an additional risk since learning procedure can produce errors.",20518
experimental results anticipation on data organisms. some These by explain of environment recent changes,These results explain recent experimental data on anticipation of environment changes by some organisms.,20519
analogous in propose We and pooling graph domain. to those operators the convolution regular,We propose graph convolution and pooling operators analogous to those in the regular domain.,20520
"take the in applying form graph Signal filters convolution spectral multipliers, of the spectral domain.","Signal filters take the form of spectral multipliers, applying convolution in the graph spectral domain.",20521
of multiple representation levels and abstraction. with,with multiple levels of representation and abstraction.,20522
is implemented The HSMClust. method package an in R,The method is implemented in an R package HSMClust.,20523
using video summaries SumMe well baseline approaches. the as our dataset as We evaluate,We evaluate our video summaries using the SumMe dataset as well as baseline approaches.,20524
Many to in advancements attributed Vision recent are datasets. large Computer,Many recent advancements in Computer Vision are attributed to large datasets.,20525
millions of examples days. is train a to possible over models few within It,It is possible to train models over millions of examples within a few days.,20526
code training We release to metrics. TensorFlow for model and plan computing a for,We plan to release code for training a TensorFlow model and for computing metrics.,20527
approximate kernel defined the any continuous allows us efficiently shift-invariant Log-HS This to distance. using,This allows us to efficiently approximate any continuous shift-invariant kernel defined using the Log-HS distance.,20528
"framework datasets. we on task classification the tested of our twelve image Empirically, on challenging","Empirically, we tested our framework on the task of image classification on twelve challenging datasets.",20529
numerical We view as and contributions. problem make a unique three this integration,We view this as a numerical integration problem and make three unique contributions.,20530
"a instead point low sets of using grid. First, the discrepancy we explore space","First, we explore the space using low discrepancy point sets instead of a grid.",20531
"topic. binarization remains research it highly still passe, is popular considered a Although","Although binarization is considered passe, it still remains a highly popular research topic.",20532
of is. a what we binarization propose this rethinking In paper,In this paper we propose a rethinking of what binarization is.,20533
"agent's efficiency. show data use that of the Finally, learning significantly we this improves the","Finally, we show that the use of this data significantly improves the agent's learning efficiency.",20534
to normalized. be distances nonnegative However and signal the these require,However these distances require the signal to be nonnegative and normalized.,20535
allows of framework existing application the numerical The methods.,The framework allows the application of existing numerical methods.,20536
visually coherent image proposes segmentation semantically Our segments. and,Our segmentation proposes visually and semantically coherent image segments.,20537
codes These changes non-semantic against and robust in noise the binary are very image.,These binary codes are very robust against noise and non-semantic changes in the image.,20538
of the (e.g. previous number All images were few limited papers category to the of,All the previous papers were limited to few number of category of the images (e.g.,20539
Experiments the by segmentation that margin. state-of-the-art our outperform segmentation methods non-semantic show algorithm large,Experiments show that our segmentation algorithm outperform the state-of-the-art non-semantic segmentation methods by large margin.,20540
of the search. neighborhoods solutions uses infeasible The concept heuristic the and during visits,The heuristic uses the concept of neighborhoods and visits infeasible solutions during the search.,20541
A is solution. applied an heuristic construct to greedy initial,A greedy heuristic is applied to construct an initial solution.,20542
favour often in performance. Stable prediction of overlooked is,Stable prediction is often overlooked in favour of performance.,20543
"prevails stability areas key models adopting in as Yet, critical as when healthcare.","Yet, stability prevails as key when adopting models in critical areas as healthcare.",20544
order study stabilization higher Our correlations. by scheme proposes a feature detecting,Our study proposes a stabilization scheme by detecting higher order feature correlations.,20545
correlation is modelled using features among network. an order higher autoencoder Latent,Latent higher order correlation among features is modelled using an autoencoder network.,20546
Australian hospital. experiments are heart from conducted a Our cohort on an failure,Our experiments are conducted on a heart failure cohort from an Australian hospital.,20547
Consistency measured for ratio parameters. using signal-to-noise Stability for model feature index subsets was and,Stability was measured using Consistency index for feature subsets and signal-to-noise ratio for model parameters.,20548
pedestrian be characteristics to microscopic understood. need The flow,The microscopic pedestrian flow characteristics need to be understood.,20549
"image developed. and processing semi manual systems data were Manual, automatic collection","Manual, semi manual and automatic image processing data collection systems were developed.",20550
physical based A was model simulation pedestrian microscopic developed. also,A physical based microscopic pedestrian simulation model was also developed.,20551
or characteristics The formulations microscopic are pedestrian of performance the explained. flow,The formulations of the flow performance or microscopic pedestrian characteristics are explained.,20552
are relationship and of described. between the Sensitivity performances simulation the flow,Sensitivity of the simulation and relationship between the flow performances are described.,20553
(BNs) distributions. high-dimensional Bayesian networks graphical probability are models are useful that for representing,Bayesian networks (BNs) are graphical models that are useful for representing high-dimensional probability distributions.,20554
and empirical give mixing death properties. evidence has superior that birth the process We,We give empirical evidence that the birth and death process has superior mixing properties.,20555
models functional complex regions. nicely CNNs these recognize,CNNs models nicely recognize these complex functional regions.,20556
"for the is rain that Moreover, it proposed removal. method used can expected be","Moreover, it is expected that the proposed method can be used for rain removal.",20557
"also training this Networks, optimized data. loop Deep feedback of components The are using","The components of this feedback loop are also Deep Networks, optimized using training data.",20558
information be to valuable diagnosis. reliable a can resource for images medical Medical support,Medical images can be a valuable resource for reliable information to support medical diagnosis.,20559
"be dehazing. also the near-infrared distortion considered color should for Therefore,","Therefore, the color distortion should also be considered for near-infrared dehazing.",20560
"colors images. details the dehazed on can and natural-looking Thus, induced be fine","Thus, natural-looking colors and fine details can be induced on the dehazed images.",20561
Autoencoder a We novel present for (VAE). Variational method constructing,We present a novel method for constructing Variational Autoencoder (VAE).,20562
denoising. Experimental proposed can approach results applied show to near-infrared the also that be,Experimental results also show that the proposed approach can be applied to near-infrared denoising.,20563
"astronomy other and important many for imaging, fields. regime scotopic security, The biomedical is","The scotopic regime is important for biomedical imaging, security, astronomy and many other fields.",20564
a feature of and framework. asymptotically optimal dynamic speed-accuracy A is this key tradeoff,A dynamic and asymptotically optimal speed-accuracy tradeoff is a key feature of this framework.,20565
"widely sparse dictionary for restoration coding been have and used Recently, image problems. learning","Recently, dictionary learning and sparse coding have been widely used for image restoration problems.",20566
"loss regions. the generate non-rain edge detail approach unwanted in this can and However, artifacts","However, this approach can generate unwanted edge artifacts and detail loss in the non-rain regions.",20567
"dictionaries between is correlation rain and calculated. the the matrix the regions, rain non-rain In","In the rain regions, the correlation matrix between the rain and non-rain dictionaries is calculated.",20568
are soft biometric recognition. for complementary gender and traits Age face,Age and gender are complementary soft biometric traits for face recognition.,20569
Max-SAT extent properties. what We differ random their in the SAT to problems and analyze,We analyze to what extent the random SAT and Max-SAT problems differ in their properties.,20570
local results random outperforms solvers search illustrate benchmarks. state-of-the-art many on Experimental ProMS Max-SAT that,Experimental results illustrate that ProMS outperforms many state-of-the-art local search solvers on random Max-SAT benchmarks.,20571
model From rate strategies. for (fitness) analytically non-sleeping population this sleeping and growth calculate we,From this model we analytically calculate population growth rate (fitness) for sleeping and non-sleeping strategies.,20572
"European forest quite below, to cover time the close Romania's is average. at this but","Romania's forest cover is at this time below, but quite close to the European average.",20573
diagnostic retrieval experts. medical image decisions by clinical can Content-based support,Content-based medical image retrieval can support diagnostic decisions by clinical experts.,20574
a identification. fingerprint role play in major Minutiae,Minutiae play a major role in fingerprint identification.,20575
latent which poor quality. fingerprints of usually difficult are for Extracting is minutiae reliable,Extracting reliable minutiae is difficult for latent fingerprints which are usually of poor quality.,20576
a map to a fingerprints fixed minutia-score with are mapped Raw correspondingly-sized stride.,Raw fingerprints are mapped to a correspondingly-sized minutia-score map with a fixed stride.,20577
And of extracted number large a minutiae be thus threshold. given a through will,And thus a large number of minutiae will be extracted through a given threshold.,20578
layers convolutional fully up. shares CNN to network speed the convolutional with The,The CNN shares convolutional layers with the fully convolutional network to speed up.,20579
"real deep occur is could unclear learning in However, brain. the whether it","However, it is unclear whether deep learning could occur in the real brain.",20580
available of databases on two power approach. the publicly our show Experiments,Experiments on two publicly available databases show the power of our approach.,20581
"study actions. we of paper, the dexterous prediction In this","In this paper, we study the prediction of dexterous actions.",20582
on evaluate trained We three benchmarks. challenging models,We evaluate trained models on three challenging benchmarks.,20583
"learning. analytically supervised for models processes are Gaussian yet tractable powerful,","Gaussian processes are powerful, yet analytically tractable models for supervised learning.",20584
"propose Graph-based TV this we Adaptive In paper, (AGTV).","In this paper, we propose Adaptive Graph-based TV (AGTV).",20585
"it the in gradient domains. wavelet promotes graph sparsity and Moreover,","Moreover, it promotes sparsity in the wavelet and graph gradient domains.",20586
is to most representative used Boosting features. the meta learning find local,Boosting meta learning is used to find the most representative local features.,20587
One is analysis as (PGA). generalization principal known such geodesic,One such generalization is known as principal geodesic analysis (PGA).,20588
highlighted novelties contributions are and reviewed of discussed. approaches and Main the,Main contributions and novelties of the reviewed approaches are highlighted and discussed.,20589
the using competitive are current conducted solid simple are baselines state-of-the-arts. Experiments that yet with,Experiments are conducted using simple yet solid baselines that are competitive with the current state-of-the-arts.,20590
"the discuss of of the manipulating Moreover, we through possibility computational approaches. aesthetics images","Moreover, we discuss the possibility of manipulating the aesthetics of images through computational approaches.",20591
"of thus complex handling remaining is The framework models, while domain-size independent. capable","The framework is thus capable of handling complex models, while remaining domain-size independent.",20592
series are the estimating for time considered. of methods functional long-run stationary covariance Bootstrap,Bootstrap methods for estimating the long-run covariance of stationary functional time series are considered.,20593
"decision In theory makers. this prospect based paper, for adopt models we","In this paper, we adopt prospect theory based models for decision makers.",20594
been single super-resolution has recent image interest. much (SR) of Sparsity constrained,Sparsity constrained single image super-resolution (SR) has been of much recent interest.,20595
correlation exploited channel are Edge similarities RGB as amongst constraints. color bands cross,Edge similarities amongst RGB color bands are exploited as cross channel correlation constraints.,20596
suggests evidence that neural criticality. cortex at operate Mounting of cerebral networks,Mounting evidence suggests that neural networks of cerebral cortex operate at criticality.,20597
variability largely neuronal spiking impacts of network this operating unknown. How near state is the,How operating near this network state impacts the variability of neuronal spiking is largely unknown.,20598
"and highly are To achieve typically accuracy, using numerous complex high features algorithms. parameterized extracted","To achieve high accuracy, numerous features are typically extracted using complex and highly parameterized algorithms.",20599
"and tertiary binding sites. functions, defines Protein classification its structure","Protein tertiary structure defines its functions, classification and binding sites.",20600
often similar structural lead proteins thereof. Similar characteristics two characteristics to between the,Similar structural characteristics between two proteins often lead to the similar characteristics thereof.,20601
real structural accurately similarity research time crucial issue. is Determining a in,Determining structural similarity accurately in real time is a crucial research issue.,20602
vision. pattern is recognition and inspired from Our computer scoring scheme,Our scoring scheme is inspired from pattern recognition and computer vision.,20603
structures. is tested on method benchmark effectiveness of The our standard,The effectiveness of our method is tested on standard benchmark structures.,20604
"robust While (e.g. analysis quality nowadays many the in sufficient is reconstruction cases,","While reconstruction quality nowadays is sufficient in many cases, the robust analysis (e.g.",20605
has recognition. in incredible scene object success achieved and DCNN,DCNN has achieved incredible success in object and scene recognition.,20606
"recognition Furthermore, the be classification model. potentially can and spatial through multiscale improved accuracy resolution","Furthermore, the recognition accuracy and spatial resolution can be potentially improved through multiscale classification model.",20607
contribution paper threefold. this is of The,The contribution of this paper is threefold.,20608
"First, for using indoor robust DCNN. and we a navigation parsimonious approach propose","First, we propose a robust and parsimonious approach for indoor navigation using DCNN.",20609
of require not does probability the data. test The the proposed generating distribution knowledge underlying,The proposed test does not require knowledge of the underlying probability distribution generating the data.,20610
validate major procedure e-commerce from data online a We website. this on,We validate this procedure on data from a major online e-commerce website.,20611
'things'. to abstraction we of represent In this as scene propose paper a an,In this paper we propose to represent a scene as an abstraction of 'things'.,20612
We in contributions work. three this make,We make three contributions in this work.,20613
"of it simple we and call things syntax. study observable First, 'things', properties","First, we study simple observable properties of 'things', and call it things syntax.",20614
The vectors proposed feature builds and them three database. into system MySQL stores,The proposed system builds three feature vectors and stores them into MySQL database.,20615
weakly learning from We approach for present an transcriptions. actions of human supervised video,We present an approach for weakly supervised learning of human actions from video transcriptions.,20616
The clusters approach remaining individual then the candidates. $\mu$C in our step final,The final step in our approach then clusters the remaining individual $\mu$C candidates.,20617
practice. model in found This the fails to behavior however of non-Gaussian noises assumption,This assumption however fails to model the behavior of non-Gaussian noises found in practice.,20618
years. significant achieved recent in has progress Planning,Planning has achieved significant progress in recent years.,20619
a of This unknown instances images paper hash provides framework classes. to object containing,This paper provides a framework to hash images containing instances of unknown object classes.,20620
"problems, recognition huge might of we amount object have data. to many In access","In many object recognition problems, we might have access to huge amount of data.",20621
function combine gradients score results Carlo gradients. in This new gradients that and reparameterization Monte,This results in new Monte Carlo gradients that combine reparameterization gradients and score function gradients.,20622
probabilistic approach two complex demonstrate We our on inference models. variational for,We demonstrate our approach on variational inference for two complex probabilistic models.,20623
"capabilities at similar could we paper, be In machine look in how developed vision. this","In this paper, we look at how similar capabilities could be developed in machine vision.",20624
variety vast of illusions Geometric-optical of a subclass illusions. a (GOI) are visual,Geometric-optical illusions (GOI) are a subclass of a vast variety of visual illusions.,20625
the circular of with specifically targets. Here deal case we,Here we specifically deal with the case of circular targets.,20626
first parallel. saliency learns location image on each in local automatically DSCLRCN related features,DSCLRCN first automatically learns saliency related local features on each image location in parallel.,20627
testing. efficiently end-to-end when be network whole The can trained and works,The whole network can be trained end-to-end and works efficiently when testing.,20628
eosin and stained general morphology. for Sections haematoxylin were with,Sections were stained with haematoxylin and eosin for general morphology.,20629
was divided and regions. into The stomach pyloric fundic,The stomach was divided into fundic and pyloric regions.,20630
"into the was which mucosa The into thrown opened. glands pits, gastric gastric","The mucosa was thrown into gastric pits, into which the gastric glands opened.",20631
villi poorly small many but intestinal The showed developed occasional intestine crypts were.,The small intestine showed many villi but occasional poorly developed intestinal crypts were.,20632
"we localization problem generation Specifically, navigation map respectively. and into visual the decompose","Specifically, we decompose the visual navigation problem into localization and map generation respectively.",20633
in shows improvement a remarkable the classification and objects. recognition of It,It shows a remarkable improvement in the recognition and classification of objects.,20634
"As however, interesting training the CNN in yet learning, challenging. is deep other","As in other deep learning, however, training the CNN is interesting yet challenging.",20635
"metaheuristic In another different algorithms proposed, paper, of i.e. this been strategy type has with","In this paper, another type of metaheuristic algorithms with different strategy has been proposed, i.e.",20636
Microcanonical Neural Convolutional to optimize Network. Annealing,Microcanonical Annealing to optimize Convolutional Neural Network.,20637
an cell of becoming and tool application in stem are research. study essential the Biomaterials,Biomaterials are becoming an essential tool in the study and application of stem cell research.,20638
networks state-of-the-art images. labeled neural these train we Then with,Then we train state-of-the-art neural networks with these labeled images.,20639
"well. we cross-country several gender-related as Along the uncover patterns way,","Along the way, we uncover several gender-related cross-country patterns as well.",20640
video-to-language integrated a be propose models. any high-level that with We detector concept word can,We propose a high-level concept word detector that can be integrated with any video-to-language models.,20641
important The proposed word properties. has detector two,The proposed word detector has two important properties.,20642
"training. for First, external require knowledge any does sources not it","First, it does not require any external knowledge sources for training.",20643
"comparable also description. performance attain task, for movie We the other","We also attain comparable performance for the other task, movie description.",20644
methods previous approach optimizing complementary a new curve introduce geometry. by to We,We introduce a new approach complementary to previous methods by optimizing curve geometry.,20645
"we iterative the theory derivatives. on particular, based a novel In shape propose algorithm of","In particular, we propose a novel iterative algorithm based on the theory of shape derivatives.",20646
is and performance. related This networks to applies very also closely their to residual,This also applies to residual networks and is very closely related to their performance.,20647
Our tumor approach on categorical pathologist-level prediction. severity performance achieves three-class,Our approach achieves pathologist-level performance on three-class categorical tumor severity prediction.,20648
system integral approaching aimed camera traffic. The is at,The integral system camera is aimed at approaching traffic.,20649
"early with pedestrians warning approaching achieves promising of sufficiently experimental system results, vehicles. The providing","The system achieves promising experimental results, providing pedestrians with sufficiently early warning of approaching vehicles.",20650
to This redundant huge of lots data. with leads datasets,This leads to huge datasets with lots of redundant data.,20651
of assembly A to data is this prior advisable. filtering,A filtering of this data prior to assembly is advisable.,20652
"filtering faster algorithm. Bignorm, read We a and quality-conscious present","We present Bignorm, a faster and quality-conscious read filtering algorithm.",20653
"using Specifically, as known of the example Switch-Side Chain-Chess chess the (SSCC). variant","Specifically, using the example of the chess variant known as Switch-Side Chain-Chess (SSCC).",20654
solution implemented We therefore a tested programming the using language. propose C++ and,We therefore propose a solution implemented and tested using the C++ programming language.,20655
other It at interest in of spotting aims cameras. a person,It aims at spotting a person of interest in other cameras.,20656
"reported. In were hand-crafted evaluation small-scale days, early the algorithms predominantly and","In the early days, hand-crafted algorithms and small-scale evaluation were predominantly reported.",20657
of two way matching novel studied. A images is shifting with transformation,A novel way of matching two images with shifting transformation is studied.,20658
indicate of proposed the the preliminary method. experimental The effectiveness results,The preliminary experimental results indicate the effectiveness of the proposed method.,20659
is to minimization rank representation norm minimize of adopted Nuclear the the matrix.,Nuclear norm minimization is adopted to minimize the rank of the representation matrix.,20660
"learned model able reward-based the skills are In mitigate risk misspecification. addition, to aware","In addition, the learned risk aware skills are able to mitigate reward-based model misspecification.",20661
Matlab and provide code both Gwyddion. We implementations for,We provide code implementations for both Matlab and Gwyddion.,20662
"HRRSIR. generate feature two representations proposed effective powerful to are for schemes end, To this","To this end, two effective schemes are proposed to generate powerful feature representations for HRRSIR.",20663
to on is The with applied database. compare the a data records resulting,The resulting data is applied to compare with the records on a database.,20664
system be This security and can control. also used for traffic,This system can also be used for security and traffic control.,20665
"initial required practice, is avoid In often of careful to tuning convergence values issues.","In practice, careful tuning of initial values is often required to avoid convergence issues.",20666
Networks time during We deal budgets (DNNs) application. Neural which Impatient with propose Deep dynamic,We propose Impatient Deep Neural Networks (DNNs) which deal with dynamic time budgets during application.,20667
"architectures, In evaluate approach our for and experiments, we distributions, budget different datasets.","In experiments, we evaluate our approach for different budget distributions, architectures, and datasets.",20668
expected compared show baselines. gain accuracy Our results a significant to in common,Our results show a significant gain in expected accuracy compared to common baselines.,20669
the segmentation. guarantees method in the mechanism self-balancing of stability model The the in,The self-balancing mechanism in the model guarantees the stability of the method in segmentation.,20670
Generative (GANs) adversarial generative successful are networks models. deep,Generative adversarial networks (GANs) are successful deep generative models.,20671
minimax based are game. on GANs a two-player,GANs are based on a two-player minimax game.,20672
the We and that novel algorithm a estimation ratio f-divergence repeats propose density minimization.,We propose a novel algorithm that repeats the density ratio estimation and f-divergence minimization.,20673
ratio divergence is density and useful. what stable relative is,what divergence is stable and relative density ratio is useful.,20674
expectation variational distributions. a on based derive approach truncated novel We posterior maximization,We derive a novel variational expectation maximization approach based on truncated posterior distributions.,20675
achieve While results not look good and performance tracking real. cartoonish they the,While they achieve good tracking performance the results look cartoonish and not real.,20676
"algorithmic In we a components. setup, these live newly-introduced apply","In a live setup, we apply these newly-introduced algorithmic components.",20677
of studied. and transformation approach rotating image A is presented for matching novel,A novel approach of image matching for rotating transformation is presented and studied.,20678
force is by inspired The interaction between currents. physical approach electromagnetic,The approach is inspired by electromagnetic interaction force between physical currents.,20679
detection. fast robust neural We and deep network pedestrian propose architecture for fusion a,We propose a deep neural network fusion architecture for fast and robust pedestrian detection.,20680
of for network proposed parallel processing architecture allows fusion The speed. for multiple networks,The proposed network fusion architecture allows for parallel processing of multiple networks for speed.,20681
"when and than detecting existing Our state-of-the-arts, method performs especially small-size occluded better pedestrians.","Our method performs better than existing state-of-the-arts, especially when detecting small-size and occluded pedestrians.",20682
also than all other faster It is methods.,It is also faster than all other methods.,20683
sense problem some to in the target clinicians. similar aid A image can,A problem similar in some sense to the target image can aid clinicians.,20684
"improves evidence-based administration, diagnosis, CBIR healthcare. retrieval teaching, and text-based research and in complements","CBIR complements text-based retrieval and improves evidence-based diagnosis, administration, teaching, and research in healthcare.",20685
help and data comparing diagnostic. Metrics visual improve,Metrics help comparing visual data and improve diagnostic.,20686
architectures from can the Specially scenario. designed benefit application,Specially designed architectures can benefit from the application scenario.,20687
and visual aspects in required This recommends handle content complex healthcare. important to chapter,This chapter recommends important and complex aspects required to handle visual content in healthcare.,20688
deep problem. widely schemes training networks used Data when the suffer insufficient are augmentation data,Data augmentation schemes are widely used when deep networks suffer the insufficient training data problem.,20689
feature object both for be soft-biometrics a estimation Height can and useful tracking.,Height estimation can be a useful feature for both soft-biometrics and object tracking.,20690
temporal combination We of fusion late and to learn spatial also incorporate cues. a,We also incorporate late fusion to learn a combination of temporal and spatial cues.,20691
kind. synthetic demonstrate data existing performance sets on gains methods Tests of this over significant,Tests on synthetic data sets demonstrate significant performance gains over existing methods of this kind.,20692
Group-level on face-level few investigated works A existing Emotion (GER) Recognition have information. on,A few existing works on Group-level Emotion Recognition (GER) have investigated on face-level information.,20693
"may challenging the not face enough to GER. information Due environments, provide to","Due to the challenging environments, face may not provide enough information to GER.",20694
GER. studies Relatively multi-modal investigated have few,Relatively few studies have investigated multi-modal GER.,20695
"for GER environments. upperbody the robustness against scene and fuse We of challenging information face,","We fuse face, upperbody and scene information for robustness of GER against the challenging environments.",20696
results our achieves performance for demonstrate that framework Experimental promising very GER.,Experimental results demonstrate that our framework achieves very promising performance for GER.,20697
"been not complete there review on a However, localization. camera has image-based","However, there has not been a complete review on image-based camera localization.",20698
the to topic people is field this enter map It help quickly. to urgent,It is urgent to map this topic to help people enter the field quickly.,20699
"In overview based paper, image camera presented. this is localization of an","In this paper, an overview of image based camera localization is presented.",20700
future Trends also are development for the discussed.,Trends for the future development are also discussed.,20701
engineers It researchers not be useful but interested. people and only other to also will,It will be useful to not only researchers but also engineers and other people interested.,20702
(LF) are micro-lens cameras arrays. technologies And field one on emerging based light of the,And one of the emerging technologies are light field (LF) cameras based on micro-lens arrays.,20703
"a More classes interestingly, task features with more (e.g. trained for the","More interestingly, the features trained for a task with more classes (e.g.",20704
(e.g. with ID) task then in classes another and used fewer,ID) and then used in another task with fewer classes (e.g.,20705
the trained features for outperforms itself. other the task race),race) outperforms the features trained for the other task itself.,20706
relations to Knowledge completion information. infer aims existing from base new,Knowledge base completion aims to infer new relations from existing information.,20707
information TransR paths. we Then regularize of relation with,Then we regularize TransR with information of relation paths.,20708
"entity PTransR on experiment, task our of we prediction. the In evaluate","In our experiment, we evaluate PTransR on the task of entity prediction.",20709
Experimental models. show that previous PTransR outperforms results,Experimental results show that PTransR outperforms previous models.,20710
The decoder. a comprises and encoder twin-tailed network an,The network comprises an encoder and a twin-tailed decoder.,20711
code. The encoder graphics a generates disentangled,The encoder generates a disentangled graphics code.,20712
"adversarial on conduct large-scale cross-model experiments portability. Finally, we","Finally, we conduct large-scale experiments on cross-model adversarial portability.",20713
"stability In this of out-of-phase inhibition rhythms. leads the greater two context, to","In this context, inhibition leads to greater stability of the two out-of-phase rhythms.",20714
"support results complex. pre-B\""otzinger in computational recordings We with mouse our vitro from","We support our computational results with in vitro recordings from mouse pre-B\""otzinger complex.",20715
"of excitation increased inhibition. variability, but rhythmic recovers following Partial this blockade leads block to","Partial excitation block leads to increased rhythmic variability, but this recovers following blockade of inhibition.",20716
brain mental a function. The human network that is supports complex,The human brain is a complex network that supports mental function.,20717
"the we and discuss of control. mind challenges, ethics, Finally, promises","Finally, we discuss the challenges, ethics, and promises of mind control.",20718
reproduce a The performance of human earthquake to is idea the prediction. expert in,The idea is to reproduce the performance of a human expert in earthquake prediction.,20719
detection paper and implements compares recognition. face This and different for techniques,This paper implements and compares different techniques for face detection and recognition.,20720
novel We learning batch a deep for systematic propose effects. removing approach,We propose a novel deep learning approach for removing systematic batch effects.,20721
in have been proteasomal resist to Knots proposed proteins degradation.,Knots in proteins have been proposed to resist proteasomal degradation.,20722
degradation associates with neurodegeneration. evidence Ample proteasomal,Ample evidence associates proteasomal degradation with neurodegeneration.,20723
possibility toxicity. indeed is machinery interesting this One conformers that knotted to leading stall,One interesting possibility is that indeed knotted conformers stall this machinery leading to toxicity.,20724
either The periodically. force is applied or constant,The force is either constant or applied periodically.,20725
The a in translocated fashion. coarse-grained are modelled proteins,The translocated proteins are modelled in a coarse-grained fashion.,20726
"well-known Huntingtin disease, Huntington a is associated with neurodegenerative disease. genetically-determined","Huntingtin is associated with Huntington disease, a well-known genetically-determined neurodegenerative disease.",20727
and the that presence a of translocation. even sometimes knot hinders We jams show,We show that the presence of a knot hinders and sometimes even jams translocation.,20728
"object and Traditionally, as problems separate treated segmentation tracking are two independently. solved and","Traditionally, object tracking and segmentation are treated as two separate problems and solved independently.",20729
signal. turns a Our valuable blur unwanted often into defocus algorithm the,Our algorithm turns the often unwanted defocus blur into a valuable signal.,20730
can turning the manually of lens focus ring during achieved the by be This recording.,This can be achieved by manually turning the focus ring of the lens during recording.,20731
impact acquiring decreases the Thus just potential of threat. proposed the image the security,Thus just acquiring the image decreases the potential impact of the proposed security threat.,20732
"examples. completely almost as binarization, adversarial preprocessing, neutralizes Normal such text","Normal preprocessing, such as text binarization, almost completely neutralizes adversarial examples.",20733
method the simplify proposed speed. and procedure the the improve Thus can,Thus the proposed method can simplify the procedure and improve the speed.,20734
is standard heterogeneity a species. analysis decipher to procedure diversification bacterial and Pan-genome of genome,Pan-genome analysis is a standard procedure to decipher genome heterogeneity and diversification of bacterial species.,20735
"synteny, obtain deletion Genomes were aligned duplications. information and insertion/inversion, about to","Genomes were aligned to obtain information about synteny, insertion/inversion, deletion and duplications.",20736
Strategies and involve screening biopsies. tissue surveillance for periodic,Strategies for screening involve periodic surveillance and tissue biopsies.,20737
"eight-directional SSDCNN the other in way. incorporate hand, a On can features natural","On the other hand, SSDCNN can incorporate eight-directional features in a natural way.",20738
largely unknown. this was reasons remain so why deadly pandemic The,The reasons why this pandemic was so deadly remain largely unknown.,20739
prevent help another sequences may in this singular Monitoring pandemic. dramatic viral mutation,Monitoring this singular mutation in viral sequences may help prevent another dramatic pandemic.,20740
reproductive calls do. greater to males success a Complex simple lead than calls,Complex calls lead males to a greater reproductive success than simple calls do.,20741
points models we the analyze replicator We that propose. of the the stable equations of,We analyze the stable points of the replicator equations of the models that we propose.,20742
Engystomops pustulosus We the frog the reproductive strategy that is male conclude of rational.,We conclude that the reproductive strategy of the male frog Engystomops pustulosus is rational.,20743
is  advancing visual of to develop methods research recognition. our  The automatic goal,  The goal of our research is to develop methods advancing automatic visual recognition.,20744
interactions due to captioning challenging scenes. in Automatic the complex dynamic is real video,Automatic video captioning is challenging due to the complex interactions in dynamic real scenes.,20745
existent and experimentally is An proposed the algorithm polynomial-time to analyzed. improvement,An improvement to the existent polynomial-time algorithm is proposed and experimentally analyzed.,20746
"was the of improvement unstable hardly PUSH-and-SWAP. However, case and predictable the in","However, the improvement was unstable and hardly predictable in the case of PUSH-and-SWAP.",20747
"geo-spatial is capture modeling achieving for When critical high spatial data, it accuracy. correlations to","When modeling geo-spatial data, it is critical to capture spatial correlations for achieving high accuracy.",20748
"efficacy SAR limited However, factors. the two is of by","However, the efficacy of SAR is limited by two factors.",20749
"Second, observations models. that can linear be assumes by it explained the","Second, it assumes that the observations can be explained by linear models.",20750
framework of data. An our the of assumption linear the relaxation upside is additional on,An additional upside of our framework is the relaxation of linear assumption on the data.,20751
right of what to capture approach imagery? (ii) the satellite multiple is resolutions,(ii) what is the right approach to capture multiple resolutions of satellite imagery?,20752
what other can the spatial data-sources (iii) and estimation of improve correlations? help,and (iii) what other data-sources can help improve the estimation of spatial correlations?,20753
be is Our general observations. can quite and applied to noisy noiseless framework and both,Our framework is quite general and can be applied to both noisy and noiseless observations.,20754
through also on synthetic corroborate data. theory experiments our We extensive,We also corroborate our theory through extensive experiments on synthetic data.,20755
the Melanoma skin cancer. is form of deadliest,Melanoma is the deadliest form of skin cancer.,20756
properties. anisotropy articular and crucial its play a attributing cartilage Inhomogeneity role in,Inhomogeneity and anisotropy play a crucial role in attributing articular cartilage its properties.,20757
of shown are in the results and of pressure terms approach contact The bones.,The results are shown in terms of contact pressure and approach of the bones.,20758
derived the be directly The area. to is proportional contact latter to,The latter is derived to be directly proportional to the contact area.,20759
the Existing view model in formulation. data experimental current are of reinterpreted,Existing experimental data are reinterpreted in view of the current model formulation.,20760
union the subspaces. of from refers data Subspace drawn of problem clustering a to segmenting,Subspace clustering refers to the problem of segmenting data drawn from a union of subspaces.,20761
State-of-the-art follow solving a for two-stage approach. problem this approaches,State-of-the-art approaches for solving this problem follow a two-stage approach.,20762
"the found second step, spectral affinity. this In clustering is applying by segmentation to the","In the second step, the segmentation is found by applying spectral clustering to this affinity.",20763
"requires learning, i.e. adaptation approach Our unsupervised supervised domain neither nor","Our approach requires neither supervised nor unsupervised domain adaptation learning, i.e.",20764
available target the no data domains. from,no data available from the target domains.,20765
characterize co-occurrences of to propose actions. We information the human usage,We propose the usage of co-occurrences information to characterize human actions.,20766
"information between local co-occurrence Next, is extract co-occurrence technique to spatio-temporal labeled a applied features.","Next, a spatio-temporal co-occurrence technique is applied to extract co-occurrence information between labeled local features.",20767
for constructed. are Novel then characterizations actions human,Novel characterizations for human actions are then constructed.,20768
Multi-channel classification. is (support for utilized vector kernel machine) SVM,Multi-channel kernel SVM (support vector machine) is utilized for classification.,20769
"and This manual tedious, intensive. is labour checking expensive","This manual checking is tedious, expensive and labour intensive.",20770
"paper, address In challenging novel metric we a propose attribute this this to problem. meaningfulness","In this paper, we propose a novel attribute meaningfulness metric to address this challenging problem.",20771
"user proposed validate a of efficacy study. the conducted we the method, further To","To further validate the efficacy of the proposed method, we conducted a user study.",20772
plan. a this outline we paper research PhD In,In this paper we outline a PhD research plan.,20773
assessment in a problem Summary: is Protein quality bioinformatics. long-standing,Summary: Protein quality assessment is a long-standing problem in bioinformatics.,20774
theory. Experiments on and synthetic our real datasets corroborate both,Experiments on both synthetic and real datasets corroborate our theory.,20775
diversified. the in is optimization domain combinatorial techniques of The number and large,The number of optimization techniques in the combinatorial domain is large and diversified.,20776
"optimization still to lack validate a is there algorithms. of real Nevertheless, benchmarks","Nevertheless, there is still a lack of real benchmarks to validate optimization algorithms.",20777
scenarios a the Such as of characterized VRP. optimization multi-objective were,Such scenarios were characterized as a multi-objective optimization of the VRP.,20778
reasonable a a This occurs. loop process until as works convergence,This process works as a loop until a reasonable convergence occurs.,20779
on Poisson statistics. likelihood functions the Thus depend the,Thus the likelihood functions depend on the Poisson statistics.,20780
"end, adaptation a augmentation based simple data this we strategy. To domain use","To this end, we use a data augmentation based simple domain adaptation strategy.",20781
biology. systems problem modeling post-transcriptional Quantitative a in of is regulation challenging process,Quantitative modeling of post-transcriptional regulation process is a challenging problem in systems biology.,20782
model Rigorous required identify methods to parameters. are,Rigorous methods are required to identify model parameters.,20783
"to like laws. biophysical Humans, fundamental organisms, are all subject","Humans, like all organisms, are subject to fundamental biophysical laws.",20784
lower our at hunter-gatherers mammal show size. a than occurred densities of We that,We show that hunter-gatherers occurred at lower densities than a mammal of our size.,20785
Cones (UOPC) This of the Polyhedral model referred model. Union is to as,This model is referred to as the Union of Polyhedral Cones (UOPC) model.,20786
"periodicity, orientation differ their of ellipticity. and Different properties in neurons that triangular grids feature","Different neurons feature triangular grids that differ in their properties of periodicity, orientation and ellipticity.",20787
"show that i.e. grid are modular, Experiments cells","Experiments show that grid cells are modular, i.e.",20788
"periodicity, which orientation ellipticity. there have groups of similar with and are neurons grids","there are groups of neurons which have grids with similar periodicity, orientation and ellipticity.",20789
without with gaps) by system grid the high probability.,without gaps) by the grid system with high probability.,20790
Both provided. theoretical empirical are and analyses,Both theoretical and empirical analyses are provided.,20791
expensive not all categories visual Collecting is impractical. but images also for training only,Collecting training images for all visual categories is not only expensive but also impractical.,20792
"this solution problem. attributes, learning (ZSL), pragmatic a offers Zero-shot using to especially","Zero-shot learning (ZSL), especially using attributes, offers a pragmatic solution to this problem.",20793
consuming time requires Providing is often these associations knowledge. specific and domain,Providing these associations is time consuming and often requires domain specific knowledge.,20794
"an work, carry In classification this to in unsupervised we manner. attribute-based out zero-shot aim","In this work, we aim to carry out attribute-based zero-shot classification in an unsupervised manner.",20795
We approach that attributes. class an propose embeddings learn corresponding to their with relations couples,We propose an approach to learn relations that couples class embeddings with their corresponding attributes.,20796
"effort. our facilitates model sets additional across data Furthermore, attributes without transferring","Furthermore, our model facilitates transferring attributes across data sets without additional effort.",20797
in from sources multiple knowledge in Integrating additional a significant improvement results performance.,Integrating knowledge from multiple sources results in a significant additional improvement in performance.,20798
two aPascal/aYahoo. public on sets: Animals We and Attributes evaluate data with,We evaluate on two public data sets: Animals with Attributes and aPascal/aYahoo.,20799
"solution In this a a single commodity that uses propose paper, RGB-D we real-time camera.","In this paper, we propose a real-time solution that uses a single commodity RGB-D camera.",20800
contacts. uses energy to regularizers The occlusions novel hand-object address and alignment,The alignment energy uses novel regularizers to address occlusions and hand-object contacts.,20801
experiments hand-object dataset. introduce extensive and datasets new existing conducted a several annotated We on,We conducted extensive experiments on several existing datasets and introduce a new annotated hand-object dataset.,20802
"advantages accuracy, key results qualitative and speed, our robustness. Quantitative method: show of and the","Quantitative and qualitative results show the key advantages of our method: speed, accuracy, and robustness.",20803
Deep paper This Identity-Aware facial convolutional a model Transfer attributes. (DIAT) for network presents of,This paper presents a Deep convolutional network model for Identity-Aware Transfer (DIAT) of facial attributes.,20804
experimental validate of results proposed the the The effectiveness method.,The experimental results validate the effectiveness of the proposed method.,20805
as systems physical be high-dimensional Many rank density interest of can low matrices. represented quantum,Many quantum systems of physical interest can be represented as high-dimensional low rank density matrices.,20806
Several CPF propositional formulae experimentally suggested encodings into of and evaluated. are,Several encodings of CPF into propositional formulae are suggested and experimentally evaluated.,20807
last performance recognition the has improved remarkably Face in decade.,Face recognition performance has improved remarkably in the last decade.,20808
"active the their implications in for and delays Specifically, visuo-oculomotor loop consider inference. we","Specifically, we consider delays in the visuo-oculomotor loop and their implications for active inference.",20809
paper. in evaluated Different data-driven models are this statistical,Different data-driven statistical models are evaluated in this paper.,20810
under The evaluated curve. in of system ROC the was terms area,The system was evaluated in terms of area under the ROC curve.,20811
method generality the of it adapt to The kinds of any to surgeries. allows,The generality of the method allows to adapt it to any kinds of surgeries.,20812
Network decision to this easily comes not The way is comprehensible. a,The way this Network comes to a decision is not easily comprehensible.,20813
hot proven innovative. It is both area useful a learning machine and,It is a hot machine learning area proven both useful and innovative.,20814
the operation The endoscopic phase consists recognizing videos. for of in frames latter each,The latter consists in recognizing the operation phase for each frames of endoscopic videos.,20815
"temporal several report, smoothing methods. this we models compare technical classification In and","In this technical report, we compare several classification models and temporal smoothing methods.",20816
approximates number more of MT than phenomena Our previous models. a closely model,Our model approximates a number of MT phenomena more closely than previous models.,20817
networks underlie stable Gene multiple cell with regulatory dynamics fate-decisions. states characterized by,Gene regulatory networks with dynamics characterized by multiple stable states underlie cell fate-decisions.,20818
"method number, network phenotypes, of lifetimes The identifies long-lived states. the and","The method identifies the number, phenotypes, and lifetimes of long-lived network states.",20819
also we sufficiently smooth kernels provide of Mat\'ern functions For convergence. rates and,For Mat\'ern kernels and sufficiently smooth functions we also provide rates of convergence.,20820
extend random multiple Our variables. results to of functions,Our results extend to functions of multiple random variables.,20821
"results cover i.i.d. embeddings mean both In our based either on case,","In either case, our results cover both mean embeddings based on i.i.d.",20822
"in expansion points. expansions ""reduced as dependent well of set"" samples terms as","samples as well as ""reduced set"" expansions in terms of dependent expansion points.",20823
and accurate Effective is analysis. in data model problem important modern selection an,Effective and accurate model selection is an important problem in modern data analysis.,20824
"the we Watkins-Q SARSA, particular, function MDP. TD with and solve In approximation use to","In particular, we use SARSA, Watkins-Q and TD with function approximation to solve the MDP.",20825
pixel-level accuracy. object-level information segmentation as requires both challenging it is and Semantic,Semantic segmentation is challenging as it requires both object-level information and pixel-level accuracy.,20826
"this Recently, great improvement area. systems in gained FCN-based","Recently, FCN-based systems gained great improvement in this area.",20827
how of these proposed number to show features. have been models A use to,A number of models have been proposed to show how to use these features.,20828
until reverberates the the recovers the The as synchrony system end burst. of,The synchrony recovers as the system reverberates until the end of the burst.,20829
alignment. temporal video a present computing scalable for and We a non-linear robust method,We present a scalable and robust method for computing a non-linear temporal video alignment.,20830
best map. A the self-organizing Kohonen final neuro-fuzzy is chosen result by,A final result is chosen by the best neuro-fuzzy self-organizing Kohonen map.,20831
of constraint paper emerging Intelligence Cohort two metaheuristic proposes for This an handling (CI). approaches,This paper proposes two constraint handling approaches for an emerging metaheuristic of Cohort Intelligence (CI).,20832
several solving test by been approaches have constrained The tested problems.,The approaches have been tested by solving several constrained test problems.,20833
"masses excised their and stomach Gastrointestinal were liver and estimated. tract, were","Gastrointestinal tract, stomach and liver were excised and their masses were estimated.",20834
affected and were duodenum most organs. The the liver,The liver and duodenum were the most affected organs.,20835
"vacuolation, liver infiltrations. of The cellular showed glycogen, hyperemia depletion and","The liver showed depletion of glycogen, vacuolation, hyperemia and cellular infiltrations.",20836
changes and epithelialis the in infiltrations showed propria. lamina villi Duodenal degenerative cellular lamina in,Duodenal villi showed degenerative changes in lamina epithelialis and cellular infiltrations in the lamina propria.,20837
a important vision. in non-trivial human computer Understanding actions but continuous problem is,Understanding continuous human actions is a non-trivial but important problem in computer vision.,20838
"of basic manipulation. extracts elements event our further chains, each method Using parsed primitive","Using event chains, our method further extracts basic primitive elements of each parsed manipulation.",20839
recent datasets various framework. robustness the experiments on extensive the conduct validate to of We,We conduct extensive experiments on various recent datasets to validate the robustness of the framework.,20840
are tool available The www.scenenn.net. dataset and at,The tool and dataset are available at www.scenenn.net.,20841
important Lensless an and challenging problem. is imaging,Lensless imaging is an important and challenging problem.,20842
on imaging based method ultrafast sensing. we compressive lensless present Here a for,Here we present a method for lensless imaging based on compressive ultrafast sensing.,20843
to with lensless regular pixel compared patterns single allows This imaging. imaging fewer significantly,This allows lensless imaging with significantly fewer patterns compared to regular single pixel imaging.,20844
possible show with efficient sensing. compressive ultrafast that measurement is We imaging and lensless,We show that efficient lensless imaging is possible with ultrafast measurement and compressive sensing.,20845
available. appearance our We detection and feature publicly make,We make our detection and appearance feature publicly available.,20846
"First, properties. analyze we theoretical their main","First, we analyze their main theoretical properties.",20847
"Second, we alternatives. a compare performance state-of-the-art against of their variety","Second, we compare their performance against a variety of state-of-the-art alternatives.",20848
substrings visual word-based matching. are refine These to used,These substrings are used to refine visual word-based matching.,20849
"the results. check false system does to Furthermore, our not positive lead with convexity","Furthermore, our system with the convexity check does not lead to false positive results.",20850
"might images contain which partly disconnected, often curvilinear Natural or be structures, occluded.","Natural images contain often curvilinear structures, which might be disconnected, or partly occluded.",20851
se How of in update face de beliefs de we dicto should evidence? the,How should we update de dicto beliefs in the face of de se evidence?,20852
"divides camps, halfers into philosophers thirders. Sleeping The two Beauty problem and","The Sleeping Beauty problem divides philosophers into two camps, halfers and thirders.",20853
"these egregious constitute particular, an violation credences the In of Reflection Principle.","In particular, these credences constitute an egregious violation of the Reflection Principle.",20854
then discuss general. the halfing for I in consequences,I then discuss the consequences for halfing in general.,20855
different instances. optima for computational integration we alternative Here analyze data to present approaches context-specific,Here we present computational approaches to analyze alternative optima for different context-specific data integration instances.,20856
"is as information such prior Additionally, sparsity, usually available. the important data, about","Additionally, important prior information about the data, such as sparsity, is usually available.",20857
leads to Inappropriate thus unreasonable and centers unreasonable scale cluster usually results.,Inappropriate scale usually leads to unreasonable cluster centers and thus unreasonable results.,20858
examples. in computed demonstrate We four gains efficiency these,We demonstrate these efficiency gains in four computed examples.,20859
of are interval-valued Some overlap and operators OWA interval-valued functions derived the properties analysed.,Some properties of interval-valued overlap functions and the derived interval-valued OWA operators are analysed.,20860
specially focus the homogeneity migrativity on and We properties.,We specially focus on the homogeneity and migrativity properties.,20861
"One correlations systematically in with frequently observation increase is made experiments, that firing rate. can","One observation frequently made in experiments, is that correlations can increase systematically with firing rate.",20862
This describes a text paper scenes. dataset from everyday small containing of images,This paper describes a dataset containing small images of text from everyday scenes.,20863
and kernels. Our of enables geometry framework space normalized us identify to of expound,Our framework enables us to expound and identify geometry of space of normalized kernels.,20864
"the this are task. made In three-sides paper, accomplish to efforts","In this paper, three-sides efforts are made to accomplish the task.",20865
"the to bag-of-words with features visual single not First, characterize image. are extracted representation, vector,","First, visual features with bag-of-words representation, not single vector, are extracted to characterize the image.",20866
"information is be complementary three to from The together. can views each other, which unified","The information from three views is complementary to each other, which can be unified together.",20867
alternate method optimization by The solved can iteratively. be,The method can be solved by alternate optimization iteratively.,20868
positive ascent utilized iterations. in semi-definite and the projection Gradient are,Gradient ascent and positive semi-definite projection are utilized in the iterations.,20869
is function new prior previous distance to the functions. bag comparisons that verified Distance,Distance comparisons verified that the new bag distance function is prior to previous functions.,20870
of consistency Experiments on the number influence and of instance parameters the validate method.,Experiments on influence of parameters and instance number validate the consistency of the method.,20871
leading image Deep learning becoming methodology is for analysis. quickly the medical,Deep learning is quickly becoming the leading methodology for medical image analysis.,20872
for part eye detector system is screening. mobile the Messidor of pathology This,This detector is part of the Messidor system for mobile eye pathology screening.,20873
"first In based estimation. estimated on the strategy, is the parameter alternating","In the first strategy, the parameter is estimated based on alternating estimation.",20874
"jointly coefficients. second In sparse strategy, estimated with along it is the the","In the second strategy, it is jointly estimated along with the sparse coefficients.",20875
"a datasets covariates. We high-dimensional number with consider large setting, of biomedical a containing","We consider a high-dimensional setting, with datasets containing a large number of biomedical covariates.",20876
early of the propose We patients then by a event. assessing score adverse risk,We then propose a score by assessing the patients risk of early adverse event.,20877
"method methods the the our experiments, In datasets. two on real-world current outperformed","In the experiments, our method outperformed the current methods on two real-world datasets.",20878
"and are robust part-based occlusion tracking, they attractive In deformation. visual trackers are against since","In visual tracking, part-based trackers are attractive since they are robust against occlusion and deformation.",20879
"due to Nevertheless, power. their lack difficult tracking of discriminative is superpixels","Nevertheless, tracking superpixels is difficult due to their lack of discriminative power.",20880
"configurations. enumerative and trees trees, we results species on For matching obtain ancestral gene","For matching gene trees and species trees, we obtain enumerative results on ancestral configurations.",20881
The a species of the gene provide combinatorial to trees study contribution and results trees.,The results provide a contribution to the combinatorial study of gene trees and species trees.,20882
of on rely algorithms sensors. Most active low-cost currently available the depth,Most of the currently available algorithms rely on low-cost active depth sensors.,20883
"and for devices. mobile a suitable are environments currently result, they outdoor not As","As a result, they are currently not suitable for outdoor environments and mobile devices.",20884
using limitations. This paper these stereo aims at passive which avoids tracking/estimating hand poses,This paper aims at tracking/estimating hand poses using passive stereo which avoids these limitations.,20885
also images. outputs CNN of be for We single how sensitivity can predicted show the,We also show how the sensitivity of CNN outputs can be predicted for single images.,20886
future of suggest the avenues project. in We work,We suggest avenues of future work in the project.,20887
to aimed We short whilst a its version preserving psychometric quality. design,We aimed to design a short version whilst preserving its psychometric quality.,20888
"spectrum. diagnostic IADL between dementia in scores across the we investigated Lastly, differences groups","Lastly, we investigated differences in IADL scores between diagnostic groups across the dementia spectrum.",20889
from to dementia. impairment normal cognition increased IADL scores across spectrum the,IADL impairment scores increased across the spectrum from normal cognition to dementia.,20890
A-IADL-Q. quality original the psychometric maintained The of A-IADL-Q-SV has the,The A-IADL-Q-SV has maintained the psychometric quality of the original A-IADL-Q.,20891
"of it functional As a concise measure is decline. such,","As such, it is a concise measure of functional decline.",20892
images intervals sky regular cameras. time of hemispherical of using capture We ground-based at the,We capture hemispherical images of the sky at regular intervals of time using ground-based cameras.,20893
and images. cloud resolution They provide a localized high,They provide a high resolution and localized cloud images.,20894
can several recognition systems. reduce factors the gait accuracy There confounding of that are,There are several confounding factors that can reduce the accuracy of gait recognition systems.,20895
"a this dataset. With we methodology, multi-modal generated","With this methodology, we generated a multi-modal dataset.",20896
similar. an visually is are different scenes when the classifier challenging Learning but semantically event,Learning an event classifier is challenging when the scenes are semantically different but visually similar.,20897
"However, handle painlessly as such our we adding humans, tasks typically background semantic by knowledge.","However, as humans, we typically handle such tasks painlessly by adding our background semantic knowledge.",20898
comparisons Pairwise measuring between for a are a decision-maker. preferences of alternatives well-known method,Pairwise comparisons between alternatives are a well-known method for measuring preferences of a decision-maker.,20899
countries from are pollution. severe air Many suffering,Many countries are suffering from severe air pollution.,20900
air pollutants making relevant different accumulate policies. critical to public propagate is Understanding how and,Understanding how different air pollutants accumulate and propagate is critical to making relevant public policies.,20901
analytically state-derivatives are they integrated The modelled out are latent as variables. since,The state-derivatives are integrated out analytically since they are modelled as latent variables.,20902
"variables noise. themselves also they latent are because state by However, contaminated are the variables","However, the state variables themselves are also latent variables because they are contaminated by noise.",20903
the tractable. integrating out analytically since them variables work sampled \textit{not} state Previous is,Previous work sampled the state variables since integrating them out is \textit{not} analytically tractable.,20904
"will We, compare a then, real-world on the dataset. models injury","We, then, will compare the models on a real-world injury dataset.",20905
"an (CNN). using convolutional this automated two-stage In neural propose paper, deep networks method we","In this paper, we propose an automated two-stage method using deep convolutional neural networks (CNN).",20906
considerably has this show method good performance We that and benefit readers. can,We show that this method has good performance and can considerably benefit readers.,20907
neural convolutional fully a first candidates. to We initial network use detect,We first use a fully convolutional neural network to detect initial candidates.,20908
"we approaches this for In compare based angle two propose and spectral classification. paper, spatial-spectral","In this paper, we propose and compare two spectral angle based approaches for spatial-spectral classification.",20909
"a this sequential for new data approach modeling develop we regression streams. In paper,","In this paper, we develop a new sequential regression modeling approach for data streams.",20910
pressure information We a for under time with develop decision-making endogenous Bayesian model acquisition.,We develop a Bayesian model for decision-making under time pressure with endogenous information acquisition.,20911
"show that has policy optimal We rendezvous i.e. structure, the a","We show that the optimal policy has a rendezvous structure, i.e.",20912
"i.e. the observing information, in posterior drift suspense, new her after and belief","the drift in her posterior belief after observing new information, and suspense, i.e.",20913
two occurs samples. time event adverse the in probability that interval between the the information,the probability that the adverse event occurs in the time interval between two information samples.,20914
current realization series. time the of the,the current realization of the time series.,20915
"on (affect), motivation, concentration defined arousal and energetic loadings Task task as (cognition). engagement is","Task engagement is defined as loadings on energetic arousal (affect), task motivation, and concentration (cognition).",20916
features and were were extracted. EEG signals power preprocessed spectral The,The EEG signals were preprocessed and power spectral features were extracted.,20917
The tested labeled were data. models then on remaining the,The models were then tested on the remaining labeled data.,20918
"subspace clustering we the address paper, multi-view this In problem.","In this paper, we address the multi-view subspace clustering problem.",20919
major distinguished by is Our contributions. work two,Our work is distinguished by two major contributions.,20920
algorithm to novel We multi-scale triangle meshes. on fields compute curvature a present,We present a novel algorithm to compute multi-scale curvature fields on triangle meshes.,20921
algorithm propose an suitable way. We radii finds that in automatic an,We propose an algorithm that finds suitable radii in an automatic way.,20922
"produced In image-based algorithm our systems. by meshes to is applicable reconstruction particular,","In particular, our algorithm is applicable to meshes produced by image-based reconstruction systems.",20923
of cyanobacteria the rhythm and tied diurnal of are to Growth inherently metabolism availability. light,Growth and metabolism of cyanobacteria are inherently tied to the diurnal rhythm of light availability.,20924
"parameters approach. algorithm implementation, trial original In and error of by selected were the its","In its original implementation, parameters of the algorithm were selected by trial and error approach.",20925
that Using symmetric ACS. fits framework TSP the problems common the demonstrate for we well,Using common symmetric TSP problems we demonstrate that the framework fits well for the ACS.,20926
overall field. of the highlight trends use various techniques in the We to development visualization,We use various visualization techniques to highlight overall trends in the development of the field.,20927
to ambiguities. are equation models exogenous resolve incorporating of Structural inherent directional (SEMs) inputs capable,Structural equation models (SEMs) are capable of incorporating exogenous inputs to resolve inherent directional ambiguities.,20928
invasive is fast choice preferred becoming nature. due treatment the FESS minimally its of to,FESS is fast becoming the preferred choice of treatment due to its minimally invasive nature.,20929
anatomically video-CT multiple algorithm video features. We incorporates registration constrained an that present,We present an anatomically constrained video-CT registration algorithm that incorporates multiple video features.,20930
in presence Our outliers. the of is algorithm robust,Our algorithm is robust in the presence of outliers.,20931
The increasing years have few interest past witnessed feature covariance-based representation. on research,The past few years have witnessed increasing research interest on covariance-based feature representation.,20932
exploiting the demonstrates This value of sparsity covariance-based for feature representation. result well structure,This result well demonstrates the value of exploiting structure sparsity for covariance-based feature representation.,20933
The hand discussion elements the intelligence. the at canonical of idea is about constitute that,The discussion at hand is about the elements that constitute the canonical idea of intelligence.,20934
back rule systems? Are era the to of based harking we expert,Are we harking back to the era of rule based expert systems?,20935
of progress. be real may difficult Even emulating in insects terms,Even emulating insects may be difficult in terms of real progress.,20936
using representations be in product Our accuracy. loss also with compressed can little quantization heavily,Our representations can also be heavily compressed using product quantization with little loss in accuracy.,20937
to submission report Contest. Process This the presents a Discovery,This report presents a submission to the Process Discovery Contest.,20938
ends submitted pictures with the report of models. the process The,The report ends with the pictures of the submitted process models.,20939
algorithm This presents evolutionary diagram-based (VorEAl). paper Voronoi the,This paper presents the Voronoi diagram-based evolutionary algorithm (VorEAl).,20940
in input Voronoi VorEAl using diagrams. space abnormal/normal partitions subsets,VorEAl partitions input space in abnormal/normal subsets using Voronoi diagrams.,20941
validated approaches. VorEAl similar and paper As part of the is experimentally contrasted with,As part of the paper VorEAl is experimentally validated and contrasted with similar approaches.,20942
a characterizing DVF. the We for Jacobian spectral also displacement introduce of measures three,We also introduce three spectral measures of the displacement Jacobian for characterizing a DVF.,20943
analysis-based with Inversion quantitative results show also remarkable agreement predictions.,Inversion results also show remarkable quantitative agreement with analysis-based predictions.,20944
Adaptive highly effective and in control the is necessary of non-small presence NTDCs.,Adaptive control is necessary and highly effective in the presence of non-small NTDCs.,20945
"computer (CNN) image used are Convolutional networks classification. in especially in vision, widely neural","Convolutional neural networks (CNN) are widely used in computer vision, especially in image classification.",20946
is activation modifying func- before tion pooling. behav- step obtained traditional by the This ior,This behav- ior is obtained by modifying the traditional activation func- tion step before pooling.,20947
"unforeseen it Hence, handle observations the phase. efficiently test can during","Hence, it can efficiently handle unforeseen observations during the test phase.",20948
(EER). equal of the terms The method outperforms methods in proposed MSHMM error compared rate,The proposed MSHMM method outperforms the compared methods in terms of equal error rate (EER).,20949
"Additionally, parameters proposed different on of are effects discussed. the method the","Additionally, the effects of different parameters on the proposed method are discussed.",20950
the of (INWFE) extraction nonparametric is weighted feature version NWFE. online Incremental,Incremental nonparametric weighted feature extraction (INWFE) is the online version of NWFE.,20951
of distribution complex in is performs well and class distributed independent It the data.,It is independent of the class distribution and performs well in complex distributed data.,20952
matrix. effects outliers the due are nature The its of scatter to nonparametric reduced of,The effects of outliers are reduced due to the nature of its nonparametric scatter matrix.,20953
"samples it possible Furthermore, asynchronously, i.e. is add to new","Furthermore, it is possible to add new samples asynchronously, i.e.",20954
of execution NWFE are compared terms classification and Results time. in accuracy with,Results are compared with NWFE in terms of classification accuracy and execution time.,20955
is speech Interpretation: increased connectedness randomness The graph of structural in Schizophrenia.,Interpretation: The structural randomness of speech graph connectedness is increased in Schizophrenia.,20956
propose Gaussian (SpInGP) temporal formulation a sparse process parallelizable for models. We inverse,We propose a parallelizable sparse inverse formulation Gaussian process (SpInGP) for temporal models.,20957
function. RKHS the as of We model an intensity root square the,We model the square root of the intensity as an RKHS function.,20958
projection from Usually by a deviations corrected are perspective the perfect undistortion.,Usually the deviations from a perfect perspective projection are corrected by undistortion.,20959
"field-of-view severe artifacts if gets however, the Latter, introduces cameras larger.","Latter, however, introduces severe artifacts if the cameras field-of-view gets larger.",20960
proposed to estimators hypothesis us intervals confidence construct Both exact asymptotically tests. and allow,Both proposed estimators allow us to construct asymptotically exact confidence intervals and hypothesis tests.,20961
agreement in are simulations. numerical almost the The analytic results perfect with,The analytic results are in almost perfect agreement with the numerical simulations.,20962
of the colorectal/intestinal We epithelium). the (at discuss in model cancer our context,We discuss our model in the context of colorectal/intestinal cancer (at the epithelium).,20963
an be (HMD) of augmented head-mounted display could system. important reality A component,A head-mounted display (HMD) could be an important component of augmented reality system.,20964
matter digital media become concern. of safeguarding has of copyright Thus,Thus safeguarding copyright of digital media has become matter of concern.,20965
existing video systems. Finally performance we detection copy of evaluate,Finally we evaluate performance of existing video copy detection systems.,20966
a data exploration. popular tool Factorization (NMF) Non-negative Matrix is for,Non-negative Matrix Factorization (NMF) is a popular tool for data exploration.,20967
also NMF characterize Bayesian promises factorization. uncertainty to in the,Bayesian NMF promises to also characterize uncertainty in the factorization.,20968
on based moment two methods new propose We of the matching. approximation idea conjugacy,We propose two new conjugacy approximation methods based on the idea of moment matching.,20969
prior Both of them formulas. updating closed-form yield Bayesian,Both of them yield closed-form Bayesian prior updating formulas.,20970
norm and estimation procedure thresholding. in periodogram sits variation Our between atomic/total classic methods,Our estimation procedure sits in between classic periodogram methods and atomic/total variation norm thresholding.,20971
"can sample guarantees our Under for conditions, be procedure. derived finite suitable","Under suitable conditions, finite sample guarantees can be derived for our procedure.",20972
results literature. open expand existing spectral questions These estimation in and some resolve,These resolve some open questions and expand existing results in spectral estimation literature.,20973
Encouraging cross dataset. leave-one-out the on training by validation are evaluation obtained results,Encouraging results are obtained by leave-one-out cross validation evaluation on the training dataset.,20974
any prior or on distributions knowledge parameters. does not It require,It does not require any prior knowledge on distributions or parameters.,20975
of performance high-dimensional empirical state-space to the on models. We the low- various compare algorithms,We compare the empirical performance of the various algorithms on low- to high-dimensional state-space models.,20976
our present of we In this laparoscopic technical collected report dataset (LapChole). cholecystectomies,In this technical report we present our collected dataset of laparoscopic cholecystectomies (LapChole).,20977
are used Why function? different perform same to the mechanisms two,Why are two different mechanisms used to perform the same function?,20978
relationship mechanism. for apparent coupling Our this suggests ratio a and between analysis possible explanation,Our analysis suggests a possible explanation for this apparent relationship between coupling ratio and mechanism.,20979
The is real algorithm and on data. experiments simulated with evaluated extensive,The algorithm is evaluated with extensive experiments on simulated and real data.,20980
of multi-scale cues encode effectively connectivity maps target with the The Boolean different granularities.,The Boolean maps effectively encode multi-scale connectivity cues of the target with different granularities.,20981
line therapy degenerative treatment is of with from for first symptoms Physical the spondylolisthesis. adults,Physical therapy is the first line of treatment for adults with symptoms from degenerative spondylolisthesis.,20982
adequately Surgical management when is not have offered symptoms. options relieved nonoperative,Surgical management is offered when nonoperative options have not adequately relieved symptoms.,20983
The each age of extracted a were mean entry. single study used and,The mean age of each study were extracted and used a single entry.,20984
filter our reported double/multiple out data. to We best tried,We tried our best to filter out double/multiple reported data.,20985
surgical have pattern. indicates and different practice Japan Korea study may This South,This study indicates Japan and South Korea may have different surgical practice pattern.,20986
of warrant different cost-effectiveness approaches surgeons. spine professional by analysis may further The,The cost-effectiveness of different approaches may warrant further analysis by professional spine surgeons.,20987
The compatible approaches. with framework various existing is quantization proposed vector,The proposed framework is compatible with various existing vector quantization approaches.,20988
theoretical the that are consistent literature. with We show these also from results strategies,We also show that these strategies are consistent with theoretical results from the literature.,20989
"human loss inconsequential, may accidents Potholes though resulting in life. cause of seem","Potholes though seem inconsequential, may cause accidents resulting in loss of human life.",20990
"selection. was issue Sparse address proposed of GCCA variable to the Further, (SGCCA)","Further, Sparse GCCA (SGCCA) was proposed to address the issue of variable selection.",20991
"streaming use summarization. for video this diverse paper, In we sampling","In this paper, we use diverse sampling for streaming video summarization.",20992
the samples. is The as measured diversity convex a of around volume the hull,The diversity is measured as the volume of a convex hull around the samples.,20993
model. occlusions a detect statistical sudden-appearance-changes Markovian and hidden The proposed can using system,The proposed system can detect sudden-appearance-changes and occlusions using a hidden Markovian statistical model.,20994
sudden-appearance our and system results reliably. experimental detect the confirm that The changes occlusions,The experimental results confirm that our system detect the sudden-appearance changes and occlusions reliably.,20995
CNNs performance detection in in improved photographs. massively have object,CNNs have massively improved performance in object detection in photographs.,20996
in limited. detection remains research into artwork object However,However research into object detection in artwork remains limited.,20997
levels. Prominent miRNA using are tests them expression among,Prominent among them are tests using miRNA expression levels.,20998
at miRNAs Genomics a hundreds of expression of and proteomics provide levels now time.,Genomics and proteomics now provide expression levels of hundreds of miRNAs at a time.,20999
"these in methods be must form, data numeric Because are quantitative developed.","Because these data are in numeric form, quantitative methods must be developed.",21000
forest challenges. new propose addresses detection a random that We action online these (RF)-based framework,We propose a new random forest (RF)-based online action detection framework that addresses these challenges.,21001
joint features. computationally efficient uses algorithm Our skeletal,Our algorithm uses computationally efficient skeletal joint features.,21002
the network results. results detection the only tool show presence improves The slightly that multi-task,The results show that the multi-task network only slightly improves the tool presence detection results.,21003
opportunity This BCI us use gives an practice. in everyday to,This gives us an opportunity to use BCI in everyday practice.,21004
area segmentation currently active research of Motion an computer Vision. is in,Motion segmentation is currently an active area of research in computer Vision.,21005
"as ""Which objects are such Questions moving?","Questions such as ""Which objects are moving?",21006
motivate for sparsity. chain to causal We introduce it use and rule directed information,We introduce chain rule for directed information and use it to motivate causal sparsity.,21007
the resolved on time gene expression method evaluate We data.,We evaluate the method on time resolved gene expression data.,21008
local integration. and also decreased network global find increased We,We also find increased local and decreased global network integration.,21009
broadly with Our brain the are hypothesis. consistent results entropic,Our results are broadly consistent with the entropic brain hypothesis.,21010
"detection GLOSS, outlier algorithm an introduce using subspace performs neighbourhoods. local We that therefore global","We therefore introduce GLOSS, an algorithm that performs local subspace outlier detection using global neighbourhoods.",21011
the which sliding stamps. We has data construct along dictionary time with their training a,We construct a sliding dictionary which has the training data along with their time stamps.,21012
"the benchmark several on and methodology illustrate we Finally, datasets.","Finally, we illustrate and benchmark the methodology on several datasets.",21013
available online numerical for Code are and experiments reproducible research.,Code and numerical experiments are available online for reproducible research.,21014
"activity. example application brain an real measurements we magnetoencephalographic on Finally, of report","Finally, we report an example application on real magnetoencephalographic measurements of brain activity.",21015
and been as biometric considered Gait has person unique identification. for a promising,Gait has been considered as a promising and unique biometric for person identification.,21016
"image correction and details. for techniques introduced in distortion analyzed are paper, this hardware-compatible In","In this paper, hardware-compatible techniques for image distortion correction are introduced and analyzed in details.",21017
estimated also for resources of approach. The required amount is each considered hardware,The required amount of hardware resources is also estimated for each considered approach.,21018
for useful parameters. may of estimation structured spatially temporally penalties sparse or Such model-based be,Such model-based penalties may be useful for sparse estimation of spatially or temporally structured parameters.,21019
of parcel computer and many are measures vision part algorithms. Distance,Distance measures are part and parcel of many computer vision algorithms.,21020
distance information. the of bin-to-bin bin-to-bin a for which PBR dispersion accounts is,PBR is a bin-to-bin distance which accounts for the dispersion of bin-to-bin information.,21021
"the these relation the of of single neurons to is However, properties unclear. models","However, the relation of these models to the properties of single neurons is unclear.",21022
trees the set a taxa set. A overlapping Supertree of carrying phylogenetic synthesizes topologies of,A Supertree synthesizes the topologies of a set of phylogenetic trees carrying overlapping taxa set.,21023
problem NP-hard. a Such to is be proved,Such a problem is proved to be NP-hard.,21024
"parsimony, on search, local cut, maximum heuristics etc. Various graph","Various heuristics on local search, maximum parsimony, graph cut, etc.",21025
"and space on the size complexities however, of methods, these subtree Time considered. depend","Time and space complexities of these methods, however, depend on the subtree size considered.",21026
interaction. is enabling Facial analysis key a for technology human-machine,Facial analysis is a key technology for enabling human-machine interaction.,21027
"etc. or a is have bald, is mustache, the does person male female, he she/he","is the person male or female, is she/he bald, does he have a mustache, etc.",21028
success. have achieved (CNN) great networks Deep neural convolutional,Deep convolutional neural networks (CNN) have achieved great success.,21029
"On other structural in proved information critical many hand, been modeling problems. vision has the","On the other hand, modeling structural information has been proved critical in many vision problems.",21030
interest It great is integrate to effectively. of them,It is of great interest to integrate them effectively.,21031
"neural Finally, provided. of a implementation end-to-end network CRF-CNN learning is","Finally, a neural network implementation of end-to-end learning CRF-CNN is provided.",21032
on two demonstrated datasets. experiments Its through benchmark effectiveness is,Its effectiveness is demonstrated through experiments on two benchmark datasets.,21033
detection the (SOD) has object been decades. salient Image-based studied past extensively in,Image-based salient object detection (SOD) has been extensively studied in the past decades.,21034
matching this of addressed The work. in is problem mesh,The problem of mesh matching is addressed in this work.,21035
is mesh foundation matching method provided. of Theoretical,Theoretical foundation of mesh matching method is provided.,21036
This dissociation one attractions and induces water R-groups. hydrophobic increases between of the the,This one induces the dissociation of water and increases the hydrophobic attractions between R-groups.,21037
H-bonds. allows the surrounding to recreate with Coupling clusters water,Coupling with the surrounding water clusters allows to recreate H-bonds.,21038
the in present involve should H-bonds clusters. the decrease This number a in of,This should involve a decrease in the number of H-bonds present in the clusters.,21039
We embeddings. a loss suggest learning for deep,We suggest a loss for learning deep embeddings.,21040
suitable optimization. makes embeddings the learning loss stochastic using This deep proposed for,This makes the proposed loss suitable for learning deep embeddings using stochastic optimization.,21041
"performs recently In proposed the the favourably loss new compared to alternatives. experiments,","In the experiments, the new loss performs favourably compared to recently proposed alternatives.",21042
theoretically another thus network copying network to is We propose one connectivity that from possible.,We thus propose that copying network connectivity from one network to another is theoretically possible.,21043
Data camera. collected a chest-mounted GoPro was using,Data was collected using a chest-mounted GoPro camera.,21044
"for We convolutional feature of classification and neural a extraction, network developed frames. automatic reduction,","We developed a convolutional neural network for automatic feature extraction, reduction, and classification of frames.",21045
on It motion. variations brightness spatio-temporal due to relies,It relies on spatio-temporal brightness variations due to motion.,21046
with common Our approach issues a within adaptive proposed deals framework. these regularized,Our proposed adaptive regularized approach deals with these issues within a common framework.,21047
"be types error. sources can representing model, our various of different handled, noise of In","In our model, various types of noise can be handled, representing different sources of error.",21048
the and the fourth functional importance determines relative norm makes of the parameter This convex.,This parameter determines the relative importance of the fourth norm and makes the functional convex.,21049
indicate flow. of estimates Experiments the optical approach this that provides robust,Experiments indicate that this approach provides robust estimates of the optical flow.,21050
(LV) we ventricle framework left segmentation propose Here for a cardiac cine-MRI. from,Here we propose a framework for left ventricle (LV) segmentation from cardiac cine-MRI.,21051
the the book notion of introduce for this time neutrosophic authors first In strong graphs.,In this book authors for the first time introduce the notion of strong neutrosophic graphs.,21052
They and neutrosophic the graphs from graphs. are very different usual,They are very different from the usual graphs and neutrosophic graphs.,21053
Using topological subgraph defined. spaces are special new these structures,Using these new structures special subgraph topological spaces are defined.,21054
are defined subgraphs described. of special Further of lattice these graphs and graph,Further special lattice graph of subgraphs of these graphs are defined and described.,21055
properties Several neutrosophic a obtained. subgraphs strong using interesting of are graph,Several interesting properties using subgraphs of a strong neutrosophic graph are obtained.,21056
(HSIC). Maps (SMs) Sensitivity independence the introduce criterion for the We Hilbert-Schmidt,We introduce the Sensitivity Maps (SMs) for the Hilbert-Schmidt independence criterion (HSIC).,21057
also the and map both sensitivity of Convergence measure the bounds provided. are,Convergence bounds of both the measure and the sensitivity map are also provided.,21058
factors. model several required successful critical The,The successful model required several critical factors.,21059
"architectures, was studies, layer and one These only however, relatively trainable. used shallow","These studies, however, used relatively shallow architectures, and only one layer was trainable.",21060
"thus designed SNN, We pooling STDP) and comprising with (trainable deep layers. several convolutional a","We thus designed a deep SNN, comprising several convolutional (trainable with STDP) and pooling layers.",21061
network was to The natural exposed images.,The network was exposed to natural images.,21062
per a examples label Only needed. category few was of tens were required no and,Only a few tens of examples per category were required and no label was needed.,21063
is selection established well likelihood model Bayesian The in criterion a marginal statistics.,The marginal likelihood is a well established model selection criterion in Bayesian statistics.,21064
"approaches, however For likelihoods latent modeling complex marginal many compute. including are to difficult models,","For many complex models, including latent modeling approaches, marginal likelihoods are however difficult to compute.",21065
the presence context address microRNA bimodal regulation. in We phenotypes of theoretically the (miRNA)-mediated of,We theoretically address the presence of bimodal phenotypes in the context of microRNA (miRNA)-mediated regulation.,21066
the are of expression downregulate small noncoding target that mRNAs. their MiRNAs molecules RNA,MiRNAs are small noncoding RNA molecules that downregulate the expression of their target mRNAs.,21067
We half-life. the distributions dependence protein on protein characterise furthermore,We furthermore characterise the protein distributions dependence on protein half-life.,21068
model. described topic behaviours dynamic a by activities The and are,The activities and behaviours are described by a dynamic topic model.,21069
of parameters posterior Theoretical derivations of given. model the are,Theoretical derivations of the posterior of model parameters are given.,21070
on the video of detailed A comparison presented algorithms data. is real learning,A detailed comparison of the learning algorithms is presented on real video data.,21071
"We propose the framework. anomaly localisation procedure, elegantly modeling an topic in also embedded","We also propose an anomaly localisation procedure, elegantly embedded in the topic modeling framework.",21072
common most (MS) neurological the disabling is the young disorders adults. Multiple in sclerosis,Multiple sclerosis (MS) is the most common disabling neurological disorders in the young adults.,21073
"spite of the mainly many cause of not remained the understood. researches, has disease In","In spite of many researches, the cause of the disease has remained mainly not understood.",21074
evidences etiology. disease factors indicate key All risk roles in that this play environmental,All evidences indicate that environmental risk factors play key roles in this disease etiology.,21075
model for introduce semi-parametric We analysis. a Bayesian survival,We introduce a semi-parametric Bayesian model for survival analysis.,21076
"in Furthermore, survival mechanisms our model handles common left, interval censoring analysis. right and","Furthermore, our model handles left, right and interval censoring mechanisms common in survival analysis.",21077
by learn optical a deep classical combining We formulation learning. spatial-pyramid compute flow to with,We learn to compute optical flow by combining a classical spatial-pyramid formulation with deep learning.,21078
applications. for This appropriate makes embedded it and efficient more,This makes it more efficient and appropriate for embedded applications.,21079
The needs larger to and further varied progress. computer make datasets academic community vision more,The academic computer vision community needs larger and more varied datasets to make further progress.,21080
human for curated We provide bounding faces. boxes,We provide human curated bounding boxes for faces.,21081
contexts. different Here on explore introduce and its we in it own possible applications,Here we introduce it on its own and explore possible applications in different contexts.,21082
"web-based a quantitative the for Summary: of prediction interactions. introduce tool RBPBind, RNA-protein We","Summary: We introduce RBPBind, a web-based tool for the quantitative prediction of RNA-protein interactions.",21083
on RBPBind the effect fully secondary interactions. of protein-RNA RNA thus incorporates structure,RBPBind thus fully incorporates the effect of RNA secondary structure on protein-RNA interactions.,21084
generally There is to unified intelligent building significant a approaches of lack machines.,There is a significant lack of unified approaches to building generally intelligent machines.,21085
the that flow. experiments optical Preliminary this robust provides indicate of estimates approach,Preliminary experiments indicate that this approach provides robust estimates of the optical flow.,21086
show we that some methods (e.g. While,While we show that some methods (e.g.,21087
end future work. with for questions some We,We end with some questions for future work.,21088
NMU-based robustly to parametric dataset. We fit multiple to then an models a present algorithm,We then present an NMU-based algorithm to robustly fit multiple parametric models to a dataset.,21089
for multiway decomposition high-order Tensor capturing interactions is important technique data. among an the,Tensor decomposition is an important technique for capturing the high-order interactions among multiway data.,21090
methods known. Another these rank assumption predefined of should that is a be,Another assumption of these methods is that a predefined rank should be known.,21091
"the with for to hard values. rank especially of estimate, However, missing tensors is cases","However, the rank of tensors is hard to estimate, especially for cases with missing values.",21092
"issues, decomposition. we for generative tensor design model these a address Bayesian To","To address these issues, we design a Bayesian generative model for tensor decomposition.",21093
estimates require algorithms covariance learning Many precise machine matrices. of,Many machine learning algorithms require precise estimates of covariance matrices.,21094
for recognition from are sensor generated action mature. Techniques data,Techniques for action recognition from sensor generated data are mature.,21095
"has activities. actions and However, been little there between work the gap relatively on bridging","However, there has been relatively little work on bridging the gap between actions and activities.",21096
demonstrate on a dataset experiments our the of Extensive real-world work. effectiveness,Extensive experiments on a real-world dataset demonstrate the effectiveness of our work.,21097
algorithm. also any minimum error possible we Further address for the learning,Further we also address the minimum possible error for any learning algorithm.,21098
"on applications. typically break for Algorithms profile real-time faces, down slow or are too","Algorithms typically break down on profile faces, or are too slow for real-time applications.",21099
branching regression that propose of generalization call We a cascaded CSRs (BCR). we conventional,We propose a generalization of conventional CSRs that we call branching cascaded regression (BCR).,21100
"introduce our in-the-wild faces approach. profile of to validate a new Additionally, dataset we","Additionally, we introduce a new in-the-wild dataset of profile faces to validate our approach.",21101
the representation. the of such methods The is success key feature to,The key to the success of such methods is the feature representation.,21102
"the that effectiveness thoroughly. impact features factors However, are still explored not the different of","However, the different factors that impact the effectiveness of features are still not explored thoroughly.",21103
of much best discussion about There less are combination the them.,There are much less discussion about the best combination of them.,21104
"other used multiple On trackers object in tracking. hand, not the are typically visual","On the other hand, visual trackers are not typically used in multiple object tracking.",21105
based multiple a on system background tracking and KCF We object subtraction. build,We build a multiple object tracking system based on KCF and background subtraction.,21106
KCF most the trustworthy outputs are (e.g. Sometimes,Sometimes KCF outputs are the most trustworthy (e.g.,21107
"outputs. occlusion), it while subtraction background other some case, the in during is","during occlusion), while in some other case, it is the background subtraction outputs.",21108
"and morphologic, syntactic, For model, are type semantic characterized. errors of each","For each type of model, morphologic, syntactic, and semantic errors are characterized.",21109
recently shallow of been behave like has that ensembles relatively Networks shown It networks. Residual,It has recently been shown that Residual Networks behave like ensembles of relatively shallow networks.,21110
the it. explain behavior behind demonstrate driving force this We and,We explain this behavior and demonstrate the driving force behind it.,21111
is damage with of Aging accumulation the life. associated a persons throughout,Aging is associated with the accumulation of damage throughout a persons life.,21112
Index the Frailty be health assessed Individual by can (FI).,Individual health can be assessed by the Frailty Index (FI).,21113
the The with age. exponent significantly rates mortality of affects growth $\alpha$ network,The network exponent $\alpha$ significantly affects the growth of mortality rates with age.,21114
"find spectrum, information the that i.e. also We","We also find that the information spectrum, i.e.",21115
by features DCNN face two produced We algorithms. the recognition analyzed,We analyzed the DCNN features produced by two face recognition algorithms.,21116
function increased quality from monotonically Image of distance as the origin. a,Image quality increased monotonically as a function of distance from the origin.,21117
"of reason, For (e.g. in object countries been it's this many regulation","For this reason, it's been object of regulation in many countries (e.g.",21118
applications Such signals imagery and intelligence. intelligence include,Such applications include signals intelligence and imagery intelligence.,21119
factors neural key convolutional the success the (CNNs). is of one networks behind of Depth,Depth is one of the key factors behind the success of convolutional neural networks (CNNs).,21120
"in (BP) the backpropagation In we issue existing supervision first this investigate paper, methods. vanishing","In this paper, we first investigate the supervision vanishing issue in existing backpropagation (BP) methods.",21121
model. variable a of representation This the latent uses,This uses a latent variable representation of the model.,21122
can In contrast Carlo viewed as more standard using ABC Monte estimate. a be naive,In contrast standard ABC can be viewed as using a more naive Monte Carlo estimate.,21123
"user-independent, faster the are but and automated is methods registration critical reliability. Fully point","Fully automated methods are faster and user-independent, but the critical point is registration reliability.",21124
Datalog. We evaluation the PTime query into consider also connection between rewritability and (monadic),We also consider the connection between PTime query evaluation and rewritability into (monadic) Datalog.,21125
use How problem. video-based for open effectively an is ConvNets recognition still to,How to effectively use ConvNets for video-based recognition is still an open problem.,21126
"targeting be can is used it approach scalable streaming and and over quick, data. The","The targeting approach is scalable and quick, and it can be used over streaming data.",21127
on cases. and for multilingual-lingual not suitable approaches texts Previous focus mainly are monolingual,Previous approaches mainly focus on monolingual texts and are not suitable for multilingual-lingual cases.,21128
methods. the show results previous that experimental outperforms significantly our model The,The experimental results show that our model significantly outperforms the previous methods.,21129
"abnormalities ultrasound. can visualized size often be in Interestingly, using kidney CKD-related","Interestingly, CKD-related abnormalities in kidney size can often be visualized using ultrasound.",21130
similar measured calculated area. parenchymal renal is the remarkably The to area,The calculated renal parenchymal area is remarkably similar to the measured area.,21131
descriptors of manifolds) formed from generated these local by sets images. the,manifolds) formed by the sets of local descriptors generated from these images.,21132
applications of learning. and in range Game theory finds machine broad nowadays engineering a,Game theory finds nowadays a broad range of applications in engineering and machine learning.,21133
"based novel propose solving this Gaussian-process a games we for approach context. in Here,","Here, we propose a novel Gaussian-process based approach for solving games in this context.",21134
"framework, decisions optimization acquisition sampling based with sequential follow classical We Bayesian on functions. a","We follow a classical Bayesian optimization framework, with sequential sampling decisions based on acquisition functions.",21135
genetic sweeps through Selective hitchhiking. affect neutral diversity,Selective sweeps affect neutral genetic diversity through hitchhiking.,21136
Convolutional for human recognition in have Networks (ConvNets) videos. performance strong shown action Two-stream,Two-stream Convolutional Networks (ConvNets) have shown strong performance for human action recognition in videos.,21137
"to as deep architectures. arisen new extremely a Residual train Networks Recently, (ResNets) have technique","Recently, Residual Networks (ResNets) have arisen as a new technique to train extremely deep architectures.",21138
"we introduce a two these ResNets combination as this approaches. paper, of spatiotemporal In","In this paper, we introduce spatiotemporal ResNets as a combination of these two approaches.",21139
of is The allow features. learning spatiotemporal hierarchical model to trained complex end-to-end whole,The whole model is trained end-to-end to allow hierarchical learning of complex spatiotemporal features.,21140
attention last have neural the Convolutional received increasing years. few (CNNs) networks over,Convolutional neural networks (CNNs) have received increasing attention over the last few years.,21141
"categorization recently proposed. have semantic been networks, many By adapting labeling CNNs","By adapting categorization networks, many semantic labeling CNNs have been recently proposed.",21142
is an contribution of in-depth architectures. Our analysis first these,Our first contribution is an in-depth analysis of these architectures.,21143
"combine addition features features. In learning these learns how different it at to to resolutions,","In addition to learning features at different resolutions, it learns how to combine these features.",21144
of enables Speech vocal intentions emotional communication prosody of modulation via intonations.,Speech prosody enables communication of emotional intentions via modulation of vocal intonations.,21145
in performance recognition Networks recognition tasks. Convolutional achieved visual (ConvNets) Neural various have excellent,Convolutional Neural Networks (ConvNets) have achieved excellent recognition performance in various visual recognition tasks.,21146
important success. training large set labeled A one of is its the most for factors,A large labeled training set is one of the most important factors for its success.,21147
"labels, these ambiguous tasks Fortunately, which there makes traditional information is among different from classification.","Fortunately, there is ambiguous information among labels, which makes these tasks different from traditional classification.",21148
two-fold. paper is in The made contribution this,The contribution made in this paper is two-fold.,21149
"microwave optimization. formulated of in GR-based is convex the imaging Afterwards, sparsity-regularized the","Afterwards, the GR-based microwave imaging is formulated in the convex of sparsity-regularized optimization.",21150
results Selected the to proposed performance methodology. state-of-art are simulation provided of demonstrate,Selected simulation results are provided to demonstrate the state-of-art performance of proposed methodology.,21151
present method of for freedom image localization. accurate robust We an and six degree,We present an accurate and robust method for six degree of freedom image localization.,21152
segmentation The method in image medical applications. used numerous widely is,The method is widely used in numerous medical image segmentation applications.,21153
"accuracy than the voxel, unit single Thus i.e. segmentation cannot an greater attain a","Thus the segmentation cannot attain an accuracy greater than a single unit voxel, i.e.",21154
the in nodes two adjoining between graph distance space.,the distance between two adjoining nodes in graph space.,21155
"all highly results. yielded the cases, approach accurate In","In all cases, the approach yielded highly accurate results.",21156
be extended to readily higher-dimensional Our can approach segmentations.,Our approach can be readily extended to higher-dimensional segmentations.,21157
"literature, of data multilevel recommended In pan been data. MI for has missing the","In the missing data literature, pan has been recommended for MI of multilevel data.",21158
Our for allows unbounded nevertheless multi-sets over encoding with multiplicities. framework preferences,Our framework nevertheless allows for encoding preferences over multi-sets with unbounded multiplicities.,21159
disorders. performed of between elucidate Mendelian casual the (MR) to randomization two relationships We groups,We performed Mendelian randomization (MR) to elucidate casual relationships between the two groups of disorders.,21160
genome-wide based meta-analyses The studies on of (GWAS). analysis was association large-scale,The analysis was based on large-scale meta-analyses of genome-wide association studies (GWAS).,21161
"However, showed SCZ. genetic polygenic with link inverse BMI correlation and","However, BMI showed inverse genetic correlation and polygenic link with SCZ.",21162
"other more to On the hand, are abnormalities secondary. be BD in likely CM","On the other hand, CM abnormalities in BD are more likely to be secondary.",21163
"time While ortho-rectification consuming. to expensive tends done, be it be and can","While ortho-rectification can be done, it tends to be expensive and time consuming.",21164
"more offers approach. low-tech Image economical, stitching a","Image stitching offers a more economical, low-tech approach.",21165
rural for Implications various projects. in and industries of terms urban both explained Irish are,Implications for various Irish industries are explained in terms of both urban and rural projects.,21166
parallelizability. Fractal algorithms extreme generation image exhibit,Fractal image generation algorithms exhibit extreme parallelizability.,21167
is A the sequential made between of comparison and algorithm. implementations parallel,A comparison is made between sequential and parallel implementations of the algorithm.,21168
between is of a still definition missing. accepted widely causal formal However observables influence,However a widely accepted formal definition of causal influence between observables is still missing.,21169
We under estimators normal show proposed the and conditions. mild are consistent asymptotically,We show the proposed estimators are consistent and asymptotically normal under mild conditions.,21170
samples need do source-domain available. not to be,source-domain samples do not need to be available.,21171
experiments estimates the robust this approach optical indicate that flow. provides Preliminary of,Preliminary experiments indicate that this approach provides robust estimates of the optical flow.,21172
biometrics and tasks. monitoring for is requirement important an Animal conservation,Animal biometrics is an important requirement for monitoring and conservation tasks.,21173
The techniques manage the offer alternatives biometrics to aforementioned problems. non-invasive,The non-invasive biometrics techniques offer alternatives to manage the aforementioned problems.,21174
millepunctatus species. lizard Diploglossus endangered an is,Diploglossus millepunctatus is an endangered lizard species.,21175
through how deals our experiments. numerical proposed complex careful demonstrate variational deformations solution with We,We demonstrate how our proposed variational solution deals with complex deformations through careful numerical experiments.,21176
"highly rather The format easy and new made parse. compact, readable, to is","The new format is made compact, highly readable, and rather easy to parse.",21177
queries precise The sophisticated for instances selecting can user very from criteria. make,The user can make sophisticated queries for selecting instances from very precise criteria.,21178
"pedestrian around-the-clock applications, surveillance driving. and essential detection autonomous e.g., Multispectral for is","Multispectral pedestrian detection is essential for around-the-clock applications, e.g., surveillance and autonomous driving.",21179
"In and classification. Autoencoders particular, apply image Convolutional we for Networks x-ray Neural scattering Convolutional","In particular, we apply Convolutional Neural Networks and Convolutional Autoencoders for x-ray scattering image classification.",21180
Synchronization represent communication to distant brain fluctuations of neural activity between believed is processes. spatially,Synchronization of brain activity fluctuations is believed to represent communication between spatially distant neural processes.,21181
two study Genant's and current ABQ SQ method method these in cohorts. compared The,The current study compared Genant's SQ method and ABQ method in these two cohorts.,21182
mutation Duplication is and our strategy. of phrases,Duplication and mutation of phrases is our strategy.,21183
strategy is seen of phenomena. wide classes living in a Such,Such a strategy is seen in wide classes of living phenomena.,21184
Experiments restore transformations proposed and wonderful can architecture robustness with the planar effectiveness. show,Experiments show the architecture proposed can restore planar transformations with wonderful robustness and effectiveness.,21185
in demographic applications require Several information of people scenarios. unconstrained ordinary,Several applications require demographic information of ordinary people in unconstrained scenarios.,21186
due a is significant This not appearance trivial human to task variations.,This is not a trivial task due to significant human appearance variations.,21187
clothes lead time. approach a face to segmentation with detection real to Combining close us,Combining with face detection lead us to a clothes segmentation approach close to real time.,21188
study dataset VOC the Pascal experiments. evaluation challenging segmentation uses The for,The study uses the challenging Pascal VOC dataset for segmentation evaluation experiments.,21189
"stability biological key reproduction of is survival cells. in and a Genetic factor maintaining,","Genetic stability is a key factor in maintaining, survival and reproduction of biological cells.",21190
analyze approach a We developed theoretical to homology the search. quantitative,We developed a quantitative theoretical approach to analyze the homology search.,21191
quality is missing Segmenting of a low challenging problem. with or data images,Segmenting images of low quality or with missing data is a challenging problem.,21192
present variety promising a We on of data results sets.,We present promising results on a variety of data sets.,21193
Unsupervised hugely generative with successful. adversarial (GANs) has learning networks proven,Unsupervised learning with generative adversarial networks (GANs) has proven hugely successful.,21194
function. sigmoid cross entropy classifier with discriminator the a as Regular loss hypothesize the GANs,Regular GANs hypothesize the discriminator as a classifier with the sigmoid cross entropy loss function.,21195
GANs. LSGANs benefits of two over There regular are,There are two benefits of LSGANs over regular GANs.,21196
"are generate images than quality to First, GANs. regular able LSGANs higher","First, LSGANs are able to generate higher quality images than regular GANs.",21197
"learning the LSGANs during perform stable more process. Second,","Second, LSGANs perform more stable during the learning process.",21198
"trial-to-trial model most common case, variability. In that a the Poisson is of process","In that case, a Poisson process is the most common model of trial-to-trial variability.",21199
hold. generally not does this Numerous relationship shown studies that have,Numerous studies have shown that this relationship does not generally hold.,21200
quantify particularly model is is A to suited over-dispersion that well distribution model. Negative-Binomial the,A model that is particularly well suited to quantify over-dispersion is the Negative-Binomial distribution model.,21201
is but been model This only applied and used widely neuroscience. recently to well-studied has,This model is well-studied and widely used but has only recently been applied to neuroscience.,21202
"main address paper, In this we issues. three","In this paper, we address three main issues.",21203
tests along We test pathway. visual the neurophysiological apply three to this,We apply this test to three neurophysiological tests along the visual pathway.,21204
"investigated effect the input of methods. preprocessing we also image Moreover,","Moreover, we also investigated the effect of input image preprocessing methods.",21205
sparse representation of reduction images considered. data context X-Ray within Effective the is of medical,Effective sparse representation of X-Ray medical images within the context of data reduction is considered.,21206
"scales. image the Specifically, different satellite warp multiple firstly into original we","Specifically, we firstly warp the original satellite image into multiple different scales.",21207
employed are images convolutional neural each (DCNN). network train scale The to in a deep,The images in each scale are employed to train a deep convolutional neural network (DCNN).,21208
"However, time-consuming. multiple training simultaneously is DCNNs","However, simultaneously training multiple DCNNs is time-consuming.",21209
"To spatial pooling we pyramid with address (SPP-net). this issue, explore DCNN","To address this issue, we explore DCNN with spatial pyramid pooling (SPP-net).",21210
on of data huge platform Categorisation a crucial task. the multimedia of is amount,Categorisation of huge amount of data on the multimedia platform is a crucial task.,21211
by captured This the Analysis (CCA). was using Correlation aforementioned features synergy projecting Canonical,This synergy was captured by projecting the aforementioned features using Canonical Correlation Analysis (CCA).,21212
results Experimental the the effectiveness shows dataset proposed on this approach. of,Experimental results on this dataset shows the effectiveness of the proposed approach.,21213
"detection present In applications. a for framework fast robotics real-world work, target we this","In this work, we present a fast target detection framework for real-world robotics applications.",21214
dimensionality structure low-dimensional in frequently reduction used identify to data. is high-dimensional Spectral,Spectral dimensionality reduction is frequently used to identify low-dimensional structure in high-dimensional data.,21215
"computationally from the expensive. streaming is learning However, especially manifolds, and data, memory","However, learning manifolds, especially from the streaming data, is computationally and memory expensive.",21216
stable manifold the Identifying is the point which key step. the is at transition,Identifying the transition point at which the manifold is stable is the key step.,21217
potential existence of nonnegative the for and epidemic an outbreak) endemic the solutions.,the potential for an epidemic outbreak) and the existence of nonnegative endemic solutions.,21218
"the Finally, health policy we and briefly public for discuss implications our planning. results of","Finally, we briefly discuss the implications of our results for public health policy and planning.",21219
the outlets. deal It great discussions generated in media Internet a some of and,It generated a great deal of discussions in the Internet and some media outlets.,21220
in attention (CNNs) attracted Convolutional sensing community. the have neural networks remote increasing,Convolutional neural networks (CNNs) have attracted increasing attention in the remote sensing community.,21221
"of Indisputable, clothing most e-commerce popular the is business. categories one","Indisputable, one of the most popular e-commerce categories is clothing business.",21222
considered is of reaction-diffusion of thermodynamics non-linear non-equilibrium Dynamic microtubules instability using systems. and frameworks,Dynamic instability of microtubules is considered using frameworks of non-linear thermodynamics and non-equilibrium reaction-diffusion systems.,21223
developed is kinetic Additional limitations in to reaction-diffusion-precipitation dynamics. model account microtubule for,Additional reaction-diffusion-precipitation model is developed to account for kinetic limitations in microtubule dynamics.,21224
by autowaves. scale shown that is microtubules concentration fluctuations accompanied are large periodic It length,It is shown that large scale periodic microtubules length fluctuations are accompanied by concentration autowaves.,21225
"diagram built stationary, and mapping parametric corresponding We metastable of solutions. areas non-stationary","We built corresponding parametric diagram mapping areas of stationary, non-stationary and metastable solutions.",21226
the The happens Andronov-Hopf. loss stationary through for of solutions bifurcation of stability,The loss of stability for the stationary solutions happens through bifurcation of Andronov-Hopf.,21227
results State-of-the are obtained these datasets. art on,State-of-the art results are obtained on these datasets.,21228
"the influences In the trial previous trace both current the decision. cases, of","In both cases, the trace of the previous trial influences the current decision.",21229
by of rotation powered filaments flagellar helical motors. rotary mobility by Bacterial is driven,Bacterial mobility is powered by rotation of helical flagellar filaments driven by rotary motors.,21230
data peak flagellar aligned bundles structure-factor predicated from line-shape. the confirmed Scattering scattering,Scattering data from aligned flagellar bundles confirmed the predicated structure-factor scattering peak line-shape.,21231
in a of are gesture implementation system. hand challenges recognition There,There are challenges in implementation of a hand gesture recognition system.,21232
high and as has requirements It accuracy to response. such fulfill fast,It has to fulfill requirements such as high accuracy and fast response.,21233
"a method a be proposed system. result, embedded implemented on naive the As could","As a result, the proposed method could be implemented on a naive embedded system.",21234
"correlation (DCF) discriminatively object attention community. learned tracking in has much Recently, drawn filters visual","Recently, discriminatively learned correlation filters (DCF) has drawn much attention in visual object tracking community.",21235
"Convolutional framework visual (CRT). tracking paper, a for In propose this Regression we","In this paper, we propose a Convolutional Regression framework for visual tracking (CRT).",21236
explicitly matching between degree captures model users. and questions The expertise,The model explicitly captures expertise matching degree between questions and users.,21237
"and model. proposed to QA-Expert datasets: of genres Paper-Reviewer, use the validate different two We","We use two different genres of datasets: QA-Expert and Paper-Reviewer, to validate the proposed model.",21238
"the review aim constraint-based we In this to causal paper, discovery algorithms.","In this paper, we aim to review the constraint-based causal discovery algorithms.",21239
"approaches. discuss constraint-based the Firstly, we learning of paradigm the","Firstly, we discuss the learning paradigm of the constraint-based approaches.",21240
"and surveyed with Secondly are algorithms analysis. inference constraint-based state-of-the-art primarily, casual detailed the the","Secondly and primarily, the state-of-the-art constraint-based casual inference algorithms are surveyed with the detailed analysis.",21241
"related software are packages Thirdly, repositories several briefly benchmark summarized. data and open-source","Thirdly, several related open-source software packages and benchmark data repositories are briefly summarized.",21242
"problems outlined a open causal are constraint-based As some future for discovery conclusion, in research.","As a conclusion, some open problems in constraint-based causal discovery are outlined for future research.",21243
convolutional for fully a color approach We image network(FCN) present restoration. based,We present a fully convolutional network(FCN) based approach for color image restoration.,21244
remarkable segmentation. like semantic shown have high-level for problem performance recently vision FCNs,FCNs have recently shown remarkable performance for high-level vision problem like semantic segmentation.,21245
by returned extract is the metric information. spatial to image The used query,The image returned by the spatial query is used to extract metric information.,21246
"resolution, e.g. useful linguistic In that for coreference this are features task, traditional","In this task, traditional linguistic features that are useful for coreference resolution, e.g.",21247
"effective no and information, anymore. context gender are longer","context and gender information, are no longer effective anymore.",21248
"diverse viewpoints. camera to may show and large object backgrounds One variations due class illuminations,","One object class may show large variations due to diverse illuminations, backgrounds and camera viewpoints.",21249
unconstrained worse Traditional video environments. methods often detection under perform object,Traditional object detection methods often perform worse under unconstrained video environments.,21250
"problem, appearance many To approaches representations modern address for deep this detection. hierarchical object model","To address this problem, many modern approaches model deep hierarchical appearance representations for object detection.",21251
set. training require Most methods sample process on large manual of these a labelling timeconsuming,Most of these methods require a timeconsuming training process on large manual labelling sample set.,21252
not detector are data needed. training labeled generic the a Here human or,Here the human labeled training data or a generic detector are not needed.,21253
"applications the multi-scene employed Thus, with in object proposed be can minimal detection framework supervision.","Thus, the proposed framework can be employed in multi-scene object detection applications with minimal supervision.",21254
Networks Convolutional Neural as features. emerged have learning powerful discriminative for (CNN) image tools,Convolutional Neural Networks (CNN) have emerged as powerful tools for learning discriminative image features.,21255
identification on MRI. structures The proposed of allows framework high-level tumor,The proposed framework allows identification of high-level tumor structures on MRI.,21256
proposed comparing demonstrate of state-of-the-art performance Extensive approach to encouraging experimental results the methods. the,Extensive experimental results demonstrate encouraging performance of the proposed approach comparing to the state-of-the-art methods.,21257
"image single segmentation. work However, the improving focused of most on previous","However, most of the previous work focused on improving single image segmentation.",21258
both online proposed in method and segmentation. can Our applied batch be,Our proposed method can be applied in both online and batch segmentation.,21259
binary for segmentation tasks. tested architecture This semantic video is and both,This architecture is tested for both binary and semantic video segmentation tasks.,21260
baseline network recurrent improved of performance networks all our fully Using the convolutional in experiments.,Using recurrent fully convolutional networks improved the baseline network performance in all of our experiments.,21261
The depends network. fully of baseline its RFCN network on convolutional performance the,The performance of the RFCN network depends on its baseline fully convolutional network.,21262
Action to anticipation computer applications. key many success recognition of vision and are the,Action recognition and anticipation are key to the success of many computer vision applications.,21263
Optimal will called method Regression resulting The Scaling Regression. be (Regularized ROS,The resulting method will be called ROS Regression (Regularized Optimal Scaling Regression.,21264
either mentioned. regression regularization of or category for coefficients Alternative quantifications are options,Alternative options for regularization of either regression coefficients or category quantifications are mentioned.,21265
indispensable are networks neural Deep algorithms. vision computer state-of-the-art to convolutional (CNNs),Deep convolutional neural networks (CNNs) are indispensable to state-of-the-art computer vision algorithms.,21266
"back-propagation. the After are layers fine-tuned globally is pruned, further network using all entire","After all layers are pruned, the entire network is further globally fine-tuned using back-propagation.",21267
"click. users platforms a with content or ""like"" allow to favor certain online Nowadays, single","Nowadays, online platforms allow users to ""like"" or favor certain content with a single click.",21268
aesthetics validate our model study. user derived a in We of,We validate our derived model of aesthetics in a user study.,21269
a understanding pressing is biology. of biophysics The in molecular channel and issue functions,The understanding of channel functions is a pressing issue in molecular biophysics and biology.,21270
of One its our and focuses generalizations. (PNP) model Poisson-Nernst-Planck the is,One of our focuses is the Poisson-Nernst-Planck (PNP) model and its generalizations.,21271
"For model Poisson-Boltzmann-Kohn-Sham channel, multiscale proton (PBKS) presented. and is multiphysics a","For proton channel, a multiscale and multiphysics Poisson-Boltzmann-Kohn-Sham (PBKS) model is presented.",21272
designed was to forward a less computations than propagation. require it,it was designed to require less computations than a forward propagation.,21273
in theoretical findings experimental agreement the stand with results. Our,Our theoretical findings stand in agreement with the experimental results.,21274
similar results magnitude visualization achieves obtains speed-ups. it order simultaneously and of,it obtains similar visualization results and simultaneously achieves order of magnitude speed-ups.,21275
costly a and annotation constraint the makes in scalability. process limited Such,Such a constraint makes the annotation process costly and limited in scalability.,21276
motion SfM. (CMSs) we rolling-shutter (RS) critical In sequences consider of paper this,In this paper we consider critical motion sequences (CMSs) of rolling-shutter (RS) SfM.,21277
"formulation, the representation of In CMSs. general derive we a","In the formulation, we derive a general representation of CMSs.",21278
state-of-the-art pixel-wise network into can architecture This be integrated that produces easily technique predictions. any,This technique can be easily integrated into any state-of-the-art network architecture that produces pixel-wise predictions.,21279
"strategy. deconvolution a We artifacts using grid-like also during ""shift-and-average"" avoid","We also avoid grid-like artifacts during deconvolution using a ""shift-and-average"" strategy.",21280
player's paper presents from first-person a to method video. a his/her This assess basketball performance,This paper presents a method to assess a basketball player's performance from his/her first-person video.,21281
first-person to camera We the address leverage challenge. this,We leverage the first-person camera to address this challenge.,21282
in (SBs) neuronal common are complex with structures cultures. bursts Synchronized,Synchronized bursts (SBs) with complex structures are common in neuronal cultures.,21283
"on in multi-electrode a of we these system. culture properties investigate SBs Here, array the","Here, we investigate the properties of these SBs in a culture on multi-electrode array system.",21284
usually setting supervised in involving trained pixel-level of They are fully annotations a objects.,They are usually trained in a fully supervised setting involving pixel-level annotations of objects.,21285
bottom-up with selected is The top-down saliency combined the map saliency map.,The selected bottom-up saliency map is combined with the top-down saliency map.,21286
"typical vision low-level Previous this problem a i.e. as problem, methods cast","Previous methods cast this problem as a typical low-level vision problem, i.e.",21287
"oriented feature offer a this Based framework. blind insight, novel face on inpainting we","Based on this insight, we offer a novel feature oriented blind face inpainting framework.",21288
"this a consists which Specifically, three implement novel of we by establishing parts. DeMeshNet,","Specifically, we implement this by establishing a novel DeMeshNet, which consists of three parts.",21289
an end-to-end the All implemented network optimization. that parts facilitates efficient within three are,All the three parts are implemented within an end-to-end network that facilitates efficient optimization.,21290
"framework directly. allows deeper the Furthermore, training end-to-end of networks","Furthermore, the framework allows end-to-end training of deeper networks directly.",21291
"two-stepped strategy a transfer knowledge from Second, developed is auxiliary fine-tuning to datasets.","Second, a two-stepped fine-tuning strategy is developed to transfer knowledge from auxiliary datasets.",21292
benchmarks show model existing outperforms models. Extensive that the on our four experiments significantly,Extensive experiments on four benchmarks show that our model significantly outperforms the existing models.,21293
to applications. of moving in track vehicle is ability importance crucial a The numerous,The ability to track a moving vehicle is of crucial importance in numerous applications.,21294
filters informative. Particle highly poorly perform observations when are,Particle filters perform poorly when observations are highly informative.,21295
are this presented. and phenomena resolution its Theoretical results explaining,Theoretical results explaining this phenomena and its resolution are presented.,21296
"all frames temporally independently, are results stable. coherent the are Although processed and","Although all frames are processed independently, the results are temporally coherent and stable.",21297
"(JNA) this To in is these study. a address \textbf{a}ttention \textbf{j}oint proposed based \textbf{n}etwork problems,","To address these problems, a \textbf{j}oint \textbf{n}etwork based \textbf{a}ttention (JNA) is proposed in this study.",21298
"We highly a present network classification. image modularized architecture for simple,","We present a simple, highly modularized network architecture for image classification.",21299
The online. and available models are code publicly,The code and models are publicly available online.,21300
machine tasks computer Clustering vision among and is most in learning. the fundamental,Clustering is among the most fundamental tasks in computer vision and machine learning.,21301
"present stitch In algorithm register we an fields. multiple and paper, this light to","In this paper, we present an algorithm to register and stitch multiple light fields.",21302
datasets importance. of is changes crucial FC the independent therefore of Investigating on PD-related reproducibility,Investigating the reproducibility of PD-related FC changes on independent datasets is therefore of crucial importance.,21303
"to datasets. be turned changes functional non-reproducible out Unfortunately, across connectivity PD-related","Unfortunately, PD-related functional connectivity changes turned out to be non-reproducible across datasets.",21304
"heterogeneity, could to to but due also disease be technical This differences.","This could be due to disease heterogeneity, but also to technical differences.",21305
it improvements can consistent that obtain show towards Our experiments the tasks. transferring,Our experiments show that it can obtain consistent improvements towards the transferring tasks.,21306
bounding refine initial facial output first face The landmarks. boxes two and stages intermediate,The first two stages refine initial face bounding boxes and output intermediate facial landmarks.,21307
demonstrate DAC-CSR Experimental datasets results the of obtained merits on challenging our over state-of-the-art. the,Experimental results obtained on challenging datasets demonstrate the merits of our DAC-CSR over the state-of-the-art.,21308
image networks neural enabled accurate real-time. in have Convolutional super-resolution,Convolutional neural networks have enabled accurate image super-resolution in real-time.,21309
"obvious. proper is criteria a CDCL in off-the-shelf However, such integration of implementations not","However, a proper integration of such criteria in off-the-shelf CDCL implementations is not obvious.",21310
"ASP domains. industrial experimented in on we implemented an and We solver, two strategy our","We implemented our strategy in an ASP solver, and we experimented on two industrial domains.",21311
crucial of step Unsupervised quality evaluation segmentation image is in segmentation applications. a,Unsupervised evaluation of segmentation quality is a crucial step in image segmentation applications.,21312
usually multi-scale lacked Previous methods unsupervised the to segmentation. adaptability evaluation,Previous unsupervised evaluation methods usually lacked the adaptability to multi-scale segmentation.,21313
"into region. them are equivalent distances of standardized predefined a Subsequently, of spectral both","Subsequently, both of them are standardized into equivalent spectral distances of a predefined region.",21314
global It to has similarity. so image been limited far,It has been so far limited to global image similarity.,21315
regions. allows in optional number reduction efficient stage of An the off-line stored,An efficient off-line stage allows optional reduction in the number of stored regions.,21316
a case of retrieval. been failure have Small objects common CNN-based,Small objects have been a common failure case of CNN-based retrieval.,21317
the pathway. is select then employed information mechanism best Attention flow to,Attention mechanism is then employed to select the best information flow pathway.,21318
imaging. a cost-effective light Micro-lens cameras based light provide field to field array approach (MLA),Micro-lens array (MLA) based light field cameras provide a cost-effective approach to light field imaging.,21319
"high-resolution propose Deep Feature new data-driven baseline transformation. (DFI), a We automatic image for Interpolation","We propose Deep Feature Interpolation (DFI), a new data-driven baseline for automatic high-resolution image transformation.",21320
"advocate we paper, In this models. the use of stratified probabilistic for logical theories representing","In this paper, we advocate the use of stratified logical theories for representing probabilistic models.",21321
"panoramas. the To of dataset outdoor a CNN, train large we leverage","To train the CNN, we leverage a large dataset of outdoor panoramas.",21322
is consistently It state-of-the-art visual methods. SCA-CNN that significantly outperforms attention-based image captioning observed,It is consistently observed that SCA-CNN significantly outperforms state-of-the-art visual attention-based image captioning methods.,21323
stages. consists end-to-end multiple of system Our,Our end-to-end system consists of multiple stages.,21324
for finding a missing in method We solution. the optimal features alternate Lasso propose,We propose a method for finding alternate features missing in the Lasso optimal solution.,21325
"this Lasso. by the relevant can selected not However, features possibly overlook","However, this can overlook possibly relevant features not selected by the Lasso.",21326
that alternate can features computed redundant avoiding by We be computations. show efficiently such,We show that such alternate features can be computed efficiently by avoiding redundant computations.,21327
(SC-DCNNs). design framework of comprehensive DCNNs presents SC-based optimization first the This and paper,This paper presents the first comprehensive design and optimization framework of SC-based DCNNs (SC-DCNNs).,21328
theoretical both methods. and proposed empirical results to present justify our We,We present both theoretical and empirical results to justify our proposed methods.,21329
of efficacy proposed BBN. on our benchmark and SVB datasets Experiments image classification the show,Experiments on benchmark image classification datasets show the efficacy of our proposed SVB and BBN.,21330
promise results Our also preliminary show ImageNet the large-scale learning. in on,Our preliminary results on ImageNet also show the promise in large-scale learning.,21331
the more to potential for the rapid discuss model. We approximations Wright-Fisher,We discuss the potential for more rapid approximations to the Wright-Fisher model.,21332
vision. computer in task is recognition action important Human an,Human action recognition is an important task in computer vision.,21333
use disease-related of of is groups proteins. over variants these several checked potential The,The potential use of these variants is checked over several groups of disease-related proteins.,21334
Our similar to more point are that than main to IDPs. IDRPs NDPs out conclusions,Our main conclusions point out that IDRPs are more similar to NDPs than to IDPs.,21335
present IDPs proteins. are IDRPs differently human and among disease-related,IDRPs and IDPs are differently present among human disease-related proteins.,21336
"role disease. can liver in emergence cancer, play thyroid the of IDPs a neurodegenerative, and","IDPs can play a role in the emergence of cancer, neurodegenerative, thyroid and liver disease.",21337
real-world and apartments. stores all collected from images We at different settings,We collected all images from real-world settings at different stores and apartments.,21338
"of contain Overall, object different images thousands our instances.","Overall, our images contain thousands of different object instances.",21339
"large artifact-free evaluated, datasets images. high-quality and on of trained, These commonly networks are annotated","These networks are commonly trained, and evaluated, on large annotated datasets of artifact-free high-quality images.",21340
promising further results and is investigation. worth adopted the that obtained The are indicate path,The obtained results are promising and indicate that the adopted path is worth further investigation.,21341
(e.g. Porting to state compute resource constrained of the deep algorithms art platforms learning,Porting state of the art deep learning algorithms to resource constrained compute platforms (e.g.,21342
combinations. involves a set linear of LCNN jointly learning Training and dictionary a small,Training LCNN involves jointly learning a dictionary and a small set of linear combinations.,21343
dictionary traces naturally accuracy. the of size between The and trade-offs a spectrum of efficiency,The size of the dictionary naturally traces a spectrum of trade-offs between efficiency and accuracy.,21344
"also but it speed ups only not training. efficient offers enables inference, LCNN dramatic at","LCNN not only offers dramatic speed ups at inference, but it also enables efficient training.",21345
We descriptors. zones propose efficient an the algorithm define compute and fuzzy the to,We define the fuzzy zones and propose an efficient algorithm to compute the descriptors.,21346
are an wherein compared. follow the taxonomy advantages and We limitations each approach-based of approach,We follow an approach-based taxonomy wherein the advantages and limitations of each approach are compared.,21347
survey also sensors. maritime The discusses approaches use systems commercial EO and that multisensor,The survey also discusses multisensor approaches and commercial maritime systems that use EO sensors.,21348
among traditional true both asymmetric is armies symmetric conflicts ones. latter as The well as,The latter is true both traditional symmetric conflicts among armies as well as asymmetric ones.,21349
by This features. building is image invariant usually achieved,This is usually achieved by building invariant image features.,21350
"invariance rotation achieving of large angles the largely for However, remains unexplored. problem","However, the problem of achieving invariance for large rotation angles remains largely unexplored.",21351
"Neural we detect train image a Convolutional to rotations. First, (CNN) Network","First, we train a Convolutional Neural Network (CNN) to detect image rotations.",21352
find suitable are this for architectures CNN generic purpose. that We not,We find that generic CNN architectures are not suitable for this purpose.,21353
helps algorithm the for of Such systems. automatize commercial maps elaboration to navigation,Such algorithm helps to automatize the elaboration of maps for commercial navigation systems.,21354
image treat attribute pedestrian classification multi-label problem. methods as recognition State-of-the-art a,State-of-the-art methods treat pedestrian attribute recognition as a multi-label image classification problem.,21355
"in a localization In weakly-supervised this we paper, the framework. attribute formulate task","In this paper, we formulate the task in a weakly-supervised attribute localization framework.",21356
"the magnitudes. Secondly, detection by of are response predicted attribute labels regression","Secondly, attribute labels are predicted by regression of the detection response magnitudes.",21357
"two datasets, i.e. attribute pedestrian are currently the Extensive largest experiments on performed","Extensive experiments are performed on the two currently largest pedestrian attribute datasets, i.e.",21358
dataset. and dataset RAP PETA the the,the PETA dataset and the RAP dataset.,21359
outperforms large Our margin. a state-of-the-art by methods method,Our method outperforms state-of-the-art methods by a large margin.,21360
present deep a We dense learning-based approach matching. for stereo new,We present a new deep learning-based approach for dense stereo matching.,21361
"accuracy. resulting The efficient method test highly at good while matching time, is achieving","The resulting method is highly efficient at test time, while achieving good matching accuracy.",21362
nature a man-made in is and designs. theme common Symmetry in prevalent,Symmetry is prevalent in nature and a common theme in man-made designs.,21363
"Detecting useful therefore, applications. number of mirror images a in is, for data symmetry and","Detecting mirror symmetry in images and data is, therefore, useful for a number of applications.",21364
resulting of exactness the depends registration accuracy. The solution the entirely on,The exactness of the resulting solution depends entirely on the registration accuracy.,21365
Tracking to on performance are compared well-known datasets. video benchmark state-of-the-art algorithms,Tracking performance are compared to state-of-the-art algorithms on well-known benchmark video datasets.,21366
"directions. study, along both observed In difficulties however, our we","In our study, however, we observed difficulties along both directions.",21367
"Moreover, sequences. tested can datasets and on trained containing be varying-length it","Moreover, it can be trained and tested on datasets containing varying-length sequences.",21368
"in code a.k.a. witnessed learning, attention years have Recent extensive binary","Recent years have witnessed extensive attention in binary code learning, a.k.a.",21369
"visual results benchmark three i.e. datasets, large-scale search on Experimental","Experimental results on three large-scale visual search benchmark datasets, i.e.",21370
should we tends fundamental to all forgotten. at hypothesis of test The whether be question,The fundamental question of whether we should test hypothesis at all tends to be forgotten.,21371
"can opinion description, single Several together a be bias. our in an biases of grouped","Several of our biases can be grouped together in a single description, an opinion bias.",21372
"or despite no has Also, believing philosophical feelings, grounds. our solid logical","Also, despite our feelings, believing has no solid logical or philosophical grounds.",21373
when hypothesis. Testing should are we have place no about thinking,Testing should have no place when we are thinking about hypothesis.,21374
observation. tensor randomly corrupted from paper recovering multi-dimensional focus incomplete called This on data,This paper focus on recovering multi-dimensional data called tensor from randomly corrupted incomplete observation.,21375
"type-specific fit on unbiased an an can By be obtained. appropriate conditioning prior,","By conditioning on an appropriate prior, an unbiased type-specific fit can be obtained.",21376
Finding vision local key visual to problems. computer between is many features correspondence,Finding visual correspondence between local features is key to many computer vision problems.,21377
"tasks. network scale-attention this propose to optimize trade-off We explicitly AutoScaler, correspondence visual a in","We propose AutoScaler, a scale-attention network to explicitly optimize this trade-off in visual correspondence tasks.",21378
for tasks. The network end-to-end visual is siamese in a framework trained correspondence entire,The entire network is trained end-to-end in a siamese framework for visual correspondence tasks.,21379
"generate maps. attention interpretable our Finally, can network scale visually attention","Finally, our attention network can generate visually interpretable scale attention maps.",21380
with consistent consistently all this. approximations treated are Models procedures and are as,Models are consistently treated as approximations and all procedures are consistent with this.,21381
the They true. treat being do as not model,They do not treat the model as being true.,21382
distinguished regions are defined confidence regions. from Approximation and,Approximation regions are defined and distinguished from confidence regions.,21383
strategy convolutional We two-step train (CNN). to a bootstrap multi-label a employ network neural,We employ a two-step bootstrap strategy to train a multi-label convolutional neural network (CNN).,21384
for regularizer discriminator. a acts as standard task This of the training supervised,This task acts as a regularizer for standard supervised training of the discriminator.,21385
NetC layers classify those fully to connected utilises object within category. the features,NetC utilises those features within fully connected layers to classify the object category.,21386
"this we In to overcome different limitation. this two ways work, present","In this work, we present two different ways to overcome this limitation.",21387
"of DSAC, We this the differentiable counterpart RANSAC. call approach","We call this approach DSAC, the differentiable counterpart of RANSAC.",21388
"component. learning any use can as pipeline In a deep future, the robust optimization DSAC","In the future, any deep learning pipeline can use DSAC as a robust optimization component.",21389
to limitations two due models respective The different functions. and have loss advantages their,The two models have their respective advantages and limitations due to different loss functions.,21390
"a new loss propose identification network and computes Specifically, that siamese we simultaneously loss. verification","Specifically, we propose a new siamese network that simultaneously computes identification loss and verification loss.",21391
"the re-ID simple, learned performance state-of-the-art benchmarks. improves on two the public Albeit person embedding","Albeit simple, the learned embedding improves the state-of-the-art performance on two public person re-ID benchmarks.",21392
"be architecture Further, can show image also applied we retrieval. in our","Further, we show our architecture can also be applied in image retrieval.",21393
sentence matching global depends visual-semantic Effective on image their well similarity. and how measure to,Effective image and sentence matching depends on how to well measure their global visual-semantic similarity.,21394
"also the maintain organism connectivity grow. To dendrites to continue grows, must as","To maintain connectivity as the organism grows, dendrites must also continue to grow.",21395
approach. to discriminative corresponding The compared a proposed is generative method,The proposed generative method is compared to a corresponding discriminative approach.,21396
databases. Experiments of are carried out a queries different range three in using,Experiments are carried out using a range of queries in three different databases.,21397
first to (e.g. is component of The the function objects localize,The function of the first component is to localize objects (e.g.,21398
houses image carrying pixel-level a out by an classification. in or infrastructure),houses or infrastructure) in an image by carrying out a pixel-level classification.,21399
a recognition few visual Learning on Grassmann quite is tasks. in manifolds popular representations,Learning representations on Grassmann manifolds is popular in quite a few visual recognition tasks.,21400
statistical hypothesis We tests. introduce two novel non-parametric,We introduce two novel non-parametric statistical hypothesis tests.,21401
Dependence measured Criterion (HSIC). is Independence via Hilbert-Schmidt the,Dependence is measured via the Hilbert-Schmidt Independence Criterion (HSIC).,21402
"are and and consistent favorable have convergence unbiased, resulting tests properties. The","The resulting tests are consistent and unbiased, and have favorable convergence properties.",21403
"is it. no study to tried that Surprisingly, work there prior","Surprisingly, there is no prior work that tried to study it.",21404
"the accuracy accuracy, Precision existing metric we Average (mAP). mean extend For","For accuracy, we extend the existing accuracy metric mean Average Precision (mAP).",21405
stability. error of Each represents aspect one,Each error represents one aspect of stability.,21406
"accuracy low correlation the that with we stability has demonstrate metric. Furthermore, metric","Furthermore, we demonstrate that the stability metric has low correlation with accuracy metric.",21407
"indeed it different Thus, of captures perspective a quality.","Thus, it indeed captures a different perspective of quality.",21408
library knowledge long valuable  are collections and and learning. resources Physical standing  for,  Physical library collections are valuable and long standing resources for knowledge and learning.,21409
respiratory and multiple respiration (RR) applications. has Non-contact rate estimation pattern (RP) of,Non-contact estimation of respiratory pattern (RP) and respiration rate (RR) has multiple applications.,21410
in unlimited words metadata the of unique number an almost have tags.,have an almost unlimited number of unique words in the metadata tags.,21411
"sets. data validation, test divided into training, We and our","We divided our data into training, validation, and test sets.",21412
each We epoch. augmented data in shuffled the counter overfitting to and training,We shuffled and augmented the training data to counter overfitting in each epoch.,21413
which chromatic humans can varies The differences with throughout small detect accuracy space. color,The accuracy with which humans can detect small chromatic differences varies throughout color space.,21414
photoreceptor the level. at encoding marginally Subsequent the only stages chromatic discriminability modify,Subsequent encoding stages modify only marginally the chromatic discriminability at the photoreceptor level.,21415
problem. as CNN-based tracking treat a classification trackers Most existing,Most existing CNN-based trackers treat tracking as a classification problem.,21416
"to it use distractors. information object we problem, address this from To distinguish of self-structure","To address this problem, we use self-structure information of object to distinguish it from distractors.",21417
"Furthermore, convex a relaxations. to leads of discretization the approach it to continuous systematic","Furthermore, it leads to a systematic approach to the discretization of continuous convex relaxations.",21418
and MRFs. study existing We to discrete-continuous relationship to the discretizations,We study the relationship to existing discretizations and to discrete-continuous MRFs.,21419
in capabilities our semantic dense experiments the also of captioning. reveal model Qualitative,Qualitative experiments also reveal the semantic capabilities of our model in dense captioning.,21420
"the data, estimated. index is a on thermal Based (TI)","Based on the data, a thermal index (TI) is estimated.",21421
rating useful be comparisons. and The can TI material for,The TI can be useful for material rating and comparisons.,21422
"this ADDT detailed the package. we chapter, to introduction provide a In","In this chapter, we provide a detailed introduction to the ADDT package.",21423
provide the functions a illustration package. the We step-by-step use for in of,We provide a step-by-step illustration for the use of functions in the package.,21424
illustrations. datasets Publicly used available for are,Publicly available datasets are used for illustrations.,21425
"different deep MldrNet combines i.e. representations proposed The levels, of","The proposed MldrNet combines deep representations of different levels, i.e.",21426
in improve components framework unified recognition an significantly the two accuracy. These,These two components in an unified framework significantly improve the recognition accuracy.,21427
a region approach the that (e.g. The rule enforces proposed,The proposed approach enforces the rule that a region (e.g.,21428
the this index. measure call Jaccard We subset-matched,We call this measure the subset-matched Jaccard index.,21429
environments. often infrared) in (near Surveillance images NIR today low-light cameras capture,Surveillance cameras today often capture NIR (near infrared) images in low-light environments.,21430
"spectrum. to VIS only the available matched be are in galleries Second, mostly face","Second, face galleries to be matched are mostly available only in the VIS spectrum.",21431
deep NIR Cross-spectral a VIS produces images learning approach. faces through hallucination from,Cross-spectral hallucination produces VIS faces from NIR images through a deep learning approach.,21432
"reduction. are propose ResFeats feature PCA for use to large vectors, the Since we dimensionality","Since the ResFeats are large feature vectors, we propose to use PCA for dimensionality reduction.",21433
instance an proposed Conclusion: effective is for The segmentation. algorithm multichannel deep gland method,Conclusion: The proposed deep multichannel algorithm is an effective method for gland instance segmentation.,21434
to the vision solve have information that Computer task. often helpful available side tasks is,Computer vision tasks often have side information available that is helpful to solve the task.,21435
an emerging from an walking pattern his/her of individual identification field. is,identification of an individual from his/her walking pattern is an emerging field.,21436
to obtained are the together set. well-endowed fused final The features form feature,The obtained features are fused together to form the final well-endowed feature set.,21437
available framework i.e. of the publicly three Experimental proposed datasets evaluation on,Experimental evaluation of the proposed framework on three publicly available datasets i.e.,21438
to investigate networks translation conditional solution general-purpose problems. as image-to-image adversarial We a,We investigate conditional adversarial networks as a general-purpose solution to image-to-image translation problems.,21439
structure time is a from discover widespread for can that series There data. techniques need,There is a widespread need for techniques that can discover structure from time series data.,21440
networks cascaded problem. convolutional cope class with neural the We propose to therefore imbalanced,We therefore propose cascaded convolutional neural networks to cope with the class imbalanced problem.,21441
"neural nodule convolutional a set trained with calculates a balanced network Successively, data probabilities.","Successively, a convolutional neural network trained with a balanced data set calculates nodule probabilities.",21442
"introduce pooling, we context Further, rich manner. residual efficient captures chained an in which background","Further, we introduce chained residual pooling, which captures rich background context in an efficient manner.",21443
experiments datasets. public out comprehensive We results seven set on new state-of-the-art and carry,We carry out comprehensive experiments and set new state-of-the-art results on seven public datasets.,21444
"further robustness and changes. study our the towards of appearance, viewpoint, volumetric model We","We further study the robustness of our model towards viewpoint, appearance, and volumetric changes.",21445
segmentation Our image semantically segments. and proposes visually coherent,Our segmentation proposes visually and semantically coherent image segments.,21446
were All of the of to few limited papers the previous number (e.g. category images,All the previous papers were limited to few number of category of the images (e.g.,21447
"empirical ""impact analysis an First, give the problem. we about vanishing""","First, we give an empirical analysis about the ""impact vanishing"" problem.",21448
a Understanding certain science why many crucial made model data is in fields. a prediction,Understanding why a model made a certain prediction is crucial in many data science fields.,21449
provide model into engender improved. how trust Interpretable and the be appropriate predictions insight may,Interpretable predictions engender appropriate trust and provide insight into how the model may be improved.,21450
Natural environment. follow (visual) physical the our by images of inherited statistics structure,Natural images follow statistics inherited by the structure of our physical (visual) environment.,21451
"Over machine years, staple ensemble of the a become have learning. methods","Over the years, ensemble methods have become a staple of machine learning.",21452
"ensembles as possibility. of proposed have a Recently, been GLMs","Recently, ensembles of GLMs have been proposed as a possibility.",21453
"GLMs the downside, approach On this possess. that loses interpretability the","On the downside, this approach loses the interpretability that GLMs possess.",21454
"required. is Therefore, no generate a question to odd-one-out manual annotation","Therefore, to generate a odd-one-out question no manual annotation is required.",21455
"multi-stream convolutional as implemented neural which machine Our network, is is end-to-end. learning learned","Our learning machine is implemented as multi-stream convolutional neural network, which is learned end-to-end.",21456
"flexibility, ease ease is use extension. and designed QUESO portability, of of for","QUESO is designed for flexibility, portability, ease of use and ease of extension.",21457
uniprocessor environments. multiprocessor run It over can and,It can run over uniprocessor and multiprocessor environments.,21458
"encodes feature video representation, It a this aggregated learning. end-to-end robust via information into","It encodes this aggregated information into a robust video feature representation, via end-to-end learning.",21459
current state-of-the-art that The both experiments TLE show datasets. on methods outperforms,The experiments show that TLE outperforms current state-of-the-art methods on both datasets.,21460
is for modeling Log-linear a of analysis data. contingency table method the popular,Log-linear modeling is a popular method for the analysis of contingency table data.,21461
effects. $F$ mitigating to is crucial these the determining Knowledge of cells,Knowledge of the cells determining $F$ is crucial to mitigating these effects.,21462
Our like uncertainty-based random classical sampling entropy algorithms outperform methods the and active learning,Our methods outperform random sampling and the classical uncertainty-based active learning algorithms like entropy,21463
implicitly of and handles truncation persons. proposed occlusion The approach,The proposed approach implicitly handles occlusion and truncation of persons.,21464
"on new dataset. the our methods approach evaluate we several Finally, baseline proposed and","Finally, we evaluate the proposed approach and several baseline methods on our new dataset.",21465
"the of VLAD. representations bag-of-visual-words, vector kind the the this Popular include and Fisher","Popular representations of this kind include the bag-of-visual-words, the Fisher vector and the VLAD.",21466
"i.e. with kernels, In has match one interference, deal to","In match kernels, one has to deal with interference, i.e.",21467
per-descriptor weights. the These methods including of stage aggregation a by modify set,These methods modify the aggregation stage by including a set of per-descriptor weights.,21468
They to the is objective weights. those that optimised function compute differ by,They differ by the objective function that is optimised to compute those weights.,21469
second match vector. of to aggregated equalising the single one a descriptor involves the The,The second one involves equalising the match of a single descriptor to the aggregated vector.,21470
difficult nonlinearity by solve existing The methods. is of to inherent and problem,The problem is of inherent nonlinearity and difficult to solve by existing methods.,21471
structure is an fields study space. with data non-Euclidean a underlying that Many scientific,Many scientific fields study data with an underlying structure that is a non-Euclidean space.,21472
impressive face achieved hallucination Deep for tasks. performance have models,Deep models have achieved impressive performance for face hallucination tasks.,21473
extensively YTF our method and LFW evaluate datasets. We on,We extensively evaluate our method on LFW and YTF datasets.,21474
instance-aware task. solution semantic for first the fully convolutional present segmentation end-to-end We,We present the first fully convolutional end-to-end solution for instance-aware semantic segmentation task.,21475
of proposal. inherits the all It semantic instance for merits segmentation FCNs and mask,It inherits all the merits of FCNs for semantic segmentation and instance mask proposal.,21476
classification jointly. instance and It performs mask prediction,It performs instance mask prediction and classification jointly.,21477
in and proposed accuracy both and efficiency. network state-of-the-art integrated achieves highly The performance is,The proposed network is highly integrated and achieves state-of-the-art performance in both accuracy and efficiency.,21478
tasks. the the performance on We standard evaluate recognition,We evaluate the performance on the standard recognition tasks.,21479
achieved neutral success image convolutional tasks. Deep great recognition have networks on,Deep convolutional neutral networks have achieved great success on image recognition tasks.,21480
"framework present and for video We feature a flow, deep accurate fast recognition.","We present deep feature flow, a fast and accurate framework for video recognition.",21481
is flow as It significant relatively achieves speedup fast. computation,It achieves significant speedup as flow computation is relatively fast.,21482
The the accuracy. of training recognition end-to-end significantly the boosts architecture whole,The end-to-end training of the whole architecture significantly boosts the recognition accuracy.,21483
general. feature Deep is and flexible flow,Deep feature flow is flexible and general.,21484
on two scale It datasets. large video recent validated is,It is validated on two recent large scale video datasets.,21485
large makes video towards step a recognition. It practical,It makes a large step towards practical video recognition.,21486
is vessel a semi-automatic method adapted for method The devised segmentation. from,The method is adapted from a semi-automatic method devised for vessel segmentation.,21487
fashion. train We network the end-to-end entire in,We train the entire network in end-to-end fashion.,21488
an We method systematic literature mapping to mapping. use survey our established conduct,We use an established systematic mapping survey method to conduct our literature mapping.,21489
for aquaculture fisheries systems support decision exists? Our questions are: research What and,Our research questions are: What decision support systems for fisheries and aquaculture exists?,21490
Do analytics? real- current fisheries any DSS provide for time,Do any current DSS for fisheries provide real- time analytics?,21491
these in detail the we The paper how employ then systematic mapping questions. method answering,The paper then detail how we employ the systematic mapping method in answering these questions.,21492
analysis the done of research in collected. the studies provide We an,We provide an analysis of the research done in the studies collected.,21493
"challenges approach and directional aliasing, address. statistics, we This regarding preconditioning, which unconventional introduces","This unconventional approach introduces challenges regarding aliasing, directional statistics, and preconditioning, which we address.",21494
descriptions generating Automatically videos plays language for fundamental natural vision a of challenge computer community.,Automatically generating natural language descriptions of videos plays a fundamental challenge for computer vision community.,21495
"MSVD, and M-VAD public i.e., Extensive datasets, are MPII-MD. on conducted three experiments","Extensive experiments are conducted on three public datasets, i.e., MSVD, M-VAD and MPII-MD.",21496
"to improve estimates, Second, we prediction a employ upon scheme. further initial coarse-to-fine","Second, to further improve upon initial estimates, we employ a coarse-to-fine prediction scheme.",21497
category. segmentation detailed labeling of object Semantic by requires pixels a image,Semantic segmentation requires a detailed labeling of image pixels by object category.,21498
"is and ambiguous result this can noisy in information However, labels.","However, this information is ambiguous and can result in noisy labels.",21499
present. image capture instead semantic of can general the inference Global content concepts,Global inference of image content can instead capture the general semantic concepts present.,21500
We on empirically semantic four datasets. network segmentation our standard evaluate,We empirically evaluate our network on four standard semantic segmentation datasets.,21501
"a The (FMO), Fast Moving i.e. notion of Object","The notion of a Fast Moving Object (FMO), i.e.",21502
"may, angular speed. rotate high FMOs do, and with typically","FMOs may, and typically do, rotate with high angular speed.",21503
"not rare are in common are very but FMOs sports videos, elsewhere.","FMOs are very common in sports videos, but are not rare elsewhere.",21504
"barely objects frame, streaks. appear In visible and are often single as semi-transparent a such","In a single frame, such objects are often barely visible and appear as semi-transparent streaks.",21505
and tracking proposed. detection is method FMOs the for of A,A method for the detection and tracking of FMOs is proposed.,21506
proposed annotated on new The is method evaluated dataset. a,The proposed method is evaluated on a new annotated dataset.,21507
"Two applications highlighting, super-resolution and are localization, presented. temporal of","Two applications of localization, temporal super-resolution and highlighting, are presented.",21508
"better required. are To top the differentiate criteria performers, additional","To better differentiate the top performers, additional criteria are required.",21509
"the conditions test (e.g. images, Therefore, challenging mimicking `harder'","Therefore, `harder' test images, mimicking the challenging conditions (e.g.",21510
"which recognize need utilized be occlusion) benchmarking. to objects, humans in routinely for","occlusion) in which humans routinely recognize objects, need to be utilized for benchmarking.",21511
"we two address concerns the To above, contributions. make mentioned","To address the concerns mentioned above, we make two contributions.",21512
"performance (i.e., While outstanding exhibit visible? is recognition what these networks","While these networks exhibit outstanding recognition performance (i.e., what is visible?",21513
"precisely they (i.e., localization accuracy where lack ), is something located?","), they lack localization accuracy (i.e., where precisely is something located?",21514
a features undergoes The recognition. for obtain sequence robust to operations pooling other of stream,The other stream undergoes a sequence of pooling operations to obtain robust features for recognition.,21515
using resolution streams The image the are two at coupled full residuals.,The two streams are coupled at the full image resolution using residuals.,21516
segmentation In approach training this semantic an to adversarial models. train propose paper we,In this paper we propose an adversarial training approach to train semantic segmentation models.,21517
"concerns conducting three competitiveness cost, as precision. and such aspects about It ranking tournament development","It concerns about three aspects such as tournament conducting cost, competitiveness development and ranking precision.",21518
using a proposes new tree It to matches. progress detect then throwaway potential method,It then proposes a new method using progress tree to detect potential throwaway matches.,21519
and using reveals performed the tournament analysis structures. proposed of weaknesses strengths the The method,The analysis performed using the proposed method reveals the strengths and weaknesses of tournament structures.,21520
precision reliable for all participants. Round-robin ranking gives,Round-robin gives reliable ranking precision for all participants.,21521
"conduction to competitiveness and it maintain is fails cost However, its development. very high,","However, its conduction cost is very high, and it fails to maintain competitiveness development.",21522
training The requires phase simple and is proposed no to approach implement.,The proposed approach requires no training phase and is simple to implement.,21523
"parameter settings. significantly, inter-database it satisfactorily under More quite performs","More significantly, it performs quite satisfactorily under inter-database parameter settings.",21524
relative the ZSR features of is scarcity key One of source-domain challenges (e.g. the in,One of the key challenges in ZSR is the relative scarcity of source-domain features (e.g.,21525
helps smoothness better achieving This output in and generalization.,This helps in achieving better output smoothness and generalization.,21526
"EMD computationally problems However usually techniques is require solving optimization expensive.Moreover, EMD like complex lasso.","However EMD is computationally expensive.Moreover, solving EMD optimization problems usually require complex techniques like lasso.",21527
limit machine These large EMD-based properties learning. scale the of approaches in applicability,These properties limit the applicability of EMD-based approaches in large scale machine learning.,21528
support claims experiments with datasets. We real our on,We support our claims with experiments on real datasets.,21529
"can The modular, different proposed be framework instantiated is variants. into scalable, and","The proposed framework is modular, scalable, and can be instantiated into different variants.",21530
studied and are Two instantiations higher-order pairwise exploring constraints.,Two instantiations are studied exploring pairwise and higher-order constraints.,21531
is challenge of incentives rational agents many for because self-preservation. formulations strong create a This,This is a challenge because many formulations of rational agents create strong incentives for self-preservation.,21532
this (R in itself no incentive off switch to has also setting.),(R also has no incentive to switch itself off in this setting.),21533
problem in a is hand-held Motion from camera by blur shake videos devices. major captured,Motion blur from camera shake is a major problem in videos captured by hand-held devices.,21534
best As rely result the on performing aligning methods a frames. nearby,As a result the best performing methods rely on aligning nearby frames.,21535
critical are in data systems. face Training recognition,Training data are critical in face recognition systems.,21536
"face a tedious. for very domain particular is a However, scale large labeling data","However, labeling a large scale face data for a particular domain is very tedious.",21537
for the identity. maximal we use Then that cleaned as subgraph the data,Then we use the maximal subgraph as the cleaned data for that identity.,21538
conventional modification first for deconvolution. blind a introduce a joint function energy We to slight,We first introduce a slight modification to a conventional joint energy function for blind deconvolution.,21539
the instance-aware work semantic segmentation. addresses This of task,This work addresses the task of instance-aware semantic segmentation.,21540
the on evaluate challenging our dataset. CityScapes We approach,We evaluate our approach on the challenging CityScapes dataset.,21541
importance Ariel (UAV) vehicle for greater has security. border the army in Unmanned An,An Unmanned Ariel vehicle (UAV) has greater importance in the army for border security.,21542
"detecting infiltrated attacking targets. are for UAVs ground Currently, the and used","Currently, UAVs are used for detecting and attacking the infiltrated ground targets.",21543
UAV. damages unwanted of avoid project and aims such This collisions to,This project aims to avoid such unwanted collisions and damages of UAV.,21544
is and that UAV detect also used to for track humans. uses Voila-jones surveillance algorithm,UAV is also used for surveillance that uses Voila-jones algorithm to detect and track humans.,21545
vision. function object algorithm cascade and This uses detector,This algorithm uses cascade object detector function and vision.,21546
time. main code advantage processing The of is the this reduced,The main advantage of this code is the reduced processing time.,21547
Neural new exciting Style of very has image manipulation. enabling shown results forms Transfer,Neural Style Transfer has shown very exciting results enabling new forms of image manipulation.,21548
the covering study types We the among relationships based approximation three operators. of $\gamma-$fuzzy also,We also study the relationships among the three types of $\gamma-$fuzzy covering based approximation operators.,21549
investigate among $\gamma-$fuzzy types coverings of operators. the based We these also relationships approximation,We also investigate the relationships among these types of $\gamma-$fuzzy coverings based approximation operators.,21550
"applications. contrast, many there despite of In models are animals, such few","In contrast, there are few such models of animals, despite many applications.",21551
are humans. cooperative much that challenge less animals main than is The,The main challenge is that animals are much less cooperative than humans.,21552
registration an the shape to novel initial We compute model employ a to scans. part-based,We employ a novel part-based shape model to compute an initial registration to the scans.,21553
"and fit Animal model, data. animated, can posed, from sampled shapes be the to","Animal shapes can be sampled from the model, posed, animated, and fit to data.",21554
"model (i.e. training data does this not Crucially, labelled require pairwise","Crucially, this model does not require pairwise labelled training data (i.e.",21555
is challenge cognitive dynamics onto the functions. neuroimaging neurophysiological mapping A major of understanding in,A major challenge in neuroimaging is understanding the mapping of neurophysiological dynamics onto cognitive functions.,21556
on geometric leverage representational these convolutional the neural We compute quantities. power of to networks,We leverage on the representational power of convolutional neural networks to compute these geometric quantities.,21557
"representation metric multi-scale a novel learning Finally, paradigm. a in develop we similarity","Finally, we develop a novel multi-scale representation in a similarity metric learning paradigm.",21558
allows of region to This enlarge both tracking improves search and objects. the non-rectangular,This both allows to enlarge the search region and improves tracking of non-rectangular objects.,21559
a CSR-DCF CPU. runs The on in real-time,The CSR-DCF runs in real-time on a CPU.,21560
prior of addressed. issue carefully inputs is The,The issue of prior inputs is carefully addressed.,21561
"The through is package Comprehensive CRAN. R the Archive Network, available","The package is available through the Comprehensive R Archive Network, CRAN.",21562
several the illustrate with use BayesVarSel We data examples. of,We illustrate the use of BayesVarSel with several data examples.,21563
For evaluation. a performed independent task and classification subject-dependent each we,For each classification task we performed a subject-dependent and independent evaluation.,21564
nonlocal and the decomposition. self-similarity combines a proposed This sparse method partition modified,This method combines the nonlocal self-similarity partition and a proposed modified sparse decomposition.,21565
nonlocal partition data a The of structure-similarity method sets. series groups,The nonlocal partition method groups a series of structure-similarity data sets.,21566
over-complete a learning has an sparsity dictionary in data Each set sparse for good representation.,Each data set has a good sparsity for learning an over-complete dictionary in sparse representation.,21567
atoms. is set data performed with the over Despeckling principal principal on dictionary each,Despeckling is performed on each data set over the principal dictionary with principal atoms.,21568
"detector. For first the recent CNN-based a use we step,","For the first step, we use a recent CNN-based detector.",21569
This of the accuracy ZSL on generalisation compromises the data. recognition target,This compromises the generalisation accuracy of ZSL recognition on the target data.,21570
"decomposition. important role Approximation processing, in deconvolution and theory image an especially plays image","Approximation theory plays an important role in image processing, especially image deconvolution and decomposition.",21571
from videos present detecting esports live game We a streaming highlights matches. technique for of,We present a technique for detecting highlights from live streaming videos of esports game matches.,21572
is Yahoo currently of on production Part our technique deployed Esports. in,Part of our technique is currently deployed in production on Yahoo Esports.,21573
that proposition is certain its acquired true. Subsequently evidence FE communicates confirmation a recently,Subsequently FE communicates its recently acquired confirmation that a certain evidence proposition is true.,21574
belief mediated state. likelihood own TOF ratio transfer Then their thereby performs reasoning revising,Then TOF performs likelihood ratio transfer mediated reasoning thereby revising their own belief state.,21575
How in? about was single does taken much image the environment a it reveal,How much does a single image reveal about the environment it was taken in?,21576
proposed of model jointly method environments statistics and allows material properties. the us to The,The proposed method allows us to jointly model the statistics of environments and material properties.,21577
"from training our but applicability demonstrate data. system We train real-world to synthesized data, its","We train our system from synthesized training data, but demonstrate its applicability to real-world data.",21578
objects. materials yet recognition set methods labels Many like as treat of recent material another,Many recent material recognition methods treat materials as yet another set of labels like objects.,21579
object and CNNs. independent network integrate estimates to this We place from,We integrate object and place estimates to this network from independent CNNs.,21580
"based. work the advantages also segmentation has transparency, all being its and Our in simplicity,","Our work also has the advantages in its transparency, simplicity, and being all segmentation based.",21581
"available description of provided. codes datasets is different resources the a such and Finally, as","Finally, a description of the different resources available such as datasets and codes is provided.",21582
"viewing physical better the stereoscopic Immersive, phenomena. enables visualized to analyze scientists structures spatial of","Immersive, stereoscopic viewing enables scientists to better analyze the spatial structures of visualized physical phenomena.",21583
"presented traditional core cannot However, their properly these findings which in media, lack be attributes.","However, their findings cannot be properly presented in traditional media, which lack these core attributes.",21584
"scientific be can renderings impractical equipment, require to high-end which obtain. often Immersive","Immersive scientific renderings often require high-end equipment, which can be impractical to obtain.",21585
uniquely signals learning leverage first two videos. in person We embedded,We leverage two learning signals uniquely embedded in first person videos.,21586
"person link videos individuals follow can to attention first Collectively, that a group. the joint","Collectively, first person videos follow joint attention that can link the individuals to a group.",21587
"Thus, a this approach considered be analysis. as multi-scale can","Thus, this approach can be considered as a multi-scale analysis.",21588
"that argumentation graphs This weighted of i.e. discusses paper the semantics are biplor,","This paper discusses the semantics of weighted argumentation graphs that are biplor, i.e.",21589
"Ben-Naim Amgoud, et. The previous work work on by builds","The work builds on previous work by Amgoud, Ben-Naim et.",21590
human understanding present social image sequences. framework for behaviors in We a raw unified,We present a unified framework for understanding human social behaviors in raw image sequences.,21591
is Recurrent temporal matching The Neural handled via person-level consistency a Network.,The temporal consistency is handled via a person-level matching Recurrent Neural Network.,21592
state-of-the-art multiple of algorithm demonstrate publicly available performance on We our benchmarks.,We demonstrate state-of-the-art performance of our algorithm on multiple publicly available benchmarks.,21593
their major To two we limitations developments. introduce overcome,To overcome their limitations we introduce two major developments.,21594
completion has considered work Previous depth of semantic maps scene separately. and labeling,Previous work has considered scene completion and semantic labeling of depth maps separately.,21595
"intertwined. tightly these However, problems are we two observe that","However, we observe that these two problems are tightly intertwined.",21596
image great in has achieved Sparse studies. coding a success various processing,Sparse coding has achieved a great success in various image processing studies.,21597
rank sparsity of the paper strategy based This the minimization. on analyzes the group of,This paper analyzes the sparsity of group based on the strategy of the rank minimization.,21598
"for group Firstly, adaptive designed. an dictionary each is","Firstly, an adaptive dictionary for each group is designed.",21599
"a well-designed network, train amounts is data indispensable. To deep such tremendous of","To train such a well-designed deep network, tremendous amounts of data is indispensable.",21600
Deterministic walk attracted has (DTW) tourist interest computer in increasing vision.,Deterministic tourist walk (DTW) has attracted increasing interest in computer vision.,21601
"were methods different static textures analysis for proposed. the last years, In of and dynamic","In the last years, different methods for analysis of dynamic and static textures were proposed.",21602
"very little. the In decreases rate correct addition, classification","In addition, the correct classification rate decreases very little.",21603
derived growth. of and We ectotherms intake body their energy calculated endotherms mass from functions,We calculated energy intake functions of ectotherms and endotherms derived from their body mass growth.,21604
"been techniques, developed. algorithms have theories, rapidly New","New theories, techniques, algorithms have been rapidly developed.",21605
in the literature in are a time. proposed period short of,are proposed in the literature in a short period of time.,21606
"a automatically aim we actors this in TV series. labeling In work, at","In this work, we aim at automatically labeling actors in a TV series.",21607
is the uniformity methods. main one criteria of superpixel Size of,Size uniformity is one of the main criteria of superpixel methods.,21608
an rarely the image. of varying uniformity conforms content But to size,But size uniformity rarely conforms to the varying content of an image.,21609
is that a uniformity. appropriate creating information propose more image segments criterion We for,We propose that a more appropriate criterion for creating image segments is information uniformity.,21610
introduce on based an image We segmenting this novel criterion. method for a,We introduce a novel method for segmenting an image based on this criterion.,21611
the to to it segmentation superior comparison proves be On benchmarks state-of-the-art.,On segmentation comparison benchmarks it proves to be superior to the state-of-the-art.,21612
are when representations. to results reported compared Superior state-of-the-art,Superior results are reported when compared to state-of-the-art representations.,21613
expected encountered sufficiently the to data cover environment. must the be in variability This,This data must sufficiently cover the variability expected to be encountered in the environment.,21614
terms Markov Reinforcement often processes are in of decision problems (RL) (MDPs). learning phrased,Reinforcement learning (RL) problems are often phrased in terms of Markov decision processes (MDPs).,21615
"Bayesian AIXI. particular results on negative agents, We RL establish in","We establish negative results on Bayesian RL agents, in particular AIXI.",21616
"Moreover, of the optimal. all class every computable Pareto is in policy environments","Moreover, in the class of all computable environments every policy is Pareto optimal.",21617
undermines This for optimality existing AIXI. properties all,This undermines all existing optimality properties for AIXI.,21618
Thompson sampling sublinear environments. in these achieves Hence regret,Hence Thompson sampling achieves sublinear regret in these environments.,21619
great dense semantic including (FCN) have Convolution tasks achieved success Networks in prediction segmentation. Fully,Fully Convolution Networks (FCN) have achieved great success in dense prediction tasks including semantic segmentation.,21620
"we our Network present (IFCN). Next, Convolution Fully Improved","Next, we present our Improved Fully Convolution Network (IFCN).",21621
"the segmentation modifications Empirically, architecture to proven performance. are enhance be to significant those","Empirically, those architecture modifications are proven to be significant to enhance the segmentation performance.",21622
and given inertial poses. measurements a We the recovers that propose method camera scale metric,We propose a method that recovers the metric scale given inertial measurements and camera poses.,21623
"with can reconstruction Therefore, combined any existing software. visual our be easily solution","Therefore, our solution can be easily combined with any existing visual reconstruction software.",21624
"In an unsupervised method learns this untrimmed that videos. from we paper, temporal propose modeling","In this paper, we propose an unsupervised temporal modeling method that learns from untrimmed videos.",21625
"slowly. or e.g., run constantly, may motion of The speed a varies man quickly","The speed of motion varies constantly, e.g., a man may run quickly or slowly.",21626
variance. capable the learned dealing This makes with more motion of learning speed process model,This learning process makes the learned model more capable of dealing with motion speed variance.,21627
"are functional or unimodal of findings these neuroimaging accessible analyses to Importantly, alone. anatomical not","Importantly, these findings are not accessible to unimodal analyses of functional or anatomical neuroimaging alone.",21628
"between nodes are realistic many In networks, the representing the edges the interactions time-varying.","In many realistic networks, the edges representing the interactions between the nodes are time-varying.",21629
in efficacy We seizures. point removal the spreaders of limiting out of also such,We also point out the efficacy of removal of such spreaders in limiting seizures.,21630
of certain have sensing to their evolved region spectrum. electromagnetic restrict to capabilities Animals,Animals have evolved to restrict their sensing capabilities to certain region of electromagnetic spectrum.,21631
"fantastic typography. problem we In for the generating the special-effects explore this work, of","In this work, we explore the problem of generating fantastic special-effects for the typography.",21632
statistically. two first estimates features their Our these method correlation derives and,Our method first estimates these two features and derives their correlation statistically.,21633
"typography algorithm with extensive In library our the validate addition, we of artistic generation. effectiveness","In addition, we validate the effectiveness of our algorithm with extensive artistic typography library generation.",21634
"a layer referred includes to An DAE, detector(ND). unique contribution single as novelty","An unique contribution includes a single layer DAE, referred to as novelty detector(ND).",21635
a patches non-lesion reconstruct squared error loss was accurately trained mean ND function. using to,ND was trained to accurately reconstruct non-lesion patches using a mean squared error loss function.,21636
The lesions. containing were data reconstruction used of identify test to error maps regions,The reconstruction error maps of test data were used to identify regions containing lesions.,21637
registration Deformable analysis. medical ubiquitous is image in,Deformable registration is ubiquitous in medical image analysis.,21638
"However, dataset exists. model such for no training sufficient a","However, no sufficient dataset for training such a model exists.",21639
was training for pre-training Unsupervised ago. years neural a networks deep critical technique,Unsupervised pre-training was a critical technique for training deep neural networks years ago.,21640
"usually to scale. is However, data easier unlabeled of obtain very and large","However, unlabeled data is easier to obtain and usually of very large scale.",21641
is a connections. with use we The auto-encoder convolutional shortcut network architecture symmetric,The architecture we use is a convolutional auto-encoder network with symmetric shortcut connections.,21642
on This uptake a the broader puts of CP. brakes,This puts the brakes on a broader uptake of CP.,21643
(reified) new Our variables. include contributions numerical constraints of and on constraints decompositions,Our contributions include new decompositions of (reified) constraints and constraints on numerical variables.,21644
"neurons influenced the entire population by external In input. the states, is globally of such","In such states, the entire population of neurons is globally influenced by the external input.",21645
low-level of this responsiveness We form constitute suggest sensory awareness. that enhanced a network-level,We suggest that this network-level enhanced responsiveness constitute a low-level form of sensory awareness.,21646
for on built image semantic approaches (CNNs). Convolutional Networks are segmentation State-of-the-art Neural,State-of-the-art approaches for semantic image segmentation are built on Convolutional Neural Networks (CNNs).,21647
refine to Fields) the Conditional predictions. model Random,Conditional Random Fields) to refine the model predictions.,21648
"we problem the paper, deal of with this segmentation. DenseNets In semantic extend to","In this paper, we extend DenseNets to deal with the problem of semantic segmentation.",21649
"annotating massive expensive dataset However, and is a time-consuming.","However, annotating a massive dataset is expensive and time-consuming.",21650
"paper, a generalization define of probabilistic approach. In this this we","In this paper, we define a probabilistic generalization of this approach.",21651
"latent can done arithmetic. be expression time, test editing vector simply using At","At test time, expression editing can be done simply using latent vector arithmetic.",21652
point-count animal detection require accurate estimates of surveys Abundance estimates probabilities. from,Abundance estimates from animal point-count surveys require accurate estimates of detection probabilities.,21653
"with model estimation. of is analysis unknown-p survival abundance thus Our time-to-event unknown-N, a combination","Our model is thus a combination of survival time-to-event analysis with unknown-N, unknown-p abundance estimation.",21654
"explore two-parameter of families e.g. specifically We TTDDs,","We specifically explore two-parameter families of TTDDs, e.g.",21655
network design for network optimizing Stochastic connectivity. framework general is a,Stochastic network design is a general framework for optimizing network connectivity.,21656
"large networks. to our scales approach Empirically, well river","Empirically, our approach scales well to large river networks.",21657
are impressive. qualitatively in reconstruction Results face,Results in face reconstruction are qualitatively impressive.,21658
"serious species considered a threat to Therefore, public health worldwide. is this","Therefore, this species is considered a serious threat to public health worldwide.",21659
on to the Due that the Ae. fact hand one,Due to the fact on the one hand that Ae.,21660
"is all inefficient. However,pixel-wisely frames tedious annotating and","However,pixel-wisely annotating all frames is tedious and inefficient.",21661
"To segment considered. is preceding frame, parse the jointly video one particular frame the","To parse one particular frame, the video segment preceding the frame is jointly considered.",21662
two superiority Experimental state-of-the-arts. the show over video surveillance datasets results of SVP on,Experimental results on two surveillance video datasets show the superiority of SVP over state-of-the-arts.,21663
be things words important can temporal for that for ventral seen. regions,ventral temporal regions important for words for things that can be seen.,21664
ANG. Semantic strongly is with associated the believed be to processing,Semantic processing is believed to be strongly associated with the ANG.,21665
closely related anterior SMG. the ability has the Phonological to of been surfaces,Phonological ability has been closely related to the anterior surfaces of the SMG.,21666
ANG. thought Pathways link posterior the to and the are SMG the through anterior SMG,Pathways through the posterior SMG are thought to link the anterior SMG and the ANG.,21667
Precentral are retrieval. with phonological structures associated,Precentral structures are associated with phonological retrieval.,21668
"of attributed between Furthermore, different can anatomies to demands be differences languages. vocabulary brain","Furthermore, differences between brain anatomies can be attributed to vocabulary demands of different languages.",21669
optimizing is function. cost as matching a registration by image commonly an metric Image performed,Image registration is commonly performed by optimizing an image matching metric as a cost function.,21670
approach. provided visualizations and studies to Ablation our understand are,Ablation studies and visualizations are provided to understand our approach.,21671
"approximate on structured Variational for in- provides powerful tool complex, a models. ference inference probabilistic","Variational inference provides a powerful tool for approximate probabilistic in- ference on complex, structured models.",21672
inference vari- largely the design of This ational methods. limits and implementation,This largely limits the design and implementation of vari- ational inference methods.,21673
to often slow algorithms polynomial-time are lead practice. they in While convex could relaxations,While convex relaxations could lead to polynomial-time algorithms they are often slow in practice.,21674
"other On theoretical exist the non-convex limited hand, guarantees for methods.","On the other hand, limited theoretical guarantees exist for non-convex methods.",21675
important tomography computed are images. findings Oral lesions (CT) on,Oral lesions are important findings on computed tomography (CT) images.,21676
"for this introduce algorithm we CRFs. minimization To an deficiency, alleviate efficient dense LP","To alleviate this deficiency, we introduce an efficient LP minimization algorithm for dense CRFs.",21677
We variables that of block can each optimized. be efficiently show,We show that each block of variables can be efficiently optimized.,21678
"optimized the the block, For problem conditional gradient via descent. other is","For the other block, the problem is optimized via conditional gradient descent.",21679
challenging for detection real-world computer is vision important Facial yet landmark an applications. task,Facial landmark detection is an important yet challenging task for real-world computer vision applications.,21680
"baseline method refine features convolutional by keypoints pooling We improving the iteratively, significantly.","We refine keypoints by pooling convolutional features iteratively, improving the baseline method significantly.",21681
Sequence has powerful models lead modeling with data. music networks symbolic neural to of,Sequence modeling with neural networks has lead to powerful models of symbolic music data.,21682
strategy and a compare beam search with addition In conditional we for provide probability maximisation.,In addition we provide and compare with a beam search strategy for conditional probability maximisation.,21683
algorithms convincingly capable are re-harmonising famous of musical Our works.,Our algorithms are capable of convincingly re-harmonising famous musical works.,21684
"test audio quantitative experiments, we demonstrate To visualisations, a this listening human examples. provide and","To demonstrate this we provide visualisations, quantitative experiments, a human listening test and audio examples.",21685
"effective, in both procedures character. sampling complementary be and the optimisation yet find We to","We find both the sampling and optimisation procedures to be effective, yet complementary in character.",21686
of other applications. algorithms countless The generality our creative permits,The generality of our algorithms permits countless other creative applications.,21687
algorithms literature. learning landcover based Deep have been proposed recently in classification,Deep learning based landcover classification algorithms have recently been proposed in literature.,21688
training independent and lesser requires of thus data. architecture number fewer connection The weights has,The architecture has fewer independent connection weights and thus requires lesser number of training data.,21689
steps. two involves object an the to Fitting model,Fitting the model to an object involves two steps.,21690
"limitation boundaries. However, conventional at in the layer representing have models occlusion occurring a interactions","However, conventional models have a limitation in representing the layer interactions occurring at occlusion boundaries.",21691
a network. algorithm a new interpolation sparse-to-dense based propose fully We on data-driven convolutional,We propose a new data-driven sparse-to-dense interpolation algorithm based on a fully convolutional network.,21692
the importance also We learning contour of the process. to the image show,We also show the importance of the image contour to the learning process.,21693
and is these interpreting essential relationships such language for natural expressions. Understanding grounding,Understanding these relationships is essential for interpreting and grounding such natural language expressions.,21694
"on tasks. all on datasets, referential expression approaches evaluate outperforming CMNs state-of-the-art We multiple","We evaluate CMNs on multiple referential expression datasets, outperforming state-of-the-art approaches on all tasks.",21695
"(SISR). various single-image been approaches (DNN)-based deep-neural-network proposed super-resolution have for Recently,","Recently, various deep-neural-network (DNN)-based approaches have been proposed for single-image super-resolution (SISR).",21696
"We method, style our algorithm. call (HR) transfer high-resolution","We call our method, high-resolution (HR) style transfer algorithm.",21697
Image in important image processing. an pattern digital is recognition area,Image pattern recognition is an important area in digital image processing.,21698
"termination. until standard parameters In the these AFSA, constant remain algorithm two","In standard AFSA, these two parameters remain constant until the algorithm termination.",21699
four benchmark have The the based approaches well-known proposed on functions. evaluated been,The proposed approaches have been evaluated based on the four well-known benchmark functions.,21700
AFSA. positive Experimental considerable impact results show performance the of on,Experimental results show considerable positive impact on the performance of AFSA.,21701
"In framework. unified constitute that novel parsing contribute two a methods we particular,","In particular, we contribute two novel methods that constitute a unified parsing framework.",21702
"we epitope. introduced of Using anchor algorithm, Layers residues the concept identify to","Using Layers algorithm, we introduced the concept of anchor residues to identify epitope.",21703
search space optimized the for epitope identification. We reduction,We optimized the search space reduction for epitope identification.,21704
selected using clustered are means Fuzzy C The algorithm. features,The selected features are clustered using Fuzzy C means algorithm.,21705
"log-likelihood maximizes the ratio we rate the that this (i.e. note, investigate expected In","In this note, we investigate the rate that maximizes the expected log-likelihood ratio (i.e.",21706
Ultrasound are segmentation in surveyed Three-dimensional paper. this methods image,Three-dimensional Ultrasound image segmentation methods are surveyed in this paper.,21707
"be lens forward model needs the to For distortion known. rectifying distortion,","For rectifying the lens distortion, forward distortion model needs to be known.",21708
"the inverse model. methods many estimate However, self-calibration distortion","However, many self-calibration methods estimate the inverse distortion model.",21709
propose rectification a distortion directly. that the uses novel distortion method model inverse We,We propose a novel distortion rectification method that uses the inverse distortion model directly.,21710
are set points locations with triangulated. subpixel of resulting The,The resulting set of points with subpixel locations are triangulated.,21711
the of pixel triangulation. on image this rectified The interpolated values are linearly based,The pixel values of the rectified image are linearly interpolated based on this triangulation.,21712
"process rewards. are the usually by In Markov decision evaluated policies cumulative model, expected","In the Markov decision process model, policies are usually evaluated by expected cumulative rewards.",21713
horizons infinite finite are considered. Both and,Both finite and infinite horizons are considered.,21714
challenging. is in pose the wild estimation Multi-person,Multi-person pose estimation in the wild is challenging.,21715
invader's taken Fickian. The as neutral is diffusion,The invader's diffusion is taken as neutral Fickian.,21716
"hinders it environmental multiplicative studied Furthermore, is fosters noise or invasion. how the","Furthermore, it is studied how multiplicative environmental noise fosters or hinders the invasion.",21717
the accuracy. their of is Critical the data utility to,Critical to the utility of the data is their accuracy.,21718
and thoroughly implementation details quality. technique the on varied outline of GPS We data validate,We outline implementation details and thoroughly validate the technique on GPS data of varied quality.,21719
"a disjoint The the split method two adds in sub-networks. resulting to network,","The method adds a split to the network, resulting in two disjoint sub-networks.",21720
"the the extract from signal. input entire Together, sub-networks features","Together, the sub-networks extract features from the entire input signal.",21721
achieves performance learning benchmarks. on This transfer several state-of-the-art large-scale method,This method achieves state-of-the-art performance on several large-scale transfer learning benchmarks.,21722
"in long-standing learning. been goal a images photo-realistic machine Generating high-resolution, has","Generating high-resolution, photo-realistic images has been a long-standing goal in machine learning.",21723
"performs at of image task inpainting. our well we model Finally, reasonably the show that","Finally, we show that our model performs reasonably well at the task of image inpainting.",21724
"specific in the Moreover, videos. values/characteristics few surveillance large-scale explore algorithms","Moreover, few algorithms explore the specific values/characteristics in large-scale surveillance videos.",21725
in us areas This to global structure predict accuracy. high at unknown allows,This allows us to predict global structure in unknown areas at high accuracy.,21726
model estimation. hybrid and propose CNN+CRF novel for principled a stereo We,We propose a novel and principled hybrid CNN+CRF model for stereo estimation.,21727
frequency-scaling These not consistent display which with media. are properties resistive signals,These signals display frequency-scaling properties which are not consistent with resistive media.,21728
recently impedance measurements conditions. physiological scaling in was a Such by also found,Such a scaling was also found recently by impedance measurements in physiological conditions.,21729
"However, essentially is that extracellular the medium other resistive. measurements suggest","However, other measurements suggest that the extracellular medium is essentially resistive.",21730
converting into process documents Unicode refers Recognition Character the translated to Text Text. of printed,Character Recognition refers to the process of converting printed Text documents into translated Unicode Text.,21731
"The in papers, of available printed the form magazines, documents etc. books,","The printed documents available in the form of books, papers, magazines, etc.",21732
produce which standard the using an are document. scanned image scanned of scanners,are scanned using standard scanners which produce an image of the scanned document.,21733
by Lines and top identify are where line. we bottom identifying an of algorithm,Lines are identifying by an algorithm where we identify top and bottom of line.,21734
"available For is publicly is database used this evaluation, which field. HFH in","For evaluation, HFH database is used which is publicly available in this field.",21735
"deep systems. It, neural however, ignored network in has detection been based largely","It, however, has been largely ignored in deep neural network based detection systems.",21736
region-based aware method convolutional network the (ARC-R-CNN). aspect called ratio context proposed is The and,The proposed method is called the aspect ratio and context aware region-based convolutional network (ARC-R-CNN).,21737
high obtains mAP IoU thresholds using on It better both significantly datasets. performance,It obtains significantly better mAP performance using high IoU thresholds on both datasets.,21738
"a model propose for captioning. novel ""Areas attention-based of We image automatic Attention"",","We propose ""Areas of Attention"", a novel attention-based model for automatic image captioning.",21739
"training. detector to are associations captions, During object these weakly-supervised image-level training from akin inferred","During training these associations are inferred from image-level captions, akin to weakly-supervised object detector training.",21740
testing. These improve captioning localizing to regions corresponding the by during associations help,These associations help to improve captioning by localizing the corresponding regions during testing.,21741
We as tasks distinct formulate variables correspond poses/cells. where integer to programs these,We formulate these distinct tasks as integer programs where variables correspond to poses/cells.,21742
of and spaces of exploration in the poses cells. This results efficient,This results in efficient exploration of the spaces of poses and cells.,21743
are a to of results popular improve way Ensembles CNNs. discriminative,Ensembles are a popular way to improve results of discriminative CNNs.,21744
from significantly. different combination The initializations starting networks several of trained improves results,The combination of several networks trained starting from different initializations improves results significantly.,21745
of this the ensembles investigate usage In we of paper GANs.,In this paper we investigate the usage of ensembles of GANs.,21746
several nature up The opens specific new ways GANs ensembles. construct to of,The specific nature of GANs opens up several new ways to construct ensembles.,21747
faster ensembles self ensembles. train so-called than to much These are traditional,These so-called self ensembles are much faster to train than traditional ensembles.,21748
illustrating our also presented results. Examples are,Examples are also presented illustrating our results.,21749
dynamic feature This maps. generalizing layer allows to images dynamic new,This new layer allows generalizing dynamic images to dynamic feature maps.,21750
worldwide. of cancer cause cancer third (CRC) Colorectal death the is,Colorectal cancer (CRC) is the third cause of cancer death worldwide.,21751
for a driving. autonomous task is Object detection crucial,Object detection is a crucial task for autonomous driving.,21752
"model which Our small leads energy model a to fully-convolutional, size and efficiency. better is","Our model is fully-convolutional, which leads to a small model size and better energy efficiency.",21753
Many based are patch-based processing. methods self-similarity image presented denoising or state-of-the-art the on,Many presented state-of-the-art denoising methods are based on the self-similarity or patch-based image processing.,21754
"has increasing Specifically, Non-Local attention a in (NLM) years. recent gained as filter patch-based Means","Specifically, Non-Local Means (NLM) as a patch-based filter has gained increasing attention in recent years.",21755
"Visual and addressed traditionally i.e. fashion, pipeline recognition are in restoration","Visual restoration and recognition are traditionally addressed in pipeline fashion, i.e.",21756
matching. We new unsupervised action of task action a propose by detection,We propose a new task of unsupervised action detection by action matching.,21757
share A human pair video matched of if they the are action. same segments,A pair of video segments are matched if they share the same human action.,21758
to in videos action detection Unsupervised matching by align meaningful a us manner. action allows,Unsupervised action detection by action matching allows us to align videos in a meaningful manner.,21759
"useful videos. pre-processing for e.g., highlights, Moreover, step sports a generating it from is video","Moreover, it is a useful pre-processing step for generating video highlights, e.g., from sports videos.",21760
effective unsupervised action efficient an present for detection. method We and,We present an effective and efficient method for unsupervised action detection.,21761
"uncertainty task inherent cases, is some the In in itself.","In some cases, uncertainty is inherent in the task itself.",21762
"for example, are future In distinct outcomes prediction, valid. many equally","In future prediction, for example, many distinct outcomes are equally valid.",21763
"from other data the labeled. uncertainty In is way cases, arises","In other cases, uncertainty arises from the way data is labeled.",21764
approach this such a focus for scenarios. handling we principled on In work,In this work we focus on a principled approach for handling such scenarios.,21765
by Each or is label a super-pixel parameter. characterized,Each super-pixel is characterized by a label or parameter.,21766
the textures. useful showing images This vesicular quite is studying for approach,This approach is quite useful for studying the images showing vesicular textures.,21767
several field to receptive novel modelling. contributions offers This,This offers several novel contributions to receptive field modelling.,21768
in the accounted brain for. is across haemodynamic Variability the response,Variability in the haemodynamic response across the brain is accounted for.,21769
the critical computer a Power of systems. for consumption vision deployment factor is embedded,Power consumption is a critical factor for the deployment of embedded computer vision systems.,21770
"recognition. gesture face pose regression, object These and recognition, include head detection,","These include object recognition, head pose regression, face detection, and gesture recognition.",21771
"to ""physical rational This cameras. framework realizations"" builds on this study paper of","This paper builds on this framework to study ""physical realizations"" of rational cameras.",21772
"leads algorithms to and this turn, In for structure motion from self-calibration. new","In turn, this leads to new algorithms for structure from motion and self-calibration.",21773
"In feature this learning we explore work, video. unlabeled from unsupervised","In this work, we explore unsupervised feature learning from unlabeled video.",21774
based accuracy. chosen is best the The and model on prediction reconstruction,The best model is chosen based on the reconstruction and prediction accuracy.,21775
to video shown be for effective tool are and modeling sequences. predicting an units Conv-LSTM,Conv-LSTM units are shown to be an effective tool for modeling and predicting video sequences.,21776
"state-of-the-art two the experiments, our In datasets. popular method on outperforms","In experiments, our method outperforms the state-of-the-art on two popular datasets.",21777
logical languages. We inference examine by Bayesian complexity networks the of specified in,We examine the complexity of inference in Bayesian networks specified by logical languages.,21778
have space time data domain. on plentiful Traffic flow characteristics and both,Traffic flow data have plentiful characteristics on both time and space domain.,21779
"flow LSTM However, limited. on and applications of approaches traffic are CNN","However, applications of CNN and LSTM approaches on traffic flow are limited.",21780
"performed fusion to Given features, forecasting. short-term achieve those the is meaningful feature-level","Given those meaningful features, the feature-level fusion is performed to achieve short-term forecasting.",21781
proposed The other open an CLTFP is on compared methods datasets. forecasting with popular,The proposed CLTFP is compared with other popular forecasting methods on an open datasets.,21782
the advantages indicate Experimental in has considerable traffic that flow results forecasting. CLTFP,Experimental results indicate that the CLTFP has considerable advantages in traffic flow forecasting.,21783
unrestricted is diverse parsing for vocabulary Scene open and scenes. challenging,Scene parsing is challenging for unrestricted open vocabulary and diverse scenes.,21784
on state-of-the-art The approach performance achieves proposed datasets. various,The proposed approach achieves state-of-the-art performance on various datasets.,21785
porous converting into biological hybrids. a for and method is translucent CLARITY tissues hydrogel-tissue,CLARITY is a method for converting biological tissues into translucent and porous hydrogel-tissue hybrids.,21786
for adopt via diffeomorphic information mutual large Thus matching. images we deformation approach a registering,Thus we adopt a large deformation diffeomorphic approach for registering images via mutual information matching.,21787
is a scale dataset such large labeling Manually infeasible.,Manually labeling such a large scale dataset is infeasible.,21788
"different our create ground to maps of utilize sources high-precision we Instead, truth. propose to","Instead, we propose to utilize different sources of high-precision maps to create our ground truth.",21789
Recognition document words be in and of handwritten analysis important continues an to recognition. problem,Recognition of handwritten words continues to be an important problem in document analysis and recognition.,21790
word from poorly with extract data images--which approaches perform Existing features can new sets. hand-engineered,Existing approaches extract hand-engineered features from word images--which can perform poorly with new data sets.,21791
including and results they recognition. character have in scene yielded state-of-the-art tasks classification Moreover recognition,Moreover they have yielded state-of-the-art results in classification tasks including character recognition and scene recognition.,21792
"deep for propose recognition. this we (deep Random using Fields CRFs) paper, Conditional word In","In this paper, we propose using deep Conditional Random Fields (deep CRFs) for word recognition.",21793
"higher-dimensional flow. to we this introduce To a improve process, increase hint information layer","To improve this process, we introduce a higher-dimensional hint layer to increase information flow.",21794
but expensive. is current quality segmentation in defines the prohibitively gold standard Manual,Manual segmentation defines the current gold standard in quality but is prohibitively expensive.,21795
"compare simple our system to Additionally, trained criteria. humans with we","Additionally, we compare our system to humans trained with simple criteria.",21796
of Our to use system healthcare clinical for is the providers. potential intended capabilities augment,Our system is intended for potential clinical use to augment the capabilities of healthcare providers.,21797
super-resolution computational introduce EM. method for fusion multi-resolution (MDF) a data We novel,We introduce a novel multi-resolution data fusion (MDF) method for super-resolution computational EM.,21798
performance achieved on been State-of-the-art datasets. all has benchmark,State-of-the-art performance has been achieved on all benchmark datasets.,21799
"state-of-the-art this their already algorithms, used good several performance. in further with When increases conjunction","When used in conjunction with several state-of-the-art algorithms, this further increases their already good performance.",21800
within types to a This day work night and times. recognize technique vehicle proposes feature-based,This work proposes a feature-based technique to recognize vehicle types within day and night times.,21801
time Thermal for used were experiments. the images night,Thermal images were used for the night time experiments.,21802
applications well image-based prediction examine for We manipulation. shape also as as,We also examine applications for image-based prediction as well as shape manipulation.,21803
a is to R \texttt{rCOSA} interfaced language. software the package,\texttt{rCOSA} is a software package interfaced to the R language.,21804
on attributes objects implements data. multivariate clustering subsets for It statistical techniques in of,It implements statistical techniques for clustering objects on subsets of attributes in multivariate data.,21805
"up-to-date clear an guidance we a technique. and for advanced software package this introduce Here,","Here, we introduce an up-to-date software package and a clear guidance for this advanced technique.",21806
"problem, attention . cocktail auditory BCI, classification Keywords: party auditory","Keywords: auditory BCI, cocktail party problem, auditory attention classification .",21807
the to vision importance implement Time-efficient Semantic link of central the of is discovery Web.,Time-efficient link discovery is of central importance to implement the vision of the Semantic Web.,21808
"works, to models have used been newer linear the planners. fastest the runtime In estimate","In newer works, linear models have been used to estimate the runtime the fastest planners.",21809
"models has of for studied been far. this purpose other no category so However,","However, no other category of models has been studied for this purpose so far.",21810
"for non-linear estimation study In this we runtime runtime estimation. functions paper,","In this paper, we study non-linear runtime estimation functions for runtime estimation.",21811
understanding actions. is their when have for people crucial false Recognizing beliefs,Recognizing when people have false beliefs is crucial for understanding their actions.,21812
abstract incorrect identifying when We have of the problem in scenes people introduce novel beliefs.,We introduce the novel problem of identifying when people in abstract scenes have incorrect beliefs.,21813
control a constitutes of resistance Artemisinin major to programs success for threat the continued malaria.,Artemisinin resistance constitutes a major threat to the continued success of control programs for malaria.,21814
geometric type cloud Point structure. is an important data of,Point cloud is an important type of geometric data structure.,21815
"data causes and This, however, issues. renders voluminous unnecessarily","This, however, renders data unnecessarily voluminous and causes issues.",21816
"and simple, efficient is Though effective. PointNet highly","Though simple, PointNet is highly efficient and effective.",21817
"of the Empirically, art. state than or par performance even better on it strong shows","Empirically, it shows strong performance on par or even better than state of the art.",21818
"such, user-generated advocate the large-scale in As approach we systems. of data for harnessing","As such, we advocate for the approach of harnessing user-generated data in large-scale systems.",21819
movement common the Essential neurological is disorder. Tremor most,Essential Tremor is the most common neurological movement disorder.,21820
as become disorder may also Medication the ineffective progresses.,Medication may also become ineffective as the disorder progresses.,21821
usually a creation requires a camera carefully Cinemagraph scene. static a of configured The in,The creation of a Cinemagraph usually requires a static camera in a carefully configured scene.,21822
camera regular with moving becomes and for challenging objects. a video The a task,The task becomes challenging for a regular video with a moving camera and objects.,21823
into warps camera. of the Our reference input approach first viewpoint an video a,Our approach first warps an input video into the viewpoint of a reference camera.,21824
"sequence Lastly, techniques produce Cinemagraph the video processing to algorithm applies movie. a a of","Lastly, the algorithm applies a sequence of video processing techniques to produce a Cinemagraph movie.",21825
numerous scenes. approach challenging on have tested the real proposed We,We have tested the proposed approach on numerous challenging real scenes.,21826
"flat cannot from they well example, discriminate ones. For blurred regions","For example, they cannot well discriminate flat regions from blurred ones.",21827
we propose blur. is critical identifying information Here local in successfully semantic high-level that,Here we propose that high-level semantic information is critical in successfully identifying local blur.,21828
frames the occluded video. in of We the formation captured the model,We model the formation of the occluded frames in the captured video.,21829
internal steps. The proposed consists of method three,The proposed method consists of three internal steps.,21830
"H&E colors is of in unmixing First, used preprocessing the step.","First, unmixing colors of H&E is used in the preprocessing step.",21831
requires image to Medical expert segmentations be segmentation ground multiple derived truth annotations. consensus from,Medical image segmentation requires consensus ground truth segmentations to be derived from multiple expert annotations.,21832
getting in minima. of pose techniques the trapped risk local Such,Such techniques pose the risk of getting trapped in local minima.,21833
missing global to considering SSL by and features predict used consistency. annotations is image local,SSL is used to predict missing annotations by considering global features and local image consistency.,21834
without procedure. maximum Graph an cut a obtains global iterative,Graph cut obtains a global maximum without an iterative procedure.,21835
exact linear The constraints SfM in our formulation. relationships detected line-based geometric provide,The detected relationships provide exact geometric constraints in our line-based linear SfM formulation.,21836
"outperforming reconstruction datasets, validated state-of-the-art techniques. our We have on various challenging algorithm","We have validated our algorithm on challenging datasets, outperforming various state-of-the-art reconstruction techniques.",21837
automatic This detection techniques provides an surveillance event article overview . of,This article provides an overview of automatic surveillance event detection techniques .,21838
a key communities. mechanism expansion for Migration is of,Migration is a key mechanism for expansion of communities.,21839
each fitness Knowledge of determines individual. the environments about the,Knowledge about the environments determines the fitness of each individual.,21840
Communication within population. between knowledge to can knowledge the beneficial spread of individuals help,Communication of knowledge between individuals can help to spread beneficial knowledge within the population.,21841
results human patterns of gene flow. with for of consistent are The genetic divergence asymmetry,The results for asymmetry of genetic divergence are consistent with patterns of human gene flow.,21842
"two Dirichlet However, of problems. have standard formulations Latent Allocation supervised","However, standard formulations of supervised Latent Dirichlet Allocation have two problems.",21843
scalable to both a This addressing truly solution re-id proposes problems. by work,This work proposes a truly scalable solution to re-id by addressing both problems.,21844
"dubbed of is a Swarm method. Fusion, natural the Move framework, The Fusion generalization inference","The inference framework, dubbed Swarm Fusion, is a natural generalization of the Fusion Move method.",21845
case of multi-threading maintains and thread a solution. environments) (in Every a updates,Every thread (in a case of multi-threading environments) maintains and updates a solution.,21846
to approach We (IQA). network-based deep neural quality a assessment present image,We present a deep neural network-based approach to image quality assessment (IQA).,21847
"superpixel low-level in vision. available By publicly have into algorithms today, tools standard turned","By today, publicly available superpixel algorithms have turned into standard tools in low-level vision.",21848
"explicitly discuss optimization this we parameter the of enforcing and connectivity. strictly end, To importance","To this end, we explicitly discuss parameter optimization and the importance of strictly enforcing connectivity.",21849
case both This out of propositions. chapter the lays for these,This chapter lays out the case for both of these propositions.,21850
by bistability phosphorylation is this energy. shown Previous has work governed that,Previous work has shown that this bistability is governed by phosphorylation energy.,21851
camera matching Feature problems. scalability from suffers localization point for,Feature point matching for camera localization suffers from scalability problems.,21852
"high recognition, accuracy. general Convolutional Deep object Networks Neural On achieve (DCNNs)","On general object recognition, Deep Convolutional Neural Networks (DCNNs) achieve high accuracy.",21853
"ResNet its improvements particular, In rate lowest error the records. broken have and","In particular, ResNet and its improvements have broken the lowest error rate records.",21854
"learning, used. reinforcement current the state and widely function action is In on reward","In reinforcement learning, the reward function on current state and action is widely used.",21855
"total result be distribution, if However, involves reward wrong. the objective the the will","However, if the objective involves the total reward distribution, the result will be wrong.",21856
The numbers in circuits living in functional enormous are any independently motifs cell. found,The motifs are independently functional circuits found in enormous numbers in any living cell.,21857
adaptation II The in but performs efficiently information very in poor ICFFL processing.,The ICFFL II performs very efficiently in adaptation but poor in information processing.,21858
under pressure Which them the the of among natural is fittest selection?,Which is the fittest among them under the pressure of natural selection?,21859
"I puzzle, from sort take efficiency. To this the Pareto help out multi-objective","To sort out this puzzle, I take help from the multi-objective Pareto efficiency.",21860
inevitable. ranked of becomes re-id model results Human verification automatic,Human verification of automatic model ranked re-id results becomes inevitable.,21861
"nature, (constraints) influences i.e. in The social are quadratic","The social influences (constraints) are quadratic in nature, i.e.",21862
"between occur through of individuals, and collision avoidance. and pairs modeled grouping","occur between pairs of individuals, and modeled through grouping and collision avoidance.",21863
framework solving an constraints for into these incorporate re-identification. energy human We minimization,We incorporate these constraints into an energy minimization framework for solving human re-identification.,21864
"for single multiple grey-level colorizations an with viable a image. ambiguous Colorization is problem,","Colorization is an ambiguous problem, with multiple viable colorizations for a single grey-level image.",21865
"only most However, single probable methods produce the previous colorization.","However, previous methods only produce the single most probable colorization.",21866
embedding learn fields a variational of (VAE). using low color a dimensional autoencoder We,We learn a low dimensional embedding of color fields using a variational autoencoder (VAE).,21867
colorization. result from conditional model diverse this Samples in,Samples from this conditional model result in diverse colorization.,21868
in Detecting people a problem. is challenging images,Detecting people in images is a challenging problem.,21869
and show variation These less measurements considerably more meaningful provide information.,These measurements show considerably less variation and provide more meaningful information.,21870
calibration to provide We a and introduction a used sensor technology method. the brief propose,We provide a brief introduction to the sensor technology used and propose a calibration method.,21871
to The are in camera a detect people data fixed used efficiently resulting fused set-up.,The resulting fused data are used to efficiently detect people in a fixed camera set-up.,21872
person tracking. to include The is system extended,The system is extended to include person tracking.,21873
"detecting simultaneously we paper, multiple photographic of problem introduce the defects. this In","In this paper, we introduce the problem of simultaneously detecting multiple photographic defects.",21874
model than better user our from performs an Our also study. average human,Our model also performs better than an average human from our user study.,21875
revisited has context Stereo within images from the recently deep rectified reconstruction learning. been of,Stereo reconstruction from rectified images has recently been revisited within the context of deep learning.,21876
"detector accurate this we convolutional paper, propose features edge In (RCF). an richer using","In this paper, we propose an accurate edge detector using richer convolutional features (RCF).",21877
convolutional gradually receptive become features fields And coarser the increasing. with,And the convolutional features gradually become coarser with receptive fields increasing.,21878
is first in adopt features tasks. such vision the computer It to rich convolutional attempt,It is the first attempt to adopt such rich convolutional features in computer vision tasks.,21879
GC codon contents and their utilization patterns). in similarity,similarity in their GC contents and codon utilization patterns).,21880
images. in prostate challenges Boundary to ultrasound incompleteness raises automatic great segmentation,Boundary incompleteness raises great challenges to automatic prostate segmentation in ultrasound images.,21881
"with In we attempt paper, framework. address this issues novel to those a","In this paper, we attempt to address those issues with a novel framework.",21882
Our is of composed framework key three modules.,Our framework is composed of three key modules.,21883
module is be corroborated boundary in dealing incompleteness. to the This with effective,This module is corroborated to be effective in dealing with the boundary incompleteness.,21884
"that loss we mimics coarse-to-fine multi-scale Together, approaches. conventional present function","Together, we present multi-scale loss function that mimics conventional coarse-to-fine approaches.",21885
pervade their some these robust behaviour. of fragile architecture and of webs to The seems,The architecture of these webs seems to pervade some of their robust and fragile behaviour.,21886
rest? where sports photos favorite team blends with player of those your the Or one,Or one of those team sports photos where your favorite player blends with the rest?,21887
the considers and image. of internal approach an color that the properties propose saliency We,We propose an approach that considers the internal color and saliency properties of the image.,21888
"decluttering. Applications and include object attenuation distractors enhancement, background","Applications include object enhancement, distractors attenuation and background decluttering.",21889
prove only single-brood model we equilibrium the hybrid equilibrium. the the that Within stable is,Within the hybrid model we prove that the single-brood equilibrium is the only stable equilibrium.,21890
approach. Expectation-Maximization uses and adopts (CNNs) Our method convolutional deep neural networks an (EM) based,Our method uses deep convolutional neural networks (CNNs) and adopts an Expectation-Maximization (EM) based approach.,21891
argument this visual and We empirical support with results.,We support this argument with empirical and visual results.,21892
"performance an In bin random investigated. industrial this context their picking is paper, in","In this paper, their performance in an industrial random bin picking context is investigated.",21893
representative generate new method to proposed. datasets is synthetic A,A new method to generate representative synthetic datasets is proposed.,21894
discuss in their provide and proposed weaknesses. of solutions and literature an overview We strengths,We provide an overview of solutions proposed in literature and discuss their strengths and weaknesses.,21895
demonstrated estimation cast The flow as FlowNet a be problem. can that optical learning,The FlowNet demonstrated that optical flow estimation can be cast as a learning problem.,21896
"Particularly with and on data, small variational methods. compete cannot real-world displacements FlowNet","Particularly on small displacements and real-world data, FlowNet cannot compete with variational methods.",21897
"sub-network by motions. a Third, specializing small elaborate small introducing on on we displacements","Third, we elaborate on small displacements by introducing a sub-network specializing on small motions.",21898
"while methods, frame rates. running at with It on par performs state-of-the-art interactive","It performs on par with state-of-the-art methods, while running at interactive frame rates.",21899
"approach propose of we this fix paper, alternative an the In to SLIC. limitations inherent","In this paper, we propose an alternative approach to fix the inherent limitations of SLIC.",21900
"high perform also rate. jointly update allowing the convergence We assignment steps, and","We also jointly perform the assignment and update steps, allowing high convergence rate.",21901
Our the the define twofold: fleet. cheapest routing and adequate most goal is,Our goal is twofold: define the cheapest routing and the most adequate fleet.,21902
"each two Programming component. Mixed formulations we First, Integer for (MIP) define","First, we define two Mixed Integer Programming (MIP) formulations for each component.",21903
form and symmetry constraints. several breaking valid improvements cuts in we discuss the of Then,Then we discuss several improvements in the form of valid cuts and symmetry breaking constraints.,21904
with experiments. strengths weaknesses their We highlight and extensive,We highlight their strengths and weaknesses with extensive experiments.,21905
"a explore Constraint integration (CP). with we lightweight Finally, Programming","Finally, we explore a lightweight integration with Constraint Programming (CP).",21906
of and set rough modal logics: survey. systems A Interface,Interface of rough set systems and modal logics: A survey.,21907
injection an in (NI) Noise (NNs). over-fitting neural networks technique efficient mitigate is to,Noise injection (NI) is an efficient technique to mitigate over-fitting in neural networks (NNs).,21908
suited preference elicitation for we approach propose this paper In to this scenario. an,In this paper we propose an approach to preference elicitation suited for this scenario.,21909
"collects example query We extend manipulative Coactive to which Learning, critiques. optionally feedback, iteratively","We extend Coactive Learning, which iteratively collects manipulative feedback, to optionally query example critiques.",21910
the space. model integrated into by the extending User critiques learning are feature dynamically,User critiques are integrated into the learning model by dynamically extending the feature space.,21911
"tasks, is formulation catalogue learning supports on-the-fly. option constructive generated Our natively where the","Our formulation natively supports constructive learning tasks, where the option catalogue is generated on-the-fly.",21912
average regret We present bound the the learner. suffered an on by upper,We present an upper bound on the average regret suffered by the learner.,21913
of analysis empirical the highlights Our our approach. promise,Our empirical analysis highlights the promise of our approach.,21914
(CNN) Convolutional pre-trained ImageNet backbone are neural approaches. most of networks the on state-of-the-art,Convolutional neural networks (CNN) pre-trained on ImageNet are the backbone of most state-of-the-art approaches.,21915
with All outperform previous same architecture. models the models,All models outperform previous models with the same architecture.,21916
to decomposition. obtain data a multi-layer hard to It is learn,It is hard to obtain data to learn a multi-layer decomposition.,21917
"method the generates each from layers, model, that Our one image. explain then image","Our method then generates image layers, one from each model, that explain the image.",21918
images. innovation approach on in models Our for generative rests,Our approach rests on innovation in generative models for images.,21919
"does is The interpretation. admit require that general, a layers and not approach physical","The approach is general, and does not require that layers admit a physical interpretation.",21920
boundary detection We that segmentation improves semantic show CNNs. significantly with,We show that boundary detection significantly improves semantic segmentation with CNNs.,21921
"organization and gene Important discussed. are structure expression, implications for chromatin genome","Important implications for gene expression, chromatin structure and genome organization are discussed.",21922
"time-of-flight infrared low-cost provides range opportunities. and Combining thermal new new, sensors","Combining new, low-cost thermal infrared and time-of-flight range sensors provides new opportunities.",21923
two A sensors distinction that significantly. of first spectral the sensitivity the is differs,A first distinction is that the spectral sensitivity of the two sensors differs significantly.,21924
"all. range In there is at fact, overlap no sensitivity","In fact, there is no sensitivity range overlap at all.",21925
"their second that requires is special A distinction very is attention. low, resolution typically which","A second distinction is that their resolution is typically very low, which requires special attention.",21926
"hence, to in need medical uncertainties models. Such considered be computerized","Such uncertainties hence, need to be considered in computerized medical models.",21927
"more recent computation. particular, incorporated and of have this more In versions course","In particular, recent versions of this course have incorporated more and more computation.",21928
"learned representation classifier are the encoding The simultaneously. the all and dictionaries for features,","The features, dictionaries and the encoding representation for the classifier are all learned simultaneously.",21929
recognition. texture orderless particularly is and and useful The for therefore representation is material,The representation is orderless and therefore is particularly useful for material and texture recognition.,21930
for The the publicly code are source system available. complete,The source code for the complete system are publicly available.,21931
"includes posterior strategies. the distribution elicitation, summaries of This computational prior and","This includes prior elicitation, summaries of the posterior distribution and computational strategies.",21932
"stage. at an compensated errors can and be way, early detected rapidly this In possibly","In this way, errors can be rapidly detected and possibly compensated at an early stage.",21933
"information. qualitative system quantitative and implemented both Thus, to able provide is our","Thus, our implemented system is able to provide both quantitative and qualitative information.",21934
global both adaptation specific Our of techniques. and consists category method,Our method consists of both global and category specific adaptation techniques.,21935
large much scale retrieval. Hashing attracted attention methods for image have,Hashing methods have attracted much attention for large scale image retrieval.,21936
"equally. hashing existing deep methods treat all However, hash bits","However, existing deep hashing methods treat all hash bits equally.",21937
for parameter regularization maximum anisotropic on focus denoising. total-variation the We,We focus on the maximum regularization parameter for anisotropic total-variation denoising.,21938
a motion as formulate learning structure In this we problem. paper from,In this paper we formulate structure from motion as a learning problem.,21939
spatial component of relative A crucial the loss based a on is differences. training approach,A crucial component of the approach is a training loss based on spatial relative differences.,21940
"traditional motion and two-frame robust. to more accurate structure more methods, results from Compared are","Compared to traditional two-frame structure from motion methods, results are more accurate and more robust.",21941
systems. of valuation-based models in by Jirousek Compositional were introduce the framework general and Shenoy,Compositional models were introduce by Jirousek and Shenoy in the general framework of valuation-based systems.,21942
electric remains active issue. in weakly a electrolocation fish challenging Understanding,Understanding active electrolocation in weakly electric fish remains a challenging issue.,21943
The large spaces. been five has proposed evaluated challenging indoor on approach,The proposed approach has been evaluated on five challenging large indoor spaces.,21944
"three We annotation. of different obtain independently results competitive type datasets, the on from input","We obtain competitive results on three different datasets, independently from the type of input annotation.",21945
methods. critical the are to of Hierarchical architectures learning scalability reinforcement,Hierarchical architectures are critical to the scalability of reinforcement learning methods.,21946
"of execute primitive sequences comprising actions hierarchical actions. frameworks macro-actions Current with serially,","Current hierarchical frameworks execute actions serially, with macro-actions comprising sequences of primitive actions.",21947
in terms of other significantly factors. These architectures CNN differ and works,These works differ significantly in terms of CNN architectures and other factors.,21948
"factors on impact unclear. performance of Based these is results the alone, the reported","Based on the reported results alone, the performance impact of these factors is unclear.",21949
"for field. research this consequently advancing basis, bottlenecks we this identify On directions existing and","On this basis, we identify existing bottlenecks and consequently directions for advancing this research field.",21950
this low-cost system. imaging paper we scan In high-speed line a propose,In this paper we propose a low-cost high-speed imaging line scan system.,21951
"a closely real-life lab, method in the our cases. we applied representing Finally,","Finally, we applied our method in a lab, closely representing the real-life cases.",21952
understanding. image for visual similarity Measuring critical is,Measuring visual similarity is critical for image understanding.,21953
"are reason similar However, much more complex. images why the is","However, the reason why images are similar is much more complex.",21954
"this on concept propose visual we similarity. of Based the contextual observation,","Based on this observation, we propose the concept of contextual visual similarity.",21955
vision few arisen future A with for is questions discussion. illustrated,A future vision is illustrated with few questions arisen for discussion.,21956
a estimation. and of introduction flow survey This scene analysis general presents,This survey presents a general introduction and analysis of scene flow estimation.,21957
"information we complex of with explosive witnessed content. and images have the Recently, growth","Recently, we have witnessed the explosive growth of images with complex information and content.",21958
verified our of general numerical has The experiment methodology. feasibility,The numerical experiment has verified general feasibility of our methodology.,21959
other state-of-the-art the enhance with overall algorithms. compared the accuracy We,We enhance the overall accuracy compared with the other state-of-the-art algorithms.,21960
ensemble An into for learning of is region. cyst region detected methodRandom classification learnt forests,An ensemble learning methodRandom forests is learnt for classification of detected region into cyst region.,21961
detection a The unified and achieves in segmentation setting. method,The method achieves detection and segmentation in a unified setting.,21962
"filters. compute local information using curvilinear we To steerable oriented features, extract gradient","To extract local curvilinear features, we compute oriented gradient information using steerable filters.",21963
an objects of labels in image-level localizes by interest learning manner. end-to-end It categorical from,It localizes objects of interest by learning from image-level categorical labels in an end-to-end manner.,21964
step. second work on the focuses This,This work focuses on the second step.,21965
"existing e.g. hypotheses all While via pool reasoning, approaches generate local the","While all existing approaches generate the hypotheses pool via local reasoning, e.g.",21966
in We impact the the number achieving empirically also of this demonstrate goal. of views,We also empirically demonstrate the impact of the number of views in achieving this goal.,21967
(FSI) capable reconstructing and proven high-quality images. three-dimensional two-dimensional of has single-pixel Fourier imaging,Fourier single-pixel imaging (FSI) has proven capable of reconstructing high-quality two-dimensional and three-dimensional images.,21968
a are which a black-and-white binary device generate only DMDs pattern instance. each can at,DMDs are a binary device which can only generate a black-and-white pattern at each instance.,21969
allows high-speed imaging proposed The technique high-quality broad potentially wavebands. and over,The proposed technique potentially allows high-quality and high-speed imaging over broad wavebands.,21970
in detecting at Feature systems objects pyramids different component recognition a basic are for scales.,Feature pyramids are a basic component in recognition systems for detecting objects at different scales.,21971
a meaningful from strategy introduce to imagery. for semantically extract features learning novel aerial We,We introduce a novel strategy for learning to extract semantically meaningful features from aerial imagery.,21972
"properties be can analytic the real all Accordingly, in exploited field. complex","Accordingly, all real analytic properties can be exploited in the complex field.",21973
this elucidated clearly The effectiveness will in of framework in later sections this be paper.,The effectiveness of this framework will be clearly elucidated in later sections in this paper.,21974
an RNA The origin of life is of research world. paradigm in that dominant,The dominant paradigm in origin of life research is that of an RNA world.,21975
multiple Managing drugs. often results in multimorbidity of with prescription the patients polypharmacy:,Managing patients with multimorbidity often results in polypharmacy: the prescription of multiple drugs.,21976
"of typically and combinations unknown. of the However, are specific diseases drugs effects long-term","However, the long-term effects of specific combinations of drugs and diseases are typically unknown.",21977
"condition the result may one effects in drugs for for prescribed other. In particular, adverse","In particular, drugs prescribed for one condition may result in adverse effects for the other.",21978
"drug of Moreover, affect one diseases group. disease group for may a another treatment","Moreover, a drug treatment for one disease group may affect diseases of another group.",21979
"for essential related many e.g. tasks, detection is hand Hand","Hand detection is essential for many hand related tasks, e.g.",21980
"which robotics parsing gesture, for and understanding hand pose, are human-computer extremely useful interaction.","parsing hand pose, understanding gesture, which are extremely useful for robotics and human-computer interaction.",21981
estimation mutually other. can each rotation We that show benefit further classification and,We further show that rotation estimation and classification can mutually benefit each other.,21982
Removing problem. challenging the to pixel-wise nature due blur heterogeneous of is motion ill-posed the,Removing pixel-wise heterogeneous motion blur is challenging due to the ill-posed nature of the problem.,21983
image latent image prior all content. Learning possible over a require would modeling the,Learning a prior over the latent image would require modeling all possible image content.,21984
the on that demonstrate the proposed experiments outperforms state-of-the-art. challenging method blurred realistic Extensive images,Extensive experiments on challenging realistic blurred images demonstrate that the proposed method outperforms the state-of-the-art.,21985
error intended reducing is algorithms. of a the device prediction for learning Bagging,Bagging is a device intended for reducing the prediction error of learning algorithms.,21986
or This rough holds even is raw functional the unstable. if,This holds even if the raw functional is rough or unstable.,21987
Two-Person dataset used was The of for the methodology. interaction this development,The Two-Person interaction dataset was used for the development of this methodology.,21988
"the We via problem allowing deeper address kernels, small architectures.","We address the problem via small kernels, allowing deeper architectures.",21989
performed available two publicly datasets. comprehensive over experiments We,We performed comprehensive experiments over two publicly available datasets.,21990
"dataset. a state-of-the-art demonstrate we performance ISBR First, on the","First, we demonstrate a state-of-the-art performance on the ISBR dataset.",21991
for makes neuroanatomical studies. This massive convenient it imaging multi-site,This makes it convenient for massive multi-site neuroanatomical imaging studies.,21992
"propagate map their geodesic distances activation we for using co-localization. superpixel combined Then,","Then, we propagate their combined activation map using superpixel geodesic distances for co-localization.",21993
mediated benefit and co-alleviation is stress. This of physiologic by toxic,This benefit is mediated by co-alleviation of toxic and physiologic stress.,21994
of modelling systems. conclusions The types are and diverse holobiont applicable framework to,The conclusions and modelling framework are applicable to diverse types of holobiont systems.,21995
on the the flow optical improvement KITTI demonstrate estimation in We real-world benchmark.,We demonstrate the improvement in optical flow estimation on the real-world KITTI benchmark.,21996
"flow in Additionally, test we the classification indirectly scenario. an optical action","Additionally, we test the optical flow indirectly in an action classification scenario.",21997
this model is strings used to elementary random-walk on test hypothesis. An binary,An elementary random-walk model on binary strings is used to test this hypothesis.,21998
human plays Attention visual role critical experience. a in,Attention plays a critical role in human visual experience.,21999
"paper, propose position dynamic images for this scenes. algorithm automatic an of we de-fencing In","In this position paper, we propose an automatic de-fencing algorithm for images of dynamic scenes.",22000
"of large numbers challenging. rounds, remains elimination that infeasibly with We treatment even find","We find that even with infeasibly large numbers of treatment rounds, elimination remains challenging.",22001
"with well-defined classes Semantic e.g. a either things shape, (objects can be","Semantic classes can be either things (objects with a well-defined shape, e.g.",22002
"e.g. car, (amorphous person) regions, or background stuff","car, person) or stuff (amorphous background regions, e.g.",22003
the appears problems computational of wide optimization in Manifold in a applied variety sciences.,Manifold optimization appears in a wide variety of computational problems in the applied sciences.,22004
"the ROPTLIB. C++ that package article an This ManifoldOptim, R library wraps presents","This article presents ManifoldOptim, an R package that wraps the C++ library ROPTLIB.",22005
theories sensory of it processing suggest Several whitens stimuli. early that sensory,Several theories of early sensory processing suggest that it whitens sensory stimuli.,22006
"decisions however, non-differentiable. computational allocation are of to resources This candidates, allows such promising","This allows allocation of computational resources to promising candidates, however, such decisions are non-differentiable.",22007
"train to are As an in end-to-end algorithms a fashion. these hard result,","As a result, these algorithms are hard to train in an end-to-end fashion.",22008
"computation that Additionally, efficient algorithm training we reduces present dramatically an time.","Additionally, we present an efficient training algorithm that dramatically reduces computation time.",22009
of in have the images recently learning diagnosis Computer-aided gained increasing machine community. attention retinal,Computer-aided diagnosis of retinal images have recently gained increasing attention in the machine learning community.,22010
uncertainty contour computation integrals. of methods and maps contains of for Gaussian It also quantification,It also contains methods for uncertainty quantification of contour maps and computation of Gaussian integrals.,22011
in the computational the methods used package. theoretical describes This and article,This article describes the theoretical and computational methods used in the package.,22012
USED al. provided by dataset Ahmad et,USED dataset provided by Ahmad et al.,22013
"time-consuming, be administered are tests These must professional. subjective, and a by","These tests are subjective, time-consuming, and must be administered by a professional.",22014
"the As result, primary goal (i.e. a","As a result, the primary goal (i.e.",22015
using different methods and description. A is detection keypoint for proposed approach hybrid,A hybrid approach is proposed using different methods for keypoint detection and description.,22016
"training into taking consideration. the design lymphocytes a new Also, prior scheme about knowledge we","Also, we design a new training scheme taking the prior knowledge about lymphocytes into consideration.",22017
ontology and hard an task. is time-consuming Modeling a,Modeling an ontology is a hard and time-consuming task.,22018
"subjective. very the Moreover, still is evaluation","Moreover, the evaluation is still very subjective.",22019
ontologists of ontologies the domain. and using evaluated same The methodology empirically different was,The methodology was empirically evaluated using different ontologists and ontologies of the same domain.,22020
is for heterogeneous managing Ontohub engine repository distributed a ontologies.,Ontohub is a repository engine for managing distributed heterogeneous ontologies.,22021
to enables communities exchange contributions The distributed share their and easily. nature,The distributed nature enables communities to share and exchange their contributions easily.,22022
integrate nature ontologies it in various languages. heterogeneous to possible makes The ontology written,The heterogeneous nature makes it possible to integrate ontologies written in various ontology languages.,22023
The produces. body and all movements the for human thoughts brain ultimately responsible is that,The human brain is ultimately responsible for all thoughts and movements that the body produces.,22024
This environment. their successfully humans to interact with allows,This allows humans to successfully interact with their environment.,22025
human the not can functioning is be abilities damaged. If properly brain many of,If the brain is not functioning properly many abilities of human can be damaged.,22026
analysis brain learn cerebral of function. The goal is to about signal,The goal of cerebral signal analysis is to learn about brain function.,22027
to their synchronize firing requires This timing assembly rates. specific an neurons within,This specific timing requires neurons within an assembly to synchronize their firing rates.,22028
providing inference Bayesian present BayesBD package of boundaries for the images. noisy We,We present the BayesBD package providing Bayesian inference for boundaries of noisy images.,22029
that interactive The lens an supports EditLens the is of graphs. technique editing,The EditLens is an interactive lens technique that supports the editing of graphs.,22030
satisfy of fully Existing the EditLens. needs cannot the algorithms,Existing algorithms cannot fully satisfy the needs of the EditLens.,22031
"algorithm is that, indicate than in the alternative general, Tests better solutions.","Tests indicate that, in general, the algorithm is better than alternative solutions.",22032
"Each also layout textures. has lighting, camera and trajectories, random","Each layout also has random lighting, camera trajectories, and textures.",22033
is thumbnail neural automatic based system deep on an networks. generation Fast-AT,Fast-AT is an automatic thumbnail generation system based on deep neural networks.,22034
"depending appropriate the of inference, the the thumbnail. During target on filter dimensions is selected","During inference, the appropriate filter is selected depending on the dimensions of the target thumbnail.",22035
"but previous utilize saliency Unlike the most not directly. work, Fast-AT does addresses problem","Unlike most previous work, Fast-AT does not utilize saliency but addresses the problem directly.",22036
"it to map. addition, the saliency need search the In conduct on region eliminates","In addition, it eliminates the need to conduct region search on the saliency map.",22037
to results show We in techniques. existing comparison competitive,We show competitive results in comparison to existing techniques.,22038
defined stochastic spheres and are objects processes cones. Many geometrical special like on,Many stochastic processes are defined on special geometrical objects like spheres and cones.,22039
"tools harmonic We analysis, from describe how i.e.","We describe how tools from harmonic analysis, i.e.",22040
these processing. their and We will an signal applications overview in over properties give,We will give an overview over these properties and their applications in signal processing.,22041
illustrate two will color image the theory processing. vision and low-level from with examples We,We will illustrate the theory with two examples from low-level vision and color image processing.,22042
"clustering a non-parametric paper, this novel technique. we present In","In this paper, we present a novel non-parametric clustering technique.",22043
for is robotics A critical reliable stereo algorithm many applications.,A reliable stereo algorithm is critical for many robotics applications.,22044
easily and by failure can specular textureless But regions cause feature making difficult. matching,But textureless and specular regions can easily cause failure by making feature matching difficult.,22045
Understanding important. to whether an robust is is hazardous these algorithm regions,Understanding whether an algorithm is robust to these hazardous regions is important.,22046
the with perceptual We also a compare models study.,We also compare the models with a perceptual study.,22047
Reconstruction is sensed an problem. ill-posed compressively signals of measurements from,Reconstruction of signals from compressively sensed measurements is an ill-posed problem.,22048
with projected using inference inputs perform to method. MAP back-propagation and We the RIDE gradient,We perform MAP inference with RIDE using back-propagation to the inputs and projected gradient method.,22049
an well. for images approach entropy in preserving based thresholding propose We texture,We propose an entropy thresholding based approach for preserving texture in images well.,22050
"on we logic. fuzzy this method a paper, based In propose novel","In this paper, we propose a novel method based on fuzzy logic.",22051
"developed newly objective study, describe that In we measures. real-time provides system this a CAD","In this study, we describe a newly developed CAD system that provides real-time objective measures.",22052
"the approaches literature. in conventional Moreover, surpass found the proposed algorithms most other","Moreover, the proposed approaches surpass most other conventional algorithms found in the literature.",22053
lattice. cunx proposed design on The quin- is,The design is proposed on quin- cunx lattice.,22054
have shown on images. been Simulation some results standard,Simulation results have been shown on some standard images.,22055
human Micro-facial deception. event can regarded an that are expressions behavioural emotional important highlight as,Micro-facial expressions are regarded as an important human behavioural event that can highlight emotional deception.,22056
"facial motion in distance local Chi-square we regions. the subtle to Then, find use","Then, we use Chi-square distance to find subtle facial motion in the local regions.",22057
coded and is on two validated datasets: performance FACS The SAMM II. CASME,The performance is validated on two FACS coded datasets: SAMM and CASME II.,22058
"runtime lower this this project average compared configuration, project's a best achieved to proxdd. Using","Using this project's best configuration, this project achieved a lower average runtime compared to proxdd.",22059
approach different train a detectors We take separate different for and scales.,We take a different approach and train separate detectors for different scales.,22060
demonstrate on datasets We and state-of-the-art results WIDER (FDDB massively-benchmarked face FACE).,We demonstrate state-of-the-art results on massively-benchmarked face datasets (FDDB and WIDER FACE).,22061
a play Video descriptors significant these in role systems.,Video descriptors play a significant role in these systems.,22062
The pixel the descriptor every computed is video. for in,The descriptor is computed for every pixel in the video.,22063
videos This good us find to correspondences across more efficiently. allow,This allow us to find good correspondences across videos more efficiently.,22064
"we and human into head, legs. hands In arms, body divide torso, parts: detail, five","In detail, we divide human body into five parts: head, torso, arms, hands and legs.",22065
combine models and with classifiers Search Tree both We Carlo test image Monte (e.g. shallow,We combine Monte Carlo Tree Search with image classifiers and test both shallow models (e.g.,22066
Neural multinomial and (e.g. deep logistic Networks regression) Convolutional,multinomial logistic regression) and deep Convolutional Neural Networks (e.g.,22067
of interest medical has the Spectral enormous imaging imaging modalities. field in received,Spectral imaging has received enormous interest in the field of medical imaging modalities.,22068
of tool and It the different provides powerful organs a non-invasive for tissues. analysis,It provides a powerful tool for the analysis of different organs and non-invasive tissues.,22069
"RGB to framework spectral from research, video. a generate propose this we In neurosurgery video","In this research, we propose a framework to generate neurosurgery spectral video from RGB video.",22070
frames. RGB applied A estimation video each technique spectral is on,A spectral estimation technique is applied on each RGB video frames.,22071
to A technique best is training set. used searching identify the,A searching technique is used to identify the best training set.,22072
indentify best method. Five techniques to different spectrum are experimented the estimation,Five different spectrum estimation techniques are experimented to indentify the best method.,22073
"multi-scale a single representation. However, pooled these features convolutional are into scale-invariant","However, these multi-scale convolutional features are pooled into a single scale-invariant representation.",22074
Our multi-scale explicitly the encodes during image approach encoding stage. information,Our approach encodes multi-scale information explicitly during the image encoding stage.,22075
We strategies in representation. multi-scale explicitly the image final to propose two encode information,We propose two strategies to encode multi-scale information explicitly in the final image representation.,22076
networks(BPNN). the strong of mapped make use of the neural We ability non-linearity back-propagation,We make use of the strong non-linearity mapped ability of the back-propagation neural networks(BPNN).,22077
sample Training method. by are got images undersampled,Training sample images are got by undersampled method.,22078
inputs pixels to elements the are chose BPNN as of The referred means(NL-Means). the Non-local,The elements chose as the inputs of the BPNN are pixels referred to Non-local means(NL-Means).,22079
efficient We heuristic search tackle devising by query issue this methods.,We tackle this issue by devising efficient heuristic query search methods.,22080
"methods that returned existing understandability, cardinality realize. of the or query cannot best minimal","minimal cardinality or best understandability, of the returned query that existing methods cannot realize.",22081
"system. this diagnosed is the) of of approach (complexity independent of the performance Hence, the","Hence, the performance of this approach is independent of the (complexity of the) diagnosed system.",22082
loss predictions. for We learning propose function reflectance a CNN of dense,We propose a loss function for CNN learning of dense reflectance predictions.,22083
This surpass. sets other only two a which baseline competitive approaches,This sets a competitive baseline which only two other approaches surpass.,22084
Our over-estimated the so suggest effect have that may learning-based far. been findings approaches of,Our findings suggest that the effect of learning-based approaches may have been over-estimated so far.,22085
"holds For its application analyses connectomic well. great reason, promise, as to this","For this reason, its application to connectomic analyses holds great promise, as well.",22086
We Process Bernoulli call a (BAR). this Autoregressive,We call this a Bernoulli Autoregressive Process (BAR).,22087
machine frames structure a causal task. learning estimation This as paper,This paper frames causal structure estimation as a machine learning task.,22088
may sub-optimal two-stage process produce This encoding.,This two-stage process may produce sub-optimal encoding.,22089
"SIFT binary approximation a Second, features. propose we to","Second, we propose a binary approximation to SIFT features.",22090
different that patterns. with the misclassified accuracy have is It may proved algorithms overall same,It is proved that algorithms with the same overall accuracy may have different misclassified patterns.,22091
ridge Tuning regression. critical parameter kernel of selection for importance is,Tuning parameter selection is of critical importance for kernel ridge regression.,22092
"exchangeable, traditionally sparsity. demonstrate vertex can edge-exchangeable models, are models exhibit unlike that We that","We demonstrate that edge-exchangeable models, unlike models that are traditionally vertex exchangeable, can exhibit sparsity.",22093
"in the vision matching. We graph also known study as problem, quadratic computer assignment","We study the quadratic assignment problem, in computer vision also known as graph matching.",22094
of Kernelized the one trackers. Correlation state-of-the-art Filter is object (KCF),Kernelized Correlation Filter (KCF) is one of the state-of-the-art object trackers.,22095
"large variations, expressions, occlusions, strong low facial resolutions, extremely etc. illumination","large occlusions, extremely low resolutions, facial expressions, strong illumination variations, etc.",22096
"The proposed face databases, challenging detection is benchmarked on i.e. approach two","The proposed approach is benchmarked on two challenging face detection databases, i.e.",22097
video propagates through technique information data. that We propose forward a,We propose a technique that propagates information forward through video data.,22098
manner. propose video We Network' processes an frames that Propagation adaptive a 'Video in,We propose a 'Video Propagation Network' that processes video frames in an adaptive manner.,22099
faces. description attributes are their human of to due detailed Face interesting,Face attributes are interesting due to their detailed description of human faces.,22100
"with the modification. In modest this way, can efficiently be pixel operated manipulation","In this way, the manipulation can be operated efficiently with modest pixel modification.",22101
our based Adversarial framework the is on Generative approach Network. of The,The framework of our approach is based on the Generative Adversarial Network.,22102
two networks consists of network. a discriminative transformation and image It,It consists of two image transformation networks and a discriminative network.,22103
each from to We learning to dual other. networks allow learn apply transformation also,We also apply dual learning to allow transformation networks to learn from each other.,22104
for learned Experiments can effectively attribute show manipulations. be that residual and images used,Experiments show that residual images can be effectively learned and used for attribute manipulations.,22105
remain generated The attribute-irrelevant of the in most details areas. images,The generated images remain most of the details in attribute-irrelevant areas.,22106
to evidence conflict theory. still issue in Dempster-Shafer manage an open is How,How to manage conflict is still an open issue in Dempster-Shafer evidence theory.,22107
"functions existing belief some However, have coefficients of shortcomings. correlation","However, existing correlation coefficients of belief functions have some shortcomings.",22108
"correlation proposed desirable coefficient this a many In paper, new is with properties.","In this paper, a new correlation coefficient is proposed with many desirable properties.",22109
functions. the conflict its of among degree measure to is belief applications One,One of its applications is to measure the conflict degree among belief functions.,22110
examples Some of effectiveness numerical correlation coefficient. and demonstrate comparisons the the,Some numerical examples and comparisons demonstrate the effectiveness of the correlation coefficient.,22111
examples These known to could either classes. belong already or undiscovered to yet,These examples could either belong to already known or to yet undiscovered classes.,22112
"Furthermore, yield active efficient show approximations that for selection. easy-to-implement techniques we","Furthermore, we show easy-to-implement approximations that yield efficient techniques for active selection.",22113
currently heuristics. that outperforms used experiments Empirical method show our,Empirical experiments show that our method outperforms currently used heuristics.,22114
"acquisition, effects side radiation patients. during may cause which exposes However, CT to","However, CT exposes radiation during acquisition, which may cause side effects to patients.",22115
"In this a this approach challenging we problem. data-driven to paper, propose address","In this paper, we propose a data-driven approach to address this challenging problem.",22116
"given MR a generate fully image. an train we network CT convolutional to Specifically,","Specifically, we train a fully convolutional network to generate CT given an MR image.",22117
implement We AutoContext context-aware adversarial Model further generative a to apply network.,We further apply AutoContext Model to implement a context-aware generative adversarial network.,22118
The to an shapes. system primate visual ability discriminate has partially exquisite occluded,The primate visual system has an exquisite ability to discriminate partially occluded shapes.,22119
inference probabilistic that We occurs steps. propose two in,We propose that probabilistic inference occurs in two steps.,22120
"datasets external as (e.g., OpenImages) large-scale takes method some bank. machine-annotated an Our memory","Our method takes some large-scale machine-annotated datasets (e.g., OpenImages) as an external memory bank.",22121
model. core new our is The method a of co-attention proposed,The core of our proposed method is a new co-attention model.,22122
appealing ground-truth data annotations where generated alternative One automatically. synthetic is are rendering,One appealing alternative is rendering synthetic data where ground-truth annotations are generated automatically.,22123
"often rendered on generalize models to real Unfortunately, fail images. purely to images trained","Unfortunately, models trained purely on rendered images often fail to generalize to real images.",22124
"training. object the to we that classes process demonstrate during generalizes Finally, unseen adaptation","Finally, we demonstrate that the adaptation process generalizes to object classes unseen during training.",22125
parameter free computationally is is Our and laborious of settings framework efficient.,Our framework is free of laborious parameter settings and is computationally efficient.,22126
for These difficult concentrate it alarms drivers to driving. false their can on make,These false alarms can make it difficult for drivers to concentrate on their driving.,22127
large from Motion key and Accuracy scale efficiency are two (SfM). problems incremental in Structure,Accuracy and efficiency are two key problems in large scale incremental Structure from Motion (SfM).,22128
"better higher speedup, that method significant Experiments our accuracy show and completeness. achieves","Experiments show that our method achieves significant speedup, higher accuracy and better completeness.",22129
REPPlab The projection pursuit. using sets R-package one-dimensional unsupervised data multivariate explore to is designed,The R-package REPPlab is designed to explore multivariate data sets using one-dimensional unsupervised projection pursuit.,22130
"in REPPlab exploration Following the into strategy original is EPPlab, divided software the two steps.","Following the original software EPPlab, the exploration strategy in REPPlab is divided into two steps.",22131
processing. vision and remained topic computer Regularization-based image active in an restoration research has image,Regularization-based image restoration has remained an active research topic in computer vision and image processing.,22132
guidance captured a additional an fields signal in as It leverages different often cue.,It often leverages a guidance signal captured in different fields as an additional cue.,22133
in steps from end-to-end a an large are Both dataset manner. learned,Both steps are learned from a large dataset in an end-to-end manner.,22134
"predict CNN the joint modeled the human Thus, a using activity can we which data.","Thus, we modeled a CNN which can predict the human activity using the joint data.",22135
makes simpler faster our and RNN models. the This than model,This makes our model simpler and faster than the RNN models.,22136
improve a methods three of that previously-developed CNN. propose We semi-supervised the performance,We propose three methods that improve the performance of a previously-developed semi-supervised CNN.,22137
memory. to (NMDARs) learning brain N-methyl-D-aspartate central in and glycoproteins receptors the are,N-methyl-D-aspartate receptors (NMDARs) are glycoproteins in the brain central to learning and memory.,22138
dynamics the The of on effects of and glycosylation structure largely are unknown. NMDARs,The effects of glycosylation on the structure and dynamics of NMDARs are largely unknown.,22139
"results NMDA glycans of suggest on our an intramolecular Overall, receptors. potentiating role","Overall, our results suggest an intramolecular potentiating role of glycans on NMDA receptors.",22140
surveillance have due Distributed to popular in become concerns. systems security recent years,Distributed surveillance systems have become popular in recent years due to security concerns.,22141
"dimensional challenge. systems in becomes high distributed transmitting data bandwidth-limited major However, a","However, transmitting high dimensional data in bandwidth-limited distributed systems becomes a major challenge.",22142
recognize along textures developed methods have many the To years. been,To recognize textures many methods have been developed along the years.,22143
are to used approach. evaluate our datasets Four,Four datasets are used to evaluate our approach.,22144
"orientation and occlusion, and of are invariance, clutter responses Statisics less tuning to similar. scale","Statisics of scale invariance, responses to clutter and occlusion, and orientation tuning are less similar.",22145
quite of object different. Statistics selectivity are,Statistics of object selectivity are quite different.,22146
"to remains comparisons in learning make subtle attributes yet visual valuable, Distinguishing differences is non-trivial.","Distinguishing subtle differences in attributes is valuable, yet learning to make visual comparisons remains non-trivial.",22147
propose images. to sparsity the supervision overcome generated synthetically problem We of via,We propose to overcome the sparsity of supervision problem via synthetically generated images.,22148
diary applications represent Food a tantalizing market.,Food diary applications represent a tantalizing market.,22149
"the cast to architectures Thus, are specific not the problem.","Thus, the architectures are not cast to the specific problem.",22150
architecture conducted experimental To proposed results evaluate our benchmark datasets. on we three have,To evaluate our proposed architecture we have conducted experimental results on three benchmark datasets.,22151
for is images. characteristic types Texture of many important an,Texture is an important characteristic for many types of images.,22152
"challenges some future presented. Finally, are for research","Finally, some future challenges for research are presented.",22153
correlations has shown tasks. EEG creative activity positive alpha and Previous performing between research,Previous research has shown positive correlations between EEG alpha activity and performing creative tasks.,22154
approach sampling Carlo The of the of may variance Monte simple outcomes sampled make inadequate.,The variance of sampled outcomes may make the simple approach of Monte Carlo sampling inadequate.,22155
an head eye tracking continuously display. Our mounted algorithm calibrates system for a,Our algorithm continuously calibrates an eye tracking system for a head mounted display.,22156
calibration nearly explicit an as step. is Accuracy as with good achieved,Accuracy is nearly as good as achieved with an explicit calibration step.,22157
feature extraction to the above attempt every We process. employed in describe measure,We attempt to describe every feature extraction measure employed in the above process.,22158
"nature musculoskeletal to robustness stability, interconnection injury. of movement, and facilitates voluntary The","The nature of musculoskeletal interconnection facilitates stability, voluntary movement, and robustness to injury.",22159
parameter estimation the chunked-and-averaged (CA) for estimator. divide-and-conquer A method is,A divide-and-conquer method for parameter estimation is the chunked-and-averaged (CA) estimator.,22160
estimators for identically CA have univariate been independent under sampling. parameters and distributed (IID) studied,CA estimators have been studied for univariate parameters under independent and identically distributed (IID) sampling.,22161
We sampling. estimators CA non-IID under the vector study of parameters and,We study the CA estimators of vector parameters and under non-IID sampling.,22162
"the existence The of assume equilibrium, i.e. feasible approaches a","The approaches assume the existence of a feasible equilibrium, i.e.",22163
"has However, been key not assumption tested. this","However, this key assumption has not been tested.",22164
"Robust a real-world computer vision visual is problem, applications. tracking many with challenging","Robust visual tracking is a challenging computer vision problem, with many real-world applications.",22165
"Color features, Most HOG existing Names. such appearance employ approaches as or hand-crafted","Most existing approaches employ hand-crafted appearance features, such as HOG or Color Names.",22166
"RGB have deep networks from features extracted applied been for successfully convolutional tracking. neural Recently,","Recently, deep RGB features extracted from convolutional neural networks have been successfully applied for tracking.",22167
"their features these Despite capture only appearance information. success,","Despite their success, these features only capture appearance information.",22168
"features show motion We deep information. RGB, and complementary deep that hand-crafted, contain further","We further show that hand-crafted, deep RGB, and deep motion features contain complementary information.",22169
model But and these? reason do we how about,But how do we model and reason about these?,22170
breakdown high-correlation This points to the causes leading data of backprop between algorithm.,This causes high-correlation between data points leading to breakdown of the backprop algorithm.,22171
"problem, address number works synthetic proposed of data. To a using this","To address this problem, a number of works proposed using synthetic data.",22172
"how of synthetic is is However, systematic missing. generated study a such data","However, a systematic study of how such synthetic data is generated is missing.",22173
as to set We tokenize space semantic discrete part the of propose states. a,We propose to tokenize the semantic space as a discrete set of part states.,22174
"problem. proposed During the are to numerous solve past algorithms this decade, hashing","During the past decade, numerous hashing algorithms are proposed to solve this problem.",22175
algorithm Every proposed outperform other methods. claims state-of-the-art hashing,Every proposed algorithm claims outperform other state-of-the-art hashing methods.,22176
to Microalgae measure counting used is biomass quantity.,Microalgae counting is used to measure biomass quantity.,22177
knowledge those the the quantity improves The of of of bioprocess quality elements applications.,The knowledge of the quantity of those elements improves the quality of bioprocess applications.,22178
publicly available for database download. in used is The this paper,The database used in this paper is publicly available for download.,22179
"time. in and that are space and both of ), far terms states future goal","), and future goal states that are far in terms of both space and time.",22180
blur illumination. by space-varying simultaneous images often motion uneven and Aerial are degraded,Aerial images are often degraded by space-varying motion blur and simultaneous uneven illumination.,22181
the is uneven Thereafter corrected illumination well.,Thereafter the uneven illumination is corrected well.,22182
are an complementary objective integrated into regularizers function. elegantly These,These complementary regularizers are elegantly integrated into an objective function.,22183
"experiments, stimulus. the explore of a given possible Using we space selectively perturbations closed-loop around","Using closed-loop experiments, we explore selectively the space of possible perturbations around a given stimulus.",22184
an important organisms. living in differentiation process is Cell,Cell differentiation is an important process in living organisms.,22185
have dynamics stochastic differentiation deterministic both components. The and,The differentiation dynamics have both deterministic and stochastic components.,22186
"also detection was Moreover, the the individual investigated. combining on performance systems impact of","Moreover, the impact of combining individual systems on the detection performance was also investigated.",22187
the detection potential the algorithms improve combining of performance. highlights to This,This highlights the potential of combining algorithms to improve the detection performance.,22188
set development further of systems. this for released We CAD additional nodules of,We released this set of additional nodules for further development of CAD systems.,22189
challenging patterns task learning We videos. by motion address in this,We address this challenging task by learning motion patterns in videos.,22190
"evaluate segmentation on achieving Berkeley state-of-the-art We the motion database, results. also","We also evaluate on the Berkeley motion segmentation database, achieving state-of-the-art results.",22191
recognition dictionaries. process The with which coupled is decoding character driven lexicon generally integrates a,The character recognition is generally coupled with a lexicon driven decoding process which integrates dictionaries.,22192
This technique. kinematics a inverse paper shoulder (IK) proposes,This paper proposes a shoulder inverse kinematics (IK) technique.,22193
"ribs, is comprised complex sternum, of four humerus, the clavicle, and joints. Shoulder scapula,","Shoulder complex is comprised of the sternum, clavicle, ribs, scapula, humerus, and four joints.",22194
"The motion pattern, rhythm. shows Scapulo such complex shoulder as humeral specific","The shoulder complex shows specific motion pattern, such as Scapulo humeral rhythm.",22195
motions generate animators IK method to allows easily. shoulder novice The shoulder natural,The shoulder IK method allows novice animators to generate natural shoulder motions easily.,22196
"of technique As quality this animation. a character the improves result,","As a result, this technique improves the quality of character animation.",22197
"a we greedy measure, quality algorithm. this Using downsampling propose","Using this quality measure, we propose a greedy downsampling algorithm.",22198
measurements. deeply addresses the compressive from of extracting work problem This learned directly features,This work addresses the problem of extracting deeply learned features directly from compressive measurements.,22199
area. no There this has been in work,There has been no work in this area.,22200
techniques These signal the to be first. reconstructed require,These techniques require the signal to be reconstructed first.,22201
"In deep all sensing of blind our the envisaged. proposed superiority can be cases, compressed","In all cases, the superiority of our proposed deep blind compressed sensing can be envisaged.",22202
"In edge particular, the construction. in nonnegative obtained seeking graph by low-rank a weights are","In particular, edge weights in the graph are obtained by seeking a nonnegative low-rank construction.",22203
Many state-of-the-art rely learning algorithms. machine supervised methods delineation on,Many state-of-the-art delineation methods rely on supervised machine learning algorithms.,22204
"training obtain. is data, a which As tedious result, to annotated they require manually","As a result, they require manually annotated training data, which is tedious to obtain.",22205
"classification final even the minor may significantly Furthermore, result. the affect errors topology of","Furthermore, even minor classification errors may significantly affect the topology of the final result.",22206
on demonstrate the We microscopy neurons. images effectiveness and of blood depicting our vessels approach,We demonstrate the effectiveness of our approach on microscopy images depicting blood vessels and neurons.,22207
improved Deep greatly recent recognition in visual learning has years.,Deep learning has greatly improved visual recognition in recent years.,22208
cascade to A designed adversarials. classifier detect efficiently was,A cascade classifier was designed to efficiently detect adversarials.,22209
Accurate fingertips interaction. for detection in critical depth human-computer image of is,Accurate detection of fingertips in depth image is critical for human-computer interaction.,22210
extracted forest. raw using is from depth Firstly edge image random image,Firstly edge image is extracted from raw depth image using random forest.,22211
combined depth the structure. information edge is in CNN with information Then our,Then the edge information is combined with depth information in our CNN structure.,22212
"different optical in of is, images. images from many non-optical that ways, Processing","Processing non-optical images is, in many ways, different from that of optical images.",22213
and different CNN We categories. test train on our object,We train and test our CNN on different object categories.,22214
"display random, oscillatory patterns. and dynamic and cell migration Cell directional polarization persistent can","Cell polarization and directional cell migration can display random, persistent and oscillatory dynamic patterns.",22215
method and a classification. on Finally object detection train to we jointly propose,Finally we propose a method to jointly train on object detection and classification.,22216
the validate ImageNet on approach task. our detection We,We validate our approach on the ImageNet detection task.,22217
in Nowadays task. used practical the for widely is CNN applications classification image,Nowadays the CNN is widely used in practical applications for image classification task.,22218
very are experiment proved evaluation accurate and The the reliable. results that methods,The experiment results proved that the evaluation methods are very accurate and reliable.,22219
"K-means logo are In images each clustering looking using algorithm. similar class, clustered the","In each class, the similar looking logo images are clustered using K-means clustering algorithm.",22220
form of data. logo are represented the images interval referenced Thus in,Thus referenced logo images are represented in the form of interval data.,22221
suitable classified is symbolic then classifier. image using logo sample A,A sample logo image is then classified using suitable symbolic classifier.,22222
"are precision, classification validated The help results and the accuracy, recall, F-measure, of with time.","The classification results are validated with the help of accuracy, precision, recall, F-measure, and time.",22223
"end-to-end. the trained and can be jointly, temporal flow, recognition propagation the gated modules The","The flow, the recognition and the gated temporal propagation modules can be trained jointly, end-to-end.",22224
"of and images. the exist misalignment CMR mis-registration MS slices in intra-image Furthermore, inter-image there","Furthermore, there exist inter-image mis-registration and intra-image misalignment of slices in the MS CMR images.",22225
apply Theory We interest model creative propose Simplicity situations. (ST) to in to,We propose to apply Simplicity Theory (ST) to model interest in creative situations.,22226
ST interest in been has designed predict to communication. and describe,ST has been designed to describe and predict interest in communication.,22227
regarded The an of creativity. form decision produces what rule be elementary as can,The decision rule produces what can be regarded as an elementary form of creativity.,22228
is proof study meant of a principle. This as,This study is meant as a proof of principle.,22229
be by may for that suggests creative the simplicity. search some unexpected It motivated actions,It suggests that some creative actions may be motivated by the search for unexpected simplicity.,22230
unknown method expectation-maximization estimate applied the is based parameters. to algorithm on An,An algorithm based on expectation-maximization method is applied to estimate the unknown parameters.,22231
method complexity the of linear pixels. number Our of to with is respect,Our method is of linear complexity with respect to the number of pixels.,22232
performed an electrode. Their extracellular measurements an and intracellular between were electrode,Their measurements were performed between an intracellular electrode and an extracellular electrode.,22233
recent been recognition years. have developed significantly techniques in Face,Face recognition techniques have been developed significantly in recent years.,22234
abdominal scans. been adopted have organ for automatic networks from Deep segmentation widely CT neural,Deep neural networks have been widely adopted for automatic organ segmentation from abdominal CT scans.,22235
new annotating feasible is domain a solution. Manually test each data for not,Manually annotating new data for each test domain is not a feasible solution.,22236
"we improved training. for domain discriminator multi-connected propose Furthermore, a adversarial","Furthermore, we propose a multi-connected domain discriminator for improved adversarial training.",22237
"persistent functions we structure multiscale this biomolecular introduce for paper, In characterization.","In this paper, we introduce multiscale persistent functions for biomolecular structure characterization.",22238
achieved scales can value. be tuning by its Various,Various scales can be achieved by tuning its value.,22239
"evaluation. persistent entropy can our be multiscale used in entropy conformation Physically,","Physically, our multiscale persistent entropy can be used in conformation entropy evaluation.",22240
"is that a our found natural specifically, it in More scheme. classification incorporates it method","More specifically, it is found that our method incorporates in it a natural classification scheme.",22241
"method in our can protein successfully be Moreover, used classification.","Moreover, our method can be successfully used in protein classification.",22242
"can information used in Essentially, to PSI systems. the describe any ""regularity"" be","Essentially, PSI can be used to describe the ""regularity"" information in any systems.",22243
a of and generative both discriminative model a MARTA $G$ GANs consists model $D$.,MARTA GANs consists of both a generative model $G$ and a discriminative model $D$.,22244
$D$ We treat feature as extractor. a,We treat $D$ as a feature extractor.,22245
"canonical structure so, by century-old we do the induced metric. Riemannian exploit To Cartan's","To do so, we exploit the Riemannian structure induced by Cartan's century-old canonical metric.",22246
the of bandwidth. proposed We of limit theorems provide to behavior functions characterize also our,We also provide limit theorems to characterize the behavior of our proposed functions of bandwidth.,22247
tracking in main appearance and variation The difficulties occlusion. the are field,The main difficulties in the tracking field are appearance variation and occlusion.,22248
We objects our frames. in video method that effective through tracking is single various show,We show that our method is effective in tracking various single objects through video frames.,22249
 are to microbial In nutrient subject fluctuating nature populations levels. ,  In nature microbial populations are subject to fluctuating nutrient levels.,22250
phenomenological observations. model recapitulates A our experimental,A phenomenological model recapitulates our experimental observations.,22251
a fuzzy discriminant (FC-LDA). introduce paper linear analysis constraint we this In,In this paper we introduce a fuzzy constraint linear discriminant analysis (FC-LDA).,22252
when results problems. algorithm different dealing the with success of verify the Obtained,Obtained results verify the success of the algorithm when dealing with different problems.,22253
task. classification Comparing LDA FC-LDA in superiority and shows,Comparing FC-LDA and LDA shows superiority in classification task.,22254
"can predictive place. schemes if However, this situation in averted monitoring be adequate are put","However, this situation can be averted if adequate predictive monitoring schemes are put in place.",22255
in a Visual fundamental vision. computer is tracking problem,Visual tracking is a fundamental problem in computer vision.,22256
"achieving been Recently, have performances. record-breaking deep-learning-based some tracking algorithms","Recently, some deep-learning-based tracking algorithms have been achieving record-breaking performances.",22257
is sequence-based to to apply problem. an effectively open data still ConvNets How,How to effectively apply ConvNets to sequence-based data is still an open problem.,22258
multiply of improved through JTMs. fusion the three The further score is recognition final,The final recognition is further improved through multiply score fusion of the three JTMs.,22259
"very low to EIT leads with However, resolution. images","However, EIT leads to images with very low resolution.",22260
"quality EIT image resampling for we Super-Resolution based work, improvement. a propose method In this","In this work, we propose a resampling based Super-Resolution method for EIT image quality improvement.",22261
Such to are brink collapse. pushing of ecosystems the threats,Such threats are pushing ecosystems to the brink of collapse.,22262
network structure alterations Alzheimer's disease causes function. of and brain (AD),Alzheimer's disease (AD) causes alterations of brain network structure and function.,22263
frequency processes connectivity latter of at consists between changes channels. different oscillatory The,The latter consists of connectivity changes between oscillatory processes at different frequency channels.,22264
based realistic person in re-identification and role plays video a central Video security surveillance.,Video based person re-identification plays a central role in realistic security and video surveillance.,22265
experiments back synthetic our up theory. on data Thorough,Thorough experiments on synthetic data back up our theory.,22266
noisy and both applicable algorithm Our settings. is noiseless to,Our algorithm is applicable to both noisy and noiseless settings.,22267
corroborate proposed the of the algorithm superiority the Experiments over algorithms. synthetic on data state-of-the-art,Experiments on synthetic data corroborate the superiority of the proposed algorithm over the state-of-the-art algorithms.,22268
The in regimes expansion) travelling are wave retreats parameter advances identified. (range the and which,The parameter regimes in which the travelling wave advances (range expansion) and retreats are identified.,22269
agreement obtained in on budding populations. The with are the results observations recent experimental yeast,The results obtained are in agreement with the recent experimental observations on budding yeast populations.,22270
"'s follows coli, relation as a Schaechter Growth law. particular, growth of Escherichia known in","Growth of Escherichia coli, in particular, follows a relation known as Schaechter 's growth law.",22271
"of perturbations. range the law over growth conditions valid remained Moreover, growth a dimension and","Moreover, the growth law remained valid over a range of growth conditions and dimension perturbations.",22272
alterations. as is caused Cancer known disease a mainly by gene,Cancer is known as a disease mainly caused by gene alterations.,22273
We their apply ComMDP and efficiency. to first validate to simulated SpeMDP data,We first apply ComMDP and SpeMDP to simulated data to validate their efficiency.,22274
these more In candidate genes processes also found. cancer are,In these processes more candidate cancer genes are also found.,22275
frequently Protein fragments occurred protein are sequences. motifs conserved in,Protein motifs are conserved fragments occurred frequently in protein sequences.,22276
"significant functions, enzyme. active such of They an as have site","They have significant functions, such as active site of an enzyme.",22277
Search motifs clustering intensive. and protein computational sequence are,Search and clustering protein sequence motifs are computational intensive.,22278
"called sequence new protein and a HSEARCH. clustering motif present finding algorithm, We","We present a new protein sequence motif finding and clustering algorithm, called HSEARCH.",22279
(RBC) introduced Radon barcodes The idea has of recently. been,The idea of Radon barcodes (RBC) has been introduced recently.,22280
"deliver in not long vector multiple Nevertheless, using will improvements. projections anticipated a","Nevertheless, using multiple projections in a long vector will not deliver anticipated improvements.",22281
propose on analysis. kernel an based integration component We unsupervised data principal method,We propose an unsupervised data integration method based on kernel principal component analysis.,22282
of techniques component analysis. used Principal data widely in analysis the is most one,Principal component analysis is one of the most widely used techniques in data analysis.,22283
"of a each we to present input matrix. the scoring impact determine function Therefore,","Therefore, we present a scoring function to determine the impact of each input matrix.",22284
approach clustering the integrated subsequent for identification. The data cancer enables subtype visualizing and,The approach enables visualizing the integrated data and subsequent clustering for cancer subtype identification.,22285
"have set. no be to free nature method, the of parameters Due to the","Due to the nature of the method, no free parameters have to be set.",22286
"usually feedforward networks, through multilayer neural is actualized e.g. This","This is usually actualized through feedforward multilayer neural networks, e.g.",22287
"forms ConvNets, layer of where representations. successive such each one","ConvNets, where each layer forms one of such successive representations.",22288
"and basis Learning. a for provides Curriculum it new taxonomy), a","a taxonomy), and it provides a new basis for Curriculum Learning.",22289
mechanisms (e.g. several investigate feedback also We architectures in,We also investigate several mechanisms in feedback architectures (e.g.,22290
choices and time) (e.g. connections in design skip,skip connections in time) and design choices (e.g.,22291
a This within is sampling rejection framework. done,This is done within a rejection sampling framework.,22292
in image analysis. Appearance in result variations many face difficulties,Appearance variations result in many difficulties in face image analysis.,22293
models This tensor contrasts the of variation-specific the with classical AAM.,This contrasts with the variation-specific models of the classical tensor AAM.,22294
"cascaded-regression-based use we fitting. effective an Last, UT-AAM method for","Last, we use an effective cascaded-regression-based method for UT-AAM fitting.",22295
"With these in utility advancements, the of enhanced. practice is UT-AAM considerably","With these advancements, the utility of UT-AAM in practice is considerably enhanced.",22296
blind is Joint technique data-fusion. source (J-BSS) for an separation emerging multi-set data-driven,Joint blind source separation (J-BSS) is an emerging data-driven technique for multi-set data-fusion.,22297
"perspective. paper, is In from this tensorial addressed J-BSS a","In this paper, J-BSS is addressed from a tensorial perspective.",22298
algorithm This of set a CPDs. to overdetermined converts DC-CPD a possibly underdetermined,This algorithm converts a possibly underdetermined DC-CPD to a set of overdetermined CPDs.,22299
latter eigenvalue based decomposition generalized algebraically via scheme. solved can be a The,The latter can be solved algebraically via a generalized eigenvalue decomposition based scheme.,22300
"case. in this solution is Therefore, the returns and exact algorithm noiseless the deterministic","Therefore, this algorithm is deterministic and returns the exact solution in the noiseless case.",22301
"the In can DC-CPD used effectively it based noisy to optimization case, algorithms. initialize be","In the noisy case, it can be used to effectively initialize optimization based DC-CPD algorithms.",22302
modified deep obtain convolution a We representation relying this architectures. on develop operator to,We develop a modified convolution operator relying on this representation to obtain deep architectures.,22303
algorithms The and representation independent. are domain entirely,The representation and algorithms are entirely domain independent.,22304
"algorithm case study also on network. NYC, transportation Manhattan, of the a We our present","We also present a case study of our algorithm on the Manhattan, NYC, transportation network.",22305
generation. order These that with collisions higher are coalescences consistent features multiple per permit,These features are consistent with higher order coalescences that permit multiple collisions per generation.,22306
"Collisions, Asymmetry. Tree Integer Selection, Keywords: Coalescence, Multiple Partitions,","Keywords: Selection, Coalescence, Integer Partitions, Multiple Collisions, Tree Asymmetry.",22307
is extend. consuming and Such to time difficult data to acquire,Such data is time consuming to acquire and difficult to extend.,22308
"multiobjective sequential we preference-based decision-making. a this In link paper, between present and","In this paper, we present a link between preference-based and multiobjective sequential decision-making.",22309
"of low-rank approximations. context the attempts several extended have to DMD Recently,","Recently, several attempts have extended DMD to the context of low-rank approximations.",22310
"in for extension is reduced-order domains, of particular e.g. interest This modeling applicative various","This extension is of particular interest for reduced-order modeling in various applicative domains, e.g.",22311
"climate study to micro-electromechanical or devices. dynamics for prediction, molecular","for climate prediction, to study molecular dynamics or micro-electromechanical devices.",22312
non-convex This problem. of optimization low-rank form takes a extension the,This low-rank extension takes the form of a non-convex optimization problem.,22313
A state-of-the-art the the compared techniques. gain algorithm proposed illustrates toy-example in of performance to,A toy-example illustrates the gain in performance of the proposed algorithm compared to state-of-the-art techniques.,22314
this is major training of reason annotated A limited the data. for amount,A major reason for this is the limited amount of annotated training data.,22315
Then pre-trained the data. for learning transfer utilize CNN on a we converted imagery,Then we utilize a pre-trained CNN for transfer learning on the converted imagery data.,22316
evaluate We a floor dataset our gait method pressure surface mapping. on of,We evaluate our method on a gait dataset of floor surface pressure mapping.,22317
popularity sparse increasing Recent Networks. learning have years an the \emph{changes} seen in Markov of,Recent years have seen an increasing popularity of learning the sparse \emph{changes} in Markov Networks.,22318
"develop for aimed we quantifying this Here, signal method a to transduction.","Here, we aimed to develop a method for quantifying this signal transduction.",22319
of the applicable the channel capacity cascade. MAPK to calculate method Our was,Our method was applicable to calculate the channel capacity of the MAPK cascade.,22320
"quantitative analyses our for of signal In conclusion, is method transduction. suitable","In conclusion, our method is suitable for quantitative analyses of signal transduction.",22321
"overfitting However, from the the suffers problem. baseline underfitting the fitting or itself","However, the baseline fitting itself suffers from the underfitting or the overfitting problem.",22322
"baseline we in gradient this a K-fold estimation policy paper, develop In method algorithms. for","In this paper, we develop a K-fold method for baseline estimation in policy gradient algorithms.",22323
"easy can is to The proposed DWS-MIL, as be efficiently. and trained abbreviated implement algorithm,","The proposed algorithm, abbreviated as DWS-MIL, is easy to implement and can be trained efficiently.",22324
efficient anomalies videos. present an in We method for detecting,We present an efficient method for detecting anomalies in videos.,22325
"and are require signals. However, as learning labels supervised neural networks convolutional","However, convolutional neural networks are supervised and require labels as learning signals.",22326
for videos anomaly crowded architecture We in detection spatiotemporal scenes. including propose a,We propose a spatiotemporal architecture for anomaly detection in videos including crowded scenes.,22327
to approaches We these efficiency. and respect make comparisons computational and statistical with consider,We consider these approaches and make comparisons with respect to statistical and computational efficiency.,22328
neurons. for networks by inspired were biological originally example Neural,Neural networks for example were originally inspired by biological neurons.,22329
the investigate to architectures. and Extensive restoration performance network between are conducted relation different experiments,Extensive experiments are conducted to investigate the relation between restoration performance and different network architectures.,22330
"encode circuits remains how time-varying unclear complex it neural However, patterns.","However, it remains unclear how neural circuits encode complex time-varying patterns.",22331
our to discuss We simulation synthetic using and relevance demonstrate approach its several biology. studies,We demonstrate our approach using several simulation studies and discuss its relevance to synthetic biology.,22332
"identify However, the be parameters. updating in methods uncertainty to quantification uncertainty the used can","However, uncertainty quantification methods can be used to identify the uncertainty in the updating parameters.",22333
"However, among the presence nonlinear methods spectra. in those mixing fail of observed will the","However, those methods will fail in the presence of nonlinear mixing among the observed spectra.",22334
of to number transformation and results. parameters more reduced constrained desirable The leads matching,The reduced number of transformation parameters leads to more constrained and desirable matching results.,22335
over better robustness the method state-of-the-art methods. of demonstrate proposed results Experimental,Experimental results demonstrate better robustness of the proposed method over state-of-the-art methods.,22336
"deppG this challenge. provided is to called this novel algorithm a paper, tackle In","In this paper, a novel algorithm called deppG is provided to tackle this challenge.",22337
recognizing Understanding isolation. individual beyond a in scene goes objects visual,Understanding a visual scene goes beyond recognizing individual objects in isolation.,22338
scene. objects information the constitute between about semantic Relationships rich also,Relationships between objects also constitute rich semantic information about the scene.,22339
show leverage contained clean We this approach does set. not fully in the the information,We show this approach does not fully leverage the information contained in the clean set.,22340
For is two. into split of the sake readability paper this,For the sake of readability this paper is split into two.,22341
"outcome general, In more is important. process is measure, to easier","In general, process is easier to measure, outcome is more important.",22342
fourfold. contribution study present is the The of original,The original contribution of the present study is fourfold.,22343
implementation conceptual highlighted overcome to Their quality are and improvement. drawbacks be for,Their conceptual and implementation drawbacks are highlighted to be overcome for quality improvement.,22344
superpixel segmentation introduced. developed map-guided for method hyperspectral and A imagery is,A map-guided superpixel segmentation method for hyperspectral imagery is developed and introduced.,22345
during are and acquisition. Image image noise distortions common image blur,Image blur and image noise are common distortions during image acquisition.,22346
"First, distortions. we under of examine performance four DNN the types classifier","First, we examine the DNN classifier performance under four types of distortions.",22347
"More of A.I. generally, an put forward ethical we view","More generally, we put forward an ethical view of A.I.",22348
provide A.I. relating near-term examples We to,We provide examples relating to near-term A.I.,22349
as well agents. systems superintelligent as hypothetical,systems as well as hypothetical superintelligent agents.,22350
unsupervised We dependencies compactly motion present approach representation an videos. learning that encodes the in,We present an unsupervised representation learning approach that compactly encodes the motion dependencies in videos.,22351
Encoder-Decoder of these based framework predict Network Recurrent sequences use a to flows. Neural We,We use a Recurrent Neural Network based Encoder-Decoder framework to predict these sequences of flows.,22352
"modality, i.e., Our framework RGB, and input generic is any to videos. Depth, RGB-D","Our framework is generic to any input modality, i.e., RGB, Depth, and RGB-D videos.",22353
advance. detectors in signature detection specific Signature-based target on the for hyperspectral rely target knowing,Signature-based detectors for hyperspectral target detection rely on knowing the specific target signature in advance.,22354
"often difficult However, to signature target obtain. impossible or are","However, target signature are often difficult or impossible to obtain.",22355
"be applied based signatures, can test target detector After signature then data. learning on any","After learning target signatures, any signature based detector can then be applied on test data.",22356
Gaussian the Different literature. model suggested approaches for in graphical are selection,Different approaches for Gaussian graphical model selection are suggested in the literature.,22357
individual is based on them of of family the independence considering conditional tests. One,One of them is based on considering the family of individual conditional independence tests.,22358
procedures sample for error a characteristic size. An its important is these rate given of,An important characteristic of these procedures is its error rate for a given sample size.,22359
risk a losses. measured of by is additive function procedures with quality statistical The,The quality of statistical procedures is measured by a risk function with additive losses.,22360
of errors account. be to losses allow Additive taken into both types,Additive losses allow both types of errors to be taken into account.,22361
"intelligence with breakthroughs milestones. recent several seen in has often Artificial games serving years, as","Artificial intelligence has seen several breakthroughs in recent years, with games often serving as milestones.",22362
games feature common have players of perfect is A these information. that,A common feature of these games is that players have perfect information.,22363
"We introduce DeepStack, information algorithm for an imperfect settings.","We introduce DeepStack, an algorithm for imperfect information settings.",22364
"been now. denoising simple, image until exploited in this However has observation not","However simple, this observation has not been exploited in image denoising until now.",22365
"our successfully generation faces. well-formed Furthermore, pipeline creates","Furthermore, our generation pipeline successfully creates well-formed faces.",22366
modeling for noise in Poisson used is photon-limited distribution imaging.,Poisson distribution is used for modeling noise in photon-limited imaging.,22367
"Restricted form computational is thus absorbed called factor the for. amount post-processing light, of limits","Restricted form factor limits the amount of absorbed light, thus computational post-processing is called for.",22368
and in of inflow each region every a city. outflow) and,inflow and outflow) in each and every region of a city.,22369
of end-to-end structure of on an design unique data. based properties We ST-ResNet spatio-temporal,We design an end-to-end structure of ST-ResNet based on unique properties of spatio-temporal data.,22370
"according high-level classifying sentiment. images this framework we to introduce paper, for In a","In this paper, we introduce a framework for classifying images according to high-level sentiment.",22371
"parameter tracking multi-object for model is tuning prerequisite applications, a reliable In performance.","In multi-object tracking applications, model parameter tuning is a prerequisite for reliable performance.",22372
with Performance real real-world for videos the effectiveness of algorithm demonstrates evaluation tracking scenarios. the,Performance evaluation with real videos demonstrates the effectiveness of the tracking algorithm for real-world scenarios.,22373
from Iterative parts positive samples. using (IHF) only Our learnt Hough Forest is extracted the,Our Iterative Hough Forest (IHF) is learnt using parts extracted only from the positive samples.,22374
information HoCP inference. scale-variant provided rich is discriminative the leveraged during by features The,The rich discriminative information provided by the scale-variant HoCP features is leveraged during inference.,22375
"adjusting not temperature other Therefore, value to setting a value. is according the indoor","Therefore, adjusting the indoor temperature is not setting a value according to other value.",22376
the paper of sake readability is this two. split For into,For the sake of readability this paper is split into two.,22377
An variable proposed. original association quantitative measure pair is categorical also of,An original quantitative measure of categorical variable pair association is also proposed.,22378
employed the aim classifier. modeling boosted of We commonly study limitations trees the to decision,We aim to study the modeling limitations of the commonly employed boosted decision trees classifier.,22379
"the recognition models by success data-hungry Inspired large, visual (e.g. of","Inspired by the success of large, data-hungry visual recognition models (e.g.",22380
"more particular becomes solving Often, when used. is easier than modality a one problem","Often, solving a particular problem becomes easier when more than one modality is used.",22381
technique regression models. classification image a We linear using set propose novel,We propose a novel image set classification technique using linear regression models.,22382
the test each subspaces. using estimate regression class specific image gallery models for We,We estimate regression models for each test image using the class specific gallery subspaces.,22383
Images regression of the then reconstructed using set the models. are test,Images of the test set are then reconstructed using the regression models.,22384
digital entertainment sketch role synthesis enforcement. face in and Exemplar-based plays both an important law,Exemplar-based face sketch synthesis plays an important role in both digital entertainment and law enforcement.,22385
two selection parts: and weight of It neighbor generally reconstruction consists representation.,It generally consists of two parts: neighbor selection and reconstruction weight representation.,22386
"search time the for online increases the synthesis. Actually, consuming","Actually, the online search increases the time consuming for synthesis.",22387
appearance-based geometric of Facial expression and use methods features. combination a recognition,Facial expression recognition methods use a combination of geometric and appearance-based features.,22388
"information. derived Spatial landmarks, geometric are features from carry displacements and facial of","Spatial features are derived from displacements of facial landmarks, and carry geometric information.",22389
"These knowledge, a features dimension-reduced from or based large either on are prior pool. selected","These features are either selected based on prior knowledge, or dimension-reduced from a large pool.",22390
"forward of using descriptive for Among features selection. a search these, sequential we subset","Among these, we search for a descriptive subset of features using sequential forward selection.",22391
"(ii) it requires machine Automated, user interaction. no i.e.,","(ii) Automated, i.e., it requires no user machine interaction.",22392
"size. linearly real a increasing with image computational Near (iii) with time, complexity the","(iii) Near real time, with a computational complexity increasing linearly with the image size.",22393
"to in Implemented tile (iv) with cope massive streaming mode, images.","(iv) Implemented in tile streaming mode, to cope with massive images.",22394
representations. are effective (DCNNs) of Convolution unprecedentedly Neural Deep image capable Networks learning,Deep Convolution Neural Networks (DCNNs) are capable of learning unprecedentedly effective image representations.,22395
"remains However, significant limited. local rotations ability in and global handling their image","However, their ability in handling significant local and global image rotations remains limited.",22396
"from all is using back-propagation, rotated collectively its versions. an errors ARF updated During","During back-propagation, an ARF is collectively updated using errors from all its rotated versions.",22397
benchmarks. used report best on results commonly the We several,We report the best results on several commonly used benchmarks.,22398
information we partners? partners Can of over population the to obtain about average,Can we average over the population to obtain information about partners of partners?,22399
"dependencies demographic changes, partners If, additionally we between allow for arise. however,","If, however, we additionally allow for demographic changes, dependencies between partners arise.",22400
to is the of demography. attention the in Particular approximation accuracy setting the given with,Particular attention is given to the accuracy of the approximation in the setting with demography.,22401
and solely annotators fits. Human bad good sort,Human annotators solely sort good and bad fits.,22402
"are for code and models data, research available purposes. The","The data, code and models are available for research purposes.",22403
"present cues. we temporal dependencies long-term online that an method multiple encodes However, across","However, we present an online method that encodes long-term temporal dependencies across multiple cues.",22404
outperforms works method previous on datasets publicly the including available MOT challenging multiple Our benchmark.,Our method outperforms previous works on multiple publicly available datasets including the challenging MOT benchmark.,22405
These design. enabling strain potential have rational in studies,These studies have potential in enabling rational strain design.,22406
"been show data. demonstrated methods poor with recently these has However, that agreement it experimental","However, it has been recently demonstrated that these methods show poor agreement with experimental data.",22407
will regulation novel uncovering REMEP insights useful cellular also in control. and into prove,REMEP will also prove useful in uncovering novel insights into cellular regulation and control.,22408
Four distinguished. series time are methods classification,Four time series classification methods are distinguished.,22409
are methods their and benefits are The and drawbacks discussed. explained,The methods are explained and their benefits and drawbacks are discussed.,22410
simple set-up natural environment. This patchy a similar to is,This simple set-up is similar to a natural patchy environment.,22411
collective spatial departure. a organisation a This before emerge the seconds in few group,This spatial organisation emerge in the group a few seconds before a collective departure.,22412
the on the current ImageNet. of networks Residual are state art,Residual networks are the current state of the art on ImageNet.,22413
"and qualitative connections serve. analysis, empirical that explore residual purpose visualization the skip we Through","Through qualitative visualization and empirical analysis, we explore the purpose that residual skip connections serve.",22414
"layers expected. refine as Our force that shortcut connections assessments show features, the residual to","Our assessments show that the residual shortcut connections force layers to refine features, as expected.",22415
"yet consisting two simple, powerful, is approach stages. It a top-down of","It is a simple, yet powerful, top-down approach consisting of two stages.",22416
keypoint fully dense type predict ResNet. each convolutional a and using For we heatmaps offsets,For each keypoint type we predict dense heatmaps and offsets using a fully convolutional ResNet.,22417
"this in two significant advance work, In them we ways.","In this work, we advance them in two significant ways.",22418
isolated occur rarely  systems. in Plants ,  Plants rarely occur in isolated systems.,22419
"the fail can Unfortunately, often because poor. arbitrarily be approximation methods such","Unfortunately, such methods often fail because the approximation can be arbitrarily poor.",22420
this effectiveness demonstrate parsing two scene method datasets. of We on point cloud the,We demonstrate the effectiveness of this method on two point cloud scene parsing datasets.,22421
in Predictive allow inference neuroimaging disease related when subject-specific alterations analyzing models data.,Predictive models allow subject-specific inference when analyzing disease related alterations in neuroimaging data.,22422
"Given can i.e. two data, made levels: be at subject's a global, inference","Given a subject's data, inference can be made at two levels: global, i.e.",22423
"subject, presence condition i.e. and local, identifiying for the","identifiying condition presence for the subject, and local, i.e.",22424
effect detecting condition individual the measurement subject's on each data. extracted from,detecting condition effect on each individual measurement extracted from the subject's data.,22425
"description explained. the in Firstly, employed levels different are usually of the models","Firstly, the different levels of description usually employed in the models are explained.",22426
"Secondly, the mathematical them presented. relationships are among","Secondly, the mathematical relationships among them are presented.",22427
"terminology are on will used notation and the be that Finally, explained. later","Finally, the notation and terminology that will be used later on are explained.",22428
may These models serve ecosystems. example an as of,These models may serve as an example of ecosystems.,22429
devising summary thesis research lines. future and a The with ends,The thesis ends with a summary and devising future research lines.,22430
in this data multidimensional clustering article. A for fuzzy proposed is algorithm,A fuzzy clustering algorithm for multidimensional data is proposed in this article.,22431
of results obtained efficiency proposed approach. confirm The the the,The obtained results confirm the efficiency of the proposed approach.,22432
"images. poorly These on perform contrast however, operators, low","These operators, however, perform poorly on low contrast images.",22433
the FPGA-based have through Experimental proved method effective. an implementation proposed results obtained,Experimental results obtained through an FPGA-based implementation have proved the proposed method effective.,22434
Savage utility expected standard provided an theory. act-based of axiomatization,Savage provided an act-based axiomatization of standard expected utility theory.,22435
"expected axiomatization a utility nonstandard of In article, we theory. this Savage-like provide","In this article, we provide a Savage-like axiomatization of nonstandard expected utility theory.",22436
"electrophysiological are intensity. characterized In as stimulus to proportional fact, commonly inversely signals","In fact, electrophysiological signals are commonly characterized as inversely proportional to stimulus intensity.",22437
"support of to statistical release probability of the transitions In providesadditional hypothesis particular, neurotransmitters. heterogeneous","In particular, statistical transitions providesadditional support to the hypothesis of heterogeneous release probability of neurotransmitters.",22438
morphological Our biophysical coherent electrophysiological into and investigations scenario. findings also conciliate a,Our findings also conciliate morphological and electrophysiological investigations into a coherent biophysical scenario.,22439
"phases. two ideas resulted algorithm, a two real-time over new developed in stabilization These","These two new ideas resulted in a real-time stabilization algorithm, developed over two phases.",22440
two active the aspects. learning approach methods existing Our in advances,Our approach advances the existing active learning methods in two aspects.,22441
"networks First, we deep neural incorporate learning. active into convolutional","First, we incorporate deep convolutional neural networks into active learning.",22442
"high these pseudo-labels. assigned and iteratively confidence are samples selected Specifically, automatically","Specifically, these high confidence samples are automatically selected and iteratively assigned pseudo-labels.",22443
(LAP) People Looking the of events. reviews paper historic This ChaLearn at,This paper reviews the historic of ChaLearn Looking at People (LAP) events.,22444
LAP perspectives a ChaLearn of also provide discussion We on activities.,We also provide a discussion on perspectives of ChaLearn LAP activities.,22445
using for this paper a recognition logo propose we method deep In learning.,In this paper we propose a method for logo recognition using deep learning.,22446
have data algorithms are proposed. Many the clustering been when recently or functions curves,Many clustering algorithms when the data are curves or functions have been recently proposed.,22447
been simulation method has a through proposed evaluated study. The,The proposed method has been evaluated through a simulation study.,22448
"to problem given. is data an a real application Finally,","Finally, an application to a real data problem is given.",22449
image. and We comprehension consider natural language generation for of expression an objects in referring,We consider generation and comprehension of natural language referring expression for objects in an image.,22450
generation benchmark lead expression multiple improved show that on datasets. both We to approaches referring,We show that both approaches lead to improved referring expression generation on multiple benchmark datasets.,22451
mixture Bayesian nonparametric models this a provide dependent flexible to address goal. approach,Bayesian nonparametric dependent mixture models provide a flexible approach to address this goal.,22452
"require demanding formulations several computationally posterior However, inference. algorithms for","However, several formulations require computationally demanding algorithms for posterior inference.",22453
why such provide better is action We a representation intuitions recognition. on for,We provide intuitions on why such a representation is better for action recognition.,22454
low-resolution The of (LR) a from sample image patches. set is LR dictionary trained,The low-resolution (LR) dictionary is trained from a set of LR sample image patches.,22455
error The square error. the used here mean criterion is,The error criterion used here is the mean square error.,22456
our show the compared effectiveness the results state-of-art algorithms. SR Simulation to of method,Simulation results show the effectiveness of our method compared to the state-of-art SR algorithms.,22457
been tasks. applied of Convolutional variety to a networks wide have computer neural vision,Convolutional neural networks have been applied to a wide variety of computer vision tasks.,22458
to image Recent medical semantic their in have advances segmentation application enabled segmentation.,Recent advances in semantic segmentation have enabled their application to medical image segmentation.,22459
method Our is unique many in respects.,Our method is unique in many respects.,22460
"siamese the the departing First, from SOMAnet on architecture, Inception usual framework. based is","First, SOMAnet is based on the Inception architecture, departing from the usual siamese framework.",22461
in of data with combination avenues Inception opens synthetic research up architectures The re-identification. new,The combination of synthetic data with Inception architectures opens up new research avenues in re-identification.,22462
"we automated and for addition, (KMARF). introduce a management knowledge framework reasoning In","In addition, we introduce a framework for knowledge management and automated reasoning (KMARF).",22463
on illustrate planning approach a transportation the We use case.,We illustrate the approach on a transportation planning use case.,22464
using Internet Computer for of sensors? use we Could in Vision Things the pictures as,Could we use Computer Vision in the Internet of Things for using pictures as sensors?,22465
principal This to is want hypothesis the we resolve. that,This is the principal hypothesis that we want to resolve.,22466
"homes, safety Currently, cameras. people areas, IP to cities, order create or in use","Currently, in order to create safety areas, cities, or homes, people use IP cameras.",22467
"the world. there and are Homes Smart Cities Furthermore, around many Smart","Furthermore, there are many Smart Cities and Smart Homes around the world.",22468
"a Computer Notwithstanding, complicated very Vision is field.","Notwithstanding, Computer Vision is a very complicated field.",22469
"integrate the constraints to we cross-modality incorporate modalities Moreover, collaboratively. consistent different","Moreover, we incorporate the cross-modality consistent constraints to integrate different modalities collaboratively.",22470
of seem in Potential measurement several similarity image applications evident the aspects. ultrasound,Potential applications of the ultrasound image similarity measurement seem evident in several aspects.,22471
"To a region abnormal function etc. imaging detection, ultrasound name few, assessment, quality","To name a few, ultrasound imaging quality assessment, abnormal function region detection, etc.",22472
population in segregation The bacterial is of plasmids investigated. a,The segregation of plasmids in a bacterial population is investigated.,22473
the influence numbers. mainly large basic plasmid at rate) reproduction distribution plasmid,basic plasmid reproduction rate) influence the distribution mainly at large plasmid numbers.,22474
with theoretical results. compare experimental the findings We,We compare the theoretical findings with experimental results.,22475
alleles sampling in frequencies effects. population fluctuations due Genetic drift a stochastic is to of,Genetic drift is stochastic fluctuations of alleles frequencies in a population due to sampling effects.,22476
depend does fitness not Population on population average size.,Population average fitness does not depend on population size.,22477
the A formulas is required derivation of provided. detailed,A detailed derivation of the required formulas is provided.,22478
the classifiers. updated fine-tuned on The based are neural nets later,The neural nets are later fine-tuned based on the updated classifiers.,22479
"human in with accords mode finely learning the ""instructor-student-collaborative"" model new education. The","The new model finely accords with the ""instructor-student-collaborative"" learning mode in human education.",22480
user also state-of-the-art is A dramatic other achieved active techniques. reduction learning effort of over,A dramatic reduction of user effort is also achieved over other state-of-the-art active learning techniques.,22481
"and our two on We framework challenging demonstrate datasets, results. very promising evaluate","We evaluate our framework on two challenging datasets, and demonstrate very promising results.",22482
has been to vision liquids containers. made computer attention little and their in Very,Very little attention in computer vision has been made to liquids and their containers.,22483
"argue several issues critical we this this Nevertheless, that FOL purpose. has in for paper,","Nevertheless, in this paper, we argue that FOL has several critical issues for this purpose.",22484
"assertional logic, than although We more is that FOL. argue and expressive extensible simpler,","We argue that assertional logic, although simpler, is more expressive and extensible than FOL.",22485
"dependent. time antibiotic is In in clinic, treatment particular, the","In particular, in the clinic, antibiotic treatment is time dependent.",22486
are These may be and clinical significance. testable of experimentally predictions,These predictions are experimentally testable and may be of clinical significance.,22487
"of GPs scenario learning. However, few of are multi-view the applications very in there","However, there are very few applications of GPs in the scenario of multi-view learning.",22488
"GP multi-view we this new learning. paper, a for In present model","In this paper, we present a new GP model for multi-view learning.",22489
techniques general intelligence. re-energized advancement learning artificial has research The rapid of into machine,The rapid advancement of machine learning techniques has re-energized research into general artificial intelligence.,22490
"a further sharper our if we inductive bias But, constrain expect should environment. we","But, we should expect a sharper inductive bias if we further constrain our environment.",22491
becoming passing with day. each is pervasive cancer Breast,Breast cancer is becoming pervasive with each passing day.,22492
"early in its the is patient. big saving of any detection Hence, step life a","Hence, its early detection is a big step in saving the life of any patient.",22493
common cancer breast a is in diagnosis. tool Mammography,Mammography is a common tool in breast cancer diagnosis.,22494
and The as is patches important classification here normal-abnormal most of step mammogram benign-malignant.,The most important step here is classification of mammogram patches as normal-abnormal and benign-malignant.,22495
classifications. these in mammogram plays a Texture a breast in a patch role significant of,Texture of a breast in a mammogram patch plays a significant role in these classifications.,22496
mammogram of patch useful. may not a features be All,All features of a mammogram patch may not be useful.,22497
"(DP). we apply feature a selection called Discrimination Potentiality technique Hence,","Hence, we apply a feature selection technique called Discrimination Potentiality (DP).",22498
"descriptors. with resulting compared DP-PB-DCT, the descriptors, are standard and DP-HOT Our","Our resulting descriptors, DP-HOT and DP-PB-DCT, are compared with the standard descriptors.",22499
"is patch Density of a has been important not studied mammogram and exhaustively. for classification,","Density of a mammogram patch is important for classification, and has not been studied exhaustively.",22500
achieved progress. have recently significant networks Deep neural,Deep neural networks have recently achieved significant progress.,22501
"rights to time, necessary models. is the the same of protect trained it shared At","At the same time, it is necessary to protect the rights of shared trained models.",22502
"new networks. watermarks into a deep formulate embedding Firstly, problem: neural we","Firstly, we formulate a new problem: embedding watermarks into deep neural networks.",22503
"neural types define situations, for also networks. embedding requirements, and We attack deep watermarking to","We also define requirements, embedding situations, and attack types for watermarking to deep neural networks.",22504
embedded. is performance a Our which the networks approach of into hurt watermark not does,Our approach does not hurt the performance of networks into which a watermark is embedded.,22505
"method for of = analysis, = utilization direction principal-singular-vector modal MODE estimation.) (PUMA","(PUMA = principal-singular-vector utilization for modal analysis, MODE = method of direction estimation.)",22506
develop We two of information operating seeking. models active,We develop two operating models of active information seeking.,22507
links We the postulates. and then these axioms rationality between consider the,We then consider the links between these axioms and the rationality postulates.,22508
least relationship a algorithm between Our on squares. is based kernel CCA and,Our algorithm is based on a relationship between kernel CCA and least squares.,22509
has in for learning. images models Recently in generative enormous there deep been interest an,Recently there has been an enormous interest in generative models for images in deep learning.,22510
Variational paper (ViGAN). InfoGAN we a this propose model called new In,In this paper we propose a new model called Variational InfoGAN (ViGAN).,22511
using models in algorithms color different have been literature. Several the proposed,Several algorithms using different color models have been proposed in the literature.,22512
"considerably the accelerated. to solution is analytical of formulation the to derivatives, optimization Due","Due to the analytical formulation of derivatives, optimization to the solution is considerably accelerated.",22513
synthetic both We data. using algorithm verify our real and,We verify our algorithm using both real and synthetic data.,22514
"weakness. Multimodal We Auxiliary such propose LSTM DNN (am-LSTM), a model, overcome to","We propose a DNN model, Auxiliary Multimodal LSTM (am-LSTM), to overcome such weakness.",22515
and The are extensibility consideration. take into flexibility also,The extensibility and flexibility are also take into consideration.,22516
"of departure analyse \textit{Danio the in zebrafish Here, we initiation rerio}. collective","Here, we analyse the initiation of collective departure in zebrafish \textit{Danio rerio}.",22517
size. same into of blocks a the split The is tensor number of,The tensor is split into a number of blocks of the same size.,22518
conditions the We give guarantee to block tensor successful incoherence position. decom- the,We give the block tensor incoherence conditions to guarantee the successful decom- position.,22519
deterministic undiagnosed features Our single-outbreak that is analysis with model a formulated comparative an class.,Our comparative analysis is formulated with a single-outbreak deterministic model that features an undiagnosed class.,22520
high-order naturally as a represented be time A can skeleton sequence tensor series.,A skeleton sequence can be naturally represented as a high-order tensor time series.,22521
learning sparse Then we perform classification Grassmann manifold. over coding using and dictionary,Then we perform classification using dictionary learning and sparse coding over Grassmann manifold.,22522
this and with theory. a life-history game of We investigate combination question evolutionary,We investigate this question with a combination of life-history and evolutionary game theory.,22523
a industrial Optimization systems. becoming in alternative involving crucial sustainable is element energy applications,Optimization is becoming a crucial element in industrial applications involving sustainable alternative energy systems.,22524
"(e.g. engineer/decision noise During the systems, often design would the of factors maker such encounter","During the design of such systems, the engineer/decision maker would often encounter noise factors (e.g.",22525
environment. fluctuations) their and with system ambient interacts temperature insolation solar when the,solar insolation and ambient temperature fluctuations) when their system interacts with the environment.,22526
"multivariate, problem noisy, multiobjective. This and nonlinear is","This problem is multivariate, noisy, nonlinear and multiobjective.",22527
graphic design common is images task photography. Exploring colors and a editing and in in,Exploring and editing colors in images is a common task in graphic design and photography.,22528
colors. means interactive and provide individual manipulating layers an for These intuitive,These layers provide an interactive and intuitive means for manipulating individual colors.,22529
fast Our layer extraction to a iterative color approach reduces system. linear,Our approach reduces color layer extraction to a fast iterative linear system.,22530
not follow a of design do Regression models datasets built are using experiment. which increasingly,Regression models are increasingly built using datasets which do not follow a design of experiment.,22531
gathered monitoring system. of by a an automated technical,gathered by an automated monitoring of a technical system.,22532
"correlations, data clusters patterns. show can other input The or","The input data can show correlations, clusters or other patterns.",22533
"Further, reliability models. the input of distribution of influences the regression data","Further, the distribution of input data influences the reliability of regression models.",22534
gesammelt. Systeme technischer Beobachtung werden die die Stattdessen durch passive beispielsweise Daten,Stattdessen werden die Daten beispielsweise durch die passive Beobachtung technischer Systeme gesammelt.,22535
"Systems Eingangsdaten Verteilungsannahmen. die Ph\""anomene des und statistischen Damit bereits bilden ab widersprechen","Damit bilden bereits die Eingangsdaten Ph\""anomene des Systems ab und widersprechen statistischen Verteilungsannahmen.",22536
"Regressionsmodells. hat die Zuverl\""assigkeit der eines Die Eingangsdaten auf Verteilung Einfluss","Die Verteilung der Eingangsdaten hat Einfluss auf die Zuverl\""assigkeit eines Regressionsmodells.",22537
"systems case, some to be in may not this the initializations. random However, restricted especially","However, this may not be the case, especially in some systems restricted to random initializations.",22538
more scale with also constraints. Improvements up sparsity lenient,Improvements also scale up with more lenient sparsity constraints.,22539
into research. hope future We can that provide mapping insights this priorities for,We hope that this mapping can provide insights into priorities for future research.,22540
to The is written any avoid performance in C mostly package unnecessary overhead.,The package is mostly written in C to avoid any unnecessary performance overhead.,22541
and widespread causing disease Falciparum parasitic major morbidity mortality malaria is a globally.,Falciparum malaria is a major parasitic disease causing widespread morbidity and mortality globally.,22542
largest obtained based is on that the inlier single Its response support. the hypothesis,Its response is based on the single hypothesis that obtained the largest inlier support.,22543
"the which we Furthermore, improves quality outline the query results. of matching of process semantic","Furthermore, we outline the process of semantic matching which improves the quality of query results.",22544
Random data (RF) algorithms missing forest missing dealing are data. for with attractive an approach,Random forest (RF) missing data algorithms are an attractive approach for dealing with missing data.,22545
was by ability data Performance assessed to algorithms impute accurately. of,Performance of algorithms was assessed by ability to impute data accurately.,22546
robust correlation. with generally to performance findings be improving reveal increasing RF with Our imputation,Our findings reveal RF imputation to be generally robust with performance improving with increasing correlation.,22547
texture previously tasks. intensity features We multi-center Gaussian show used classification features in that outperform,We show that Gaussian texture features outperform intensity features previously used in multi-center classification tasks.,22548
"resistance can promote spread. drug it However, also","However, it can also promote drug resistance spread.",22549
synthetic to found counterfactual forests. One is be promising method especially,One method found to be especially promising is counterfactual synthetic forests.,22550
"drug analysis connections important behavior, sexual and between The risky reveals usage, risk.","The analysis reveals important connections between risky behavior, drug usage, and sexual risk.",22551
end-to-end generating an learning propose framework for We segmentations. foreground object,We propose an end-to-end learning framework for generating foreground object segmentations.,22552
"regime these this In to work, new propose mitigate drawbacks. a we","In this work, we propose a new regime to mitigate these drawbacks.",22553
time can Therefore reduced be capture significantly. the,Therefore the capture time can be reduced significantly.,22554
important pose video Finding is in an egocentric the many camera applications. step,Finding the camera pose is an important step in many egocentric video applications.,22555
with loop use closures profile to specific wearer's short aligned footsteps. this motion We perform,We use this specific motion profile to perform short loop closures aligned with wearer's footsteps.,22556
"noisy. usually is depth egocentric For estimation videos,","For egocentric videos, depth estimation is usually noisy.",22557
become patients resistant. may drug In addition some,In addition some patients may become drug resistant.,22558
"clinically are Since premature. its statements not effectiveness proposed the been has on tested, psychotherapy","Since the psychotherapy proposed has not been clinically tested, statements on its effectiveness are premature.",22559
"For this Dirichlet Allocation. the extend Latent of we purpose, techniques significantly","For this purpose, we significantly extend the techniques of Latent Dirichlet Allocation.",22560
predictions. frame-level step first-order uses Usually of this pooling action statistics the,Usually this pooling step uses the first-order statistics of frame-level action predictions.,22561
"the these To kernel descriptors, linearization. generate of idea use we","To generate these descriptors, we use the idea of kernel linearization.",22562
"datasets all are controlled under conditions. captured However,","However, all datasets are captured under controlled conditions.",22563
"are results facial standard qualitative Complementary ""in-the-wild"" databases. demonstrated on reconstruction","Complementary qualitative reconstruction results are demonstrated on standard ""in-the-wild"" facial databases.",22564
technique source of Project. An is our of Menpo as open part the released implementation,An open source implementation of our technique is released as part of the Menpo Project.,22565
local these a without can learn How abstractions teacher? a robot,How can a robot learn these local abstractions without a teacher?,22566
this A called MISFA Curious problem. recently addresses proposed Dr. algorithm,A recently proposed algorithm called Curious Dr. MISFA addresses this problem.,22567
on artificial The based is two slowness. underlying curiosity and called learning principles algorithm,The algorithm is based on two underlying learning principles called artificial curiosity and slowness.,22568
simulator. We the and method efficacy in demonstrate our vision-based of a robot stability,We demonstrate the efficacy and stability of our method in a vision-based robot simulator.,22569
in objects segmenting learning framework for an generic We videos. propose end-to-end,We propose an end-to-end learning framework for segmenting generic objects in videos.,22570
on Code are and project models the available pre-trained website.,Code and pre-trained models are available on the project website.,22571
properties made Its applications some which need it degrees. processes for different in diagonal appropriate,Its diagonal properties made it appropriate for some applications which need processes in different degrees.,22572
Radon needs of arithmetic compute-intensive operations algorithm. transform a which makes computation it lot a,Radon transform computation needs a lot of arithmetic operations which makes it a compute-intensive algorithm.,22573
But and complex need to memory. arbitrary this is accesses algorithm,But this algorithm is complex and need arbitrary accesses to memory.,22574
memory an this In algorithm to paper sequentially. we proposed accesses which,In this paper we proposed an algorithm which accesses to memory sequentially.,22575
works. the that previous some a limitations introduces addresses This of technique paper,This paper introduces a technique that addresses some limitations of the previous works.,22576
The is on channels identify segmentation several to regions. order homogeneous color performed Otsu in,The Otsu segmentation is performed on several color channels in order to identify homogeneous regions.,22577
"viable in therefore, therapeutic targets represent pathologies. STATs, numerous","STATs, therefore, represent viable therapeutic targets in numerous pathologies.",22578
to number STAT circuits. signaling proposed of is The model a applicable,The proposed model is applicable to a number of STAT signaling circuits.,22579
"the statistics. match not generated and are non-photorealistic image do However, real images","However, the generated images are non-photorealistic and do not match real image statistics.",22580
dynamics with understand of pests. information to Population is spread the spatial applied,Population dynamics with spatial information is applied to understand the spread of pests.,22581
a pests describing discrete introduce how model spread space. in We,We introduce a model describing how pests spread in discrete space.,22582
impact. to its modeled the features in are The negative explicitly MRF minimize unmatched,The unmatched features are explicitly modeled in the MRF to minimize its negative impact.,22583
"i.e. comparisons feature finds that Experimental proposed correspondences, better algorithm the show","Experimental comparisons show that the proposed algorithm finds better feature correspondences, i.e.",22584
the algorithms solved be This to by characters. can hand-crafting prescale,This can be solved by hand-crafting algorithms to prescale the characters.,22585
singular not values the to Theses compressed with suitable decomposition. images are metrics,Theses metrics are not suitable to images compressed with the singular values decomposition.,22586
assessments areas quality were of image Three identified.,Three areas of image quality assessments were identified.,22587
our contributions. on Experiments real and respective show instances conducted interest the of random,Experiments conducted on real and random instances show the respective interest of our contributions.,22588
and nonlinear applies linear principle general This to both classifiers. selection,This general selection principle applies to both linear and nonlinear classifiers.,22589
idea. to prototypical classifiers example above a illustrate as are used the Large-margin,Large-margin classifiers are used as a prototypical example to illustrate the above idea.,22590
"brightness there Moreover, and between and VCI. serious discrepancies edge exist NGI","Moreover, there exist serious edge and brightness discrepancies between NGI and VCI.",22591
"with a is this fusion proposed transfer-based for To problem, deal removal. method new noise","To deal with this problem, a new transfer-based fusion method is proposed for noise removal.",22592
resultant uniform with confidence smoothed a filter. is The map,The resultant confidence map is smoothed with a uniform filter.,22593
onto map smoothed A confidence the non-maximal is suppression to obtain applied peaks.,A non-maximal suppression is applied onto the smoothed confidence map to obtain peaks.,22594
"images drastically vision such performance systems. In degraded affect addition, of","In addition, such degraded images drastically affect performance of vision systems.",22595
"solve it the is de-raining/de-snowing. to Hence, important image single of problem","Hence, it is important to solve the problem of single image de-raining/de-snowing.",22596
"solve is its due difficult this problem inherent nature. to to a However, ill-posed","However, this is a difficult problem to solve due to its inherent ill-posed nature.",22597
prior well-posed into it attempt problem. Existing to convert information a introduce approaches to,Existing approaches attempt to introduce prior information to convert it into a well-posed problem.,22598
"Regrettably, take to considerable approaches remain scenes. either or time both capture, to synthetic limited","Regrettably, both approaches either take considerable time to capture, or remain limited to synthetic scenes.",22599
be in decomposition then image any manipulation software This manipulated and off-the-shelf back. can recomposited,This decomposition can then be manipulated in any off-the-shelf image manipulation software and recomposited back.,22600
using perform data. network learning convolutional synthetic a We trained by a decomposition neural such,We perform such a decomposition by learning a convolutional neural network trained using synthetic data.,22601
"settings non-small useful and online for is Therefore, SSG in sizes. sample","Therefore, SSG is useful in online settings and for non-small sample sizes.",22602
The implications and extensions potential outlined. are,The potential implications and extensions are outlined.,22603
neither did not paper Luria-Delbruck of the also the experiment. model fitting possibility The consider,The Luria-Delbruck paper also did not consider the possibility of neither model fitting the experiment.,22604
"in environment. in difficulties challenging to unconstrained Image shooting alignment photography, consumer-level due stitching is","Image stitching is challenging in consumer-level photography, due to alignment difficulties in unconstrained shooting environment.",22605
studies generated show by relieve misalignment. approaches effectively artifacts seam-cutting Recent can local that,Recent studies show that seam-cutting approaches can effectively relieve artifacts generated by local misalignment.,22606
"into stitching be our easily can seam-cutting addition, implemented In pipelines. composition other","In addition, our seam-cutting composition can be easily implemented into other stitching pipelines.",22607
our that state-of-the-art DHN demonstrate SR results notably achieves Extensive better methods. than performance experimental,Extensive experimental results demonstrate that our DHN achieves notably better performance than state-of-the-art SR methods.,22608
addresses This characters. paper font typographic automatic a the subset from a generation of of,This paper addresses the automatic generation of a typographic font from a subset of characters.,22609
"we Specifically, characters. of extrapolate a a use additional to font subset typographic","Specifically, we use a subset of a typographic font to extrapolate additional characters.",22610
"sufficient obtain complete number a we Consequently, use. of characters a containing font daily for","Consequently, we obtain a complete font containing a number of characters sufficient for daily use.",22611
proposed uses for font generation. from strokes given method samples The,The proposed method uses strokes from given samples for font generation.,22612
"skeleton construct strokes, character a are dataset. by exploiting characters, which extracted The we from","The strokes, from which we construct characters, are extracted by exploiting a character skeleton dataset.",22613
the are handmade similar and generated verify characters to evaluations that Objective characters. subjective,Objective and subjective evaluations verify that the generated characters are similar to handmade characters.,22614
unsupervised a learning. Clustering approach for central is,Clustering is a central approach for unsupervised learning.,22615
"compare to fundamental is applied, is analysis After quantitatively most clusterings. the clustering","After clustering is applied, the most fundamental analysis is to quantitatively compare clusterings.",22616
two Guessing and problem. Matrix introduce Problem provide solve the We this Binary algorithms to,We introduce the Binary Matrix Guessing Problem and provide two algorithms to solve this problem.,22617
"better. is but slower compared less to and generalizes first algorithm significantly one, This restrictive","This algorithm is significantly slower compared to first one, but less restrictive and generalizes better.",22618
numerical compare computational and performance We results. of provide both algorithms,We compare computational performance of both algorithms and provide numerical results.,22619
"To to do so. this authors, the attempt the of is first best the","To the best of the authors, this is the first attempt to do so.",22620
"efficient is system skin proposed. this paper, detection an In","In this paper, an efficient skin detection system is proposed.",22621
algorithm The speed. processing been to obtain in real-time implemented has FPGA,The algorithm has been implemented in FPGA to obtain real-time processing speed.,22622
for keypoint used multi-resolution is learning A descriptors. ConvNet,A multi-resolution ConvNet is used for learning keypoint descriptors.,22623
been images. and Many estimation multi-view proposed in human for have approaches RGB single pose,Many approaches have been proposed for human pose estimation in single and multi-view RGB images.,22624
and sampled is common practice supervision. frames as A to inputs video as use labels,A common practice is to use sampled frames as inputs and video labels as supervision.,22625
on shown and both are COCO detection. VOC Results PASCAL,Results are shown on both PASCAL VOC and COCO detection.,22626
effective during recognition been face past methods the proposed of have for Plenty decade.,Plenty of effective methods have been proposed for face recognition during the past decade.,22627
"face In independently. implemented systems, most the module alignment is","In most systems, the face alignment module is implemented independently.",22628
This models. difficulties training has and end-to-end the recognition of in actually designing face caused,This has actually caused difficulties in the designing and training of end-to-end face recognition models.,22629
large approach in up kernel-based Random for speeding features datasets. sinusoidal popular are a inference,Random sinusoidal features are a popular approach for speeding up kernel-based inference in large datasets.,22630
experiments. efficacy We approach support of via the numerical our,We support the efficacy of our approach via numerical experiments.,22631
define to image some we able metrics characterize and the decay.,and we define some metrics able to characterize the image decay.,22632
to robust modelling in widely mixture possessing thick due Heavy-tailed are tails. distributions used,Heavy-tailed distributions are widely used in robust mixture modelling due to possessing thick tails.,22633
results. scene achieves LDPO classification excellent quantitative The,The LDPO achieves excellent quantitative scene classification results.,22634
"detritus ecosystem made presented. is oxygen An dissolved plants, nutrients, of and","An ecosystem made of nutrients, plants, detritus and dissolved oxygen is presented.",22635
a algorithm. to of proposed simulations of the We also provide range illustrate performance the,We also provide a range of simulations to illustrate the performance of the proposed algorithm.,22636
"called parameter. SVDQuartets a been recover to has this Moreover, statistically reconstruction developed method consistent","Moreover, a statistically consistent reconstruction method called SVDQuartets has been developed to recover this parameter.",22637
was expenditure step For required recorded. time each the,For each step the required time expenditure was recorded.,22638
in ecology important forest Recognising imagery individual and has trees sensed within management. applications remotely,Recognising individual trees within remotely sensed imagery has important applications in forest ecology and management.,22639
"from is with the hyperspectral a data. Second, graph point LiDAR built the fused cloud,","Second, a graph is built from the LiDAR point cloud, fused with the hyperspectral data.",22640
"PCA. computational hyperspectral is robust of For reduced imagery using space feature the efficiency,","For computational efficiency, the feature space of hyperspectral imagery is reduced using robust PCA.",22641
Multiclass We (MCRC). followed by approach this Recursive Cut call Cut,We call this approach Multiclass Cut followed by Recursive Cut (MCRC).,22642
behavioral simple parametric cloning. Traditional have methods and modeling models employed,Traditional modeling methods have employed simple parametric models and behavioral cloning.,22643
are bilateral spatial considered. high-dimensional and kernels standard Both,Both standard spatial and high-dimensional bilateral kernels are considered.,22644
"addition, mean-field algorithm. commonly In inference used the method our compare to we","In addition, we compare our inference method to the commonly used mean-field algorithm.",22645
in task Semantic segmentation is understanding. fundamental image image a,Semantic image segmentation is a fundamental task in image understanding.,22646
into can but segmentation any be easily simple systems. semantic module This effective integrated,This simple but effective module can be easily integrated into any semantic segmentation systems.,22647
"classroom We this develop using capacity Studies."" activity of ""Study a describe the to","We describe a classroom activity to develop this capacity using the ""Study of Studies.""",22648
group critiques shared are and the with instructor. Their class the,Their group critiques are shared with the class and the instructor.,22649
included the not summary. noted key Students that published details were in,Students noted that key details were not included in the published summary.,22650
published They conclusions. the generally skeptical of were,They were generally skeptical of the published conclusions.,22651
detect network under unconstrained variations. allows severe pose consideration faces to This and occlusion our,This consideration allows our network to detect faces under severe occlusion and unconstrained pose variations.,22652
stitching. in is attentions of extensive naturalness gaining The image warps,The naturalness of warps is gaining extensive attentions in image stitching.,22653
"is global a thus relies homography, totally proposed Because on warp it our parameter-free. only","Because our proposed warp only relies on a global homography, thus it is totally parameter-free.",22654
"study to comparing and wins users' most user homography it that favor, SPHP. demonstrates A","A user study demonstrates that it wins most users' favor, comparing to homography and SPHP.",22655
We extending time microevolutionary scales. models dynamics competition macroevolutionary long study to by,We study macroevolutionary dynamics by extending microevolutionary competition models to long time scales.,22656
evolutionary to different the use statistical a We investigate approach many of systems. dynamics,We use a statistical approach to investigate the evolutionary dynamics of many different systems.,22657
numerous method advantages over Our approaches. has traditional,Our method has numerous advantages over traditional approaches.,22658
"is embedded available. algorithm proposed in application, smart-phone The is publicly a which","The proposed algorithm is embedded in a smart-phone application, which is publicly available.",22659
(i.e. notions based on plans compare We decision-theoretic,We compare plans based on decision-theoretic notions (i.e.,22660
sound implementation complete. proven is The be and to,The implementation is proven to be sound and complete.,22661
close-to-singular validated These nominal and benefits through simulations geometric Monte-Carlo for configurations. both extensive are,These benefits are validated through extensive Monte-Carlo simulations for both nominal and close-to-singular geometric configurations.,22662
The of activating loss typical the ability is of called of geriatric sarcopenia. syndrome,The loss of ability of activating is typical of the geriatric syndrome called sarcopenia.,22663
"function. a transverse described energy polyconvex, The hyperelastic, by isotropic is strain material","The material is described by a hyperelastic, polyconvex, transverse isotropic strain energy function.",22664
discussed results the implemented model graphically and and numerically are obtained The is represented.,The model is implemented numerically and the obtained results are discussed and graphically represented.,22665
labelling semantic in There the task have recent improvements remarkable in been the years.,There have been remarkable improvements in the semantic labelling task in the recent years.,22666
"on pixel-level of However, large-scale art the the state rely annotations. methods","However, the state of the art methods rely on large-scale pixel-level annotations.",22667
"charge MNIST. to available, free symbols a HASY of is of publicly similar single dataset","HASY is a publicly available, free of charge dataset of single symbols similar to MNIST.",22668
Markov random effects are latent (GMRF). whose on field focuses INLA a models Gaussian,INLA focuses on models whose latent effects are a Gaussian Markov random field (GMRF).,22669
"conditioning the model fitted with parameters Hence, the be package. could on R-INLA these","Hence, conditioning on these parameters the model could be fitted with the R-INLA package.",22670
are of light differing explanations We the that studied best in these posit models.,We posit that the explanations are best studied in light of these differing models.,22671
"a results. we re-rank the paper, propose k-reciprocal In re-ID this encoding method to","In this paper, we propose a k-reciprocal encoding method to re-rank the re-ID results.",22672
The network (CNN) in convolutional studied be note. will this neural,The convolutional neural network (CNN) will be studied in this note.,22673
"transform we building as introduce CNNs. basic a To with, of block begin a RECOS","To begin with, we introduce a RECOS transform as a basic building block of CNNs.",22674
"is The ""REctified-COrrelations on a Sphere"". ""RECOS"" an for acronym","The ""RECOS"" is an acronym for ""REctified-COrrelations on a Sphere"".",22675
"is we in how provided discuss backpropagation guidance by the through training. labels (BP) Second,","Second, we discuss how guidance is provided by labels through backpropagation (BP) in the training.",22676
"visualizing we for Python high-dimensional large, manipulating Here toolbox a HypeTools, datasets. and present","Here we present HypeTools, a Python toolbox for visualizing and manipulating large, high-dimensional datasets.",22677
designed point notion trajectories data of The is toolbox the around clouds. and,The toolbox is designed around the notion of data trajectories and point clouds.,22678
be The plotted visualized static plots interactive may animations. as or trajectories as,The trajectories may be plotted as interactive static plots or visualized as animations.,22679
in reveal also dimensionality (e.g. structure reduction datasets and static These algorithms alignment can same,These same dimensionality reduction and alignment algorithms can also reveal structure in static datasets (e.g.,22680
images is from difficult The animals segmentation task. of camera-trap a,The segmentation of animals from camera-trap images is a difficult task.,22681
approach multi-layer background for analysis (multi-layer We proposed principal a robust RPCA) subtraction. component,We proposed a multi-layer robust principal component analysis (multi-layer RPCA) approach for background subtraction.,22682
of an RPCA multi-layer optimized exhaustive with search. parameters The our were,The parameters of our multi-layer RPCA were optimized with an exhaustive search.,22683
specifications enable evaluation in alternative of empirical two computational an terms These efficiency.,These two alternative specifications enable an empirical evaluation in terms of computational efficiency.,22684
w.r.t. show semantics the Experimental hybrid is more efficient that results,Experimental results show that the hybrid semantics is more efficient w.r.t.,22685
"sequences, semantics w.r.t. two the whereas same follows behaviour the","sequences, whereas the two semantics follows the same behaviour w.r.t.",22686
(although better denotational absolute in branchings the terms). performs one,branchings (although the denotational one performs better in absolute terms).,22687
produce the states behavioral adaptation. distinct to of inherently cognitive traversal Learning requires,Learning requires the traversal of inherently distinct cognitive states to produce behavioral adaptation.,22688
"compromise will alignment, matching the learning background With noise the feature and significantly process. bad","With bad alignment, the background noise will significantly compromise the feature learning and matching process.",22689
"provide focusing of these aspects. In we survey, review a particular root-cause analysis, this on","In this survey, we provide a review of root-cause analysis, focusing on these particular aspects.",22690
continues finding challenge robust solution It to it. in pose a to a,It continues to pose a challenge in finding a robust solution to it.,22691
"catering extends law enforcement commercial scope to security, the applications. Its and","Its scope extends to catering the security, commercial and law enforcement applications.",22692
imitated but cumbersome artificially. be A for trivial brain to task,A trivial task for brain but cumbersome to be imitated artificially.,22693
a the From devise function procedure nonparametrically. estimating hazard this gradient generic boosting we for,From this we devise a generic gradient boosting procedure for estimating the hazard function nonparametrically.,22694
"To avoid regularization employs devices. several overfitting, boosting","To avoid overfitting, boosting employs several regularization devices.",22695
agrees fluid computational with simulations. This dynamic result,This result agrees with computational fluid dynamic simulations.,22696
"the close from comes pixels to and improvement In edges situated particular, corners.","In particular, the improvement comes from pixels situated close to edges and corners.",22697
popular seven benchmark SCN is bottleneck of an Effectiveness on verified with modules several datasets.,Effectiveness of an SCN with seven bottleneck modules is verified on several popular benchmark datasets.,22698
"In (GAN) unlabeled this used is the to samples. generate adversarial generative work, network","In this work, the generative adversarial network (GAN) is used to generate unlabeled samples.",22699
for outliers We smoothing label the (LSRO). propose regularization,We propose the label smoothing regularization for outliers (LSRO).,22700
person re-identification problem: We method the a practical proposed (re-ID). on verify,We verify the proposed method on a practical problem: person re-identification (re-ID).,22701
other a retrieve aims task person cameras. to query from This,This task aims to retrieve a query person from other cameras.,22702
protein be tools bioinformatics can existent fiDPD annotation. used for complement as a to function,fiDPD can be used as a complement to existent bioinformatics tools for protein function annotation.,22703
a provide way for a determine to parameter. population values plausible Confidence intervals,Confidence intervals provide a way to determine plausible values for a population parameter.,22704
involving articles in omnipresent are They analyses. research statistical,They are omnipresent in research articles involving statistical analyses.,22705
in frequency single is high super-resolution task. details Hallucinating a image image challenging,Hallucinating high frequency image details in single image super-resolution is a challenging task.,22706
selectivity (color); an label and We image feature an membership). (class image for: explore indexes,We explore selectivity indexes for: an image feature (color); and an image label (class membership).,22707
"focus respectively on stable well-founded We semantics, on models. and on based two","We focus on two semantics, respectively based on stable and on well-founded models.",22708
inference We then programs. examine complexity the with logic probabilistic of,We then examine the complexity of inference with probabilistic logic programs.,22709
is condition common musculoskeletal after clinical typically radiographic Osteoarthritis from assessment diagnosed a examination. (OA),Osteoarthritis (OA) is a common musculoskeletal condition typically diagnosed from radiographic assessment after clinical examination.,22710
examination. the radiographic diagnostics Computer-aided could of improve objectivity knee (CAD),Computer-aided diagnostics (CAD) could improve the objectivity of knee radiographic examination.,22711
OA The step essential knee joint CAD area. of localize first to automatically the is,The first essential step of knee OA CAD is to automatically localize the joint area.,22712
"remains However, this literature challenging. itself to the task according","However, according to the literature this task itself remains challenging.",22713
for analyses. that demonstrate suitable method Our the developed results is large-scale,Our results demonstrate that the developed method is suitable for large-scale analyses.,22714
"acid arc-annotated This sequences. the in Ribonucleic finds applications, particular, comparison (RNA) in of problem","This problem finds applications, in particular, in the comparison of arc-annotated Ribonucleic acid (RNA) sequences.",22715
"In simple, this algorithm we this propose problem. a to hybrid tackle work evolutionary","In this work we propose a simple, hybrid evolutionary algorithm to tackle this problem.",22716
crossover concerns of The merging. solution this a on operator most important based feature algorithm,The most important feature of this algorithm concerns a crossover operator based on solution merging.,22717
heuristic outperforms is the literature. experimentally algorithm a from that shown the It proposed,It is experimentally shown that the proposed algorithm outperforms a heuristic from the literature.,22718
texts. important applications old images restoring of classical and document have Recognition in,Recognition of document images have important applications in restoring old and classical texts.,22719
the document such is to resolution improve Thus option only the of images.,Thus the only option is to improve the resolution of such document images.,22720
module are samples in using remaining feature. gallery ranked the the SR local The,The remaining gallery samples are ranked in the SR module using the local feature.,22721
"important Further, increase we most parameters find of for each the EDS. the","Further, we find the most important parameters for the increase of each EDS.",22722
growth rapid past a seen in has The omics technologies. decade,The past decade has seen a rapid growth in omics technologies.,22723
for Genome-wide traits. susceptibility a variants uncovered have complex association (GWAS) studies of variety,Genome-wide association studies (GWAS) have uncovered susceptibility variants for a variety of complex traits.,22724
"significance of not are discovered understood. most variants the fully still However, functional","However, the functional significance of most discovered variants are still not fully understood.",22725
The framework applied five complex was diseases. to,The framework was applied to five complex diseases.,22726
segmentation detection represents many in basis computer vision. and tasks Object for machine the and,Object detection and segmentation represents the basis for many tasks in computer and machine vision.,22727
from web gathered (a.k.a. Our on dataset experiments a the,Our experiments on a dataset gathered from the web (a.k.a.,22728
number display complex large between systems interactions of a components. of patterns Networked,Networked systems display complex patterns of interactions between a large number of components.,22729
network Our system's a architecture. of expectation the ground behavior dynamical its in results,Our results ground the expectation of a system's dynamical behavior in its network architecture.,22730
"denoising much last During researchers. methods attracted have decades, of the attention","During the last decades, denoising methods have attracted much attention of researchers.",22731
designed model. to of infer A Bayes posterior proposed algorithm the variational the is,A variational Bayes algorithm is designed to infer the posterior of the proposed model.,22732
in updated closed-form. parameters All recursively can involved be,All involved parameters can be recursively updated in closed-form.,22733
"CNNs literature, proposed their empirical. Although the have is been many in design","Although many CNNs have been proposed in the literature, their design is empirical.",22734
state-of-the-art among methods. The performance offers VCNN the all volume-based CNN proposed,The proposed VCNN offers the state-of-the-art performance among all volume-based CNN methods.,22735
"recorded standard typically from cameras. estimated sequences by optical frame is flow miniature However,","However, optical flow is typically estimated from frame sequences recorded by standard miniature cameras.",22736
real This event for is sequences. method validated,This method is validated for real event sequences.,22737
landing The a developed controller of a are constant on-board algorithms in divergence implemented quadrotor.,The developed algorithms are implemented in a constant divergence landing controller on-board of a quadrotor.,22738
maneuvers. perform landing very quadrotor enables fast to the This,This enables the quadrotor to perform very fast landing maneuvers.,22739
"a propose Siamese To we network. achieve this, novel","To achieve this, we propose a novel Siamese network.",22740
perform vehicle detection surveillance from traffic cameras. We fast,We perform fast vehicle detection from traffic surveillance cameras.,22741
technique co-segmentation collections. image We present a space-time for co-located,We present a co-segmentation technique for space-time co-located image collections.,22742
foreground technique collection. across belief background and the models Our entire progressively the expands,Our technique progressively expands the foreground and background belief models across the entire collection.,22743
"non-negative, is information, always which information interaction Unlike be (conditional) can mutual negative.","Unlike (conditional) mutual information, which is always non-negative, interaction information can be negative.",22744
efficiency increase medical in to image fusion medical Multimodal diagnosis. helps,Multimodal medical image fusion helps to increase efficiency in medical diagnosis.,22745
decomposition frequency is of coefficients. used low into and high for images the DTCWT,DTCWT is used for decomposition of the images into low and high frequency coefficients.,22746
by image fused The is DTCWT. reconstructed inverse,The fused image is reconstructed by inverse DTCWT.,22747
an question. human remains essential learners information exploit How this,How human learners exploit this information remains an essential question.,22748
"constraints process. such focus on govern Here, we a the temporal that","Here, we focus on the temporal constraints that govern such a process.",22749
feature extraction classification. and Seizure involves stage detection,Seizure detection stage involves feature extraction and classification.,22750
retrospective patient-specific. in and detection this paper Seizure is,Seizure detection in this paper is retrospective and patient-specific.,22751
"Portal. data Electro-physiology via is by provided Epilepsy International Kaggle, accessed iEEG","iEEG data is accessed via Kaggle, provided by International Epilepsy Electro-physiology Portal.",22752
"demonstrates variability the volume. high anatomical both very shape in and inter-patient pancreas Specifically, its","Specifically, the pancreas demonstrates very high inter-patient anatomical variability in both its shape and volume.",22753
with paper (e.g. capture deals of motion chains This kinematic,This paper deals with motion capture of kinematic chains (e.g.,22754
cameras. monocular taken skeletons) by uncalibrated from human sequences image,human skeletons) from monocular image sequences taken by uncalibrated cameras.,22755
on space based (KCS). a method chain present We observation a kinematic projecting an into,We present a method based on projecting an observation into a kinematic chain space (KCS).,22756
data state-of-the-art benchmark results We achieve world bases and real different on scenes.,We achieve state-of-the-art results on different benchmark data bases and real world scenes.,22757
a experts. as Our is mixture formulated of model,Our model is formulated as a mixture of experts.,22758
for Each a saliency related is expert of closely images. trained to network set predict,Each expert network is trained to predict saliency for a set of closely related images.,22759
predict scene is This guided global network by to information weights. gating,This gating network is guided by global scene information to predict weights.,22760
The simultaneously network are gating in trained an manner. and the end-to-end expert networks,The expert networks and the gating network are trained simultaneously in an end-to-end manner.,22761
several in and control Genetic biological cells processes. of oscillators many the are organisms present,Genetic oscillators are present in the cells of many organisms and control several biological processes.,22762
transcription of robustness relationship and the study the We protein bursty here oscillations. between,We study here the relationship between bursty transcription and the robustness of protein oscillations.,22763
disregard local information usually Pairwise structures. attributes but graph in involved measurements contextual consider,Pairwise measurements usually consider local attributes but disregard contextual information involved in graph structures.,22764
between issue address contextual of this proposing nodes. similarities by pairs We,We address this issue by proposing contextual similarities between pairs of nodes.,22765
of The set. COCO MS subset objects represent the label a,The objects represent a subset of the MS COCO label set.,22766
PDF the see file find the the to Please download URL to data.,Please see the PDF file to find the URL to download the data.,22767
samples representative based for of generating are Existing on posterior the estimation techniques posterior.,Existing techniques for posterior estimation are based on generating samples representative of the posterior.,22768
consider likelihood methods Such in efficiency not of evaluations. do terms,Such methods do not consider efficiency in terms of likelihood evaluations.,22769
an treat active framework. posterior query order to be estimation we In regression in efficient,In order to be query efficient we treat posterior estimation in an active regression framework.,22770
is deblurring. for framework Residual) PROBE Blur Removal blind (Progressive a recursive of,PROBE (Progressive Removal of Blur Residual) is a recursive framework for blind deblurring.,22771
"PROBE that analysis convergence Remarkably, itself lends properties. reveals its to","Remarkably, PROBE lends itself to analysis that reveals its convergence properties.",22772
"However, by factors. is time-dependent process exploration highly visual many a shaped dynamic","However, visual exploration is a highly dynamic process shaped by many time-dependent factors.",22773
be were and found Different and lifetimes discussed. values will anisotropy,Different lifetimes and anisotropy values were found and will be discussed.,22774
"dependency, the metric based information mutual presented. causal For quantifying is a","For quantifying the causal dependency, a mutual information based metric is presented.",22775
STPN approach based energy prediction is proposed An on subsequently the framework.,An energy prediction approach is subsequently proposed based on the STPN framework.,22776
"(FCSS), descriptor, convolutional correspondence. dense a present semantic fully We self-similarity called for","We present a descriptor, called fully convolutional self-similarity (FCSS), for dense semantic correspondence.",22777
outperforms that descriptors various Experiments FCSS on conventional descriptors handcrafted CNN-based and demonstrate benchmarks.,Experiments demonstrate that FCSS outperforms conventional handcrafted descriptors and CNN-based descriptors on various benchmarks.,22778
"In learning a we the scribble-based this as cast interactive semi-supervised segmentation problem. image paper,","In this paper, we cast the scribble-based interactive image segmentation as a semi-supervised learning problem.",22779
that images. of a are natural Experiments wide variety datasets carried contain out on,Experiments are carried out on datasets that contain a wide variety of natural images.,22780
We state-of-the-art quantitative segmentation compared achieve results better qualitative and interactive to algorithms.,We achieve better qualitative and quantitative results compared to state-of-the-art interactive segmentation algorithms.,22781
regression on spherical variables Spherical between relationships explores domains.,Spherical regression explores relationships between variables on spherical domains.,22782
We a nonparametric model uses itself. that sphere from to develop map a a diffeomorphic,We develop a nonparametric model that uses a diffeomorphic map from a sphere to itself.,22783
settings. diffeomorphisms to several restriction mapping natural is of this The in,The restriction of this mapping to diffeomorphisms is natural in several settings.,22784
in is maximum-likelihood optimization. penalized framework a The model using estimated gradient-based,The model is estimated in a penalized maximum-likelihood framework using gradient-based optimization.,22785
"goal, using that penalty Jacobian specify roughness first-order the a we diffeomorphisms. Towards of","Towards that goal, we specify a first-order roughness penalty using the Jacobian of diffeomorphisms.",22786
capturing found variables. model This to spherical relationships in between others is outperform,This model is found to outperform others in capturing relationships between spherical variables.,22787
results. Two the schemes algorithm colour presented of are restoration and improved drastically for enhancement,Two colour restoration and enhancement schemes of the algorithm are presented for drastically improved results.,22788
"spindle additional, these of forces position. by novel, the a found regulation We","We found a novel, additional, regulation of these forces by the spindle position.",22789
the ascribed at to cortex. microtubule dynamics the control this We,We ascribed this control to the microtubule dynamics at the cortex.,22790
"modulates turn, In pulling activity. generator it force","In turn, it modulates pulling force generator activity.",22791
generator superimposes that of This regulation processivity. force,This regulation superimposes that of force generator processivity.,22792
"this than in rather cell robustness biochemical originates mechanics Interestingly, networks.","Interestingly, this robustness originates in cell mechanics rather than biochemical networks.",22793
drop the Network presence rates recognition (CNN) noise. Convolutional of Neural in,Convolutional Neural Network (CNN) recognition rates drop in the presence of noise.,22794
"favorably and recognition computational system complexity in robustness, Our rate. of terms compares","Our system compares favorably in terms of robustness, computational complexity and recognition rate.",22795
depends the its and cross-section This source on term area cable derivates.,This source term depends on the cable cross-section area and its derivates.,22796
"swelling solutions. and different cables their In study we numerical addition, with provide","In addition, we study different cables with swelling and provide their numerical solutions.",22797
"the generalized extend we the paper, distribution. this to distribution In Poisson-binomial (GPB) Poisson-binomial","In this paper, we extend the Poisson-binomial distribution to the generalized Poisson-binomial (GPB) distribution.",22798
study computational We settings. time in algorithm the of the various parameter also,We also study the computational time of the algorithm in various parameter settings.,22799
geo-localization features. a on local problem the cast image We clustering as,We cast the geo-localization as a clustering problem on local image features.,22800
in be or tested that predictions results can constitute in vivo vitro. specific Our,Our results constitute specific predictions that can be tested in vivo or in vitro.,22801
constants Optical chemical with are used associated measure biosensors often reactions. to rate kinetic,Optical biosensors are often used to measure kinetic rate constants associated with chemical reactions.,22802
sensor components process multiple to experiments interacting involve on Biosensor this surface. the study,Biosensor experiments to study this process involve multiple interacting components on the sensor surface.,22803
modeled a equations of multiple-component experiment set with nonlinear is This biosensor (IDEs). integrodifferential,This multiple-component biosensor experiment is modeled with a set of nonlinear integrodifferential equations (IDEs).,22804
of on network. a study spreading a We model stochastic infection,We study a stochastic model of infection spreading on a network.,22805
"neighbor If node is the is the infected susceptible, infected. and the becomes neighbor","If the node is infected and the neighbor is susceptible, the neighbor becomes infected.",22806
task. application It requirement detection essential for is the an saliency,It is an essential application requirement for the saliency detection task.,22807
simplified is deep neural any It pre-processings an extremely without and network post-processings. end-to-end,It is an extremely simplified end-to-end deep neural network without any pre-processings and post-processings.,22808
a like procedure. pipeline detection simplifies compact works the It effectively which,It works like a compact pipeline which effectively simplifies the detection procedure.,22809
primary of autonomous the vehicles. for one Safe with drivers is human challenges interaction,Safe interaction with human drivers is one of the primary challenges for autonomous vehicles.,22810
the are recognition systems is used that widely a Iris world. mature throughout technology,Iris recognition systems are a mature technology that is widely used throughout the world.,22811
"method effectiveness image the In shown. denoising this the for is literature, of","In the literature, the effectiveness of this method for image denoising is shown.",22812
bilateral filter domain the new framework. denoising a wavelet image the complex Using in forms,Using the bilateral filter in the complex wavelet domain forms a new image denoising framework.,22813
aesthetics composition is important factor the photography. affecting Photo in an,Photo composition is an important factor affecting the aesthetics in photography.,22814
maritime in This vessels. autonomous resides the architecture,This architecture resides in the autonomous maritime vessels.,22815
"Once are least parameters unknown estimated by regression model selected, a squares. is","Once a model is selected, unknown regression parameters are estimated by least squares.",22816
paper forward This in analyzes linear regression high-dimensional models. sparse,This paper analyzes forward regression in high-dimensional sparse linear models.,22817
for Probabilistic number bounds of prediction proved. are norm selected error covariates and,Probabilistic bounds for prediction error norm and number of selected covariates are proved.,22818
several explored gender the researchers. based texture by has been Predicting iris a person's on,Predicting a person's gender based on the iris texture has been explored by several researchers.,22819
affect sources that We discuss and the generated data various representations content. resulting,We discuss various data sources and representations that affect the resulting generated content.,22820
of prior remedied is This account to be domain-dependent likely knowledge. by into taking,This is likely to be remedied by taking into account of domain-dependent prior knowledge.,22821
"integrated introduce We stereo. photometric approach uncalibrated new, to a","We introduce a new, integrated approach to uncalibrated photometric stereo.",22822
optimization then the solve Multipliers Method of Alternate using Direction We (ADMM). this,We then solve this optimization using the Alternate Direction Method of Multipliers (ADMM).,22823
experimental extensive sets. on synthetic real evaluation and conduct We data,We conduct extensive experimental evaluation on real and synthetic data sets.,22824
prior We in these improvements cases. good work over show,We show good improvements over prior work in these cases.,22825
is concluded open with a description problems. of The survey some,The survey is concluded with a description of some open problems.,22826
"remains mechanism the elusive. However, of CQ action and Plasmodium resistance in falciparum","However, the mechanism of CQ action and resistance in Plasmodium falciparum remains elusive.",22827
estimate the and online through We global loop-closing in optimization. non-linear achieve consistency,We achieve global consistency in the estimate through online loop-closing and non-linear optimization.,22828
"mechanism the object visual in methods. current However, state-of-the-art missing detection a such is","However, such a mechanism is missing in the current state-of-the-art visual object detection methods.",22829
based The Time Dynamic Warping. method is on,The method is based on Dynamic Time Warping.,22830
"graph, themselves Shapes in an RDF shapes called expressed as a triples are RDF graph.","Shapes are themselves expressed as RDF triples in an RDF graph, called a shapes graph.",22831
basic in idea color the The information exploit the is image. shading to,The basic idea is to exploit the shading information in the color image.,22832
are stereo a We photometric moving from under calculating effectively illuminations. object natural,We are effectively calculating photometric stereo from a moving object under natural illuminations.,22833
image establish key set. technical challenge The the is over correspondences entire to,The key technical challenge is to establish correspondences over the entire image set.,22834
for software designed are specifically forest its BEE inventory. scanner supporting The and,The BEE scanner and its supporting software are specifically designed for forest inventory.,22835
experiments. square Four to do selected were the plots,Four square plots were selected to do the experiments.,22836
The single and the acquired plots data in four scanner the BEE scanned respectively. scan,The BEE scanner scanned in the four plots and acquired the single scan data respectively.,22837
"collected manual plots. the measured comparison, was the in data four also For","For comparison, the manual measured data was also collected in the four plots.",22838
"a from Additionally, estimate propose to approach new video. model we background","Additionally, we propose a new approach to estimate background model from video.",22839
spatial-median also post-processing utilized filtering network the We the outputs. of as,We also utilized spatial-median filtering as the post-processing of the network outputs.,22840
"processing. Furthermore, the real of time due capable to is network architecture, CNN our","Furthermore, due to the network architecture, our CNN is capable of real time processing.",22841
are models in reconstruction refined an using image an unsupervised The further fashion loss.,The models are further refined in an unsupervised fashion using an image reconstruction loss.,22842
organization stereo synthesis are standard three presented. frame The dimension features of of format of,The features of synthesis organization of standard format of three dimension stereo frame are presented.,22843
"circuit relate can this way, parameters information to processing. we In","In this way, we can relate circuit parameters to information processing.",22844
minimizes data algorithm likelihood. complexity The captures model and,The algorithm minimizes model complexity and captures data likelihood.,22845
to Process-Hidden Algorithm Markov baseline Dirichlet better is the perform (HDP-HMM). Hierarchical demonstrated than Models,Algorithm is demonstrated to perform better than the baseline Hierarchical Dirichlet Process-Hidden Markov Models (HDP-HMM).,22846
"computational complexity real time for algorithm makes it low of operations. The on-board, suitable","The low computational complexity of algorithm makes it suitable for on-board, real time operations.",22847
world. popular of language is most the languages one Arabic in the,Arabic language is one of the most popular languages in the world.,22848
Each ellipses. sequence of code a contains and lines,Each code contains a sequence of lines and ellipses.,22849
to images is challenging counting Crowd a on due problem variations. static scale,Crowd counting on static images is a challenging problem due to scale variations.,22850
to be this been shown have Recently in deep task. effective neural networks,Recently deep neural networks have been shown to be effective in this task.,22851
"automatic aging. this face propose method the we In for GAN-based work,","In this work, we propose the GAN-based method for automatic face aging.",22852
"To approach vectors. this introduce we GAN's optimization latent novel a ""Identity-Preserving"" of end, for","To this end, we introduce a novel approach for ""Identity-Preserving"" optimization of GAN's latent vectors.",22853
of We a structured-output for training propose neural networks. semi-supervised method,We propose a method for semi-supervised training of structured-output neural networks.,22854
a of the We use as unlabelled then data. error for source discriminator signal,We then use the discriminator as a source of error signal for unlabelled data.,22855
test on set. boosts network a performance a of the held out This effectively,This effectively boosts the performance of a network on a held out test set.,22856
be if full wavelet Matched designed can image available. is easily,Matched wavelet can be easily designed if full image is available.,22857
steps. is The in two recognition achieved,The recognition is achieved in two steps.,22858
"the temporal spatial from First, and we features data. extract multimodal","First, we extract spatial and temporal features from the multimodal data.",22859
are fused second in step. making features The decision for then extracted the,The extracted features are then fused for decision making in the second step.,22860
optimal squares. derived The are alignment parameters analytically least via linear,The optimal alignment parameters are derived analytically via linear least squares.,22861
"that After are minutia exhibit correspondence pairs weak alignment, each discarded.","After each alignment, minutia pairs that exhibit weak correspondence are discarded.",22862
work robust series algorithm time clustering stationary a for In this proposed. is,In this work a robust clustering algorithm for stationary time series is proposed.,22863
for to algorithm robust set of is spectral the densities. applied data A functional then,A robust algorithm for functional data is then applied to the set of spectral densities.,22864
"it often However, to leads errors. quantization","However, it often leads to quantization errors.",22865
"the max-margin work, to classifier we efficient this use hash In an learn function.","In this work, we use the max-margin classifier to learn an efficient hash function.",22866
"it implemented. effectively simple parallelism, efficiently be algorithmically it supports can is and Since","Since it is algorithmically simple and effectively supports parallelism, it can efficiently be implemented.",22867
"propose autonomous paper, braking new system a learning. reinforcement on In we based this deep","In this paper, we propose a new autonomous braking system based on deep reinforcement learning.",22868
is achieved commands This scripting within simple CHASE. by,This is achieved by scripting simple commands within CHASE.,22869
An of supplied. is functionality CHASE additional,An additional functionality of CHASE is supplied.,22870
"brain's is its architecture in connectome, The functional meso-scale reflected the diversity of i.e.","The brain's functional diversity is reflected in the meso-scale architecture of its connectome, i.e.",22871
division communities of and brain topologically-related into clusters its regions.,its division into clusters and communities of topologically-related brain regions.,22872
from conduct simple line based synthesis we Then AD-VAE. photo drawing on,Then we conduct photo synthesis from simple line drawing based on AD-VAE.,22873
"the structure network. of terms neuronal we in the the dynamics of study Furthermore, bifurcation","Furthermore, we study the neuronal dynamics in terms of the bifurcation structure of the network.",22874
extensively simulations. by numerical analytical the results All validated are,All the analytical results are extensively validated by numerical simulations.,22875
"and Jukes Cantor, They the Tamura. encompass models for instance and Kimura, of well-known","They encompass for instance the well-known models of Jukes and Cantor, Kimura, and Tamura.",22876
binary based recognition novel This patterns. a paper approach on presents local face automatic,This paper presents a novel automatic face recognition approach based on local binary patterns.,22877
compute of a neighbourhood the descriptor values. vector considers to feature pixel a This local,This descriptor considers a local neighbourhood of a pixel to compute the feature vector values.,22878
"illumination handle not different is robust method noise, to image variances very conditions. This and","This method is not very robust to handle image noise, variances and different illumination conditions.",22879
"The evaluated face on method namely is corpora, proposed UFI datasets. and FERET benchmark two","The proposed method is evaluated on two benchmark corpora, namely UFI and FERET face datasets.",22880
a descriptors Fisher-Vectors set (FV) multiple like features. local higher-order of encode statistics of SIFT,Fisher-Vectors (FV) encode higher-order statistics of a set of multiple local descriptors like SIFT features.,22881
visual good in architectures learning recognitions shallow with They performance tasks. show on already combination,They already show good performance in combination with shallow learning architectures on visual recognitions tasks.,22882
processing feature decisions resulting of considered. from is Fusion the each also,Fusion of the decisions resulting from processing each feature is also considered.,22883
that existing indicate The outperform results learning methods. approaches the new feature,The results indicate that the new feature learning approaches outperform existing methods.,22884
"models inter-pixel that process construct one-state correlations Markov first we Specifically, and anti-correlations. both a","Specifically, we first construct a one-state Markov process that models both inter-pixel correlations and anti-correlations.",22885
is Q semi-definite. G self-loops importance positive of We the in to the ensure show,We show the importance of the self-loops in G to ensure Q is positive semi-definite.,22886
"is that estimation. alpha produces reconstruction least by The the (F,B) selected, pair followed error","The (F,B) pair that produces the least reconstruction error is selected, followed by alpha estimation.",22887
unexamined. of error The that residual left significance has been,The significance of that residual error has been left unexamined.,22888
the hyperspectral and last has image analysis much development in experienced Multispectral decade.,Multispectral and hyperspectral image analysis has experienced much development in the last decade.,22889
an paper presents recognition face images. approach This gender from unconstrained for,This paper presents an approach for gender recognition from unconstrained face images.,22890
"we our To potentials method, experiments evaluate real perform limitations on data. and of","To evaluate potentials and limitations of our method, we perform experiments on real data.",22891
key problem and is in a recognition. Texture image understanding characterization pattern,Texture characterization is a key problem in image understanding and pattern recognition.,22892
"shape-based shape flexible a paper, using present patterns. we this In representation texture co-occurrence","In this paper, we present a flexible shape-based texture representation using shape co-occurrence patterns.",22893
datasets the the texture demonstrate proposed various efficiency scene datasets on method. Experiments of and,Experiments on various texture datasets and scene datasets demonstrate the efficiency of the proposed method.,22894
"solves a SSS*, it algorithm. drawbacks the all It finally perceived of practical transforming into","It solves all the perceived drawbacks of SSS*, finally transforming it into a practical algorithm.",22895
"that effect, In we show + alpha-beta tables. ransposition SSS* =","In effect, we show that SSS* = alpha-beta + ransposition tables.",22896
focuses essential attention between algorithms. It differences on the,It focuses attention on the essential differences between algorithms.,22897
"present new a instance MTD(f). of framework, (C) We the","(C) We present a new instance of the framework, MTD(f).",22898
semi-dense solution efficient paper and This presents odometry visual for a cameras. robust RGB-D,This paper presents a robust and efficient semi-dense visual odometry solution for RGB-D cameras.,22899
"measurements use features. as we anthropometric Hence, ratios of","Hence, we use ratios of anthropometric measurements as features.",22900
"predictions perform with then estimation. observable are The to along height used, features,","The predictions are then used, along with observable features, to perform height estimation.",22901
"is estimated, the height performed. is Once to mapping a classes","Once the height is estimated, a mapping to classes is performed.",22902
camera from in estimation an structure Accurate matrices of step is algorithms. motion important,Accurate estimation of camera matrices is an important step in structure from motion algorithms.,22903
signal representations into over the The spaced densely input wavelets Dual-Tree computationally scales. decompose efficient,The computationally efficient Dual-Tree wavelets decompose the input signal into densely spaced representations over scales.,22904
these of laws Sampling bias. versions a discretized the introduce typically from can,Sampling from the discretized versions of these laws can typically introduce a bias.,22905
"error. discretization, the a given of from sampling precise for level most","sampling from the most precise discretization, for a given level of error.",22906
of new the results approach. for optimality and variance We criteria regarding extensions cover,We cover extensions of results regarding variance and optimality criteria for the new approach.,22907
Both complemented numerical applications simulations. by are,Both applications are complemented by numerical simulations.,22908
deep training lack Supervised the sufficient suffers learning of often data. from,Supervised deep learning often suffers from the lack of sufficient training data.,22909
on written Rcpp R in online. the proposed available algorithm is and package based,R package based on the proposed algorithm and written in Rcpp is available online.,22910
"establish popular Recently, become regression a forests alternative such to correspondences. have","Recently, regression forests have become a popular alternative to establish such correspondences.",22911
"Compared model completely training. with is multi-model end-to-end ensemble, our","Compared with multi-model ensemble, our model is completely end-to-end training.",22912
by images (visual) follow physical environment. of inherited Natural structure the our statistics,Natural images follow statistics inherited by the structure of our physical (visual) environment.,22913
architecture We biologically-inspired cortex. the visual the of a sparse designed by coding algorithm primary,We designed a sparse coding algorithm biologically-inspired by the architecture of the primary visual cortex.,22914
of this here that coefficients We heavy-tailed a distribution. show exhibit representation,We show here that coefficients of this representation exhibit a heavy-tailed distribution.,22915
"structured-knowledge, online. a IRT-based student for temporal,  We method develop predicting T-SKIRT: responses ","  We develop T-SKIRT: a temporal, structured-knowledge, IRT-based method for predicting student responses online.",22916
"system practical novel, real-time and identification. and We introduce a accurate people tracking for","We introduce a novel, accurate and practical system for real-time people tracking and identification.",22917
system preserve lighting. work and Our with user different can privacy,Our system can preserve user privacy and work with different lighting.,22918
"such robustness CNNs other and as to transformations limited geometric have However, rotation. a scaling","However, CNNs have a limited robustness to other geometric transformations such as scaling and rotation.",22919
"established but solution. deep limits of no the performance improvement there is CNNs, the This","This limits the performance improvement of the deep CNNs, but there is no established solution.",22920
"unsuitable those fast convergence. present genetic and In algorithms a cases,","In those cases, genetic algorithms present a fast and unsuitable convergence.",22921
It problem diversification confronts greedy operator. diversity named the using the,It confronts the diversity problem using the named greedy diversification operator.,22922
system networks solve on We to designed based this problem. neural a,We designed a system based on neural networks to solve this problem.,22923
"this ground-truth With end-to-end. model labels, trained volume was","With ground-truth volume labels, this model was trained end-to-end.",22924
"To used. was only the additional with an label segmentation improve result, dataset","To improve the result, an additional dataset with only segmentation label was used.",22925
The model datasets on with different teaching alternately types of signals. two trained was these,The model was trained alternately on these two datasets with different types of teaching signals.,22926
final proposed We a estimation the method variance also prediction. for,We also proposed a variance estimation method for the final prediction.,22927
that a Face Entropy implements reference the called game requirements. We Battle present Samu,We present a reference game called Face Battle that implements the Samu Entropy requirements.,22928
"for learning image in unlabeled By semi-supervised data manner. architecture allows a from our design,","By design, our architecture allows for learning from unlabeled image data in a semi-supervised manner.",22929
"instances. to slow large Their however, algorithm, many branch-and-cut is convergence and for prone cuts","Their branch-and-cut algorithm, however, is prone to many cuts and slow convergence for large instances.",22930
"moral instances, and to larger practical tracing lineage leveraging to scale significance.","and scale to larger instances, leveraging moral lineage tracing to practical significance.",22931
"a implements models. latent several models, including spatial number used of widely R-INLA","R-INLA implements a number of widely used latent models, including several spatial models.",22932
Monte Markov take Chain the fit same model. to Carlo),Markov Chain Monte Carlo) take to fit the same model.,22933
are sea used cameras the floor. widely to observe Underwater,Underwater cameras are widely used to observe the sea floor.,22934
"summarize designing the and trends and imaging processing challenges sensors. future underwater in we Finally,","Finally, we summarize the future trends and challenges in designing and processing underwater imaging sensors.",22935
per using (SSPP) single face sample recognition Real-world task. a is person challenging a,Real-world face recognition using a single sample per person (SSPP) is a challenging task.,22936
approach show Experiments the performance. benchmark proposed state-of-the-art on using dataset a,Experiments on a benchmark dataset using the proposed approach show state-of-the-art performance.,22937
detection tasks the computer in of vision. Face is challenging one,Face detection is one of the challenging tasks in computer vision.,22938
rate decrease can face noticeably. detection variations These,These variations can decrease face detection rate noticeably.,22939
features and based primary detection The is the Adaboost algorithm. Haar-like on,The primary detection is based on Haar-like features and the Adaboost algorithm.,22940
validation Extended is Binary two-stage based Local on Patterns. process The,The validation process is based on two-stage Extended Local Binary Patterns.,22941
"absence to truth. due ground this achieve to of difficult is However,","However, this is difficult to achieve due to absence of ground truth.",22942
deployment known performance reference is unavailable. But about is the real a little after when,But little is known about the real performance after deployment when a reference is unavailable.,22943
with and approach classifiers segmentation We validate different multi-organ methods. segmentation our on,We validate our approach on multi-organ segmentation with different classifiers and segmentation methods.,22944
our of function allow is The work innovation loss the to key w.r.t. back-propagation of,The key innovation of our work is to allow back-propagation of the loss function w.r.t.,22945
labels the the generated discriminator. the (randomly to from images) each generator to assigned,the labels (randomly assigned to each generated images) to the generator from the discriminator.,22946
generalize based from that computer seen vision previously Generally data are used. methods,Generally computer vision based methods that generalize from previously seen data are used.,22947
"of For necessary. such methods, large training data amounts annotated are","For training such methods, large amounts of annotated data are necessary.",22948
"CNN can be we that workflow for Furthermore, a surgical demonstrate adapted such segmentation.","Furthermore, we demonstrate that such a CNN can be adapted for surgical workflow segmentation.",22949
contains views sign. Each name several street of image same the,Each image contains several views of the same street name sign.,22950
a constraint set answer These in new area: efforts programming. resulted research,These efforts resulted in a new research area: constraint answer set programming.,22951
by of systems transition et proposed develop Nieuwenhuis al. the an extension also We,We also develop an extension of the transition systems proposed by Nieuwenhuis et al.,22952
about extension this to We EZCSP the it. formal describe algorithm claims prove use and,We use this extension to describe the EZCSP algorithm and prove formal claims about it.,22953
in Under and of Practice Logic (TPLP). Theory Programming consideration,Under consideration in Theory and Practice of Logic Programming (TPLP).,22954
properties SAT. devote the theoretical time discussing to of considerable We,We devote considerable time to discussing the theoretical properties of SAT.,22955
has tried To before. been of the best this knowledge never our,To the best of our knowledge this has never been tried before.,22956
"by methods, consistently led find outperform that non-linear approaches. We BCSD, linear","We find that linear methods, led by BCSD, consistently outperform non-linear approaches.",22957
to for up bring hybrid model computing. promises architecture a Neuro-electronic,Neuro-electronic hybrid promises to bring up a model architecture for computing.,22958
world tasks now. solving for higher with Such demand on accuracy paradigms real is,Such paradigms for solving real world tasks with higher accuracy is on demand now.,22959
a a autonomous navigate to line. A robot particular modeled system is as here following,A robot as a autonomous system is modeled here to navigate following a particular line.,22960
"exploratory regression but challenging. become more more collections becomes analysis As data larger, important","As data collections become larger, exploratory regression analysis becomes more important but more challenging.",22961
called decision metboost We to data. clustered trees boosted for decision hierarchically propose extension an,We propose an extension to boosted decision decision trees called metboost for hierarchically clustered data.,22962
Analysis years. Expression has grown in recent Automatic of Facial The rapidly field,The field of Automatic Facial Expression Analysis has grown rapidly in recent years.,22963
"Two detection and AU occurrence, intensity. the of the are estimation AU of sub-challenges defined:","Two sub-challenges are defined: the detection of AU occurrence, and the estimation of AU intensity.",22964
"many common Over-segmentation, vision applications. computer a is generation, preliminary or for super-pixel stage","Over-segmentation, or super-pixel generation, is a common preliminary stage for many computer vision applications.",22965
algorithm We Variation new (PCLV). Point Local this call Cloud,We call this new algorithm Point Cloud Local Variation (PCLV).,22966
"(RFFPSR). framework, propose Representation Based Fourier State on the we Random with this Predictive Features","Based on this framework, we propose the Predictive State Representation with Random Fourier Features (RFFPSR).",22967
"to determine relative enables their identify and This proportions. cell to sub-types,","This enables to identify cell sub-types, and to determine their relative proportions.",22968
"whose highly high-dimensional poorly analysis datasets, manual in reproducible. time-consuming results This is and","This results in high-dimensional datasets, whose manual analysis is highly time-consuming and poorly reproducible.",22969
"independent individual are (e.g. rarely samples However, in practice,","However, in practice, individual samples are rarely independent (e.g.",22970
We community in the for available have package method made NPflow. R the the,We have made the method available for the community in the R package NPflow.,22971
"than Our slice-by-slice, patches. faster method segment could brain much on image those images based","Our method could segment brain images slice-by-slice, much faster than those based on image patches.",22972
from non-overlapping match aims Re-IDentification cameras. Person (Re-ID) person images to two captured,Person Re-IDentification (Re-ID) aims to match person images captured from two non-overlapping cameras.,22973
"learning, classification a for computer standard dataset systems. vision become MNIST The and benchmark has","The MNIST dataset has become a standard benchmark for learning, classification and computer vision systems.",22974
can refutation called be also This theorem-proving.,This can also be called refutation theorem-proving.,22975
and systems control still intervention. surveillance and human supervision require Current,Current surveillance and control systems still require human supervision and intervention.,22976
(MTL) face recognition. multi-task This for paper learning explores,This paper explores multi-task learning (MTL) for face recognition.,22977
questions and improve recognition face the of why MTL the can how answer performance. We,We answer the questions of how and why MTL can improve the face recognition performance.,22978
approach. the Multi-PIE experiments the dataset the proposed of demonstrate Extensive effectiveness entire on,Extensive experiments on the entire Multi-PIE dataset demonstrate the effectiveness of the proposed approach.,22979
an active fields. research Handwritten character area recognition numerous of applications is with in,Handwritten character recognition is an active area of research with applications in numerous fields.,22980
this concentrated recent works various languages. have Past on in and field,Past and recent works in this field have concentrated on various languages.,22981
paper. detectors popular in a mentioned contour-based corner of our number We have,We have mentioned a number of popular contour-based corner detectors in our paper.,22982
curvature introduce the method paper. a value of calculate effective We this new to in,We introduce a new effective method to calculate the value of curvature in this paper.,22983
"on ambiguously datasets. to our approach methods, labeled existing Compared improvement several demonstrates","Compared to existing methods, our approach demonstrates improvement on several ambiguously labeled datasets.",22984
in is image fundamental a Image multispectral registration processing. issue,Image registration is a fundamental issue in multispectral image processing.,22985
two spectral response challenges characteristic selective The registration. multispectral in image raises of imaging to,The selective characteristic of spectral response in multispectral imaging raises two challenges to image registration.,22986
"the individual in intensity local levels First, region channel images. a may of be different","First, the intensity levels of a local region may be different in individual channel images.",22987
between images. The NTG the two is applied difference on channel,The NTG is applied on the difference between two channel images.,22988
"experiments the we examine additive Lastly, of present empirical trend performance filtering. to","Lastly, we present experiments to examine the empirical performance of additive trend filtering.",22989
"a estimate employed. approach probabilistic To Bayesian is the a obtain interest, quantity of of","To obtain a probabilistic estimate of the quantity of interest, a Bayesian approach is employed.",22990
deriving algorithm presented. An dictionary observable is an such for,An algorithm for deriving such an observable dictionary is presented.,22991
either a queen A workers. or sterilizes chemically that behaviorally scenario long-discussed her is,A long-discussed scenario is that a queen either behaviorally or chemically sterilizes her workers.,22992
"demographic are such ecological and the unknown. However, manipulation conditions that enable","However, the demographic and ecological conditions that enable such manipulation are unknown.",22993
"dynamics that Accordingly, propose based evolutionary simple on genetics. a model of haplodiploid is we","Accordingly, we propose a simple model of evolutionary dynamics that is based on haplodiploid genetics.",22994
queen's mating conditions depend the frequency. not These on do,These conditions do not depend on the queen's mating frequency.,22995
the map addresses occupancy estimation. multi-view problem people paper of This,This paper addresses the problem of multi-view people occupancy map estimation.,22996
"subtraction background on either Existing a per-view, operate or rely for solutions this pre-processing. problem","Existing solutions for this problem either operate per-view, or rely on a background subtraction pre-processing.",22997
more as lessen detection approaches scenes performance Both the become crowded.,Both approaches lessen the detection performance as scenes become more crowded.,22998
of this the to data-set. In multi-camera large-scale is due part lack large,In large part this is due to the lack of large-scale multi-camera data-set.,22999
"available we Furthermore, a HD make publicly new three-camera data-set.","Furthermore, we make publicly available a new three-camera HD data-set.",23000
made trained under and an models Our license. open-source will source be code available,Our source code and trained models will be made available under an open-source license.,23001
data surfaces. sets modern sampled error Many complex high-dimensional are with from,Many modern data sets are sampled with error from complex high-dimensional surfaces.,23002
"and systems. be non-Gaussian method successfully to extremely the non-linear can applied Furthermore,","Furthermore, the method can be successfully applied to extremely non-linear and non-Gaussian systems.",23003
"of via analysis validate we Finally, measured the empirically our brain signals. approach","Finally, we empirically validate our approach via the analysis of measured brain signals.",23004
evaluation performance the proposed results the demonstrated method. state-of-the-art of The,The evaluation results demonstrated the state-of-the-art performance of the proposed method.,23005
"in problem. a first the learning position adaptation domain motivation, After general we transfer larger","After a general motivation, we first position domain adaptation in the larger transfer learning problem.",23006
a The camera. speed measurement presents the of smartphone report using vehicular,The report presents the measurement of vehicular speed using a smartphone camera.,23007
is to actual determined to distance. vehicle the displacement map pixel Conversion coefficient,Conversion coefficient is determined to map the pixel displacement to actual vehicle distance.,23008
truth to speeds proportional the speeds. are ground measured The,The speeds measured are proportional to the ground truth speeds.,23009
state-of-the-art. current show the performance the COCO-text on datasets over Results superior and RRC ICDAR,Results on the ICDAR RRC and the COCO-text datasets show superior performance over current state-of-the-art.,23010
reconstruction distribution resampling We error. form optimal of minimizing obtain general a by the,We obtain a general form of optimal resampling distribution by minimizing the reconstruction error.,23011
empirical proposed methods. efficiency effectiveness performance of resampling The and the validates the,The empirical performance validates the effectiveness and efficiency of the proposed resampling methods.,23012
high and convolutional Deep are for demands. their networks computational memory well-known,Deep convolutional networks are well-known for their high computational and memory demands.,23013
leads only accuracy. many minimal changes to cases smaller this significantly to In networks with,In many cases this leads to significantly smaller networks with only minimal changes to accuracy.,23014
assumptions. regression Examples structured sparsity include under,Examples include regression under structured sparsity assumptions.,23015
"analysis of instances this gap known some For our the closes theory. in continuum,","For some known instances of this continuum, our analysis closes the gap in theory.",23016
exploit further algorithms. to the We a propose continuum accelerated unification of,We further exploit the unification to propose a continuum of accelerated algorithms.,23017
that continuum whole of rate We theoretically show the the optimal attains convergence.,We show that the whole continuum attains the theoretically optimal rate of convergence.,23018
of different aspects highlight data Visualizations simultaneously the across can dimensions. different,Visualizations can highlight different aspects of the data simultaneously across different dimensions.,23019
seeking tool to fields this pairwise in envision connectivity. interactive for applications We visualize,We envision applications for this interactive tool in fields seeking to visualize pairwise connectivity.,23020
This branched residual image a for paper proposes classification. network,This paper proposes a branched residual network for image classification.,23021
"network the to high-level memory low-level sharing the features, features. can By allocate more","By sharing the low-level features, the network can allocate more memory to high-level features.",23022
is image task vision. challenging fundamental computer in but Multi-label a classification,Multi-label image classification is a fundamental but challenging task in computer vision.,23023
recent achieved relations semantic exploiting by has in progress been years. between Great labels,Great progress has been achieved by exploiting semantic relations between labels in recent years.,23024
approach. on of the three confirm power benchmarks experiments public Thorough our we different,Thorough experiments on three different public benchmarks we confirm the power of our approach.,23025
from traditional computation. algorithms image Most suffer compressive the sensing intensive reconstruction for,Most traditional algorithms for compressive sensing image reconstruction suffer from the intensive computation.,23026
"network. fully-connected the linear in layer Specifically, network mapping neural the implements","Specifically, the fully-connected layer in neural network implements the linear mapping network.",23027
"An for accompanying ground-truthing also Truth labeling, Crop, introduced. is superpixel and tool","An accompanying ground-truthing tool for superpixel labeling, Truth and Crop, is also introduced.",23028
for challenges discussed. are research and Open directions future,Open challenges and directions for future research are discussed.,23029
"effective image we segmentation. paper, fast In a for this but method simple propose","In this paper, we propose a simple but effective method for fast image segmentation.",23030
to a partitioning. We conduct function present efficiently novel energy graph,We present a novel energy function to efficiently conduct graph partitioning.,23031
"their Defect-free fabric the direction, while have direction. defects specified damage images of pattern regularity","Defect-free pattern fabric images have the specified direction, while defects damage their regularity of direction.",23032
captioning a the In we problem as formulate image multimodal this of task. translation work,In this work we formulate the problem of image captioning as a multimodal translation task.,23033
annotation a assumption. rely on Current sparsity strong approaches,Current annotation approaches rely on a strong sparsity assumption.,23034
our performance the reflection and of method real using demonstrate images. We synthesized,We demonstrate the performance of our method using synthesized and real reflection images.,23035
verify multiple object benchmarks. We on detection method our proposed,We verify our proposed method on multiple object detection benchmarks.,23036
"recover fine geometric medium shape-from-shading to on the method details. face is a Finally, applied","Finally, a shape-from-shading method is applied on the medium face to recover fine geometric details.",23037
et Statistical al. mechanics Background: results (Dauphin,Background: Statistical mechanics results (Dauphin et al.,23038
significantly in extends optimization. and convex SAGA non-convex of It applicability the,It significantly extends the applicability of SAGA in convex and non-convex optimization.,23039
positives rate It false to is control which important occur. at the,It is important to control the rate at which false positives occur.,23040
"on Carlo (MCMC). however, computationally-intensive This Markov relies method, Monte Chain","This method, however, relies on computationally-intensive Markov Chain Monte Carlo (MCMC).",23041
"This relevant at considers sequentially, data with updated each scheme items probabilities step. being","This scheme considers data items sequentially, with relevant probabilities being updated at each step.",23042
"for this a zoom-out-and-in generating object proposals. propose paper, we network In","In this paper, we propose a zoom-out-and-in network for generating object proposals.",23043
to region-level small objects. semantics contain shallow counterparts feature Deeper which maps help identify can,Deeper feature maps contain region-level semantics which can help shallow counterparts to identify small objects.,23044
In testing this approach the a for consider problem. paper we uniform debiasing,In this paper we consider a debiasing approach for the uniform testing problem.,23045
of studies illustrate our sample Simulation finite properties procedure.,Simulation studies illustrate finite sample properties of our procedure.,23046
"algorithm this for paper multi-view Accordingly, novel a averaging proposes registration. motion","Accordingly, this paper proposes a novel motion averaging algorithm for multi-view registration.",23047
is used to fine A method Levenberg-Marquardt optimization. for coarse approach with,A coarse to fine approach with Levenberg-Marquardt method is used for optimization.,23048
"problem In of paper, study the person language with this description. search we natural","In this paper, we study the problem of person search with natural language description.",23049
"Networks Region-based Human-Object To we solve (HO-RCNN). the propose Convolutional task, Neural","To solve the task, we propose Human-Object Region-based Convolutional Neural Networks (HO-RCNN).",23050
applicable based generally Our framework map to confidence reinforcement algorithms. proposed tracking other learning is,Our proposed reinforcement learning framework is generally applicable to other confidence map based tracking algorithms.,23051
disease detection motivates of the and Early outcome work. influences our the,Early detection influences the outcome of the disease and motivates our work.,23052
These users. its to platform the target Gamification findings help,These findings help the Gamification platform to target its users.,23053
"on datasets. our real we To method, experiments two evaluate perform","To evaluate our method, we perform experiments on two real datasets.",23054
hard Value-Related Dependencies and Strengths of are to specify. such assumed to imprecise be,Strengths of such Value-Related Dependencies are assumed to be imprecise and hard to specify.,23055
is a segmentation. results that iaSTAPLE These crowd image sourcing useful tool show highly for,These results show that iaSTAPLE is a highly useful tool for crowd sourcing image segmentation.,23056
raised CT Increasing radiation in dose. practice medical has concerns of associated use modern over,Increasing use of CT in modern medical practice has raised concerns over associated radiation dose.,23057
Promising in been obtained large experiments of number with images. our a have CT results,Promising results have been obtained in our experiments with a large number of CT images.,23058
present parametric dynamic a for textures. Here we model,Here we present a parametric model for dynamic textures.,23059
"a Knowledge entails i.e. (KP): Paradox This which I term paradox,","This entails a paradox, which I term Knowledge Paradox (KP): i.e.",23060
a the KP paradoxes. for sorites and unifying liar The context provides the,The KP provides a unifying context for the sorites and the liar paradoxes.,23061
"Any a sorites, is as concept i.e. viewed","Any concept is viewed as a sorites, i.e.",23062
dynamics. interpretation of knowledge of Such an scientific provides its understanding an,Such an interpretation of scientific knowledge provides an understanding of its dynamics.,23063
combined rules-based heuristics with These simple then from feature biological maps are prior derived knowledge.,These feature maps are then combined with simple rules-based heuristics derived from prior biological knowledge.,23064
network paper chained (CC-Net). presents This cascade,This paper presents chained cascade network (CC-Net).,23065
"numerical rized continuous correlative, Broadly, catego- these be discrete. and can modeling in approaches","Broadly, these numerical modeling approaches can be catego- rized in correlative, continuous and discrete.",23066
"differ in these entry data and However, data density quality. sets","However, these data sets differ in entry density and data quality.",23067
"dating research local and regional specifics uncertainties. and or they Often gaps, disregard","Often they disregard local and regional specifics and research gaps, or dating uncertainties.",23068
The in modulation method high-pass successful image method a fusion. is,The high-pass modulation method is a successful method in image fusion.,23069
image. the Using the boxcar ringing introduces filter the effect fused into,Using the boxcar filter introduces the ringing effect into the fused image.,23070
"problem, wavelet of To the boxcar with this cope transform we filters. instead use","To cope with this problem, we use the wavelet transform instead of boxcar filters.",23071
"moment, typical moment, moment, geometrical moments Tchebichef (including Zernike Twelve etc.)","Twelve typical moments (including geometrical moment, Zernike moment, Tchebichef moment, etc.)",23072
binary studied. are used performance MomentsNet whose is for construct the image to recognition,are used to construct the MomentsNet whose recognition performance for binary image is studied.,23073
"subject-predicate-object objects the detects It interactions with e.g. and a captures pair-wise triplet, their","It detects the objects and captures their pair-wise interactions with a subject-predicate-object triplet, e.g.",23074
"phrase considered visual paper, a with each this is relationship as components. three In","In this paper, each visual relationship is considered as a phrase with three components.",23075
non-maximum training and suppression strategy method also are model Corresponding proposed.,Corresponding non-maximum suppression method and model training strategy are also proposed.,23076
speed outperforms accuracy. and both in results state-of-art that show ViP-CNN the Experimental our method,Experimental results show that our ViP-CNN outperforms the state-of-art method both in speed and accuracy.,23077
"the an been to difficult towards giving However, achievable task. process has framework a this","However, giving an achievable framework to the process towards this has been a difficult task.",23078
A layered layered like DIKW nicely structure framework of the with cortex. pre-frontal aligns,A layered framework like DIKW aligns nicely with the layered structure of pre-frontal cortex.,23079
with combination norms. new finite This in to sample respect bounds stronger the results,This combination results in new finite sample bounds with respect to the stronger norms.,23080
rates our From sample bounds easily these follow. finite,From these finite sample bounds our rates easily follow.,23081
"asymptotic the in Finally, we optimality results our prove many of cases.","Finally, we prove the asymptotic optimality of our results in many cases.",23082
takes challenge. This in phase the project the segmentation part of only,This project only takes part in the segmentation phase of the challenge.,23083
and (feature of are other phases The not identification) lesion the considered. competition extraction,The other phases of the competition (feature extraction and lesion identification) are not considered.,23084
starting prediction point the for Gene analysis. genome is customarily,Gene prediction is customarily the starting point for genome analysis.,23085
both our real corroborate on datasets synthetic and theory. Experiments,Experiments on both synthetic and real datasets corroborate our theory.,23086
main Object heritage is reconstruction cultural the one of in preservation. problems,Object reconstruction is one of the main problems in cultural heritage preservation.,23087
is This in of documentation. due data to problem lack,This problem is due to lack of data in documentation.,23088
"image of algorithm, A-KAZE Using keypoints obtained. were each","Using A-KAZE algorithm, keypoints of each image were obtained.",23089
we employed descriptor to from create feature LIOP it. Then,Then we employed LIOP to create feature descriptor from it.,23090
"the was generate used. In to geometry MVS order shape,","In order to generate the geometry shape, MVS was used.",23091
"be motor after BCIs to stroke. example, function regain used may For lost","For example, BCIs may be used to regain lost motor function after stroke.",23092
controlling from that requires controlling neural activity BCI. activity dissociated limbs is the unaffected This,This requires that neural activity controlling unaffected limbs is dissociated from activity controlling the BCI.,23093
"measure effective their connectivity, of control units connectivity. Using that we a decrease observed","Using a measure of effective connectivity, we observed that control units decrease their connectivity.",23094
a activity into neural provides brain-computer enables insight the dual-control study interface. This that,This study provides insight into the neural activity that enables a dual-control brain-computer interface.,23095
"inference scale automatic scene novel a More propose we method. importantly,","More importantly, we propose a novel automatic scene scale inference method.",23096
the comprehensive We BrnoCompSpeed. speed recent on dataset evaluate for our method measurement,We evaluate our method on the recent comprehensive dataset for speed measurement BrnoCompSpeed.,23097
of discrete an deep We present distributions. estimation to probability conditional approach,We present an approach to deep estimation of discrete conditional probability distributions.,23098
"have including Such applications, several of video image, models generative data. modeling and audio,","Such models have several applications, including generative modeling of audio, image, and video data.",23099
discrete dyadic approach the Our smoothing techniques: space. combines two graph-based of partitioning and main,Our approach combines two main techniques: dyadic partitioning and graph-based smoothing of the discrete space.,23100
powerful. The and simple associations surprisingly are resulting,The resulting associations are surprisingly simple and powerful.,23101
"Distributed is optimally NP-hard. Problems (DCOPs) Constraint Nevertheless, Optimization solving","Nevertheless, solving Distributed Constraint Optimization Problems (DCOPs) optimally is NP-hard.",23102
"Therefore, incomplete algorithms are applications, complex necessary. in large-scale, DCOP","Therefore, in large-scale, complex applications, incomplete DCOP algorithms are necessary.",23103
"available. readily image-sequences, only single images, cases are most rather in However,","However, in most cases image-sequences, rather only single images, are readily available.",23104
"means probabilistic obtaining estimates of pose we our consider Finally, model. from","Finally, we consider means of obtaining probabilistic pose estimates from our model.",23105
datasets. our driving openly-available and evaluate method localization autonomous on We indoor real-world,We evaluate our method on openly-available real-world autonomous driving and indoor localization datasets.,23106
"in Nowadays instead increasingly ways, physical are of interaction. being electronic handled actions","Nowadays actions are increasingly being handled in electronic ways, instead of physical interaction.",23107
earlier times in person. used authentication is of biometrics a the From,From earlier times biometrics is used in the authentication of a person.,23108
for tool data. is PCA analyzing powerful a,PCA is a powerful tool for analyzing data.,23109
goal dimensionality of PCA The is reduction. primary,The primary goal of PCA is dimensionality reduction.,23110
(SKIG) from Gesture KInect are Frames extracted the dataset. Sheffield,Frames are extracted from the Sheffield KInect Gesture (SKIG) dataset.,23111
by then done and is set The the training implementation recognizer. a creating training,The implementation is done by creating a training set and then training the recognizer.,23112
to metric with value the used similarity recognize Euclidean the is gestures. distance threshold as,Euclidean distance with the threshold value is used as similarity metric to recognize the gestures.,23113
the problem of One is Canadian context the in (CTP). Traveler Problem this,One of the problem in this context is the Canadian Traveler Problem (CTP).,23114
that expected find the a The goal minimizes strategy is travel cost. to,The goal is to find a strategy that minimizes the expected travel cost.,23115
known P$\#$ be hard. problem The is to,The problem is known to be P$\#$ hard.,23116
the theoretically study we and CTP work In empirically. this,In this work we study the CTP theoretically and empirically.,23117
"that the optimally Gen-PAO general we Second, develop algorithm a CTP. solve","Second, we develop a general algorithm Gen-PAO that optimally solve the CTP.",23118
solving other types of CTP. of CTP Sensing-CTP two capable is and Expensive-Edges Gen-PAO called,Gen-PAO is capable of solving two other types of CTP called Sensing-CTP and Expensive-Edges CTP.,23119
"by to Along fixed rise dynamics points, defined the a the manifold ghost. slow give","Along the manifold defined by the fixed points, slow dynamics give rise to a ghost.",23120
and performance influencing of team competences factors members. Key team personality are,Key factors influencing team performance are competences and personality of team members.,23121
"analysing results empirical we when obtained that Finally, student some present we performance. preliminary","Finally, we present some preliminary empirical results that we obtained when analysing student performance.",23122
learning the feature step A extraction. approaches key of machine is,A key step of the machine learning approaches is feature extraction.,23123
is widely used method feature One (PCA). extraction analysis component principle of most the,One of the most widely used feature extraction method is principle component analysis (PCA).,23124
"signal be to However, sensitive it misalignment. can","However, it can be sensitive to signal misalignment.",23125
to entries embeddings types the constrained All of non-negative. be of relation are,All entries of the embeddings of relation types are constrained to be non-negative.,23126
model that our show significantly Experimental results completion outperforms on graph task. other knowledge baselines,Experimental results show that our model significantly outperforms other baselines on knowledge graph completion task.,23127
"more require extension spaces. an general applications some to However,","However, some applications require an extension to more general spaces.",23128
"HSU based on proposed, Today, have or statistical geometrical been model. many either a algorithms","Today, many HSU algorithms have been proposed, based either on a geometrical or statistical model.",23129
cloud to objects. their find can upload the photos server to Clients,Clients can upload their photos to the cloud server to find objects.,23130
"does content of not client photos. want the the leakage the of his/her However,","However, the client does not want the leakage of the content of his/her photos.",23131
the (RBI) We Image Base Random Representation. propose,We propose the Random Base Image (RBI) Representation.,23132
base images. is image random The divided original into,The original image is divided into random base images.,23133
base to the cloud Only images server. submitted the are randomly,Only the base images are submitted randomly to the cloud server.,23134
"image Thus, the leaked. be the not can content of","Thus, the content of the image can not be leaked.",23135
integral-image makes for acceleration. the again RBI The enable great the,The RBI makes the integral-image enable again for the great acceleration.,23136
study probabilistic problem We base reasoning. for rules of learning logical knowledge first-order the,We study the problem of learning probabilistic first-order logical rules for knowledge base reasoning.,23137
that design compose controller We system these neural to operations. learns a,We design a neural controller system that learns to compose these operations.,23138
methodology of standard the compare methods. Monte PDMC and new efficiency this to Carlo We,We compare the efficiency of this new methodology to standard PDMC and Monte Carlo methods.,23139
reduced up to scalings several dimensional exhibited. are and Accelerations magnitudes,Accelerations up to several magnitudes and reduced dimensional scalings are exhibited.,23140
This the at by regularizer convergence. induced the minimizes bias,This minimizes the bias induced by the regularizer at convergence.,23141
"state-of-the-art However, deeper larger domain. in deliver needed results to are this and networks","However, larger and deeper networks are needed to deliver state-of-the-art results in this domain.",23142
are achieved limit. relative be the to dataset pronounced particularly gains in small The shown,The achieved relative gains are shown to be particularly pronounced in the small dataset limit.,23143
"have been such no discovered rules Unfortunately, proteins. for","Unfortunately, no such rules have been discovered for proteins.",23144
vision focuses in of image. what on computer is an Most,Most of computer vision focuses on what is in an image.,23145
a convolutional network neural based on Our is structure. model,Our model is based on a convolutional neural network structure.,23146
fully It thus convolutional efficient. is highly,It is fully convolutional thus highly efficient.,23147
train our fast category for We object on detector detection data. a instance,We train a fast object category detector for instance detection on our data.,23148
Our download for available cs.unc.edu/~ammirato/active_vision_dataset_website/. at is dataset,Our dataset is available for download at cs.unc.edu/~ammirato/active_vision_dataset_website/.,23149
the model in the noise. speckle parameters all retrain then presence We of,We then retrain all the model parameters in the presence of speckle noise.,23150
"reconstruct visual we fail to look certain does Why at our patterns? system reality, when","Why does our visual system fail to reconstruct reality, when we look at certain patterns?",23151
illusions in Geometrical pathway? Where the emerge start to do visual,Where do Geometrical illusions start to emerge in the visual pathway?,23152
"our detection network. relation the end-to-end To first knowledge, is best the of VTransE","To the best of our knowledge, VTransE is the first end-to-end relation detection network.",23153
analysis quantitative allows of organs the in Computational anatomy medical images.,Computational anatomy allows the quantitative analysis of organs in medical images.,23154
"Therefore, estimation hampered instability. is their by usefulness","Therefore, their usefulness is hampered by estimation instability.",23155
statistical (ROPE): for introduce We modeling. method Penalized novel robust Estimates a network of Resampling,We introduce Resampling of Penalized Estimates (ROPE): a novel statistical method for robust network modeling.,23156
is available R an CRAN. ROPE on as package,ROPE is available as an R package on CRAN.,23157
be accuracy can a Therefore achieved. any substantial without losing speedup,Therefore a substantial speedup without losing any accuracy can be achieved.,23158
"method better only. suturing data, for performs our accelerometer For","For accelerometer data, our method performs better for suturing only.",23159
"space. be locations these obtained Originally, in at parameter expensive the simulation random would samples","Originally, these expensive simulation samples would be obtained at random locations in the parameter space.",23160
problem. challenging a remains image stitching Low-textured,Low-textured image stitching remains a challenging problem.,23161
instability and due simplicity. information to fitting process This partial model is loss prone to,This process is prone to fitting instability and partial information loss due to model simplicity.,23162
review contribution on of In J.B.S. a the,In a review on the contribution of J.B.S.,23163
features effective extract signature to used for is Path online characters.,Path signature is used to extract effective features for online characters.,23164
diagnosis and Resonance used routine Magnetic Imaging widely treatment. clinical is (MRI) in,Magnetic Resonance Imaging (MRI) is widely used in routine clinical diagnosis and treatment.,23165
"in experiments extensive hyperintensity segmentation matter questions, white To conducted these we address task.","To address these questions, we conducted extensive experiments in white matter hyperintensity segmentation task.",23166
segmentation hierarchical results final then We by automatic our refine pruning.,We then refine our final segmentation results by automatic hierarchical pruning.,23167
processes. kernel latent a that models couplings multiple We novel introduce across input-dependent,We introduce a novel kernel that models input-dependent couplings across multiple latent processes.,23168
Variational Bayesian by inference. observation classification for multiple sets Probit derived are and support,Probit classification and support for multiple observation sets are derived by Variational Bayesian inference.,23169
be to both applied vehicles. It and land can aerial,It can be applied to both aerial and land vehicles.,23170
to (FNNs) far So been used have neural feedforward only networks control. train UAV,So far only feedforward neural networks (FNNs) have been used to train UAV control.,23171
"highly is data. for control input Vision correlated prediction a its known sequential problem, based","Vision based control is a sequential prediction problem, known for its highly correlated input data.",23172
"network an makes a especially RNN. hard, The training correlation","The correlation makes training a network hard, especially an RNN.",23173
"which a not available. is requires of training end-to-end Further, often lot data","Further, end-to-end training requires a lot of data which often is not available.",23174
to explain behavior visualizations learned. the help Different,Different visualizations help to explain the behavior learned.,23175
of diffusion Image image a the for task denoising. fundamental role plays,Image diffusion plays a fundamental role for the task of image denoising.,23176
outperforms both existing over body the Our approach methods benchmarks on all all parts.,Our approach outperforms all existing methods on both benchmarks over all the body parts.,23177
speech. prediction work learning on sequence Most on for deep focuses and video,Most work on deep learning for sequence prediction focuses on video and speech.,23178
The features to are classifiers. vector then train used machine support extracted,The extracted features then are used to train support vector machine classifiers.,23179
"a final classification. classifier fused In the are obtain outputs the to stage,","In the final stage, the classifier outputs are fused to obtain a classification.",23180
method Our applicable settings. to the both semi-supervised and is weakly-supervised,Our method is applicable to both the weakly-supervised and semi-supervised settings.,23181
"the in we phase. convolution dilated framework hybrid Second, encoding a (HDC) propose","Second, we propose a hybrid dilated convolution (HDC) framework in the encoding phase.",23182
the models ConvNets This transformation of trained advantage better takes over ImageNet.,This transformation takes better advantage of the trained ConvNets models over ImageNet.,23183
X-ray We projections. to Computed discuss applications further Tomography and,We further discuss applications to Computed Tomography and X-ray projections.,23184
the is latter. example This of an thesis,This thesis is an example of the latter.,23185
the first automatic is in recognition systems. step most melanoma segmentation Lesion,Lesion segmentation is the first step in most automatic melanoma recognition systems.,23186
Large-scale been using networks magnetic have studied extensively resonance resting imaging. functional functional state,Large-scale functional networks have been extensively studied using resting state functional magnetic resonance imaging.,23187
"However, unknown. organization, and fine-scale network of the remain pattern, activity function largely","However, the pattern, organization, and function of fine-scale network activity remain largely unknown.",23188
the laws. problem inherent we For this reveal,For this problem we reveal the inherent laws.,23189
proposes Conditional This (CRFs). Fields in for novel Random quantification approach paper uncertainty a dense,This paper proposes a novel approach for uncertainty quantification in dense Conditional Random Fields (CRFs).,23190
clinical on Resonance Magnetic and synthetic method data. validated is The Imaging,The method is validated on synthetic and clinical Magnetic Resonance Imaging data.,23191
image segmented that wrongly high regions. well to of found We uncertainty areas correspond,We found that areas of high uncertainty correspond well to wrongly segmented image regions.,23192
using capturing. image eye a appears human of to The low be for sensors number,The human eye appears to be using a low number of sensors for image capturing.,23193
done did the authors. We that by more actual rather statistical analysis values previous putting,We did more statistical analysis rather that putting the actual values done by previous authors.,23194
We kind different of effect i.e. propose disjunction,We propose different kind of disjunction effect i.e.,23195
based on statistical strong our and weak analysis.,strong and weak based on our statistical analysis.,23196
amongst aims characteristics Fine-grained similar categories. visual to discriminative recognition capture visually,Fine-grained visual recognition aims to capture discriminative characteristics amongst visually similar categories.,23197
move to that Fast artefacts. imaging is avoid motion required for targets,Fast imaging is required for targets that move to avoid motion artefacts.,23198
fetal in MRI. particular is This difficult for,This is in particular difficult for fetal MRI.,23199
is small Fine-grained a recognition task variances. due challenging to intra-category the,Fine-grained recognition is a challenging task due to the small intra-category variances.,23200
of of leverage Most methods for parts fine-grained objects recognition top-performing performance. better,Most of top-performing fine-grained recognition methods leverage parts of objects for better performance.,23201
"computationally are extremely annotations required. expensive Therefore, part which are","Therefore, part annotations which are extremely computationally expensive are required.",23202
"approach we paper, In SFCN. shape also segmentation present a new based this on","In this paper, we also present a new shape segmentation approach based on SFCN.",23203
"as approach, input. are low-level our geometric features by In three triangles-to-triangles trained using SFCNs","In our approach, SFCNs are trained triangles-to-triangles by using three low-level geometric features as input.",23204
is frames. A initialization estimate step to multiple motion over required single,A single initialization step is required to estimate motion over multiple frames.,23205
robustness functions regularization enhance time. Several over,Several regularization functions enhance robustness over time.,23206
multi-sensor of This paper efficient implementation multi-Bernoulli the generalized an filter. proposes (GLMB) labeled,This paper proposes an efficient implementation of the multi-sensor generalized labeled multi-Bernoulli (GLMB) filter.,23207
of Large computer-understandable of intermediate millions proofs consist steps. logical,Large computer-understandable proofs consist of millions of intermediate logical steps.,23208
"or So been generate has generally to steps. used learning machine not filter far, these","So far, machine learning has generally not been used to filter or generate these steps.",23209
the this license. dataset We available make under publicly BSD,We make this dataset publicly available under the BSD license.,23210
played role Hashing large-scale a retrieval. pivotal image has in,Hashing has played a pivotal role in large-scale image retrieval.,23211
"great promise. (CNN), of Convolutional shown learning With Neural hashing Network the has development","With the development of Convolutional Neural Network (CNN), hashing learning has shown great promise.",23212
"propose large-scale for method novel study, In hashing retrieval. a we image this","In this study, we propose a novel hashing method for large-scale image retrieval.",23213
in of operations the photo Compositing one most editing. common is,Compositing is one of the most common operations in photo editing.,23214
"images performance. noticeably depth For real still affect however, the end discrepancies with scans","For depth images however, discrepancies with real scans still noticeably affect the end performance.",23215
"the the is this and simultaneously subject agent object of Notably, observation.","Notably, this agent is simultaneously the subject and the object of observation.",23216
adaptive often a increase in This results in approach massive rendering performance.,This adaptive approach often results in a massive increase in rendering performance.,23217
subsequent suggestions with This work. note some ends for,This note ends with some suggestions for subsequent work.,23218
"the an flow boundary filtering. improved edgelet-based Finally, with is estimate","Finally, the boundary flow estimate is improved with an edgelet-based filtering.",23219
"dataset. On boundary on estimation, the Sintel present we training the first flow results","On boundary flow estimation, we present the first results on the Sintel training dataset.",23220
of in as such reduction aspects Our SSL. discussion exposes role important the dimension,Our discussion exposes important aspects such as the role of dimension reduction in SSL.,23221
deep the limitations limit learning-based of development methods. approaches new severely These especially,These limitations severely limit the development of new approaches especially deep learning-based methods.,23222
first This provides progress. review paper of comprehensive recent a the,This paper first provides a comprehensive review of the recent progress.,23223
"ameliorate problem. judicious However, can this of model structure choice","However, judicious choice of model structure can ameliorate this problem.",23224
a in environment fluctuating gain information fueling fitness. is to process environmental a of Adaptation,Adaptation in a fluctuating environment is a process of fueling environmental information to gain fitness.,23225
structures results thermodynamic These the adaptation in clarify evolution. and information,These results clarify the information thermodynamic structures in adaptation and evolution.,23226
CIELAB is region using First the Based the phase Segmentation. segmented In Color lesion space,In the First phase the lesion region is segmented using CIELAB Color space Based Segmentation.,23227
"descriptors are as feature Then extracted. such shape, color and texture","Then feature descriptors such as shape, color and texture are extracted.",23228
"for localization categorization, The latter the for accuracy. is first multi-class improving while is","The first is for multi-class categorization, while the latter is for improving localization accuracy.",23229
class We convolutional cascaded imbalanced propose neural to networks cope with therefore the problem.,We therefore propose cascaded convolutional neural networks to cope with the class imbalanced problem.,23230
"probabilities. trained set network neural with data a Successively, calculates nodule convolutional balanced a","Successively, a convolutional neural network trained with a balanced data set calculates nodule probabilities.",23231
both provides time Transform frequency. in localization Wavelet and,Wavelet Transform provides localization in both time and frequency.,23232
"preprocessing is the the in DWT In signal. used ECG wander baseline to stage, remove","In preprocessing stage, DWT is used to remove the baseline wander in the ECG signal.",23233
"investigate of clonal repertoire aging. long-term selection on we the Lastly, effect","Lastly, we investigate the effect of long-term clonal selection on repertoire aging.",23234
"as And, regions much to as avoided be have salient irrelevant possible.","And, irrelevant salient regions have to be avoided as much as possible.",23235
what remains: outstanding question us which But principle tells state to constrain? variables an,But an outstanding question remains: what principle tells us which state variables to constrain?,23236
in Neutral distinguishing forces population. no assumes selective evolution a there variants are that different,Neutral evolution assumes that there are no selective forces distinguishing different variants in a population.,23237
from baby a apply names Australia. data approach complete our of We to set,We apply our approach to a complete data set of baby names from Australia.,23238
sparse great representation image Group-based potential denoising. shown in has,Group-based sparse representation has shown great potential in image denoising.,23239
"of prior existing However, self-similarity the (NSS) input methods image. noisy nonlocal consider most only","However, most existing methods only consider the nonlocal self-similarity (NSS) prior of noisy input image.",23240
"called this group model, prior In (GSRC). sparsity constraint paper a new residual we propose","In this paper we propose a new prior model, called group sparsity residual constraint (GSRC).",23241
recent concerns ethical intellectuals years artificial In raised of prominent have about the consequences intelligence.,In recent years prominent intellectuals have raised ethical concerns about the consequences of artificial intelligence.,23242
of the this rejects considers and This paper possibility outcome.,This paper considers and rejects the possibility of this outcome.,23243
video heavily segmentation Conventional models. approaches appearance on rely,Conventional video segmentation approaches rely heavily on appearance models.,23244
appearance complex use methods under discriminative limited that descriptors power have often Such scenarios.,Such methods often use appearance descriptors that have limited discriminative power under complex scenarios.,23245
combines geodesic It dynamic models. based also distance,It also combines geodesic distance based dynamic models.,23246
regions. refining by uncertain is The the final on obtained cutout foreground,The final foreground cutout is obtained by refining on the uncertain regions.,23247
"images to heavy CT angles data. from limited However, artifacts from projection due suffers incomplete","However, CT images from limited angles suffers from heavy artifacts due to incomplete projection data.",23248
results. deliver satisfactory but iterative Existing calculations can not require extensive methods,Existing iterative methods require extensive calculations but can not deliver satisfactory results.,23249
networks extends (FCN) fully-convolutional neural clothing This paper parsing the for problem.,This paper extends fully-convolutional neural networks (FCN) for the clothing parsing problem.,23250
disambiguate fine-grained requires and categories. parsing Clothing knowledge on higher-level to cues clothing contextual semantics,Clothing parsing requires higher-level knowledge on clothing semantics and contextual cues to disambiguate fine-grained categories.,23251
acquisition inherently The slow. of is Resonance (MRI) Imaging Magnetic,The acquisition of Magnetic Resonance Imaging (MRI) is inherently slow.,23252
this responsible The for is transformation consolidation. memory systems process called,The process responsible for this transformation is called systems memory consolidation.,23253
well memory The underlying reconsolidation neural are not systems understood. and mechanisms consolidation,The neural mechanisms underlying systems memory consolidation and reconsolidation are not well understood.,23254
detection Outlier important issue real number clustering cluster an for is data. and estimation,Outlier detection and cluster number estimation is an important issue for clustering real data.,23255
"on designed Based that outlier results. a clustering these evaluates and detection we observations, function","Based on these observations, we designed a function that evaluates clustering and outlier detection results.",23256
order We (SLM). linear model regression class a called study the of fundamental second models,We study a fundamental class of regression models called the second order linear model (SLM).,23257
"techniques capture to unable still are are and convincingly structured. these that However, generate artefacts","However, these techniques are still unable to capture and generate artefacts that are convincingly structured.",23258
present approach musical to sequences. generate we structured In this an paper,In this paper we present an approach to generate structured musical sequences.,23259
a musical sequences. sampling mechanism introduce efficiently for of variations We,We introduce a mechanism for sampling efficiently variations of musical sequences.,23260
"to some some unfamiliar However, texture radiologists. are not was fully recovered, which the of","However, some of the texture are not fully recovered, which was unfamiliar to some radiologists.",23261
This details results and paper methodology the challenge. EmotioNet of the,This paper details the methodology and results of the EmotioNet challenge.,23262
divided tracks. two challenge The was into,The challenge was divided into two tracks.,23263
limitations when algorithms The of to emotion. current apparent more trying recognize are,The limitations of current algorithms are more apparent when trying to recognize emotion.,23264
that special an forward. need in-depth provide the We of discussion points moving attention,We provide an in-depth discussion of the points that need special attention moving forward.,23265
"one melanoma cancer, of skin kind dangerous. is very As","As one kind of skin cancer, melanoma is very dangerous.",23266
based strategy detection recarbonization critical melanoma for therapy. Dermoscopy and early is,Dermoscopy based early detection and recarbonization strategy is critical for melanoma therapy.,23267
"dominant accuracy. However, diagnostic the dermatologists well-trained","However, well-trained dermatologists dominant the diagnostic accuracy.",23268
"order developing many focus solve image problem, systems. to analysis this on automatic effort In","In order to solve this problem, many effort focus on developing automatic image analysis systems.",23269
The aggregation. representations to novel extension measure saliency prior using a the proposed weights regional,The proposed extension weights the regional representations using a novel saliency measure prior to aggregation.,23270
This further retrieval to leads accuracy. improvement in,This leads to further improvement in retrieval accuracy.,23271
such study arduous. protein and interactions been The has experimental complexes of,The experimental study of such protein complexes and interactions has been arduous.,23272
methods attractive an alternative. Theoretical prediction are structure,Theoretical structure prediction methods are an attractive alternative.,23273
adjacencies accurately between spatial DCA can infer residues.,DCA can accurately infer spatial adjacencies between residues.,23274
These techniques models. high-resolution adjacencies can structure-prediction as included be to constraints in predict,These adjacencies can be included as constraints in structure-prediction techniques to predict high-resolution models.,23275
that DCA. commonly large are find We identified by interfaces,We find that large interfaces are commonly identified by DCA.,23276
agent's Recognition and to hierarchy recognize goals. the a Plan complete require explaining algorithms actions,Plan Recognition algorithms require to recognize a complete hierarchy explaining the agent's actions and goals.,23277
dynamics (SU) and is a its system uninvadable stable states. such by The of dominated,The dynamics of such a system is dominated by its stable and uninvadable (SU) states.,23278
variance. and of weak Here the the competition we problem large in consider limit,Here we consider the problem in the limit of weak competition and large variance.,23279
that much with methods in inference matches We traditional AVITM accuracy find better time.,We find that AVITM matches traditional methods in accuracy with much better inference time.,23280
"Because black new AVITM to models. readily is it topic is applied box,","Because AVITM is black box, it is readily applied to new topic models.",23281
"surveillance applications. on recognition In mainly in vehicles we this fine-grained of focus paper, traffic","In this paper, we focus on fine-grained recognition of vehicles mainly in traffic surveillance applications.",23282
be traffic from surveillance automatically bounding constructed box The data. can,The bounding box can be automatically constructed from traffic surveillance data.,23283
method that state-of-the-art show recognition. for outperforms fine-grained methods also our the We,We also show that our method outperforms the state-of-the-art methods for fine-grained recognition.,23284
in an are state-of-the-art important playing visual recognition. Deep networks role neural,Deep neural networks are playing an important role in state-of-the-art visual recognition.,23285
prior stationary also for from proposed is distinguishing objects. non-stationary The useful,The proposed prior is also useful for distinguishing stationary from non-stationary objects.,23286
work presents our paper this area. in This,This paper presents our work in this area.,23287
objective. networks achieve Both the to challenge of make algorithms use deep convolutional,Both algorithms make use of deep convolutional networks to achieve the challenge objective.,23288
"and complete CNN are our the method. The linear predictor, learned tags jointly parameters, by","The CNN parameters, linear predictor, and the complete tags are learned jointly by our method.",23289
is the minimization An problem. iterative solve developed algorithm to,An iterative algorithm is developed to solve the minimization problem.,23290
data effectiveness. show over image benchmark sets its Experiments,Experiments over benchmark image data sets show its effectiveness.,23291
resilient rates We demonstrate bit is orders-of-magnitude also that (e.g. generative more error compression to,We also demonstrate that generative compression is orders-of-magnitude more resilient to bit error rates (e.g.,23292
coding channels) than variable-length wireless noisy traditional schemes. from,from noisy wireless channels) than traditional variable-length coding schemes.,23293
in extent. Skin a is severe world-wide lesion disease,Skin lesion is a severe disease in world-wide extent.,23294
significantly in dermoscopy melanoma detection the Early survival of increases images rate.,Early detection of melanoma in dermoscopy images significantly increases the survival rate.,23295
"reasons, following to e.g. the recognition due accurate the of However, melanoma is extremely challenging","However, the accurate recognition of melanoma is extremely challenging due to the following reasons, e.g.",23296
"and between visual etc. low non-melanoma similarity and between skin, lesions, melanoma lesions contrast","low contrast between lesions and skin, visual similarity between melanoma and non-melanoma lesions, etc.",23297
proposed feature straight-forward A for dermoscopic extraction is task. the CNN,A straight-forward CNN is proposed for the dermoscopic feature extraction task.,23298
"the Experimental accuracies show of frameworks, results i.e. promising our","Experimental results show the promising accuracies of our frameworks, i.e.",23299
series available evaluate of proposal publicly our datasets. in We a,We evaluate our proposal in a series of publicly available datasets.,23300
agents' their actions. from Plan infer plans observed recognition algorithms,Plan recognition algorithms infer agents' plans from their observed actions.,23301
images. natural text a rotation-based arbitrary-oriented framework novel for in This paper scene introduces detection,This paper introduces a novel rotation-based framework for arbitrary-oriented text detection in natural scene images.,23302
"analysis. lesion a skin network neural In for deep this multi-task study, proposed is","In this study, a multi-task deep neural network is proposed for skin lesion analysis.",23303
neocortical properties (VSDi) scales. revealed processing of at Voltage-sensitive fundamental imaging has dye mesoscopic,Voltage-sensitive dye imaging (VSDi) has revealed fundamental properties of neocortical processing at mesoscopic scales.,23304
"variations graph data. have that proposed researchers Recently, CNNs can handle of","Recently, researchers have proposed variations of CNNs that can handle graph data.",23305
vision. (WOD) in detection problems computer a Weakly-supervised object is challenging,Weakly-supervised object detection (WOD) is a challenging problems in computer vision.,23306
We of using segmentation neural networks. consider the image problem convolutional semantic deep,We consider the problem of semantic image segmentation using deep convolutional neural networks.,23307
"k-space radial lines number of requires high-resolution large reconstruction. a trajectory However, radial for the","However, the radial k-space trajectory requires a large number of radial lines for high-resolution reconstruction.",23308
removes the network corrupted deep Methods: artifact artifacts images. The streaking the proposed from,Methods: The proposed deep network removes the streaking artifacts from the artifact corrupted images.,23309
"usually expensive. costs computational are However, the","However, the computational costs are usually expensive.",23310
"order by the time Furthermore, is computational of magnitude faster.","Furthermore, the computational time is by order of magnitude faster.",23311
Hungarian The solved association is method. by applying instance the,The instance association is solved by applying the Hungarian method.,23312
recognition State-of-the-art systems networks. are face based neural on deep (convolutional),State-of-the-art face recognition systems are based on deep (convolutional) neural networks.,23313
Our study deep secure recognition in the templates need face demonstrates systems. to,Our study demonstrates the need to secure deep templates in face recognition systems.,23314
hierarchical Networks a reinforcement (FuNs): We learning. for FeUdal architecture introduce novel,We introduce FeUdal Networks (FuNs): a novel architecture for hierarchical reinforcement learning.,23315
framework module. module Our Manager a Worker and employs a,Our framework employs a Manager module and a Worker module.,23316
primitive tick Worker of at the The generates every environment. actions,The Worker generates primitive actions at every tick of the environment.,23317
"a few there way. in examples However, demonstrating concrete are results these","However, there are few examples demonstrating these results in a concrete way.",23318
"particular, known the there discounting. results regarding gener- examples alised are demonstrating In no","In particular, there are no examples demonstrating the known results regarding gener- alised discounting.",23319
"a discuss subtleties. some results, number experimentally theoretical of reproduce and related We","We experimentally reproduce a number of theoretical results, and discuss some related subtleties.",23320
"update what exploration? the play in Carlo does Monte role Second,","Second, what role does the Monte Carlo update play in exploration?",23321
"in adapting Bellemare the examine al. et intrinsic particular, we In difficulties","In particular, we examine the intrinsic difficulties in adapting Bellemare et al.",23322
's the about violated. are when model approach assumptions,'s approach when assumptions about the model are violated.,23323
special apparatus. requiring and The is result practical more general no a algorithm,The result is a more practical and general algorithm requiring no special apparatus.,23324
"are produce valid. that outputs state-of-the-art methods Crucially, often not","Crucially, state-of-the-art methods often produce outputs that are not valid.",23325
large pathologists biological by expanses is currently reviewing detection tissues. performed of Metastasis,Metastasis detection is currently performed by pathologists reviewing large expanses of biological tissues.,23326
is error-prone. process intensive labor and This,This process is labor intensive and error-prone.,23327
reduce rates in false approach detection. Our metastasis negative could considerably,Our approach could considerably reduce false negative rates in metastasis detection.,23328
images short Neoclassical on furniture. outlines in results of classification paper object This research,This short paper outlines research results on object classification in images of Neoclassical furniture.,23329
A the use-case. research bottom-up Neoclassica main in data-driven framework the research routine is,A data-driven bottom-up research routine in the Neoclassica research framework is the main use-case.,23330
mode in finding algorithms popular are pattern methods shift for (MS) Mean analysis.,Mean shift (MS) algorithms are popular methods for mode finding in pattern analysis.,23331
selection approach parameter model both inference. simultaneously and performs This,This approach simultaneously performs both model selection and parameter inference.,23332
data is of The case with discussed. missing also values,The case of data with missing values is also discussed.,23333
data. interest both simulated shown and proposed criteria on real is of The,The interest of both proposed criteria is shown on simulated and real data.,23334
"accurate, Our applications. method CNN-based in challenging extraction provide can brain geometry-independent","Our CNN-based method can provide accurate, geometry-independent brain extraction in challenging applications.",23335
This segmentation for paper clustering a point clouds. hierarchical proposes mobile approach the LiDAR of,This paper proposes a hierarchical clustering approach for the segmentation of mobile LiDAR point clouds.,23336
a perform clouds We matrix. point clustering on based on unorganized proximity hierarchical the,We perform the hierarchical clustering on unorganized point clouds based on a proximity matrix.,23337
the a common of Oral (OSCC) Squamous is Carcinoma of Cell type oral cancer epithelium.,Oral Squamous Cell Carcinoma (OSCC) is a common type of cancer of the oral epithelium.,23338
Endomicroscopy structure in (CLE) analysis. vivo for Laser micro-anatomical Confocal images sub-surface records cell,Confocal Laser Endomicroscopy (CLE) records sub-surface micro-anatomical images for in vivo cell structure analysis.,23339
"solutions heuristics existing too However, either competitive consume less obtain much time. or","However, existing heuristics either obtain less competitive solutions or consume too much time.",23340
"new fusion. on To address interval was data method proposed this issue, based a","To address this issue, a new method was proposed based on interval data fusion.",23341
"sparsely in populated and all. populated, cases, may at very Some cells extreme be not","Some cells may be very sparsely populated, and in extreme cases, not populated at all.",23342
response an expected still might estimate We such like to in cells.,We might still like to estimate an expected response in such cells.,23343
(HANOVA) a such for propose data. hierarchical We novel ANOVA representation here,We propose here a novel hierarchical ANOVA (HANOVA) representation for such data.,23344
more the cells the for will interaction model lean data. Sparse lower-order on,Sparse cells will lean more on the lower-order interaction model for the data.,23345
"such involving is data, experiments. This case imaging particularly biological for as high-dimensional the problems","This is particularly the case for problems involving high-dimensional data, such as biological imaging experiments.",23346
"we use neighbour distributions. Computationally, between estimator nearest of distance a the","Computationally, we use a nearest neighbour estimator of the distance between distributions.",23347
nearest algorithm distance estimator. theoretically of properties the neighbour justify based We on the,We justify the algorithm theoretically based on properties of the nearest neighbour distance estimator.,23348
optimization networks. for units deep alleviating difficulties used building neural wildly are when Residual,Residual units are wildly used for alleviating optimization difficulties when building deep neural networks.,23349
using CRU across sharing enables factors. shared units residual different knowledge,CRU enables knowledge sharing across different residual units using shared factors.,23350
models GitHub) and trained on (Code . available are,(Code and trained models are available on GitHub) .,23351
on on generative discriminative results progresses and synthesis. promising texture shown deep Recent modeling have,Recent progresses on deep discriminative and generative modeling have shown promising results on texture synthesis.,23352
"we for synthesis. improved issues on focus In this work, solving these texture","In this work, we focus on solving these issues for improved texture synthesis.",23353
"a techniques are convergence of better important Meanwhile, diversity. suite to introduced achieve and","Meanwhile, a suite of important techniques are introduced to achieve better convergence and diversity.",23354
"it If left is pixel is processing the then uncorrupted, as unchanged. found","If the processing pixel is found as uncorrupted, then it is left unchanged.",23355
discuss the PDEs to where (e.g. a of Bayesian (CG) We coefficients formulation coarse-graining,We discuss a Bayesian formulation to coarse-graining (CG) of PDEs where the coefficients (e.g.,23356
"fine variability. random, exhibit material scale parameters)","material parameters) exhibit random, fine scale variability.",23357
in basic Our two be summarized architecture steps. could,Our architecture could be summarized in two basic steps.,23358
"to problems. networks successfully applied neural synthesis convolutional view Recently, have been (CNN)","Recently, convolutional neural networks (CNN) have been successfully applied to view synthesis problems.",23359
view the CNN-based proposed significantly We outperforms show method experimentally synthesis state-of-the-art method. the,We experimentally show the proposed method significantly outperforms the state-of-the-art CNN-based view synthesis method.,23360
updates the space be Those (\textit{i.e.} in applied feature point shape can either,Those shape updates can either be applied in the feature point space (\textit{i.e.},23361
"in or parametric explicit space. a low-dimensional, updates)","explicit updates) or in a low-dimensional, parametric space.",23362
"achieved Recently, promising fields deep learning vision. computer have approaches various in of results","Recently, deep learning approaches have achieved promising results in various fields of computer vision.",23363
predicted evolutionary evolution can Whether key biology. be is a in question,Whether evolution can be predicted is a key question in evolutionary biology.,23364
understand better to set repeatability Here evolution. of we out the,Here we set out to better understand the repeatability of evolution.,23365
"at one In most were containing alleles beneficial least large-effect mutation surviving present. populations, known","In most surviving populations, alleles containing at least one known large-effect beneficial mutation were present.",23366
Neural large-scale solution visual Convolutional is The Network deep the recognition. (CNN) state-of-the-art for,The deep Convolutional Neural Network (CNN) is the state-of-the-art solution for large-scale visual recognition.,23367
"In automatically. of structures possibility learning the paper, deep we network discuss this","In this paper, we discuss the possibility of learning deep network structures automatically.",23368
of the through extensive We proposed experiments. show effectiveness the approach,We show the effectiveness of the proposed approach through extensive experiments.,23369
characteristics and and determine efficiency Their structure effectiveness.,Their structure and characteristics determine efficiency and effectiveness.,23370
the themselves form are product Then moments. expressed invariants of of the in,Then the invariants themselves are expressed in the form of product of moments.,23371
just encoding proteins. procedure like This is DNA,This procedure is just like DNA encoding proteins.,23372
All the of can decomposed available PIs. GMIs be literature linear into combinations in,All GMIs available in the literature can be decomposed into linear combinations of PIs.,23373
"low In independent interest. uses, GMIs practical are particular of order","In practical uses, low order independent GMIs are of particular interest.",23374
This participates project in Classification the part. only Lesion,This project only participates in the Lesion Classification part.,23375
scores by step models The source images. two are prediction the different obtained from in,The scores in prediction step are obtained by two different models from the source images.,23376
face contour. along of a face alignment landmarks find set algorithms can Current robustly,Current face alignment algorithms can robustly find a set of landmarks along face contour.,23377
in of Our three initial works the curve. help method rough a with steps,Our method works in three steps with the help of a rough initial curve.,23378
"initial the along small and curve. overlapped sample squares we First,","First, we sample small and overlapped squares along the initial curve.",23379
exchangeability with functions. how We to choice investigate model,We investigate how to model exchangeability with choice functions.,23380
of variables. a on Exchangeability sequence uncertain structural a is assessment,Exchangeability is a structural assessment on a sequence of uncertain variables.,23381
problem an curve important is finance. Yield in forecasting,Yield curve forecasting is an important problem in finance.,23382
to have Processes applications. Gaussian variety functional successfully in applied of data a model been,Gaussian Processes have been successfully applied to model functional data in a variety of applications.,23383
model curve. the used Gaussian is yield A to Process,A Gaussian Process is used to model the yield curve.,23384
curve compare to forecast methods We with yield existing proposed the method. the,We compare existing methods to forecast the yield curve with the proposed method.,23385
term curve in Accuracy has practical long implications. structure yield of the important the region,Accuracy in the long term structure region of the yield curve has important practical implications.,23386
interested Analysts in are information. frequently this,Analysts are frequently interested in this information.,23387
et procedure is the by Maclaurin previous work formulation to reverse-mode linked al. Our of,Our formulation of the reverse-mode procedure is linked to previous work by Maclaurin et al.,23388
learning experiments cleaning on We present data interactions. on task and,We present experiments on data cleaning and on learning task interactions.,23389
have Previous in isolation. works problems targeted these,Previous works have targeted these problems in isolation.,23390
together. tackling them We by solution a believe effective be obtained elegant could more and,We believe a more effective and elegant solution could be obtained by tackling them together.,23391
"are sources First, on image. all the potential by located peak detection point","First, all potential point sources are located by peak detection on the image.",23392
of are image point spectral The extracted. and features potential sources then these,The image and spectral features of these potential point sources are then extracted.,23393
"classifier features. a sources point to extracted is true recognize the Finally, the build through","Finally, a classifier to recognize the true point sources is build through the extracted features.",23394
on are of approach our images real and applications demonstrated. Experiments astronomical X-ray,Experiments and applications of our approach on real X-ray astronomical images are demonstrated.,23395
"are they at introduced again In this this paper, timing.","In this paper, they are introduced again at this timing.",23396
"deeply is ""Function Modularization"" subconsciously. approach penetrated","""Function Modularization"" approach is deeply penetrated subconsciously.",23397
"""State modules. ""action in or space"" used of RL generally space"" the show functional existence","""State space"" or ""action space"" generally used in RL show the existence of functional modules.",23398
learning to module. only reinforcement That has learning limited for action-planning the,That has limited reinforcement learning to learning only for the action-planning module.,23399
Temporal action an yet challenging important problem. is localization,Temporal action localization is an important yet challenging problem.,23400
state-of-the-art select systems rank Many boundaries. classifiers and of segment-level proposal use to pre-determined segments,Many state-of-the-art systems use segment-level classifiers to select and rank proposal segments of pre-determined boundaries.,23401
in is unique and in temporal It dynamics. fine-grained jointly space-time modeling semantics action,It is unique in jointly modeling action semantics in space-time and fine-grained temporal dynamics.,23402
the in network train end-to-end manner CDC an We efficiently.,We train the CDC network in an end-to-end manner efficiently.,23403
publish will We soon. update version and source the online the camera-ready codes,We will update the camera-ready version and publish the source codes online soon.,23404
synthetic a range real our approach on We evaluate examples. of wide and,We evaluate our approach on a wide range of synthetic and real examples.,23405
"parallelization, Additional logging include proposal, batch features error-handling. and multi-point visualization,","Additional features include multi-point batch proposal, parallelization, visualization, logging and error-handling.",23406
matching we rely novel on and strategy. a matching multi-scale Instead techniques patch solely,Instead we solely rely on patch matching techniques and a novel multi-scale matching strategy.,23407
filtering. present outlier enhancements also for We,We also present enhancements for outlier filtering.,23408
their is dynamics. linking An their essential circuits understanding and toward structure neural step,An essential step toward understanding neural circuits is linking their structure and their dynamics.,23409
"relationship general, almost arbitrarily can be In this complex.","In general, this relationship can be almost arbitrarily complex.",23410
Evaluating a policy can world the deploying it costly. risky in by and real be,Evaluating a policy by deploying it in the real world can be risky and costly.,23411
caricature recognition important fundamentally of is perception. to face Studying understanding,Studying caricature recognition is fundamentally important to understanding of face perception.,23412
images and the All Web. were from face the caricatures collected,All the caricatures and face images were collected from the Web.,23413
stochastic networks systems interest. describe and Queue of complex both practical theoretical,Queue networks describe complex stochastic systems of both theoretical and practical interest.,23414
acid is sequence. sensitive amino to behavior IDP its an of phase The,The phase behavior of an IDP is sensitive to its amino acid sequence.,23415
for IDP are Ramifications of dependence separation sequence deciphering phase the of discussed. findings our,Ramifications of our findings for deciphering the sequence dependence of IDP phase separation are discussed.,23416
worth performance They training limited set. even a have with,They have worth performance even with a limited training set.,23417
find hard characteristics to together methods. more in are other even these accurate All,All these characteristics are hard to find together in other even more accurate methods.,23418
of all obtain We the them. on art state of performances,We obtain state of the art performances on all of them.,23419
"transferred of during cells from the daughter whole to differentiation. is chromatin modifications, mother cell","the whole of chromatin modifications, is transferred from mother to daughter cells during cell differentiation.",23420
"information positions the act. However, lack on genomic chromatin-modifying they which enzymes about","However, chromatin-modifying enzymes lack the information about genomic positions on which they act.",23421
channels to SRCNN statistical augments input predictability with downscaling. DeepSD in maximize multi-scale,DeepSD augments SRCNN with multi-scale input channels to maximize predictability in statistical downscaling.,23422
of one of cylindrical clip the is sequence. channel generated Each the coordinates skeleton from,Each clip is generated from one channel of the cylindrical coordinates of the skeleton sequence.,23423
presents approach a for This new parameters paper GARCH estimation. optimization of the,This paper presents a new approach for the optimization of GARCH parameters estimation.,23424
"for Firstly, we method propose the maximum. localization a the of","Firstly, we propose a method for the localization of the maximum.",23425
"In $m in sparse ignoring = runs $O(n)$. terms, time where fastRG graphs, log O(n)$,","In sparse graphs, where $m = O(n)$, ignoring log terms, fastRG runs in time $O(n)$.",23426
on in github. implementation is fastRG available An,An implementation in fastRG is available on github.,23427
contain history. a protein Natural record sequences their of,Natural protein sequences contain a record of their history.,23428
can structural force detail. the interpreted that with resulting be field We show,We show that the resulting force field can be interpreted with structural detail.,23429
The control. used in based commonly a optimal used device is on approach,The approach used is based on a device commonly used in optimal control.,23430
number. total types a and percentage to cell counted expressed cell Different the as were,Different cell types were counted and expressed as a percentage to the total cell number.,23431
the the density and cellular kinetics evaluated. Growth pellet and thickness including weight were,Growth kinetics including the pellet weight and thickness and the cellular density were evaluated.,23432
cells. immature of term to the dark instead It is use better paralysed chondrocytes,It is better to use the term immature dark chondrocytes instead of paralysed cells.,23433
"alternatives number require parameters. same the of these Moreover,","Moreover, these alternatives require the same number of parameters.",23434
we The the behaviors of presented present model previously. here paradoxical free is,The model we present here is free of the paradoxical behaviors presented previously.,23435
"of this these advances. tackle Here, the respective leveraging we problem strengths by","Here, we tackle this problem by leveraging the respective strengths of these advances.",23436
and Density vital tool in statistical both a community. ratio estimation machine learning is,Density ratio estimation is a vital tool in both machine learning and statistical community.,23437
"In present identifies trims robust which paper, a estimator and outliers. automatically we this","In this paper, we present a robust estimator which automatically identifies and trims outliers.",23438
parameter this the error settings. under high-dimensional We estimation analyze of estimator,We analyze the parameter estimation error of this estimator under high-dimensional settings.,23439
to of conducted Experiments the verify are the effectiveness estimator.,Experiments are conducted to verify the effectiveness of the estimator.,23440
distributing the at additional procedure. introduced within inference attempts approximation inference Existing have,Existing attempts at distributing inference have introduced additional approximation within the inference procedure.,23441
"partition, the given For features. finite parallel simple of the is instantiation inference","For the finite partition, parallel inference is simple given the instantiation of features.",23442
"The hybrid parallel, posterior. true sampler, being while produces from resulting asymptotically samples the","The resulting hybrid sampler, while being parallel, produces samples asymptotically from the true posterior.",23443
automatically. deep to framework Methods A was learning grade trained images the,Methods A deep learning framework was trained to grade the images automatically.,23444
the of an quality DR. Conclusion grading is Image for prerequisite essential,Conclusion Image quality is an essential prerequisite for the grading of DR.,23445
Detecting activities task. but videos is challenging important an in untrimmed,Detecting activities in untrimmed videos is an important but challenging task.,23446
"stabilizes in-phase Their superposition, nonlinear experiments. as observed however, synchronization in","Their nonlinear superposition, however, stabilizes in-phase synchronization as observed in experiments.",23447
heavily rely model recognition Current videos for training. action on methods trimmed,Current action recognition methods heavily rely on trimmed videos for model training.,23448
"expensive large-scale and trimmed acquire dataset. However, to is a video time-consuming it","However, it is expensive and time-consuming to acquire a large-scale trimmed video dataset.",23449
method and presented that Several the applications efficiency. are illustrate versatility,Several applications are presented that illustrate the method versatility and efficiency.,23450
classification problems. popular become in Gradient-based methods texture have,Gradient-based texture methods have become popular in classification problems.,23451
"use size. larger neighborhood we a Also,","Also, we use a larger neighborhood size.",23452
specificity also calculated. sensitivity are and Optimal,Optimal sensitivity and specificity are also calculated.,23453
are Three cohort used different the in datasets experiments.,Three different cohort datasets are used in the experiments.,23454
"its has hampered been of wider the by time However, algorithms. lack sampling polynomial adoption","However, its wider adoption has been hampered by the lack of polynomial time sampling algorithms.",23455
suggest boundaries. a the boundary refinement also refine We residual-based to further ob- ject,We also suggest a residual-based boundary refinement to further refine the ob- ject boundaries.,23456
process proposed The these used gradual GUN a two of commonly frameworks. utilizes instead,The proposed GUN utilizes a gradual process instead of these two commonly used frameworks.,23457
"gradual presented training the a Furthermore, is strategy GUN. for","Furthermore, a gradual training strategy is presented for the GUN.",23458
"vivid recover The and trained. easy is fine and GUN to can results, be","The GUN can recover fine and vivid results, and is easy to be trained.",23459
demonstrate effectiveness The network. proposed the results of on image several sets experimental the,The experimental results on several image sets demonstrate the effectiveness of the proposed network.,23460
a We first Linear training Embedding Quasi-curvature Locally (QLLE) proposed for set.,We first proposed a Quasi-curvature Locally Linear Embedding (QLLE) for training set.,23461
each sample. the criterion linear QLLE guarantees of neighborhood in,QLLE guarantees the linear criterion in neighborhood of each sample.,23462
"problem. is method (NM) for Then, proposed neural a out-of-sample","Then, a neural method (NM) is proposed for out-of-sample problem.",23463
in artificial considered intelligence applications. Quantum are decision being use increasingly systems for,Quantum decision systems are being increasingly considered for use in artificial intelligence applications.,23464
nodes certain be in can based distinguished correlations states. and Classical on their quantum,Classical and quantum nodes can be distinguished based on certain correlations in their states.,23465
decision This paper the properties some of tree states structure. a obtained in investigates,This paper investigates some properties of the states obtained in a decision tree structure.,23466
decision mapped these may How tree the is be to correlations considered.,How these correlations may be mapped to the decision tree is considered.,23467
are representations quantum Classical to states tree and approximations provided.,Classical tree representations and approximations to quantum states are provided.,23468
the execution oculomotor of visual optimal tasks. control We study during search the problem of,We study the problem of optimal oculomotor control during the execution of visual search tasks.,23469
. model experiments our We psychophysical with validated eye-tracking human simulations,We validated our model simulations with human psychophysical eye-tracking experiments .,23470
Gaussian be function multi-output for to model vector-valued useful process has for been shown prediction.,Gaussian process model for vector-valued function has been shown to be useful for multi-output prediction.,23471
methods proposed data real The the illustrated through is and examples. several simulated usefulness of,The usefulness of the proposed methods is illustrated through several simulated and real data examples.,23472
fast particularly it inference applications. suitable makes for real-time Its,Its fast inference makes it particularly suitable for real-time applications.,23473
to Poisson between identify were nonlinear models fitted associations regression dengue. and MEI,Poisson regression models were fitted to identify nonlinear associations between MEI and dengue.,23474
statistically but meteorological negative between We a weak GBS factors. and correlation found significant local,We found a weak but statistically significant negative correlation between GBS and local meteorological factors.,23475
paper this The three-fold. is purpose of,The purpose of this paper is three-fold.,23476
filter. a is (CoF) preserving Co-occurrence Filter boundary,Co-occurrence Filter (CoF) is a boundary preserving filter.,23477
"with extends therefore BF boundaries, the not to edges. deal The just CoF","The CoF therefore extends the BF to deal with boundaries, not just edges.",23478
the It directly image. from learns co-occurrences,It learns co-occurrences directly from the image.,23479
"the subgoal have issue, proposed. To framework option solve and been this","To solve this issue, the subgoal and option framework have been proposed.",23480
Micro-objective learning solve (MOL) We propose to problem. this,We propose Micro-objective learning (MOL) to solve this problem.,23481
games: and Montezuma's Revenge We Atari two our evaluated in algorithm Seaquest.,We evaluated our algorithm in two Atari games: Montezuma's Revenge and Seaquest.,23482
"significantly the improved With game, to scores. experiments three baseline MOL each","With three experiments to each game, MOL significantly improved the baseline scores.",23483
"in Revenge, the previous better model. state-of-the-art MOL than Montezuma's results Especially achieved times two","Especially in Montezuma's Revenge, MOL achieved two times better results than the previous state-of-the-art model.",23484
a efficiently. training algorithm to implement model train our We two-step,We implement a two-step training algorithm to train our model efficiently.,23485
"in enabling attention monitoring. it driver for instance, an For is technology automotive","For instance, it is an enabling technology in automotive for driver attention monitoring.",23486
"the is iteration, probability each estimated Monte posterior In Carlo simulation. using","In each iteration, the posterior probability is estimated using Monte Carlo simulation.",23487
"three and -- of offer measure, acceptance-rejection, methods Monte Carlo. variance reduction conditional We change","We offer three variance reduction methods -- change of measure, acceptance-rejection, and conditional Monte Carlo.",23488
suggest Monte Numerical Carlo be that method should the preferred. conditional experiments,Numerical experiments suggest that the conditional Monte Carlo method should be preferred.,23489
recognition This paper hybrid a face multi-pose using (HFFD). approach presents features face descriptors,This paper presents a multi-pose face recognition approach using hybrid face features descriptors (HFFD).,23490
each fitted city. are studies separate models where Our different approach to previous from is,Our approach is different from previous studies where separate models are fitted to each city.,23491
We between data for unit. correlation each the the simulated administrative observed computed and Pearson's,We computed the Pearson's correlation between the observed and simulated data for each administrative unit.,23492
Our both public works significance. health and have scientific,Our works have both scientific and public health significance.,23493
"Scanning in barrier these minimal applications. a bodies clothing, however, presents practical to","Scanning bodies in minimal clothing, however, presents a practical barrier to these applications.",23494
personalized exploited produce smooth methods models details. body have lacking Previous that shapes,Previous methods that have exploited body models produce smooth shapes lacking personalized details.,23495
of a personalized to approach the shape We a new person. recover contribute,We contribute a new approach to recover a personalized shape of the person.,23496
are of ones the transformed combination original The the images and analyzed.,The combination of the transformed images and the original ones are analyzed.,23497
"became a suggested introductory practice is research strategy books. Indeed, and it by","Indeed, it became research practice and is a suggested strategy by introductory books.",23498
"is not we always that prove this However, the case.","However, we prove that this is not always the case.",23499
"life. malignant of is skin cancer to threatening form Melanoma, very a","Melanoma, a malignant form of skin cancer is very threatening to life.",23500
can be derived to in used predicates model models. planning independent domain- Axioms,Axioms can be used to model derived predicates in domain- independent planning models.,23501
work planners solely state- Previous focused search on on space planners. axiom-aware,Previous work on axiom-aware planners focused solely on state- space search planners.,23502
propose and based programming planners programming. answer set We on integer axiom-aware,We propose axiom-aware planners based on answer set programming and integer programming.,23503
"context. a the local calls for of and global such both consideration Hence, task","Hence, such a task calls for the consideration of both global and local context.",23504
problem fundamental many applications. vision and Image computer is a matting has,Image matting is a fundamental computer vision problem and has many applications.,23505
superiority demonstrate the our clearly previous Experimental results over methods. algorithm of,Experimental results clearly demonstrate the superiority of our algorithm over previous methods.,23506
training. learning-based super-resolution or often network dictionary focus methods Recent on learning (SR),Recent learning-based super-resolution (SR) methods often focus on dictionary learning or network training.,23507
method proposed and learning stage a The stage. of a consists reconstructing,The proposed method consists of a learning stage and a reconstructing stage.,23508
image of results demonstrate the methods. experimental several on LPE-based sets the effectiveness The,The experimental results on several image sets demonstrate the effectiveness of the LPE-based methods.,23509
synthetic this effective and problems. benchmark real approach on show world be We to,We show this approach to be effective on synthetic and real world benchmark problems.,23510
important many in is Object computer applications. an recognition task and vision detection,Object detection and recognition is an important task in many computer vision applications.,23511
HSV better for The domain the is image processed in color detection. color,The image is processed in the HSV color domain for better color detection.,23512
"any-angle tessellated into moves the on allowed, are Agents' square is e.g. which workspace grid","Agents' workspace is tessellated into the square grid on which any-angle moves are allowed, e.g.",23513
planner part a used algorithm as This AA-SIPP(m). multi-agent prioritized of then is,This algorithm is then used as part of a prioritized multi-agent planner AA-SIPP(m).,23514
"that is show theoretical, side under AA-SIPP(m) complete On the well-defined conditions. we","On the theoretical, side we show that AA-SIPP(m) is complete under well-defined conditions.",23515
an learning word-level end-to-end deep for We speech visual propose recognition. architecture,We propose an end-to-end deep learning architecture for word-level visual speech recognition.,23516
"The system residual a is spatiotemporal bidirectional Long convolutional, Short-Term of combination and Memory networks.","The system is a combination of spatiotemporal convolutional, residual and bidirectional Long Short-Term Memory networks.",23517
A key level at in question functional from phenomena. meaning biophysical neuroscience which emerges is,A key question in neuroscience is at which level functional meaning emerges from biophysical phenomena.,23518
image-text Both considered. are retrieval modes text-image and,Both image-text and text-image retrieval modes are considered.,23519
"in class are Traditionally, testing labels identical. and sets the training","Traditionally, class labels in the training and testing sets are identical.",23520
"assumed is, classes. that it the That some query is into pre-defined usually falls","That is, it is usually assumed that the query falls into some pre-defined classes.",23521
classes are performance noticeable drop unseen when the demonstrate training. We testing during a,We demonstrate a noticeable performance drop when the testing classes are unseen during training.,23522
"The available our publicly and benchmarking train/test are split, website. on code, evaluation results","The train/test split, evaluation code, and benchmarking results are publicly available on our website.",23523
important be can particularly is diagnosis melanoma prompt cured Early excision. since with,Early diagnosis is particularly important since melanoma can be cured with prompt excision.,23524
do when change fine-tuned? How much features,How much do features change when fine-tuned?,23525
"with planner, resulting incorporates together effective The FS+, guidance. features heuristic these","The resulting planner, FS+, incorporates these features together with effective heuristic guidance.",23526
are to wide networks a years. of recent problems applied Deep in neural range,Deep neural networks are applied to a wide range of problems in recent years.,23527
Networks with levels. sizes different pooling feature different determine,Networks with different pooling sizes determine different feature levels.,23528
study. The is also computational time in discussed this,The computational time is also discussed in this study.,23529
"under Dermoscopic varying obtained are skin devices, images often with imaging conditions. different acquisition","Dermoscopic skin images are often obtained with different imaging devices, under varying acquisition conditions.",23530
"Globally, of is a major morbidity, cause and Background: influenza hospitalization mortality.","Background: Globally, influenza is a major cause of morbidity, hospitalization and mortality.",23531
effectiveness shown the substantial United vaccination Influenza has in States. protective,Influenza vaccination has shown substantial protective effectiveness in the United States.,23532
"different association Health-care has state-level robust, access with Conclusions: a positive across populations. rates vaccination","Conclusions: Health-care access has a robust, positive association with state-level vaccination rates across different populations.",23533
access. health-care a of potential This advantage highlights population-level expanding,This highlights a potential population-level advantage of expanding health-care access.,23534
"""surely are Such called (s.e. expanded"" states","Such states are called ""surely expanded"" (s.e.",23535
of the one at least then must expanded. states two be,then at least one of the two states must be expanded.,23536
"realized its purpose should and be communication together comprehensively Therefore, with functions. other","Therefore, communication should be realized comprehensively together with its purpose and other functions.",23537
among individuality agents could of the Differentiation also seen. be,Differentiation of individuality among the agents also could be seen.,23538
make model and our trained We available. to publicly code plan,We plan to make our trained model and code publicly available.,23539
"flow, maps, disparity etc. depth optical maps,","depth maps, optical flow, disparity maps, etc.",23540
is method measurement. The determine sparse dense from able a reconstruction to,The method is able to determine a dense reconstruction from sparse measurement.,23541
knowledge this and (MAP) approach. overcome problem posterior inject prior a maximum a We in,We overcome this problem and inject prior knowledge in a maximum a posterior (MAP) approach.,23542
Clusters formed scores. similarity by then thresholding the are,Clusters are then formed by thresholding the similarity scores.,23543
achieve proposed significant methods. the approach Experimental results state-of-the-art that over can improvements demonstrate,Experimental results demonstrate that the proposed approach can achieve significant improvements over state-of-the-art methods.,23544
"given relations the cross-domain unpaired address discovering data. we pairing, avoid To task costly of","To avoid costly pairing, we address the task of discovering cross-domain relations given unpaired data.",23545
behavior group outcomes. members impacts in collective Variation among often,Variation in behavior among group members often impacts collective outcomes.,23546
"optimal of increased scouts this as However, of scouts proportion the increased. persistence","However, this optimal proportion of scouts increased as the persistence of scouts increased.",23547
where modeling rank capture cases are mobility important matrices of scenarios. The one,The cases where mobility matrices are of rank one capture important modeling scenarios.,23548
"cases. global of we equilibria some analysis special Additionally, study for the","Additionally, we study the global analysis of equilibria for some special cases.",23549
hypothesis validity concern of One employed. the the is,One concern is the validity of the hypothesis employed.,23550
"and these on late, joined competition the We to we look unfortunately forward results. improving","We joined the competition unfortunately late, and we look forward to improving on these results.",23551
balancing the pipeline. processing White in a image step fundamental is,White balancing is a fundamental step in the image processing pipeline.,23552
"there importance of previous work been estimation. on illuminant has Given the problem, the much","Given the importance of the problem, there has been much previous work on illuminant estimation.",23553
image on leads our improved method performance sets. that empirically diverse We overall to show,We show empirically that overall our method leads to improved performance on diverse image sets.,23554
orthogonal problem. models the and Procrustes popular problem generalized such includes regression The as,The generalized problem includes popular models such as regression and the orthogonal Procrustes problem.,23555
"increases. performance as data proposed of Further, size improves the algorithm the relative","Further, the relative performance of the proposed algorithm improves as data size increases.",23556
methods various in Several to segment Networks are developed textures to Fully train applications. Convolutional,Several methods are developed to train Fully Convolutional Networks to segment textures in various applications.,23557
(no recognition segmentation). grass from training wood and texture datasets,wood and grass from texture recognition datasets (no training segmentation).,23558
The competition not other of phases considered. the are,The other phases of the competition are not considered.,23559
enjoy discrete of codes hashed The serveral representation IoM nature merits. indices,The discrete indices representation nature of IoM hashed codes enjoy serveral merits.,23560
"information. the biometric strong empowers Firstly, hashing concealment to IoM","Firstly, IoM hashing empowers strong concealment to the biometric information.",23561
the ground non-invertibility This contributes solid to of guarantee.,This contributes to the solid ground of non-invertibility guarantee.,23562
false multi-labels capable labels is COCO and from of dataset. method The detecting,The method is capable of detecting false labels and multi-labels from COCO dataset.,23563
processing. the preliminary or and any got We results post challenge pre submitted without to,We got and submitted preliminary results to the challenge without any pre or post processing.,23564
We resource. formation many species share where investigate of ecological stable the same networks the,We investigate the formation of stable ecological networks where many species share the same resource.,23565
a occurs stable We such as naturally result ecosystem show extinctions. of that,We show that such stable ecosystem naturally occurs as a result of extinctions.,23566
"This extinctions. describes to pressure on parameter system the by converge,","This parameter describes the pressure on system to converge, by extinctions.",23567
field computer algorithms that Machine builds is that a learning science of learn.,Machine learning is a field of computer science that builds algorithms that learn.,23568
meet. two in to been the these Less which has attention separate areas given concerns,Less attention has been given to the areas in which these two separate concerns meet.,23569
"applications and processing. problem, learning image video which has is important an Subspace in many","Subspace learning is an important problem, which has many applications in image and video processing.",23570
find low-dimensional signals to can used be and It a images. representation of,It can be used to find a low-dimensional representation of signals and images.,23571
the The proposed and outliers algorithm learn for subspace jointly the detect images. to tries,The proposed algorithm tries to jointly detect the outliers and learn the subspace for images.,23572
(RL). Two-timescale Reinforcement Approximation algorithms used in widely Stochastic (SA) Learning are,Two-timescale Stochastic Approximation (SA) algorithms are widely used in Reinforcement Learning (RL).,23573
two Their have that are using updated stepsizes. distinct iterates parts,Their iterates have two parts that are updated using distinct stepsizes.,23574
"novel work, their for a finite sample In develop we this analysis. recipe","In this work, we develop a novel recipe for their finite sample analysis.",23575
is probability'. of The obtain as bound `lock-in we known type,The type of bound we obtain is known as `lock-in probability'.,23576
"on this extract From stepsize selection. key then result, we latter insights","From this latter result, we then extract key insights on stepsize selection.",23577
"are compound Moreover, publicly few figure available. decomposition implementations for","Moreover, few implementations for compound figure decomposition are publicly available.",23578
learning as transfer this automatically well exemplars. using training synthesizing overcome as challenge We,We overcome this challenge using transfer learning as well as automatically synthesizing training exemplars.,23579
about paper. discussion Numerical are in and also the rules examples given the proposed,Numerical examples and discussion about the proposed rules are also given in the paper.,23580
Structured algorithms vector tracking shown based recently. have support performance (SVM) favorable output machine,Structured output support vector machine (SVM) based tracking algorithms have shown favorable performance recently.,23581
"optimization time-consuming sampling and candidate the their limit applications. complex Nonetheless, real-time","Nonetheless, the time-consuming candidate sampling and complex optimization limit their real-time applications.",23582
"we results exploit high-confidence the corruption feedback to tracking model avoid the from problem. Thirdly,","Thirdly, we exploit the feedback from high-confidence tracking results to avoid the model corruption problem.",23583
and publicly be source available. will The experimental code made results,The source code and experimental results will be made publicly available.,23584
propose reliable localization of a new the and video detection forgeries. for We copy-move algorithm,We propose a new algorithm for the reliable detection and localization of video copy-move forgeries.,23585
can one two uncertainty model. are major There of types,There are two major types of uncertainty one can model.,23586
uncertainty captures the Aleatoric noise in observations. inherent,Aleatoric uncertainty captures noise inherent in the observations.,23587
per-pixel semantic regression tasks. under We with and study segmentation depth models framework the,We study models under the framework with per-pixel semantic segmentation and depth regression tasks.,23588
"methods Furthermore, learning machine require training data.","Furthermore, machine learning methods require training data.",23589
optimization recognition joint as and connection their cast zero-shot problem. formalize We a,We formalize their connection and cast zero-shot recognition as a joint optimization problem.,23590
"is The achieving evaluated results. zero-shot recognition proposed on four benchmark outstanding for method datasets,","The proposed method is evaluated for zero-shot recognition on four benchmark datasets, achieving outstanding results.",23591
"a two is Traditionally, step pose estimation considered problem. as","Traditionally, pose estimation is considered as a two step problem.",23592
algorithm a JET thread. real-time The CPU single runs on in,The JET algorithm runs in real-time on a single CPU thread.,23593
"rankings to rankings can cases, be such fuzzy of crisp generalization useful. a more In","In such cases, a generalization of crisp rankings to fuzzy rankings can be more useful.",23594
are in camera shift technology. cameras Event a paradigm,Event cameras are a paradigm shift in camera technology.,23595
"a camera similar We propose visual tracking odometry. formulation, in direct state-of-the-art to","We propose a direct camera tracking formulation, similar to state-of-the-art in visual odometry.",23596
add stage. convolutional a layer preprocessing in We,We add a convolutional layer in preprocessing stage.,23597
The layer. in pre-processing of three CA-CNNs convolutional lies difference the size of kernel,The difference of three CA-CNNs lies in the convolutional kernel size of pre-processing layer.,23598
The method that show satisfactory. practicable proposed and the results experimental is,The experimental results show that the proposed method is practicable and satisfactory.,23599
propose We the respondents a questions. multiple-choice model answers aggregate of answering probabilistic to,We propose a probabilistic model to aggregate the answers of respondents answering multiple-choice questions.,23600
"recognise to different robots (concrete, materials Enabling etc.) metal","Enabling robots to recognise different materials (concrete, metal etc.)",23601
"important e.g. a scene many tasks, for in is","in a scene is important for many tasks, e.g.",23602
infer usually considered phylogenies. genes to in mitochondrial are Variations,Variations in mitochondrial genes are usually considered to infer phylogenies.,23603
"investigations, Using statistical then we properties. particular a more further topology, robustness extracted which displays","Using further statistical investigations, we then extracted a particular topology, which displays more robustness properties.",23604
goodness-of-fit graphical the supply that also It Bayesian address issue model can of procedures adequacy.,It can also supply graphical Bayesian goodness-of-fit procedures that address the issue of model adequacy.,23605
attention (VFOA) conversational of recognized visual prominent cue. focus a has The as been,The visual focus of attention (VFOA) has been recognized as a prominent conversational cue.,23606
We VFOAs in are multi-party interested the estimating associated social and with interactions. tracking,We are interested in estimating and tracking the VFOAs associated with multi-party social interactions.,23607
estimation and cannot based tracking. gaze and both detection eye VFOA Consequently on be,Consequently both gaze and VFOA estimation cannot be based on eye detection and tracking.,23608
eye that between the gaze We a and head exploits movements. propose method correlation,We propose a method that exploits the correlation between eye gaze and head movements.,23609
gaze variables Both latent modeled model. Bayesian are as state-space switching and VFOA a in,Both VFOA and gaze are modeled as latent variables in a Bayesian switching state-space model.,23610
Pairwise are making. of an comparisons criteria) tool decision important modern (multiple,Pairwise comparisons are an important tool of modern (multiple criteria) decision making.,23611
been preprocessing better channel. the to choose each have Results for used,Results have been used to choose the better preprocessing for each channel.,23612
"some others. However, information been it more than known that characters contain long has","However, it has long been known that some characters contain more information than others.",23613
Recognizing objects recognition. is interact in visual with other task crucial how a each,Recognizing how objects interact with each other is a crucial task in visual recognition.,23614
"alternative, proposes To context-aware those recognition an mitigate paper this drawbacks, interaction framework.","To mitigate those drawbacks, this paper proposes an alternative, context-aware interaction recognition framework.",23615
shown is of Significant the the state larger the art datasets. improvement on of,Significant improvement of the state of the art is shown on the larger datasets.,23616
the patterns regularity in visual represent world. discernible the Visual,Visual patterns represent the discernible regularity in the visual world.,23617
scenes. the visual capture or essential nature They of objects,They capture the essential nature of visual objects or scenes.,23618
"boundary-aware In approach, not detection annotations our and achieved. bounding-box are is required","In our approach, bounding-box annotations are not required and boundary-aware detection is achieved.",23619
of methods the is semi-supervised most learning in machine popular one learning. Graph-based,Graph-based semi-supervised learning is one of the most popular methods in machine learning.,23620
"a article, study non-parametric framework. the consistency problem we this In under","In this article, we study the consistency problem under a non-parametric framework.",23621
demonstrating inconsistent. case this give counterexample estimator the for that can a be We,We give a counterexample demonstrating that the estimator for this case can be inconsistent.,23622
findings supported theoretical by numerical The are studies.,The theoretical findings are supported by numerical studies.,23623
"focused context the on research e.g. While either layers, from recent all propagating","While recent research focused on either propagating the context from all layers, e.g.",23624
"low-level or multiple (e.g. having ResNet, layers (including very loss layers) the","ResNet, (including the very low-level layers) or having multiple loss layers (e.g.",23625
"layers importance ignored. close GoogLeNet), of the the to the features higher is","GoogLeNet), the importance of the features close to the higher layers is ignored.",23626
"by $n$ AlexNet extend In the in we GoogLeNet and layers. additional top particular, connections","In particular, we extend AlexNet and GoogLeNet by additional connections in the top $n$ layers.",23627
attention recent Chinese (CFR) font in years. gained recognition significant has,Chinese font recognition (CFR) has gained significant attention in recent years.,23628
have for approach the our Experimental of confirmed new CFR. results effectiveness,Experimental results have confirmed the effectiveness of our new approach for CFR.,23629
to uses that cancer a cells. is cancer of drugs kill treatment Chemotherapy class,Chemotherapy is a class of cancer treatment that uses drugs to kill cancer cells.,23630
typical delivered drugs in protocol of weeks. A of several cycles consists chemotherapeutic three,A typical chemotherapeutic protocol consists of several drugs delivered in cycles of three weeks.,23631
protocols. results there of Our the compelling that dose-dense arguments suggest supporting are use,Our results suggest that there are compelling arguments supporting the use of dose-dense protocols.,23632
"and we protocols suggest the of discuss alternative. limitations Finally, an these","Finally, we discuss the limitations of these protocols and suggest an alternative.",23633
One of suitable function. hardest fitness a parts approach this defining is the of,One of the hardest parts of this approach is defining a suitable fitness function.,23634
to according We real-time performance. of these their each trackers accuracy tracking and ranked,We ranked each of these trackers according to their tracking accuracy and real-time performance.,23635
"on are variability. in rigid and RNNs, often based overly lacking those","those based on RNNs, are often overly rigid and lacking in variability.",23636
"""ground-truth"" encourages descriptions. while high captions to the resemblance suppressing other principle reasonable This","This principle encourages high resemblance to the ""ground-truth"" captions while suppressing other reasonable descriptions.",23637
"restrictive also such BLEU METEOR, methods. favor and","BLEU and METEOR, also favor such restrictive methods.",23638
a that sequence It generator is is nontrivial. noteworthy training,It is noteworthy that training a sequence generator is nontrivial.,23639
"Further, like electron elements etc. other synthesis, chemiosmosis, rotary transport the ATP chain, elaborate","Further, other elements like the elaborate electron transport chain, chemiosmosis, rotary ATP synthesis, etc.",23640
point are to also evaluated systems such critically complicated that non-viable. are out,are also critically evaluated to point out that such complicated systems are non-viable.,23641
necessitates a cellular explanatory The communication new paradigm respiration. for,The communication necessitates a new explanatory paradigm for cellular respiration.,23642
an in construction the learning. for state-of-the-art We machine over overview provide coreset,We provide an overview over the state-of-the-art in coreset construction for machine learning.,23643
unseen Given the side classes and seen of dataset classes of (e.g. information,Given the dataset of seen classes and side information of unseen classes (e.g.,23644
are probabilistic decision-theoretic of Influence diagrams a extension graphical models.,Influence diagrams are a decision-theoretic extension of probabilistic graphical models.,23645
solve how can this be problem. we In Goddard used to paper they the show,In this paper we show how they can be used to solve the Goddard problem.,23646
results techniques. that Experimental current method better super-resolution show proposed image single than performs the,Experimental results show that the proposed method performs better than current single image super-resolution techniques.,23647
the We of networks. discuss complexity computational sum-product in approximating posteriori inference maximum a,We discuss the computational complexity of approximating maximum a posteriori inference in sum-product networks.,23648
using empirically algorithm against max-product and proposed realistic networks. the We analyze synthetic,We empirically analyze the proposed algorithm against max-product using synthetic and realistic networks.,23649
classification. questions work learning addresses theory of the nonparametric in This for open various active,This work addresses various open questions in the theory of active learning for nonparametric classification.,23650
"previously transitions Some unconfirmed. were remained conjectured, such but","Some such transitions were previously conjectured, but remained unconfirmed.",23651
"methods. deep both we and explore traffic To density, deeply optimization based understand learning based","To deeply understand traffic density, we explore both deep learning based and optimization based methods.",23652
extensively counting and compared on tasks Both and methods evaluated are different datasets.,Both methods are extensively evaluated and compared on different counting tasks and datasets.,23653
"approximations. of most only However, finite multi-fidelity a methods number assume","However, most multi-fidelity methods assume only a finite number of approximations.",23654
"continuous be In applications however, spectrum practical might available. a approximations many of","In many practical applications however, a continuous spectrum of approximations might be available.",23655
"Bayesian we In setting. this optimisation method, a work, develop for BOCA, this","In this work, we develop a Bayesian optimisation method, BOCA, for this setting.",23656
experiments. several in synthetic other and outperforms BOCA baselines real,BOCA outperforms several other baselines in synthetic and real experiments.,23657
"Recently, for architectures. standard studied intensively semi-supervised deep learning been has CNN","Recently, semi-supervised deep learning has been intensively studied for standard CNN architectures.",23658
"image the many set state-of-the-art Fully segmentation Networks Convolutional tasks. However, for (FCNs)","However, Fully Convolutional Networks (FCNs) set the state-of-the-art for many image segmentation tasks.",23659
"hardware crucial. robust embedded and real-time energy-efficient Accordingly, is vision delivering performance","Accordingly, energy-efficient embedded vision hardware delivering real-time and robust performance is crucial.",23660
This chips measurements algorithms. two these supported analysis is by implement from that,This analysis is supported by measurements from two chips that implement these algorithms.,23661
vectors that correlated. highly is are the observed weight It usually,It is observed that the weight vectors are usually highly correlated.,23662
of the RGB layout focuses on from room monocular estimation task image. This paper a,This paper focuses on the task of room layout estimation from a monocular RGB image.,23663
of engaging interactions. probabilities nodes individual in cooperation) the of,probabilities of cooperation) of the individual nodes engaging in interactions.,23664
long or Deliberating learning. large spaces continuous challenges state in been on have standing reinforcement,Deliberating on large or continuous state spaces have been long standing challenges in reinforcement learning.,23665
R package an nimble algorithms on for inference and models. is constructing hierarchical conducting,nimble is an R package for constructing algorithms and conducting inference on hierarchical models.,23666
SMC models an overview We provide algorithms. of commonly-used state-space and first,We first provide an overview of state-space models and commonly-used SMC algorithms.,23667
C++ compiled more into execution. in for run or be algorithms R efficient can These,These algorithms can be run in R or compiled into C++ for more efficient execution.,23668
extend nimble's code. users methods SMC in how high-level easily illustrates can This,This illustrates how users can easily extend nimble's SMC methods in high-level code.,23669
and tongue purposes. conjunctiva work anemia analyzes of sites for This pallor screening the,This work analyzes the pallor sites of conjunctiva and tongue for anemia screening purposes.,23670
"the are for segmented. the site outer Similarly, images, tongue pallor and tongue regions inner","Similarly, for the tongue pallor site images, the inner and outer tongue regions are segmented.",23671
is segments video videos actions) important an large-scale analysis. from untrimmed human step for,human actions) segments from untrimmed videos is an important step for large-scale video analysis.,23672
novel Unit Temporal Regression model. a We propose (TURN) Network,We propose a novel Temporal Unit Regression Network (TURN) model.,23673
"organ The performed and detection, detection is including steps, segmentation. three lesion in process lesion","The process is performed in three steps, including organ detection, lesion detection and lesion segmentation.",23674
high lesion to of handling its be extremely proves in diversity. method The robust,The method proves to be robust in its handling of extremely high lesion diversity.,23675
"as knowledge a We parallel embedding. our ParaGraphE, for library framework provides name graph which","We name our framework as ParaGraphE, which provides a library for parallel knowledge graph embedding.",23676
platform(e.g. given special a exploiting features of By hardware the,By exploiting the special hardware features of a given platform(e.g.,23677
conflicting to find ways Artificial autonomous goals. cope may to systems also with need,Artificial autonomous systems may also need to find ways to cope with conflicting goals.,23678
and un-ordered by and razor evolutionary is Ockham's scheme perspectives. favored proposed The,The proposed scheme is un-ordered and favored by Ockham's razor and evolutionary perspectives.,23679
plant with threatens for and populations demand products extinction. animal Consumer many,Consumer demand for plant and animal products threatens many populations with extinction.,23680
"We instance framework for general conceptually a present flexible, object and segmentation. simple,","We present a conceptually simple, flexible, and general framework for object instance segmentation.",23681
target attribute learning and common descriptions. recognition (re-ID) Person re-identification at pedestrian a share,Person re-identification (re-ID) and attribute recognition share a common target at learning pedestrian descriptions.,23682
existing labels into of only Most methods re-ID identity consideration. take pedestrians,Most existing re-ID methods only take identity labels of pedestrians into consideration.,23683
"addition, correlations we the and In considering re-weight predictions among the attribute the dependencies attributes.","In addition, we re-weight the attribute predictions considering the dependencies and correlations among the attributes.",23684
theories. have impact These techniques economic on,These techniques have impact on economic theories.,23685
more impact We expensive) picking of test the models. also deeper (and,We also test the impact of picking deeper (and more expensive) models.,23686
"rate of these such is with Even convergence slow. the tricks, methods","Even with these tricks, the convergence rate of such methods is slow.",23687
This loss behaved. empirically better is proxy-based,This proxy-based loss is empirically better behaved.,23688
The method art methods recognition known accuracy. the the state-of-the of proposed terms in outperforms,The proposed method outperforms the known state-of-the art methods in terms of the recognition accuracy.,23689
illustrative provided of development is overview this An via examples.,An overview of this development is provided via illustrative examples.,23690
"under provided. into various fusion Furthermore, parameter also are contexts insights selection","Furthermore, insights into parameter selection under various fusion contexts are also provided.",23691
"We Segment introduce (SegLink), Linking method. text an oriented detection","We introduce Segment Linking (SegLink), an oriented text detection method.",23692
"Both by scales fully-convolutional multiple end-to-end network. are elements densely an at detected neural trained,","Both elements are detected densely at multiple scales by an end-to-end trained, fully-convolutional neural network.",23693
detections links. are segments connected combining produced by Final by,Final detections are produced by combining segments connected by links.,23694
the makes more convenient output notion approach pathways. to This the categorize,This notion makes the approach more convenient to categorize the output pathways.,23695
the are discussed. Implications of method networks metabolic proposed on,Implications of the proposed method on metabolic networks are discussed.,23696
"limits application. However, practical process, optimization its framework a their requires iterative which slow","However, their framework requires a slow iterative optimization process, which limits its practical application.",23697
been neural to speed networks with up neural approximations Fast feed-forward proposed transfer. have style,Fast approximations with feed-forward neural networks have been proposed to speed up neural style transfer.,23698
"via gradient applying gradient Inspired works, the obtain we filters. prior by information","Inspired by the prior works, we obtain gradient information via applying gradient filters.",23699
dataset. our Transient Experiments that performance Attributes state-of-the-art approach confirm achieves the on,Experiments confirm that our approach achieves the state-of-the-art performance on Transient Attributes dataset.,23700
recognition detection work system. plate and fully license automated details Sighthounds This,This work details Sighthounds fully automated license plate detection and recognition system.,23701
they (e.g. robust different so The CNNs are conditions under are and that fine-tuned trained,The CNNs are trained and fine-tuned so that they are robust under different conditions (e.g.,23702
a (e.g. variety work and of plate across license templates can,and can work across a variety of license plate templates (e.g.,23703
"a Generative (MSG-Net), build performance. which the With real-time Layer, achieves Network Multi-style CoMatch we","With the CoMatch Layer, we build a Multi-style Generative Network (MSG-Net), which achieves real-time performance.",23704
comparing Our to quality has method state-of-the-art approaches. superior achieved image,Our method has achieved superior image quality comparing to state-of-the-art approaches.,23705
"be implementations frameworks and available. and pre-trained PyTorch models Our will Torch, MXNet publicly for","Our implementations and pre-trained models for Torch, PyTorch and MXNet frameworks will be publicly available.",23706
Privacy for problem distributed traditionally has a motivation major solving. been,Privacy has traditionally been a major motivation for distributed problem solving.,23707
utilitarian agents. assuming Here computation that problem performed by the among is we approach,Here we approach the problem by assuming that computation is performed among utilitarian agents.,23708
"improve we appearance to information of integrate In SORT. paper, this performance the","In this paper, we integrate appearance information to improve the performance of SORT.",23709
"queries associations nearest in online measurement-to-track visual neighbor establish space. During application, we using appearance","During online application, we establish measurement-to-track associations using nearest neighbor queries in visual appearance space.",23710
"dataset depth, multi-modality and Radiation including Skeleton. sources, provides data Infrared Our RGB, also","Our dataset also provides multi-modality data sources, including RGB, depth, Infrared Radiation and Skeleton.",23711
detection believe large-scale action will dataset benefit the future researches community. on this for We,We believe this large-scale dataset will benefit future researches on action detection for the community.,23712
of effectiveness the proposed superior demonstrate approach. Extensive experiments and the performance,Extensive experiments demonstrate the effectiveness and superior performance of the proposed approach.,23713
task. this This makes search for neighbor Euclidean biased the nearest,This makes the Euclidean nearest neighbor search biased for this task.,23714
a filtering in equivalent graph the signal linear of This frequency domain. to sparse is,This is equivalent to linear graph filtering of a sparse signal in the frequency domain.,23715
"we basis graph compute the To offline. approximate search, an speed of online Fourier up","To speed up online search, we compute an approximate Fourier basis of the graph offline.",23716
"these datasets are First, in content i.e., some covered. classes lacking are pre-defined only diversity,","First, these datasets are lacking in content diversity, i.e., only some pre-defined classes are covered.",23717
"method OCR-based show experiment, improves performance. we that In proposed the baseline the","In experiment, we show that the proposed OCR-based method improves the baseline performance.",23718
"statistics. paper, class introduce novel CNNs that exploit of In a this we second-order","In this paper, we introduce a novel class of CNNs that exploit second-order statistics.",23719
"are over settings, many modern at acquired iteratively data rather than all In once. time,","In many modern settings, data are acquired iteratively over time, rather than all at once.",23720
"offline as as online, batch. Such opposed known settings are or to","Such settings are known as online, as opposed to offline or batch.",23721
decades. challenging Recognition Facial Expression been has for Automated (FER) a task,Automated Facial Expression Recognition (FER) has been a challenging task for decades.,23722
often hyperparameter achieve require tuning rigorous These methods good results. to,These methods often require rigorous hyperparameter tuning to achieve good results.,23723
traditional to methods Deep object Neural (DNN) have recognition. Recently shown visual in outperform Networks,Recently Deep Neural Networks (DNN) have shown to outperform traditional methods in visual object recognition.,23724
"available viz. publicly network databases, evaluate three on proposed our We","We evaluate our proposed network on three publicly available databases, viz.",23725
manners. Experiments and are in subject-independent cross-database performed,Experiments are performed in subject-independent and cross-database manners.,23726
might Local issues. approaches to relieve help these,Local approaches might help to relieve these issues.,23727
universal consistent. kernel there We that show are methods will learnt locally,We will show there that locally learnt kernel methods are universal consistent.,23728
 is Brassica  Background: a pest Mustard aphid of oilseeds. major,  Background: Mustard aphid is a major pest of Brassica oilseeds.,23729
. presently resistance in juncea source is available for aphid Brassica No,No source for aphid resistance is presently available in Brassica juncea .,23730
"Brassica to A is be fruticulosa wild to known resistant crucifer, aphid. mustard","A wild crucifer, Brassica fruticulosa is known to be resistant to mustard aphid.",23731
Some consistent genotypes resistant reaction. of test showed the,Some of the test genotypes showed consistent resistant reaction.,23732
both temporal into information procedure. context the We the incorporate and spatial training,We incorporate both the spatial and temporal context information into the training procedure.,23733
"significantly our various algorithms. outperforms datasets, our on algorithm the model and evaluate We compared","We evaluate our model on various datasets, and our algorithm significantly outperforms the compared algorithms.",23734
"within these topics. there of In literature, been development recent has both the exciting","In the recent literature, there has been exciting development within both of these topics.",23735
"the information interest. remain algorithms of limits estimation of precise In graphon estimation, and context","In the context of graphon estimation, precise information limits and estimation algorithms remain of interest.",23736
obviously additional requires these tasks Incorporating efforts.,Incorporating these tasks obviously requires additional efforts.,23737
model exploits Our CNN learning the residual framework.,Our CNN model exploits the residual learning framework.,23738
"it compute uses to the features loss. Additionally, normalized","Additionally, it uses normalized features to compute the loss.",23739
extensive on Our different datasets. show excellent generalization experiments,Our extensive experiments show excellent generalization on different datasets.,23740
The method for beneficial scene is proposed incidental particularly localizing texts.,The proposed method is particularly beneficial for localizing incidental scene texts.,23741
"standard On other the state-of-the-art also focused method texts, performance. reaches scene our datasets with","On other standard datasets with focused scene texts, our method also reaches the state-of-the-art performance.",23742
different conducted The experiment viz. two sets in was,The experiment was conducted in two different sets viz.,23743
"a in with replications. unprotected, protected/sprayed design, three block and randomized","protected/sprayed and unprotected, in a randomized block design, with three replications.",23744
of learning approach human present supervised for We an weakly actions.,We present an approach for weakly supervised learning of human actions.,23745
"of image problem geo-localization. address the cross-view paper, this we In","In this paper, we address the problem of cross-view image geo-localization.",23746
"detect the query the reference R-CNN employ Faster and to we in images. buildings First,","First, we employ the Faster R-CNN to detect buildings in the query and reference images.",23747
"proposed. capture recurrent to information, connection sufficiently spectral a is bidirectional the Besides,","Besides, to sufficiently capture the spectral information, a bidirectional recurrent connection is proposed.",23748
are presented verify and theory. examples Illustrative to the clarify,Illustrative examples are presented to verify and clarify the theory.,23749
"shortcuts the or avoid are mapping connections exploding/vanishing gradient problem. to Correspondingly, utilized skip identity","Correspondingly, the skip connections or identity mapping shortcuts are utilized to avoid gradient exploding/vanishing problem.",23750
"addition, connections better performance. centered naturally the the have activation skip In which to led","In addition, the skip connections have naturally centered the activation which led to better performance.",23751
residual image architectures have been also compared. super-resolution for Different,Different residual architectures for image super-resolution have also been compared.,23752
more text is of challenging often problem The much classification. classification image than,The problem of image classification is often much more challenging than text classification.,23753
of input are the up-convolution the introduced preserve The to image. information layers fine-grained,The up-convolution layers are introduced to preserve the fine-grained information of the input image.,23754
coarse-to-fine boosts is new proposal on a vehicle architecture object the detection. Its based that,Its architecture is based on a new coarse-to-fine object proposal that boosts the vehicle detection.,23755
"on reside manifold which images is underlying those investigated. However, rarely the","However, the underlying manifold which those images reside on is rarely investigated.",23756
potential also of DBL learning machine describe optimization. various of subfields to and applications We,We also describe potential applications of DBL to various subfields of machine learning and optimization.,23757
object work from for easy-to-define class This boundaries interactions. deviates,This work deviates from easy-to-define class boundaries for object interactions.,23758
data can mixture statistical a a How we on set? model massive train,How can we train a statistical mixture model on a massive data set?,23759
coresets for how we work In of this show mixtures to Gaussians. construct,In this work we show how to construct coresets for mixtures of Gaussians.,23760
We up-to-scale motions. a SfM to relative use of in chain it compute,We use it to compute SfM in a chain of up-to-scale relative motions.,23761
dense region segmentation. These eventually constitute learning regions complete localized object for a and semantic,These localized regions eventually constitute a dense and complete object region for learning semantic segmentation.,23762
problem. this We step addressing take towards a,We take a step towards addressing this problem.,23763
"challenges, (MPN-COV) we address To Normalized present Matrix method. a these Power Covariance","To address these challenges, we present a Matrix Power Normalized Covariance (MPN-COV) method.",23764
"the advantage both In its quantitatively addition, Log-Euclidean analyze and over we metric. qualitatively well-known","In addition, we analyze both qualitatively and quantitatively its advantage over the well-known Log-Euclidean metric.",23765
"is, suitable for approach style transfer. as is photorealistic not this However,","However, as is, this approach is not suitable for photorealistic style transfer.",23766
appearance share source similar target (e.g. The might classes and,The source and target classes might share similar appearance (e.g.,23767
(e.g. bear or cat appear fur background fur) against is to similar similar,bear fur is similar to cat fur) or appear against similar background (e.g.,23768
applications vision. computer approach found has traditional wide The of in range a bag-of-words,The traditional bag-of-words approach has found a wide range of applications in computer vision.,23769
"amounts however, of suffers Given large model from lack the of a power. discriminative data,","Given large amounts of data, however, the model suffers from a lack of discriminative power.",23770
"large Moreover, can datasets. on very be kernel the computation expensive","Moreover, the kernel computation can be very expensive on large datasets.",23771
and In general fitting. model methods and complex these building inefficient for require pipelines,In general these methods require complex and inefficient pipelines for model building and fitting.,23772
the to learned retrieve used classification and are features medical results images. The,The learned features and the classification results are used to retrieve medical images.,23773
"best class are used. achieved For when results are based predictions retrieval,","For retrieval, best results are achieved when class based predictions are used.",23774
multimodal to proposed organs. method images medical retrieve best different body is for suited The,The proposed method is best suited to retrieve multimodal medical images for different body organs.,23775
Within differences. academic there gender WMC were fields no,Within academic fields there were no WMC gender differences.,23776
human challenge in lies large the One variations view captured main in actions.,One main challenge lies in the large view variations in captured human actions.,23777
three Our the on significant approaches benchmark model improvement state-of-the-art over achieves datasets.,Our model achieves significant improvement over the state-of-the-art approaches on three benchmark datasets.,23778
networks study of We images neural classification deep quality for with distortions.,We study deep neural networks for classification of images with quality distortions.,23779
"for. networks quality they trained However, have been not on poorly fine-tuned perform that distortions","However, fine-tuned networks perform poorly on quality distortions that they have not been trained for.",23780
"our on The distortion. trained a particular model type ""experts"" in of are","The ""experts"" in our model are trained on a particular type of distortion.",23781
"detection untrimmed of activity video streams. address in We continuous, problem the","We address the problem of activity detection in continuous, untrimmed video streams.",23782
entire is with and model classification The jointly end-to-end losses. optimized localization trained,The entire model is trained end-to-end with jointly optimized localization and classification losses.,23783
two Video estimation interpolation motion steps: pixel involves synthesis. and frame typically,Video frame interpolation typically involves two steps: motion estimation and pixel synthesis.,23784
estimation. the Such motion depends of two-step heavily on approach a quality,Such a two-step approach heavily depends on the quality of motion estimation.,23785
robotic object actively in applications. is an Planar studied vision-based problem tracking,Planar object tracking is an actively studied problem in vision-based robotic applications.,23786
to the quality. carefully The truth ground ensure annotated semi-manually is,The ground truth is carefully annotated semi-manually to ensure the quality.,23787
planar proposed benchmark future benefit expect studies We to tracking. on object the,We expect the proposed benchmark to benefit future studies on planar object tracking.,23788
dataset be but fit tend fails the custom to Solutions used to to generalize.,Solutions tend to be custom fit to the dataset used but fails to generalize.,23789
fusion provides of experts stability different among classification rates in datasets. The,The fusion of experts provides stability in classification rates among different datasets.,23790
the of transient cortical raises question targeting. localization The its of dynein,The transient cortical localization of dynein raises the question of its targeting.,23791
"either be directed dynein tracks the can At having classified cortex, motion. diffusive-like as or","At the cortex, dynein tracks can be classified as having either directed or diffusive-like motion.",23792
subjectivity method this human categorization robust freeing by segmentation Automatic error. and it makes from,Automatic segmentation makes this categorization method robust by freeing it from subjectivity and human error.,23793
"semi-automatic In for we this segmentation. decomposition used thesis, method mode empirical","In this thesis, we used empirical mode decomposition method for semi-automatic segmentation.",23794
comparative of results. All not parameters did desired give these,All of these parameters did not give desired comparative results.,23795
on We proposed the show that outperforms our model model benchmark baseline datasets.,We show that our proposed model outperforms the baseline model on benchmark datasets.,23796
is (EMR). in growing to There methods interest Medical Records learning applying machine Electronic,There is growing interest in applying machine learning methods to Electronic Medical Records (EMR).,23797
"EMR vary Across quality can institutions, widely. different however,","Across different institutions, however, EMR quality can vary widely.",23798
"documents. approach the query-by-string segmentation-free of this for we handwritten word problem paper, In spotting","In this paper, we approach the problem of segmentation-free query-by-string word spotting for handwritten documents.",23799
preliminary that experiments promising validate method task. We on this provide our,We provide promising preliminary experiments that validate our method on this task.,23800
"of neuroscientists to comprehensively. cortical identify In connectivity field seek the connectomics,","In the field of connectomics, neuroscientists seek to identify cortical connectivity comprehensively.",23801
"fully we multi-stage multi-recursive-input this paper, to convolutional propose this problem. address In networks","In this paper, we propose multi-stage multi-recursive-input fully convolutional networks to address this problem.",23802
of justice Discussions precision. assessments  fairness lack  in typically criminal conceptual Objectives: risk,  Objectives: Discussions of fairness in criminal justice risk assessments typically lack conceptual precision.,23803
often substitutes Rhetoric for analysis. too careful,Rhetoric too often substitutes for careful analysis.,23804
empirical data from illustration arraignments. an using provide also We,We also provide an empirical illustration using data from arraignments.,23805
"across In protected different practice, base complication major groups. a legally rates different is","In practice, a major complication is different base rates across different legally protected groups.",23806
consider There challenging tradeoffs. to a need is,There is a need to consider challenging tradeoffs.,23807
"In an for present LV work, iterative delineation. algorithm this we segmentation","In this work, we present an iterative segmentation algorithm for LV delineation.",23808
many research Our metrics. over the previous outperforms method,Our method outperforms the previous research over many metrics.,23809
of models. computation enables exact This the GMM IVQR for the estimators,This enables exact computation of the GMM estimators for the IVQR models.,23810
shown to is classification Entropy performance. contribute to,Entropy is shown to contribute to classification performance.,23811
studies computational question. Simulation statistical be the to discuss will tradeoff presented and,Simulation studies will be presented to discuss the statistical and computational tradeoff question.,23812
"the majority the surgical In cases, automatic step of of first segmentation tools. is the","In the majority of cases, the first step is the automatic segmentation of surgical tools.",23813
work two previous upon major ways. in We improve,We improve upon previous work in two major ways.,23814
a analysis. genome of large fundamental in genomic sequences computational Alignment task is,Alignment of large genomic sequences is a fundamental task in computational genome analysis.,23815
on underlying therefore methods points. and quality Speed anchor depend of these the,Speed and quality of these methods therefore depend on the underlying anchor points.,23816
generalized proposed. are for models linear oracle inequalities Non-asymptotic,Non-asymptotic oracle inequalities for generalized linear models are proposed.,23817
approach illustrate of datasets. our on the experiments several good Numerical performances,Numerical experiments illustrate the good performances of our approach on several datasets.,23818
data mineable from high-dimensional extraction a Quantitative radiomics. of images is process medical known as,Quantitative extraction of high-dimensional mineable data from medical images is a process known as radiomics.,23819
"success in has achieved recent many In great years, applications. learning vision deep computer","In recent years, deep learning has achieved great success in many computer vision applications.",23820
networks image (CNNs) classification. as neural to major a Convolutional lately emerged have approach,Convolutional neural networks (CNNs) have lately emerged as a major approach to image classification.,23821
"a this unit paper, the (ACU). we introduce convolution called active convolution unit In","In this paper, we introduce a convolution unit called the active convolution unit (ACU).",23822
backpropagation during Its shape learned training. be can through,Its shape can be learned through backpropagation during training.,23823
has unit proposed few Our a advantages.,Our proposed unit has a few advantages.,23824
the on Experimental proposed demonstrate the scheme. of effectiveness image results clustering,Experimental results on image clustering demonstrate the effectiveness of the proposed scheme.,23825
"based with heuristic, Boltzmann on divide-and-conquer. is along probabilistic distribution sampling sampling the The","The sampling distribution is based on the Boltzmann sampling heuristic, along with probabilistic divide-and-conquer.",23826
regret (a.k.a. sampling the of analyze Thompson (TS) We,We analyze the regret of Thompson sampling (TS) (a.k.a.,23827
that observe (envelope). We amplitude the with specific beta patterns correlate,We observe that specific patterns correlate with the beta amplitude (envelope).,23828
"the stochastic we Next, mutual optimize information to show descent. using how also objective gradient","Next, we also show how to optimize the mutual information objective using stochastic gradient descent.",23829
have been introduced recognition. recently gender networks Local neural for deep,Local deep neural networks have been recently introduced for gender recognition.,23830
"computationally good are performance they very to Although, train. they expensive achieve very","Although, they achieve very good performance they are very computationally expensive to train.",23831
We more classification. and are investigated also which gender for have age patches discriminative,We have also investigated which patches are more discriminative for age and gender classification.,23832
"discriminative more it embedding. Thus, ensures a","Thus, it ensures a more discriminative embedding.",23833
scans and CXR voluminous medical workloads These radiologists on significant practitioners. place,These voluminous CXR scans place significant workloads on radiologists and medical practitioners.,23834
effective Organ CXR. computer-aided detection segmentation step is on obtain to a crucial,Organ segmentation is a crucial step to obtain effective computer-aided detection on CXR.,23835
that natural segmentation. and accurate produces Extensive show highly experiments method our,Extensive experiments show that our method produces highly accurate and natural segmentation.,23836
time-consuming difficult makes encompass during all This and training. to it and conditions tasks,This makes it time-consuming and difficult to encompass all tasks and conditions during training.,23837
"Celebrity FG-NET, Faces the and MORPH, (AGFW), Cross-Age Dataset Wild in AginG (CACD).","FG-NET, MORPH, AginG Faces in the Wild (AGFW), and Cross-Age Celebrity Dataset (CACD).",23838
memory much representing budget. a outputs with resolution limited allows This higher,This allows representing much higher resolution outputs with a limited memory budget.,23839
for fibril-like forming found The cooperative be structures. the transition sequences most is to,The transition is found to be the most cooperative for sequences forming fibril-like structures.,23840
define regions. predefined strategy a on image to This segmentation step supervised depends interest,This strategy depends on a supervised image segmentation step to define predefined interest regions.,23841
to task a used the environment. analyze practical trees was counting real A considering sample,A real sample was used to analyze the counting trees task considering a practical environment.,23842
incorporate network Our image networks. stylization different can,Our network can incorporate different image stylization networks.,23843
proposed the We quantitatively. method both clearly qualitatively show per-frame the baseline that outperforms and,We show that the proposed method clearly outperforms the per-frame baseline both qualitatively and quantitatively.,23844
been detection. infrared have for state-of-the-art Many methods small proposed target,Many state-of-the-art methods have been proposed for infrared small target detection.,23845
They high-contrast well work homogeneous the with backgrounds images and targets. on,They work well on the images with homogeneous backgrounds and high-contrast targets.,23846
in progress substantial object research related vision. for tasks and localization computer Recent has demonstrated,Recent research has demonstrated substantial progress in object localization and related tasks for computer vision.,23847
with research. issues We the current these address,We address these issues with the current research.,23848
from an neural distance object. approximate convolutional a offset We the network train to target,We train a convolutional neural network to approximate an offset distance from the target object.,23849
Bayesian of for search employ then the localization We accurate active a target.,We then employ a Bayesian active search for accurate localization of the target.,23850
improvement method exhibits over substantial regression baseline a method. Our this,Our method exhibits a substantial improvement over this baseline regression method.,23851
interactions for (i.e. object bounds annotations of temporal Manual,Manual annotations of temporal bounds for object interactions (i.e.,23852
"training detection and and typical end times) input to algorithms. start recognition, are localization","start and end times) are typical training input to recognition, localization and detection algorithms.",23853
method. proposes counting crowd This a paper,This paper proposes a crowd counting method.,23854
"regression Conventional multi-class (e,g., counting predictor and methods one generally crowd utilize classifier).","Conventional crowd counting methods generally utilize one predictor (e,g., regression and multi-class classifier).",23855
"one can predictor such count only large However, appearance not with well. targets changes","However, such only one predictor can not count targets with large appearance changes well.",23856
"By proposed the the robustness selected changes. large method has integrating the CNNs, to appearance","By integrating the selected CNNs, the proposed method has the robustness to large appearance changes.",23857
"predictor a we appearance. that to confirm automatically each Moreover, specialized specific","Moreover, we confirm that each predictor automatically specialized to a specific appearance.",23858
uncertainty and prior incorporation information. tractable naturally facilitate These methods quantification of,These methods naturally facilitate tractable uncertainty quantification and incorporation of prior information.,23859
"First, explicit derive we (geometric) for the bounds upper of rate convergence.","First, we derive explicit upper bounds for the (geometric) rate of convergence.",23860
trace implications. has class and theoretical The practical useful property,The trace class property has useful theoretical and practical implications.,23861
"compact, corresponding is eigenvalues It operator Markov the are that implies summable. and its","It implies that the corresponding Markov operator is compact, and its eigenvalues are summable.",23862
several implemented algorithms package the learning machine multilabel We in mlr. classification,We implemented several multilabel classification algorithms in the machine learning package mlr.,23863
"there versions access is of classification the to rFerns. Moreover, and randomForestSRC multilabel","Moreover, there is access to the multilabel classification versions of randomForestSRC and rFerns.",23864
face a pose challenge detection. variations unconstrained Large-scale in still,Large-scale variations still pose a challenge in unconstrained face detection.,23865
remove included is negative further This step samples. to hard,This step is included to further remove hard negative samples.,23866
input task output changes in by the causal main changes is Estimating analysis.,Estimating output changes by input changes is the main task in causal analysis.,23867
changes SOM output changes. when input Such pattern of the a the weighted,Such a weighted SOM pattern of the output changes when the input changes.,23868
for been measure dissimilarity Many methods of proposed have the patterns.,Many methods have been proposed for the dissimilarity measure of patterns.,23869
"it a change. challenge However, when how major remains the to patterns attempting measure","However, it remains a major challenge when attempting to measure how the patterns change.",23870
"that each class-wise Moreover, learn the appears image. for trees object decision we in","Moreover, we learn class-wise decision trees for each object that appears in the image.",23871
polynomial show this and be it. algorithm We a problem to for develop NP-hard,We show this problem to be NP-hard and develop a polynomial algorithm for it.,23872
"grades use our For a Science students' Computer in real department. we on experiments, data","For our experiments, we use real data on students' grades in a Computer Science department.",23873
re-identification Person across people to aims (re-id) non-overlapping match views. camera,Person re-identification (re-id) aims to match people across non-overlapping camera views.,23874
far widely most So is existing RGB-based appearance the used in works.,So far the RGB-based appearance is widely used in most existing works.,23875
"into from test fused At views time, are keyframes. predictions multiple","At test time, predictions from multiple views are fused into keyframes.",23876
several enforcing training multi-view for and methods We testing. during and propose analyze consistency,We propose and analyze several methods for enforcing multi-view consistency during training and testing.,23877
and accuracy Both of quantitative approach. our the qualitative and the robustness analyses demonstrate,Both qualitative and quantitative analyses demonstrate the accuracy and the robustness of our approach.,23878
metastases. Resection of accelerated growth primary by of often tumors followed is,Resection of primary tumors is often followed by accelerated growth of metastases.,23879
when latent infinite of are components We also investigate introduced. models well-definedness the the,We also investigate the well-definedness of the models when infinite latent components are introduced.,23880
a We efficiently. present structure a explores algorithm space pragmatic larger search which,We present a pragmatic search algorithm which explores a larger structure space efficiently.,23881
long a segmentation Semantic in standing been has task challenging computer vision.,Semantic segmentation has been a long standing challenging task in computer vision.,23882
"datasets, approaches visual on benchmarking our tested several We challenging i.e.","We tested our approaches on several challenging benchmarking visual datasets, i.e.",23883
earlier extends scene works the based on these dynamics. This paper,This paper extends these earlier works based on the scene dynamics.,23884
fast lines. observation computation this of epipolar on precise and Based achieve we,Based on this observation we achieve fast and precise computation of epipolar lines.,23885
improvements We research state-of-the-art previous over result. this introduce lead to three that,We introduce three improvements over previous research that lead to this state-of-the-art result.,23886
"of evaluate the feature feature statistical sets, protocols, modeling and match matches. We","We evaluate feature sets, match protocols, and the statistical modeling of feature matches.",23887
"on the set, change proposals makes benchmark On of algorithm precise, our two-thirds accurate scenes.","On the benchmark set, our algorithm makes precise, accurate change proposals on two-thirds of scenes.",23888
becomes the number objects/parts large. expensive This once scalable of not computationally is and,This is computationally expensive and not scalable once the number of objects/parts becomes large.,23889
"localization. part nonparametric paper, object and In we propose for method a data-driven this","In this paper, we propose a nonparametric data-driven method for object and part localization.",23890
"transfer iterative In predicted propose an particular, refine the gradually bounding that boxes. we strategy","In particular, we propose an iterative transfer strategy that gradually refine the predicted bounding boxes.",23891
"are located the extracted convolutional recognition. on features deep parts, Based objects for and","Based on the located objects and parts, deep convolutional features are extracted for recognition.",23892
for multiple both object approach Our and segmentation. suitable single is,Our approach is suitable for both single and multiple object segmentation.,23893
into from annihilation This structured a $k-t$ low formed matrix relation translates samples. the rank,This annihilation relation translates into a structured low rank matrix formed from the $k-t$ samples.,23894
This is by for challenge pattern introducing a selection. approached constraint,This challenge is approached by introducing a constraint for pattern selection.,23895
"stability are these delta-measure of and closed itemsets, Cosine the among interest, constraints. associated robustness,","Cosine interest, robustness, stability of closed itemsets, and the associated delta-measure are among these constraints.",23896
state-of-the-art from Extending to challenging. object is image detectors video,Extending state-of-the-art object detectors from image to video is challenging.,23897
"present flow-guided accurate detection. video feature framework and an for aggregation, learning end-to-end We object","We present flow-guided feature aggregation, an accurate and end-to-end learning framework for video object detection.",23898
temporal instead. level on coherence feature leverages It,It leverages temporal coherence on feature level instead.,23899
"validity comparison numerical methods. demonstrate results other Eventually, its with and efficiency in numerical","Eventually, numerical results demonstrate its validity and efficiency in comparison with other numerical methods.",23900
"we In study paper, the this completing appearance problem of objects. the of occluded challenging","In this paper, we study the challenging problem of completing the appearance of occluded objects.",23901
of cues incorporates model objects. their context to detect these in all the parts Our,Our model incorporates all these cues to detect parts in the context of their objects.,23902
its performance showing against We robustness this by technique the evaluate uncertainties. of also,We also evaluate the performance of this technique by showing its robustness against uncertainties.,23903
Outdoor dynamic extremely range. lighting high has,Outdoor lighting has extremely high dynamic range.,23904
"alternative we propose work, an In this approach.","In this work, we propose an alternative approach.",23905
based report smartphone values. the of application of to learning colorimetric pH machine We detection,We report the application of machine learning to smartphone based colorimetric detection of pH values.,23906
"networks. different different this require approach, Under problems","Under this approach, different problems require different networks.",23907
to them. between the gap classification) due,classification) due to the gap between them.,23908
a predicted consistency PI maintains loss Our latent distribution. and classification between PI-based,Our PI-based classification loss maintains a consistency between latent PI and predicted distribution.,23909
The expectation-maximization updated and estimated are network and iteratively PI in an procedure. latent,The latent PI and network are iteratively estimated and updated in an expectation-maximization procedure.,23910
important goal towards to this collect human is and An learn step perceptions.,An important step towards this goal is to collect and learn human perceptions.,23911
using deep easily frameworks. This can implemented be learning module existing,This module can be easily implemented using existing deep learning frameworks.,23912
than to of variance. posterior generalises functions wide proof The range a cost other the,The proof generalises to a wide range of cost functions other than the posterior variance.,23913
"deep, a embedding model. metric based fully Our convolutional on is similarity","Our similarity metric is based on a deep, fully convolutional embedding model.",23914
benchmark. We competitive VOC segmentation instance show on results Pascal the,We show competitive results on the Pascal VOC instance segmentation benchmark.,23915
"propose this level we pruning scheme In for paper, pruning. neuron a gradually global","In this paper, we propose a gradually global pruning scheme for neuron level pruning.",23916
scheme given performance. Our find original in network automatically under would sub-network a thinner a,Our scheme would automatically find a thinner sub-network in original network under a given performance.,23917
in Google navigation UAV We framework for propose Map a GPS-denied environment. aided,We propose a framework for Google Map aided UAV navigation in GPS-denied environment.,23918
navigation provides Geo-referenced closures. does and not drift-free require loop localization,Geo-referenced navigation provides drift-free localization and does not require loop closures.,23919
"efficient. The via initialized which UAV correlation, simple position is is and","The UAV position is initialized via correlation, which is simple and efficient.",23920
then subsequent predict optical in to frames. We flow its use position,We then use optical flow to predict its position in subsequent frames.,23921
to UAV. search to to filter fine the employ localize particle conduct We a coarse,We employ particle filter to conduct a coarse to fine search to localize the UAV.,23922
Syllogism individual. commonsense ingredient of essential an reasoning for is an,Syllogism is an essential ingredient for commonsense reasoning of an individual.,23923
have population complex introgression Human of and changing history populations of a size.,Human populations have a complex history of introgression and of changing population size.,23924
process people. a about Lifelogging of information life rich of daily is of collecting source,Lifelogging is a process of collecting rich source of information about daily life of people.,23925
"constancy research. camera paper, color In a we designed dataset provide novel invariant this for","In this paper, we provide a novel dataset designed for camera invariant color constancy research.",23926
also The under scenes captured different illuminations. five lab are,The lab scenes are also captured under five different illuminations.,23927
This the of as color effect research shading on well. allows,This allows research on the effect of color shading as well.,23928
developed. for An posterior method is efficient MCMC computation,An efficient MCMC method for posterior computation is developed.,23929
the simulated proposed data. method investigated on performance is The of,The performance of the proposed method is investigated on simulated data.,23930
"illustrate Finally, we EU-SILC. from application its real on data","Finally, we illustrate its application on real data from EU-SILC.",23931
our We of efficiency and approach evaluated in terms precision.,We evaluated our approach in terms of efficiency and precision.,23932
"but corresponding scarce. theoretical exist, results are Popular heuristics","Popular heuristics exist, but corresponding theoretical results are scarce.",23933
with performance. is corresponding proposed algorithm strong empirical A,A corresponding algorithm is proposed with strong empirical performance.,23934
(who's as We problem pairwise the formulate better?),We formulate the problem as pairwise (who's better?),23935
"collections, ranking deep video supervised ranking. of using","ranking of video collections, using supervised deep ranking.",23936
We of via the analysis parameters. robustness our of its demonstrate approach sensitivity,We demonstrate the robustness of our approach via sensitivity analysis of its parameters.,23937
parameters. to gradient We can modified tune algorithm a descent apply therefore,We can therefore apply a modified gradient descent algorithm to tune parameters.,23938
system is polynomial a of each equations. a point to data solution,each data point is a solution to a system of polynomial equations.,23939
to cancer of brain preferred one techniques emerged the patients. treat therapy has Radiation as,Radiation therapy has emerged as one of the preferred techniques to treat brain cancer patients.,23940
"a very narrow is a dose to area. high radiation treatment, delivered very During of","During treatment, a very high dose of radiation is delivered to a very narrow area.",23941
"still usually Nevertheless, which performed, delineation and manually operator-dependent. inefficient is task is","Nevertheless, delineation task is usually still manually performed, which is inefficient and operator-dependent.",23942
attempts automatizing reported. Several of this have process,Several attempts of automatizing this process have reported.,23943
"the results optic region. marginal in however, when analyzing organs","however, marginal results when analyzing organs in the optic region.",23944
yielded features segmentation. on of proposed improvements to the Incorporation,Incorporation of proposed features yielded to improvements on the segmentation.,23945
respect contours. to proposed manual evaluated the with approach of also Performance was,Performance of the proposed approach was also evaluated with respect to manual contours.,23946
locations receives set instant. from network each image input a LSTM at of An,An LSTM network receives input from a set of image locations at each instant.,23947
each held features important cues Appearance on hand in hand. motion on and objects give,Appearance features give important cues on hand motion and on objects held in each hand.,23948
LSTM over to learns time. temporal mechanism Finally features how attention a fuse,Finally a temporal attention mechanism learns how to fuse LSTM features over time.,23949
"accelerators implemented Thus, been have FPGAs. successfully multiple CNN on","Thus, multiple CNN accelerators have been successfully implemented on FPGAs.",23950
"or units elements FPGA as limited. logic resources DSP Unfortunately, remain such","Unfortunately, FPGA resources such as logic elements or DSP units remain limited.",23951
Biological subject are environmental populations to fluctuating conditions.,Biological populations are subject to fluctuating environmental conditions.,23952
"this we model analyze growth. problem Here, in simple mathematical a of population","Here, we analyze this problem in a simple mathematical model of population growth.",23953
Those by pre-trained Approximations DNNs the works (LRA). Low-Rank decomposed directly,Those works directly decomposed the pre-trained DNNs by Low-Rank Approximations (LRA).,23954
"and is in AlexNet, comprehensively ResNets, our evaluated of approach GoogLeNet. The effectiveness","The effectiveness of our approach is comprehensively evaluated in ResNets, AlexNet, and GoogLeNet.",23955
outperforms our procedure in competitive We experiments baselines show the that proposed model segmentation. in,We show in our experiments that the proposed model outperforms competitive baselines in procedure segmentation.,23956
"two a a We region-of-interest corner and based CNN deconvolution introduce novelties, estimator model. based","We introduce two novelties, a corner based region-of-interest estimator and a deconvolution based CNN model.",23957
object suggests Further desirable. finegrained model is analysis our well particularly when localization performs,Further analysis suggests our model performs particularly well when finegrained object localization is desirable.,23958
more kind This stable. the of function asymmetric makes GAN training loss,This kind of asymmetric loss function makes the GAN training more stable.,23959
We allows for interaction alternative formulations. our tensor over significantly how better performance show,We show how our interaction tensor allows for significantly better performance over alternative formulations.,23960
robots. classify is The to objects fundamental for ability,The ability to classify objects is fundamental for robots.,23961
architectures. them transfer learning pre-trained deep processable This makes procedure by,This transfer learning procedure makes them processable by pre-trained deep architectures.,23962
"Such instances in aerial commonly photography. for example, arise,","Such instances commonly arise, for example, in aerial photography.",23963
"a constant of) small factor within $\varepsilon(g,\mathcal{S})$.","within a small constant factor of) $\varepsilon(g,\mathcal{S})$.",23964
"which scenes. we non-planar view techniques selection a Finally, our method extends to multi-resolution present","Finally, we present a multi-resolution view selection method which extends our techniques to non-planar scenes.",23965
"to mammal small dynamics, plant dynamics) trophic understand effects. community","small mammal dynamics, plant community dynamics) to understand trophic effects.",23966
an predicted refines also the additional We optional that use poses. step,We also use an optional additional step that refines the predicted poses.,23967
"on by we proof-of-concept, a visual induced walking. focus As linking experiences","As a proof-of-concept, we focus on linking visual experiences induced by walking.",23968
images walking person distinctive connects first This two through paths. similar,This connects two distinctive first person images through similar walking paths.,23969
complex such of images Histopathology to as the are crucial diseases study cancer.,Histopathology images are crucial to the study of complex diseases such as cancer.,23970
"a prognosis of in key histologic role and nuclei play diagnosis, The disease characteristics analysis.","The histologic characteristics of nuclei play a key role in disease diagnosis, prognosis and analysis.",23971
first vision is unsupervised applications. for the network CAE computer detection Our,Our CAE is the first unsupervised detection network for computer vision applications.,23972
methods Here we extend exploit unknown the derivative function. to from these information,Here we extend these methods to exploit derivative information from the unknown function.,23973
"networks baseline such networks. Further, different two-stream often use","Further, such networks often use different baseline two-stream networks.",23974
"data not training tasks, will many paired available. for However, be","However, for many tasks, paired training data will not be available.",23975
Quantitative approach. demonstrate methods superiority against prior comparisons several the our of,Quantitative comparisons against several prior methods demonstrate the superiority of our approach.,23976
can labels given defining object specify tree a order hierarchy. partial Any over a,Any given tree can specify a partial order over object labels defining a hierarchy.,23977
"such do use not allow models. full However, existing optimization of methods","However, existing optimization methods do not allow full use of such models.",23978
"these for methods trees cases. In work does applying to practice, arbitrary not except simple","In practice, applying these methods to arbitrary trees does not work except for simple cases.",23979
many complex trees. state-of-the-art segmentation of biomedical diverse examples for show We,We show state-of-the-art biomedical segmentation for many diverse examples of complex trees.,23980
"for tree this a (MCTS). In present paper, Carlo we new algorithm Monte search parallel","In this paper, we present a new algorithm for parallel Monte Carlo tree search (MCTS).",23981
approach structured parallel pattern for pipeline to programming provides The MCTS. first the,The pipeline pattern provides for the first structured parallel programming approach to MCTS.,23982
Stixel representation. World discussed in this a such The paper is,The Stixel World discussed in this paper is such a representation.,23983
generalization Stixel mathematical is a the embedded model. novel into Our formulation for,Our generalization is embedded into a novel mathematical formulation for the Stixel model.,23984
"normal which may degrade LRS performance. Additionally, tissue blurs registration regions,","Additionally, LRS blurs normal tissue regions, which may degrade registration performance.",23985
"switch triggers adaptation the between to dynamic thereby conditions. The both allowing cellular modes, microenvironment","The cellular microenvironment triggers the switch between both modes, thereby allowing adaptation to dynamic conditions.",23986
"invasion. specific vivo tumor for Importantly, characteristic these conditions in are","Importantly, these specific conditions are characteristic for in vivo tumor invasion.",23987
learning Deep has object and for results classification demonstrated excellent image detection. been to achieve,Deep learning has been demonstrated to achieve excellent results for image classification and object detection.,23988
"learning However, video deep (e.g. on analysis of the impact","However, the impact of deep learning on video analysis (e.g.",23989
"and handle two-stream feature separately. spatial these employ methods framework CNN temporal to Also,","Also, these methods employ two-stream CNN framework to handle spatial and temporal feature separately.",23990
the for on recognition Fields algorithms based Many are of state-of-the-art (CRFs). gesture Random Conditional,Many of the state-of-the-art algorithms for gesture recognition are based on Conditional Random Fields (CRFs).,23991
values. use to the it values then the to assign latent label We,We then use it to assign the latent values to the label values.,23992
the streams. from evaluate our method We human on video gestures task recognizing of,We evaluate our method on the task of recognizing human gestures from video streams.,23993
"performed different grouping datasets, by in experiments The binary were labels. generated","The experiments were performed in binary datasets, generated by grouping different labels.",23994
feature information paper that algorithms. segmentation video global for a proposes geodesic-distance-based encodes improved This,This paper proposes a geodesic-distance-based feature that encodes global information for improved video segmentation algorithms.,23995
of algorithms. can used feature part The be various as and is generic,The feature is generic and can be used as part of various algorithms.,23996
"fast introduces method. image deformable paper This a Quicksilver, registration","This paper introduces Quicksilver, a fast deformable image registration method.",23997
model. network A encoder-decoder deep used is the prediction as,A deep encoder-decoder network is used as the prediction model.,23998
software. available as open-source is freely Quicksilver an,Quicksilver is freely available as an open-source software.,23999
approaches tracking tracking-by-detection popular Adaptive are for arbitrary objects.,Adaptive tracking-by-detection approaches are popular for tracking arbitrary objects.,24000
"of heavily and their these the effectiveness invested efficiency However, are approaches detectors. in","However, these approaches are heavily invested in the efficiency and effectiveness of their detectors.",24001
"accumulating samples in misclassification errors of borderline detector the introduce in tracking. Furthermore,","Furthermore, misclassification of borderline samples in the detector introduce accumulating errors in tracking.",24002
"rate usually and estimation non-differentiable, is quantizer discrete entropy control. is for the However, required","However, the quantizer is non-differentiable, and discrete entropy estimation usually is required for rate control.",24003
compression a convolutional very challenging (CNN)-based image system. network it These to make develop,These make it very challenging to develop a convolutional network (CNN)-based image compression system.,24004
potential makers. public health decision Ensemble to to predictions reliable more the offer methods deliver,Ensemble methods offer the potential to deliver more reliable predictions to public health decision makers.,24005
problem. tackle to construct this networks We adversarial,We construct adversarial networks to tackle this problem.,24006
"fake detect network a the images. other On hand, aims discriminator to","On the other hand, a discriminator network aims to detect fake images.",24007
"adversarial and information. trained consider The spatial, is appearance, network to with losses three","The network is trained with three losses to consider spatial, appearance, and adversarial information.",24008
of The parts are locations determines the spatial predicted correct. whether loss,The spatial loss determines whether the locations of predicted parts are correct.,24009
output loss adversarial images The realistic. are ensures,The adversarial loss ensures output images are realistic.,24010
involve of economics estimation of number a simultaneous parameters. empirical in Many settings large applied,Many applied settings in empirical economics involve simultaneous estimation of a large number of parameters.,24011
similarity. of The spectral pair-wise clustering capturing central the is process in point,The central point in spectral clustering is the process of capturing pair-wise similarity.,24012
notion Node density to exploited the local of Topological structure. are Features and define,Topological Node Features are exploited to define the notion of density and local structure.,24013
construction These affinity into of the are local incorporated features the matrix.,These local features are incorporated into the construction of the affinity matrix.,24014
outer mucus a layer. Colonic loose a layer consists dense epithelium-attached of and,Colonic mucus consists of a loose outer layer and a dense epithelium-attached layer.,24015
The human of outer layer microbiota representatives (HGM). gut is inhabited by various the,The outer layer is inhabited by various representatives of the human gut microbiota (HGM).,24016
mucin glycans. The demonstrated degrade different analyzed to abilities (iv) known genomes,(iv) The analyzed genomes demonstrated different abilities to degrade known mucin glycans.,24017
methods. approach significantly the existing state-of-the-art outperforms Our,Our approach significantly outperforms the existing state-of-the-art methods.,24018
demonstrate networks. that We strategy completely synthetically trained this outperforms,We demonstrate that this strategy outperforms completely synthetically trained networks.,24019
"Finally, several reconstructions approaches. we state-of-the-art and show to compare high-quality approach our","Finally, we show high-quality reconstructions and compare our approach to several state-of-the-art approaches.",24020
image for the This expands domain-invariant approach feature samples. space smoke,This approach expands the domain-invariant feature space for smoke image samples.,24021
"the detector. design architectures choices smoke with several to applied are different Experimentally, deep","Experimentally, several deep architectures with different design choices are applied to the smoke detector.",24022
the set. get a satisfactory can on ultimate test framework The result,The ultimate framework can get a satisfactory result on the test set.,24023
"of framework. adopted to discrimination specific SVMs the Then, is linear enhance template","Then, template specific linear SVMs is adopted to enhance the discrimination of framework.",24024
"templates merged different results. final the Finally, matching are corresponding scores multiple as","Finally, multiple matching scores corresponding different templates are merged as the final results.",24025
simple framework unified IJB-A exhibits excellent This dataset. on performance,This simple unified framework exhibits excellent performance on IJB-A dataset.,24026
crucial a tracking instrument computer-assisted for various interventions. is Real-time requirement,Real-time instrument tracking is a crucial requirement for various computer-assisted interventions.,24027
illumination. for natural a introduce shape-from-shading variational multi-view method under We,We introduce a variational method for multi-view shape-from-shading under natural illumination.,24028
This for setting breast intraoperative lumpectomy. simulates cancer work an,This work simulates an intraoperative setting for breast cancer lumpectomy.,24029
when small methods Commonly can fail database is used available. only a,Commonly used methods can fail when only a small database is available.,24030
"is configurations. collected several our to method applied on data data locally Moreover,","Moreover, our method is applied to locally collected data on several data configurations.",24031
trials. over the reported The average are results the different,The reported results are the average over the different trials.,24032
robust competing Function yielded norm regularization than higher results and more methods.,Function norm regularization yielded higher and more robust results than competing methods.,24033
image years impressive In single made were the advances super-resolution. recent for,In the recent years impressive advances were made for single image super-resolution.,24034
this success. is a of learning Deep behind big part,Deep learning is behind a big part of this success.,24035
the Deep(er) key external priors and modeling ingredients. design are architecture,Deep(er) architecture design and external priors modeling are the key ingredients.,24036
especially Our images large adaptation under structures with or favors repetitive resolutions.,Our adaptation especially favors images with repetitive structures or under large resolutions.,24037
data. be applied previously can learned to The then model unseen online,The learned model can then be applied online to previously unseen data.,24038
Results: protocol. a knee approach clinical on The network evaluated imaging variational is,Results: The variational network approach is evaluated on a clinical knee imaging protocol.,24039
and on proteins the have demonstrated fold Recent can simulations ribosome. that experiments,Recent experiments and simulations have demonstrated that proteins can fold on the ribosome.,24040
"questions. the fitness from co-translational remain generality However, of and effects extent folding open resulting","However, the extent and generality of fitness effects resulting from co-translational folding remain open questions.",24041
Here a we analysis evolutionary of selection uncovers for folding. evidence report that genome-wide co-translational,Here we report a genome-wide analysis that uncovers evidence of evolutionary selection for co-translational folding.,24042
image We from images. encoder-decoder deep for architecture a deformation multimodal introduce prediction,We introduce a deep encoder-decoder architecture for image deformation prediction from multimodal images.,24043
faster than Our order of is an LDDMM. approach optimization-based magnitude,Our approach is an order of magnitude faster than optimization-based LDDMM.,24044
"pathway, network. our by student neural consisting training of a a deep We goal achieve","We achieve our goal by training a student pathway, consisting of a deep neural network.",24045
)and scores. to various alignments the compare juxtaposing by obtained,)and to compare various alignments by juxtaposing the obtained scores.,24046
"background. Discrminative to target approach classification its trackers, the a from separate employ","Discrminative trackers, employ a classification approach to separate the target from its background.",24047
sequences The proposed on outperformed bearing trackers tracking challenges. tracker various state-of-the-art different,The proposed tracker outperformed state-of-the-art trackers on different sequences bearing various tracking challenges.,24048
subnet. Layout-Aware subnet and a Multi-Patch a,a Multi-Patch subnet and a Layout-Aware subnet.,24049
"is object in supervised object late, weakly great detection importance Of recognition. with","Of late, weakly supervised object detection is with great importance in object recognition.",24050
"deep detectors on learning, many Based achieved weakly have results. promising supervised","Based on deep learning, weakly supervised detectors have achieved many promising results.",24051
overall architecture. explore property of random the further the CNN-DCN We,We further explore the property of the overall random CNN-DCN architecture.,24052
"quality. inverted be with images satisfactory Surprisingly, can","Surprisingly, images can be inverted with satisfactory quality.",24053
well theoretical empirical evidence are Extensive analysis as as provided.,Extensive empirical evidence as well as theoretical analysis are provided.,24054
"optionally users objects for the of can In customizing hyperlapses. interest choose addition,","In addition, users can optionally choose objects of interest for customizing the hyperlapses.",24055
has and robust real-time to monocular image for be effective learning Deep shown relocalisation.,Deep learning has shown to be effective for robust and real-time monocular image relocalisation.,24056
"it tuning. naive However, require function, using loss expensive hyper-parameters was which trained with a","However, it was trained using a naive loss function, with hyper-parameters which require expensive tuning.",24057
"In treatment. problem the a fundamental we this theoretical paper, more give","In this paper, we give the problem a more fundamental theoretical treatment.",24058
"In this we a approach. work, propose structure-aware regression","In this work, we propose a structure-aware regression approach.",24059
joints. bones using adopts a representation pose It of reparameterized instead,It adopts a reparameterized pose representation using bones instead of joints.,24060
validates the effectiveness evaluation our approach. of Comprehensive,Comprehensive evaluation validates the effectiveness of our approach.,24061
"exploited measure ), can better be with queries. which similarity for","), which can be exploited for better similarity measure with queries.",24062
provide results. is Our work to first such the,Our work is the first to provide such results.,24063
rates We provide and convergence both in expectation high-probability. with,We provide convergence rates both in expectation and with high-probability.,24064
be These by low-altitude. and detection algorithms change at tasks supported can achieved mosaicking,These tasks can be supported by mosaicking and change detection algorithms achieved at low-altitude.,24065
"these datasets for available. algorithms testing are Currently, not public","Currently, public datasets for testing these algorithms are not available.",24066
system is convolutional Our formed two neural cascade networks. by the of,Our system is formed by the cascade of two convolutional neural networks.,24067
of charge text. convolutional detecting first network containing and in is fully The is areas,The first network is fully convolutional and is in charge of detecting areas containing text.,24068
results possibly but a This inaccurate segmentation very image. of reliable in the input,This results in a very reliable but possibly inaccurate segmentation of the input image.,24069
been by Srinivasa foreshadowed proposal may have This Ramanujan.,This proposal may have been foreshadowed by Srinivasa Ramanujan.,24070
resting-state diagnosis biomarkers. MRI include candidates functional (rs-fcMRI)-based for connectivity data-driven Promising,Promising candidates for data-driven diagnosis include resting-state functional connectivity MRI (rs-fcMRI)-based biomarkers.,24071
large distinct and numbers of reproducible cell cells. requires Unbiased cataloging of types,Unbiased and reproducible cataloging of distinct cell types requires large numbers of cells.,24072
"processes aggregating the are independently. previous work, designed hashing the In and","In previous work, the aggregating and the hashing processes are designed independently.",24073
retrieval This more discriminative leads codes improved and hash to accuracy. binary,This leads to more discriminative binary hash codes and improved retrieval accuracy.,24074
We several SIFT experiments on datasets both benchmark perform features. and with extensive convolutional retrieval,We perform extensive retrieval experiments on several benchmark datasets with both SIFT and convolutional features.,24075
social only spread of and effect The makes the media more severe. problem the this,The spread and the effect of the social media only makes this problem more severe.,24076
can image detection understanding as forgery problem. Image be a seen of subset,Image forgery detection can be seen as a subset of image understanding problem.,24077
forgery performance We detection compare for and problem. algorithm of automated humans the an,We compare the performance of an automated algorithm and humans for forgery detection problem.,24078
"as context). same the in such used be ""Easy/difficult Terms image"" will","Terms such as ""Easy/difficult image"" will be used in the same context).",24079
ideal only for features. GP parameters using the searching image allows,GP allows searching for ideal parameters using only the image features.,24080
commonplace life. networking everyday become mobile on a devices has of Social,Social networking on mobile devices has become a commonplace of everyday life.,24081
"photo become process to In in has mobile advances due the imaging. trivial capturing addition,","In addition, photo capturing process has become trivial due to the advances in mobile imaging.",24082
lot people capture a be visually-attractive. they Hence and want them photos everyday of to,Hence people capture a lot of photos everyday and they want them to be visually-attractive.,24083
"This tools. automated, rise enhancement one-touch to given has","This has given rise to automated, one-touch enhancement tools.",24084
(e.g. The existing typical heuristically methods machine-learned,The existing typical machine-learned methods heuristically (e.g.,24085
features the way integrate MF propose a in formulation. also to the We image,We also propose a way to integrate the image features in the MF formulation.,24086
predictions. CNN components Predicted using are cells found by the on sperm connected,Predicted sperm cells are found by using connected components on the CNN predictions.,24087
We the of a size optimization threshold of on detected investigate components. parameter,We investigate optimization of a threshold parameter on the size of detected components.,24088
"scale investigate large we this In between denoising paper, the dataset. classification interaction and on","In this paper, we investigate the interaction between denoising and classification on large scale dataset.",24089
it classification to beneficial conclude that incorporate is models. We in,We conclude that it is beneficial to incorporate in classification models.,24090
"we affect hand, the also performance. classification noise how other On study","On the other hand, we also study how noise affect classification performance.",24091
"we end, to come counter-intuitive. a number of interesting In being the conclusions, some","In the end, we come to a number of interesting conclusions, some being counter-intuitive.",24092
traffic congestion pricing systems routing Dynamic in recent getting research. importance are and,Dynamic traffic routing systems and congestion pricing are getting importance in recent research.,24093
estimation prediction vehicle important component of and systems. an such Lane is density,Lane prediction and vehicle density estimation is an important component of such systems.,24094
the results reduction It in also in number false detections. large of,It also results in large reduction in the number of false detections.,24095
found is entire results. The give tested real-world videos approach on and to is acceptable,The entire approach is tested on real-world videos and is found to give acceptable results.,24096
effectiveness such The of with is computational mining intractable requirements. associated often methods,The effectiveness of such mining methods is often associated with intractable computational requirements.,24097
state-of-the-art Current offline tailored action systems detection are applications. batch-processing for,Current state-of-the-art action detection systems are tailored for offline batch-processing applications.,24098
of the allows Epigenomics human identification signatures epigenetic in diseases.,Epigenomics allows the identification of epigenetic signatures in human diseases.,24099
are Ontology frequently interaction between protein-protein score alignments to (PPI) networks. terms (GO) Gene used,Gene Ontology (GO) terms are frequently used to score alignments between protein-protein interaction (PPI) networks.,24100
is not specific a Our generic on depend proposal and action method. approach does,Our approach is generic and does not depend on a specific action proposal method.,24101
recognition problem paper video human we from In sequences. action address this the of,In this paper we address the problem of human action recognition from video sequences.,24102
We their weaknesses. extensively and also evaluate strengths,We also extensively evaluate their strengths and weaknesses.,24103
Computational active visual an research recently has aesthetics become area.,Computational visual aesthetics has recently become an active research area.,24104
images determining aesthetic of more the appropriate. is ranking Hence the,Hence determining the aesthetic ranking of the images is more appropriate.,24105
Depth of (DFF) is classical problems the in vision. from computer focus inverse ill-posed one,Depth from focus (DFF) is one of the classical ill-posed inverse problems in computer vision.,24106
us allows stacks to This varying digitally focal of sizes. create,This allows us to digitally create focal stacks of varying sizes.,24107
and for acquired workstation transferred Images were post-processing. a to,Images were acquired and transferred to a workstation for post-processing.,24108
ex were for verification. vivo of animals different A set used,A different set of animals were used for ex vivo verification.,24109
"Various spleen, exsanguination. and including procured organs, liver, brain, heart, stomach lungs, were following intestines,","Various organs, including brain, lungs, heart, liver, spleen, stomach and intestines, were procured following exsanguination.",24110
determinations. measurements value vivo with the of were activity ex consistent image-based The,The ex vivo measurements of activity were consistent with the image-based value determinations.,24111
new of testing radionanotheranostic agents. in This allow vivo quantitative work may for,This work may allow for quantitative in vivo testing of new radionanotheranostic agents.,24112
little such They a single from very make image. color can predictions input data as,They can make predictions from very little input data such as a single color image.,24113
The with resolution objects interior can the voxels. exterior and of coarse be represented,The exterior and interior of the objects can be represented with coarse resolution voxels.,24114
type. on is dependent input a specific not approach Our,Our approach is not dependent on a specific input type.,24115
high predictions shows Our accurate analysis are resolution resolution predictions. more than that low our,Our analysis shows that our high resolution predictions are more accurate than low resolution predictions.,24116
tool era. regression increased standard is importance with Big Data statistical Nonparametric a the in,Nonparametric regression is a standard statistical tool with increased importance in the Big Data era.,24117
used local points can Boundary but polynomial difficulties to pose be alleviate regression additional them.,Boundary points pose additional difficulties but local polynomial regression can be used to alleviate them.,24118
"multimodal fusion, the Vector use For Support we Machine. traditional","For multimodal fusion, we use the traditional Support Vector Machine.",24119
Our approach a and collision generative discriminatively with combines learned points. salient detection model,Our approach combines a generative model with collision detection and discriminatively learned salient points.,24120
"(DR), vision-threatening diabetes risk a retinopathy All developing the have people diabetic complication. of with","All people with diabetes have the risk of developing diabetic retinopathy (DR), a vision-threatening complication.",24121
can and timely Early due the reduce to of blindness DR. treatment occurrence detection,Early detection and timely treatment can reduce the occurrence of blindness due to DR.,24122
the speed accuracy of diagnosis the DR Computer-aided has benefit detection. potential and improving in,Computer-aided diagnosis has the potential benefit of improving the accuracy and speed in DR detection.,24123
network proposed cascaded stages. two comprises The,The proposed network comprises two cascaded stages.,24124
show achieves performance. Experimental algorithm the BMCNN results that proposed state-of-the-art,Experimental results show that the proposed BMCNN algorithm achieves state-of-the-art performance.,24125
"detail, repetitive restore can In both BMCNN structures. and irregular","In detail, BMCNN can restore both repetitive and irregular structures.",24126
by context a region regions boundary. classifiers considering Our segmentation a around potentially-erroneous detect large,Our classifiers detect potentially-erroneous regions by considering a large context region around a segmentation boundary.,24127
that We a fully-automatic merge/split mode probability make present a uses decisions. threshold also to,We also present a fully-automatic mode that uses a probability threshold to make merge/split decisions.,24128
from software work is this and upon request. data The available,The software and data from this work is available upon request.,24129
for illumination adverse detecting presents a method This  under pedestrians  novel paper conditions.,  This paper presents a novel method for detecting pedestrians under adverse illumination conditions.,24130
lessons. An reveals historical interesting application set mortality data a to,An application to a mortality data set reveals interesting historical lessons.,24131
determine factors this parameters temporal-network of furthermore structural convergence. We the influencing the,We furthermore determine the temporal-network structural factors influencing the parameters of this convergence.,24132
The representations. deep features are better learnt shared through,The deep features are better learnt through shared representations.,24133
of from estimation image. depth This single problem the addresses still paper a,This paper addresses the problem of depth estimation from a single still image.,24134
on the state-of-the-art algorithm public two achieve evaluation and We performance. benchmarks our,We evaluation our algorithm on two public benchmarks and achieve the state-of-the-art performance.,24135
dataset be pose The available. synthetic hand will,The synthetic hand pose dataset will be available.,24136
"potentials However, the employed translational of motion power to pairwise models discriminative limit the differences.","However, pairwise potentials limit the discriminative power of the employed motion models to translational differences.",24137
field while the not pyramid can enlarge Our structure greatly sacrificing computation efficiency.,Our pyramid structure can greatly enlarge the field while not sacrificing computation efficiency.,24138
benefit VGA-size upsampling on progressive Extra quasi-realtime network input. testing adaptive includes depth and for,Extra benefit includes adaptive network depth and progressive upsampling for quasi-realtime testing on VGA-size input.,24139
"In convolution-deconvolution detection. propose network for a keypoint novel we facial this paper,","In this paper, we propose a novel convolution-deconvolution network for facial keypoint detection.",24140
routing are function and a functions implemented framework. transform Both multi-task through convolutions the in,Both transform functions and the routing function are implemented through convolutions in a multi-task framework.,24141
"makes learning for features estimation. a intensity approach this for Potentially, AU facial VAEs suitable","Potentially, this makes VAEs a suitable approach for learning facial features for AU intensity estimation.",24142
"the existing from separately VAE-based apply methods classifiers learned Yet, most encoded features.","Yet, most existing VAE-based methods apply classifiers learned separately from the encoded features.",24143
"use recognition. Therefore, to are for seis- prestack facies seismic data we increasingly inclined mic","Therefore, we are increasingly inclined to use prestack seismic data for seis- mic facies recognition.",24144
"or clustering conventional using Then, techniques (e.g. classification","Then, using conventional classification or clustering techniques (e.g.",24145
"or recognition. achieve self- we extracted seismic maps) K-means features, can on the organizing facies","K-means or self- organizing maps) on the extracted features, we can achieve seismic facies recognition.",24146
to data prestack region. We method model our physical from and the LZB applied,We applied our method to the prestack data from physical model and LZB region.,24147
the is superior that our approach to re- shows sult The conventionals.,The re- sult shows that our approach is superior to the conventionals.,24148
priors. learning The of latter complex shape allows,The latter allows learning of complex shape priors.,24149
challenge. method on CREMI known the gives segmentation best The the results seeded segmentation,The method gives the best known seeded segmentation results on the CREMI segmentation challenge.,24150
video captioning differs existing technical in three from aspects. systems It,It differs from existing video captioning systems in three technical aspects.,24151
scheme is to in the weakly winner-takes-all phase. associate sentences training region-sequences A adopted to,A winner-takes-all scheme is adopted to weakly associate sentences to region-sequences in the training phase.,24152
is for a recognition. fundamental Facial landmark module face pose-invariant localization,Facial landmark localization is a fundamental module for pose-invariant face recognition.,24153
The proposal-refinement distance-aware and softmax proposed end-to-end dual-pathway architecture. relies system on functions,The proposed end-to-end system relies on distance-aware softmax functions and dual-pathway proposal-refinement architecture.,24154
improves into performance. We our model this description which supervision its then provide (noisy) greatly,We then provide this (noisy) supervision into our description model which greatly improves its performance.,24155
work Our proposed over description model w.r.t. prior improves,Our proposed description model improves over prior work w.r.t.,24156
provides resolution. and co-reference generated quality additionally grounding local description and,generated description quality and additionally provides grounding and local co-reference resolution.,24157
images axial microscopy in lateral versus resolution. show Fluorescence anisotropy severe usually,Fluorescence microscopy images usually show severe anisotropy in axial versus lateral resolution.,24158
quantitative automatic the extraction of data. biological,the automatic extraction of quantitative biological data.,24159
data. isotropic a We readily acquired method propose from to resolution anisotropic recover,We propose a method to recover isotropic resolution from readily acquired anisotropic data.,24160
jointly multi-scale their enhance Both and promotions. to mutual features are to learn multi-view employed,Both are jointly employed to learn multi-view and multi-scale features to enhance their mutual promotions.,24161
for differences subtle subcategories. the exploit to the and are employed distinguishing Both jointly local,Both are jointly employed to exploit the subtle and local differences for distinguishing the subcategories.,24162
"large-scale widely Recently, methods hashing been used in retrieval. have image","Recently, hashing methods have been widely used in large-scale image retrieval.",24163
progress. achieved research detection rapid Semantic object recently segmentation and have,Semantic segmentation and object detection research have recently achieved rapid progress.,24164
adapt object segments of detectors instead to approaches Most boxes. produce,Most approaches adapt object detectors to produce segments instead of boxes.,24165
"some cannot to pixel related work, belong Therefore, multiple a unlike instances.","Therefore, unlike some related work, a pixel cannot belong to multiple instances.",24166
for of movie recommendation. may modeling human The a possibility a paper taste suggest,The paper may suggest a possibility of modeling human taste for a movie recommendation.,24167
similarities or on relationships pairs between Deep samples. learning of is of based often triplets,Deep learning of similarities is often based on relationships between pairs or triplets of samples.,24168
a jointly. and in procedure are integrated learning similarity The and grouping single optimized model,The similarity learning and grouping procedure are integrated in a single model and optimized jointly.,24169
object estimation and on performance pose The classification. approach shows proposed competitive detailed unsupervised,The proposed unsupervised approach shows competitive performance on detailed pose estimation and object classification.,24170
"the results, affected of Despite by cost. the adversely better scalability is eye sensor movements","Despite better results, the scalability of eye movements is adversely affected by the sensor cost.",24171
"combination. their and from features visual assess eye We features, movements,","We assess features from eye movements, visual features, and their combination.",24172
coupled paper for a learning (CDL) This the approach proposes heterogeneous face deep matching.,This paper proposes a coupled deep learning (CDL) approach for the heterogeneous face matching.,24173
CDL The parts. includes of mainly function objective two,The objective function of CDL mainly includes two parts.,24174
"method CDL. Besides, is an to alternating of update parameters iteratively employed minimization the","Besides, an alternating minimization method is employed to iteratively update the parameters of CDL.",24175
Remote processing important sensing is geo-sciences. image in so,Remote sensing image processing is so important in geo-sciences.,24176
might are by types sensors initially of which obtained Images unrecognizable. be different,Images which are obtained by different types of sensors might initially be unrecognizable.,24177
of sensing types different types to processing are There remote images. the of according,There are different types of processing according to the types of remote sensing images.,24178
components second and Both third lower the the complexity. drastically the,Both the second and the third components drastically lower the complexity.,24179
real data. are for Qualitative microscopic results shown,Qualitative results are shown for real microscopic data.,24180
manipulating and interesting attributal important high-level problems. human using and Generating controls images facial are,Generating and manipulating human facial images using high-level attributal controls are important and interesting problems.,24181
"SL-GAN. an we iterative algorithm training present for Finally,","Finally, we present an iterative training algorithm for SL-GAN.",24182
effectiveness datasets framework. and the validate CelebA on of experiments CASIA-WebFace The our proposed recent,The experiments on recent CelebA and CASIA-WebFace datasets validate the effectiveness of our proposed framework.,24183
"available. make models and also will We data, code pre-trained","We will also make data, pre-trained models and code available.",24184
"vary. on semantics, hand, do not the Their other","Their semantics, on the other hand, do not vary.",24185
chose therefore mention not authors Some textbook it. to,Some textbook authors therefore chose not to mention it.,24186
show scheme Experiments several recognition on activity datasets to that leads state-of-the-art our performance.,Experiments on several activity recognition datasets show that our scheme leads to state-of-the-art performance.,24187
"to a address algorithm problem. simple this work, and this propose we In elegant","In this work, we propose a simple and elegant algorithm to address this problem.",24188
"become loss a deep for triplet Recently, person networks framework a learning common ReID. with","Recently, deep learning networks with a triplet loss become a common framework for person ReID.",24189
"the orders on obtaining training loss main on the attentions However, correct triplet pays set.","However, the triplet loss pays main attentions on obtaining correct orders on the training set.",24190
"we sensing of remote the super-resolution irregularly-sampled images. address Here,","Here, we address the super-resolution of irregularly-sampled remote sensing images.",24191
"examined Here, under ${\it we expression protein environmental various global Escherichia perturbations. in coli}$","Here, we examined global protein expression in ${\it Escherichia coli}$ under various environmental perturbations.",24192
evolution. is result of hypothesized that We such proportionality a,We hypothesized that such proportionality is a result of evolution.,24193
complex still Scene is environments. controlled to but attentional viewing used in selection study,Scene viewing is used to study attentional selection in complex but still controlled environments.,24194
"to we methods fixations. point the pairs spatial between relationship of investigate Here, apply process","Here, we apply spatial point process methods to investigate the relationship between pairs of fixations.",24195
"ordered of group flies In another of contrast, the patterns. highly exhibit dipteran retinas","In contrast, the retinas of another group of dipteran flies exhibit highly ordered patterns.",24196
"paper, description storytelling In an egocentric as this tackle problem. we sequences","In this paper, we tackle storytelling as an egocentric sequences description problem.",24197
"Furthermore, our description. outperforms methods proposal attentional video we classical that for prove encoder-decoder","Furthermore, we prove that our proposal outperforms classical attentional encoder-decoder methods for video description.",24198
"spaces the dual observations of are expressed, and Mathematically attributes. these","Mathematically expressed, these are the dual spaces of observations and attributes.",24199
Adaptation Vision. Computer Domain in actively problem an researched is,Domain Adaptation is an actively researched problem in Computer Vision.,24200
"where is the data undersampling. case aggressive we Cartesian particular, acquired address In using","In particular, we address the case where data is acquired using aggressive Cartesian undersampling.",24201
"to we selection. approach In propose view data-driven this set a paper,","In this paper, we propose a data-driven approach to view set selection.",24202
"algorithms. selection semantic compared to view alternative improve Furthermore, performance on views selected the segmentation","Furthermore, the selected views improve performance on semantic segmentation compared to alternative view selection algorithms.",24203
align reference. super-resolution   Previous approaches need to multiple the frames video to CNN-based,  Previous CNN-based video super-resolution approaches need to align multiple frames to the reference.,24204
propose in (SPMC) a CNN a `sub-pixel motion We compensation' accordingly framework. layer,We accordingly propose a `sub-pixel motion compensation' (SPMC) layer in a CNN framework.,24205
the video in of suitability experiments layer this and SR. Analysis show,Analysis and experiments show the suitability of this layer in video SR.,24206
"audits, planning investment, (energy applications Urban etc.)","Urban planning applications (energy audits, investment, etc.)",24207
for We researchers for make applications. use learning machine to our other available dataset remote-sensing,We make our dataset available for other machine learning researchers to use for remote-sensing applications.,24208
in crucial image role Relationships a among objects understanding. play,Relationships among objects play a crucial role in image understanding.,24209
"often classification problem, this considering of relationship treat as type a each methods (e.g. Previous","Previous methods often treat this as a classification problem, considering each type of relationship (e.g.",24210
"phrase each visual ""ride"") or distinct (e.g.","""ride"") or each distinct visual phrase (e.g.",24211
to propose this framework problem. an We tackle integrated,We propose an integrated framework to tackle this problem.,24212
"state-of-the-art. proposed improvement two On large achieves the substantial over method datasets,","On two large datasets, the proposed method achieves substantial improvement over state-of-the-art.",24213
"discrete, matrices thereby are Unfortunately, gradient-based methods. posing for difficulties permutation","Unfortunately, permutation matrices are discrete, thereby posing difficulties for gradient-based methods.",24214
with We networks learnable spatio-temporal two-stream feature do by aggregation. integrating state-of-the-art so,We do so by integrating state-of-the-art two-stream networks with learnable spatio-temporal feature aggregation.,24215
whole-video end-to-end classification. architecture for trainable is The resulting,The resulting architecture is end-to-end trainable for whole-video classification.,24216
"associates interpretable units CNN model, different the an explicit with AOG parts. object As different","As an interpretable model, the AOG associates different CNN units with different explicit object parts.",24217
during active-learning new encode to discovered We the process. AOG the grow incrementally knowledge,We incrementally grow the AOG to encode new knowledge discovered during the active-learning process.,24218
"method efficiency. experiments, In our learning high exhibits","In experiments, our method exhibits high learning efficiency.",24219
neurons network consisting noise-induced which small-world consider the Watts-Strogatz of We subthreshold spikings. exhibit,We consider the Watts-Strogatz small-world network consisting of subthreshold neurons which exhibit noise-induced spikings.,24220
spike-timing-dependent the governed strengths by has adaptive synaptic dynamic This plasticity (STDP). neuronal network,This neuronal network has adaptive dynamic synaptic strengths governed by the spike-timing-dependent plasticity (STDP).,24221
do learn to an and detector is occlusions we that deformations? How object invariant,How do we learn an object detector that is invariant to occlusions and deformations?,24222
these can hope examples learn use to The final the is classifier that invariances.,The hope is that the final classifier can use these examples to learn invariances.,24223
possible it But all really in the occlusions to see a is dataset?,But is it really possible to see all the occlusions in a dataset?,24224
"We object follow like occlusions and argue that also deformations a long-tail. categories,","We argue that like categories, occlusions and object deformations also follow a long-tail.",24225
"this propose we paper, solution. an In alternative","In this paper, we propose an alternative solution.",24226
examples deformations. occlusions We to an that propose generates adversarial learn network and with,We propose to learn an adversarial network that generates examples with occlusions and deformations.,24227
both are a manner. our framework in joint and detector In adversary original learned the,In our framework both the original detector and adversary are learned in a joint manner.,24228
paper. this release also for We code the,We also release the code for this paper.,24229
the characteristic to often-overlooked wander. its mind of An is propensity human,An often-overlooked characteristic of the human mind is its propensity to wander.,24230
a statistical segmentation. as serves for video clustering-based ADWIN the upper bound,ADWIN serves as a statistical upper bound for the clustering-based video segmentation.,24231
proposed length. generated The of used be assess sequences arbitrary can protocols to,The proposed protocols can be used to assess generated sequences of arbitrary length.,24232
an applications. segmentation is task for clinical many images Automatic medical important of,Automatic segmentation of medical images is an important task for many clinical applications.,24233
"modalities. wide different anatomical of using a structures are imaging visualised In practice, range","In practice, a wide range of anatomical structures are visualised using different imaging modalities.",24234
signal explore We visual context for spatial power learning of self-supervisory the a as representations.,We explore the power of spatial context as a self-supervisory signal for learning visual representations.,24235
"spatially such, contextual network the learns a conditioned representation. As","As such, the network learns a spatially conditioned contextual representation.",24236
on complexity motion features the represent of limited handcrafted are to approaches Traditional based patterns.,Traditional approaches based on handcrafted features are limited to represent the complexity of motion patterns.,24237
structures explore We for RNN temporal and stacked stream: RNN. two different the hierarchical,We explore two different structures for the temporal stream: stacked RNN and hierarchical RNN.,24238
RNN body human according is to kinematics. Hierarchical designed,Hierarchical RNN is designed according to human body kinematics.,24239
a single only. image requires The MR anatomical method,The method requires a single anatomical MR image only.,24240
explored transferring Going beyond we style. categorization,Going beyond categorization we explored transferring style.,24241
"stylistic elements. representative and illustrations we discovered Furthermore, discriminative","Furthermore, we discovered representative illustrations and discriminative stylistic elements.",24242
evaluation. the experimental We proposed show an in the extensive method superiority of,We show the superiority of the proposed method in an extensive experimental evaluation.,24243
"gradient for In enhancements outlier some framework, we in special filtering present this matching.","In this framework, we present some special enhancements for outlier filtering in gradient matching.",24244
detection performances already promising achieved benchmarks. text Previous approaches across have scene various for,Previous approaches for scene text detection have already achieved promising performances across various benchmarks.,24245
The fully to pose protocol cover space. natural capture aims the hand,The capture protocol aims to fully cover the natural hand pose space.,24246
study the estimation problem of for in the wild. images depth We single-image,We study the problem of single-image depth estimation for images in the wild.,24247
with training We two for loss annotations. normal propose novel functions surface,We propose two novel loss functions for training with surface normal annotations.,24248
works typically video focus Existing monocular deblurring.,Existing works typically focus monocular video deblurring.,24249
"deblurring we novel from paper, videos. approach stereo to this In propose a","In this paper, we propose a novel approach to deblurring from stereo videos.",24250
talk reality. mapping present invited for augmented recent projection technologies This will,This invited talk will present recent projection mapping technologies for augmented reality.,24251
"(AR) technologies using Second, introduced. projection reality are augmented mapping applications","Second, augmented reality (AR) applications using projection mapping technologies are introduced.",24252
first static presents the study human dynamics paper on forecasting images. from This,This paper presents the first study on forecasting human dynamics from static images.,24253
by front In propagation the collapse. a transient global phase observes extinction followed one,In the extinction phase one observes transient front propagation followed by a global collapse.,24254
interactively. classification plots This constructing explore paper forest and describes to data models structuring,This paper describes structuring data and constructing plots to explore forest classification models interactively.,24255
"ensemble, an produced is trees. bagging an multiple classifier A by forest example of","A forest classifier is an example of an ensemble, produced by bagging multiple trees.",24256
compression whenever is a artifacts algorithm in Compression arise applied. images lossy,Compression artifacts arise in images whenever a lossy compression algorithm is applied.,24257
"our networks. MSE obtains trained task, better SSIM In performance or method GAN this than","In this task, our GAN method obtains better performance than MSE or SSIM trained networks.",24258
and easily audiences rooms. between transport Artists different can themselves,Artists and audiences can easily transport themselves between different rooms.,24259
painting volumetric casting ray result. render Then we use the algorithms to GPU-based,Then we use GPU-based ray casting algorithms to render the volumetric painting result.,24260
upon priori segmentation heavily hand-crafted features user. depend and of Traditional the knowledge approaches a,Traditional segmentation approaches depend heavily upon hand-crafted features and a priori knowledge of the user.,24261
"to adopt As clinical a within environment. these difficult are such, methods","As such, these methods are difficult to adopt within a clinical environment.",24262
"among different to learn discriminative classes is features it more challenging Therefore, for FCNs.","Therefore, it is challenging to learn more discriminative features among different classes for FCNs.",24263
new answering baseline a task. paper question for visual This presents,This paper presents a new baseline for visual question answering task.,24264
Reproducing strictly the scoring We Hilbert Space. study rules Kernel in proper,We study strictly proper scoring rules in the Reproducing Kernel Hilbert Space.,24265
Scoring and a rule associated Kernel We general Kernel propose Divergence.,We propose a general Kernel Scoring rule and associated Kernel Divergence.,24266
We strictly which under is conditions proper. the Score Kernel consider,We consider conditions under which the Kernel Score is strictly proper.,24267
"analysis statistical for distributions unknown distributions. data with possible Pearson Thus, made","Thus, Pearson distributions made statistical analysis possible for data with unknown distributions.",24268
Image-language attracted matching in computer of tasks the have lot vision recently a attention field.,Image-language matching tasks have recently attracted a lot of attention in the computer vision field.,24269
these for paper similarity networks the between This two-branch data investigates learning two modalities. neural,This paper investigates two-branch neural networks for learning the similarity between these two data modalities.,24270
network propose different representations. that We structures output two produce,We propose two network structures that produce different output representations.,24271
a problems. useful very Such other segmentation breadth functions and vision discrete are of in,Such discrete functions are very useful in segmentation and a breadth of other vision problems.,24272
image widely a is Bag-of-Words The representing model used (BoW) part in descriptor one.,The Bag-of-Words (BoW) model is a widely used image representing descriptor in part one.,24273
visual features generated by Its is space. simply clustering in Euclidian codebook,Its codebook is simply generated by clustering visual features in Euclidian space.,24274
"clustered proposed In particular, under is distance Mahalanobis supervised. which the learned is codebook","In particular, the proposed codebook is clustered under Mahalanobis distance which is learned supervised.",24275
our prove is proposed that experiments effective. method Extensive,Extensive experiments prove that our proposed method is effective.,24276
the integration induces advantage learning. discriminative obtained denoiser an Such is via considerable when,Such an integration induces considerable advantage when the denoiser is obtained via discriminative learning.,24277
"fast lacking. prior of However, discriminative with study is denoiser integration the still","However, the study of integration with fast discriminative denoiser prior is still lacking.",24278
"of use methods large external propose i.e. unlabeled We existing (i) corpora which text,","We propose methods which use large existing external corpora of (i) unlabeled text, i.e.",24279
"visual based and object with tagged images question classes, to books, novel (ii) achieve answering.","books, and (ii) images tagged with classes, to achieve novel object based visual question answering.",24280
majority benchmarks crucial are for vision computer applications. of the Standardized,Standardized benchmarks are crucial for the majority of computer vision applications.,24281
(AU) for Action becomes detection essential Unit facial analysis.,Action Unit (AU) detection becomes essential for facial analysis.,24282
translation for Generative made Conditional Networks progress have cross-domain recently. image-to-image Adversarial (GANs) much,Conditional Generative Adversarial Networks (GANs) for cross-domain image-to-image translation have made much progress recently.,24283
"it huge Second, learning transfer document a from i.e. proposes images, set of","Second, it proposes transfer learning from a huge set of document images, i.e.",24284
"approach. learning tackle end-to-end problem, we attribute this To unsupervised propose an","To tackle this problem, we propose an end-to-end unsupervised attribute learning approach.",24285
has development Face learning. with the progress recognition great made deep of,Face recognition has made great progress with the development of deep learning.,24286
attention Mean and a is generalized with extension Discrepancy promising Maximum It weighting. memory of,It is a promising and generalized extension of Maximum Mean Discrepancy with memory attention weighting.,24287
"into Then CNNs, end-to-end an naturally learning ASML integrated scheme. is in resulting","Then ASML is naturally integrated into CNNs, resulting in an end-to-end learning scheme.",24288
this Collaborative In Low-Rank Subspace Clustering. present paper we,In this paper we present Collaborative Low-Rank Subspace Clustering.,24289
a representation phenomenon Given unified observations a multiple matrix. of learn we,Given multiple observations of a phenomenon we learn a unified representation matrix.,24290
the datasets. benchmark performance of is The evaluated on a proposed of algorithm SISR number,The performance of the proposed algorithm is evaluated on a number of SISR benchmark datasets.,24291
methods as observations combinations multivariate decompose linear factorisation latent Matrix feature vectors. of,Matrix factorisation methods decompose multivariate observations as linear combinations of latent feature vectors.,24292
has feature independent entries. with work Previous focussed vectors latent on,Previous work has focussed on latent feature vectors with independent entries.,24293
model to smoothness. such as We representing structures covariance latent extend characteristics the nondiagonal include,We extend the model to include nondiagonal latent covariance structures representing characteristics such as smoothness.,24294
a investigate we In simple this type. the article of Hodgkin-Huxley conductance-based two-component model,In this article we investigate a simple two-component conductance-based model of the Hodgkin-Huxley type.,24295
special of projections case The AFM. are a,The projections are a special case of AFM.,24296
An polynomials. achieved compared to trigonometric magnitude is speed-up order of traditional,An order of magnitude speed-up is achieved compared to traditional trigonometric polynomials.,24297
Variational segmentation. in Set (LS) method Level has been used a widely medical,Variational Level Set (LS) has been a widely used method in medical segmentation.,24298
"world. is However, the limited with real objects multi-instance it when dealing in","However, it is limited when dealing with multi-instance objects in the real world.",24299
"hide we any During to patches. not need do testing,","During testing, we do not need to hide any patches.",24300
localization. action to our weakly-supervised demonstrate be extended that We easily also can framework,We also demonstrate that our framework can be easily extended to weakly-supervised action localization.,24301
of We motion the address the of epipolar geometry problem using silhouettes.,We address the problem of epipolar geometry using the motion of silhouettes.,24302
four was providing validated on calibrations across method datasets different very Our standard viewpoints. accurate,Our method was validated on four standard datasets providing accurate calibrations across very different viewpoints.,24303
"we threshold optimal this an design framework. In paper, present","In this paper, we present an optimal threshold design framework.",24304
"set performance. maximally a oracle results achievable the theoretically to we derive inform of First,","First, we derive a set of oracle results to theoretically inform the maximally achievable performance.",24305
that should intensity. threshold pixel match the with show oracle exactly underlying We the,We show that the oracle threshold should match exactly with the underlying pixel intensity.,24306
"Experimentally, of rate methods. existing threshold than convergence achieves new design better method the","Experimentally, the new threshold design method achieves better rate of convergence than existing methods.",24307
experienced models carefully textures mathematical users. normally from generated are parameters selected by Procedural with,Procedural textures are normally generated from mathematical models with parameters carefully selected by experienced users.,24308
estimation We the consider intensity. task facial automated of of expression,We consider the task of automated estimation of facial expression intensity.,24309
arises intensity co-occurrence statistically from AU patterns of structure Their levels. induced,Their structure arises from statistically induced co-occurrence patterns of AU intensity levels.,24310
used to Resonance Magnetic and (CMR) cardiac Cardiac function. commonly imaging is structure assess,Cardiac Magnetic Resonance (CMR) imaging is commonly used to assess cardiac structure and function.,24311
exams CMR One of post-processing is tedious. that disadvantage is of,One disadvantage of CMR is that post-processing of exams is tedious.,24312
evaluated the HVSMR challenge. was Performance within,Performance was evaluated within the HVSMR challenge.,24313
"I build I cannot understand do 'If it.' it, not","'If I cannot build it, I do not understand it.'",24314
rare. that Achieving is of level understanding,Achieving that level of understanding is rare.,24315
understand to phenomenon a simulate understand is it sufficiently fully to To computationally. it,To understand a phenomenon sufficiently to fully simulate it is to understand it computationally.,24316
se. to computers does 'Computation' not refer per,'Computation' does not refer to computers per se.,24317
"it that them the and work. underlying methods make principles to refers Rather,","Rather, it refers to the underlying principles and methods that make them work.",24318
"fabric. thin look to tends flexible A draped, heavy stiff from different when fabric, a","A thin flexible fabric, when draped, tends to look different from a heavy stiff fabric.",24319
"the we (CNN) network Then, with popular card models, convolutional images. train two neural","Then, we train two popular convolutional neural network (CNN) models, with the card images.",24320
fully Collecting expensive. datasets annotated and challenging image is,Collecting fully annotated image datasets is challenging and expensive.,24321
are mode: questions visual on about one focus that We unexplored images. particular asked,We focus on one particular unexplored mode: visual questions that are asked about images.,24322
"instance, of dog?"" question the is ""what For breed the the","For instance, the question ""what is the breed of the dog?""",24323
single-image super-resolution. networks reconstruction recently for demonstrated Convolutional neural high-quality have,Convolutional neural networks have recently demonstrated high-quality reconstruction for single-image super-resolution.,24324
"techniques automation, reproducibility. available full currently limiting However, lack","However, currently available techniques lack full automation, limiting reproducibility.",24325
"challenging scene datasets, The conducted labeling i.e. on experiments four are","The experiments are conducted on four challenging scene labeling datasets, i.e.",24326
"CamVid, scene various and labeling datasets, background against Stanford and methods. SUN SiftFlow, compared state-of-the-art","SiftFlow, CamVid, Stanford background and SUN datasets, and compared against various state-of-the-art scene labeling methods.",24327
"approach. much general simpler the Furthermore, more than new previous method is and our","Furthermore, our new method is much simpler and more general than the previous approach.",24328
"and matte. lighting, and normals), an albedo, alpha","normals), albedo, and lighting, and an alpha matte.",24329
problem the to before And are status raise reservable filtering. they intriguingly the if we,And we raise the intriguingly problem if they are reservable to the status before filtering.,24330
proposed. algorithm is very zero-order effective yet A simple,A very simple yet effective zero-order algorithm is proposed.,24331
computational filters most reverse to practically with It able cost. is low,It is able to practically reverse most filters with low computational cost.,24332
applications. can method be generalized other solve to new This and inverse also problems enables,This method can also be generalized to solve other inverse problems and enables new applications.,24333
the method object detects basic We that persons. a detector supplement using,We supplement the method using a basic object detector that detects persons.,24334
these pose. of just obtain With components the a rough we estimate human,With just these components we obtain a rough estimate of the human pose.,24335
the primary were DT-MRI heart angles from of The fibre obtained the eigenvalues. orientation,The fibre orientation angles of the heart were obtained from the DT-MRI primary eigenvalues.,24336
"Irritation it cooperation. in partners a breakdown apparently and when arises precipitate causes defect,","Irritation arises when partners apparently defect, and it causes a precipitate breakdown in cooperation.",24337
one's inefficiency. it propensity partner's to economic for Failing leads model to substantial,Failing to model one's partner's propensity for it leads to substantial economic inefficiency.,24338
"zero entropy is Information but summarize used count contained transcriptome to data, ignoring them. data","Information entropy is used to summarize transcriptome data, but ignoring zero count data contained them.",24339
"single only as guess"" a not a Algorithms sufficient. are that yield ""best result","Algorithms that only yield a single ""best guess"" as a result are not sufficient.",24340
challenging. extremely is instance-level and context object-object Modeling relationships,Modeling instance-level context and object-object relationships is extremely challenging.,24341
"locations bounding of different requires \etc. about classes, reasoning boxes It","It requires reasoning about bounding boxes of different classes, locations \etc.",24342
"reasoning detections. requires on conditional spatial previous all, inherently distributions instance-level Above modeling","Above all, instance-level spatial reasoning inherently requires modeling conditional distributions on previous detections.",24343
still detect parallel non-maximal by all object The followed state-of-the-art detectors (NMS). suppression in object,The state-of-the-art object detectors still detect all object in parallel followed by non-maximal suppression (NMS).,24344
our show two accuracy on results Evaluation of the benchmark approach. robustness and datasets,Evaluation results on two benchmark datasets show the robustness and accuracy of our approach.,24345
amount handle of limited Most noise. methods existing,Most existing methods handle limited amount of noise.,24346
convergent. We always formally technique is prove our also that,We also formally prove that our technique is always convergent.,24347
the outperform significantly our They that existing demonstrate techniques. techniques,They demonstrate that our techniques significantly outperform the existing techniques.,24348
This but for datasets. SSCs burden has all use prohibited smallest the,This burden has prohibited SSCs use for all but the smallest datasets.,24349
provide k-SSC. success We for of analysis bounds for theoretical the,We provide theoretical analysis for the bounds of success for k-SSC.,24350
"obtain adopted. approximate the CRF the strategy solution N-best is model, the proposed To of","To obtain the approximate solution of the proposed CRF model, the N-best strategy is adopted.",24351
witnessed has neural saliency rapid Image progress convolutional detection deep due recently networks. to,Image saliency detection has recently witnessed rapid progress due to deep convolutional neural networks.,24352
Image based The and learning system is on domains. processing machine,The system is based on Image processing and machine learning domains.,24353
comparison other with EUSBoost techniques paper. of Also the shown is in,Also comparison of EUSBoost with other techniques is shown in the paper.,24354
"nuclear we for address constrained a norm issue, problems. To propose rank-drop method this","To address this issue, we propose a rank-drop method for nuclear norm constrained problems.",24355
analysts. Data is that more data ever become Big an by setting has commonplace encountered,Big Data has become an ever more commonplace setting that is encountered by data analysts.,24356
"techniques cannot useful of are discarded. Unfortunately, many and traditional be these","Unfortunately, many of these traditional techniques are useful and cannot be discarded.",24357
technique test (KS) for goodness-of-fit the such Kolmogorov-Smirnov One is (GoF).,One such technique is the Kolmogorov-Smirnov (KS) test for goodness-of-fit (GoF).,24358
via Data (CA) derived paradigm. and A stream-appropriate Big is the KS-type test chunked-and-averaged estimator,A Big Data and stream-appropriate KS-type test is derived via the chunked-and-averaged (CA) estimator paradigm.,24359
test. GoF termed The the CAKS test is new,The new test is termed the CAKS GoF test.,24360
"in neural model, single being achieves computationally while these efficient. end-to-end network a DESIRE trainable","DESIRE achieves these in a single end-to-end trainable neural network model, while being computationally efficient.",24361
and further accuracy. mechanism the A prediction feedback iterates boost over ranking the refinement to,A feedback mechanism iterates over the ranking and refinement to further boost the prediction accuracy.,24362
on KITTI available Drone and our model Dataset. two datasets: Stanford evaluate We publicly,We evaluate our model on two publicly available datasets: KITTI and Stanford Drone Dataset.,24363
We ignore magnify to linear and motion acceleration. propose,We ignore linear motion and propose to magnify acceleration.,24364
link filtering acceleration temporal magnification. spatial to We second-order derivative,We link temporal second-order derivative filtering to spatial acceleration magnification.,24365
show and where to apply motion We magnification. we method moving objects color our magnification,We apply our method to moving objects where we show motion magnification and color magnification.,24366
poor the to rainfall. visibility intense accumulation Rain remote to due of effect objects refers,Rain accumulation effect refers to poor visibility of remote objects due to the intense rainfall.,24367
the only. also It BCC on layer enforces the background,It also enforces the BCC on the background layer only.,24368
"specifically this To dealing the knowledge, rain. our optical with flow first method is","To our knowledge, this is the first optical flow method specifically dealing with rain.",24369
Intelligence. Vision intense has as study language Artificial emerged and understanding a undergoing in subject,Vision and language understanding has emerged as a subject undergoing intense study in Artificial Intelligence.,24370
optimization distributed problem multi-agent consider edges. in networks the We accelerating by sequentially of adding,We consider the problem of accelerating distributed optimization in multi-agent networks by sequentially adding edges.,24371
"this paper, by selection designing In edge we topologies and via begin network scheduling.","In this paper, we begin by designing network topologies via edge selection and scheduling.",24372
dynamics of edge evolution then The network scheduling. is incurred by,The dynamics of network evolution is then incurred by edge scheduling.,24373
the address problem videos. of temporal in We action localization,We address the problem of temporal action localization in videos.,24374
Experimental effectively. can show that prediction the agent work FML-based results,Experimental results show that the FML-based prediction agent can work effectively.,24375
"obtained text recognized are a system. words reading the scene by First,","First, the recognized words are obtained by a scene text reading system.",24376
fitting We for method a a against points. set describe data of ellipsoid minimum generalised,We describe a generalised method for ellipsoid fitting against a minimum set of data points.,24377
paper technological of scientific systems discusses The integrated expert development. and dynamic problems,The paper discusses scientific and technological problems of dynamic integrated expert systems development.,24378
dynamic systems for considered. development of problem-oriented methodology are integrated expert Extensions,Extensions of problem-oriented methodology for dynamic integrated expert systems development are considered.,24379
and paid knowledge processing. Attention the temporal is to representation,Attention is paid to the temporal knowledge representation and processing.,24380
Number identified of affects the databases reference identification. sequences precision of the in,Number of identified sequences in the reference databases affects the precision of identification.,24381
"barcodes nematode approaches. phylogeny-based marine This alignment-based, tree-based using paper and of compares identification the","This paper compares the identification of marine nematode barcodes using alignment-based, tree-based and phylogeny-based approaches.",24382
"retrieve fraction of as intractable Moreover, is backbone it to well. fixed the a","Moreover, it is intractable to retrieve a fixed fraction of the backbone as well.",24383
gradient one learning. algorithms in Boosting method descent machine is popular as,Boosting as gradient descent algorithms is one popular method in machine learning.,24384
"its generalized both LBI simplicity, theory the in popular Lasso Split Despite and outperforms experiments.","Despite its simplicity, Split LBI outperforms the popular generalized Lasso in both theory and experiments.",24385
"arrangement mainly within cytoplasm). nuclei and orientation shape, cell spatial the","mainly cell nuclei shape, orientation and spatial arrangement within the cytoplasm).",24386
network is An important transcription (GRN). direction infer to this the step regulatory in factor-gene,An important step in this direction is to infer the transcription factor-gene regulatory network (GRN).,24387
consistency by higher NetREX networks predicted alternative biological The approaches. than showed,The networks predicted by NetREX showed higher biological consistency than alternative approaches.,24388
problems. these class This and formulating introduces probabilistic for programs solving paper a of,This paper introduces a class of probabilistic programs for formulating and solving these problems.,24389
"the ability and samples. its Consequently, informative more improves detection progressively localizes positive detector","Consequently, the detector progressively improves its detection ability and localizes more informative positive samples.",24390
We software. data detection of anomaly benchmark from a using algorithms monitoring propose Cloud,We propose a benchmark of anomaly detection algorithms using data from Cloud monitoring software.,24391
this deny et would of Wang results like we In al. to the paper,In this paper we would like to deny the results of Wang et al.,24392
part the an detection of object integral pipeline. suppression is Non-maximum,Non-maximum suppression is an integral part of the object detection pipeline.,24393
"detection boxes scores. all First, it of sorts basis their the on","First, it sorts all detection boxes on the basis of their scores.",24394
process boxes. remaining on is This applied the recursively,This process is recursively applied on the remaining boxes.,24395
MobileNets We mobile a models called vision of embedded for efficient present and applications. class,We present a class of efficient models called MobileNets for mobile and embedded vision applications.,24396
introduce global off two between simple trade latency hyper-parameters accuracy. and efficiently We that,We introduce two simple global hyper-parameters that efficiently trade off between latency and accuracy.,24397
"bounding boxes very drawing consuming. time However, manually is","However, manually drawing bounding boxes is very time consuming.",24398
future approach a for of We frames. making long-term propose predictions hierarchical,We propose a hierarchical approach for making long-term predictions of future frames.,24399
image Localization for is prerequisite structures a medical in analysis. of anatomical many tasks,Localization of anatomical structures is a prerequisite for many tasks in medical image analysis.,24400
"different sizes, pyramid spatial pooling slices analyze of To applied. to the is ConvNet allow","To allow the ConvNet to analyze slices of different sizes, spatial pyramid pooling is applied.",24401
localization defined (e.g. with structures results The in of clearly best were boundaries achieved,The best results were achieved in localization of structures with clearly defined boundaries (e.g.,24402
boundary aortic clearly the structure the and when not arch) worst visible was (e.g.,aortic arch) and the worst when the structure boundary was not clearly visible (e.g.,24403
The method in structures. was and multiple localization accurate more robust,The method was more robust and accurate in localization multiple structures.,24404
doors open. expected until all to The time algorithm aims the minimize,The algorithm aims to minimize the expected time until all doors open.,24405
optimal are We that identify algorithms to up universal also constant a factor.,We also identify algorithms that are optimal up to a universal constant factor.,24406
"we study cascading doors. case memoryless two Here, the of","Here, we study the case of two cascading memoryless doors.",24407
classifications. of Most approach implements convolutional neural the for (CNNs) bottom-up traditional (feed-forward) image networks,Most of the traditional convolutional neural networks (CNNs) implements bottom-up approach (feed-forward) for image classifications.,24408
problem. paper This image action based approach presents recognition an for video classification skeleton-based,This paper presents an image classification based approach for skeleton-based video action recognition problem.,24409
the popular benchmard achieve state-of-the-art on We e.g. results the datasets,We achieve the state-of-the-art results on the popular benchmard datasets e.g.,24410
object instances. approach propose for of semi-automatic an annotation We,We propose an approach for semi-automatic annotation of object instances.,24411
unseen capabilities to of further generalization approach datasets. our show We,We further show generalization capabilities of our approach to unseen datasets.,24412
"such as fast interaction, and Many human-computer tasks, classification. require real-time gender efficient facial","Many real-time tasks, such as human-computer interaction, require fast and efficient facial gender classification.",24413
"a obtained FSITM higher by In TMQI performance the and combining indices. addition, is","In addition, a higher performance is obtained by combining the FSITM and TMQI indices.",24414
inspire and We our in benefit area. believe this important idea research would future,We believe our idea would inspire and benefit future research in this important area.,24415
query our bounds and optimal. Our that total is complexity label almost show lower,Our lower bounds show that our label and total query complexity is almost optimal.,24416
image computer applications. step vision in an important of is binary many Matching different features,Matching of binary image features is an important step in many different computer vision applications.,24417
was using database ground well-known The known truth. a tested with image method,The method was tested using a well-known image database with known ground truth.,24418
relations the human of life. Social foundation daily are,Social relations are the foundation of human daily life.,24419
tracking. multi-view proposed detection pedestrian phases: two consists system The of and,The proposed system consists of two phases: multi-view pedestrian detection and tracking.,24420
"blob. to subtraction detection utilizes background First, segment foreground the pedestrian","First, pedestrian detection utilizes background subtraction to segment the foreground blob.",24421
"performed Second, individual two tracking is at pedestrian subject. the the Tracking classification phases: and","Second, the Tracking is performed at two phases: pedestrian classification and tracking the individual subject.",24422
features. for used The classification neural networks PHOG with is,The neural networks is used for classification with PHOG features.,24423
the focus more now second-stage on FCN segmentation This of can organs. detailed,This second-stage FCN can now focus on more detailed segmentation of the organs.,24424
"for the tracking performance and However, classification measures of are different.","However, the measures of performance for tracking and classification are different.",24425
systems not classification. tracking are evaluating Data be may sets suitable for that appropriate for,Data sets that are suitable for evaluating tracking systems may not be appropriate for classification.,24426
based EEG state numerous brain decoding has applications.,EEG based brain state decoding has numerous applications.,24427
explore work generality al. of we the Edelman this In et,In this work we explore the generality of Edelman et al.,24428
face by decoding hypothesis considering recognition. of,hypothesis by considering decoding of face recognition.,24429
"decoding. the Hence, evidence imaging our supporting that presents hypothesis work source improves for","Hence, our work presents supporting evidence for the hypothesis that source imaging improves decoding.",24430
a filtering using layer. We directional learnable CNN architecture first novel a conduct,We first conduct a CNN architecture using a novel learnable directional filtering layer.,24431
"for novel paper, image we a propose descriptors. this evaluating benchmark In local","In this paper, we propose a novel benchmark for evaluating local image descriptors.",24432
"This allows for thus application more comparisons more different and scenarios. realistic, reliable in","This allows for more realistic, and thus more reliable comparisons in different application scenarios.",24433
their evaluate and analyse state-of-the-art descriptors of performance properties. We several the,We evaluate the performance of several state-of-the-art descriptors and analyse their properties.,24434
the addresses objects segmenting task of in videos. paper moving unconstrained This,This paper addresses the task of segmenting moving objects in unconstrained videos.,24435
module explicit network We neural to this. achieve a memory introduce with two-stream an novel,We introduce a novel two-stream neural network with an explicit memory module to achieve this.,24436
illuminants spectra. multiple Real-world often with of lighting different consists,Real-world lighting often consists of multiple illuminants with different spectra.,24437
"is availability training main the the bottleneck limited of However, for re-ID labeled fully-supervised samples.","However, the main bottleneck for fully-supervised re-ID is the limited availability of labeled training samples.",24438
"learning performance. we relationships improves further pairwise which addition, the iteratively, refine pseudo In the","In addition, we refine the pseudo pairwise relationships iteratively, which further improves the learning performance.",24439
from webpages many Extracting in domains. a geographical is application well-motivated tags,Extracting geographical tags from webpages is a well-motivated application in many domains.,24440
important challenging actions in yet Detecting an untrimmed is videos task.,Detecting actions in untrimmed videos is an important yet challenging task.,24441
a registration we In (DIRNet). image work deformable for network propose this learning deep,In this work we propose a deep learning network for deformable image registration (DIRNet).,24442
"exploit do information, techniques spatial not these i.e. All","All these techniques do not exploit spatial information, i.e.",24443
the the consider features all location image. in corresponding their and descriptors the without embedding,consider all the features and the corresponding descriptors without embedding their location in the image.,24444
the simple This existing approach. greater an than accuracy state-of-the-art variant shows,This simple variant shows an accuracy greater than the existing state-of-the-art approach.,24445
balanced a version proposed. Morever of is ZuBuD more,Morever a more balanced version of ZuBuD is proposed.,24446
shaped of a tree. proposed like The consists model deep a CNN,The proposed model consists of a deep CNN shaped like a tree.,24447
layers the classes. includes of common a convolutional of tree stem all to The sequence,The stem of the tree includes a sequence of convolutional layers common to all classes.,24448
class of These to learned individual the subset for use. each determine features gates,These learned gates determine for each individual class the subset of features to use.,24449
a view reformulating by consistency ray differentiable using We term. (DRC) do so consistency,We do so by reformulating view consistency using a differentiable ray consistency (DRC) term.,24450
"semantics etc. foreground depth, color masks, images,","foreground masks, depth, color images, semantics etc.",24451
of empirical technique setting. our controlled present We analysis a in,We present empirical analysis of our technique in a controlled setting.,24452
"face deep we effective an completion In using a this generative algorithm model. paper, propose","In this paper, we propose an effective face completion algorithm using a deep generative model.",24453
What inside raga. is and particular when composing performers a performing brain the he happens,What happens inside the performers brain when he is performing and composing a particular raga.,24454
perplexed long are time. for questions the neuroscientists a that These,These are the questions that perplexed neuroscientists for a long time.,24455
investigating Symmetry sides inside an image plane. similar is an by composition feature important,Symmetry is an important composition feature by investigating similar sides inside an image plane.,24456
crucial the or man-made nature It universe. effect recognize a within to objects has,It has a crucial effect to recognize man-made or nature objects within the universe.,24457
"A despeckling is filters. quantitative and then defined, state-of-the-art index new to image-quality applied","A new quantitative image-quality index is then defined, and applied to state-of-the-art despeckling filters.",24458
multipliers used of the method Alternating direction solve to model. is,Alternating direction method of multipliers is used to solve the model.,24459
and It also finer provides much details. discernable,It also provides much finer and discernable details.,24460
how extent detection We on the performance transfer. depends the study of,We study how the detection performance depends on the extent of transfer.,24461
how intermediate images network. investigate response of from the We our also evolution,We also investigate how the evolution of intermediate response images from our network.,24462
These methods. were previous single the results stage by reached not,These results were not reached by the previous single stage methods.,24463
ROS slow down view popular. getting that The aging is,The view that ROS slow down aging is getting popular.,24464
"critical a performance. verification typical step for feature a is normalization method, boosting face In","In a typical face verification method, feature normalization is a critical step for boosting performance.",24465
introduce of training. the and study This us during effect to motivates normalization,This motivates us to introduce and study the effect of normalization during training.,24466
"despite is But normalization this non-trivial, find we being differentiable.","But we find this is non-trivial, despite normalization being differentiable.",24467
using Based strategies propose analysis on normalized this features. we training for two,Based on this analysis we propose two strategies for training using normalized features.,24468
"is cosine loss, instead optimizes softmax modification of similarity of inner-product. first which The a","The first is a modification of softmax loss, which optimizes cosine similarity instead of inner-product.",24469
"answer to and of balls number there boxes?"" For example, ""is an equal","For example, to answer ""is there an equal number of balls and boxes?""",24470
"them, results. we boxes, for balls, look and can the look compare for count","we can look for balls, look for boxes, count them, and compare the results.",24471
enables be learning flexible the The model to for custom proposed framework network a design.,The proposed learning framework enables the network model to be flexible for a custom design.,24472
"for dependency on the classification. the trained network it Moreover, alleviates","Moreover, it alleviates the dependency on the network trained for classification.",24473
"we trained learning distinct rates. network text During classification phase, with","During text classification phase, we trained network with distinct learning rates.",24474
on reported encouraging characters recognition Arabic approach images. segmented Arabic scene from of Our results,Our approach reported encouraging results on recognition of Arabic characters from segmented Arabic scene images.,24475
usefulness problems. in on implement SCP We prototype a its system and a demonstrate CP,We implement a prototype in a CP system and demonstrate its usefulness on SCP problems.,24476
"Also, included. has scenarios case use been study a land for","Also, a case study for land use scenarios has been included.",24477
"interest descriptions in generating automatically Recently, image. for of an been has a lot there","Recently, there has been a lot of interest in automatically generating descriptions for an image.",24478
preferences. that image generation enables user This better description accommodates,This enables image description generation that better accommodates user preferences.,24479
"Location instance as is ""street view"" on visual commonly imagery. retrieval treated recognition","Location recognition is commonly treated as visual instance retrieval on ""street view"" imagery.",24480
"The dataset queries i.e. views, panoramic and are items","The dataset items and queries are panoramic views, i.e.",24481
groups of location. a single images at taken,groups of images taken at a single location.,24482
"case, views used multiple queries. are either as In","In either case, multiple views are used as queries.",24483
with reach a standard benchmark perfect views. recognition location query on We only four near,We reach near perfect location recognition on a standard benchmark with only four query views.,24484
"has multi-layer a graph reason, proposed that structure-based been algorithm For recently.","For that reason, a multi-layer graph structure-based algorithm has been proposed recently.",24485
attributes separating can avoid the into layers. by problem multiple It effectively,It can effectively avoid the problem by separating attributes into multiple layers.,24486
"relaxation multi-layer for solve procedure the using convex-concave problem structure. the we Then, a","Then, we solve the problem using a convex-concave relaxation procedure for the multi-layer structure.",24487
The proposed on algorithm based algorithms state-of-the-art single-layer the better than exhibits performance structure.,The proposed algorithm exhibits better performance than state-of-the-art algorithms based on the single-layer structure.,24488
observations These are influential in helpful data. the to promote insights about,These are helpful to promote insights about influential observations in the data.,24489
proposed framework network general account. environments multi-camera The takes into,The proposed framework takes general multi-camera network environments into account.,24490
processes estimators. Gaussian (GPs) function powerful non-parametric are,Gaussian processes (GPs) are powerful non-parametric function estimators.,24491
"procedures. inference expensive the computational largely limited by cost of applications their the are However,","However, their applications are largely limited by the expensive computational cost of the inference procedures.",24492
faces the CNN. the Experiments performance showed face boosted with recognition the training of augmented,Experiments showed training with the augmented faces boosted the face recognition performance of the CNN.,24493
"axis-aligned shape. accurately cannot Nevertheless, boxes an capture oriented or object's","Nevertheless, axis-aligned or oriented boxes cannot accurately capture an object's shape.",24494
"gap, the measure. To Union this (rIoU) over close accuracy introduce Intersection relative we","To close this gap, we introduce the relative Intersection over Union (rIoU) accuracy measure.",24495
(max) used. the zero-th or are (average) Typically first-order statistics,Typically zero-th (max) or the first-order (average) statistics are used.,24496
"of this paper, In second-order the explore benefits we statistics. using","In this paper, we explore the benefits of using second-order statistics.",24497
We challenge this deep network developing neural convolutional a method. (DCNN) by participate in,We participate in this challenge by developing a deep convolutional neural network (DCNN) method.,24498
"hand system based In we gesture. design project, interaction human-computer this real-time a on","In this project, we design a real-time human-computer interaction system based on hand gesture.",24499
stacking generate attention-aware Residual by is which features. Modules built Network Our Attention Attention,Our Residual Attention Network is built by stacking Attention Modules which generate attention-aware features.,24500
different going features attention-aware The change from layers deeper. adaptively modules as,The attention-aware features from different modules change adaptively as layers going deeper.,24501
is that demonstrates experiment network noisy labels. against also The our robust,The experiment also demonstrates that our network is robust against noisy labels.,24502
is industry. different sector Healthcare from totally other,Healthcare sector is totally different from other industry.,24503
social though not budget. percentage consume it achieve It expectation did huge even of,It did not achieve social expectation even though it consume huge percentage of budget.,24504
of medical being data done Mostly expert. the interpretations by medical is,Mostly the interpretations of medical data is being done by medical expert.,24505
We observed an interacting behavior study coli E. in anomalous populations.,We study an anomalous behavior observed in interacting E. coli populations.,24506
reproduce Simulations results. experimental the of model Keller-Segel the chemotaxis,Simulations of the Keller-Segel chemotaxis model reproduce the experimental results.,24507
data show prediction experiments with and Dublin Beijing. from accurate The,The experiments show accurate prediction with data from Dublin and Beijing.,24508
compressive sparse shown in great potential Patch-based representation reconstruction. sensing modeling image (CS) has,Patch-based sparse representation modeling has shown great potential in image compressive sensing (CS) reconstruction.,24509
reconstruction. Experimental state-of-the-art demonstrate outperforms the image results for many method CS proposed techniques that,Experimental results demonstrate that the proposed method outperforms many state-of-the-art techniques for image CS reconstruction.,24510
heritability Our a missing why among varies provide simple for results interpretation it traits. and,Our results provide a simple interpretation for missing heritability and why it varies among traits.,24511
is human-centric that We by model a driven approach. propose a novel,We propose a novel model that is driven by a human-centric approach.,24512
relations. geographical and non-geographical topological Few using data manage algorithms,Few algorithms manage geographical and non-geographical data using topological relations.,24513
convolutional for facial developed networks have task. expression a neural We (CNN) recognition,We have developed convolutional neural networks (CNN) for a facial expression recognition task.,24514
gray-scale with images. CNN models trained using depth We different,We trained CNN models with different depth using gray-scale images.,24515
showed real results data. with experiments world The competitive,The experiments showed competitive results with real world data.,24516
"challenging. precise some for OAR remains delineation However,","However, for some OAR precise delineation remains challenging.",24517
a The among poor with and shape structures. these is esophagus contrast versatile,The esophagus with a versatile shape and poor contrast is among these structures.,24518
shape assessed were overlap spatial similarities. Results and regarding,Results were assessed regarding spatial overlap and shape similarities.,24519
Rolling capability. become of imaging cameras Shutter because popularized low-cost have (RS),Rolling Shutter (RS) cameras have become popularized because of low-cost imaging capability.,24520
"produces cameras ego-motion with estimates. (MVO) Monocular For inaccurate RS Visual that reason, Odometry","For that reason, Monocular Visual Odometry (MVO) with RS cameras produces inaccurate ego-motion estimates.",24521
solve images works from RS this sensors. problem with inertial motion prediction distortion Previous and/or,Previous works solve this RS distortion problem with motion prediction from images and/or inertial sensors.,24522
fast vibration of motion causes extremely cameras instantaneously). mobile,vibration of mobile cameras causes extremely fast motion instantaneously).,24523
This problem the denoising addresses in article strong of image situations the noise.,This article addresses the image denoising problem in the situations of strong noise.,24524
decomposition We a dual propose sparse method.,We propose a dual sparse decomposition method.,24525
sub-dictionary decomposition. decomposition a the the on over-complete dictionary sparse method in makes This,This method makes a sub-dictionary decomposition on the over-complete dictionary in the sparse decomposition.,24526
exploiting We for manifold an algorithm outlier-robust IRLS this structure. pose provide generic filtering,We provide an outlier-robust IRLS algorithm for generic pose filtering exploiting this manifold structure.,24527
the pools the under the convolutional body on It layers guidance of joint positions. feature,It pools the feature on the convolutional layers under the guidance of body joint positions.,24528
of schemes discussed. feature convolutional body Two pooling into maps joints are for mapping,Two schemes of mapping body joints into convolutional feature maps for pooling are discussed.,24529
joint can be off-the-shelf obtained body from any estimation positions algorithm. skeleton The,The body joint positions can be obtained from any off-the-shelf skeleton estimation algorithm.,24530
based (PSO). species Swarm the Particle We Optimization track the recognized objects using,We track the recognized objects using the species based Particle Swarm Optimization (PSO).,24531
"the of multiple of interest images. object training we from alignment First, our extract","First, we extract the multiple alignment of our object of interest from training images.",24532
"based competition with species Gaussian we Subsequently, the repulsion annealed PSO. analyze among and","Subsequently, we analyze the competition and repulsion among species with annealed Gaussian based PSO.",24533
"this a function in design In dictionary framework learning. we paper, cost with Bayesian","In this paper, we design a cost function in Bayesian framework with dictionary learning.",24534
"of properties methods However, theoretical largely these the unexplored. remain","However, the theoretical properties of these methods remain largely unexplored.",24535
"images before propose object generating proposals. object estimate of scales to Therefore, we","Therefore, we propose to estimate object scales of images before generating object proposals.",24536
proposed ScaleNet. called scales is object The predicting method for,The proposed method for predicting object scales is called ScaleNet.,24537
large by margin. The the the datasets state-of-the-art method supermarket a on outperforms previous resulted,The resulted method outperforms the previous state-of-the-art on the supermarket datasets by a large margin.,24538
present operates optical on flow estimation We four-dimensional an approach cost full the volume. that,We present an optical flow estimation approach that operates on the full four-dimensional cost volume.,24539
further setting. to exploit then We regularity semi-global four-dimensional by this the adapting matching,We then exploit this regularity further by adapting semi-global matching to the four-dimensional setting.,24540
"To we this contributions. make following the end,","To this end, we make the following contributions.",24541
use face segmentations unprecedented our enable under swapping We to (b) conditions. robust,(b) We use our segmentations to enable robust face swapping under unprecedented conditions.,24542
"work, allow our robust extensive previous to is quantitative tests. Unlike swapping (c) enough for","(c) Unlike previous work, our swapping is robust enough to allow for extensive quantitative tests.",24543
This is vision systems. effect was demonstrated quantitatively machine time this first the for,This is the first time this effect was quantitatively demonstrated for machine vision systems.,24544
added improve connections Residual were to training.,Residual connections were added to improve training.,24545
"algorithms However, can achieve this effectively criterion. few existing","However, few existing algorithms can effectively achieve this criterion.",24546
"of angular be by $m$. Moreover, the size parameter can margin quantitatively adjusted a","Moreover, the size of angular margin can be quantitatively adjusted by a parameter $m$.",24547
criterion. feature ideal $m$ to derive further specific the approximate We,We further derive specific $m$ to approximate the ideal feature criterion.,24548
made The available. also been code has publicly,The code has also been made publicly available.,24549
plays imaging blood fundus role diagnosis. in in medical Segmenting vessels an important,Segmenting blood vessels in fundus imaging plays an important role in medical diagnosis.,24550
take supervised advantages This of us approach labeling. to allows learning without,This approach allows us to take advantages of supervised learning without labeling.,24551
test a model to it. build also naive We DCNN,We also build a naive DCNN model to test it.,24552
bronchial Some airway diseases lung and related to structures are morphology.,Some lung diseases are related to bronchial airway structures and morphology.,24553
to segment the the in each branch. uses VOI It CEF,It uses the CEF in the VOI to segment each branch.,24554
"unify we tree. of airway an bronchial regions the integrated to form extracted all Finally,","Finally, we unify all of the extracted bronchial regions to form an integrated airway tree.",24555
based state-of-the-art recognition approaches action mostly skeleton-based (RNN). to neural recurrent networks on Current are,Current state-of-the-art approaches to skeleton-based action recognition are mostly based on recurrent neural networks (RNN).,24556
designed automatically. joints novel transformer rearrange A and skeleton important to module is skeleton select,A novel skeleton transformer module is designed to rearrange and select important skeleton joints automatically.,24557
"available A a data are first open and source implementation set, our fully results online.","A first data set, our results and a fully open source implementation are available online.",24558
a Information for does hypothesis foundation uncertainty better testing provides than quantification. accounting,Information accounting provides a better foundation for hypothesis testing than does uncertainty quantification.,24559
viewing scores in a subjective experiment using interactive setup. acquired an were The light-field,The scores were acquired in a subjective experiment using an interactive light-field viewing setup.,24560
Classifiers tested databases different on perform given data settings. trained in poorly when acquired on,Classifiers trained on given databases perform poorly when tested on data acquired in different settings.,24561
data. video operates methodology on Our directly,Our methodology operates directly on video data.,24562
"we on the focus and step distortion. important image reducing contents preserving Especially, image projection","Especially, we focus on the projection step preserving important image contents and reducing image distortion.",24563
"projection based our projection is Basically, on model. method Pannini","Basically, our projection method is based on Pannini projection model.",24564
"is for producing temporally consistency temporal stable videos. normal-view for projection Finally, image the enforced","Finally, the temporal consistency for image projection is enforced for producing temporally stable normal-view videos.",24565
aim comfortable. passengers control systems climate thermally Vehicle keep to,Vehicle climate control systems aim to keep passengers thermally comfortable.,24566
"adenoid streams spindle cells, of etc. islands,","streams of spindle cells, adenoid islands, etc.",24567
Rectifier been deep networks. have convolutional used in (ReLUs) units widely neuron,Rectifier neuron units (ReLUs) have been widely used in deep convolutional networks.,24568
"has however, Exploiting proven information, visual the challenging.","Exploiting the visual information, however, has proven challenging.",24569
controlled recognition only the limit speech in We visual conditions. study of,We study the limit of visual only speech recognition in controlled conditions.,24570
"performance are visual recognition compared system. with automatic human of Finally, abilities speech a the","Finally, human abilities are compared with the performance of a visual automatic speech recognition system.",24571
"but statistical normal-hearing participants tests, hearing-impaired the participants reaching significance. without outperformed our In","In our tests, hearing-impaired participants outperformed the normal-hearing participants but without reaching statistical significance.",24572
"In is of crucial the importance. network approach, nodes choice of this","In this approach, the choice of network nodes is of crucial importance.",24573
voxels is fMRI One as option nodes. consider to,One option is to consider fMRI voxels as nodes.,24574
assumes are within that ROIs approach This voxels similar. functionally,This approach assumes that voxels within ROIs are functionally similar.,24575
does of assumption the ROI dynamics Therefore generally similar within hold. each voxel not,Therefore the assumption of similar voxel dynamics within each ROI does not generally hold.,24576
AI continue arguments is How Lighthill's to explained. current apply to,How Lighthill's arguments continue to apply to current AI is explained.,24577
and networks Neumann's are evolutionary complexity from against John natural arguments von algorithms neural discussed.,John von Neumann's arguments from natural complexity against neural networks and evolutionary algorithms are discussed.,24578
Dataology. computation Peter as paper advocating concludes Naur's studying by The,The paper concludes by advocating studying computation as Peter Naur's Dataology.,24579
enhances the perception and Incorporating dynamic semantics of scene. motion overall the,Incorporating semantics and motion enhances the overall perception of the dynamic scene.,24580
even information The incorporated. without more challenging being stereo task becomes,The task becomes even more challenging without stereo information being incorporated.,24581
for important replacements on Wear their is criterion (TKRs) performance an knee characteristics. total,Wear on total knee replacements (TKRs) is an important criterion for their performance characteristics.,24582
the simulations has attention wear of years. Numerical seen over such increasing last,Numerical simulations of such wear has seen increasing attention over the last years.,24583
comparing choice can by of experimental be only with results. validated The the correctness,The correctness of the choice can only be validated by comparing with experimental results.,24584
successfully rotations camera and SfM-Net frame-to-frame translations. extracts and estimates estimates meaningful depth,SfM-Net extracts meaningful depth estimates and successfully estimates frame-to-frame camera rotations and translations.,24585
the stimulus shape that responses. This systematically network investigates network properties and study,This study systematically investigates the network and stimulus properties that shape network responses.,24586
"effective oscillatory of the network we Subsequently, analyze the amplification by structure. stimuli","Subsequently, we analyze the amplification of oscillatory stimuli by the effective network structure.",24587
"and teams contacts. case and cases, Operationally, treated genetic samples, workers tracked health surveillance collected","Operationally, health workers and surveillance teams treated cases, collected genetic samples, and tracked case contacts.",24588
computationally Use of makes a intensive patch method. ordering TSPs large,Use of large TSPs makes patch ordering a computationally intensive method.,24589
image for A ordering proposed. denoising is method patch modified,A modified patch ordering method for image denoising is proposed.,24590
"training as are Finally, new the reprojected the to improve triangulations labeled detector. used data","Finally, the reprojected triangulations are used as new labeled training data to improve the detector.",24591
"We each iteration. data in this repeat labeled generating more process,","We repeat this process, generating more labeled data in each iteration.",24592
detector hand single is used images. The method train to keypoint for a,The method is used to train a hand keypoint detector for single images.,24593
"clustering to the applied state-of-the-art. also problem, the SAE is beats supervised when it Furthermore,","Furthermore, when the SAE is applied to supervised clustering problem, it also beats the state-of-the-art.",24594
The method as proposed be optimized basic problem. eigenvalue a can,The proposed method can be optimized as a basic eigenvalue problem.,24595
robustness demonstrate of our and the method. Results segmentation generalisability,Results demonstrate the robustness and generalisability of our segmentation method.,24596
dictionary. its Usually is own or each described component subspace by,Usually each component is described by its own subspace or dictionary.,24597
problem. requires a it binary to solve optimization Therefore,Therefore it requires to solve a binary optimization problem.,24598
"is open important automatic is of and Thus, study still problem. the an lip-reading","Thus, the study of automatic lip-reading is important and is still an open problem.",24599
"the (PCA, system analyzed. of DCT and performance different compare SIFT) are the descriptors To","To compare the performance of the system different descriptors (PCA, DCT and SIFT) are analyzed.",24600
test We system a continuous corpus speech. of in Spanish our,We test our system in a Spanish corpus of continuous speech.,24601
"this problems algorithms. thesis, study In on based two clustering we","In this thesis, we study two problems based on clustering algorithms.",24602
"as performance the We we clustering overall. observe that more increases attributes, add","We observe that as we add more attributes, the clustering performance increases overall.",24603
"specially are reduce pollutant to bicycles emissions, in Aiming urban popularity regaining areas.","Aiming to reduce pollutant emissions, bicycles are regaining popularity specially in urban areas.",24604
partitioning case. works the than in worst faster exploration the the cyclical algorithm Hence,Hence the partitioning algorithm works faster than the cyclical exploration in the worst case.,24605
colorize solution practical to raw A art have line automatically many would applications.,A solution to automatically colorize raw line art would have many practical applications.,24606
"the information in methods to We color present generalization. improving scheme, limit passed processing","We present processing methods to limit information passed in the color scheme, improving generalization.",24607
"biologically these predictions produced. Under pose circumstances, may be implausible","Under these circumstances, biologically implausible pose predictions may be produced.",24608
typically Explicit is of learning such challenging. constraints,Explicit learning of such constraints is typically challenging.,24609
is conservation. implies violation of human that a energy associated consciousness This with,This implies that human consciousness is associated with a violation of energy conservation.,24610
"""simplest"" of to This case problem is in even the NP-complete three known genomes. be","This problem is known to be NP-complete even in the ""simplest"" case of three genomes.",24611
to are approximations of exact executed algorithms the provide Heuristic usually solution.,Heuristic algorithms are usually executed to provide approximations of the exact solution.,24612
and role prominent planning a business plays strategy. business Sales forecasting in,Sales forecasting plays a prominent role in business planning and business strategy.,24613
"semiconductor this Corporation, CPU a considered. was sales industry, forecasting multinational of paper In Intel","In this paper CPU sales forecasting of Intel Corporation, a multinational semiconductor industry, was considered.",24614
with ensemble models smaller selected form were The errors validation the to model.,The models with smaller validation errors were selected to form the ensemble model.,24615
represent selected validation condition. automatically closely The market the which models moving current windows process,The moving windows validation process automatically selected the models which closely represent current market condition.,24616
allowed to The cadence schema forecasting effectively market weekly to model the response fluctuation.,The weekly cadence forecasting schema allowed the model to response effectively to market fluctuation.,24617
Generic interpretability. to developed analysis increase importance variable was the model also,Generic variable importance analysis was also developed to increase the model interpretability.,24618
"full a paper, recognition. In page we for present text approach new this","In this paper, we present a new approach for full page text recognition.",24619
"usually analyses Cortical carried projected onto where reference are subsequent are out. data a mesh,","Cortical data are usually projected onto a reference mesh, where subsequent analyses are carried out.",24620
"rely on very estimation. computational either require complexity, crude large However, or they","However, they either require large computational complexity, or rely on very crude estimation.",24621
proposed of Experiments superiority with the existing our algorithms. inference demonstrate compared algorithm,Experiments demonstrate the superiority of our proposed algorithm compared with existing inference algorithms.,24622
Why menstruate females some all? do at,Why do some females menstruate at all?,24623
weapon? to Is access data training secret their good/big,Is access to good/big training data their secret weapon?,24624
are components an intensity in important image. and two Color,Color and intensity are two important components in an image.,24625
"discovery. formulated object new A which color to for matrix, distortions, robust is is affinity","A new affinity matrix, which is robust to color distortions, is formulated for object discovery.",24626
the method proposed to Then generation. apply proposal both we saliency detection and object,Then we apply the proposed method to both saliency detection and object proposal generation.,24627
effective point sets. an proposes of scaling the for paper registration approach This $m$-D,This paper proposes an effective approach for the scaling registration of $m$-D point sets.,24628
"designs novel the problem. registration a paper function Therefore, for scaling this objective","Therefore, this paper designs a novel objective function for the scaling registration problem.",24629
Experimental demonstrate efficiency approaches superiority its on over robustness. results previous and,Experimental results demonstrate its superiority over previous approaches on efficiency and robustness.,24630
semantically boundaries. characteristics crosses Attributes applicability category whose are widely meaningful,Attributes are semantically meaningful characteristics whose applicability widely crosses category boundaries.,24631
"they human-computer used describable, be human interaction. are Additionally, can efficient for attributes since","Additionally, since attributes are human describable, they can be used for efficient human-computer interaction.",24632
"improve In semantic to paper, attribute to this propose we prediction. segmentation facial employ","In this paper, we propose to employ semantic segmentation to improve facial attribute prediction.",24633
attributes fact many in facial The that local lies idea describe the core properties.,The core idea lies in the fact that many facial attributes describe local properties.,24634
build segmentation with model a semantic jointly We our facial attribute deep prediction network.,We build our facial attribute prediction model jointly with a deep semantic segmentation network.,24635
jointly interconnected reaffirms need the these That model tasks. to two,That reaffirms the need to jointly model these two interconnected tasks.,24636
are approaches modern pipelines. Many two-staged detection for object,Many modern approaches for object detection are two-staged pipelines.,24637
the of stage are The classified which in regions identifies stage. then second first interest,The first stage identifies regions of interest which are then classified in the second stage.,24638
we R-CNN company of paper this apply In the task Faster logo detection. to,In this paper we apply Faster R-CNN to the task of company logo detection.,24639
on feature stages. those the We of resolution influence performance of the investigate map,We investigate the influence of feature map resolution on the performance of those stages.,24640
"summarization. can root used we the show Finally, how in equivalence notion of be argumentation","Finally, we show how the notion of root equivalence can be used in argumentation summarization.",24641
"trending imaging sensing. towards these Recently, encoded technologies wavelet have been sparse/compressive and","Recently, these technologies have been trending towards wavelet encoded imaging and sparse/compressive sensing.",24642
recognized It is well that central cars. has area drivable to detecting been self-driving,It has been well recognized that detecting drivable area is central to self-driving cars.,24643
"more of for feature robust fusion Besides, a performance. the also applied is level","Besides, a fusion of the feature level is also applied for more robust performance.",24644
"ROAD-KITTI hypothesis, is and yet Our unsupervised avoids common gets benchmark. results state-of-the-art on approach","Our approach is unsupervised and avoids common hypothesis, yet gets state-of-the-art results on ROAD-KITTI benchmark.",24645
underpinning of examples. referee described are expected with checks The an article data,The expected referee checks of data underpinning an article are described with examples.,24646
should be diffraction If the necessary reprocessed referee the deems data (e.g. images,If the referee deems necessary the diffraction data images should be reprocessed (e.g.,24647
different resolution the than a submission). diffraction authors to,to a different diffraction resolution than the authors submission).,24648
voluntary organisation such e.g. of can The certification training,The organisation of such voluntary certification training can e.g.,24649
such via by IUCr recognised duly to certificates. the be those issue crystallography associations,be via those crystallography associations duly recognised by the IUCr to issue such certificates.,24650
"people this flow in large aim we public infrastructures. paper, monitor of the In to","In this paper, we aim to monitor the flow of people in large public infrastructures.",24651
measure that data-irregularity assesses descriptor model. define We a each a how data well fits,We define a data-irregularity measure that assesses how well each descriptor fits a data model.,24652
distinctive This flow from highly allows rank to patterns us very to common (outliers) ones.,This allows us to rank flow patterns from highly distinctive (outliers) to very common ones.,24653
"configurations By more key discarding (classes). reliable outliers, we obtain","By discarding outliers, we obtain more reliable key configurations (classes).",24654
show standard that Synthetic method the superior is to experiments methods. clustering proposed,Synthetic experiments show that the proposed method is superior to standard clustering methods.,24655
work photography. are of within systems Computer the vision designed well everyday to context,Computer vision systems are designed to work well within the context of everyday photography.,24656
"not the However, render artists them that resemble in world photographs. around do often ways","However, artists often render the world around them in ways that do not resemble photographs.",24657
"with media. content, imagery Behance rich emotions, annotate artistic labels for attribute We and","We annotate Behance imagery with rich attribute labels for content, emotions, and artistic media.",24658
the method to cuts boundary. surface Our ridges extracts these form and,Our method extracts and cuts these ridges to form the surface boundary.,24659
that show of paths. is resulting the a collection surface minimal We,We show that the resulting surface is a collection of minimal paths.,24660
optimal generation. for We query different methods testing be will,We will be testing different methods for optimal query generation.,24661
The will dataset able learn respond and the be model using human human-like. will dialogs,The model will learn using the human dialogs dataset and will be able respond human-like.,24662
semi-supervised even supporting its experiences learn agent from past The will learning.,The agent will even learn from its past experiences supporting semi-supervised learning.,24663
forecasting a model insight at to of higher problem level abstraction. is the Our,Our insight is to model the forecasting problem at a higher level of abstraction.,24664
real-time in segmentation this semantic We focus paper. challenging of task the on,We focus on the challenging task of real-time semantic segmentation in this paper.,24665
our applications. energy in the usefulness medical We of demonstrate several,We demonstrate the usefulness of our energy in several medical applications.,24666
is (LEGGM) magnitude local Gabor The new gradient edge descriptor coined pattern.,The new descriptor is coined local edge gradient Gabor magnitude (LEGGM) pattern.,24667
"model, intrinsic some a generative it however as suffers feature from deficiencies. Behaved","Behaved as a generative feature model, it however suffers from some intrinsic deficiencies.",24668
"errors. images detected means from unavoidable captured suffer contours depth in acquisition noise However, that","However, acquisition noise in captured depth images means that detected contours suffer from unavoidable errors.",24669
"and propose classical deep architecture planning. We an combining learning LatPlan, unsupervised","We propose LatPlan, an unsupervised architecture combining deep learning and classical planning.",24670
normoglycemia to be hypoglycemia accurately avoid predicted disastrous and should Daytime situations. to achieve,Daytime hypoglycemia should be accurately predicted to achieve normoglycemia and to avoid disastrous situations.,24671
"daytime an abnormally into nocturnal hypoglycemia divided hypoglycemia. glucose Hypoglycemia, level, is blood low and","Hypoglycemia, an abnormally low blood glucose level, is divided into daytime hypoglycemia and nocturnal hypoglycemia.",24672
deal Many with prevention hypoglycemia. hypoglycemia of studies nocturnal,Many studies of hypoglycemia prevention deal with nocturnal hypoglycemia.,24673
We and (CART) classification tree as prediction apply a regression method.,We apply classification and regression tree (CART) as a prediction method.,24674
"relevant that BeME-WithFun might identify coherent helped modules be cancer functionally In for progression. particular,","In particular, BeME-WithFun helped identify functionally coherent modules that might be relevant for cancer progression.",24675
The makes to criterion minima. more local robust proposed tracking resulting boundary the,The proposed criterion makes the resulting boundary tracking more robust to local minima.,24676
the also fixation find enhance mutations that probability a can beneficial We nonmutator. of,We also find that beneficial mutations can enhance the fixation probability of a nonmutator.,24677
"boundaries difficult and It illumination object are occlusions vary, when weak. more and is","It is more difficult when occlusions and illumination vary, and object boundaries are weak.",24678
"a GazeDirector, approach new eye gaze present that model-fitting. redirection uses We for","We present GazeDirector, a new approach for eye gaze redirection that uses model-fitting.",24679
end-to-end network We for testing. and form training an,We form an end-to-end network for training and testing.,24680
accuracy. efficiency designed considering module and Each is,Each module is designed considering efficiency and accuracy.,24681
benefits applications. many video processing Our framework,Our framework benefits many video processing applications.,24682
"system images we which are STATIC, that ambiguous. develop a Using automatically predicts","Using STATIC, we develop a system that automatically predicts which images are ambiguous.",24683
then object. the localize in attempt an votes aggregated are The to target,The votes are then aggregated in an attempt to localize the target object.,24684
"collaborative this In tracking. information an we for exchange study, active propose scheme ensemble","In this study, we propose an active collaborative information exchange scheme for ensemble tracking.",24685
the results The by tracking. for evaluations algorithm promising the demonstrate realized online real-world proposed,The evaluations demonstrate promising results realized by the proposed algorithm for the real-world online tracking.,24686
state-of-the-art networks have semantic the image convolutional tasks. Deep segmentation for achieved,Deep convolutional networks have achieved the state-of-the-art for semantic image segmentation tasks.,24687
"CNN incurs during extra in the training. conditional fields incorporation Furthermore, of random little time","Furthermore, the incorporation of conditional random fields in CNN incurs little extra time during training.",24688
effective and are handcrafted-feature-based features proven in skeleton-based methods. There have several RNN-based that,There are several skeleton-based features that have proven effective in RNN-based and handcrafted-feature-based methods.,24689
"whether unknown they are approaches. remains CNN-based suitable for However, it","However, it remains unknown whether they are suitable for CNN-based approaches.",24690
encoding different skeleton paper with encode five features methods. to proposes spatial into This images,This paper proposes to encode five spatial skeleton features into images with different encoding methods.,24691
"joints feature the implication is addition, extraction for of performance studied. In used different","In addition, the performance implication of different joints used for feature extraction is studied.",24692
between finds two dense correspondences semantically-meaningful input images. technique Our,Our technique finds semantically-meaningful dense correspondences between two input images.,24693
for to compute coarse-to-fine nearest-neighbor used the generating field A results. strategy is the,A coarse-to-fine strategy is used to compute the nearest-neighbor field for generating the results.,24694
"Captions, dense-captioning also We events. introduce a ActivityNet for large-scale benchmark","We also introduce ActivityNet Captions, a large-scale benchmark for dense-captioning events.",24695
"localization. video performances dense-captioning and events, model of retrieval report we for our Finally,","Finally, we report performances of our model for dense-captioning events, video retrieval and localization.",24696
organisms increasingly viewed used a increase their is fitness. resource being to Information by as,Information is increasingly being viewed as a resource used by organisms to increase their fitness.,24697
"information assumed However, all collection this cost-free. of that work is","However, all of this work assumed that information collection is cost-free.",24698
"of the these ignorant. the in remain it organism instances, In is interest to counter-intuitively,","In these instances, counter-intuitively, it is in the interest of the organism to remain ignorant.",24699
use as encoding a architectures pooling step. final average Most feature recent CNN,Most recent CNN architectures use average pooling as a final feature encoding step.,24700
"novel addition, a these present by made way to visualize we approaches. In decisions","In addition, we present a novel way to visualize decisions made by these approaches.",24701
pooling. allow moving the us when to between average analyze bilinear and difference contributions Both,Both contributions allow us to analyze the difference when moving between average and bilinear pooling.,24702
effectiveness experiments the exemplify these numerical Our techniques. of,Our numerical experiments exemplify the effectiveness of these techniques.,24703
preference differentiated phenotypes and to populations. different persistence the Mating of can lead emergence for,Mating preference for different phenotypes can lead to the emergence and persistence of differentiated populations.,24704
such alleles Dominance between in divergent processes. interfere can encoding for phenotypes,Dominance between alleles encoding for divergent phenotypes can interfere in such processes.,24705
"differentiation preference. When arises compared assuming is co-dominance, as when limited only migration to","When assuming co-dominance, differentiation arises only when migration is limited as compared to preference.",24706
dependent when Such migration is assuming between haplotypes. differentiation less dominance on strict,Such differentiation is less dependent on migration when assuming strict dominance between haplotypes.,24707
"dominance alleles, associated with (i.e. distribution However, initial the advantage of on this depending","However, depending on the initial distribution of alleles, this advantage associated with dominance (i.e.",24708
populations. Haldane's to may the of lead throughout both dominant allele fixation sieve),Haldane's sieve) may lead to fixation of the dominant allele throughout both populations.,24709
system. were anterior the into showing and of activity septum divided an Areas AChE posterior,Areas of the septum showing AChE activity were divided into an anterior and posterior system.,24710
man is applicability The findings in animals neurophysiologic some of discussed. to,The applicability to man of some neurophysiologic findings in animals is discussed.,24711
developed. is to looped from of parameters recover A being objects structural multiframes method jointed,A method to recover structural parameters of looped jointed objects from multiframes is being developed.,24712
"form, conventional as stated multiple rotation its constraints. In averaging is a over minimization rotation","In its conventional form, rotation averaging is stated as a minimization over multiple rotation constraints.",24713
"generally globally. to considered is As problem these solve non-convex, challenging are constraints this","As these constraints are non-convex, this problem is generally considered challenging to solve globally.",24714
this use circumvent show the of Lagrangian duality. to how difficulty We through,We show how to circumvent this difficulty through the use of Lagrangian duality.,24715
missing This information. to ambiguities the has more far task due depth,This task has far more ambiguities due to the missing depth information.,24716
discussed. are for Different methods Dimensionality reduction,Different methods for Dimensionality reduction are discussed.,24717
"used and Neural Network, RandomForest. SVM this Classifiers work in are","Classifiers used in this work are Neural Network, SVM and RandomForest.",24718
us regularize motion. allows the This structure of instead also the to,This allows us to also regularize the structure instead of the motion.,24719
capturing for data We seek image-based automate modeling. to,We seek to automate data capturing for image-based modeling.,24720
L\'evy in many applications stretched and distributions fractional calculus. have and fractal Gaussian,L\'evy and stretched Gaussian distributions have many applications in fractional and fractal calculus.,24721
error scores in tractography. benefits reduced further the tangible that produce show downstream We,We further show that the reduced error scores produce tangible benefits in downstream tractography.,24722
observed The RWLS perform wavelet transforms. the is to learned than standard better existing dyadic,The learned RWLS is observed to perform better than the existing standard dyadic wavelet transforms.,24723
"formulate classification Bayesian problem we a HSI from perspective. the First,","First, we formulate the HSI classification problem from a Bayesian perspective.",24724
"labels. smoothness spatial on Next, by a spatial the considered placing further prior information is","Next, spatial information is further considered by placing a spatial smoothness prior on the labels.",24725
problem. an detection in Temporal important long action is videos,Temporal action detection in long videos is an important problem.,24726
on problem sliding applying this by address methods action classifiers windows. State-of-the-art,State-of-the-art methods address this problem by applying action classifiers on sliding windows.,24727
with adapt Boundary pipeline We Cascaded temporal Regression (CBR) a model. detection two-stage action,We adapt a two-stage temporal action detection pipeline with Cascaded Boundary Regression (CBR) model.,24728
Class-agnostic specific are the second stage. first proposals actions and detected in and the respectively,Class-agnostic proposals and specific actions are detected respectively in the first and the second stage.,24729
the temporal sliding uses refine coordinate CBR the boundaries temporal to of windows. regression,CBR uses temporal coordinate regression to refine the temporal boundaries of the sliding windows.,24730
"The high under remarkable IoU thresholds, is e.g. especially performance gain","The performance gain is especially remarkable under high IoU thresholds, e.g.",24731
We analyzing system plant in indoor automated present for growth an conditions.,We present an automated system for analyzing plant growth in indoor conditions.,24732
variable classification. proposed methods selection Relevant of have model-based and clustering been in,Relevant methods of variable selection have been proposed in model-based clustering and classification.,24733
"this in weighted this To issue, model is a address MMD paper. proposed","To address this issue, a weighted MMD model is proposed in this paper.",24734
of over superiority MMD MMD adaptation. for demonstrate experiments Extensive the conventional our weighted domain,Extensive experiments demonstrate the superiority of our weighted MMD over conventional MMD for domain adaptation.,24735
in and vision the of peripheral scene recognition? roles What central are human,What are the roles of central and peripheral vision in human scene recognition?,24736
"for both and used and EDI We investigate update. how, revision argue can that, be","We argue that, and investigate how, EDI can be used for both revision and update.",24737
class of imbalance boundaries. errors along addresses The challenges and severe SD-Net,The SD-Net addresses challenges of severe class imbalance and errors along boundaries.,24738
Our inclusion results of significant labels the auxiliary ECB yields show improvements. and that,Our results show that the inclusion of auxiliary labels and ECB yields significant improvements.,24739
the It state-of-the-art latest models. F-CNN outperforms also,It also outperforms the latest state-of-the-art F-CNN models.,24740
part layers The the deep filters discriminative utilizes convolutional detectors. as PWA proposed of,The proposed PWA utilizes the discriminative filters of deep convolutional layers as part detectors.,24741
"can joint performed. Thus, cross-domain disentanglement be jointly adaptation and feature","Thus, cross-domain joint feature disentanglement and adaptation can be jointly performed.",24742
"capability. verify we the In provide results qualitative experiments, to our disentanglement","In the experiments, we provide qualitative results to verify our disentanglement capability.",24743
frequency) recover. we to want that high,high frequency) that we want to recover.,24744
recently methods show proposed and solution. of state-of-the-art with Results published the comparisons the superiority,Results and comparisons with recently published state-of-the-art methods show the superiority of the proposed solution.,24745
validate We the effectiveness the of on datasets. approach Visual and our COCO Microsoft Genome,We validate the effectiveness of our approach on the Microsoft COCO and Visual Genome datasets.,24746
for binary to annotation. best how on focus location Existing approaches labels the object utilise,Existing approaches focus on how to best utilise the binary labels for object location annotation.,24747
"overlapping, on vertex sets. networks non-identical two Consider","Consider two networks on overlapping, non-identical vertex sets.",24748
and of We data examples. real through applicability the methodology demonstrate simulations our,We demonstrate the applicability of our methodology through simulations and real data examples.,24749
We compatible sketch. for which model colors can a the propose automatically auto-painter generate,We propose the auto-painter model which can automatically generate compatible colors for a sketch.,24750
of algorithm. In provide we this and convergence of correctness a paper detailed analysis this,In this paper we provide a detailed analysis of convergence and correctness of this algorithm.,24751
We extend al. and ideas generalize of et [Springborn,We generalize and extend ideas of [Springborn et al.,24752
localization temporal of on focuses untrimmed paper videos. in actions This,This paper focuses on temporal localization of actions in untrimmed videos.,24753
localize propose queries. activities natural to We by language,We propose to localize activities by natural language queries.,24754
queries complex also in test. for We Charades-STA sentence build,We also build complex sentence queries in Charades-STA for test.,24755
that Experimental methods outperforms show on datasets. previous CTRL both results significantly,Experimental results show that CTRL outperforms previous methods significantly on both datasets.,24756
"methods flickering to highly However, videos. real-time in when are visible applied unstable resulting","However, real-time methods are highly unstable resulting in visible flickering when applied to videos.",24757
in deal the action of predicting we this progress In with videos. paper problem,In this paper we deal with the problem of predicting action progress in videos.,24758
"actions result, define we a their categorization of and a As phases.","As a result, we define a categorization of actions and their phases.",24759
"our will of paper, the acceptance available. layer data publicly augmentation be made Upon robot","Upon acceptance of the paper, our robot data augmentation layer will be made publicly available.",24760
concept We introduce component-trees an of derivate-based arbitrary number the images with of channels. for,We introduce the concept of derivate-based component-trees for images with an arbitrary number of channels.,24761
approach component-tree devoted gray-scale extension is the to The of a images. classical natural,The approach is a natural extension of the classical component-tree devoted to gray-scale images.,24762
networks tested on trained on when Deep distortions. pristine perform such images poorly,Deep networks trained on pristine images perform poorly when tested on such distortions.,24763
"In that work learning. a of for present distributed, this online allows we version OLED","In this work we present a version of OLED that allows for distributed, online learning.",24764
achieves state-of-the-art on approach performance Our datasets. both,Our approach achieves state-of-the-art performance on both datasets.,24765
over Training variables. stochastic stochastic intractable networks involves marginalization an,Training stochastic networks involves an intractable marginalization over stochastic variables.,24766
"modes. multiple characterizations the In analytic to of rigorous we contrast provide literature, existing","In contrast to the existing literature, we provide rigorous analytic characterizations of multiple modes.",24767
This and the the of method can cost. reduce increase computational effectiveness,This can reduce the effectiveness of the method and increase computational cost.,24768
with for labels. converting We propose hierarchically method geometric parts shapes segmented part a into,We propose a method for converting geometric shapes into hierarchically segmented parts with part labels.,24769
"untapped annotations geometry. enormous, These represent of source freely-available information on an","These freely-available annotations represent an enormous, untapped source of information on geometry.",24770
proposed Simulations demonstrate the results efficacy formulation. of the,Simulations results demonstrate the efficacy of the proposed formulation.,24771
is a (re-id). person Large-scale in trend re-identification,Large-scale is a trend in person re-identification (re-id).,24772
important gallery. in that performed large real-time It a is be search,It is important that real-time search be performed in a large gallery.,24773
challenging proposed We the method the on test VOC datasets. PASCAL,We test the proposed method on the challenging PASCAL VOC datasets.,24774
provided. research for recommendations are Conclusions and further,Conclusions and recommendations for further research are provided.,24775
significant Facial and expressions communication human behavior. in a role play,Facial expressions play a significant role in human communication and behavior.,24776
the between long studied have facial expressions and relationship Psychologists emotions.,Psychologists have long studied the relationship between facial expressions and emotions.,24777
"framework estimation landmark localisation. has aggregation, detection, box and The four stages: pose bounding face","The framework has four stages: face detection, bounding box aggregation, pose estimation and landmark localisation.",24778
The classification. for reconstruction class uses algorithm errors SRC,The SRC algorithm uses class reconstruction errors for classification.,24779
a into recognition the testing hypothesis problem we simplify of problems. open set Then set,Then we simplify the open set recognition problem into a set of hypothesis testing problems.,24780
"realistic many unattainable. collecting However, images is sufficient ever-growing for applications, in classes new","However, in many realistic applications, collecting sufficient images for ever-growing new classes is unattainable.",24781
"recognition Hereafter, problem, the converted is ZSL into i.e. conventional supervised","Hereafter, ZSL recognition is converted into the conventional supervised problem, i.e.",24782
visual can classifiers features as straightforwardly such SVM. to typical be the fed synthesised,the synthesised visual features can be straightforwardly fed to typical classifiers such as SVM.,24783
"we data. of using On unseen four benefit benchmark datasets, demonstrate synthesised the","On four benchmark datasets, we demonstrate the benefit of using synthesised unseen data.",24784
suggest experimental results proposed improve state-of-the-art the results. Extensive our that significantly approach,Extensive experimental results suggest that our proposed approach significantly improve the state-of-the-art results.,24785
applications. our We of efficiency a the range SG-WLS in show effectiveness and of,We show the effectiveness and efficiency of our SG-WLS in a range of applications.,24786
highly of representation reasoning at Logics limited feasible enabling in expressive languages. belief computationally aim,Logics of limited belief aim at enabling computationally feasible reasoning in highly expressive representation languages.,24787
"to we paper, this beyond go the theory. aim In","In this paper, we aim to go beyond the theory.",24788
of landscape Wright's a metaphor part evolutionary thinking. adaptive penetrates significant Sewall,Sewall Wright's adaptive landscape metaphor penetrates a significant part of evolutionary thinking.,24789
"with questions. two are Therefore, we faced","Therefore, we are faced with two questions.",24790
"yet its of understood. elegans, well the relative simplicity locomotion is not C. Despite machinery","Despite the relative simplicity of C. elegans, its locomotion machinery is not yet well understood.",24791
the bends. on dorsoventral of We body focus generation,We focus on the generation of dorsoventral body bends.,24792
"describing rank event, the videos is goal Given related accordingly. a to novel text a","Given a text describing a novel event, the goal is to rank related videos accordingly.",24793
"are task given the This to no zero-exemplar, event. video novel is examples","This task is zero-exemplar, no video examples are given to the novel event.",24794
train sources. detectors bank on Related a external works of data concept,Related works train a bank of concept detectors on external data sources.,24795
"retrieved for test videos, and confidence predict ranked accordingly. which detectors These are scores","These detectors predict confidence scores for test videos, which are ranked and retrieved accordingly.",24796
as pre-defined The event casts events. a a space novel of probability,The space casts a novel event as a probability of pre-defined events.,24797
"distance learns related measure and an it videos. between Also, event to the its","Also, it learns to measure the distance between an event and its related videos.",24798
Our end-to-end on trained publicly available model is EventNet.,Our model is trained end-to-end on publicly available EventNet.,24799
Nature Christensen Tangled modify et. Model of We the,We modify the Tangled Nature Model of Christensen et.,24800
capacity. the the affect so that carrying agents,so that the agents affect the carrying capacity.,24801
"oriented. datasets successfully time, images same three are At in the the","At the same time, images in the three datasets are successfully oriented.",24802
"a UAV efficient oblique more be can For method images, the proposed of orientation solution.","For orientation of oblique UAV images, the proposed method can be a more efficient solution.",24803
is Video processing requirements. to both these fulfil suitable,Video processing is suitable to fulfil both these requirements.,24804
"has trajectory barycentre tracking. is be to In first that the the done diving, analysis","In diving, the first analysis that has to be done is the barycentre trajectory tracking.",24805
cameras sensor. moving captured without videos any by additional,videos captured by moving cameras without any additional sensor.,24806
larval Drosophila spectral present We modeling the of complete connectome. mushroom body semiparametric,We present semiparametric spectral modeling of the complete larval Drosophila mushroom body connectome.,24807
recognition. success have great for convolutional achieved networks Deep image,Deep convolutional networks have achieved great success for image recognition.,24808
"videos, traditional not action is evident. However, advantage recognition over in so for their methods","However, for action recognition in videos, their advantage over traditional methods is not so evident.",24809
a models for in videos. present We learning and flexible action general video-level framework,We present a general and flexible video-level framework for learning action models in videos.,24810
"such of for data. training major is however, methods, challenge the One collection","One major challenge for such methods, however, is the collection of training data.",24811
training use We to independent sources. therefore propose two,We therefore propose to use two independent training sources.,24812
dynamic that Fixed distinct patterns. via networks output point networks are stimuli encode,Fixed point networks are dynamic networks that encode stimuli via distinct output patterns.,24813
"and successfully optimization combination can However, reduction the with dimension selection provide functionality. of such","However, the combination of dimension reduction with selection and optimization can successfully provide such functionality.",24814
"a present approach detection this contour In Convolutional paper, novel with we for Networks. Neural","In this paper, we present a novel approach for contour detection with Convolutional Neural Networks.",24815
measurements patches. maps overlapping contour patch-level with create uses Our to method,Our method uses patch-level measurements to create contour maps with overlapping patches.,24816
the databases proposed of the technique. results benchmark contour effectiveness the demonstrate Experimental major on,Experimental results on the major contour benchmark databases demonstrate the effectiveness of the proposed technique.,24817
can We contours. detection both achieve and show of fine-scale large-scale our good method,We show our method can achieve good detection of both fine-scale and large-scale contours.,24818
"tedious, to only process variabilities. is Such not prune also but intra/inter-observe a","Such a process is not only tedious, but also prune to intra/inter-observe variabilities.",24819
"the marketing Recently, of pathology. imaging digital slide system FDA whole first for the cleared","Recently, FDA cleared the marketing of the first whole slide imaging system for digital pathology.",24820
achieve excellent on classification standard Deep performance tasks. (DNNs) networks neural,Deep neural networks (DNNs) achieve excellent performance on standard classification tasks.,24821
"distortions noise, accuracy However, classification quality and blur poor. image such under as becomes","However, under image quality distortions such as blur and noise, classification accuracy becomes poor.",24822
"DNNs we the distorted In this subjects human with of images. compare work, performance on","In this work, we compare the performance of DNNs with human subjects on distorted images.",24823
that DNNs additionally between find errors We there is subjects. in and correlation little human,We additionally find that there is little correlation in errors between DNNs and human subjects.,24824
problem Learning. Recognition Machine common Face in is a,Face Recognition is a common problem in Machine Learning.,24825
widely used already been lives. in our technology This has,This technology has already been widely used in our lives.,24826
"Face comes expression with different occlusion. background, variant different and facial images illumination,","Face images comes with different background, variant illumination, different facial expression and occlusion.",24827
for a face of recognition. number There are the approaches large,There are a large number of approaches for the face recognition.,24828
"different databases. so, Doing with suit approaches these don't face","Doing so, these approaches don't suit with different face databases.",24829
used algorithms is for implementing MATLAB also.,MATLAB is used for implementing algorithms also.,24830
to an handle uncertainty issue. diagnosis open medical is How in,How to handle uncertainty in medical diagnosis is an open issue.,24831
"Z-numbers new based methodology In making paper, decision is this presented. a on","In this paper, a new decision making methodology based on Z-numbers is presented.",24832
"opinions the Firstly, Z-numbers. represented by experts' are","Firstly, the experts' opinions are represented by Z-numbers.",24833
"pair as = ordered numbers (A, Z fuzzy is an Z-number of B). denoted","Z-number is an ordered pair of fuzzy numbers denoted as Z = (A, B).",24834
"a method numbers proposed. ranking is Then, fuzzy for new","Then, a new method for ranking fuzzy numbers is proposed.",24835
"result, As combination by a is Dempster' information different the the rule. from combined sources","As a result, the information from different sources is combined by the Dempster' combination rule.",24836
information of decision advantage to final is fusion. reasonable more making The due the,The final decision making is more reasonable due to the advantage of information fusion.,24837
component data This a uncertainty. provides degree of tolerance for,This component provides a degree of tolerance for data uncertainty.,24838
dynamic. system is temporal the meant to This tackle,This is meant to tackle the temporal system dynamic.,24839
"manner. In now a to acknowledged occur is bursty mRNA in synthesis highly particular,","In particular, mRNA synthesis is now acknowledged to occur in a highly bursty manner.",24840
"naive would be However, intractable. a fitting procedure","However, a naive fitting procedure would be intractable.",24841
seems approach of data. expression in single-cell promising This explosion the to relation current,This approach seems promising in relation to the current explosion of single-cell expression data.,24842
is level performance Sport is exponentially for a that athletes. all trend growing analysis,Sport performance analysis is a trend that is growing exponentially for all level athletes.,24843
here based and technique The cost. is two important shown on low requirements: flexibility,The technique here shown is based on two important requirements: flexibility and low cost.,24844
video Shot manipulation. detection an (SBD) for is pre-processing boundary important step,Shot boundary detection (SBD) is an important pre-processing step for video manipulation.,24845
"frames is no classified Here, each or of segment either sharp, as gradual transition.","Here, each segment of frames is classified as either sharp, gradual or no transition.",24846
"flow prevents computations the heavy of optical this. However,","However, the heavy computations of optical flow prevents this.",24847
models. compositing The transitions synthetically are using image generated,The transitions are generated synthetically using image compositing models.,24848
assign conflicting Social a is (weighted) to abstract principled to values way argumentation arguments.,Social abstract argumentation is a principled way to assign values to conflicting (weighted) arguments.,24849
the uniqueness of of property the this note we the important discuss model. In,In this note we discuss the important property of the uniqueness of the model.,24850
"further SSSA uses goes SSA its observers, as the by relevant whereas sample observer-moments. using","SSA uses as its sample the relevant observers, whereas SSSA goes further by using observer-moments.",24851
to principle warrants conclusive. The this law-like from far which similarly degree status is,The degree to which this principle warrants similarly law-like status is far from conclusive.,24852
"been principle evaluating assumptions. Typically, requiring has additional straightforward, not this","Typically, evaluating this principle has not been straightforward, requiring additional assumptions.",24853
for size provide principle. Our the of support the broad results applicability,Our results provide support for the broad applicability of the size principle.,24854
"Epileptic syn- in abnormal, the seizures overly are activity electrical brain. by caused chronized,","Epileptic seizures are caused by abnormal, overly syn- chronized, electrical activity in the brain.",24855
"mal electrical activity the as brain. propagating waves, manifests across abnor- The","The abnor- mal electrical activity manifests as waves, propagating across the brain.",24856
to long a term network. LSTM consider accomplish We how prediction using,We consider how to accomplish long term prediction using a LSTM network.,24857
"problem, structures, To using multi-resolution two gradient both the we vanishing propose alleviate encoder-decoder-predictor representation.","To alleviate the vanishing gradient problem, we propose two encoder-decoder-predictor structures, both using multi-resolution representation.",24858
"Through kinds quantitatively. of the extensive experiments, different effects features we evaluate of extra","Through extensive experiments, we evaluate the effects of different kinds of extra features quantitatively.",24859
of on HyperLearner. experimental The pedestrian the effectiveness the results validate multiple proposed benchmarks,The experimental results on multiple pedestrian benchmarks validate the effectiveness of the proposed HyperLearner.,24860
providing improved approach occlusion and promise Our handling. localization shows,Our approach shows promise providing improved localization and occlusion handling.,24861
and page challenge dataset segmentation layouts The text that varying contains degradations methods. line,The dataset contains varying page layouts and degradations that challenge text line segmentation methods.,24862
paper based evaluation that propose we scheme a In on this is new baselines.,In this paper we propose a new evaluation scheme that is based on baselines.,24863
"line we Finally, present published text detection results by recently algorithm. a achieved","Finally, we present results achieved by a recently published text line detection algorithm.",24864
segmenting and images. presents from camera-trap method This a classifying mammal genera work for,This work presents a method for classifying and segmenting mammal genera from camera-trap images.,24865
images. detection approach camera-trap This new toward of from a shows a method animals fully-automatic,This method shows a new approach toward a fully-automatic detection of animals from camera-trap images.,24866
outperformed individual accuracy. neural fusion-based the classification networks This regarding all approach,This fusion-based approach outperformed all the individual neural networks regarding classification accuracy.,24867
image the Generative propose Net generation. Cooperative (GCN) We for,We propose the Generative Cooperative Net (GCN) for image generation.,24868
of usage a We such also tool. an as model data-augmentation discovered generative,We also discovered a usage of such generative model as an data-augmentation tool.,24869
synthesized large help It up could set easy a dataset. and very generate is to,It is easy to set up and could help generate a very large synthesized dataset.,24870
dealing is There shape substantial work in vision pattern and with computer similarity. recognition,There is substantial work in pattern recognition and computer vision dealing with shape similarity.,24871
data are often from videos as such manifold. complex Visual sampled,Visual data such as videos are often sampled from complex manifold.,24872
"The evaluated is to datasets, baselines. significant approach improvements the proposed two benchmark on showing","The proposed approach is evaluated on two benchmark datasets, showing significant improvements to the baselines.",24873
We present talking a for a a generating video face. method of,We present a method for generating a video of a talking face.,24874
is The unlabelled tens of trained model on of hours videos.,The model is trained on tens of hours of unlabelled videos.,24875
show also videos different from We person. speech using a of re-dubbing results,We also show results of re-dubbing videos using speech from a different person.,24876
networks detailed (CNNs) ChineseFoodNet. convolutional benchmark on We a several of state-of-the-art share deep neural,We share a detailed benchmark of several state-of-the-art deep convolutional neural networks (CNNs) on ChineseFoodNet.,24877
modelling. topic We framework joint based novel on propose Bayesian a,We propose a novel framework based on Bayesian joint topic modelling.,24878
VOC on challenging Extensive demonstrate our experiments the that the competitors. outperforms dataset state-of-the-art approach,Extensive experiments on the challenging VOC dataset demonstrate that our approach outperforms the state-of-the-art competitors.,24879
for practical do not applications. to Even performance existing lead the recognizers satisfactory best,Even the best existing recognizers do not lead to satisfactory performance for practical applications.,24880
Model call Attention this We model the Convolutional (CHAM). Hierarchical,We call this model the Convolutional Hierarchical Attention Model (CHAM).,24881
are labels. as data extracted coarse from the training Themes treated,Themes extracted from the training data are treated as coarse labels.,24882
knowledge. any without prior requiring Our for and labels extracts coarse system training test images,Our system extracts coarse labels for training and test images without requiring any prior knowledge.,24883
labels are with fine to Vocabulary be the images. words associated,Vocabulary words are the fine labels to be associated with images.,24884
a scores with Our words of frequencies. wide maintains for high system range precision and,Our system scores and maintains high precision for words with a wide range of frequencies.,24885
local of the global and predictions. is Final saliency cooperation those prediction achieved via,Final saliency prediction is achieved via the cooperation of those global and local predictions.,24886
biological disease. imaging and development a role plays vital in Microscopy many understanding processes in,Microscopy imaging plays a vital role in understanding many biological processes in development and disease.,24887
in their coordinate changes to genome expression Cells adjustments accommodate environment. in,Cells coordinate adjustments in genome expression to accommodate changes in their environment.,24888
significant interrupt in changes transcriptomes discovering Drastic of transcriptomics. genes comparative biologically,Drastic changes of transcriptomes interrupt discovering biologically significant genes in comparative transcriptomics.,24889
in such inspection This reverse engineering. holds as industrial assumption always or almost applications,This assumption almost always holds in applications such as industrial inspection or reverse engineering.,24890
multimodal process training set. only image requires as the a few The registered learning pairs,The learning process requires only a few registered multimodal image pairs as the training set.,24891
"different from unpaired a image data proposed. alignment term is resolutions/modalities, combine To image hetero-domain","To combine unpaired data from different image resolutions/modalities, a hetero-domain image alignment term is proposed.",24892
"been neural implemented segmentation Convolutional object detection. have and in image networks classification,","Convolutional neural networks have been implemented in image classification, segmentation and object detection.",24893
for degradation. the We suppress connections to addition help of also call residual,We also call for the addition of residual connections to help suppress degradation.,24894
"model our was trained tests, from scratch. In this","In our tests, this model was trained from scratch.",24895
the first for image-based the We full people model body. present generative in of clothing,We present the first image-based generative model of people in clothing for the full body.,24896
"a image we models learn generative Instead, database. from large","Instead, we learn generative models from a large image database.",24897
"so have pure approaches image-based not far. For been considered this reason,","For this reason, pure image-based approaches have not been considered so far.",24898
"clothing. First, segmentation a and learn to the of generate body semantic we","First, we learn to generate a semantic segmentation of the body and clothing.",24899
"images. learn we on Second, resulting creates the segments a model realistic conditional that","Second, we learn a conditional model on the resulting segments that creates realistic images.",24900
"or can conditioned pose, color. be model is differentiable and shape The on full","The full model is differentiable and can be conditioned on pose, shape or color.",24901
The styles. samples are clothing different in result items and of people,The result are samples of people in different clothing items and styles.,24902
people entirely with clothing. proposed new can model generate realistic The,The proposed model can generate entirely new people with realistic clothing.,24903
"robust CCO and CCA. the we Second, standard for kernel IF kernel derive","Second, we derive the IF for robust kernel CCO and standard kernel CCA.",24904
visual in representation Robust role feature significant plays tracking.,Robust feature representation plays significant role in visual tracking.,24905
"issue, a may many the experimental challenging factors However, it since remains affect performance.","However, it remains a challenging issue, since many factors may affect the experimental performance.",24906
"with etc. scale test deformation quences We variation, se- challenging our on illumination variation, approach","We test our approach on challenging se- quences with illumination variation, scale variation, deformation etc.",24907
Experimental methods the results demonstrate that proposed outperforms favorably. method state-of-the-art,Experimental results demonstrate that the proposed method outperforms state-of-the-art methods favorably.,24908
no the points. data mechanism handle has texts to LSH the of con-,LSH has no mechanism to handle the con- texts of the data points.,24909
contribution made The in this is article three fold.,The contribution made in this article is three fold.,24910
similar applied video sequences. method The for retrieval extracting is resultant to,The resultant method is applied to video retrieval for extracting similar sequences.,24911
stages. these We propose can compensate new that image for a design skipping sensor,We propose a new image sensor design that can compensate for skipping these stages.,24912
(ADCs). features converters analog-to-digital tunable resolution adjustable an The and design sensor,The sensor design features an adjustable resolution and tunable analog-to-digital converters (ADCs).,24913
"variety models these of As are a problems for computational practical useful applications, challenging.","As useful models for a variety of practical applications, these problems are computational challenging.",24914
idea the extends The backbone-based crossovers. crossover of backbone-based double general,The double backbone-based crossover extends the idea of general backbone-based crossovers.,24915
"each Finally, we of the key component. investigate algorithmic usefulness","Finally, we investigate the usefulness of each key algorithmic component.",24916
obtain We publicly-available results on benchmarks. state-of-the-art,We obtain state-of-the-art results on publicly-available benchmarks.,24917
"salient object demonstrated detection. deep Neural Networks RGB strong performance (CNN) Convolutional have Recently, on","Recently, deep Convolutional Neural Networks (CNN) have demonstrated strong performance on RGB salient object detection.",24918
object compared We RGB-D methods. show state-of-the-art detection improved results salient to,We show improved results compared to state-of-the-art RGB-D salient object detection methods.,24919
for A is derived evaluated. this possible and method doing,A possible method for doing this is derived and evaluated.,24920
images. for develop colorizing grayscale We natural technique a probabilistic,We develop a probabilistic technique for colorizing grayscale natural images.,24921
"Following the become research image has growing preeminent image categorization automatic digital area. usage, rapidly","Following the rapidly growing digital image usage, automatic image categorization has become preeminent research area.",24922
"(CNN), generate promising based network and approaches they results. are neural Latest convolutional on","Latest approaches are based on convolutional neural network (CNN), and they generate promising results.",24923
"addition, of to spatial use resolution. is specific this approach limited a the In","In addition, this approach is limited to the use of a specific spatial resolution.",24924
"networks problems, these fully propose we (FCN) for To convolutional remedy SIVG. exploiting","To remedy these problems, we propose exploiting fully convolutional networks (FCN) for SIVG.",24925
for FCN architectures We SIVG. two present,We present two FCN architectures for SIVG.,24926
"by chrominance denoted DeepView$_{dec}$. of one consists decoupled networks luminance The for second signals, and","The second one consists of decoupled networks for luminance and chrominance signals, denoted by DeepView$_{dec}$.",24927
We studies. evaluated subjective with our and objective both approach,We evaluated our approach with both objective and subjective studies.,24928
efficiently. method segmentation our can Cross-validation shown promising that obtain has,Cross-validation has shown that our method can obtain promising segmentation efficiently.,24929
of theories way to into a This perspective framework. single a curiosity unify offers previous,This perspective offers a way to unify previous theories of curiosity into a single framework.,24930
with deep network. We encoder-decoder algorithm develop fully a novel top-down convolutional a detection contour,We develop a novel deep contour detection algorithm with a top-down fully convolutional encoder-decoder network.,24931
proposed technologies precise clearer prediction. a more to lead and above The,The above proposed technologies lead to a more precise and clearer prediction.,24932
cellular using capture resolution train high first to We pixel CNN a information. level patches,We first train a CNN using high pixel resolution patches to capture cellular level information.,24933
recognition Deep outperformed networks convolutional detection (CNNs) have object algorithms. neural existing and,Deep convolutional neural networks (CNNs) have outperformed existing object recognition and detection algorithms.,24934
that On diverse. scenes are captures the satellite imagery other hand,On the other hand satellite imagery captures scenes that are diverse.,24935
"variables in semi-parametric, presence non-linear of latent the regression is model A introduced.","A semi-parametric, non-linear regression model in the presence of latent variables is introduced.",24936
"This a ""sparse optimization of low-rank"" plus with to structured flavor. leads problems classes convex","This leads to classes of structured convex optimization problems with a ""sparse plus low-rank"" flavor.",24937
for the are the Details of algorithm presented. model learning,Details of the algorithm for learning the model are presented.,24938
estimation. for indices a selection experimental Sobol' points design We of consider problem,We consider a problem of experimental design points selection for Sobol' indices estimation.,24939
the demonstrate of applications provide a efficiency of We the approach. set proposed that,We provide a set of applications that demonstrate the efficiency of the proposed approach.,24940
on model state-of-art. our and dataset We evaluate results the to the achieve comparable LSUN,We evaluate our model on the LSUN dataset and achieve results comparable to the state-of-art.,24941
even or its dimensions. application This infeasible in makes growth exponential high unattractive,This exponential growth makes its application unattractive or even infeasible in high dimensions.,24942
"We positively combination discrepancy. Latin also hypercube potential a affects sampling, with which discuss","We also discuss a potential combination with Latin hypercube sampling, which positively affects discrepancy.",24943
novel paper method we In a present obstacle stereo the using camera. for this avoidance,In this paper we present a novel method for obstacle avoidance using the stereo camera.,24944
avoidance obstacle their discussed. limitations are and The conventional methods,The conventional obstacle avoidance methods and their limitations are discussed.,24945
to fuzzy controller drive A commands create robot/quadcopter. for the designed is the,A fuzzy controller is designed to create the drive commands for the robot/quadcopter.,24946
that problems) been scarcely have the seventeen eight studied. of,eight of the seventeen problems) that have been scarcely studied.,24947
"a with the Specifically, method. series corresponds wave an scattered it accelerated-gradient of to expansion","Specifically, it corresponds to a series expansion of the scattered wave with an accelerated-gradient method.",24948
guarantees strongly objects. expansion even convergence the scattering for This,This expansion guarantees the convergence even for strongly scattering objects.,24949
simulated data. The method is on proposed measured both experimentally validated and,The proposed method is validated on both simulated and experimentally measured data.,24950
are domain. Then failure estimated failures probabilities iteratively to domains for original decreasing the conditional,Then iteratively conditional probabilities are estimated for failures domains decreasing to the original failure domain.,24951
"image task. important Thus, recognition challenging understanding and a is view-invariant object","Thus, view-invariant object recognition is a challenging and important image understanding task.",24952
"with objectives discussed, The are for themselves and along listed possible tradeoffs. needs","The objectives themselves are listed and discussed, along with possible needs for tradeoffs.",24953
and classes relate are categorized Manufacturing supported broadly Additive technologies to explicitly features. of devices,Additive Manufacturing technologies are broadly categorized to explicitly relate classes of devices and supported features.,24954
"optimize a analyse, to and In engineering, design, modern simulations systems. computer are popular tool","In modern engineering, computer simulations are a popular tool to analyse, design, and optimize systems.",24955
"important uncertainty engineer's of the an an workflow. efficient Hence, is aspect of quantification","Hence, an efficient quantification of uncertainty is an important aspect of the engineer's workflow.",24956
"In this in the uncertainty variables context, input characterization crucial. is the of","In this context, the characterization of uncertainty in the input variables is crucial.",24957
p-boxes. distributional) called and are of distinguished: types probability-boxes parametric free Two (also,Two types of probability-boxes are distinguished: free and parametric (also called distributional) p-boxes.,24958
In we to the predict focus attention. this work of aim driver's,In this work we aim to predict the driver's focus of attention.,24959
several outperforms learning deep with feature (e.g. loss classification methods Training,Training with classification loss outperforms several deep feature learning methods (e.g.,24960
loss) of with triplet more retrieval applications. networks for typical contrastive Siamese,Siamese networks with contrastive of triplet loss) more typical for retrieval applications.,24961
training requiring data. simple accuracy achieves while also less geolocalization approach state-of-the-art Our significantly,Our simple approach achieves state-of-the-art geolocalization accuracy while also requiring significantly less training data.,24962
Sparse in mapping many methodology has key a scientific been high-dimensional problems.,Sparse mapping has been a key methodology in many high-dimensional scientific problems.,24963
"grouping technique information such since used hidden. this is practice However, usually in limitedly is","However, in practice this technique is used limitedly since such grouping information is usually hidden.",24964
us theorem limit This central to provide associated for averages. a allows the ergodic,This allows us to provide a central limit theorem for the associated ergodic averages.,24965
can strong of pruning achieve. the they %because,%because of the strong pruning they can achieve.,24966
"widely been they adopted However, CSP by not solvers. have","However, they have not been widely adopted by CSP solvers.",24967
two algorithms. of form in This search is approach the presented,This approach is presented in the form of two search algorithms.,24968
demonstrate Experimental method. the benefits results proposed the of,Experimental results demonstrate the benefits of the proposed method.,24969
view represent models genetic lenses through which and macroecological we theory biodiversity. two primary Phylo/population,Phylo/population genetic models and macroecological theory represent two primary lenses through which we view biodiversity.,24970
test intelligence. for Turing artificial was measure the considered long,Turing test was long considered the measure for artificial intelligence.,24971
"in But with be to insufficient the measure. has proved AI, advances it","But with the advances in AI, it has proved to be insufficient measure.",24972
mea- human measure We like aim now sure to intelligence. we can intelligence machine,We can now aim to mea- sure machine intelligence like we measure human intelligence.,24973
widely is math measure accepted of test. standardized of intelligence and the One science,One of the widely accepted measure of intelligence is standardized math and science test.,24974
B-Splines approximate used and in unknown functional data to are regression commonly models forms. smooth,B-Splines are commonly used in regression models to smooth data and approximate unknown functional forms.,24975
and are knot defined order a a sequence. by B-Splines polynomial,B-Splines are defined by a polynomial order and a knot sequence.,24976
for cellular undergoing model a hybrid reorganisation. mechanically We cell present vertex/cell-centred planar monolayers simulating,We present a hybrid vertex/cell-centred model for mechanically simulating planar cellular monolayers undergoing cell reorganisation.,24977
coupled networks constraint progressively. kinematic a The through which relax to we two allow are,The two networks are coupled through a kinematic constraint which we allow to relax progressively.,24978
The the and remodelling the of events. evolution during analysed are forces EPM,The evolution of forces and the EPM are analysed during the remodelling events.,24979
cursive Every and has nature associated different script challenges.,Every cursive script has different nature and associated challenges.,24980
categorized language use. on can Urdu its script they and basis of Arabic We,We can categorized Urdu and Arabic language on basis of its script they use.,24981
"Naskh whereas, Urdu writing. style written Nastaliq follows in is mostly style Arabic of","Urdu is mostly written in Nastaliq style whereas, Arabic follows Naskh style of writing.",24982
and This Dataset database paper offline Urdu handwritten (UNHD). name Urdu-Nastaliq presents Handwritten new comprehensive,This paper presents new and comprehensive Urdu handwritten offline database name Urdu-Nastaliq Handwritten Dataset (UNHD).,24983
"dataset publicly handwritten Currently, researchers. comprehensive is there Urdu for available and no standard","Currently, there is no standard and comprehensive Urdu handwritten dataset available publicly for researchers.",24984
the We popular dataset. on our method Willow tested (iii),(iii) We tested our method on the popular Willow dataset.,24985
the outperform results art. the of Our state,Our results outperform the state of the art.,24986
"manual definition However, a most of involve boundaries. the","However, most involve a manual definition of the boundaries.",24987
"AEI was method templates. GEnI GEI, tested This and on the","This method was tested on the GEI, GEnI and AEI templates.",24988
seems to approach. segmented result when GEI the our best with exhibit,GEI seems to exhibit the best result when segmented with our approach.,24989
the recognition. depict outperforms existing implementations that significantly Experimental gait of view-invariant our results approach,Experimental results depict that our approach significantly outperforms the existing implementations of view-invariant gait recognition.,24990
Method of Multipliers Direction optimization is using problem (ADMM). efficiently The Alternating resultant the solved,The resultant optimization problem is solved efficiently using the Alternating Direction Method of Multipliers (ADMM).,24991
other sustainable very Such over is can exert processes. processing not control cognitive but,Such processing is not very sustainable but can exert control over other cognitive processes.,24992
this also discusses way. regulating of paper This possible benefits cognitive processes,This paper also discusses possible benefits of regulating cognitive processes this way.,24993
Grading extent to spread. of important of cancer is its know the,Grading of cancer is important to know the extent of its spread.,24994
"of is glandular important. grading, to structures segmentation Prior","Prior to grading, segmentation of glandular structures is important.",24995
and bias. subject observer is consuming a Manual time process segmentation is to,Manual segmentation is a time consuming process and is subject to observer bias.,24996
"segment structures. to Hence, process gland required an the is automated","Hence, an automated process is required to segment the gland structures.",24997
a glands variation in texture. show These shape large and size,These glands show a large variation in shape size and texture.,24998
"we show the demonstrate latter, of using the number a To examples. boundaries negative","To show the latter, we demonstrate the boundaries using a number of negative examples.",24999
learn proposes from work a noise-degraded measurements. representations method to new low-dimensional the This the,This work proposes a new method to learn the low-dimensional representations from the noise-degraded measurements.,25000
neural tremendous convolutional on have improvements (CNN) single based on Methods network super-resolution. demonstrated image,Methods based on convolutional neural network (CNN) have demonstrated tremendous improvements on single image super-resolution.,25001
"of a for them model up-scale most each Moreover, train specific factor.","Moreover, most of them train a specific model for each up-scale factor.",25002
"multi-scale this paper, we network. resolution super (MSSR) propose a In","In this paper, we propose a multi-scale super resolution (MSSR) network.",25003
images. kinds helps various in reconstruct property of This HR regions,This property helps reconstruct various kinds of regions in HR images.,25004
"whether that approaches computer-assisted (i.e. pre-specified anatomically-defined examine features, Unfortunately, current","Unfortunately, current computer-assisted approaches that examine pre-specified features, whether anatomically-defined (i.e.",25005
"on thickness) (i.e. based volume, cortical or pixelwise thalamic comparison","thalamic volume, cortical thickness) or based on pixelwise comparison (i.e.",25006
new ontologies mechanism integration propose import for We relations. OWL a semantic of using,We propose a new mechanism for integration of OWL ontologies using semantic import relations.,25007
gone through multi-target tracking Multi-camera yet. transformation fully this has (MCMT) not,Multi-camera multi-target (MCMT) tracking has not fully gone through this transformation yet.,25008
depth work contemporary information provided vast on The majority RGBD by cameras. relevant capitalizes of,The vast majority of contemporary relevant work capitalizes on depth information provided by RGBD cameras.,25009
"for this present foundation a image-based we lighting. In mathematical paper, new","In this paper, we present a new mathematical foundation for image-based lighting.",25010
for lambertian as and such our We derive BRDF's different Phong-like. solution,We derive our solution for different BRDF's such as lambertian and Phong-like.,25011
learn clean We first priors images. of independent natural set from external an,We first learn external priors from an independent set of clean natural images.,25012
performed noisy Extensive image on experiments are datasets. real-world several,Extensive experiments are performed on several real-world noisy image datasets.,25013
to process into partitioning image the easier significant regions segmentation is of the Image analyze.,Image segmentation is the process of partitioning the image into significant regions easier to analyze.,25014
used Hidden Field of is in Random image segmentation. several techniques one Markov model,Hidden Markov Random Field model is one of several techniques used in image segmentation.,25015
process. model provides to elegant the segmentation It an way,It provides an elegant way to model the segmentation process.,25016
modeling to of This function. an minimization the leads objective,This modeling leads to the minimization of an objective function.,25017
best optimization (CG) Conjugate one Gradient algorithm the is known techniques. of,Conjugate Gradient algorithm (CG) is one of the best known optimization techniques.,25018
objective criterion of to as used measure the Coefficient segmentation. is an The Dice quality,The Dice Coefficient is used as an objective criterion to measure the quality of segmentation.,25019
consideration (TPLP) Theory Logic of Programming in and Under . Practice,Under consideration in Theory and Practice of Logic Programming (TPLP) .,25020
address this visual algorithms. on-line we developing paper tracking the In of problem,In this paper we address the problem of developing on-line visual tracking algorithms.,25021
"development and It decouples encouraging algorithms application, re-usability. of","It decouples development of algorithms and application, encouraging re-usability.",25022
classification ordinal possess Probability can produced loss undesired the cross-entropy distributions for by problems properties.,Probability distributions produced by the cross-entropy loss for ordinal classification problems can possess undesired properties.,25023
"by them ignored as well. important are perceptual However, two properties","However, two important perceptual properties are ignored by them as well.",25024
The insight new perception a proposed indicator offers for interpreting visual behaviors.,The proposed indicator offers a new insight for interpreting visual perception behaviors.,25025
proposed for kernel classification. similarity we problem our based In time-series last,In our last problem we proposed similarity kernel for time-series based classification.,25026
graphs work. session complex or literature the of recommendation) based,session based recommendation) or complex graphs of the literature work.,25027
"However, still some of robotic challenges. capsule use endoscopy has the","However, the use of robotic capsule endoscopy still has some challenges.",25028
ECG proposed. in detection recognition of is and based signals abnormalities Computer,Computer based recognition and detection of abnormalities in ECG signals is proposed.,25029
classification techniques used represent a type of commonly applications. special SVM in medical,SVM represent a special type of classification techniques commonly used in medical applications.,25030
departments of of cardiologic Automatic work and make ECG could faster more classification efficient. the,Automatic classification of ECG could make the work of cardiologic departments faster and more efficient.,25031
"as It result, reduce false of diagnosis and, save the also number would lives. a","It would also reduce the number of false diagnosis and, as a result, save lives.",25032
compares techniques qualitatively. well experimentally that show both quantitatively our current approach We to and,We show experimentally that our approach compares well to current techniques both quantitatively and qualitatively.,25033
drug tumors alterations Genetic and initiate enable the resistance. evolution of,Genetic alterations initiate tumors and enable the evolution of drug resistance.,25034
"to more effect, load sensitive drugs. Given genetic should tumors make anticancer its negative","Given its negative effect, genetic load should make tumors more sensitive to anticancer drugs.",25035
"of cancer. genetic might vulnerability unexplored be new, being Beyond load biomarker, a a","Beyond being a biomarker, genetic load might be a new, unexplored vulnerability of cancer.",25036
efficient norm-preservation We dimensionality of consider the randomized guarantees. reduction problem with,We consider the problem of efficient randomized dimensionality reduction with norm-preservation guarantees.,25037
concrete for muscles analysis. selected Ten are,Ten muscles are selected for concrete analysis.,25038
of The attributes discussed. evaluated influence is (fatigability) subject and,The influence of subject attributes (fatigability) is evaluated and discussed.,25039
recognition. unsupervised cross-domain for method This paper visual adaptation a novel domain presents,This paper presents a novel unsupervised domain adaptation method for cross-domain visual recognition.,25040
in a The objective solved efficiently be can closed function form.,The objective function can be solved efficiently in a closed form.,25041
made application networks first several have areas which key We successful summarize inferred in predictions.,We first summarize several key application areas in which inferred networks have made successful predictions.,25042
dimensions presume unphysical. that inferred homogeneous ensemble conventional be can conformational approaches Conformational by a,Conformational dimensions inferred by conventional approaches that presume a homogeneous conformational ensemble can be unphysical.,25043
and We hubs hubs. new in situations: this without with known method two known apply,We apply this new method in two situations: with known hubs and without known hubs.,25044
measures of several Then performance. we DHGL HGL using compare with,Then we compare DHGL with HGL using several measures of performance.,25045
through the of extensive empirical real-world validation on We REMIX data. benefits demonstrate,We demonstrate the benefits of REMIX through extensive empirical validation on real-world data.,25046
"Actually, could work. this which improve practice algorithms no and there automate are clinical","Actually, there are no clinical practice algorithms which could improve and automate this work.",25047
generality due to and in power. popular discriminating their Kernel clustering methods are,Kernel methods are popular in clustering due to their generality and discriminating power.,25048
"methods, also clustering extend our kernel other We popular analysis e.g. to","We also extend our analysis to other popular kernel clustering methods, e.g.",25049
"sets, dominant cut biases take or different where density average/normalized can forms.","average/normalized cut or dominant sets, where density biases can take different forms.",25050
"objectives kernel many popular equivalent. equalization clustering density makes Moreover,","Moreover, density equalization makes many popular kernel clustering objectives equivalent.",25051
data synthetic illustrate experiments density biases and proposed real solutions. Our and,Our synthetic and real data experiments illustrate density biases and proposed solutions.,25052
natively overcome understand later to law. the algebra of The slowdown should Moore's linear,The later should understand linear algebra natively to overcome the slowdown of Moore's law.,25053
of method ex-vivo The porcine is stomach model. performance using an demonstrated the,The performance of the method is demonstrated using an ex-vivo porcine stomach model.,25054
(HFIT). This fuzzy a hierarchical paper tree of design proposes inference,This paper proposes a design of hierarchical fuzzy inference tree (HFIT).,25055
Such approximation high provides accuracy. structure a of a hierarchical degree natural,Such a natural hierarchical structure provides a high degree of approximation accuracy.,25056
construction The in two phases. of takes HFIT place,The construction of HFIT takes place in two phases.,25057
"Secondly, is algorithm applied obtained tree's to parameters. optimize differential the evolution the","Secondly, the differential evolution algorithm is applied to optimize the obtained tree's parameters.",25058
through features These extensive demonstrated vividly evaluations. experimental are,These features are vividly demonstrated through extensive experimental evaluations.,25059
in use Thisallows semi-supervisedsetting. unlabeled the of training during data a,Thisallows the use of unlabeled data during training in a semi-supervisedsetting.,25060
reconstruction or only with is super-resolution on MS images dates. different pansharpening) used (SRR) captured,pansharpening) or super-resolution reconstruction (SRR) is used with only MS images captured on different dates.,25061
method deriving subband volumes. proposed SRR is the by multitemporal obtained MS between The relations,The proposed SRR method is obtained by deriving the subband relations between multitemporal MS volumes.,25062
underlying and its observations. to model able new tracker adapt must modify A be to,A tracker must be able to modify its underlying model and adapt to new observations.,25063
characteristics. The feature persons identity center the have with to similar loss same engaged encourages,The engaged center loss encourages persons with the same identity to have similar feature characteristics.,25064
"text. recent document recognition scanned remarkable achieved years, success in In has text recognizing","In recent years, text recognition has achieved remarkable success in recognizing scanned document text.",25065
"detection end-to-end intensive The producing post-processing, technique an proposed system. word avoids","The proposed technique avoids intensive post-processing, producing an end-to-end word detection system.",25066
task. convolutional Networks (CNN) state-of-the-art Deep detection are the object Neural performers for,Deep convolutional Neural Networks (CNN) are the state-of-the-art performers for object detection task.,25067
It more detection than known object that computation and well classification. is memory requires image,It is well known that object detection requires more computation and memory than image classification.,25068
object more an is CNN-based for a consolidation detection system the challenging. of embedded Thus,Thus the consolidation of a CNN-based object detection for an embedded system is more challenging.,25069
We model. and develop end-to-end design TensorFlow(TF)-based an,We design and develop an end-to-end TensorFlow(TF)-based model.,25070
detection face use a case. We as use,We use face detection as a use case.,25071
The becomes thus model embedded amenable for proposed implementations.,The proposed model thus becomes amenable for embedded implementations.,25072
R to refinement Then accurate full-resolution a B images. performed process interpolate and is,Then a refinement process is performed to interpolate accurate full-resolution R and B images.,25073
a comprehensive RealSense of present We systems. RGBD Intel overview imaging the stereoscopic,We present a comprehensive overview of the stereoscopic Intel RealSense RGBD imaging systems.,25074
recent article a Building (James R.~Shaw. on,Building on a recent article (James R.~Shaw.,25075
se} belief rational {\em choice. De and,{\em De se} belief and rational choice.,25076
conclude speculate and the I might. not about settle this that necessarily issue does what,I conclude that this does not necessarily settle the issue and speculate about what might.,25077
completely positioning from exploration one. the The conventional of different is,The positioning of exploration is completely different from the conventional one.,25078
the exploration inside produces dynamics factors The by ChNN itself. chaotic,The chaotic dynamics inside the ChNN produces exploration factors by itself.,25079
"learning from method completely different Therefore, also the is the one. conventional","Therefore, the learning method is also completely different from the conventional one.",25080
"trace Using the the weight TD error, updated. is and","Using the trace and TD error, the weight is updated.",25081
times. fields multiple are normal) surface differentiated,surface normal) fields are differentiated multiple times.,25082
"the In tackle approach problem. this paper, we our present to","In this paper, we present our approach to tackle the problem.",25083
"of Golovin, named advantage Our domain. game takes limited the agent,","Our agent, named Golovin, takes advantage of the limited game domain.",25084
"datasets: four and GPDS, datasets. CEDAR were Extensive PUC-PR conducted on experiments Brazilian MCYT,","Extensive experiments were conducted on four datasets: GPDS, MCYT, CEDAR and Brazilian PUC-PR datasets.",25085
collection line roughness as Our segments. a expresses of surface model,Our model expresses surface roughness as a collection of line segments.,25086
variables the this to non-informative statistical distributions. quantity For well known conforms,For non-informative variables this quantity conforms to the well known statistical distributions.,25087
non-informative truly informative ones. This variables allows from for discerning,This allows for discerning truly informative variables from non-informative ones.,25088
"with the localization refined regularization. are Finally, results sparsity","Finally, the localization results are refined with sparsity regularization.",25089
accuracy. other method of outperforms Our methods terms in localization state-of-the-art,Our method outperforms other state-of-the-art methods in terms of localization accuracy.,25090
are normal) surface fields multiple differentiated times.,surface normal) fields are differentiated multiple times.,25091
that unknown encapsulates problem. the It this in ambiguity the matrix is,It is this unknown matrix that encapsulates the ambiguity in the problem.,25092
as often optimization quadratic are permutations. problems problems Correspondence modelled over,Correspondence problems are often modelled as quadratic optimization problems over permutations.,25093
"consider source of an supervision -- a i.e. we -- this alternate In paper, ""peer""","In this paper, we consider an alternate source of supervision -- a ""peer"" -- i.e.",25094
is of online unsupervised an objectives achieved adaptive The clustering and using algorithm. identification,The identification of objectives is achieved using an online and unsupervised adaptive clustering algorithm.,25095
least parallel objectives learned (at using identified The are in partially) Q-learning.,The identified objectives are learned (at least partially) in parallel using Q-learning.,25096
briefly also work extensions and implications The this possible discussed. of are,The implications and possible extensions of this work are also briefly discussed.,25097
"further as the which most HDR capture prior input process. work images complicates Moreover, requires","Moreover, most prior work requires HDR images as input which further complicates the capture process.",25098
a color-constancy design further and ambiguity. resolve function We loss the scale to,We further design a loss function to resolve the color-constancy and scale ambiguity.,25099
is for reflectance estimation. benchmark the large-scale SynBRDF first dataset,SynBRDF is the first large-scale benchmark dataset for reflectance estimation.,25100
baselines outperforms method on datasets. both consistently other Our,Our method consistently outperforms other baselines on both datasets.,25101
"efficient still this dimensions. not in high However, be may","However, this still may not be efficient in high dimensions.",25102
"accelerated strategy inference an approximate parallel. perform Next, propose to we inference in","Next, we propose an approximate inference strategy to perform accelerated inference in parallel.",25103
"capsule endoscopy robotic challenges. However, some still has","However, robotic capsule endoscopy still has some challenges.",25104
prediction problem a time regression series We as SST formulate problem. the,We formulate the SST prediction problem as a time series regression problem.,25105
relationship. time series model the utilized is layer to LSTM,LSTM layer is utilized to model the time series relationship.,25106
a to of prediction. Full-connected map layer layer utilized final output LSTM the to is,Full-connected layer is utilized to map the output of LSTM layer to a final prediction.,25107
"show updated addition, its also online we characteristics. In","In addition, we also show its online updated characteristics.",25108
(ASP) formalism is Answer combinatorial modeling Set a problems. Programming for powerful,Answer Set Programming (ASP) is a powerful modeling formalism for combinatorial problems.,25109
"trivial. models is However, ASP writing not","However, writing ASP models is not trivial.",25110
uncertain writes an marks. user with parts ASP open program while The marking question,The user writes an ASP program while marking uncertain parts open with question marks.,25111
"The sketched is model solved rewritten into ASP traditional which is program, another methods. by","The sketched model is rewritten into another ASP program, which is solved by traditional methods.",25112
"user and the As a problem. ASP result, modelling her program reusable a obtains functional","As a result, the user obtains a functional and reusable ASP program modelling her problem.",25113
action video Kinetics the DeepMind describe dataset. human We,We describe the DeepMind Kinetics human action video dataset.,25114
high-dimensional recent data. using papers for strategies LDA outlined exploiting sparsity have Several with for,Several recent papers have outlined strategies for exploiting sparsity for using LDA with high-dimensional data.,25115
"However, for scalable the methods lack problems. solution underlying of many optimization","However, many lack scalable methods for solution of the underlying optimization problems.",25116
local the surface learn We profiles. a employ approach regional to,We employ a regional approach to learn the local surface profiles.,25117
"shape, well-defined drawings'' certain patches limit. in a a For given ""line shading approach","For a given shape, certain shading patches approach ""line drawings'' in a well-defined limit.",25118
"the features desired metrics. produced discriminative signals, training back-propagating By RNN with network our","By back-propagating the training signals, our RNN network produced discriminative features with desired metrics.",25119
geometry the the relevant of also vertebrae. of approach enables The analysis the,The approach also enables the analysis of the geometry of the relevant vertebrae.,25120
approach goal is the in The of fracture to improve long-term prediction osteoporosis.,The long-term goal of the approach is to improve fracture prediction in osteoporosis.,25121
such transport example Cellular is proteins behavior. by collective motor of an,Cellular transport by motor proteins is an example of such collective behavior.,25122
processes. multi-particle microscopic clarifies complex Our theoretical analysis of dynamic aspects,Our theoretical analysis clarifies microscopic aspects of complex multi-particle dynamic processes.,25123
computationally well to algorithm large scales data efficient and sets. Our is,Our algorithm is computationally efficient and scales well to large data sets.,25124
Affine-Gradient present classification. We novel based Pattern Local for descriptor Binary (AGLBP) texture a,We present a novel Affine-Gradient based Local Binary Pattern (AGLBP) descriptor for texture classification.,25125
constrain bending image are Their points. to used based characteristics the feature candidate,Their bending characteristics are used to constrain the image feature based candidate points.,25126
"action However, straightforwardly space discrete. approach not apply when is the does this","However, this approach does not apply straightforwardly when the action space is discrete.",25127
that We scheme this show improves the performance. training,We show that this scheme improves the training performance.,25128
orientation feature computed dMRI is fiber (FO) from reconstruction. key The fiber tract for a,The fiber orientation (FO) is a key feature computed from dMRI for fiber tract reconstruction.,25129
"dictionary signals. First, FOs represent smaller we the use a encoding basis diffusion to coarse","First, we use a smaller dictionary encoding coarse basis FOs to represent the diffusion signals.",25130
"reduce is training. used dictionary Here, the cost the of computational smaller to","Here, the smaller dictionary is used to reduce the computational cost of training.",25131
the on is classification pruning compression to The with based accuracy. filters contribution least the,The compression is based on pruning filters with the least contribution to the classification accuracy.,25132
become recently vision Subspace data a many tasks. practice in common computer has representation,Subspace data representation has recently become a common practice in many computer vision tasks.,25133
classical data. algorithms for demands generalizing It machine subspace learning,It demands generalizing classical machine learning algorithms for subspace data.,25134
This possibility LRR for extending Grassmann subspace explores paper data the manifolds. of on,This paper explores the possibility of extending LRR for subspace data on Grassmann manifolds.,25135
algorithm A for is solving proposed the novel and implemented. model investigated,A novel algorithm for solving the proposed model is investigated and implemented.,25136
method new The results outperforms methods number of experimental . state-of-the-art show a the clustering,The experimental results show the new method outperforms a number of state-of-the-art clustering methods .,25137
real-world framework proposed synthetic datasets. and using assessed is The,The proposed framework is assessed using synthetic and real-world datasets.,25138
for a The path shortest is planning. highly (SSP) problem probabilistic model stochastic expressive,The stochastic shortest path problem (SSP) is a highly expressive model for probabilistic planning.,25139
"employ existing a methods to determinization. However, simplistic approach","However, existing methods employ a simplistic approach to determinization.",25140
their by other functions interacting Proteins usually with perform proteins.,Proteins perform their functions usually by interacting with other proteins.,25141
fundamental problem. proteins a interact Predicting which is,Predicting which proteins interact is a fundamental problem.,25142
"high expensive, and rate methods Experimental error. a slow, are of have","Experimental methods are slow, expensive, and have a high rate of error.",25143
have been very ones Many promising. computational among proposed sequence-based are methods which,Many computational methods have been proposed among which sequence-based ones are very promising.,25144
can the is that SPRINT entire program interactome. only human the predict,SPRINT is the only program that can predict the entire human interactome.,25145
of the but mostly input restricting the under limitation to images. binary,but mostly under the limitation of restricting the input to binary images.,25146
attention. the human models to system better helps Knowledge of visual visual of develop computational,Knowledge of the human visual system helps to develop better computational models of visual attention.,25147
adapted to age used saliency These models. results were develop,These results were used to develop age adapted saliency models.,25148
mechanical Realistic of dynamics of description clotting a properties presents challenge. patient-specific major,Realistic description of patient-specific mechanical properties of clotting dynamics presents a major challenge.,25149
datasets better (ii) than Are datasets? wider deeper,(ii) Are deeper datasets better than wider datasets?,25150
noise networks? deep (iii) improvement in adding of performance Does lead to label,(iii) Does adding label noise lead to improvement in performance of deep networks?,25151
(iv) Is alignment needed face recognition? for,(iv) Is alignment needed for face recognition?,25152
"main volume are morphological operations. and Its growing, models, deformable components","Its main components are deformable models, volume growing, and morphological operations.",25153
density for clustering. present hierarchical an accelerated We based algorithm,We present an accelerated algorithm for hierarchical density based clustering.,25154
the clustering. default density based HDBSCAN* This choice for accelerated makes,This makes accelerated HDBSCAN* the default choice for density based clustering.,25155
regression multiple learning classification multi-task with objectives. applications learning and Numerous deep benefit from,Numerous deep learning applications benefit from multi-task learning with multiple regression and classification objectives.,25156
"that demonstrating we Finally, results substantially Deep-LK show impressive GOTURN. outperforms","Finally, we show impressive results demonstrating that Deep-LK substantially outperforms GOTURN.",25157
conservative telosome mechanism. the segregates hypothesize that by We a,We hypothesize that the telosome segregates by a conservative mechanism.,25158
These presence an data chromatin active the for mechanism in suggest of yeast. segregation,These data suggest the presence of an active mechanism for chromatin segregation in yeast.,25159
is parsing research increasing attention. Human attracting,Human parsing is attracting increasing research attention.,25160
in conjunctive show in OBDA complexity. semantics answering data query makes We coNP-hard that bag,We show that bag semantics makes conjunctive query answering in OBDA coNP-hard in data complexity.,25161
"than as lines converging In or this horizontal. mortar misperceived diverging are the illusion, rather","In this illusion, the mortar lines are misperceived as diverging or converging rather than horizontal.",25162
For model a quantitative first predictions for of this time reported are the stimulus.,For the first time quantitative predictions of a model are reported for this stimulus.,25163
directly provided by gating signal The depth input. or disparity stereo from monocular estimated is,The depth gating signal is provided by stereo disparity or estimated directly from monocular input.,25164
perform segmentation. convolutional gating semantic depth-aware this integrate to network into a We recurrent neural,We integrate this depth-aware gating into a recurrent convolutional neural network to perform semantic segmentation.,25165
"acceleration be this used hardly for unstructured. can since sparsity it However, is","However, this sparsity can hardly be used for acceleration since it is unstructured.",25166
CNNs. neurons and/or removes channels convolutional in,removes neurons and/or convolutional channels in CNNs.,25167
the this unregularized. we while noise do to outputs To weights inject neurons keeping the,To do this we inject noise to the neurons outputs while keeping the weights unregularized.,25168
The be among graph may features. nodal or given measures constructed using similarity,The graph may be given or constructed using similarity measures among nodal features.,25169
waggle We segment automated honeybee and propose of monitor an different components to dance. process,We propose an automated process to monitor and segment different components of honeybee waggle dance.,25170
(CS) is an Compressive Imaging effective fast Magnetic sensing for approach (MRI). Resonance,Compressive sensing (CS) is an effective approach for fast Magnetic Resonance Imaging (MRI).,25171
output MR as reconstructed inputs sampled images. the and take They data k-space,They take the sampled k-space data as inputs and output reconstructed MR images.,25172
"MR cope our extend images. we complex-valued Moreover, network to with","Moreover, we extend our network to cope with complex-valued MR images.",25173
parameter. often distributional regularization the uncertainty shown size to is to The correspond,The distributional uncertainty size is often shown to correspond to the regularization parameter.,25174
corresponds used distributional of norm regularize) to the the shape the to uncertainty.,the norm used to regularize) corresponds to the shape of the distributional uncertainty.,25175
theoretical several contributions. key paper presents The,The paper presents several key theoretical contributions.,25176
uncertainty centered distributional The a at neighborhood the distribution. empirical is defined as,The distributional uncertainty is defined as a neighborhood centered at the empirical distribution.,25177
We propose which data-driven learns a way. methodology such a natural in neighborhood,We propose a methodology which learns such neighborhood in a natural data-driven way.,25178
framework particular regularization our show as We case. that adaptive rigorously encompasses a,We show rigorously that our framework encompasses adaptive regularization as a particular case.,25179
transfer  to images.  visual arbitrary styles Universal to aims content style transfer,  Universal style transfer aims to transfer arbitrary visual styles to content images.,25180
We neural from a convolutional network lesion labels. detection propose for weak novel,We propose a novel convolutional neural network for lesion detection from weak labels.,25181
"work how Consequently, this into to tracking detectors two fuse a system. demonstrates","Consequently, this work demonstrates how to fuse two detectors into a tracking system.",25182
"only solution such As are the can NP-hard, approximated. problems be","As such problems are NP-hard, the solution can only be approximated.",25183
"this framework, energy SHBG between our and glucose interaction uncovered calculations a In surface. remarkable","In this framework, our calculations uncovered a remarkable interaction energy between glucose and SHBG surface.",25184
"yielding components movement Surprisingly, observed toward clusters the we surrounding protein. SHBG solute also","Surprisingly, we also observed solute components movement toward SHBG yielding clusters surrounding the protein.",25185
aerialphotogrammetry powerful We per-point from a method to labels extract semantic data. class present,We present a powerful method to extract per-point semantic class labels from aerialphotogrammetry data.,25186
makes accuracies high a propose accuracy that achieving quite patch-and-scan measurement challenging. We compound,We propose a compound patch-and-scan accuracy measurement that makes achieving high accuracies quite challenging.,25187
datasets. also achieve performance good on real-world Our models,Our models also achieve good performance on real-world datasets.,25188
"variation a algorithm. Decomposition the (TSVD) classical We Truncated of Singular Value propose simple ReFACTor,","We propose ReFACTor, a simple variation of the classical Truncated Singular Value Decomposition (TSVD) algorithm.",25189
and in popular modelling. financial class This rather has of mixtures statistical recently become,This class of mixtures has recently become rather popular in statistical and financial modelling.,25190
the performance and illustrate on of procedure data. We our real simulated,We illustrate the performance of our procedure on simulated and real data.,25191
networks We the indicators. with such hybrid of a variations trained of neural series,We trained a series of such hybrid neural networks with variations of the indicators.,25192
as inter-class actions. introduced of for classification measure congruence recently of a was Self-similarity,Self-similarity was recently introduced as a measure of inter-class congruence for classification of actions.,25193
"investigate the we action classification dual of intra-class dissimilarity problem for Herein, of styles.","Herein, we investigate the dual problem of intra-class dissimilarity for classification of action styles.",25194
"one invariant is in The engineering. topics technology of central and science,","The invariant is one of central topics in science, technology and engineering.",25195
differential invariants or usually complicated. difficult of The derivation is,The derivation of differential invariants is usually difficult or complicated.,25196
Projective can also in found this with invariants way process. screening differential a be,Projective differential invariants can also be found in this way with a screening process.,25197
"guide (GAC). paper, we In a this actor-critic method actor-critic novel called propose the","In this paper, we propose a novel actor-critic method called the guide actor-critic (GAC).",25198
Our two are theoretical contributions folds. main,Our main theoretical contributions are two folds.,25199
among origins great open life our The stands time. scientific questions of of the,The origins of life stands among the great open scientific questions of our time.,25200
We weakly other methods. outperforming supervised achieve results,We achieve results outperforming other weakly supervised methods.,25201
"plasticity. activity should their feedbacks First, receive top-down driving their and they both synaptic","First, they should receive top-down feedbacks driving both their activity and their synaptic plasticity.",25202
"between maintain a interneurons tight excitation circuit. should balance Second, in the inhibition and inhibitory","Second, inhibitory interneurons should maintain a tight balance between excitation and inhibition in the circuit.",25203
have size with observed been networks. neuronal Avalanches power-law in distributed parameters,Avalanches with power-law distributed size parameters have been observed in neuronal networks.,25204
of be This the criticality manifestation self-organized a might observation (SOC).,This observation might be a manifestation of the self-organized criticality (SOC).,25205
"mechanicsm is physiological Yet, unknown. currently behavior this the of","Yet, the physiological mechanicsm of this behavior is currently unknown.",25206
"corruptions errors. with real-world or many noises, applications, data image large often come In","In many real-world applications, image data often come with noises, corruptions or large errors.",25207
"application, image task. CRT to an As we our recognition apply","As an application, we apply our CRT to image recognition task.",25208
"labels are adaptation unsupervised In assumes domain. target particular, the in domain available no","In particular, unsupervised domain adaptation assumes no labels are available in the target domain.",25209
exploits theoretically resulting function loss The manifold. on such geodesic distance sound a,The resulting loss function exploits a theoretically sound geodesic distance on such manifold.,25210
Results of the distance. show indeed Euclidean suboptimal the nature,Results show indeed the suboptimal nature of the Euclidean distance.,25211
used to The aesthetically approach appealing patterns. classify is proposed,The proposed approach is used to classify aesthetically appealing patterns.,25212
"agent on synchronization niche. the primarily sentient Therefore, its and approximate relies dynamics, such between","Therefore, such sentient dynamics, relies primarily on approximate synchronization between the agent and its niche.",25213
"with theory Finally, experiments. some we numerical complement our","Finally, we complement our theory with some numerical experiments.",25214
vision Computer from calories introduced estimate to images. been has food,Computer vision has been introduced to estimate calories from food images.,25215
"orderings. result in depth the that we depth and demonstrate segmentation superior Qualitatively, edges","Qualitatively, we demonstrate that the depth edges result in superior segmentation and depth orderings.",25216
"have with methods to solve problems. priors hand-crafted such been iterative used along optimization Traditionally,","Traditionally, hand-crafted priors along with iterative optimization methods have been used to solve such problems.",25217
and demonstrate approach performed the significance. validate its experiments We have proposed extensive to,We have performed extensive experiments to validate the proposed approach and demonstrate its significance.,25218
general assumptions discuss the some We underlying project.,We discuss some general assumptions underlying the project.,25219
present to belief Specifically DS we methodology problem. application revision the of,Specifically we present the application of DS methodology to belief revision problem.,25220
"agent. far intelligent However, a we from are constructing still generally","However, we are still far from constructing a generally intelligent agent.",25221
"unknown and optimally does one explore How learn in environments? general,","How does one explore and learn optimally in general, unknown environments?",25222
"general the sense? optimal in does be in fact, it mean What, to","What, in fact, does it mean to be optimal in the general sense?",25223
"present We a JavaScript these of implementation AIXIjs, GRL agents.","We present AIXIjs, a JavaScript implementation of these GRL agents.",25224
with well learning constrained The objects highly based captured can nature of anatomical be techniques.,The highly constrained nature of anatomical objects can be well captured with learning based techniques.,25225
"label non-linear learned of structure) shape, representations shape. the via","shape, label structure) via learned non-linear representations of the shape.",25226
(e.g. approach different show to tasks We that proposed analysis easily can be adapted the,We show that the proposed approach can be easily adapted to different analysis tasks (e.g.,25227
"improve image the enhancement, the and accuracy models. state-of-the-art of segmentation) prediction","image enhancement, segmentation) and improve the prediction accuracy of the state-of-the-art models.",25228
benchmarks. on The applicability public multi-modal shown and cardiac our datasets approach of is,The applicability of our approach is shown on multi-modal cardiac datasets and public benchmarks.,25229
"detection salient we integrating In \textit{semantic paper, object this the into propose process. priors}","In this paper, we propose integrating \textit{semantic priors} into the salient object detection process.",25230
consists basic steps. algorithm of three Our,Our algorithm consists of three basic steps.,25231
"two datasets, framework on proposed HKUIS. challenging further the evaluate and ECSSD namely, We","We further evaluate the proposed framework on two challenging datasets, namely, ECSSD and HKUIS.",25232
demonstrate that methods. The our experimental extensive outperforms other results method state-of-the-art,The extensive experimental results demonstrate that our method outperforms other state-of-the-art methods.,25233
update probabilities. rule with Cromwell's how of at odds humans our intuition appears,Cromwell's rule appears at odds with our intuition of how humans update probabilities.,25234
"approach maximising is an optimal utility. design the with adopted, A expected decision-theoretic","A decision-theoretic approach is adopted, with the optimal design maximising an expected utility.",25235
previously functionality demonstrate acebayes. unseen of examples These,These examples demonstrate previously unseen functionality of acebayes.,25236
"activations. Pairwise the Our (PC) by in overfitting reduces Confusion procedure, confusion} intentionally {introducing called","Our procedure, called Pairwise Confusion (PC) reduces overfitting by intentionally {introducing confusion} in the activations.",25237
"analysis tracking tasks, intensive like and require swift computationally execution of algorithms. visual recognition, Real-time","Real-time visual analysis tasks, like tracking and recognition, require swift execution of computationally intensive algorithms.",25238
improved a which propose gives calibrated performance We better new uncertainties. variant and dropout,We propose a new dropout variant which gives improved performance and better calibrated uncertainties.,25239
closure consistent notions Two a of introduced. fuzzy relation of are,Two notions of consistent closure of a fuzzy relation are introduced.,25240
dataset paper a of Visual Atomic video introduces localized Actions spatio-temporally This (AVA).,This paper introduces a video dataset of spatio-temporally localized Atomic Visual Actions (AVA).,25241
"recognition. and action with complexity, exposes the AVA, intrinsic action scene realistic its of difficulty","AVA, with its realistic scene and action complexity, exposes the intrinsic difficulty of action recognition.",25242
"Inception-ResNet, methods with Inception-ResNet are: These Shallow Deep Inception-ResNet, and LSTMs.","These methods are: Shallow Inception-ResNet, Deep Inception-ResNet, and Inception-ResNet with LSTMs.",25243
"learning a when representation. have consensus multi-view clustering, confidence views In may different different levels","In multi-view clustering, different views may have different confidence levels when learning a consensus representation.",25244
assigning Existing methods usually to address this distinctive by weights different views.,Existing methods usually address this by assigning distinctive weights to different views.,25245
solutions. a unified lead to a Thus weight view considering may suboptimal for,Thus considering a unified weight for a view may lead to suboptimal solutions.,25246
developed algorithm with An iterative a convergence efficient guarantee. is,An efficient iterative algorithm is developed with a convergence guarantee.,25247
the correctness benchmarks of effectiveness demonstrate Experimental and model. on four results proposed the,Experimental results on four benchmarks demonstrate the correctness and effectiveness of the proposed model.,25248
in effective extremely Our is simple video model sequence terms but of labeling.,Our model is simple but extremely effective in terms of video sequence labeling.,25249
"interpretable novel and that potential, methods statistical however, are requires Realizing analysis both predictive. this","Realizing this potential, however, requires novel statistical analysis methods that are both interpretable and predictive.",25250
railroad to image captcha of this system. China's approach applied We successfully solve the,We applied this approach to successfully solve the image captcha of China's railroad system.,25251
suspicious The coincidence. principle based approach on is of the,The approach is based on the principle of suspicious coincidence.,25252
images their train in They purchase tickets. select the order must relevant quickly to,They must quickly select the relevant images in order to purchase their train tickets.,25253
"can edges Furthermore, remain guarantee protective that present. actions will","Furthermore, protective actions can guarantee that edges will remain present.",25254
on the Previous edges independent. are approaches assumption rely that,Previous approaches rely on the assumption that edges are independent.,25255
which on package in The is method an available implemented github. is R,The method is implemented in an R package which is available on github.,25256
FRK large with software spatial/spatio-temporal R an and datasets. package modelling is for prediction,FRK is an R software package for spatial/spatio-temporal modelling and prediction with large datasets.,25257
dimension. another temporal The is component adding included by,The temporal component is included by adding another dimension.,25258
rates training obtained intermediately events. of represent abnormal accuracy classifiers the Higher,Higher training accuracy rates of the intermediately obtained classifiers represent abnormal events.,25259
"activity. brain also predictors that this, discover we branch can predict achieve To","To achieve this, we discover that branch predictors can also predict brain activity.",25260
have systems been computer scientists. inspiration contemporary an abundant to Immunological,Immunological systems have been an abundant inspiration to contemporary computer scientists.,25261
sampling. systems slow suffers from on real reinforcement (RL) deep Applying learning data,Applying deep reinforcement learning (RL) on real systems suffers from slow data sampling.,25262
"paper, for reinforcement an In learning. implicit generic reward propose this we model","In this paper, we propose an implicit generic reward model for reinforcement learning.",25263
comes deviation agents' previous from experiences. the the only It from,It only comes from the deviation from the agents' previous experiences.,25264
"for of is In clustering, partitional k-means this. important very a the algorithm unsupervised realm","In the realm of unsupervised partitional clustering, k-means is a very important algorithm for this.",25265
from can't re-estimation are which Observations step. firmly clusters identified into excluded the be,Observations which can't be firmly identified into clusters are excluded from the re-estimation step.,25266
datasets. real simulated on We both demonstrate and this,We demonstrate this on both simulated and real datasets.,25267
with in and this Python available algorithm is Our will report. implemented be,Our algorithm is implemented in Python and will be available with this report.,25268
We our consciousness. embodiment and on intelligence speculate in the own factors role driving of,We speculate on the role of embodiment factors in driving our own intelligence and consciousness.,25269
"extensively studied been of video the recognition has recently. Based on recognition, progress image","Based on the progress of image recognition, video recognition has been extensively studied recently.",25270
"and images image-level easier web In contrast, and cheaper labels to their are obtain. much","In contrast, web images and their image-level labels are much easier and cheaper to obtain.",25271
"images first We retrieve e.g. engines, search from","We first retrieve images from search engines, e.g.",25272
"as class queries, and names Flickr e.g. semantic using Google,","Flickr and Google, using semantic class names as queries, e.g.",25273
for present We framework lung a deep learning diagnosis. computer-aided cancer,We present a deep learning framework for computer-aided lung cancer diagnosis.,25274
our framework. discuss of We the and advantages challenges,We discuss the challenges and advantages of our framework.,25275
process two fast-forwarding enhances provided path training ways. parallel in The the by,The parallel path provided by fast-forwarding enhances the training process in two ways.,25276
"considerable learning attracted equivariant research representations Recently, attention. has","Recently, learning equivariant representations has attracted considerable research attention.",25277
One is creation the main of challenges models. of speaker the the,One of the main challenges is the creation of the speaker models.,25278
proposed verification the significantly the d-vector that system. traditional demonstrate We method outperforms,We demonstrate that the proposed method significantly outperforms the traditional d-vector verification system.,25279
emerged standard a Hamiltonian as Monte Carlo has for tool posterior computation.,Hamiltonian Monte Carlo has emerged as a standard tool for posterior computation.,25280
is affinities Calculating expensive. possible computationally all order higher,Calculating all possible higher order affinities is computationally expensive.,25281
a only in is most cases Hence subset used.,Hence in most cases only a subset is used.,25282
given is a from used tool commonly sampling distribution. for Langevin a diffusion,Langevin diffusion is a commonly used tool for sampling from a given distribution.,25283
when rate also absent. the We the strong-convexity assumption convergence study is,We also study the convergence rate when the strong-convexity assumption is absent.,25284
experiments areas. Randomized to assist have decision-making many used been in,Randomized experiments have been used to assist decision-making in many areas.,25285
population guarantee. with people the treatment test help certain the They select statistical optimal for,They help people select the optimal treatment for the test population with certain statistical guarantee.,25286
"in significant can subjects heterogeneity to However, response treatments. show","However, subjects can show significant heterogeneity in response to treatments.",25287
for modeling key A data feature that uplift the is is unlabeled.,A key feature for uplift modeling is that the data is unlabeled.,25288
training the of both challenge evaluation a models. to This the uplift and presents,This presents a challenge to both the training and the evaluation of uplift models.,25289
forest present a We which uplift of creates randomized algorithm trees. new a,We present a new uplift algorithm which creates a forest of randomized trees.,25290
Our of gracefully. number that scales needed trajectories show the analysis,Our analysis show that the number of trajectories needed scales gracefully.,25291
investigate We representations of human the action recognition. by task such motivated,We investigate such representations motivated by the task of human action recognition.,25292
"result scanning by the flight constrained quadcopters. of this time restricted is setting, In the","In this setting, the scanning result is constrained by the restricted flight time of quadcopters.",25293
of region to the to interest can tool and Our used specify trajectories. be plan,Our tool can be used to specify the region of interest and to plan trajectories.,25294
from of images. structured Interpretation Image the task Semantic extracting descriptions (SII) is semantic,Semantic Image Interpretation (SII) is the task of extracting structured semantic descriptions from images.,25295
a standard The is approach on image evaluated processing proposed benchmark.,The proposed approach is evaluated on a standard image processing benchmark.,25296
"In a of scene surgical proposed. dense reconstruction quasi algorithm this is paper,","In this paper, a quasi dense reconstruction algorithm of surgical scene is proposed.",25297
"this person dataset To we construct natural end, for language benchmark a first retrieval.","To this end, we first construct a benchmark dataset for natural language person retrieval.",25298
Faster adopt then a We a generator. candidate region proposal network region R-CNN in as,We then adopt a region proposal network in Faster R-CNN as a candidate region generator.,25299
normalized is by correlation finding correspondences. cross widely Template matching used (NCC) for image,Template matching by normalized cross correlation (NCC) is widely used for finding image correspondences.,25300
quantified is from images patches microscopy. The of serial improvement using section electron brain,The improvement is quantified using patches of brain images from serial section electron microscopy.,25301
"significantly Relative networks to a parameter-tuned bandpass filter, siamese false matches. convolutional reduce","Relative to a parameter-tuned bandpass filter, siamese convolutional networks significantly reduce false matches.",25302
their well-known difficult. However to be is training,However their training is well-known to be difficult.,25303
We transformations self-labeling. them combining image bi-directional jointly optimize target with,We jointly optimize bi-directional image transformations combining them with target self-labeling.,25304
"or deep are In methods than cases, equivalent often these traditional to methods. learning better","In these cases, traditional methods are often better than or equivalent to deep learning methods.",25305
"network a Therefore, residual deep train classes. surrogate using we","Therefore, we train a deep residual network using surrogate classes.",25306
The as CNN from penultimate activations classification serve for subsequent tasks. layer features the,The activations from the penultimate CNN layer serve as features for subsequent classification tasks.,25307
We evaluate on datasets. two representations the available feature publicly,We evaluate the feature representations on two publicly available datasets.,25308
parametric families consider line. real on We of $K$ exponential the dimension,We consider parametric exponential families of dimension $K$ on the real line.,25309
"were the forgotten in and apparently community. Lai, bandit","Lai, and were apparently forgotten in the bandit community.",25310
"example, optimization when host a genes This heterologous organism. problem important, re-engineering in is for","This optimization problem is important, for example, when re-engineering heterologous genes in a host organism.",25311
of the and describe of aspects results We to our translation. implications other this,We describe the implications of our results to this and other aspects of translation.,25312
"transmission, under that find the We assumption frequency-dependent of i.e.","We find that under the assumption of frequency-dependent transmission, i.e.",25313
"semi-automated paper, we for cardiac multitype method In propose this indices a estimation.","In this paper, we propose a semi-automated method for multitype cardiac indices estimation.",25314
function These in the advantages clinical assessment. cardiac a potential proposed great endow method,These advantages endow the proposed method a great potential in clinical cardiac function assessment.,25315
In this time. different we route question a of paper to rather take,In this paper we take rather a different route to question of time.,25316
understanding come to up We of perception. with theoretical a attempt first time,We first attempt to come up with a theoretical understanding of time perception.,25317
work an extensive Hence our algebraic has flavor.,Hence our work has an extensive algebraic flavor.,25318
investigations direction. our believe We more in work that this initiate will,We believe that our work will initiate more investigations in this direction.,25319
"inference present We doubly stochastic not layers. force algorithm, which independence does a between variational","We present a doubly stochastic variational inference algorithm, which does not force independence between layers.",25320
"rate unique also protein In particular, production converges the steady-state. a to","In particular, the protein production rate also converges to a unique steady-state.",25321
"goals define In desirable within learning, often states. reinforcement we specifying by rewards","In reinforcement learning, we often define goals by specifying rewards within desirable states.",25322
previous isolation. of Most sub-problems the in solving one above has on work just focused,Most previous work has focused on solving just one of the above sub-problems in isolation.,25323
"the diving of classification. address problem Specifically, we","Specifically, we address the problem of diving classification.",25324
recorded sports. generalized be from footage video proposed to easily The techniques can other,The techniques proposed can be easily generalized to video footage recorded from other sports.,25325
labeled Deep exploits data of models. learning learn to powerful large volumes,Deep learning exploits large volumes of labeled data to learn powerful models.,25326
notion robustness problems. study in We the stable matching of,We study the notion of robustness in stable matching problems.,25327
"introducing first by (a,b)-supermatches. robustness define We","We first define robustness by introducing (a,b)-supermatches.",25328
outperforms evaluation empirical other local approaches. instances show on search that the large Our,Our empirical evaluation on large instances show that local search outperforms the other approaches.,25329
cases relative these The in investigated. are three performances,The relative performances in these three cases are investigated.,25330
both The cases. in most are efficient discussed methods,The most efficient methods are discussed in both cases.,25331
consists components: Automata Cellular (SCA) main Single-layer Cuboid Cellular HCA Automata (CCA). and of two,HCA consists of two main components: Single-layer Cellular Automata (SCA) and Cuboid Cellular Automata (CCA).,25332
values updated well-defined are update cells The according a rule. saliency iteratively to all of,The saliency values of all cells are iteratively updated according to a well-defined update rule.,25333
"propagation Therefore, unified integration HCA. and multi-layer single-layer jointly modeled in our are","Therefore, single-layer propagation and multi-layer integration are jointly modeled in our unified HCA.",25334
activations. propose to then supervise a loss function multi-label fused the We,We then propose a multi-label loss function to supervise the fused activations.,25335
fundamental reasoning Abstraction about systems. complex is tool for a,Abstraction is a fundamental tool for reasoning about complex systems.,25336
for great Program to analyzing abstraction has effect programs. deterministic utilized been,Program abstraction has been utilized to great effect for analyzing deterministic programs.,25337
"abstractions, not however, are typically Program probabilistic.","Program abstractions, however, are typically not probabilistic.",25338
non-deterministic explicitly the generalize We probabilistic abstractions program abstractions program non-deterministic choices. to by quantifying,We generalize non-deterministic program abstractions to probabilistic program abstractions by explicitly quantifying the non-deterministic choices.,25339
upgrades framework abstractions context. and key probabilistic the Our of to definitions properties,Our framework upgrades key definitions and properties of abstractions to the probabilistic context.,25340
for general abstractions on also inference preliminary programs. and probabilistic performing We ideas discuss probabilistic,We also discuss preliminary ideas for performing inference on probabilistic abstractions and general probabilistic programs.,25341
With performance steadily. networks neural power counting convolutional has the of the improved deep (CNNs),With the power of deep convolutional neural networks (CNNs) the counting performance has improved steadily.,25342
"density better found lower-resolution maps counting that sometimes performance. we have our experiments, In the","In our experiments, we found that the lower-resolution density maps sometimes have better counting performance.",25343
by mortality low fleet. catchability purse-seine implied The minimal the fishing,The low catchability implied minimal fishing mortality by the purse-seine fleet.,25344
a of realism. degree GANs with synthesize allow to images high,GANs allow to synthesize images with a high degree of realism.,25345
choice of It nonlinearities layers affects the when certain architectures. designing and,It affects the choice of certain layers and nonlinearities when designing architectures.,25346
"GAN layers. fully-connected particular, to does In it not convolutional models with train allow hidden","In particular, it does not allow to train convolutional GAN models with fully-connected hidden layers.",25347
deterministic existing semantic exploit in-between space. methods Most learning via embedding ZSL transfer an,Most existing ZSL methods exploit deterministic transfer learning via an in-between semantic embedding space.,25348
"try perspective. we attack from this paper, modelling problem a In this generative to probabilistic","In this paper, we try to attack this problem from a generative probabilistic modelling perspective.",25349
"any observed category, We for assume the e.g. representation,","We assume for any category, the observed representation, e.g.",25350
three proposed results. achieve experiments show state-of-the-art on our benchmark Extensive datasets can model,Extensive experiments on three benchmark datasets show our proposed model can achieve state-of-the-art results.,25351
proposed an have model analytical MC-WNNM not solution. The does,The proposed MC-WNNM model does not have an analytical solution.,25352
can has convergence solution guaranteed. step updating closed-form alternative the Each and be,Each alternative updating step has closed-form solution and the convergence can be guaranteed.,25353
"norm regularizer. often decay used weight To weight as shrinks ensure is that the generalization,","To ensure generalization, weight decay that shrinks the weight norm is often used as regularizer.",25354
"until pedestrian algorithm are improved simultaneously Progressively, CNN convergence. clustering model the and","Progressively, pedestrian clustering and the CNN model are improved simultaneously until algorithm convergence.",25355
formulated self-paced This naturally process as learning. is,This process is naturally formulated as self-paced learning.,25356
that directions further promising improvement. point then may We lead out to,We then point out promising directions that may lead to further improvement.,25357
devoted paper is settling question. this This to,This paper is devoted to settling this question.,25358
that changing solving ASP continuously processes and programs. solving logic with deal grounding features Multi-shot,Multi-shot ASP solving features grounding and solving processes that deal with continuously changing logic programs.,25359
"doing operative remain In they a seamless accommodate so, in and way. changes","In doing so, they remain operative and accommodate changes in a seamless way.",25360
evolutionary to it brings ASP operations. about state another changing since aspect dimension This adds,This evolutionary aspect adds another dimension to ASP since it brings about state changing operations.,25361
input accomplishes with language clingo this complementing ASP's capacities. by declarative control,clingo accomplishes this by complementing ASP's declarative input language with control capacities.,25362
"an choosing question. nodes---remains model the open selection---even number However, of","However, model selection---even choosing the number of nodes---remains an open question.",25363
"data large real-world corruptions, come or many errors with applications, In outliers.","In many real-world applications, data come with corruptions, large errors or outliers.",25364
regularization. There main proposed for are features the two outlier,There are two main features for the proposed outlier regularization.,25365
search it scalable is approach an as approximate made treating by neighbour nearest problem. Our,Our approach is made scalable by treating it as an approximate nearest neighbour search problem.,25366
"of known cancer. of biological the especially Vasculature in key significance, is be to study","Vasculature is known to be of key biological significance, especially in the study of cancer.",25367
"vessels or extracted thinning pipeline, Typically, a by are tracking. and either segmentation by direct","Typically, vessels are extracted by either a segmentation and thinning pipeline, or by direct tracking.",25368
are suited microscopy methods these tumor well of Neither images vasculature. of to,Neither of these methods are well suited to microscopy images of tumor vasculature.,25369
aims provide programs computer to about our to commonsense knowledge with Artificial Intelligence reason world.,Artificial Intelligence aims to provide computer programs with commonsense knowledge to reason about our world.,25370
have Intelligence Artificial always (AI). popular Games for been testbeds,Games have always been popular testbeds for Artificial Intelligence (AI).,25371
in used be The tasks. can annotated images image classification remote sensing,The annotated images can be used in remote sensing image classification tasks.,25372
"classification. large-scale worldwide for method, a sensing construct benchmark image remote on Based we this","Based on this method, we construct a worldwide large-scale benchmark for remote sensing image classification.",25373
procedures time or to hard generalize Such consuming across well either are datasets.,Such procedures are either time consuming or hard to generalize well across datasets.,25374
our available. make will We publicly code,We will make our code publicly available.,25375
"the optimization. and successful easy RE implement high-dimensional algorithm Moreover, is in to","Moreover, the RE algorithm is easy to implement and successful in high-dimensional optimization.",25376
classifiers that are and respond locations. scales different at They,They are classifiers that respond at different scales and locations.,25377
system The in tracker state. different each has governing rules,The tracker system has different governing rules in each state.,25378
"""hierarchical to called pooling extend representation, pooling"". rank high a we video capacity Then, rank","Then, we extend rank pooling to a high capacity video representation, called ""hierarchical rank pooling"".",25379
our demonstrate action of recognition. task We the approach on and activity,We demonstrate our approach on the task of action and activity recognition.,25380
work model-based for this (optimal) sequential present we In in selection strategies measurement diagnosis.,In this work we present strategies for (optimal) measurement selection in model-based sequential diagnosis.,25381
"variables. involving consider outcome this fair problem paper, of In statistical the we inference","In this paper, we consider the problem of fair statistical inference involving outcome variables.",25382
can be constrained learned solving then a problem. outcome fair by A model optimization,A fair outcome model can then be learned by solving a constrained optimization problem.,25383
"Also, Faster-RCNN RFCN. is faster than it and","Also, it is faster than Faster-RCNN and RFCN.",25384
image is due (SBIR) and retrieval the to domain-gap between inherent challenging sketch Sketch-based photo.,Sketch-based image retrieval (SBIR) is challenging due to the inherent domain-gap between sketch and photo.,25385
tackle cross-modal SBIR not existing problem. do studies Most directly this,Most existing SBIR studies do not directly tackle this cross-modal problem.,25386
research. we to insights few a drive future key addition In draw,In addition we draw a few key insights to drive future research.,25387
Those expensive and to mitigate. problems require complex algorithms,Those problems require complex and expensive algorithms to mitigate.,25388
"Essentially, influence effectively algorithm model mutual subjects. interacting between good should the two the a","Essentially, a good algorithm should effectively model the mutual influence between the two interacting subjects.",25389
"Also, in region identifying only the interaction. for scene small on-going the discriminative is a","Also, only a small region in the scene is discriminative for identifying the on-going interaction.",25390
"relative explicitly we In a difficulties. work, to address model this attention propose these","In this work, we propose a relative attention model to explicitly address these difficulties.",25391
a applications. Part-based to for visual representation effective proven has of variety been be,Part-based representation has been proven to be effective for a variety of visual applications.,25392
"object/part-level of discovery without annotations automatic However, parts challenging. is discriminative","However, automatic discovery of discriminative parts without object/part-level annotations is challenging.",25393
on Experiments existing method approaches. benchmark our the conducted datasets demonstrate superiority of over,Experiments conducted on benchmark datasets demonstrate the superiority of our method over existing approaches.,25394
detection pairs images. to in interactions capture objects Visual between relationship aims of,Visual relationship detection aims to capture interactions between pairs of objects in images.,25395
this the This first been time issue addressed. is has,This is the first time this issue has been addressed.,25396
"can model detecting Finally, robots how study misspecification. we such start","Finally, we study how robots can start detecting such model misspecification.",25397
(PS) regulation of space a is key component social effective engagement. Personal,Personal space (PS) regulation is a key component of effective social engagement.,25398
"summary, control In than in is BPD larger subjects. PID","In summary, PID is larger in BPD than control subjects.",25399
sensors. commodity the depth of re-identification from address problem We person,We address the problem of person re-identification from commodity depth sensors.,25400
scarcity. for is One data depth-based recognition challenge,One challenge for depth-based recognition is data scarcity.,25401
"attention reinforcement unit Since this stochastic, parameters learning. trained are using temporal the is","Since this unit is stochastic, the temporal attention parameters are trained using reinforcement learning.",25402
method sequences. our person validate the Extensive experiments accuracy in depth re-identification of from,Extensive experiments validate the accuracy of our method in person re-identification from depth sequences.,25403
of Deep hierarchical dimensional variables learning predictors. hidden layers applies high construct to nonlinear,Deep learning applies hierarchical layers of hidden variables to construct nonlinear high dimensional predictors.,25404
spatio-temporal Our modeling. goal for train develop architectures to learning deep is and,Our goal is to develop and train deep learning architectures for spatio-temporal modeling.,25405
"research. Finally, we directions future with for conclude","Finally, we conclude with directions for future research.",25406
diagnosing used area. Chest (CXRs) abnormalities X-Rays in are for and the heart lung widely,Chest X-Rays (CXRs) are widely used for diagnosing abnormalities in the heart and lung area.,25407
detecting abnormalities processes. high real enhance with these diagnosis greatly accuracy Automatically could world,Automatically detecting these abnormalities with high accuracy could greatly enhance real world diagnosis processes.,25408
find perform We same the doesn't architecture across well DCN all abnormalities. that,We find that the same DCN architecture doesn't perform well across all abnormalities.,25409
features. Shallow higher provide deep consistently to detection features accuracy earlier compared layers or,Shallow features or earlier layers consistently provide higher detection accuracy compared to deep features.,25410
classification We ensemble found also single significantly compared have to model. to models improve,We have also found ensemble models to improve classification significantly compared to single model.,25411
argumentation. static of is growing the to Our and dynamic aim interest complement in coordination,Our aim is to complement the growing interest in coordination of static and dynamic argumentation.,25412
a logic-based ontologies Formal formalism. are axiomatizations in,Formal ontologies are axiomatizations in a logic-based formalism.,25413
on detection the the redundancies given of in Our methodology axiomatization. based is inference-based,Our methodology is based on the detection of inference-based redundancies in the given axiomatization.,25414
"pattern plays very of the aperture role. critical design a However,","However, design of the aperture pattern plays a very critical role.",25415
major in challenges problem. this face two We,We face two major challenges in this problem.,25416
"known are results Consequently, segmentation not of enough. accurate FCN-type models the","Consequently, segmentation results of the known FCN-type models are not accurate enough.",25417
"pixel-wise ground truth is difficult obtain. to Second,","Second, pixel-wise ground truth is difficult to obtain.",25418
"construction preliminary any This presents for assumptions. without paper solution graph strict a undirected MPP,","This paper presents a preliminary undirected graph construction solution for MPP, without any strict assumptions.",25419
on based evolutionary an propose dynamics Here we explanation graphs. on,Here we propose an explanation based on evolutionary dynamics on graphs.,25420
"rely Unlike results explanations populations. homogeneous hold even our previous crucially for heterogeneity, that on","Unlike previous explanations that rely crucially on heterogeneity, our results hold even for homogeneous populations.",25421
in low-level a flow tracklets We link approach detections in network to use high-level. and,We use a network flow approach to link detections in low-level and tracklets in high-level.,25422
reduce and switch the show effect identity. detections framework missed our of recover to We,We show the effect of our framework to recover missed detections and reduce switch identity.,25423
lower competitive with state-of-the-art. on comparing achieve datasets We results identity to several switches,We achieve competitive results with lower identity switches on several datasets comparing to state-of-the-art.,25424
"A requiring of samples. key millions training current is data limitation state-of-the-art approaches efficiency, with","A key limitation is data efficiency, with current state-of-the-art approaches requiring millions of training samples.",25425
"available. not yet data However, human is readily demonstration","However, human demonstration data is not yet readily available.",25426
problem this follows. present addresses The work as,The present work addresses this problem as follows.,25427
"boosting data settings in than with Furthermore, can high-dimensional observations. used be covariates more","Furthermore, boosting can be used in high-dimensional data settings with more covariates than observations.",25428
"visualization provide a the tutorial of We fitting and results. hands-on on including tuning, model","We provide a hands-on tutorial on model fitting and tuning, including the visualization of results.",25429
present thoroughly numeric results. a some analysis We and analytical,We present a thoroughly numeric analysis and some analytical results.,25430
this of allows with Inference over to structure sequences data. posterior recover a us smoothed,Inference with this structure allows us to recover a posterior smoothed over sequences of data.,25431
GPSSM. the kernels This arbitrary within the allows use for of further,This further allows for the use of arbitrary kernels within the GPSSM.,25432
proof system The theory theory. based is on both probability and,The system is based on both probability theory and proof theory.,25433
from system is The epistemology. derived Chisholm's,The system is derived from Chisholm's epistemology.,25434
and by Chisholm's (i.e. system concretize undefined We grounding primitive his,We concretize Chisholm's system by grounding his undefined and primitive (i.e.,25435
probability and foundational) proof theory. of in concept reasonablenes,foundational) concept of reasonablenes in probability and proof theory.,25436
are for are Counterfactuals false. agent that sure statements knows an,Counterfactuals are statements that an agent knows for sure are false.,25437
"actions to their when cases, have other to are explain users. Among systems counterfactuals useful","Among other cases, counterfactuals are useful when systems have to explain their actions to users.",25438
naturally system. fall out counterfactuals from our Uncertainties for,Uncertainties for counterfactuals fall out naturally from our system.,25439
is logic reasoning just Efficient hard in first-order problem. a simple,Efficient reasoning in just simple first-order logic is a hard problem.,25440
for algorithm a sketch We first-order novel resolution. a reasoning that of present extends,We present a sketch of a novel algorithm for reasoning that extends first-order resolution.,25441
finished described work address progress; is lack. this and in will once The below,The work described below is in progress; and once finished will address this lack.,25442
"usually focuses on daytime. However, the it","However, it usually focuses on the daytime.",25443
"harder images are analyze. darker thus and sky/cloud to noisier, Nighttime and","Nighttime sky/cloud images are darker and noisier, and thus harder to analyze.",25444
"this a nighttime to superpixel-based we images. propose sky/cloud paper, method In segment","In this paper, we propose a superpixel-based method to segment nighttime sky/cloud images.",25445
We nighttime the first release to community. database segmentation image research the also sky/cloud,We also release the first nighttime sky/cloud image segmentation database to the research community.,25446
images. the our results nighttime efficacy proposed for algorithm of experimental show The,The experimental results show the efficacy of our proposed algorithm for nighttime images.,25447
"the decreased structure. to dramatically breaking However, of SRDCF's due circulant efficiency","However, SRDCF's efficiency dramatically decreased due to the breaking of circulant structure.",25448
tracking. propose Correlation Faster Regularized Discriminative (FSRDCF) We Spatially Filters for,We propose Faster Spatially Regularized Discriminative Correlation Filters (FSRDCF) for tracking.,25449
achieves equivalent three Our performance to on approach all SRDCF the datasets. tracker baseline,Our approach achieves equivalent performance to the baseline tracker SRDCF on all three datasets.,25450
"superior approach trackers. comparison, state-of-the-art other performance compared non-spatial-regularization demonstrates to our For","For state-of-the-art comparison, our approach demonstrates superior performance compared to other non-spatial-regularization trackers.",25451
visual the (e.g. tagged structure automatically data of Discovering semantic,Discovering automatically the semantic structure of tagged visual data (e.g.,25452
and made by challenging tags. The is sparse task incomplete inherently more,The task is made more challenging by inherently sparse and incomplete tags.,25453
"Symmetry meaning fundamental science, technology, and are engineering. and discrimination detection in of","Symmetry detection and discrimination are of fundamental meaning in science, technology, and engineering.",25454
concept paper impact'. defining looks alternative general a This of approach: at an `low,This paper looks at an alternative approach: defining a general concept of `low impact'.,25455
metaheuristic common algorithms effective are and searches. tool a for Multi-start,Multi-start algorithms are a common and effective tool for metaheuristic searches.,25456
"(CNN) novel mesh First, we for network neural segmentation. propose a convolutional","First, we propose a novel convolutional neural network (CNN) for mesh segmentation.",25457
expensive and realworld. many also labeling trial-and-errors incurs efforts human many for in It,It also incurs many human efforts for labeling and many expensive trial-and-errors in realworld.,25458
target. restore occasionally when It losing can tracking the,It can restore tracking when occasionally losing the target.,25459
"online have learn everyday become skills. resource to rich videos humans for a instructional Therefore,","Therefore, online instructional videos have become a rich resource for humans to learn everyday skills.",25460
"it observe is evaluate difficult by steps every to expert. and activity However,","However, it is difficult to observe and evaluate every activity steps by expert.",25461
"LSTM from are trajectories First, units action encoded network. with dense","First, action units are encoded from dense trajectories with LSTM network.",25462
The unit variable-length LSTM features a then action evaluated are Siamese by network.,The variable-length action unit features are then evaluated by a Siamese LSTM network.,25463
our that benchmark results shown methods have state-of-the-art on outperforms method Experimental datasets. current,Experimental results have shown that our method outperforms current state-of-the-art methods on benchmark datasets.,25464
"be and biological learning graph-based as can social, medical, framed challenges engineering Numerous tasks.","Numerous social, medical, engineering and biological challenges can be framed as graph-based learning tasks.",25465
"a propose network approach we new to classification. Here, based feature","Here, we propose a new feature based approach to network classification.",25466
between to be These different of discriminating classes for suitable signatures out features networks. turn,These features turn out to be suitable signatures for discriminating between different classes of networks.,25467
method is Our evaluated established empirically network on benchmarks.,Our method is evaluated empirically on established network benchmarks.,25468
exploit the estimate correlations LMCs multiple across outputs. and,LMCs estimate and exploit correlations across the multiple outputs.,25469
results. real An datasets developed shows and integration which system in promising evaluated is then,An integration system is then developed and evaluated in real datasets which shows promising results.,25470
our Experiments that filters results. show state-of-the-art produce new extensions and,Experiments show that our new filters and extensions produce state-of-the-art results.,25471
surveillance outperforms tasks. demonstrate state-of-the-art various MT-VSN in recognition We that,We demonstrate that MT-VSN outperforms state-of-the-art in various surveillance recognition tasks.,25472
NP-hard is prove propose for We problem. the and problem an solving the approach,We prove the problem is NP-hard and propose an approach for solving the problem.,25473
"addition, solving diversified algorithm we In exactly the MaxSAT. an partial top-k MEMKC present","In addition, we present an algorithm MEMKC exactly solving the diversified top-k partial MaxSAT.",25474
"from involves, dynamical theory This side, networks. such the analyzing systems arising on the","This involves, on the theory side, analyzing the dynamical systems arising from such networks.",25475
"for Our are into sub-problems, efficient decomposes known. solutions solution this which problem two","Our solution decomposes this problem into two sub-problems, for which efficient solutions are known.",25476
interest. of image underlying P-stage modeling allows the continuous of The for,The P-stage allows for continuous modeling of the underlying image of interest.,25477
a given which measurements are well image space is projected The represented. the onto in,The given measurements are projected onto a space in which the image is well represented.,25478
the corresponding proximal to solve optimization algorithms We develop further problem. efficient,We further develop efficient proximal algorithms to solve the corresponding optimization problem.,25479
learning pattern for machine. crucial Subjective is,Subjective pattern is crucial for learning machine.,25480
is for pattern. algebraic expression subjective X-form,X-form is algebraic expression for subjective pattern.,25481
"of machine. center of form space, is representation Collection which internal X-forms learning","Collection of X-forms form internal representation space, which is center of learning machine.",25482
without discuss teaching. We teaching learning by and,We discuss learning by teaching and without teaching.,25483
internal space its learning its dynamics. representation and,its internal representation space and its learning dynamics.,25484
defects. simulated is augmented image The data training by,The training data is augmented by simulated image defects.,25485
"the deformations always scene, exist images simultaneously. In real shape and color in","In the real scene, color and shape deformations always exist in images simultaneously.",25486
invariants be qualified situation. shape can color not for this Simple or invariants,Simple shape invariants or color invariants can not be qualified for this situation.,25487
just of method make a two The to the simple conventional is linear factors. combination,The conventional method is just to make a simple linear combination of the two factors.,25488
"is weights a Meanwhile, issue. selection of complex manual the","Meanwhile, the manual selection of weights is a complex issue.",25489
Our construction is integration the based on framework. method multiple,Our construction method is based on the multiple integration framework.,25490
to show how are across consistent sections. brains We spatially predictions and transferable new,We show how predictions are transferable to new brains and spatially consistent across sections.,25491
"for and still chip However, manipulation. fabrication researchers to droplet barrier biological skill handle exists","However, skill barrier still exists for biological researchers to handle chip fabrication and droplet manipulation.",25492
sensitivity. We is efficiency that higher demonstrated with and icMDA to able SNVs accurately detect,We demonstrated that icMDA is able to accurately detect SNVs with higher efficiency and sensitivity.,25493
"proposed. been reported strategies the literature, in As several clauses have learned deletion","As reported in the literature, several learned clauses deletion strategies have been proposed.",25494
"Laplacians, (SPD) covariances, matrices etc.) positive graph (e.g., definite Symmetric","Symmetric positive definite (SPD) matrices (e.g., covariances, graph Laplacians, etc.)",25495
relationship of used spatial the to domain. are or temporal widely model,are widely used to model the relationship of spatial or temporal domain.,25496
"embedded matrices Nevertheless, SPD theoretically Riemannian are manifolds. on","Nevertheless, SPD matrices are theoretically embedded on Riemannian manifolds.",25497
much more that The state-of-the-art. DMT-Net competitive demonstrate over is our experimental results,The experimental results demonstrate that our DMT-Net is much more competitive over state-of-the-art.,25498
"approach is supervised In interval novel features feature this paper, selection valued proposed. a for","In this paper, a novel feature selection approach for supervised interval valued features is proposed.",25499
takes features care specific interval the The clustering. through selecting K-Means proposed of approach class,The proposed approach takes care of selecting the class specific features through interval K-Means clustering.,25500
valued The interval algorithm data. is adapt clustering kernel of modified to K-Means,The kernel of K-Means clustering algorithm is modified to adapt interval valued data.,25501
"features of are each K corresponding number there to Hence, class.","Hence, there are K number of features corresponding to each class.",25502
"chosen. class, are to the Subsequently, cluster representatives corresponding each","Subsequently, corresponding to each class, the cluster representatives are chosen.",25503
This procedure for of classes. is the all samples remaining repeated,This procedure is repeated for all the samples of remaining classes.,25504
"interval standard For supervised datasets are experimentation, four used.","For experimentation, four standard supervised interval datasets are used.",25505
action for develop videos. a novel in localization framework We,We develop a novel framework for action localization in videos.,25506
and of mostly eliminates need Our non-residual parameterization in networks. initialization careful the also residual,Our parameterization also mostly eliminates the need of careful initialization in residual and non-residual networks.,25507
explored into recently visual knowledge recognition. Numerous have incorporate semantic been models embedding to,Numerous embedding models have been recently explored to incorporate semantic knowledge into visual recognition.,25508
"propose this we discriminative to learn constraints difference the In embeddings. structured and paper, visual-semantic","In this paper, we propose the structured discriminative and difference constraints to learn visual-semantic embeddings.",25509
for constraints of separability The discriminative image encourage different instances classes.,The discriminative constraints encourage separability for image instances of different classes.,25510
preserve The help word embeddings. the regularize to difference semantic relationships among constraints image embeddings,The difference constraints help regularize image embeddings to preserve the semantic relationships among word embeddings.,25511
stimulations. the studied inhibitory response optogenetic excitatory and and network to stimulation electrical We,We studied the network response to electrical stimulation and excitatory and inhibitory optogenetic stimulations.,25512
exhibit the stimulations Different with network. interactions activity pathological in different,Different stimulations exhibit different interactions with pathological activity in the network.,25513
parameter for stimulation network and studied these values. different We interactions,We studied these interactions for different network and stimulation parameter values.,25514
rhythmicity. stimulation electrical than be efficient suppressing stimulation found pathological Optogenetic more was to in,Optogenetic stimulation was found to be more efficient than electrical stimulation in suppressing pathological rhythmicity.,25515
actions detection interest. and temporal of are of Action videos topics segmentation in increasing,Action detection and temporal segmentation of actions in videos are topics of increasing interest.,25516
"great are segmentation methods Thus, of temporal importance. and supervised action weakly detection","Thus, weakly supervised action detection and temporal segmentation methods are of great importance.",25517
contaminated as (SAR) Radar are speckle. multiplicative Synthetic often a by Aperture known noise images,Synthetic Aperture Radar (SAR) images are often contaminated by a multiplicative noise known as speckle.,25518
makes the images processing Speckle SAR interpretation difficult. of and,Speckle makes the processing and interpretation of SAR images difficult.,25519
"aging this we In aim way. work, a to personalized render automatically in faces","In this work, we aim to automatically render aging faces in a personalized way.",25520
"two in Moreover, learning factors are the process. dictionary consideration into taken","Moreover, two factors are taken into consideration in the dictionary learning process.",25521
"characteristics, the personalized person aging each e.g. dictionaries, have facial may First, beyond extra","First, beyond the aging dictionaries, each person may have extra personalized facial characteristics, e.g.",25522
"in are mole, which invariant the process. aging","mole, which are invariant in the aging process.",25523
property fast for We this develop methods exploit to posterior approximate inference.,We exploit this property to develop fast approximate methods for posterior inference.,25524
data data. EEG We on simulated illustrate approach model with a our using neuroscience mechanistic,We illustrate our approach using simulated data on a mechanistic neuroscience model with EEG data.,25525
the real performance on strategy of video improved This shows our sequences.,This shows the improved performance of our strategy on real video sequences.,25526
"testes cell germ is and differentiation. Sertoli the""nurse""in that cell regulates proliferation","Sertoli cell is the""nurse""in testes that regulates germ cell proliferation and differentiation.",25527
of Sertoli cells during One germ processes. supports cell a number these certain,One Sertoli cell supports a certain number of germ cells during these processes.,25528
"Thus, the it capability. of male determinant is reproductive","Thus, it is the determinant of male reproductive capability.",25529
present on We end-to-end based compression image an system sensing. compressive,We present an end-to-end image compression system based on compressive sensing.,25530
"the system. sampling pixel camera, compressive to single hardware a construct","the single pixel camera, to construct a hardware compressive sampling system.",25531
the This bridges irreconcilable approaches paper to updates. seemingly,This paper bridges the seemingly irreconcilable approaches to updates.,25532
involves sub-second functional Cognition reconfiguration time networks dynamic brain at of scale.,Cognition involves dynamic reconfiguration of functional brain networks at sub-second time scale.,25533
A reconfigurations objects precise of to categorize visual elusive. tracking these remains,A precise tracking of these reconfigurations to categorize visual objects remains elusive.,25534
"similarity low showed inter-category network values. intra-category a results Second, using and high algorithm, very","Second, using a network similarity algorithm, results showed high intra-category and very low inter-category values.",25535
"robust a propose we In report, Faster approach detection on deep based this R-CNN. face","In this report, we propose a robust deep face detection approach based on Faster R-CNN.",25536
"we suited call so Face is proposed R-CNN. detection, it The well approach for face","The proposed approach is well suited for face detection, so we call it Face R-CNN.",25537
an economic and supports intelligence important Artificial social (AI) activities. life that daily technology is,Artificial intelligence (AI) is an important technology that supports daily social life and economic activities.,25538
Japan's social greatly and of solves economy sustainable various growth It contributes the to problems.,It contributes greatly to the sustainable growth of Japan's economy and solves various social problems.,25539
"are AI patterns, many developed excels extracting certain limitations. certainly recently there in Although technology","Although recently developed AI technology certainly excels in extracting certain patterns, there are many limitations.",25540
smoothing has out a This the process effect of program.,This process has the effect of smoothing out a program.,25541
We several of be used for the can the parts program. approximations give different that,We give several approximations that can be used for the different parts of the program.,25542
We the procedural problem this bandlimiting framework automatically to shader programs. of apply,We apply this framework to the problem of automatically bandlimiting procedural shader programs.,25543
amount in deal they sensory the moment. each data expert are of Humans with,Humans are expert in the amount of sensory data they deal with each moment.,25544
"present architecture data synthesis. analysis big Here, we for an intuition and based","Here, we present an intuition based architecture for big data analysis and synthesis.",25545
"unsupervised In the two domain unmatched mapping, is $B$. $A$ and datasets learner given","In unsupervised domain mapping, the learner is given two unmatched datasets $A$ and $B$.",25546
"this learning $G_{AB}$ $G_{BA}$. method we a learning In of present without work,","In this work, we present a method of learning $G_{AB}$ without learning $G_{BA}$.",25547
reproduced for The other methodology detection. be species systematically provided can,The provided methodology can be systematically reproduced for other species detection.,25548
technique the new a commonly in the label We extend move used alpha-expansion-based space. with,We extend the commonly used alpha-expansion-based technique with a new move in the label space.,25549
"not augmentation data has deeply However, investigated wearable been for sensor data yet.","However, data augmentation for wearable sensor data has not been deeply investigated yet.",25550
"methods data sensor paper, are for this proposed. augmentation various In data wearable","In this paper, various data augmentation methods for wearable sensor data are proposed.",25551
also of adopts encourage realistic constraint faces/sketches. an to adversarial r-BTN the generation,r-BTN also adopts an adversarial constraint to encourage the generation of realistic faces/sketches.,25552
experiments and performance. Our learning improved policy demonstrate,Our experiments demonstrate improved learning and policy performance.,25553
task Image or object recognition computer vision. an is important in,Image or object recognition is an important task in computer vision.,25554
ruled a design deals kind of of with a This paper surface.,This paper deals with a kind of design of a ruled surface.,25555
geometric aided concepts from combines computer It kinematics. fields and design of the,It combines concepts from the fields of computer aided geometric design and kinematics.,25556
"Until microbial was hidden view. world of from much recently, the","Until recently, much of the microbial world was hidden from view.",25557
"similarities independent diversification, find the of tempo universal quantitative habitat Moreover, in we of type.","Moreover, we find universal quantitative similarities in the tempo of diversification, independent of habitat type.",25558
Dubois by exercise an al. et Following proposed,Following an exercise proposed by Dubois et al.,25559
estimation We for flow unsupervised optical motion. ego-centric address,We address unsupervised optical flow estimation for ego-centric motion.,25560
"geometric is a a dense First, imposing with computed prior strong constraints. flow pixel-level spatial","First, a dense pixel-level flow is computed with a geometric prior imposing strong spatial constraints.",25561
second the The then deeper prediction feeding stage second network. refines a,The second stage then refines the prediction feeding a second deeper network.,25562
final the Our combination without outperforms state-of-the-art fine-tuning network architectures. employing RGB of or ensemble,Our final combination outperforms the state-of-the-art without employing fine-tuning or ensemble of RGB network architectures.,25563
person representations for end-to-end group for trajectory We approach propose activity analysis. an learning,We propose an end-to-end approach for learning person trajectory representations for group activity analysis.,25564
"and groups people of shot) pass, (teams).","pass, shot) and groups of people (teams).",25565
"called method introduced. robust GC-RANSAC in for A RANSAC, novel short, is estimation, Graph-Cut","A novel method for robust estimation, called Graph-Cut RANSAC, GC-RANSAC in short, is introduced.",25566
"is proposed LO globally to implement, The conceptually step easy simple, optimal and efficient.","The proposed LO step is conceptually simple, easy to implement, globally optimal and efficient.",25567
"essential fundamental estimation. affine and fitting, matrix line transformation, homography,","line fitting, homography, affine transformation, fundamental and essential matrix estimation.",25568
and investigate of connections We effect the dropout. skip,We investigate the effect of skip connections and dropout.,25569
person's period shown another image'. the 'retrieving age by similar someone's at image,'retrieving someone's image at the similar age period shown by another person's image'.,25570
"The experiments show outperforming results, methods. hierarchical promising","The experiments show promising results, outperforming hierarchical methods.",25571
recordings been in-vivo feline methods and applied have These seizure results. promising yielded to,These methods have been applied to in-vivo feline seizure recordings and yielded promising results.,25572
captioning. framework based video using for in Recent encoder-decoder made has progress been attention,Recent progress has been made in using attention based encoder-decoder framework for video captioning.,25573
"tested datasets. been CASIA and SDUMLA-HMT, proposed has The system using","The proposed system has been tested using SDUMLA-HMT, and CASIA datasets.",25574
compare distributions The (i.e. to probability two ability degenerate,The ability to compare two degenerate probability distributions (i.e.,25575
"many Face and attribute estimation retrieval, social potential in video surveillance, face media. has applications","Face attribute estimation has many potential applications in video surveillance, face retrieval, and social media.",25576
person-person interactions of over the dynamics This how neglects change inter-related time.,This neglects the inter-related dynamics of how person-person interactions change over time.,25577
"development. exploration to Creative and science, art is central cognitive","Creative exploration is central to science, art and cognitive development.",25578
"by is research automated of creative paradigms. lack exploration a However, on limited high-resolution","However, research on creative exploration is limited by a lack of high-resolution automated paradigms.",25579
"as and digits, discovered letters, airplanes. such shape categories Players","Players discovered shape categories such as digits, letters, and airplanes.",25580
"once exploited explore category, so again, and They it each on. then to dropped","They exploited each category, then dropped it to explore once again, and so on.",25581
paradigm study high-resolution of creative way the This opens automated exploration. for,This paradigm opens the way for automated high-resolution study of creative exploration.,25582
table-top To in our test scenario. experiments framework present manipulation object a we,To test our framework we present experiments in a table-top object manipulation scenario.,25583
well single Is performs datasets? that model a across there,Is there a single model that performs well across datasets?,25584
"characteristics the not, dataset If and determine of of models? what performance a MF TF","If not, what characteristics of a dataset determine the performance of MF and TF models?",25585
TF+MF Is a on that model datasets? robustly performs all there joint,Is there a joint TF+MF model that performs robustly on all datasets?,25586
We models also characteristics. and results based on the analyze dataset explain and,We also analyze and explain the results based on models and dataset characteristics.,25587
"obtains Our strong robust, across model results datasets. all best and is","Our best model is robust, and obtains strong results across all datasets.",25588
as Here these to refer Genetic technologies Engineering (EGE). I Evolutionary collectively,Here I collectively refer to these technologies as Evolutionary Genetic Engineering (EGE).,25589
"of humanitarian, is for conservation and EGE's. economic, great potential There positive applications","There is great potential for positive humanitarian, economic, and conservation applications of EGE's.",25590
"the To of best problem is knowledge, this our unsolved.","To the best of our knowledge, this problem is unsolved.",25591
correspondence-based approach linear derived techniques point local The transformations. from proposed constraints with extends affine,The proposed approach extends point correspondence-based techniques with linear constraints derived from local affine transformations.,25592
the efficiently by obtained The system is technique. polynomial hidden-variable solved multivariate,The obtained multivariate polynomial system is efficiently solved by the hidden-variable technique.,25593
"the conditions we local eliminating roots. geometry of Observing novel affinities, invalid introduce","Observing the geometry of local affinities, we introduce novel conditions eliminating invalid roots.",25594
the in A solution paper. Matlab included is proposed implementation the of,A Matlab implementation of the proposed solution is included in the paper.,25595
N-polyhedron without an general computes of intermediate the for tree result boolean QuickCSG solids. expressions,QuickCSG computes the result for general N-polyhedron boolean expressions without an intermediate tree of solids.,25596
"combination the the approach. improves our in approach, vertex-centric of speed This with simplifying assumption,","This simplifying assumption, in combination with our vertex-centric approach, improves the speed of the approach.",25597
development. popular computer area Pathfinding in game very is a,Pathfinding is a very popular area in computer game development.,25598
future The with the algorithms advantages along proposed and also directions. are discussed of shortcomings,The advantages and shortcomings of the proposed algorithms are also discussed along with future directions.,25599
"change a data. set of a requires training However, training classifier place-specific","However, training a change classifier requires a set of place-specific training data.",25600
"novel we we this In address discovery. place study, term this unsupervised problem, which","In this study, we address this novel problem, which we term unsupervised place discovery.",25601
domain. support methods large guarantee deconvolution a Most kernel blind size usually the pre-define to,Most blind deconvolution methods usually pre-define a large kernel size to guarantee the support domain.,25602
"in yielding error to results. be is introduced, Blur deblurring likely kernel artifacts estimation severe","Blur kernel estimation error is likely to be introduced, yielding severe artifacts in deblurring results.",25603
we And algorithm efficient it. an propose optimization solve to,And we propose an efficient optimization algorithm to solve it.,25604
method. results blurry effectiveness further on deblurring proposed validate The the real-world images the of,The deblurring results on real-world blurry images further validate the effectiveness of the proposed method.,25605
in Sparse learning a data building is block core and machine coding many pipelines. analysis,Sparse coding is a core building block in many data analysis and machine learning pipelines.,25606
"functions. These methods non-smooth, in convex optimal are methods the for of class first-order","These methods are optimal in the class of first-order methods for non-smooth, convex functions.",25607
we study reasons In paper acceleration. for the such this,In this paper we study the reasons for such acceleration.,25608
are and consuming Pixel-level to time obtain. annotations expensive,Pixel-level annotations are expensive and time consuming to obtain.,25609
"have semantic significant tags impact only using supervision could segmentation. image weak Hence, a in","Hence, weak supervision using only image tags could have a significant impact in semantic segmentation.",25610
"proposed Recently, networks image have CNN-based fine-tune using to tags. methods pre-trained","Recently, CNN-based methods have proposed to fine-tune pre-trained networks using image tags.",25611
"accuracy. Without information, this localization additional leads poor to","Without additional information, this leads to poor localization accuracy.",25612
"was however, This priors generate of problem, objectness making foreground/background masks. to use alleviated by","This problem, however, was alleviated by making use of objectness priors to generate foreground/background masks.",25613
"or annotations/bounding yield still either object inaccurate Unfortunately boxes, boundaries. these priors pixel-level require","Unfortunately these priors either require pixel-level annotations/bounding boxes, or still yield inaccurate object boundaries.",25614
"stability this an presented. anaerobic the analysis In digestion paper, of process is","In this paper, the stability analysis of an anaerobic digestion process is presented.",25615
"the optimal and are safety In operation regions addition, bioreactor both of established.","In addition, both safety and optimal operation regions of the bioreactor are established.",25616
are one clinically. Less vertebral of third fractures diagnosed compression than,Less than one third of vertebral compression fractures are diagnosed clinically.,25617
Computed spine an compression scans. fractures detecting automated present in (CT) for method We Tomography,We present an automated method for detecting spine compression fractures in Computed Tomography (CT) scans.,25618
The of algorithm processes. three composed is,The algorithm is composed of three processes.,25619
"the are spinal extracted. and is First, patches column segmented sagittal","First, the spinal column is segmented and sagittal patches are extracted.",25620
using patches a are Network Neural binary (CNN). classified Convolutional then The,The patches are then binary classified using a Convolutional Neural Network (CNN).,25621
brain from challenge. tested segmentation SegAN the datasets using our MICCAI BRATS method We tumor,We tested our SegAN method using datasets from the MICCAI BRATS brain tumor segmentation challenge.,25622
this we recognition approach scene an propose images. lexicon-free In in to of paper text,In this paper we propose an approach to lexicon-free recognition of text in scene images.,25623
a LSTM-based Our relies convolutional model visual on attention features. from soft learned approach,Our approach relies on a LSTM-based soft visual attention model learned from convolutional features.,25624
spatial permits into This encoding of image representation. the information,This permits encoding of spatial information into the image representation.,25625
be end-to-end annotations. level using Training done word only can,Training can be done end-to-end using only word level annotations.,25626
"the simple algorithms models. Next, behavior illustrated is HMC two of the using probability distribution","Next, the behavior of the two HMC algorithms is illustrated using simple probability distribution models.",25627
Each services and uploaded media photographs platforms. to of are photo-sharing day social billions,Each day billions of photographs are uploaded to photo-sharing services and social media platforms.,25628
the are people information world. around with about packed live how images These,These images are packed with information about how people live around the world.,25629
instance PASCAL segmentation demonstrate We state-of-the-art accuracy VOC. on,We demonstrate state-of-the-art instance segmentation accuracy on PASCAL VOC.,25630
relaxation. segment propose convex fit a data to novel multi-structural method and via We,We propose a novel method to fit and segment multi-structural data via convex relaxation.,25631
energy. state-of-the-art Our use a is energy global similar minimisation which to techniques approach,Our approach is similar to state-of-the-art energy minimisation techniques which use a global energy.,25632
"CDJ structure convolutional maintains the (CNNs), standard neural networks of i.e. The","The CDJ maintains the structure of standard convolutional neural networks (CNNs), i.e.",25633
multiple layers multiple response fully connected. of maps,multiple layers of multiple response maps fully connected.,25634
via sample specific soft nodes. travels each activated a of those data route,each data sample travels via a specific soft route of those activated nodes.,25635
and conventional network The by combining losses each softmax in proposed learnt layer. entropy is,The network is learnt by combining conventional softmax and proposed entropy losses in each layer.,25636
to present new approach We a learning. ensemble,We present a new approach to ensemble learning.,25637
has submitted also Menpo part for been method of the evaluation as Our challenge.,Our method has also been submitted for evaluation as part of the Menpo challenge.,25638
important imaging breast role Ultrasound plays an lesion differentiation. in,Ultrasound imaging plays an important role in breast lesion differentiation.,25639
"depends However, on accuracy diagnostic experience. ultrasonographer","However, diagnostic accuracy depends on ultrasonographer experience.",25640
boundary. assess combining with features morphological accomplished the by BI-RADS was which This lesion,This was accomplished by combining the BI-RADS with morphological features which assess lesion boundary.,25641
"feature Additionally, performing the morphological was best subset indicated.","Additionally, the best performing morphological feature subset was indicated.",25642
features. a combining the classification six obtained by We BI-RADS better with morphological,We obtained a better classification by combining the BI-RADS with six morphological features.,25643
object an can background. tracking algorithm multiple proposes on-line in This that unknown paper operate,This paper proposes an on-line multiple object tracking algorithm that can operate in unknown background.,25644
brute is Chemical so force for large are new searches space molecules that interesting infeasible.,Chemical space is so large that brute force searches for new interesting molecules are infeasible.,25645
solution These that BO. show for is parallel results large-scale PDTS a successful,These results show that PDTS is a successful solution for large-scale parallel BO.,25646
to knowledge we integrate neural how unsupervised networks. this paper In into show symbolic,In this paper we show how to integrate symbolic knowledge into unsupervised neural networks.,25647
A problems. contribution for second optimization joint consider methods is to these,A second contribution is to consider optimization methods for these joint problems.,25648
examples and problems of potential Synthetic practical the approach. larger-scale illustrate the,Synthetic problems and larger-scale practical examples illustrate the potential of the approach.,25649
"are of learning instance robust versions investigated. miSVM and multiple Two (MIL) MILES, classifiers,","Two robust versions of multiple instance learning (MIL) classifiers, miSVM and MILES, are investigated.",25650
classifier performed patient classification. better emphysema than on both and MILES miSVM The,The miSVM classifier performed better than MILES on both patient and emphysema classification.,25651
is radiologist. and by only the classifier The second the between correlation the outperformed PFT,The correlation between the classifier and the PFT is only outperformed by the second radiologist.,25652
and facilitating is therefore method for The inter-observer promising of emphysema variability. reducing assessment,The method is therefore promising for facilitating assessment of emphysema and reducing inter-observer variability.,25653
"recently gathered attention. Epistemic much logic knowledge with operators, non-standard especially ""knowing-value"" operator, the has","Epistemic logic with non-standard knowledge operators, especially the ""knowing-value"" operator, has recently gathered much attention.",25654
axiomatization multiagent given this case. in A is,A multiagent axiomatization is given in this case.,25655
"tracked based attributes. object Kalman is Finally, using its on filter each","Finally, each object is tracked using Kalman filter based on its attributes.",25656
match pedestrians (re-id) observed camera aims by to Person views. disjoint re-identification,Person re-identification (re-id) aims to match pedestrians observed by disjoint camera views.,25657
computer surveillance It attracts in to due to its increasing importance system. attention vision,It attracts increasing attention in computer vision due to its importance to surveillance system.,25658
Experiments by state-of-the-art on benchmarks achieved our results show method.,Experiments on benchmarks show state-of-the-art results achieved by our method.,25659
one Battle most is of Online Multiplayer played genres nowadays. the game Arena (MOBA),Multiplayer Online Battle Arena (MOBA) is one of the most played game genres nowadays.,25660
agent problem for we MOBA address paper this games. the In development of,In this paper we address the problem of agent development for MOBA games.,25661
a mechanics. navigation We and that architecture handles implement both game two-layered agent,We implement a two-layered architecture agent that handles both navigation and game mechanics.,25662
of at research and aimed learn have modeling features how Decades psychological categories. people been,Decades of psychological research have been aimed at modeling how people learn features and categories.,25663
based with theories validation these on artificial simple representations. of stimuli is empirical The often,The empirical validation of these theories is often based on artificial stimuli with simple representations.,25664
method results in indicate our standardized of potential aiding assessments. the MI These proposed,These results indicate the potential of our proposed method in aiding standardized MI assessments.,25665
Finding Multi-Agent (MAPF) in NP-hard robotics. studied problem intelligence Path artificial is an well and,Multi-Agent Path Finding (MAPF) is an NP-hard problem well studied in artificial intelligence and robotics.,25666
many which various for MAPF solvers applications heuristics. real-world has existing use It,It has many real-world applications for which existing MAPF solvers use various heuristics.,25667
various MTE demonstrated. of probabilistic of Weaknesses proposals belief are of interpretation functions,Weaknesses of various proposals of probabilistic interpretation of MTE belief functions are demonstrated.,25668
interpretations. presented frequency-based various interpretation drawbacks is earlier new overcoming of A,A new frequency-based interpretation is presented overcoming various drawbacks of earlier interpretations.,25669
"on Furthermore, the system. to model use possible gain we the biological parameters in insight","Furthermore, we use the model to gain insight on possible biological parameters in the system.",25670
"and on in still These have performance models achieved detection super-human localisation, object images. recognition,","These models have achieved super-human performance on object recognition, localisation, and detection in still images.",25671
a adversarial using generative (GAN). method compelling we propose Here networks,Here we propose a compelling method using generative adversarial networks (GAN).,25672
identification image community. The important topic is in the vascular medical of an networks analysis,The identification of vascular networks is an important topic in the medical image analysis community.,25673
Our based concentric points spherical a on cloud of distributed approach over layers. sampling is,Our approach is based on a cloud of sampling points distributed over concentric spherical layers.,25674
The using both images. real and synthetic tested method was,The method was tested using both synthetic and real images.,25675
visual our networks approach the also vascular potential show results identifying of Further topologies. for,Further visual results also show the potential of our approach for identifying vascular networks topologies.,25676
"holds aspect Finally, improvements. and for modular adjustments potential problem-oriented method's the","Finally, the method's modular aspect holds potential for problem-oriented adjustments and improvements.",25677
a of our There visual is understanding awareness world. fundamental the puzzle in,There is a fundamental puzzle in understanding our awareness of the visual world.,25678
"when However, actually little. observers test we surprisingly perception, know","However, when we actually test perception, observers know surprisingly little.",25679
"on have solving effectiveness shown methods problem. Recently, this deep-learning-based their","Recently, deep-learning-based methods have shown their effectiveness on solving this problem.",25680
and such training introduced. network advantages hierarchies structure that of Novel are strategies take label,Novel training strategies and network structure that take advantages of such label hierarchies are introduced.,25681
object perform collaborative smartphones application MirBot users that recognition. allows for is a to,MirBot is a collaborative application for smartphones that allows users to perform object recognition.,25682
features network (CNN). search from a extracted of means neural similarity convolutional by using,by means of similarity search using features extracted from a convolutional neural network (CNN).,25683
"available users' grows for publicly and dataset feedback, thanks is research. This the to continuously","This dataset grows continuously thanks to the users' feedback, and is publicly available for research.",25684
Play Google is available and app The stores. freely Apple the at,The app is freely available at the Apple and Google Play stores.,25685
presents model paper digital This for sparse and intelligent and recovery. adaptive sampling an image,This paper presents an adaptive and intelligent sparse model for digital image sampling and recovery.,25686
purposes. are codes reproduction Simulation for online available,Simulation codes are available online for reproduction purposes.,25687
networks are of interpretation. from for belief for construction this data Algorithms derived,Algorithms for construction of belief networks from data are derived for this interpretation.,25688
form universal As show an application of weak we some consistency.,As an application we show some weak form of universal consistency.,25689
"present view action a human video Okutama-Action, dataset new aerial concurrent for detection. We","We present Okutama-Action, a new video dataset for aerial view concurrent human action detection.",25690
in make tractable. models dynamic convex developments approach Recent efficient large-scale of the optimization,Recent developments in efficient convex optimization of large-scale dynamic models make the approach tractable.,25691
is Code for experiments the and approach available. publicly,Code for the approach and experiments is publicly available.,25692
"CoMaL been proposed handle have Recently, that a such case. Features","Recently, CoMaL Features have been proposed that handle such a case.",25693
and points. efficiently CoMaL propose accurately novel to tracking algorithm track a We,We propose a novel tracking algorithm to accurately and efficiently track CoMaL points.,25694
learning deep works prior study point Few on sets.,Few prior works study deep learning on point sets.,25695
makes practical RBG This for in often and various GGP. developments research suitable more,This often makes RBG more suitable for various research and practical developments in GGP.,25696
multiple is signals. study relationship to way the connectivity popular Functional between neural one,Functional connectivity is one popular way to study the relationship between multiple neural signals.,25697
are for well-used set connectivity currently estimation. of functional techniques a Correlation-based methods,Correlation-based methods are a set of currently well-used techniques for functional connectivity estimation.,25698
also calcium recordings local We field imaging. expand method the (LFP) to potential and simulated,We also expand the method to simulated local field potential (LFP) recordings and calcium imaging.,25699
neural analyze class to methods signals. alternative new This provides ways of,This new class of methods provides alternative ways to analyze neural signals.,25700
human system. This visual attention is the to the similar of mechanism,This is similar to the attention mechanism of the human visual system.,25701
and stones shapes reconstructed The kidney visible sizes. on identifiable with planes are,The kidney stones are visible on reconstructed planes with identifiable shapes and sizes.,25702
signals. varying fast extracts features SFA from slowly varying,SFA extracts slowly varying features from fast varying signals.,25703
"to leach dynamic SFA is representations textures. invariant from Fortunately, capable","Fortunately, SFA is capable to leach invariant representations from dynamic textures.",25704
"only learned Therefore, varying, slowly but predictable. the not partly features are also","Therefore, the learned features are not only slowly varying, but also partly predictable.",25705
guarantees shortest generation the produces and also therefore of paths. consistent heuristics and admissible FastMap,FastMap also produces admissible and consistent heuristics and therefore guarantees the generation of shortest paths.,25706
algorithm Our different under scattering conditions. solutions to illumination render isotropic is fluorescent applied highly,Our algorithm is applied to render highly scattering isotropic fluorescent solutions under different illumination conditions.,25707
causes world. the of in retinopathy is the one Diabetic preventable of blindness leading,Diabetic retinopathy is one of the leading causes of preventable blindness in the world.,25708
"lesions, are groups and earliest a Its both red term sign microaneurysms general hemorrhages. that","Its earliest sign are red lesions, a general term that groups both microaneurysms and hemorrhages.",25709
"using daily by are clinical practice, In detected lesions fundus photographs. physicians these manually","In daily clinical practice, these lesions are manually detected by physicians using fundus photographs.",25710
are incorporating crafted by by learned Features augmented hand a CNN features.,Features learned by a CNN are augmented by incorporating hand crafted features.,25711
source online. An implementation is system of publicly our open available,An open source implementation of our system is publicly available online.,25712
compare performance current We the on programs with the ToxTrac of web. accessible,We compare the performance of ToxTrac with current accessible programs on the web.,25713
suggest independent research using ToxTrac behavior of future of animal studies for area. We,We suggest using ToxTrac for future studies of animal behavior independent of research area.,25714
"the path-finding. research, this we In investigate subject of","In this research, we investigate the subject of path-finding.",25715
is algorithm cases. on path-finding various The proposed demonstrated,The algorithm proposed is demonstrated on various path-finding cases.,25716
several multi-view our demonstrate clearly outperforms on state-of-the-art that the methods. datasets experiments Rigorous approach,Rigorous experiments on several multi-view datasets demonstrate that our approach clearly outperforms the state-of-the-art methods.,25717
problem and computer challenging vision. is in re-identification open Person an,Person re-identification is an open and challenging problem in computer vision.,25718
"of generate methods existing Unfortunately, the none practically results. satisfying","Unfortunately, none of the existing methods generate practically satisfying results.",25719
We directions promising application. propose several practical for,We propose several promising directions for practical application.,25720
"appearance based the Generally, target. learn of model rigid a trackers DCF","Generally, DCF based trackers learn a rigid appearance model of the target.",25721
"a deformable In filter. propose unified we learning paper, convolution formulation this for a","In this paper, we propose a unified formulation for learning a deformable convolution filter.",25722
"as the framework, deformable our linear combination sub-filters. In represented of is filter a","In our framework, the deformable filter is represented as a linear combination of sub-filters.",25723
locations are coefficients Both formulation. sub-filter and jointly our the their relative in inferred,Both the sub-filter coefficients and their relative locations are inferred jointly in our formulation.,25724
"improves state-of-the-art. performance approach to to leading the comparable Our baseline method,","Our approach improves the baseline method, leading to performance comparable to state-of-the-art.",25725
and a simulation We challenging number algorithm both of real cases. with our validate also,We also validate our algorithm with both simulation and a number of challenging real cases.,25726
Monte Hamiltonian energy in Carlo algorithm different affect consider performance. We kinetic choices how of,We consider how different choices of kinetic energy in Hamiltonian Monte Carlo affect algorithm performance.,25727
motions. of Major the ambiguity high potential and problems pixel of from dimensionality come space,Major problems come from high dimensionality of pixel space and the ambiguity of potential motions.,25728
"samples. can latent method sampling our video variables, Through output different imaginary different","Through sampling different latent variables, our method can output different imaginary video samples.",25729
The unsupervised an with learning. way trained framework adversarial is in,The framework is trained in an adversarial way with unsupervised learning.,25730
"$RIQA$. we assessment propose a For metric evaluation, new","For evaluation, we propose a new assessment metric $RIQA$.",25731
image performance quality Our promising framework assessment. achieves in,Our framework achieves promising performance in image quality assessment.,25732
which yet suffer understood. its causality from is Schizophrenic patients fully not hallucination,Schizophrenic patients suffer from hallucination which its causality is not yet fully understood.,25733
of approach This from attempts theories. mechanical mystery the this perspective paper to quantum,This paper attempts to approach this mystery from the perspective of quantum mechanical theories.,25734
in artificial domains from analysis. software important to intelligence is Constrained counting ranging,Constrained counting is important in domains ranging from artificial intelligence to software analysis.,25735
constraints. already approaches are There a over of few counting various models for types,There are already a few approaches for counting models over various types of constraints.,25736
"theoretical scalability, hashing-based achieve rely guarantees approaches but solution Recently, still enumeration. both on and","Recently, hashing-based approaches achieve both theoretical guarantees and scalability, but still rely on solution enumeration.",25737
also A stopping dynamic presented. criterion is variant with a,A variant with a dynamic stopping criterion is also presented.,25738
This interpolation. video challenging of task the considers paper long-term,This paper considers the challenging task of long-term video interpolation.,25739
of benefit our joint the We approach the intensity-depth context illustrate imaging. in of,We illustrate the benefit of our approach in the context of joint intensity-depth imaging.,25740
"attribute-centric underperformed state-of-the-art thus However, approaches. existing conventional far have approaches","However, existing attribute-centric approaches have thus far underperformed state-of-the-art conventional approaches.",25741
annotation. (camera) their specific to domain need is due for non-scalable This extensive,This is due to their non-scalable need for extensive domain (camera) specific annotation.,25742
datasets existing either trained -- is or labelled. model Our weakly photography fashion strongly on,Our model is trained on existing fashion photography datasets -- either weakly or strongly labelled.,25743
used forgeries. can contrast Inconsistency to enhancement in image expose be,Inconsistency in contrast enhancement can be used to expose image forgeries.,25744
of efficiency to We perform method. extensive the demonstrate and method efficacy experimental evaluations our,We perform extensive experimental evaluations to demonstrate the efficacy and efficiency of our method method.,25745
SATFC. we paper process This built by describes the which,This paper describes the process by which we built SATFC.,25746
experiments scientific and special Plant instruments. through proven laboratory aliveness is,Plant aliveness is proven through laboratory experiments and special scientific instruments.,25747
"functioning, is general Sleep cognitive and performance for daytime well-being. crucial","Sleep is crucial for daytime functioning, cognitive performance and general well-being.",25748
We that with as find LRTCs progresses. both deprivation decline measures sleep,We find with both measures that LRTCs decline as sleep deprivation progresses.,25749
into when appropriate decline changes in evident taking power This consideration. becomes signal,This decline becomes evident when taking changes in signal power into appropriate consideration.,25750
the maintain Our in the importance sleep LRTCs brain. of results demonstrate to human,Our results demonstrate the importance of sleep to maintain LRTCs in the human brain.,25751
"In the emerge state. of critical naturally a vicinity networks, complex LRTCs in","In complex networks, LRTCs naturally emerge in the vicinity of a critical state.",25752
(SSCs). the and Spermatogenesis cells is differentiation of proliferation stem by spermatogonial sustained,Spermatogenesis is sustained by the proliferation and differentiation of spermatogonial stem cells (SSCs).,25753
embryonic (PGCs) are primordial derived during development. germ cells SSCs from,SSCs are derived from primordial germ cells (PGCs) during embryonic development.,25754
"further molecules However, still processes clarification. these controlling need the","However, the molecules controlling these processes still need further clarification.",25755
caused bicycles accidents to serious have safety. become the The a by threat public,The accidents caused by bicycles have become a serious threat to the public safety.,25756
video detection in surveillance of traffic bicycle one major China. So is task systems,So bicycle detection is one major task of traffic video surveillance systems in China.,25757
"systems. surveillance video therefore, traffic for real-time applicable It is,","It is, therefore, applicable for real-time traffic video surveillance systems.",25758
"RSIR also evaluating, Benchmark approaches. developing, for critical comparing are datasets and","Benchmark datasets are also critical for developing, evaluating, and comparing RSIR approaches.",25759
"the this data, Interpreting thorough analyses. study the multi-modal in depth paper presents","Interpreting the depth data, the study in this paper presents thorough multi-modal analyses.",25760
"What steps while for robotics improving in the take handling should next ""autonomy"" community objects?","What next steps should the community take for improving ""autonomy"" in robotics while handling objects?",25761
Markov consider evaluating task decision policy a process a We of (MDP). the for,We consider the task of evaluating a policy for a Markov decision process (MDP).,25762
interacting cells. novel this paper a modeling propose we populations of framework for In computational,In this paper we propose a novel computational framework for modeling populations of interacting cells.,25763
"to and multiscale efficient. scalable is nature, simulation extremely its the highly algorithm resulting Thanks","Thanks to its multiscale nature, the resulting simulation algorithm is extremely scalable and highly efficient.",25764
materials. challenge face movie always of reproducing game The and industries the,The game and movie industries always face the challenge of reproducing materials.,25765
illumination or This procedural models and problem various patterns). textures combining by is tackled (painted,This problem is tackled by combining illumination models and various textures (painted or procedural patterns).,25766
as Long lines may visual artifacts. appear,Long lines may appear as visual artifacts.,25767
this algorithm set new We stochastic to a use in procedural patterns. wall tile generate,We use this tile set in a new procedural algorithm to generate stochastic wall patterns.,25768
"we constrained For hash purpose, specific functions Wang tiling. implementing a introduce this","For this purpose, we introduce specific hash functions implementing a constrained Wang tiling.",25769
"heartbeat a that pattern. ballistocardiogram ""heartbeat represents DL-FUMI individual's personal an estimates concept""","DL-FUMI estimates a ""heartbeat concept"" that represents an individual's personal ballistocardiogram heartbeat pattern.",25770
is ubiquitous analysis. image medical Supervised in learning,Supervised learning is ubiquitous in medical image analysis.,25771
t-tests. two-sample tests to Wilcoxon Rank-based paired the and popular distribution-free alternatives are,Wilcoxon Rank-based tests are distribution-free alternatives to the popular two-sample and paired t-tests.,25772
"stats For coin. data, in several as they available packages such independent R are and","For independent data, they are available in several R packages such as stats and coin.",25773
Exact provided also some on permutations are methods. for based tests,Exact tests based on permutations are also provided for some methods.,25774
different schools reviewed. the of are Details of major briefly methods,Details of the major schools of different methods are briefly reviewed.,25775
proximation to block-diagonal Gauss-Newton feedforward networks. for efficient present the ap- matrix We an neural,We present an efficient block-diagonal ap- proximation to the Gauss-Newton matrix for feedforward neural networks.,25776
programming each vertex We shader program. represent a use as genetic to computer,We use genetic programming to represent each vertex shader as a computer program.,25777
the We and current development challenges. project's status discuss,We discuss the project's current status and development challenges.,25778
(i.e. domain complex of continuous the paper problem this in detection event In the,In this paper the problem of complex event detection in the continuous domain (i.e.,25779
events locations) ending and addressed. starting with unknown is,events with unknown starting and ending locations) is addressed.,25780
"However, with the vulnerable to similar this the e.g. makes concepts events model","However, this makes the model vulnerable to the events with similar concepts e.g.",25781
vision. core problem computer is detection a Object in,Object detection is a core problem in computer vision.,25782
"the development has been object With improved. deep performance ConvNets, the detectors of dramatically of","With the development of deep ConvNets, the performance of object detectors has been dramatically improved.",25783
"stochastic from result as show processes species-range expansions. may pattern during Alternatively, we here, this","Alternatively, as we show here, this pattern may result from stochastic processes during species-range expansions.",25784
"that starts with asexual wave. sexual conditions an eventually recruitment, Even under colonisation favour","Even under conditions that eventually favour sexual recruitment, colonisation starts with an asexual wave.",25785
"a after sexual colonisation erodes dominance. is wave clonal completed, Long","Long after colonisation is completed, a sexual wave erodes clonal dominance.",25786
"are in few-shot interested In learning paper, problem. we the this","In this paper, we are interested in the few-shot learning problem.",25787
"recognition number tasks, \textit{e.g. A of pattern","A number of pattern recognition tasks, \textit{e.g.",25788
"learning proposed characteristics feature learning, vMF the achieves The of discriminative \textit{i.e.","The proposed vMF feature learning achieves the characteristics of discriminative learning, \textit{i.e.",25789
"faces, YouTube and IJB-A. LFW, CACD },","}, LFW, YouTube faces, CACD and IJB-A.",25790
properties. several presents The model proposed important,The proposed model presents several important properties.,25791
"criteria, which often metadata, are collect. Large hand-labeled are or databases expensive to by organized","Large databases are often organized by hand-labeled metadata, or criteria, which are expensive to collect.",25792
segmentation and until convergence. is the iteratively graph dynamically GC The based progresses localized updated,The localized graph is dynamically updated and the GC based segmentation iteratively progresses until convergence.,25793
"study with the of case illustrate we in Finally, details pancreatic approach a our cancer.","Finally, we illustrate the details of our approach with a case study in pancreatic cancer.",25794
"are defined that as goal-directed. and perceive, act, entities Agents commonly are","Agents are commonly defined as entities that act, perceive, and are goal-directed.",25795
completely automaton) a dynamics. transition cellular the the determines matrix,a cellular automaton) the transition matrix completely determines the dynamics.,25796
a within acting such possibility to This system. of entities contradict seems the,This seems to contradict the possibility of acting entities within such a system.,25797
entity-sets. based multivariate on within perceptions chains we and Here Markov actions definitions present of,Here we present definitions of actions and perceptions within multivariate Markov chains based on entity-sets.,25798
"loop the is multivariate also an perception-action entity-set and chain. Importantly, a Markov induces","Importantly, the perception-action loop also induces an entity-set and is a multivariate Markov chain.",25799
is rapidly The in growing to Sweden immigration. due population,The population in Sweden is growing rapidly due to immigration.,25800
"of importance. of upgrades services telecommunication issue this light, provide is infrastructure to In the","In this light, the issue of infrastructure upgrades to provide telecommunication services is of importance.",25801
"paper, the we In second explore this alternative.","In this paper, we explore the second alternative.",25802
in correlation have their geo-demographic Infrastructure-Stressing light population with studied We segments. the of the,We have studied the Infrastructure-Stressing population in the light of their correlation with geo-demographic segments.,25803
to day. per requires all Obesity intakes treatment patients food obese record,Obesity treatment requires obese patients to record all food intakes per day.,25804
Computer vision calories estimate has introduced images. from food to been,Computer vision has been introduced to estimate calories from food images.,25805
"of and top is view calorie view side a needed. estimate food, To","To estimate calorie of food, a top view and side view is needed.",25806
is food object. calibration Faster used detect the and to R-CNN,Faster R-CNN is used to detect the food and calibration object.,25807
used each GrabCut to algorithm contour. food's get is,GrabCut algorithm is used to get each food's contour.,25808
and is food estimated corresponding Then with the the object. volume,Then the volume is estimated with the food and corresponding object.,25809
is experiment estimation the method results And show our effective.,And the experiment results show our estimation method is effective.,25810
"physical information world. still detailed Videos, contain the a wealth images, of unlike about","Videos, unlike still images, contain a wealth of detailed information about the physical world.",25811
the We this data crowd-sourcing describe also challenges in scale. at,We also describe the challenges in crowd-sourcing this data at scale.,25812
inference. both is used model same for training and The,The same model is used both for training and inference.,25813
configurations. classification give structures the Blum of for of properties generic We a linking,We give a classification of the properties of Blum linking structures for generic configurations.,25814
"and from EEG collected paper, AD we subjects. In patients data resting-state this control healthy","In this paper, we collected EEG resting-state data from AD patients and healthy control subjects.",25815
"differences properties, the graph-theoretical network in assess analysis performed. was To","To assess the differences in network properties, graph-theoretical analysis was performed.",25816
"In based this hierarchical novel meta-descriptors paper, present distribution pixel features. we of a on","In this paper, we present novel meta-descriptors based on a hierarchical distribution of pixel features.",25817
Gaussians by another of Gaussian The set of are again characteristics distribution. described the,The characteristics of the set of Gaussians are again described by another Gaussian distribution.,25818
"decision. is print a a collecting crucial images muzzle Therefore, database","Therefore, collecting a muzzle print images database is a crucial decision.",25819
has collected images. to A of quality special the been the care given,A special care has been given to the quality of the collected images.,25820
"data often not either published or experimental results, reproducing However, observations is viable.","However, reproducing published results, either experimental data or observations is often not viable.",25821
a Image analysis. fundamental segmentation biomedical in is image problem,Image segmentation is a fundamental problem in biomedical image analysis.,25822
benchmarks. learning in promising Recent advances achieved deep image on have results biomedical many segmentation,Recent advances in deep learning have achieved promising results on many biomedical image segmentation benchmarks.,25823
between allows faces. retains of clusters of dynamic number selection a This similarity pairwise and,This allows a dynamic selection of number of clusters and retains pairwise similarity between faces.,25824
"convolutional vision. in powerful paradigm Recently, (CNN) network neural a become has computer","Recently, convolutional neural network (CNN) has become a powerful paradigm in computer vision.",25825
outperforms in a margin methods model available by Our datasets. publicly large existing,Our model outperforms existing methods by a large margin in publicly available datasets.,25826
"progress limited. on task-to-task Nonetheless, remains transfer","Nonetheless, progress on task-to-task transfer remains limited.",25827
recognizing at classification aims Gender a gender. person's,Gender classification aims at recognizing a person's gender.,25828
brain is the the of In memory. easy identify physical to not it substrate,In the brain it is not easy to identify the physical substrate of memory.,25829
previous excellent state-of-the-art obtain results improve results. significantly over correspondence that We experimental shape,We obtain excellent experimental results that significantly improve over previous state-of-the-art shape correspondence results.,25830
model a refer Bags. as We System a such to of,We refer to such a model as a System of Bags.,25831
Parameters linear dynamic from suffer of ambiguities. systems,Parameters of linear dynamic systems suffer from ambiguities.,25832
Bags. as dissimilarity of We employ it on measure a Systems,We employ it as a dissimilarity measure on Systems of Bags.,25833
"our Messidor. outperform methods and EyePACS datasets, the on that Experiments two algorithm state-of-the-art show","Experiments show that our algorithm outperform the state-of-the-art methods on two datasets, EyePACS and Messidor.",25834
the between increasing inherent granularity. advantage at We hierarchy labels take of,We take advantage of the inherent hierarchy between labels at increasing granularity.,25835
"against demonstrate of results effectiveness models. significant achieving the baseline Our BINN, improvements","Our results demonstrate the effectiveness of BINN, achieving significant improvements against baseline models.",25836
"subsequent DCNN for positive Then, reduction. the presented is a false three-dimensional","Then, a three-dimensional DCNN is presented for the subsequent false positive reduction.",25837
been problem studied classification Video has years. many,Video classification problem has been studied many years.,25838
need of dataset Both data. approaches large input a,Both approaches need a large dataset of input data.,25839
approach. on The pooling is frame first based model,The first model is based on frame pooling approach.,25840
Two LSTM based other on models networks.,Two other models based on LSTM networks.,25841
been imbalanced set training handling conducted. for The experiments of has data,The set of experiments for handling imbalanced training data has been conducted.,25842
introduce class net to forms. model penalty both an applies We that elastic,We introduce an elastic net penalty class that applies to both model forms.,25843
"be ordinal toward can counterpart. this a model its Additionally, penalty shrink used to non-ordinal","Additionally, this penalty can be used to shrink a non-ordinal model toward its ordinal counterpart.",25844
"which the implements model for the ordinalNet, introduce package this we Finally, class. R algorithm","Finally, we introduce the R package ordinalNet, which implements the algorithm for this model class.",25845
"in high-level vision handled are image separately vision. and Conventionally, computer tasks denoising","Conventionally, image denoising and high-level vision tasks are handled separately in computer vision.",25846
convolutional First image a network propose state-of-the-art denoising neural performance. for which achieves the we,First we propose a convolutional neural network for image denoising which achieves the state-of-the-art performance.,25847
methods. with demonstrate our recent images results state-of-the-art We approach on compare and face,We demonstrate results on face images and compare our approach with recent state-of-the-art methods.,25848
self-ensembling the domain problems. This of paper for use visual explores adaptation,This paper explores the use of self-ensembling for visual domain adaptation problems.,25849
of its transition sampler bouncy in dynamics. We terms generalize particle,We generalize bouncy particle sampler in terms of its transition dynamics.,25850
inefficiently entire temporal video detectors for search specific the State-of-the-art action actions.,State-of-the-art temporal action detectors inefficiently search the entire video for specific actions.,25851
We application an problem. temporal consider localization of as action the action spotting,We consider temporal action localization as an application of the action spotting problem.,25852
to weighted networks extendable also edges. framework count-valued Our easily is with model,Our framework is also easily extendable to model weighted networks with count-valued edges.,25853
a benchmark is challenge in for large-scale multi-label The classification. dataset this video used,The dataset used in this challenge is a large-scale benchmark for multi-label video classification.,25854
that better the interactions We Chaining capture labels. propose network structure called can between a,We propose a network structure called Chaining that can better capture the interactions between labels.,25855
"our Also, report multi-scale we approaches with and in information dealing attention pooling.","Also, we report our approaches in dealing with multi-scale information and attention pooling.",25856
"vectors contrast extracted information. superpixel are texture cover color, to from regional Feature each and","Feature vectors are extracted from each superpixel to cover regional color, contrast and texture information.",25857
representing highly The geometric of framework spaces a provides conceptual influential of way knowledge.,The highly influential framework of conceptual spaces provides a geometric way of representing knowledge.,25858
It symbolic at between gap aims subsymbolic bridging and processing. the,It aims at bridging the gap between symbolic and subsymbolic processing.,25859
"that the By we get experimental find results. comparing results, SCDMIs better","By comparing the experimental results, we find that SCDMIs get better results.",25860
been and learning. synthesis has Face fascinating yet a problem challenging machine computer vision in,Face synthesis has been a fascinating yet challenging problem in computer vision and machine learning.,25861
"results. to photo-realistic Generative preserving is (GAN) and identity highlighted Particularly, Net generate Adversarial","Particularly, Generative Adversarial Net (GAN) is highlighted to generate photo-realistic and identity preserving results.",25862
"details. introduced are evaluation databases and public the metrics Furthermore, in available","Furthermore, the public available databases and evaluation metrics are introduced in details.",25863
promising discussing research. and future for directions unsolved difficulties with the end review We,We end the review with discussing unsolved difficulties and promising directions for future research.,25864
"process, results an is iterative relying new previous with building incremental, upon ones. and Research","Research is an incremental, iterative process, with new results relying and building upon previous ones.",25865
"We the automatically discovering, organizing, present retrieving for trackr results. framework and annotating,","We present the trackr framework for organizing, automatically annotating, discovering, and retrieving results.",25866
computing. image synthesis a Cross-modal topical problem image medical in is,Cross-modal image synthesis is a topical problem in medical image computing.,25867
and in challenging two DOTE tasks: image extensively We evaluate synthesis. cross-modality super-resolution,We extensively evaluate DOTE in two challenging tasks: image super-resolution and cross-modality synthesis.,25868
methods. results state-of-the-art performance superior method over The of experimental other our demonstrate,The experimental results demonstrate superior performance of our method over other state-of-the-art methods.,25869
a provides interface. for with surgical informative navigation The system signals,The system provides informative signals for navigation with a surgical interface.,25870
ex animal/ vivo/in Validation demonstration and system is experiments. this on human vivo reported of,Validation and demonstration of this system is reported on ex vivo/in vivo animal/ human experiments.,25871
optical robotics. potentially in useful (OF) power flow low of computation is Rapid and,Rapid and low power computation of optical flow (OF) is potentially useful in robotics.,25872
software in FPGA. algorithm implemented The was both on and,The algorithm was implemented both in software and on FPGA.,25873
interval. by of sample the set speed The the is motion,The speed of the motion is set by the sample interval.,25874
of cooperative joint reinforcement multi-agent with signal. We problem a study the reward learning single,We study the problem of cooperative multi-agent reinforcement learning with a single joint reward signal.,25875
with user define constraints must that applications preferences the deal the These undoubtedly and interests.,These applications must undoubtedly deal with the constraints and preferences that define the user interests.,25876
This via strategy. pruning filter models compress simultaneously aims and accelerate CNN paper off-the-shelf to,This paper aims to simultaneously accelerate and compress off-the-shelf CNN models via filter pruning strategy.,25877
each method importance first. The entropy-based evaluated the proposed filter is by of,The importance of each filter is evaluated by the proposed entropy-based method first.,25878
to get CNN a unimportant are smaller discarded model. several Then filters,Then several unimportant filters are discarded to get a smaller CNN model.,25879
"pruning. during its adopted filter recover is fine-tuning to generalization is ability Finally, damaged which","Finally, fine-tuning is adopted to recover its generalization ability which is damaged during filter pruning.",25880
"with previous filter better our obtains importance method performance. criteria, Compared evaluation entropy-based","Compared with previous filter importance evaluation criteria, our entropy-based method obtains better performance.",25881
method an increase the can demonstrate that also selection pair significantly We effective performance.,We also demonstrate that an effective pair selection method can significantly increase the performance.,25882
"training In and overall the modifications we pipeline, objective training to addition, e.g. several defined","In addition, we defined several modifications to the training objective and overall training pipeline, e.g.",25883
"reflectance discuss data usage strategies and of indices. augmentation boundary also effect we estimation,","boundary effect estimation, also we discuss usage of data augmentation strategies and reflectance indices.",25884
Evolution of body and plans sculpts nervous both systems over time. together agents the,Evolution sculpts both the body plans and nervous systems of agents together over time.,25885
simple other conclude memory perform than of approaches. We relatively versions that networks better,We conclude that relatively simple versions of memory networks perform better than other approaches.,25886
"adequate results with shortcomings. it in this Even cases, provides though most several comes","Even though this provides adequate results in most cases, it comes with several shortcomings.",25887
The to are designed and require carefully methods outliers. regularizers prone they usually,The methods require carefully designed regularizers and they are usually prone to outliers.,25888
preserving patterns trees. on This privacy paper inducing the focuses of when sensitive decision,This paper focuses on preserving the privacy of sensitive patterns when inducing decision trees.,25889
binary for We a in datasets. sensitive adopt hiding approach record classification augmentation rules,We adopt a record augmentation approach for hiding sensitive classification rules in binary datasets.,25890
"and based two method The phases, segmentation. localisation is proposed on","The proposed method is based on two phases, localisation and segmentation.",25891
The delineate region interest. active phase the of contour model to utilises second,The second phase utilises active contour model to delineate the region of interest.,25892
Obtained approach the validity of the its results show robustness. proposed and,Obtained results show the validity of the proposed approach and its robustness.,25893
surface. at exposed cortical measurements direct of provides Electrocorticography (ECoG) postsynaptic the synchronized potentials,Electrocorticography (ECoG) provides direct measurements of synchronized postsynaptic potentials at the exposed cortical surface.,25894
component that multiple coefficients. is first unknown simultaneously The the predicts layer a (MLP) perceptron,The first component is a multiple layer perceptron (MLP) that simultaneously predicts the unknown coefficients.,25895
existing Practical fully methods. applications several not by impose are challenges which addressed,Practical applications impose several challenges which are not fully addressed by existing methods.,25896
and scenes are Especially models. scenes in there between cluttered many feature mismatches,Especially in cluttered scenes there are many feature mismatches between scenes and models.,25897
patches using the each on scatter a local We matrix's LRF surface propose eigenvectors.,We propose a LRF on each local surface patches using the scatter matrix's eigenvectors.,25898
excellent image neural potential carry fast out for Fully deep networks convolutional segmentation. accurate and,Fully convolutional deep neural networks carry out excellent potential for fast and accurate image segmentation.,25899
"state-of-the-art yields text of inclusion The results, e.g.","The inclusion of text yields state-of-the-art results, e.g.",25900
data set state-of-the-art competing methods. compared is FERET with Our evaluated method and proposed on,Our proposed method is evaluated on FERET data set and compared with state-of-the-art competing methods.,25901
Enhancing at be can challenging model performance times. a,Enhancing a model performance can be challenging at times.,25902
Detecting and resolution is objects challenging small notoriously representation. to low their noisy due,Detecting small objects is notoriously challenging due to their low resolution and noisy representation.,25903
from problem is constrained dominant graph. a caste finding tracking The sets as,The tracking problem is caste as finding constrained dominant sets from a graph.,25904
invariance have image invariance Rotation translation values in great recognition tasks. and,Rotation invariance and translation invariance have great values in image recognition tasks.,25905
of autonomous for collision Accurate prediction prerequisite is avoidance traffic participant the vehicles.,Accurate traffic participant prediction is the prerequisite for collision avoidance of autonomous vehicles.,25906
"work, by pedestrians predict this emulating own we motion planning. their In","In this work, we predict pedestrians by emulating their own motion planning.",25907
"function for infer a From possible density we mixture online observations, destinations.","From online observations, we infer a mixture density function for possible destinations.",25908
and implementation our our system. training also elaborate experience We details on on share,We also elaborate on implementation details and share our experience on training our system.,25909
models multiple time It (e.g. underlying series dynamics that patterns exhibit categorical-valued from,It models categorical-valued time series that exhibit dynamics from multiple underlying patterns (e.g.,25910
world both and data. based on Our synthetic- empirical results are real,Our empirical results are based on both synthetic- and real world data.,25911
enhance the struggled efficiency to grouped using researchers have convolution. parameters' Some,Some researchers have struggled to enhance the parameters' efficiency using grouped convolution.,25912
in new This introduces framework decision a for real-time paper video games. making,This paper introduces a new framework for real-time decision making in video games.,25913
"model to behaves the it is how important globally. also However, understand","However, it is also important to understand how the model behaves globally.",25914
"scales visual (VAS). commonly subjective evaluated Pain analog personal, a through experience is is that","Pain is a personal, subjective experience that is commonly evaluated through visual analog scales (VAS).",25915
"location to the supervision. problem However, challenging of lack is due quite this","However, this problem is quite challenging due to the lack of location supervision.",25916
"posals maps. Specifically, object highly pro- we select guidance saliency confident the of class-specific under","Specifically, we select highly confident object pro- posals under the guidance of class-specific saliency maps.",25917
"in sub-network Meanwhile, is prediction a the architecture. built saliency","Meanwhile, a saliency prediction sub-network is built in the architecture.",25918
are implicitly The prediction procedure. the localization guide results to used,The prediction results are used to implicitly guide the localization procedure.,25919
Experiments state-of-the-arts. all on VOC outperforms PASCAL our demonstrate approach that,Experiments on PASCAL VOC demonstrate that our approach outperforms all state-of-the-arts.,25920
the SqueezeNet. and on the detectors the is other based of One the XNOR-Net on,One of the detectors is based on the XNOR-Net and the other on the SqueezeNet.,25921
this other quantitatively compared methods qualitatively are Multiple also in and study.,Multiple other methods are also quantitatively and qualitatively compared in this study.,25922
"usually outliers. contain In real-world applications, many data","In many real-world applications, data usually contain outliers.",25923
to VORPCA matrix VOR Applying to a new leads factorization model.,Applying VOR to matrix factorization leads to a new VORPCA model.,25924
accurate real-valued. the The is reconstruction yields that most data,The data that yields the most accurate reconstruction is real-valued.,25925
"are Thus, mandatory to models quantisation encoding. efficient an allow","Thus, quantisation models are mandatory to allow an efficient encoding.",25926
clustering understood These as be also can problems. challenging data,These can also be understood as challenging data clustering problems.,25927
to make predict it which facts hard works feature These best.,These facts make it hard to predict which feature works best.,25928
we In on quantisation this such based popular discuss k-means. as paper methods strategies,In this paper we discuss quantisation strategies based on popular methods such as k-means.,25929
"image translate compression directly a good benefits to do performance. Surprisingly, these not","Surprisingly, these benefits do not directly translate to a good image compression performance.",25930
costs. gains ratio storage compression are to lost in The the increased due,The gains in the compression ratio are lost due to increased storage costs.,25931
"as ""black described are boxes"". often Machine-learned models","Machine-learned models are often described as ""black boxes"".",25932
We demonstrate an of approach using the online validity advertising our application.,We demonstrate the validity of our approach using an online advertising application.,25933
skin pose studied systems. challenges deformation the to reconstruction Body homogeneous texture and,Body deformation and homogeneous skin texture pose challenges to the studied reconstruction systems.,25934
two-stream a dynamic model We for introduce texture synthesis.,We introduce a two-stream model for dynamic texture synthesis.,25935
"study. our we Finally, user texture approach with quantitatively thorough a evaluate synthesis","Finally, we quantitatively evaluate our texture synthesis approach with a thorough user study.",25936
"processing. synthesis field Texture of graphics, in image and computer vision, is used the widely","Texture synthesis is widely used in the field of computer graphics, vision, and image processing.",25937
damaged can be either Each node (i.e.,Each node can be either damaged (i.e.,25938
further nodes of local connected Damage damage. facilitates,Damage of connected nodes facilitates further local damage.,25939
programmed of any mortality recovers model patterns ageing. without frailty the known Our and,Our model recovers the known patterns of frailty and mortality without any programmed ageing.,25940
"new theory. to including us Large exploit allow tools, populations analytic information model","Large model populations allow us to exploit new analytic tools, including information theory.",25941
probabilities. confidence of class is Classification dispersion the estimated by analyzing,Classification confidence is estimated by analyzing the dispersion of class probabilities.,25942
coding (CSC) learning data. dictionary sparse a coding Convolutional the sparse by from shift-invariant improves,Convolutional sparse coding (CSC) improves sparse coding by learning a shift-invariant dictionary from the data.,25943
"alleviate In these this paper, learning. we using problems online by","In this paper, we alleviate these problems by using online learning.",25944
R work This the presents with to distributions. paper gk package these,This paper presents the gk R package to work with these distributions.,25945
is compact data meta-program theories that Julia if they Presented from discovers a exist.,Presented is a Julia meta-program that discovers compact theories from data if they exist.,25946
is as metric: of the Compactness a measured space-time derivatives. number such by,Compactness is measured by a metric: such as the number of space-time derivatives.,25947
this scenes. clustered refrigerator detecting pack- aged apply We in strategy food for products,We apply this strategy for detecting pack- aged food products clustered in refrigerator scenes.,25948
achieved have networks success in recent years. a great neural Convolutional the,Convolutional neural networks have achieved a great success in the recent years.,25949
conceptual knowledge. of representing influential The highly way geometric framework a of spaces provides,The highly influential framework of conceptual spaces provides a geometric way of representing knowledge.,25950
variety imitating proven textures. Texture wide of successful at has synthesis a,Texture synthesis has proven successful at imitating a wide variety of textures.,25951
and mini-batches. stochastic gradients efficient using enables training This,This enables efficient training using stochastic gradients and mini-batches.,25952
"in make independent extra process Furthermore, the cost approximate memory assumptions of inference $N$. the","Furthermore, extra assumptions in the approximate inference process make the memory cost independent of $N$.",25953
automation. Studies jobs estimated of have around of at all current half risk are that,Studies have estimated that around half of all current jobs are at risk of automation.,25954
later typically decades were These than the those estimates non-experts. of,These estimates were typically decades later than those of the non-experts.,25955
tasks maps cognitive integration Performing structure. level different feature drastically requires the with of high,Performing high level cognitive tasks requires the integration of feature maps with drastically different structure.,25956
"this multi-dimensional a tensors. of ({MD-sketch}), extension to sketch count-sketch introduce novel we work, In","In this work, we introduce multi-dimensional sketch ({MD-sketch}), a novel extension of count-sketch to tensors.",25957
"layer estimating For another evidence adds models, difficulty. the doubly-intractable of","For doubly-intractable models, estimating the evidence adds another layer of difficulty.",25958
"the agent's utilize learn to the reinforcement Q-learning Deep decision algorithm learning, We especially policy.","We utilize reinforcement learning, especially the Deep Q-learning algorithm to learn the agent's decision policy.",25959
various (explanatory) similarity measures feature variables proteins. available of are classification for of The,The feature (explanatory) variables available for classification are various measures of similarity of proteins.,25960
"Homologous highly proteins problem. giving in rare two-class a single are candidate a set, unbalanced","Homologous proteins are rare in a single candidate set, giving a highly unbalanced two-class problem.",25961
variables. The of diverse proposed ensemble feature subsets identifies and of strong phalanxes,The proposed ensemble of phalanxes identifies strong and diverse subsets of feature variables.,25962
A of evaluation aggregates metrics. ensembling based second phase diverse classifiers on,A second phase of ensembling aggregates classifiers based on diverse evaluation metrics.,25963
phalanxes metrics. overall of and ensemble called is The result an,The overall result is called an ensemble of phalanxes and metrics.,25964
and against both close distant It homologues. robustness provide,It provide robustness against both close and distant homologues.,25965
for generating We new propose a system art.,We propose a new system for generating art.,25966
subjects on Human images generated scales. even the rated various higher,Human subjects even rated the generated images higher on various scales.,25967
methods for pre-trained features using frame-level video neural Current networks extract often (CNNs). convolutional analysis,Current methods for video analysis often extract frame-level features using pre-trained convolutional neural networks (CNNs).,25968
for video temporal revise methods representations work and aggregation. existing In we study alternative this,In this work we revise existing video representations and study alternative methods for temporal aggregation.,25969
Our experimental classification. for results show the video task the both improvements advantage of of,Our experimental results show the advantage of both improvements for the task of video classification.,25970
"mechanisms Although this its effect computational poorly is are underlying well `learning-to-learn' documented, understood.","Although this `learning-to-learn' effect is well documented, its underlying computational mechanisms are poorly understood.",25971
case from presented. located Results studies Switzerland two are in,Results from two case studies located in Switzerland are presented.,25972
"unsupervised In domain Transform the this for we Embedding introduce Nonlinear adaptation. paper, (NET)","In this paper, we introduce the Nonlinear Embedding Transform (NET) for unsupervised domain adaptation.",25973
alignment. disparity through reduces The NET cross-domain domain nonlinear,The NET reduces cross-domain disparity through nonlinear domain alignment.,25974
are also data It embeds that such domain-aligned clustered the together. similar data points,It also embeds the domain-aligned data such that similar data points are clustered together.,25975
answer question: How one similar embeddings simple two are Deep images?,Deep embeddings answer one simple question: How similar are two images?,25976
"bedrock visual these the verification, search. of learning, embeddings zero-shot is Learning and","Learning these embeddings is the bedrock of verification, zero-shot learning, and visual search.",25977
"and informative approaches. selects than sampling, traditional distance weighted more We which examples stable propose","We propose distance weighted sampling, which selects more informative and stable examples than traditional approaches.",25978
achieves Our on them. state-of-the-art performance all method of,Our method achieves state-of-the-art performance on all of them.,25979
the on semantic high-level since Those is color dependent semantics-aware mapping are methods context. the,Those methods are semantics-aware since the color mapping is dependent on the high-level semantic context.,25980
"that neural semantics-aware models photo this adjustment. propose In paper, we a the deep network","In this paper, we propose a deep neural network that models the semantics-aware photo adjustment.",25981
proposed parsing method robust task. is scene with trained using a The loss a,The proposed method is trained using a robust loss with a scene parsing task.,25982
the results on effectiveness datasets Experimental of two benchmark method. validate proposed the,Experimental results on two benchmark datasets validate the effectiveness of the proposed method.,25983
is various fields. technique eye tracking Video-based valuable a research in,Video-based eye tracking is a valuable technique in various research fields.,25984
"we to restrict contrast, do specific not reflectance model. ourselves In a","In contrast, we do not restrict ourselves to a specific reflectance model.",25985
"a of works we reflectances. that Instead, offer wide a variety method on","Instead, we offer a method that works on a wide variety of reflectances.",25986
on that unit method We normals a novel to hemisphere. maps present pixels embedding the,We present a novel embedding method that maps pixels to normals on the unit hemisphere.,25987
"In the and, initialized are then model poorly undergoes samples first training, improvement. easy generated","In training, easy samples are generated first and, then the poorly initialized model undergoes improvement.",25988
"but becomes challenging are model selected. more reliable As the discriminative, samples","As the model becomes more discriminative, challenging but reliable samples are selected.",25989
"round of model After takes place. another that, improvement","After that, another round of model improvement takes place.",25990
Variability drug in practice. effects efficacy in clinical observed and adverse are,Variability in drug efficacy and adverse effects are observed in clinical practice.,25991
learning automatic novel of architectures non-rigid We segmentation propose deep for surgical instruments. two,We propose two novel deep learning architectures for automatic segmentation of non-rigid surgical instruments.,25992
constraint the The methods proposed architecture. encode inside network multi-scale the two,The two proposed methods encode the multi-scale constraint inside the network architecture.,25993
compare convolutional the state-of-the-art We networks. fully architectures against proposed,We compare the proposed architectures against state-of-the-art fully convolutional networks.,25994
Coefficient methods. Our over results previous segmentation improved instrument show Dice a significant statistically Similarity,Our results show a statistically significant improved Dice Similarity Coefficient over previous instrument segmentation methods.,25995
improving and discuss design We choices our for key drivers analyze the accuracy.,We analyze our design choices and discuss the key drivers for improving accuracy.,25996
field systems. The the empirical work of vision in and situated is it is computer,The work is empirical and it is situated in the field of computer vision systems.,25997
object reveal Fine-grained features from between which benefit part-based can subtle visual categorization differences categories.,Fine-grained categorization can benefit from part-based features which reveal subtle visual differences between object categories.,25998
used Handcrafted features part have widely and been for classification. detection,Handcrafted features have been widely used for part detection and classification.,25999
is organ diagnosis. of region an its Automatic of prerequisite cystic computer-aided a and segmentation,Automatic segmentation of an organ and its cystic region is a prerequisite of computer-aided diagnosis.,26000
"CT segmentation focus in on In cyst this abdominal pancreatic paper, scan. we","In this paper, we focus on pancreatic cyst segmentation in abdominal CT scan.",26001
from set correspondences. of are used criteria estimating an Different for error input FMs,Different error criteria are used for estimating FMs from an input set of correspondences.,26002
"In the of error were the paper, this different efficiency aspects and studied. criteria accuracy","In this paper, the accuracy and efficiency aspects of the different error criteria were studied.",26003
"FM and RE-CG reprojection an value error desired takes As a $d$. input,","As input, RE-CG takes an FM and a desired reprojection error value $d$.",26004
"having error generates a random As output, that correspondence RE-CG value.","As output, RE-CG generates a random correspondence having that error value.",26005
"based enhancement. Finally, we detection develop and for algorithm fiber these on principles, an","Finally, based on these principles, we develop an algorithm for fiber detection and enhancement.",26006
exemplify microscopy to reveal images. nerve enhance and light its We utility in axons,We exemplify its utility to reveal and enhance nerve axons in light microscopy images.,26007
efforts (AUs). units action devoted to recognizing facial Extensive have been,Extensive efforts have been devoted to recognizing facial action units (AUs).,26008
this of unique strategy information addition not the In opponent's does assess state.,In addition this unique strategy does not assess information of the opponent's state.,26009
of evolutionary method analysis the stability. standard bypasses The,The method bypasses the standard analysis of evolutionary stability.,26010
findings could of kinases some These salient protein conventional C. . features explain,These findings could explain some salient features of conventional protein kinases C. .,26011
agreed measurements (e.g. study Cardiac with report structure values,Cardiac structure measurements agreed with study report values (e.g.,26012
"we introduce Office-Home, this domain adaptation dataset, algorithms. to a paper, In new evaluate first","In this paper, we first introduce a new dataset, Office-Home, to evaluate domain adaptation algorithms.",26013
images contains variety from everyday of dataset multiple objects of domains. a The,The dataset contains images of a variety of everyday objects from multiple domains.,26014
Convolutional kernels Neural vital basic Networks of components deep and Convolutional (CNN). are,Convolutional kernels are basic and vital components of deep Convolutional Neural Networks (CNN).,26015
"learnable to other are addition in shapes weights. In words, parameters","In other words, shapes are learnable parameters in addition to weights.",26016
validate to our ICNN. the semantic proposed Experiments implemented of are for effectiveness segmentation,Experiments for semantic segmentation are implemented to validate the effectiveness of our proposed ICNN.,26017
been structural helps understand Investigating how time. genomes these rearrangements modified to over have,Investigating these structural rearrangements helps to understand how genomes have been modified over time.,26018
restoration. cranial implants surgery of are necessary Patient-specific the cranial defect important and in,Patient-specific cranial implants are important and necessary in the surgery of cranial defect restoration.,26019
"design cranial are and methods However, implants complicated of traditional manual time-consuming. of","However, traditional methods of manual design of cranial implants are complicated and time-consuming.",26020
a convolutional neural for Rectified networks. (ReLU) widely deep unit is linear function used activation,Rectified linear unit (ReLU) is a widely used activation function for deep convolutional neural networks.,26021
"negative the networks the ReLU benefits miss zero-hard because rectification, from of However, values.","However, because of the zero-hard rectification, ReLU networks miss the benefits from negative values.",26022
less uncertain. have States observed more are that frequently features deemed,States that have less frequently observed features are deemed more uncertain.,26023
Aided Design. Design used Geometric and curve Aided in widely Computer Computer is NURBS,NURBS curve is widely used in Computer Aided Design and Computer Aided Geometric Design.,26024
computer-assisted a Brain key plays tumour segmentation surgery. in role,Brain tumour segmentation plays a key role in computer-assisted surgery.,26025
derived equation malaria. disease an to blood-stage capture differential ordinary dynamics during We model the,We derived an ordinary differential equation model to capture the disease dynamics during blood-stage malaria.,26026
model age-structured was an equation The directly model. differential earlier from derived partial,The model was directly derived from an earlier age-structured partial differential equation model.,26027
The original was due simplified model experimental to constraints.,The original model was simplified due to experimental constraints.,26028
the using with objective a we Here multiple model calibrated data simplified genetic algorithm. experimental,Here we calibrated the simplified model with experimental data using a multiple objective genetic algorithm.,26029
"particular, be or may the subspace a In concentrated lower-dimensional near data manifold.","In particular, the data may be concentrated near a lower-dimensional subspace or manifold.",26030
components analysis (SPCA). We spherical principal develop,We develop spherical principal components analysis (SPCA).,26031
to recent show in performance We state-of-the-art approaches. comparison,We show state-of-the-art performance in comparison to recent approaches.,26032
detection introduces approach paper new This real-time Yes-Net. a object named,This paper introduces a new real-time object detection approach named Yes-Net.,26033
various applications. emerge of optimal practical choice Problems in often,Problems of optimal choice emerge often in various practical applications.,26034
We the to two solve propose problem. approaches,We propose two approaches to solve the problem.,26035
reach processing solutions good Both application. of a on real signal them data life from,Both of them reach good solutions on real life data from a signal processing application.,26036
learning activity We salamander driven retina. spiking in by simulate recorded,We simulate learning driven by spiking activity recorded in salamander retina.,26037
spike timing-dependent to change weights learning according during rules period. a training Input,Input weights change according to spike timing-dependent learning rules during a training period.,26038
covered Images of variation. world clinical real range a,Images covered a range of real world clinical variation.,26039
"neural In views classify simultaneously can networks accuracy. deep echocardiographic high conclusion, essential with and","In conclusion, deep neural networks can classify essential echocardiographic views simultaneously and with high accuracy.",26040
Our interpretation. deep assisted for a results learning complex provide echocardiographic more foundation,Our results provide a foundation for more complex deep learning assisted echocardiographic interpretation.,26041
a progress significant Human last in Recognition the witnessed has decade. Activity,Human Activity Recognition has witnessed a significant progress in the last decade.,26042
is learning highest the and more classification Results has accuracy. show deep effective that,Results show that deep learning is more effective and has the highest classification accuracy.,26043
Practice Under Programming (TPLP) . and Theory consideration of in Logic,Under consideration in Theory and Practice of Logic Programming (TPLP) .,26044
"for for such, great potential low they power robots. and have fast As vision algorithms","As such, they have great potential for fast and low power vision algorithms for robots.",26045
iCub We robot. data algorithm from neuromorphic validate the the collected on,We validate the algorithm on data collected from the neuromorphic iCub robot.,26046
alternates the the Our and method shape labels. updating between semantic,Our method alternates between updating the shape and the semantic labels.,26047
"semantic prior surface this assumptions classes. includes of shape about different step Also, the","Also, this step includes prior assumptions about the surface shape of different semantic classes.",26048
"This in infections. experiment two resulted severe mild two and infections,","This experiment resulted in two severe infections, and two mild infections.",26049
"detections. decade, Neural for Convolutional demonstrated object (CNNs) In Networks the successful have been past","In the past decade, Convolutional Neural Networks (CNNs) have been demonstrated successful for object detections.",26050
"objects. degrades Moreover, detecting small when performance","Moreover, performance degrades when detecting small objects.",26051
achieved pyramid. the by on an SOS-CNN applying Scale invariance is image,Scale invariance is achieved by applying the SOS-CNN on an image pyramid.,26052
learning zero-shot other not in generalize techniques well Prevalent problem scenarios. related to do,Prevalent techniques in zero-shot learning do not generalize well to other related problem scenarios.,26053
"problems. few-shot unified zero-shot, present and conventional we a generalized Here, approach learning for zero-shot","Here, we present a unified approach for conventional zero-shot, generalized zero-shot and few-shot learning problems.",26054
"class. for Given direction produces method an image, one our principal each seen","Given an image, our method produces one principal direction for each seen class.",26055
transfer information to class-adaptive from allows classes. and efficient This unseen seen,This allows efficient and class-adaptive information transfer from seen to unseen classes.,26056
animals by Learning occurs supervision. and largely and children effortlessly without obvious,Learning by children and animals occurs effortlessly and largely without obvious supervision.,26057
correlation. the mutual Dependence be can measure information information-theoretic characterized total called multivariate using,Dependence can be characterized using the information-theoretic multivariate mutual information measure called total correlation.,26058
approach formulation. We multi-object network apply tracking flow with to a this,We apply this approach to multi-object tracking with a network flow formulation.,26059
to resources individuals responses describe in between used ecology. are exchange Functional interactions and widely,Functional responses are widely used to describe interactions and resources exchange between individuals in ecology.,26060
"Our framework applications many empirical conceptual, has theoretical purposes. and for","Our framework has many applications for conceptual, theoretical and empirical purposes.",26061
SR in context deep the learning. surveys the paper of literature This,This paper surveys the SR literature in the context of deep learning.,26062
observations. from the noisy tracking estimating of lane Automatic underlying sequence involves signal a signal,Automatic lane tracking involves estimating the underlying signal from a sequence of noisy signal observations.,26063
well that on models. is works used a Filter Gaussian widely Kalman method The linear,The Kalman Filter is a widely used method that works well on linear Gaussian models.,26064
Filter a a observation We Particle multiple using model. of top on novel propose mode,We propose using a Particle Filter on top of a novel multiple mode observation model.,26065
Experiments our that performance Filter. Kalman conventional show a method produces to superior,Experiments show that our method produces superior performance to a conventional Kalman Filter.,26066
"sufficient considerable be Using boxes annotated for we find approach, gains. to this performance weakly","Using this approach, we find weakly annotated boxes to be sufficient for considerable performance gains.",26067
Gaussian variables of models. in latent graphical study the learning problem We,We study the problem of learning latent variables in Gaussian graphical models.,26068
"proper this problem. learning We algorithms introduce fast, for","We introduce fast, proper learning algorithms for this problem.",26069
"are approaches, our In manifestly contrast algorithms with non-convex. existing","In contrast with existing approaches, our algorithms are manifestly non-convex.",26070
complement We experiments. with several theory numerical our,We complement our theory with several numerical experiments.,26071
summaries parameterized personalized matches. approach produce of variable soccer length present We to a,We present a parameterized approach to produce personalized variable length summaries of soccer matches.,26072
"discovery, big annotation factors Two genes IS have effects main functionality namely: prediction. and on","Two main factors have big effects on IS discovery, namely: genes annotation and functionality prediction.",26073
us to study results. Our several led,Our study led us to several results.,26074
non-rigid for structure-from-motion solving (NRSfM). smoothness based spatial-temporal proposes a dense paper simple This method,This paper proposes a simple spatial-temporal smoothness based method for solving dense non-rigid structure-from-motion (NRSfM).,26075
"recognize videos. to task actions the pre-segmented is happening within Given videos,","Given pre-segmented videos, the task is to recognize actions happening within videos.",26076
"task features recognition. hand Historically, to used video action crafted of were address the","Historically, hand crafted video features were used to address the task of action recognition.",26077
"of for the a approach image The ResNet, extends model state art classification.","The approach extends ResNet, a state of the art model for image classification.",26078
on for face have Face relied datasets detection methods training.,Face detection methods have relied on face datasets for training.,26079
LSLF Our and multi-view of large unconstrained dataset number a faces. of occluded consists partially,Our LSLF dataset consists of a large number of unconstrained multi-view and partially occluded faces.,26080
"Second, CrowedFaces our we CrowedNonFaces image introduce datasets. and","Second, we introduce our CrowedFaces and CrowedNonFaces image datasets.",26081
from CrowedNonFaces The datasets include scenes. crowed faces crowedFaces and and images non-faces,The crowedFaces and CrowedNonFaces datasets include faces and non-faces images from crowed scenes.,26082
"(PAN) Networks principled transformation image-to-image we Perceptual tasks. In propose for Adversarial this a paper,","In this paper, we propose a principled Perceptual Adversarial Networks (PAN) for image-to-image transformation tasks.",26083
"etc.) transformation image image-to-image several Experiments on image evaluated tasks inpainting, (e.g., de-raining,","Experiments evaluated on several image-to-image transformation tasks (e.g., image de-raining, image inpainting, etc.)",26084
related state-of-the-art show PAN outperforms proposed that methods. the many,show that the proposed PAN outperforms many related state-of-the-art methods.,26085
trained from A warp then first second deep-network the to is the to view. detections,A deep-network is then trained to warp detections from the first to the second view.,26086
lead pedestrian very poor detection when pedestrian can using This detectors. performance to standard,This can lead to very poor pedestrian detection performance when using standard pedestrian detectors.,26087
active years. Video has been few prediction an topic research past in of the,Video prediction has been an active topic of research in the past few years.,26088
"we prediction. project, video for this approach a use long-term hierarchical In","In this project, we use a hierarchical approach for long-term video prediction.",26089
al. adopted et the Villegas is largely from work by method Our,Our method is largely adopted from the work by Villegas et al.,26090
analogy-based with is combination auto-encoder networks. built of The convolutional and LSTMs a method,The method is built with a combination of LSTMs and analogy-based convolutional auto-encoder networks.,26091
"realistic we loss. predictions, to more also adversarial adopt Additionally, frame order in generate","Additionally, in order to generate more realistic frame predictions, we also adopt adversarial loss.",26092
"the we Go! Way In to this paper, introduce","In this paper, we introduce the Way to Go!",26093
for generating a automatically wayfinding given for a design layout. approach,approach for automatically generating a wayfinding design for a given layout.,26094
"deep to In we this learning synergic model a issue. this paper, address propose (SDL)","In this paper, we propose a synergic deep learning (SDL) model to address this issue.",26095
'end can be SDL Our to model end'. trained,Our SDL model can be trained 'end to end'.,26096
data isometry restricted universal a recovery. (RIP) is tool The for property,The restricted isometry property (RIP) is a universal tool for data recovery.,26097
"inclined detection an get results. use to we last, non-maximum suppression At the","At last, we use an inclined non-maximum suppression to get the detection results.",26098
the in in field. emerging an obtained scheduling from has UAV Research optimization interest scientists,Research in UAV scheduling has obtained an emerging interest from scientists in the optimization field.,26099
a components learning. semi-supervised generative network for based are models with powerful approach Neural discriminative,Neural network based generative models with discriminative components are a powerful approach for semi-supervised learning.,26100
active opens for semi-supervised door the This learning. Bayesian,This opens the door for semi-supervised Bayesian active learning.,26101
novel articles. and propose for scoring ranking We uncertainty-aware of a Bayesian news model,We propose a novel Bayesian model of uncertainty-aware scoring and ranking for news articles.,26102
experiments of public the is traffic framework several datasets. demonstrated through suggested effectiveness on The,The effectiveness of the suggested framework is demonstrated through experiments on several public traffic datasets.,26103
"be evolution schemes. solved cannot integration analytically, systems practice, In numerical of requiring Hamiltonian the","In practice, the evolution of Hamiltonian systems cannot be solved analytically, requiring numerical integration schemes.",26104
of algorithm an extension al. essentially Spirtes It of CI consists et the in,It consists essentially in an extension of the CI algorithm of Spirtes et al.,26105
various computer vision Deep been area. has of extensively aspects used learning,Deep learning has been extensively used various aspects of computer vision area.,26106
detection. of recognition Instead vision network neural for using and,Instead of using neural network for vision recognition and detection.,26107
This paper an problem addresses unsupervised in localization the of image. object,This paper addresses the problem of unsupervised object localization in an image.,26108
widely recognized. been in has mechanism is epigenetic important whose role DNA development an methylation,DNA methylation is an epigenetic mechanism whose important role in development has been widely recognized.,26109
"However, in voice correlated natural highly activity with facial human communications. is","However, facial activity is highly correlated with voice in natural human communications.",26110
adaptation iterative complex change factorization of method. enables This matrix slight the the in design,This slight change in design matrix enables the adaptation of the iterative complex factorization method.,26111
The innovation in of datasets. computational genome-wide wide-spread will association use the facilitate reanalyzing BVSR,The computational innovation will facilitate the wide-spread use of BVSR in reanalyzing genome-wide association datasets.,26112
essential everyone's life. Indoor familiar environments in the are scenes and most,Indoor scenes are the most familiar and essential environments in everyone's life.,26113
"in research efforts made e.g. communities, been have Many different","Many efforts have been made in different research communities, e.g.",26114
"analyze patient and record, quantify, protocol there Unfortunately, medical to is motion. formal no","Unfortunately, there is no formal medical protocol to record, quantify, and analyze patient motion.",26115
"estimation. we In posture ""distracted this dataset new driver"" for present paper, a","In this paper, we present a new dataset for ""distracted driver"" posture estimation.",26116
genetically-weighted of The system ensemble Neural of Networks consists Convolutional a (CNNs).,The system consists of a genetically-weighted ensemble of Convolutional Neural Networks (CNNs).,26117
effect of study the elements (i.e. visual different also We,We also study the effect of different visual elements (i.e.,26118
means detection classification and and face) and hands hand localizations. distraction of by in face,hands and face) in distraction detection and classification by means of face and hand localizations.,26119
"material, The out cellular is molecular an cytoskeleton active by proteins. motor equilibrium driven of","The cellular cytoskeleton is an active material, driven out of equilibrium by molecular motor proteins.",26120
data. of and Our neontological paleontological analysis flexible joint equations allow Bayesian for,Our equations allow for flexible joint Bayesian analysis of paleontological and neontological data.,26121
"and deformation, rotation, demonstrates clutter. out-of-plane in-plane robustness scale, to and It distractors background","It demonstrates robustness to deformation, in-plane and out-of-plane rotation, scale, distractors and background clutter.",26122
reinitialize FaceTrack helps The detector to drift. also during,The detector also helps to reinitialize FaceTrack during drift.,26123
"segmentation, the of object We i.e. semi-supervised video tackle task","We tackle the task of semi-supervised video object segmentation, i.e.",26124
"Additionally, on objectness, step we is which add PASCAL. a pretraining based learned on","Additionally, we add a pretraining step based on objectness, which is learned on PASCAL.",26125
handling (CNN) diverse neural are based detectors of network inefficient Convolutional face scales. in faces,Convolutional neural network (CNN) based face detectors are inefficient in handling faces of diverse scales.,26126
"efficient Prior detection, the faces. distribution to the histogram of scale CNN predicts an","Prior to detection, an efficient CNN predicts the scale distribution histogram of the faces.",26127
Then scale and guides of image. zoom-out histogram the the the zoom-in,Then the scale histogram guides the zoom-in and zoom-out of the image.,26128
"MALF of show experiments advantages and on AFW Extensive SAFD. FDDB,","Extensive experiments on FDDB, MALF and AFW show advantages of SAFD.",26129
explore science on informatics intensively different urban data. address data Urban based issues to methods,Urban informatics explore data science methods to address different urban issues intensively based on data.,26130
part used comparative Quality of performance are and assessment as control the pipeline.,Quality control and comparative performance assessment are used as part of the pipeline.,26131
based cars cameras. detecting evaluate monitoring for on We method the,We evaluate the method for detecting cars based on monitoring cameras.,26132
relies bases processing knowledge from relations available Learning based relation the evidence on instances. on,Learning relations based on evidence from knowledge bases relies on processing the available relation instances.,26133
"the impact discover effectiveness type the scenarios Furthermore, regularizer. of that we significantly","Furthermore, we discover scenarios that significantly impact the effectiveness of the type regularizer.",26134
transfer Sonar CNNs the incorporating one We into ATR: learning. to power demonstrate avenue of,We demonstrate the power of one avenue to incorporating CNNs into Sonar ATR: transfer learning.,26135
unique are and a present challenge they These observational. data because high-dimensional,These data present a unique challenge because they are high-dimensional and observational.,26136
estimating We three treatment data. methods propose and heterogeneous for analyze effects observational using,We propose and analyze three methods for estimating heterogeneous treatment effects using observational data.,26137
This the brain paper issues multi-class of these in context investigates tumour segmentation.,This paper investigates these issues in the context of multi-class brain tumour segmentation.,26138
(FGVC) has significant Visual achieved Fine-Grained progress recently. Categorization,Fine-Grained Visual Categorization (FGVC) has achieved significant progress recently.,26139
complete large-scale fine-grained open This training issue a without perform identification an raises to set.,This raises an open issue to perform large-scale fine-grained identification without a complete training set.,26140
"images retrieves identification. framework similar semantics, with progressively fine-grained performs The and","The framework progressively retrieves images with similar semantics, and performs fine-grained identification.",26141
"this approaches. introduce model In both we that a unifies paper,","In this paper, we introduce a model that unifies both approaches.",26142
"develop data. for the Therefore, a synthesizing training framework we","Therefore, we develop a framework for synthesizing the training data.",26143
Adversarial the and employed Generative are image Networks to distributions. minimize real synthetic between gap,Generative Adversarial Networks are employed to minimize the gap between synthetic and real image distributions.,26144
a shared invariant learned to poses transfer space. CNN are high-level that then human models,CNN models are then learned that transfer human poses to a shared high-level invariant space.,26145
"classification. linear Finally, used for is SVM","Finally, linear SVM is used for classification.",26146
for exploiting aims information images. super-resolution Video from additionally the multiple,Video super-resolution aims for exploiting additionally the information from multiple images.,26147
"image the consecutive related Typically, warping. are via and flow images optical","Typically, the images are related via optical flow and consecutive image warping.",26148
"of contribution uses. practical their This face algorithms, recogni-tion implementation overview and an gives","This contribution gives an overview of face recogni-tion algorithms, their implementation and practical uses.",26149
people resulting utilized to can be face classify model in specific The individuals unknowns. or,The resulting face model can be utilized to classify people in specific individuals or unknowns.,26150
multi-channel beamformers. by adaptive for enhancement speech,for multi-channel speech enhancement by adaptive beamformers.,26151
MRI is to MRI identify from task parameters The fingerprinting tissue complex-valued of signals.,The task of MRI fingerprinting is to identify tissue parameters from complex-valued MRI signals.,26152
Extensive our performance compared state-of-the-art competitive with experiments that achieves demonstrate model models.,Extensive experiments demonstrate that our model achieves competitive performance compared with state-of-the-art models.,26153
as Person problem. re-identification image (person an mostly re-ID) viewed is retrieval,Person re-identification (person re-ID) is mostly viewed as an image retrieval problem.,26154
image a task a pool. query large to search person aims This in,This task aims to search a query person in a large image pool.,26155
"automatic to practice, usually In adopts detectors person obtain re-ID cropped pedestrian images.","In practice, person re-ID usually adopts automatic detectors to obtain cropped pedestrian images.",26156
"errors: part from excessive this two detector types of and background process However, suffers missing.","However, this process suffers from two types of detector errors: excessive background and part missing.",26157
with are that better pedestrians Visual aligned examples show PAN.,Visual examples show that pedestrians are better aligned with PAN.,26158
longstanding problem computational Image in photography. matting is a,Image matting is a longstanding problem in computational photography.,26159
generated matte matting automatically the use from to the learning We method based generate trimap.,We use learning based matting method to generate the matte from the automatically generated trimap.,26160
"Accurate surgical many medical essential for applications. segmentation other image diagnosis, is and planning","Accurate medical image segmentation is essential for diagnosis, surgical planning and many other applications.",26161
(CNNs) segmentation Neural Networks become the have state-of-the-art Convolutional methods. automatic,Convolutional Neural Networks (CNNs) have become the state-of-the-art automatic segmentation methods.,26162
"Random In integrate we a user into constraints Conditional Field. addition, hard interactions as back-propagatable","In addition, we integrate user interactions as hard constraints into a back-propagatable Conditional Random Field.",26163
is computational representation color introduced. A biologically for model plausible,A biologically plausible computational model for color representation is introduced.,26164
origin of phenotypic quantitative and We the the cell physical explanations for fates. provide two,We provide physical and quantitative explanations for the origin of the two phenotypic cell fates.,26165
multiple of the fates. reveal results mechanism new for emergence cell a Our,Our results reveal a new mechanism for the emergence of multiple cell fates.,26166
Micro-lens field light light cameras based to cost-effective field. array capture approach a offer (MLA),Micro-lens array (MLA) based light field cameras offer a cost-effective approach to capture light field.,26167
"paper, we enhancement approach. In a learning field light this based present","In this paper, we present a learning based light field enhancement approach.",26168
networks. enhanced field is and spatial of neural captured resolution light convolutional Both using angular,Both spatial and angular resolution of captured light field is enhanced using convolutional neural networks.,26169
"learning, zero-shot i.e. to the of importance Due","Due to the importance of zero-shot learning, i.e.",26170
three-fold. purpose The of this paper is,The purpose of this paper is three-fold.,26171
for a present deterministic with mathematical malaria model waning We immunity. transmission,We present a deterministic mathematical model for malaria transmission with waning immunity.,26172
system The five equations. model of non-linear consists differential of,The model consists of five non-linear system of differential equations.,26173
been established. Conditions point for endemic existence have of equilibrium,Conditions for existence of endemic equilibrium point have been established.,26174
bifurcation bifurcation the occur backward that reveals theory model. in based qualitative A on study,A qualitative study based on bifurcation theory reveals that backward bifurcation occur in the model.,26175
Recent fine-tuned to adopts descriptors. power the further work of strategies improve the discriminative,Recent work adopts fine-tuned strategies to further improve the discriminative power of the descriptors.,26176
burstiness and the can accuracy. effectively issue We demonstrate that retrieval this address improve,We demonstrate that this can effectively address the burstiness issue and improve retrieval accuracy.,26177
"Secondly, feature enhance embedding propose further to discriminability. recent aggregating and to methods we employ","Secondly, we propose to employ recent embedding and aggregating methods to further enhance feature discriminability.",26178
Extensive state-of-the-art experiments retrieval achieves framework that demonstrate accuracy. proposed our,Extensive experiments demonstrate that our proposed framework achieves state-of-the-art retrieval accuracy.,26179
curvature existing estimates demonstrate techniques. that provides than and We method better our is fast,We demonstrate that our method is fast and provides better curvature estimates than existing techniques.,26180
improvement. several particular demonstrate compare to our we the alternatives method In to,In particular we compare our method to several alternatives to demonstrate the improvement.,26181
Artificial methods solve recent control progress tasks to Intelligence continuous- years. in have made significant,Artificial Intelligence methods to solve continuous- control tasks have made significant progress in recent years.,26182
an phase. area means that active this in research This still is,This means that this area is still in an active research phase.,26183
"however, outcomes. these prone Manual and expensive is of time-series, biased to and non-reproducible evaluation","Manual evaluation of these time-series, however, is expensive and prone to biased and non-reproducible outcomes.",26184
non-conflicting paths find multiple path the In (MAPF) for finding multi-agent task is agents. to,In multi-agent path finding (MAPF) the task is to find non-conflicting paths for multiple agents.,26185
for paper variant. finding focus In the suboptimal for this we sum-of-costs on solutions MAPF,In this paper we focus on finding suboptimal solutions for MAPF for the sum-of-costs variant.,26186
in that search-based solver the show the results SAT-based solvers. case Experimental significantly many outperforms,Experimental results show that in many case the SAT-based solver significantly outperforms the search-based solvers.,26187
"susceptibles, process approximation of a depletion the is For without birth-death exact.","For a birth-death process without depletion of susceptibles, the approximation is exact.",26188
method novel benefits This two the of enjoys techniques.,This method enjoys the benefits of two novel techniques.,26189
PIO's idea features is phenomena ST basic concepts. in observed to into some formulate mechanics,PIO's basic idea is to formulate some phenomena observed in ST features into mechanics concepts.,26190
of previous the bounding-box-based box methods assume object interest. covers the segmentation tightly Most bounding,Most previous bounding-box-based segmentation methods assume the bounding box tightly covers the object of interest.,26191
too a too be common could input small. large rectangle However or it is that,However it is common that a rectangle input could be too large or too small.,26192
curve-based our input We without extends network show retraining. to,We show our network extends to curve-based input without retraining.,26193
Experiments demonstrate the benchmark on of approaches. effectiveness proposed the datasets,Experiments on benchmark datasets demonstrate the effectiveness of the proposed approaches.,26194
for formalism trackers. a novel comparison paper The Lisp-like a better of proposes,The paper proposes a novel Lisp-like formalism for a better comparison of trackers.,26195
of functional design functional decomposition and a This trackers. certain assumes,This assumes a certain functional design and functional decomposition of trackers.,26196
work future suggests concludes with promising The of research lines and work.,The work concludes with promising lines of research and suggests future work.,26197
We complementarity explore fusion among to the strategies also the models. adopt several,We also adopt several fusion strategies to explore the complementarity among the models.,26198
"training we a of images Specifically, samples. pairwise into batch the organize","Specifically, we organize the training images into a batch of pairwise samples.",26199
could induce influence and Background: mental submersible manned workload operation measures. neurophysiological Deep-sea,Background: Deep-sea manned submersible operation could induce mental workload and influence neurophysiological measures.,26200
not operation well known. submersible responses Psychophysiological are to,Psychophysiological responses to submersible operation are not well known.,26201
were these ratings mental at also stress of points. time Subjective conducted,Subjective ratings of mental stress were also conducted at these time points.,26202
arousal later period. Brain during declined the operation level,Brain arousal level declined during the later operation period.,26203
and should knowledge The visualise system processing. users help structures,The system should help users visualise knowledge structures and processing.,26204
pattern in Strengths SP recognition may be developed vision. computer the of System for,Strengths of the SP System in pattern recognition may be developed for computer vision.,26205
arithmetic of of needed representation numbers and is processes. Work on the the performance,Work is needed on the representation of numbers and the performance of arithmetic processes.,26206
based We of the effectiveness experimental solutions. present the new to demonstrate learning deep results,We present experimental results to demonstrate the effectiveness of the new deep learning based solutions.,26207
important images. an is manipulated signature Resampling of,Resampling is an important signature of manipulated images.,26208
Tampered segmentation are located regions method. Walker Random a using,Tampered regions are located using a Random Walker segmentation method.,26209
compare the both detection/localization of methods. performance We these of,We compare the performance of detection/localization of both these methods.,26210
nature. This is data in and sensor multi-modal heterogeneous,This sensor data is multi-modal and heterogeneous in nature.,26211
parameterized molecular model binding. of present We analytical alchemical a,We present a parameterized analytical model of alchemical molecular binding.,26212
complexes. assessed host-guest on has set The the a model been of of validity,The validity of the model has been assessed on a set of host-guest complexes.,26213
profiles. a for shape the The rationalization of characteristic free offers the energy model,The model offers a rationalization for the characteristic shape of the free energy profiles.,26214
phalanx for features that subset is A a well work of classification tasks.,A phalanx is a subset of features that work well for classification tasks.,26215
"we application for propose of in regression In class different phalanxes a paper, settings. this","In this paper, we propose a different class of phalanxes for application in regression settings.",26216
is for high-order problem. reduction tensors a challenging Dimensionality,Dimensionality reduction for high-order tensors is a challenging problem.,26217
"respectively. the inherent resulting or This structures high-order in will undesired tensors, destroy","This will destroy the inherent high-order structures or resulting in undesired tensors, respectively.",26218
for reduction paper a tensorial This probabilistic introduces dimensionality vectorial model data.,This paper introduces a probabilistic vectorial dimensionality reduction model for tensorial data.,26219
EM established via variational been inference Bayesian approach. the A has,A Bayesian inference has been established via the variational EM approach.,26220
an by can on-screen blinking. looking and in Users at keyboard type words,Users can type in words by looking at an on-screen keyboard and blinking.,26221
a improve can This effectively and gaze making selection the blinks. accuracy method by of,This method can effectively improve the accuracy of making a selection by gaze and blinks.,26222
"needs, removed disable we complex process. considering In the people's calibration","In considering disable people's needs, we removed the complex calibration process.",26223
run and in methods input off-screen mode The can mode. portable screen,The input methods can run in screen mode and portable off-screen mode.,26224
images of information. Colorization on gray-scale prior color relies,Colorization of gray-scale images relies on prior color information.,26225
use Examplar-based color of image such information. a as methods source,Examplar-based methods use a color image as source of such information.,26226
the gray-scale image Then image. are of source the transferred the colors to,Then the colors of the source image are transferred to the gray-scale image.,26227
"transfer literature, descriptors. by the this texture In mainly is guided","In the literature, this transfer is mainly guided by texture descriptors.",26228
common texture frequently few images the so approaches usually that contain Face fail.,Face images usually contain few texture so that the common approaches frequently fail.,26229
we new a this based propose image on method paper In morphing.,In this paper we propose a new method based on image morphing.,26230
shapes. images with able This technique a compute to is correspondence between similar map,This technique is able to compute a correspondence map between images with similar shapes.,26231
show clearly that morphing experiments based numerical outperforms state-of-the-art Our methods. our approach,Our numerical experiments show that our morphing based approach clearly outperforms state-of-the-art methods.,26232
images unseen representations Re-Identification for is discriminative for Learning (ReID). person person critical,Learning discriminative representations for unseen person images is critical for person Re-Identification (ReID).,26233
"random a number particles with varies As is the result, and time. of","As a result, the number of particles is random and varies with time.",26234
different advantage An scheme particles of evolve is independently. this can of that descendants,An advantage of this scheme is that descendants of different particles can evolve independently.,26235
simulation existing with which methods. indicate that present algorithms compete results our can the We,We present simulation results which indicate that our algorithms can compete with the existing methods.,26236
time burden. the execution wastes a imposes and computational large This,This wastes the execution time and imposes a large computational burden.,26237
"inaccuracy. lead sources to comprehensively investigate In work, that we the filtering this four","In this work, we comprehensively investigate four sources that lead to the filtering inaccuracy.",26238
pre-alignment the significantly two of by orders improves It one filtering to magnitude. of accuracy,It significantly improves the accuracy of pre-alignment filtering by one to two orders of magnitude.,26239
predictive cardiologist both value). and the performance We in (sensitivity) exceed (positive precision average recall,We exceed the average cardiologist performance in both recall (sensitivity) and precision (positive predictive value).,26240
proposes This paper convolutional for spotting. segmentation-free region network based a word neural,This paper proposes a region based convolutional neural network for segmentation-free word spotting.,26241
approach novel B-splines interactive present a volumetric enabling We Refined of Locally (LR-splines). visualization,We present a novel approach enabling interactive visualization of volumetric Locally Refined B-splines (LR-splines).,26242
approaches comparison performed. adaptive is existing with A,A comparison with existing adaptive approaches is performed.,26243
algorithms processing designed (GPU) capabilities. fully are to The graphics modern utilize unit,The algorithms are designed to fully utilize modern graphics processing unit (GPU) capabilities.,26244
"w.r.t. Theoretically, we process decision a Markov dual (MDP) find","Theoretically, we find a dual Markov decision process (MDP) w.r.t.",26245
"policy derive the solution a parametrized and environment. optimizing the to gradient agent,","the agent, and derive a policy gradient solution to optimizing the parametrized environment.",26246
"environments proposed discontinuous addressed are Furthermore, by general a framework. generative","Furthermore, discontinuous environments are addressed by a proposed general generative framework.",26247
in The under TPLP. consideration is acceptance paper for,The paper is under consideration for acceptance in TPLP.,26248
function useful the features. of behind Modes data probability geometric are and density observed ridges,Modes and ridges of the probability density function behind observed data are useful geometric features.,26249
"estimator propose the \emph{density-derivative-ratios}. these for we problems, To a with cope novel","To cope with these problems, we propose a novel estimator for the \emph{density-derivative-ratios}.",26250
"estimator. convergence the Moreover, a we proposed rate of establish","Moreover, we establish a convergence rate of the proposed estimator.",26251
applications. to problem is manipulation Causation considered variety a crucial without of discovery a,Causation discovery without manipulation is considered a crucial problem to a variety of applications.,26252
models. experiments We to conduct and contribution our the of validate analyze,We conduct experiments to validate and analyze the contribution of our models.,26253
"frames. is a loss consecutive to between triplet other hand, pursue On the appearance-smoothness employed","On the other hand, a triplet loss is employed to pursue appearance-smoothness between consecutive frames.",26254
own used papers have private their dataset. All,All papers have used their own private dataset.,26255
"of highly dataset. systems the CAD on the efficiency Therefore, depends","Therefore, the efficiency of CAD systems highly depends on the dataset.",26256
rigid from the Our moving objects technique out segments also scene.,Our technique also segments out moving objects from the rigid scene.,26257
is the for The existence important of mapping learning.,The existence of the mapping is important for learning.,26258
"to networks trained. spiking Conclusively, noise impervious correlations when sampling are","Conclusively, sampling spiking networks are impervious to noise correlations when trained.",26259
on the each equivalence side simple are reported. of from Calculations models,Calculations on simple models from each side of the equivalence are reported.,26260
"layers of is learning"". ""deepest continuous denoted use The","The use of continuous layers is denoted ""deepest learning"".",26261
continuous symmetry problems symplectic phase space. a in respect These time/layer,These problems respect a symplectic symmetry in continuous time/layer phase space.,26262
presented. versions versions Hamiltonian and are problems these of Both Lagrangian,Both Lagrangian versions and Hamiltonian versions of these problems are presented.,26263
"symplectic structure, the is well-studied discrete respected in Their time/layer, while addressed. implementation a","Their well-studied implementation in a discrete time/layer, while respected the symplectic structure, is addressed.",26264
"reconstruct. to illuminate scene (LEDs) to use we Rather, light-emitting the diodes","Rather, we use light-emitting diodes (LEDs) to illuminate the scene to reconstruct.",26265
"state-of-the-art numerical we discuss solutions. two Then,","Then, we discuss two state-of-the-art numerical solutions.",26266
"images. the to of study Finally, extend we case RGB this","Finally, we extend this study to the case of RGB images.",26267
on games. two Strong the performance is other also achieved,Strong performance is also achieved on the other two games.,26268
"agents replays, show learn strategies. game our interesting In we","In game replays, we show our agents learn interesting strategies.",26269
use conditional for the to One is methods GAN alignment. of,One of the methods is to use conditional GAN for alignment.,26270
"GAN as as conditional of previous other methods. adopting well attempts perform do not However,","However, previous attempts of adopting conditional GAN do not perform as well as other methods.",26271
"dance the Using IMFs. into be primitives can their motion decomposed EMD, motions (choreographic) or","Using the EMD, their dance motions can be decomposed into motion (choreographic) primitives or IMFs.",26272
about tool to is a transfer an (e.g. knowledge task adaptation important Domain,Domain adaptation is an important tool to transfer knowledge about a task (e.g.,26273
"learned classification) a source domain. to a domain second, target or in","classification) learned in a source domain to a second, or target domain.",26274
during data assume available task-relevant is approaches that training. Current target-domain,Current approaches assume that task-relevant target-domain data is available during training.,26275
such how domain task-relevant data no available. perform when We is adaptation target-domain to demonstrate,We demonstrate how to perform domain adaptation when no such task-relevant target-domain data is available.,26276
"source-domain Therefore, interest (e.g. task of solution the","Therefore, the source-domain task of interest solution (e.g.",26277
new management procedure a assessing with decision in making. paper fuzzy proposes application This,This paper proposes a new fuzzy assessing procedure with application in management decision making.,26278
be a using of This determined with set nonlinear can the functions. parametric membership programing,This can be determined with a set of nonlinear parametric programing using the membership functions.,26279
"derive To is careful needed. post-processing ground truth, the","To derive the ground truth, careful post-processing is needed.",26280
style and is traditional flexible It transfer with constraints. compatible the,It is flexible and compatible with the traditional style transfer constraints.,26281
"this Trust-PCL. problem, To we address propose region method, an off-policy trust","To address this problem, we propose an off-policy trust region method, Trust-PCL.",26282
"interactions ""densify"" noise. Next, and response, robustness enhancing the lateral decorrelate to","Next, lateral interactions ""densify"" and decorrelate the response, enhancing robustness to noise.",26283
using data from empirically test our Drosophila. theory We,We test our theory empirically using data from Drosophila.,26284
propositional such relations associating with models. by minimal Circumscription formalizes preference theories,Circumscription formalizes such preference relations by associating propositional theories with minimal models.,26285
acceptance consideration paper is under This for in TPLP.,This paper is under consideration for acceptance in TPLP.,26286
"an effective present and object for RON, generic detection. We framework efficient","We present RON, an efficient and effective framework for generic object detection.",26287
convergence We and for schemes. study resampling rates convergence,We study convergence and convergence rates for resampling schemes.,26288
"constraints, designs processing been this To smoothness problem, post architecture have different combat proposed. and","To combat this problem, smoothness constraints, post processing and different architecture designs have been proposed.",26289
"three of sources deconvolution loss random initialization overlap, checkerboard functions. and highlight artifacts:","highlight three sources of checkerboard artifacts: deconvolution overlap, random initialization and loss functions.",26290
introduces This paper package of implementation object-oriented funData the data. R functional as an,This paper introduces the funData R package as an object-oriented implementation of functional data.,26291
packages CRAN. on GitHub publicly are available Both and,Both packages are publicly available on GitHub and CRAN.,26292
"the to determine K data. the from difficulty found However, we","However, we found the difficulty to determine K from the data.",26293
"caused virtual and is data we this shortage. argue that measurement Using analysis, by","Using virtual measurement and analysis, we argue that this is caused by data shortage.",26294
"score the of transformation. show two orientation Finally, applications we","Finally, we show two applications of the orientation score transformation.",26295
article available materials Supplementary this are online. for,Supplementary materials for this article are available online.,26296
efforts. target Those RSV-related groups may for children be as considered mitigation potential,Those children may be considered as potential target groups for RSV-related mitigation efforts.,26297
problem approach the human This of estimation. to based learning paper a deep presents pose,This paper presents a deep learning based approach to the problem of human pose estimation.,26298
training a estimator as done. after generator pose is human The is the used,The generator is used as a human pose estimator after the training is done.,26299
as deep therefore when promising know they are architectures GPDNNs that don't know.,GPDNNs are therefore promising as deep architectures that know when they don't know.,26300
"an be image to could Based soft Gaussian maps. SGM, converted on several","Based on SGM, an image could be converted to several soft Gaussian maps.",26301
"In analysis particular, eigenbases to classical manifolds. the Laplacian generalizing allow Fourier","In particular, Laplacian eigenbases allow generalizing the classical Fourier analysis to manifolds.",26302
"local shape analysis. framework paper, this for a introduce new spectral we In","In this paper, we introduce a new framework for local spectral shape analysis.",26303
patients vasodilation demonstrated local have impaired The in heating. tests diabetic temperature response that to,The temperature tests demonstrated that diabetic patients have impaired vasodilation in response to local heating.,26304
memberships clusters the allows the to simultaneously. to number estimate and of This clusters,This allows to estimate the memberships to clusters and the number of clusters simultaneously.,26305
estimator the intensities. estimate Then maximum-likelihood developed non is parametrically a to integrated,Then a maximum-likelihood estimator is developed to estimate non parametrically the integrated intensities.,26306
Hashing image for applications methods used classification. have been large-scale retrieval and widely of,Hashing methods have been widely used for applications of large-scale image retrieval and classification.,26307
address decision in nature larger. systems Current support that are and progressively domains becoming heterogeneous,Current decision support systems address domains that are heterogeneous in nature and becoming progressively larger.,26308
paper people attributes (e.g. detecting for deep architecture propose this In we a,In this paper we propose a deep architecture for detecting people attributes (e.g.,26309
"...) clothing in gender, contexts. race, surveillance","gender, race, clothing ...) in surveillance contexts.",26310
monocular vanishing point a for uncalibrated present novel We approach images. from detection,We present a novel approach for vanishing point detection from uncalibrated monocular images.,26311
"scene. In to no observed the we contrast assumptions a state-of-the-art, priori make about","In contrast to state-of-the-art, we make no a priori assumptions about the observed scene.",26312
benchmark competitive performance method three on datasets. Our estimation horizon achieves,Our method achieves competitive performance on three horizon estimation benchmark datasets.,26313
if their a is of different the approaches dataset limits This used. benefit,This limits the benefit of their approaches if a different dataset is used.,26314
framework highly way spaces The conceptual representing of provides influential a geometric of knowledge.,The highly influential framework of conceptual spaces provides a geometric way of representing knowledge.,26315
mutually generalize categories Affordance object not exclusive. are and types across,Affordance types generalize across object categories and are not mutually exclusive.,26316
makes segmentation task. of This the affordance a objects of in difficult images regions,This makes the segmentation of affordance regions of objects in images a difficult task.,26317
"this of process, to network During predictions the need be binarized. the","During this process, the predictions of the network need to be binarized.",26318
crucial data fields for role as such plays Time-Spatial traffic different management. a,Time-Spatial data plays a crucial role for different fields such as traffic management.,26319
data sensors These as can such collected surveillance via devices tracking or be systems.,These data can be collected via devices such as surveillance sensors or tracking systems.,26320
sive. methods mas- when fail easily Those become data would,Those methods would easily fail when data become mas- sive.,26321
perform time-oriented and data on attempts Recent concern cluster how prediction to with simply information.,Recent attempts concern on how to simply cluster data and perform prediction with time-oriented information.,26322
to and visiting SOM the able is pattern. patterns cluster the detect abnormal,SOM is able to cluster the visiting patterns and detect the abnormal pattern.,26323
for its RNN predic- time architecture. the using perform can tion analysis dynamic series,RNN can perform the predic- tion for time series analysis using its dynamic architecture.,26324
the progressed Recent neural development with research networks deep of convolutional (DCNN). super-resolution on has,Recent research on super-resolution has progressed with the development of deep convolutional neural networks (DCNN).,26325
"techniques exhibit learning residual performance. In improved particular,","In particular, residual learning techniques exhibit improved performance.",26326
and tasks the signature biometrics one Offline most of is forensics. challenging document verification in,Offline signature verification is one of the most challenging tasks in biometrics and document forensics.,26327
in biometric technology used recognition Facial and security widely a verification system. is,Facial recognition and verification is a widely used biometric technology in security system.,26328
"biometrics vulnerable using attacks Unfortunately, videos. is spoofing photographs to or face","Unfortunately, face biometrics is vulnerable to spoofing attacks using photographs or videos.",26329
D-K inner image face features the The reflect correlation sequence. of the,The D-K features reflect the inner correlation of the face image sequence.,26330
tracking objects undergo is drastic changes appearance Object time. challenging often over as target,Object tracking is challenging as target objects often undergo drastic appearance changes over time.,26331
"been applied tracking. Recently, correlation object have successfully to adaptive filters","Recently, adaptive correlation filters have been successfully applied to object tracking.",26332
the representations. surrounding appropriate and We the feature size into of take account context,We take into account the appropriate size of surrounding context and the feature representations.,26333
tracking to use long-term occurs. filter output of the responses if determine We failure this,We use the output responses of this long-term filter to determine if tracking failure occurs.,26334
to are oscillations cognitive functions many processes. thought in and Cortical involved be,Cortical oscillations are thought to be involved in many cognitive functions and processes.,26335
regulate have oscillations. proposed to mechanisms been Several,Several mechanisms have been proposed to regulate oscillations.,26336
prominent is understudied but mechanism gap-junctional One coupling.,One prominent but understudied mechanism is gap-junctional coupling.,26337
interneurons. ubiquitous Gap are cortex in junctions between GABAergic,Gap junctions are ubiquitous in cortex between GABAergic interneurons.,26338
developed junction network. model a We plasticity computational a of gap cortical recurrent in,We developed a computational model of gap junction plasticity in a recurrent cortical network.,26339
plasticity functional cortex. This direct a role in in junction information gap transmission for implies,This implies a direct functional role for gap junction plasticity in information transmission in cortex.,26340
the that We readily and approach of incorporate show protein ligand. can flexibility,We show that the approach can readily incorporate flexibility of protein and ligand.,26341
"city, where was picture the a taken.","the city, where a picture was taken.",26342
"Moreover, features, that also e.g. it contextual reveals","Moreover, it reveals that also contextual features, e.g.",26343
"account environment, type, natural the methods. wall taken into by etc., automatic are","wall type, natural environment, etc., are taken into account by the automatic methods.",26344
evaluate on approach our et al. We Johnson,We evaluate our approach on Johnson et al.,26345
a image algorithm the attempts achieve manner. timely quality to in highest recovery reconstruction Any,Any image recovery algorithm attempts to achieve the highest quality reconstruction in a timely manner.,26346
work. We further look HB-MAP insights and the into analysis provide what makes to,We look to provide further analysis and insights into what makes the HB-MAP work.,26347
first removal is anisotropic for modified smooth piece-wise image. diffusion a of The framework noise,The first framework is a modified anisotropic diffusion for noise removal of piece-wise smooth image.,26348
second smoothing. The local method Total framework is image Variation activity-tuned a for (LAT-RTV) Relative,The second framework is a local activity-tuned Relative Total Variation (LAT-RTV) method for image smoothing.,26349
"we introduce recent of learning Next, briefly progress algorithms. effective some","Next, we briefly introduce some recent progress of effective learning algorithms.",26350
"promising on envisaged. multimedia are Finally, learning trends future","Finally, promising future trends on multimedia learning are envisaged.",26351
and language a image includes model MDNet an model.,MDNet includes an image model and a language model.,26352
to proposed image The ensembles model multi-scale is utilization feature efficiency. and enhance,The image model is proposed to enhance multi-scale feature ensembles and utilization efficiency.,26353
The optimization overall our developed by using is network strategy. end-to-end trained,The overall network is trained end-to-end by using our developed optimization strategy.,26354
performance obtains image model state-of-the-art datasets two well. on proposed The as CIFAR,The proposed image model obtains state-of-the-art performance on two CIFAR datasets as well.,26355
learning visual Deep are convolutional tools representations images. networks from for powerful neural,Deep convolutional neural networks are powerful tools for learning visual representations from images.,26356
"medical challenging. volumetric However, analyse to efficient architectures deep designing remains images","However, designing efficient deep architectures to analyse volumetric medical images remains challenging.",26357
"we of a Additionally, sampling uncertainty voxel-level estimation feasibility approximation dropout. using through show the","Additionally, we show the feasibility of voxel-level uncertainty estimation using a sampling approximation through dropout.",26358
"FC data reconstructed robust Furthermore, task-behavioral with provide more associations measurements.","Furthermore, reconstructed FC data provide more robust associations with task-behavioral measurements.",26359
"connections. task-sensitive map also functional extend most we the Finally, approach this to","Finally, we extend this approach to also map the most task-sensitive functional connections.",26360
"This for investigates of reliable, unconstrained detection and bimanual a gestures. techniques paper classification","This paper investigates techniques for a reliable, unconstrained detection and classification of bimanual gestures.",26361
is fully-automatic. the part proposed of classification approach The,The classification part of the proposed approach is fully-automatic.,26362
"and ), examine from scattering the theoretical the for coming consequences transform. results","), and examine the consequences coming from theoretical results for the scattering transform.",26363
Kong. been Jumping in Donkey introduction has since mechanic its important an,Jumping has been an important mechanic since its introduction in Donkey Kong.,26364
(IU) the introduces Unintentional agent. paper Intentional This,This paper introduces the Intentional Unintentional (IU) agent.,26365
article described part applied the (DPM) is in Method deformable algorithm. on models,Method described in the article is applied on deformable part models (DPM) algorithm.,26366
"accuracy gain can higher be degradation If achieved. allowable, is insignificant computational","If insignificant accuracy degradation is allowable, higher computational gain can be achieved.",26367
the filter. shortening using of filters The decomposed decomposition tensor consists convolutions and method,The method consists of filters tensor decomposition and convolutions shortening using the decomposed filter.,26368
as well method the simulation are proposed Mathematical provided. of as overview results,Mathematical overview of the proposed method as well as simulation results are provided.,26369
novel understanding. automatic propose the of We advertisement problem,We propose the novel problem of automatic advertisement understanding.,26370
"make ads symbolic and references ""), (e.g.","""), and symbolic references ads make (e.g.",26371
to theory The aside leave frequencies (or framework. is outside of) blamed its,The theory is blamed to leave frequencies outside (or aside of) its framework.,26372
Carlo algorithm. investigate between and we Gillespie algorithm the the Here Monte the relation,Here we investigate the relation between the Monte Carlo algorithm and the Gillespie algorithm.,26373
propose program synthesis. novel We automatic a for method,We propose a novel method for automatic program synthesis.,26374
the tree. Programming search prototype P-Tree probabilistic represents through single space program a,P-Tree Programming represents the program search space through a single probabilistic prototype tree.,26375
tree prototype form problem. given a this program instances evaluate which we we on From,From this prototype tree we form program instances which we evaluate on a given problem.,26376
propagated The values prototype from error are evaluations the through the tree.,The error values from the evaluations are propagated through the prototype tree.,26377
benchmarks to applied from The several iterative method symbolic is regression the literature.,The iterative method is applied to several symbolic regression benchmarks from the literature.,26378
extend. Programming It a Genetic large to outperforms standard,It outperforms standard Genetic Programming to a large extend.,26379
multiple consist of on neural networks trained Deep datasets object layers. recognition convolutional,Deep neural networks trained on object recognition datasets consist of multiple convolutional layers.,26380
levels different layers These appearance target with encode of abstraction.,These layers encode target appearance with different levels of abstraction.,26381
"target. are resolutions precisely the However, spatial their to too coarse localize","However, their spatial resolutions are too coarse to precisely localize the target.",26382
We infer coarse-to-fine in a layer locate the each manner. to of maximum response targets,We infer the maximum response of each layer to locate targets in a coarse-to-fine manner.,26383
al. by choice the functions considered We investigate coherent Seidenfeld of generalisation et a,We investigate a generalisation of the coherent choice functions considered by Seidenfeld et al.,26384
place This paper of the presents literature. recognition overview an,This paper presents an overview of the place recognition literature.,26385
"and of work we recognition. Eventually, future place discussions with close","Eventually, we close with discussions and future work of place recognition.",26386
excels with goals. Machine-learning well-defined in areas many,Machine-learning excels in many areas with well-defined goals.,26387
"in photography. goal a such forms, as available not art clear However, usually is","However, a clear goal is usually not available in art forms, such as photography.",26388
"its very photograph concept. aesthetic value, a by success is a measured The of subjective","The success of a photograph is measured by its aesthetic value, a very subjective concept.",26389
adds challenge approach. to for a learning the This machine,This adds to the challenge for a machine learning approach.,26390
"creation. introduce for deep-learning a artistic Creatism, system content We","We introduce Creatism, a deep-learning system for artistic content creation.",26391
aspect corresponds an image to operation optimized efficiently. be can Each that,Each aspect corresponds to an image operation that can be optimized efficiently.,26392
Experiments robot's a show that our can creation be with portion of confused professional work.,Experiments show that a portion of our robot's creation can be confused with professional work.,26393
"procedural useful human and content agents, players, general-game-playing generation. Game data-driven are for maps","Game maps are useful for human players, general-game-playing agents, and data-driven procedural content generation.",26394
levels. maps game manually-created screenshots hand-assembling generally made of by are These,These maps are generally made by hand-assembling manually-created screenshots of game levels.,26395
to method image-based is of method. Performance compared manual,Performance of image-based method is compared to manual method.,26396
important. choose observer An as any ecosystem can external property being,An external observer can choose any ecosystem property as being important.,26397
approach of state remains ecosystem under the challenging. this an Summarizing,Summarizing the state of an ecosystem under this approach remains challenging.,26398
extent structure. selection which ecosystem This natural act to may the on limit can,This may limit the extent to which natural selection can act on ecosystem structure.,26399
"as technique refer to We this ""super-spectral-resolution"".","We refer to this technique as ""super-spectral-resolution"".",26400
We and for reasons investigated requirement. RNA structural this protein,We investigated protein and RNA structural reasons for this requirement.,26401
Motacilla) plumage Eurasia. with genus of a Wagtails (Aves: striking are across pattern variation birds,Wagtails (Aves: Motacilla) are a genus of birds with striking plumage pattern variation across Eurasia.,26402
introduces The interface the model hydropathy through indices.,The model introduces the interface through hydropathy indices.,26403
all-atom simulations. the We by justify model,We justify the model by all-atom simulations.,26404
of the is to absent of sensitive geometry the proteins The specificity the bonds.,The geometry of the proteins is sensitive to the specificity of the absent bonds.,26405
"and A tracking moving object object generally system of consists registration, tracking modules. detection,","A moving object tracking system generally consists of registration, object detection, and tracking modules.",26406
been so monomeric their versions in far. have theoretically studied proteins These,These proteins have been studied theoretically in their monomeric versions so far.,26407
"templates, CHARDA extended or easily fitting different with model of methods, classes predicates. is","CHARDA is easily extended with different classes of model templates, fitting methods, or predicates.",26408
"been has in Recently, for interest recognition system. monitoring individual activity growing there a brain","Recently, there has been a growing interest in monitoring brain activity for individual recognition system.",26409
a amount real-time large usually (DNNs) operations Deep for inference. of networks neural demand,Deep neural networks (DNNs) usually demand a large amount of operations for real-time inference.,26410
Weight normalization and techniques are gradual pruning to the distribution degradation. applied performance mitigate,Weight distribution normalization and gradual pruning techniques are applied to mitigate the performance degradation.,26411
with deep experiments are neural networks. convolutional neural and The conducted networks fully-connected,The experiments are conducted with fully-connected deep neural networks and convolutional neural networks.,26412
images. color impulse paper noise with removal This deals from,This paper deals with impulse noise removal from color images.,26413
to representation. unsupervised We a propose approach an style-coherent learn,We propose an unsupervised approach to learn a style-coherent representation.,26414
galleries style of can requiring outfits without approach organize Our by labels. style any,Our approach can organize galleries of outfits by style without requiring any style labels.,26415
primary (WSI) slide in US. the Whole been diagnosis recently cleared imaging has for,Whole slide imaging (WSI) has recently been cleared for primary diagnosis in the US.,26416
WSI to speed. high challenge in perform focusing A critical accurate of is,A critical challenge of WSI is to perform accurate focusing in high speed.,26417
to Traditional create systems a scanning. focus map prior,Traditional systems create a focus map prior to scanning.,26418
a WSI. report surveying map for Here we focus novel method,Here we report a novel focus map surveying method for WSI.,26419
brightfield WSI. can be fluorescence It both and used for,It can be used for both brightfield and fluorescence WSI.,26420
"propose weakness, for overcome a novel homography To we technique local stereo this rectification. pixel-wise","To overcome this weakness, we propose a novel pixel-wise local homography technique for stereo rectification.",26421
Automatic various slices can for benefit medical recognition part image CT body applications.,Automatic body part recognition for CT slices can benefit various medical image applications.,26422
ordering slice volumes information fully or CT exploited. intrinsic superior-inferior is in The structural not,The intrinsic structural or superior-inferior slice ordering information in CT volumes is not fully exploited.,26423
are method novel A and presented. functions unsupervised CNN inter-sample learning two loss,A novel unsupervised learning method and two inter-sample CNN loss functions are presented.,26424
validate and its effectiveness. qualitative experiments Quantitative,Quantitative and qualitative experiments validate its effectiveness.,26425
"of In show detection. network addition, initialization anomaly applications two and we UBR in","In addition, we show two applications of UBR in network initialization and anomaly detection.",26426
"the of be cardinality. equal required As an complication, are to additional partitions","As an additional complication, the partitions are required to be of equal cardinality.",26427
"EPP. been Previously, SO- for strategies proposed solution only sub-optimal have","Previously, only sub-optimal solution strategies have been proposed for SO- EPP.",26428
"strategy. propose optimal we In the this paper, solution first","In this paper, we propose the first optimal solution strategy.",26429
"the arbitrary constraints first partitioning scheme Furthermore, the to (Constrained is support our on SO-EPP).","Furthermore, our scheme is the first to support arbitrary constraints on the partitioning (Constrained SO-EPP).",26430
"provides compared Being superior schemes. performance optimal, solution to BN-EPP existing","Being optimal, BN-EPP provides superior performance compared to existing solution schemes.",26431
"We inspired problems. additionally scale large solving novel introduce for WalkSAT algorithm BN-EPP a Walk-BN-EPP,","We additionally introduce Walk-BN-EPP, a novel WalkSAT inspired algorithm for solving large scale BN-EPP problems.",26432
an brain improve MRI to training this study CNN-based use segmentation. In we approach adversarial,In this study we use an adversarial training approach to improve CNN-based brain MRI segmentation.,26433
"attempting Currently, research is more to make research comprehensive. this our group","Currently, our research group is attempting to make this research more comprehensive.",26434
"there not the However, many this situations are is case. where","However, there are many situations where this is not the case.",26435
interest in camouflage case. is particular the Of video surveillance,Of particular interest in video surveillance is the camouflage case.,26436
stationary bands. to wavelet The the employs transform into the method image decompose proposed frequency,The proposed method employs the stationary wavelet transform to decompose the image into frequency bands.,26437
"the answer affirmative. is pipeline, carefully-designed that show we With a","With a carefully-designed pipeline, we show that the answer is affirmative.",26438
"generator generated is a annotation. without set large-scale manual GAN, using of First, the image","First, a large-scale image set is generated using the generator of GAN, without manual annotation.",26439
"model Finally, fine-tuned the real is images. on pre-trained","Finally, the pre-trained model is fine-tuned on real images.",26440
challenging test the competitive recognition accuracy datasets. that We proposed framework on show achieves,We show that the proposed framework achieves competitive recognition accuracy on challenging test datasets.,26441
of biometrics. the field challenging Gesture recognition a in is problem,Gesture recognition is a challenging problem in the field of biometrics.,26442
"environments, and has indoor outdoor. in been approach multiple The both tested","The approach has been tested in multiple environments, both indoor and outdoor.",26443
one retinotopic frequently used is Visual mapping the paradigms in fMRI. of most field or,Visual field or retinotopic mapping is one of the most frequently used paradigms in fMRI.,26444
could of basis a in only orientation. difference be patch distinguished the on These,These could only be distinguished on the basis of a difference in patch orientation.,26445
"i.e. part, classification increase to the methods are video-level proposed the classification In performance, two","In the video-level classification part, two methods are proposed to increase the classification performance, i.e.",26446
"and cancer transcriptional is aberrant landscapes. with As it a genomic associated disease,","As a cancer disease, it is associated with aberrant genomic and transcriptional landscapes.",26447
has Integration visualization topic. and been vector of an interesting scalar,Integration of scalar and vector visualization has been an interesting topic.,26448
highly streamlines. of specified then evaluated number It the selects,It then selects the specified number of highly evaluated streamlines.,26449
tasks we In reaching a goal. daily multiple decisions make many before,In many daily tasks we make multiple decisions before reaching a goal.,26450
use asked traces. we eligibility whether humans Here,Here we asked whether humans use eligibility traces.,26451
major of diseases (HF) heart problem. Management is failure chronic a as such public health,Management of chronic diseases such as heart failure (HF) is a major public health problem.,26452
It under for is in acceptance consideration TPLP.,It is under consideration for acceptance in TPLP.,26453
"change a by method propose By of regarding labels novel perspective, we words. as","By change of perspective, we propose a novel method by regarding labels as words.",26454
"models In normalizations addition, LSTM models. improve our our we to into batch adopted","In addition, we adopted batch normalizations into our models to improve our LSTM models.",26455
"classifiers. models our other they can Since extractors, used with be feature are","Since our models are feature extractors, they can be used with other classifiers.",26456
"blending is one, Analogical is reasoning another.","Analogical reasoning is one, blending is another.",26457
are vision Many cost the as problems of optimization formulated function. computer a,Many computer vision problems are formulated as the optimization of a cost function.,26458
of understanding population in problem the fluctuations. A is central stochastic ecology consequences,A central problem in population ecology is understanding the consequences of stochastic fluctuations.,26459
its great practical recently values. attention search (MLS) receives increasing for Mobile landmark,Mobile landmark search (MLS) recently receives increasing attention for its great practical values.,26460
"remains However, two still due challenges. unsolved to it important","However, it still remains unsolved due to two important challenges.",26461
Landmark-based task in in computer vision. recognition videos challenging a human action is,Landmark-based human action recognition in videos is a challenging task in computer vision.,26462
that in shows and Evidence psychology subtle. facial short responses experimental discriminative are,Evidence in experimental psychology shows that discriminative facial responses are short and subtle.,26463
"which be improves learned overall the the performance. more informative, features training, Through can end-to-end","Through end-to-end training, the learned features can be more informative, which improves the overall performance.",26464
benchmark proposed performance on datasets. several method achieved has competitive Our,Our proposed method has achieved competitive performance on several benchmark datasets.,26465
an is inputs. with in efficient deal reliability Fuzzy to logic the uncertainty method,Fuzzy logic is an efficient method to deal with the uncertainty in reliability inputs.,26466
a have resurgence witnessed in video of interest years summarization. Recent,Recent years have witnessed a resurgence of interest in video summarization.,26467
subjectiveness two The at least causes problems.,The subjectiveness causes at least two problems.,26468
"a video the to evaluate it performance of challenging very is Second, summarizer.","Second, it is very challenging to evaluate the performance of a video summarizer.",26469
"other in our training used Besides, pre-trained end-to-end models. without adopts deep model","Besides, our model adopts end-to-end training without pre-trained used in other deep models.",26470
progress all across require Advances Intelligence Artificial in computer science. of,Advances in Artificial Intelligence require progress across all of computer science.,26471
the video-level leveraged in and both submission. We features frame-level,We leveraged both video-level and frame-level features in the submission.,26472
TPLP. consideration under is paper for acceptance in This,This paper is under consideration for acceptance in TPLP.,26473
notion that of his positive has valuations. proved graphoidal properties independence Shenoy for normal,Shenoy proved that his notion of independence has graphoidal properties for positive normal valuations.,26474
especially probabilistic functions. so-called belief excludes It,It excludes especially so-called probabilistic belief functions.,26475
"These however, from are perspectives. approaches, problematic two","These approaches, however, are problematic from two perspectives.",26476
image. Most are on image of enhancement the color each channel-by-channel algorithms applied,Most of the color image enhancement algorithms are applied channel-by-channel on each image.,26477
processing image system processing of But color not original is the a such color.,But such a system of color image processing is not processing the original color.,26478
an to before anticipation it aims Action detect happens. action,Action anticipation aims to detect an action before it happens.,26479
robotics this to are applications predictive world real Many capability. related in surveillance and,Many real world applications in robotics and surveillance are related to this predictive capability.,26480
"the trend. past based is which anticipation However, a on single ignores history frame's representation,","However, anticipation is based on a single past frame's representation, which ignores the history trend.",26481
"it future a anticipate time. fixed only Besides, can","Besides, it can only anticipate a fixed future time.",26482
for Reinforced a anticipation. action propose Encoder-Decoder We (RED) network,We propose a Reinforced Encoder-Decoder (RED) network for action anticipation.,26483
approximate The GANs adversarive probability. to game an promote and jointed complex example,The GANs promote an adversarive game to approximate complex and jointed example probability.,26484
data noise generate networks approximate fake by examples The realistic to driven distributions.,The networks driven by noise generate fake examples to approximate realistic data distributions.,26485
attributes the labels. aligned coarse to classifier are images as inputs embedded and The generator,The coarse images aligned to attributes are embedded as the generator inputs and classifier labels.,26486
And circularly to of the training images. generated degeneration is prevent implemented adversarial,And adversarial training is circularly implemented to prevent degeneration of the generated images.,26487
algorithms? better we Can data machine evolve training for learning,Can we evolve better training data for machine learning algorithms?,26488
reasons. results two are for important These,These results are important for two reasons.,26489
is A scratch. sub-network convolutional pre-trained deep designed from first and,A deep convolutional sub-network is first designed and pre-trained from scratch.,26490
first this simulation the numerical Extensive to conducted method. of performance studies were evaluate,Extensive numerical simulation studies were first conducted to evaluate the performance of this method.,26491
The work demonstrate effectively that results experimental for proposed can Go FDAA applications. the,The experimental results demonstrate that the proposed FDAA can work effectively for Go applications.,26492
"approaches. length of smaller lengths expected competing than its other expected Also, is","Also, its expected length is smaller than expected lengths of other competing approaches.",26493
"Finally, two data the illustrate proposed sets. method using we real","Finally, we illustrate the proposed method using two real data sets.",26494
considered Three the establish data proof-of-concept. to are sets,Three data sets are considered to establish the proof-of-concept.,26495
obtained show the that proposed is The results effective. method,The results obtained show that the proposed method is effective.,26496
paradigm. (ASP) Set well-established is Answer Programming declarative a,Answer Set Programming (ASP) is a well-established declarative paradigm.,26497
of of of ASP successes the efficient availability is systems. the One,One of the successes of ASP is the availability of efficient systems.,26498
ground+solve approach. systems the State-of-the-art are based on,State-of-the-art systems are based on the ground+solve approach.,26499
the and benchmarks strategies. highlight Results and real strengths synthetic some of weaknesses on different,Results on real and synthetic benchmarks highlight some strengths and weaknesses of the different strategies.,26500
image approach improving for an context object detection. Using is effective,Using image context is an effective approach for improving object detection.,26501
semantic or Previously rely proposed methods used information. cues spatial on contextual that,Previously proposed methods used contextual cues that rely on semantic or spatial information.,26502
"similarity. information: of kind In inner-scene different this we work, explore a contextual","In this work, we explore a different kind of contextual information: inner-scene similarity.",26503
a by stage. uses base-scores post-detection performs provided as base a and CISS detector,CISS uses base-scores provided by a base detector and performs as a post-detection stage.,26504
the dependency a of done by base-scores pre-trained the supporters and model. This is utilizing,This is done by utilizing the base-scores of the supporters and a pre-trained dependency model.,26505
"theory,"" study. Duncan and work to a psychophysical Humphreys' This ""similarity relates","This work relates to Duncan and Humphreys' ""similarity theory,"" a psychophysical study.",26506
the detector's dataset. base a Experimental enhancement PASCAL on VOC scores results the demonstrate of,Experimental results demonstrate the enhancement of a base detector's scores on the PASCAL VOC dataset.,26507
Whittle's is likelihood to used pseudo-posterior the obtain approximation distribution.,Whittle's likelihood approximation is used to obtain the pseudo-posterior distribution.,26508
apply to mean estimate algorithm sunspot to the solar the cycle. We annual data,We apply the algorithm to annual mean sunspot data to estimate the solar cycle.,26509
"work mechanisms in e.g. study In vSLAM, we optimization of this post-processing the core one","In this work we study one of the core post-processing optimization mechanisms in vSLAM, e.g.",26510
the to solution. geometric shape causes This bias boundary strongly of the the,This causes the geometric shape of the boundary to strongly bias the solution.,26511
"the of energy. the we squared Frobenious smoothness a as using propose Instead, Hessian norm","Instead, we propose using the squared Frobenious norm of the Hessian as a smoothness energy.",26512
negatively the We a gene. underlying characterize self-regulating symmetries solutions of the exact chemically stochastic,We chemically characterize the symmetries underlying the exact solutions of a stochastic negatively self-regulating gene.,26513
symmetry breaking three causes number effects. at The molecular of low,The breaking of symmetry at low molecular number causes three effects.,26514
number from expected protein value. the deterministically Average differs,Average protein number differs from the deterministically expected value.,26515
models variables existence assuming of hidden under are development. Hence,Hence models assuming existence of hidden variables are under development.,26516
non-trivial since result is e.g. as considered to This be,This result is to be considered as non-trivial since e.g.,26517
deep is The on neural networks. convolutional method based,The method is based on deep convolutional neural networks.,26518
"catheter the the Subsequently, is segmented a of constructed from image. model centerline","Subsequently, a centerline model of the catheter is constructed from the segmented image.",26519
the underrepresented-classes) those of classes. normal to,underrepresented-classes) to those of the normal classes.,26520
"have proposed. stage using actively of multi-scale features a lot detectors Recently, been single","Recently, a lot of single stage detectors using multi-scale features have been actively proposed.",26521
well important against to worth problem researching. defense examples which an is adversarial is How,How to defense against adversarial examples is an important problem which is well worth researching.,26522
"methods far, very So to have a significant few adversarial examples. provided defense","So far, very few methods have provided a significant defense to adversarial examples.",26523
limit diffusion mutations. equation Kolmogorov the of neutral for The derived is forward case,The diffusion limit forward Kolmogorov equation is derived for the case of neutral mutations.,26524
approximate rates. of time-dependent the obtained limit solution low in An is mutation,An approximate time-dependent solution is obtained in the limit of low mutation rates.,26525
using simulations. approximate The confirmed is solution numerical,The approximate solution is confirmed using numerical simulations.,26526
"models. non-linear kernel subspace method that paper, present this In can we handle a clustering","In this paper, we present a kernel subspace clustering method that can handle non-linear models.",26527
are multitude multiscale exhibit of Biofilms behaviors. that biomechanical communities a bacteria of,Biofilms are communities of bacteria that exhibit a multitude of multiscale biomechanical behaviors.,26528
"be label by information co-occurrence jointly More model. exploited our can LSTM importantly,","More importantly, label co-occurrence information can be jointly exploited by our LSTM model.",26529
computer task challenging a is Visual object applications. with numerous real-world tracking vision,Visual object tracking is a challenging computer vision task with numerous real-world applications.,26530
we Here Tracking Spectral propose (SFT)method. a efficient but Filter simple,Here we propose a simple but efficient Spectral Filter Tracking (SFT)method.,26531
contain connections; CNNs i.e. shortcut Recent deep forward,Recent deep CNNs contain forward shortcut connections; i.e.,26532
skip connections from high low to layers.,skip connections from low to high layers.,26533
from to layers low high layers. paths,paths from high layers to low layers.,26534
label to The the used is network final Master produce predictions.,The Master network is used to produce the final label predictions.,26535
Type-I is Hybrid mixture censoring A of censoring scheme Type-II and schemes.,A Hybrid censoring scheme is mixture of Type-I and Type-II censoring schemes.,26536
"obtained. approximate the are likelihood and MLE (AMLE) (MLE), maximum estimator R of The","The maximum likelihood estimator (MLE), and the approximate MLE (AMLE) of R are obtained.",26537
asymptotic distribution is maxi- of likelihood The the obtained. R estimator mum of,The asymptotic distribution of the maxi- mum likelihood estimator of R is obtained.,26538
"R asymptotic interval tion, distribu- confidence derived. on the the Based of be can","Based on the asymptotic distribu- tion, the confidence interval of R can be derived.",26539
bootstrap confidence proposed. Two are also intervals,Two bootstrap confidence intervals are also proposed.,26540
a Analysis real also illustrative of data presented has purposes. set been for,Analysis of a real data set has also been presented for illustrative purposes.,26541
semantic propose approach weakly-supervised for We discover task. the segmentation pixels class-specific to an,We propose an approach to discover class-specific pixels for the weakly-supervised semantic segmentation task.,26542
to Modelling should accommodate neuroscience aim analysis phenomena. for multiscale methods and,Modelling and analysis methods for neuroscience should aim to accommodate multiscale phenomena.,26543
"Answer technology, paper, Set underlying mentioned only is Programming that The quickly in (ASP).","The underlying technology, only quickly mentioned in that paper, is Answer Set Programming (ASP).",26544
"we to detail problem. we ASP the used approach solve work, In this such","In this work, we detail the ASP approach we used to solve such problem.",26545
We to such Constraint problem how (CLP(FD)). address on in Programming Finite show Logic Domains,We show how to address such problem in Constraint Logic Programming on Finite Domains (CLP(FD)).,26546
Theory publication and paper possible for is Logic under consideration Programming. of Practice on This,This paper is under consideration for possible publication on Theory and Practice of Logic Programming.,26547
"Firstly, of approach propose et we proposed Germain the we in al. improvement an previous","Firstly, we propose an improvement of the previous approach we proposed in Germain et al.",26548
"we from discuss compare PAC-Bayesian results, obtain both generalization We bounds. and which","We discuss and compare both results, from which we obtain PAC-Bayesian generalization bounds.",26549
the thus to It methods. is application of importance great practical of such,It is thus of great practical importance to the application of such methods.,26550
This problems. preserving important many reflecting in renders and structure them incapable of,This renders them incapable of reflecting and preserving important structure in many problems.,26551
thus networks. spiking neuronal Coupled informs theory oscillator synchronization the of,Coupled oscillator theory thus informs the synchronization of spiking neuronal networks.,26552
"the have LSTM network explicit attention original not ability. However, does","However, the original LSTM network does not have explicit attention ability.",26553
"we a train order training network effectively. in propose scheme Moreover, our to stepwise","Moreover, we propose a stepwise training scheme in order to train our network effectively.",26554
state-of-the-art skeleton for approach achieves recognition. benchmark on datasets performance challenging based action five Our,Our approach achieves state-of-the-art performance on five challenging benchmark datasets for skeleton based action recognition.,26555
proceeds a reasoning occlusion occlusion search hypothesize-and-test labels. over by Our,Our occlusion reasoning proceeds by a hypothesize-and-test search over occlusion labels.,26556
"face and of perform extensive evaluation large unoccluded. occluded on number images, both a We","We perform extensive evaluation on a large number of face images, both occluded and unoccluded.",26557
propose image procedure and prediction. iterative an score type updating for event importance We,We propose an iterative updating procedure for event type and image importance score prediction.,26558
presents expediting generic for a This pre-processor conventional matching techniques. paper template,This paper presents a generic pre-processor for expediting conventional template matching techniques.,26559
"arbitrary reference Moreover, it scaling and effectively. of handle images rotation can","Moreover, it can handle arbitrary rotation and scaling of reference images effectively.",26560
processes. simple that combines a neural propose method Gaussian networks and We,We propose a simple method that combines neural networks and Gaussian processes.,26561
"functions Gaussian of With the processes. are mean networks the method, used neural proposed for","With the proposed method, neural networks are used for the mean functions of Gaussian processes.",26562
capture to among details networks deep learning have difficulties nuance Traditional categories. different tubule,Traditional deep learning networks have difficulties to capture nuance details among different tubule categories.,26563
approach on two our with subsets We experiments dataset. the of different two demonstrate,We demonstrate our approach with two experiments on two different subsets of the dataset.,26564
a framework way representing influential highly conceptual knowledge. spaces geometric of of The provides,The highly influential framework of conceptual spaces provides a geometric way of representing knowledge.,26565
"PIs given. simple instances are of Then, some","Then, some simple instances of PIs are given.",26566
performance These PIs. ways enhance of the,These ways enhance the performance of PIs.,26567
"out synthetic real image based on carry and datasets. Finally, experiments we some","Finally, we carry out some experiments based on synthetic and real image datasets.",26568
used commonly comparison. moment for We invariants choose,We choose commonly used moment invariants for comparison.,26569
discriminative learning. two transformation fuzzy framework consists subspace sparse clustering of and of modules This,This framework consists of two modules of fuzzy sparse subspace clustering and discriminative transformation learning.,26570
"novel, a purely matting natural image affinity-based algorithm. We present","We present a novel, purely affinity-based natural image matting algorithm.",26571
"have decoder While address relatively few the studies been side. rigorously, studied encoders","While encoders have been studied rigorously, relatively few studies address the decoder side.",26572
"visualization. semantic and image mapping, segmentation pixel-level based superpixel","superpixel mapping, pixel-level segmentation and semantic based image visualization.",26573
"learning $F$, involving targets problems study arbitrary of $Y$. functions We classes and distributions $X$","We study learning problems involving arbitrary classes of functions $F$, distributions $X$ and targets $Y$.",26574
"to And convex, procedure out turns if $F$ be is the proper.","And if $F$ is convex, the procedure turns out to be proper.",26575
and quantitative available publicly egocentric on video Qualitative our comparisons validate datasets results.,Qualitative and quantitative comparisons on publicly available egocentric video datasets validate our results.,26576
units facial (AUs) is expressions still facial from challenging a action Recognizing problem. spontaneous,Recognizing facial action units (AUs) from spontaneous facial expressions is still a challenging problem.,26577
"Most shown CNNs recently, facial on AU have recognition. promise","Most recently, CNNs have shown promise on facial AU recognition.",26578
mixture Type and censoring I Type II The of hybrid schemes. is a censoring,The hybrid censoring is a mixture of Type I and Type II censoring schemes.,26579
of consider the likelihood we unknown the estimators maximum parameters. First,First we consider the maximum likelihood estimators of the unknown parameters.,26580
form. can obtained closed estimators It observed in maximum not that likelihood the be is,It is observed that the maximum likelihood estimators can not be obtained in closed form.,26581
the using estimators We Lindley's Bayes compute technique. also approximation approximate,We also compute the approximate Bayes estimators using Lindley's approximation technique.,26582
prune maintaining is heavily edges convolutional performance. to objective retrieval while The,The objective is to heavily prune convolutional edges while maintaining retrieval performance.,26583
"introduced scheme efficient uses that novel Furthermore, extremely is a an approximation over-segmentation.","Furthermore, a novel approximation scheme is introduced that uses an extremely efficient over-segmentation.",26584
divided content into in a can be motion. signals and video Visual,Visual signals in a video can be divided into content and motion.,26585
"which motion are specifies video, describes in objects the dynamics. their content While","While content specifies which objects are in the video, motion describes their dynamics.",26586
random Each part a of part. motion and content vector consists a,Each random vector consists of a content part and a motion part.,26587
"of This logic, quantum propositional a as propositions. proposes using observables formulation Eigenlogic, named work","This work proposes a formulation of propositional logic, named Eigenlogic, using quantum observables as propositions.",26588
fashion demands The customers. effective for of recommendation shopping online services ubiquity,The ubiquity of online fashion shopping demands effective recommendation services for customers.,26589
existing method that the correlates reveals memorability. with analysis Detailed feature on findings proposed well,Detailed feature analysis reveals that the proposed method correlates well with existing findings on memorability.,26590
particular general models trained for a can like (i.e. outperform dresses) significantly vertical models,models trained for a particular vertical like dresses) can significantly outperform general models (i.e.,26591
range a cover verticals). wide that models of,models that cover a wide range of verticals).,26592
"models verticals. are often network separately for different Therefore, neural trained deep","Therefore, deep neural network models are often trained separately for different verticals.",26593
"different However, models not scalable using expensive and to is specialized for verticals deploy.","However, using specialized models for different verticals is not scalable and expensive to deploy.",26594
problem training and is from data The training tackled difficulty. two aspects:,The problem is tackled from two aspects: training data and training difficulty.,26595
our experiments. is approach in of The effectiveness shown,The effectiveness of our approach is shown in experiments.,26596
"leads often sampling patches random and model criteria training. to selection for redundant However, noisy-image","However, random sampling selection criteria often leads to redundant and noisy-image patches for model training.",26597
"automatic based EnhanceGAN, adversarial that image model learning enhancement. performs an We introduce","We introduce EnhanceGAN, an adversarial learning based model that performs automatic image enhancement.",26598
technique bone good essential characterization is texture identifying osteoporosis for A cases.,A good bone texture characterization technique is essential for identifying osteoporosis cases.,26599
resurgence. AI the has Machine a Learning been story big during success,Machine Learning has been a big success story during the AI resurgence.,26600
massive from relates out of success amount particular stand data. learning One a to,One particular stand out success relates to learning from a massive amount of data.,26601
they have a CNN single-image super-resolution. have that significant reconstruction shown Deep recently on performance,Deep CNN have recently shown that they have a significant reconstruction performance on single-image super-resolution.,26602
using layers CNN trend deeper performance. Current to is improve,Current trend is using deeper CNN layers to improve performance.,26603
"power, modifies of state. light's Each these events path, polarization","Each of these events modifies light's path, power, polarization state.",26604
material light of BRDF inside requires the Computing subsurface a polarized transport the simulating material.,Computing polarized subsurface BRDF of a material requires simulating the light transport inside the material.,26605
solution expensive. that to Computing is equation,Computing solution to that equation is expensive.,26606
Method is solving common Discrete to The (DOM) the Ordinate a VRTE. approach,The Discrete Ordinate Method (DOM) is a common approach to solving the VRTE.,26607
us This of allows solve to the independently for VRTE order each expansion.,This allows us to independently solve the VRTE for each order of expansion.,26608
with We objects the computed also of the rendering show BRDF.,We also show the rendering of objects with the computed BRDF.,26609
available for is The authors' web the public site. use through solver,The solver is available for public use through the authors' web site.,26610
"Compared scalable to human easier and obtain. more with DWEs attributes, defined are","Compared with human defined attributes, DWEs are more scalable and easier to obtain.",26611
the the of neighbourhood visual becomes that VAWEs Thus similar to domain. in structure,Thus the neighbourhood structure of VAWEs becomes similar to that in the visual domain.,26612
allows strategy features. the visual VAWE generalize methods to and learned various ZSL This to,This strategy allows the learned VAWE to generalize to various ZSL methods and visual features.,26613
robustly in to detect then These each frame. combined are people,These are then combined to detect people robustly in each frame.,26614
image detections full-body tubes. temporally Our tracking to algorithm the extract human connects,Our tracking algorithm connects the image detections temporally to extract full-body human tubes.,26615
manner. invariant representation convolutional captures second-order of a statistics The in resulting translationally features,The resulting representation captures second-order statistics of convolutional features in a translationally invariant manner.,26616
effect investigate the how also evaluation. network of training accuracy matrix function and computations We,We also investigate how the accuracy of matrix function computations effect network training and evaluation.,26617
"depth be issues, applications. to or infrared illumination real However, in required due may cameras","However, due to illumination issues, infrared or depth cameras may be required in real applications.",26618
multi-layer generalize and cases. this algorithm multi-branch further We to,We further generalize this algorithm to multi-layer and multi-branch cases.,26619
enhance reduces various method the architectures. error accumulated compatibility with and the Our,Our method reduces the accumulated error and enhance the compatibility with various architectures.,26620
a CT lesions in great detection researchers. poses for liver Automatic of challenge images,Automatic detection of liver lesions in CT images poses a great challenge for researchers.,26621
highly outperform the network. convolutional fully approach detection that results shows state-of-the-art improved Our,Our approach shows highly improved detection results that outperform the state-of-the-art fully convolutional network.,26622
"are Therefore, properties spectral connected. their tightly","Therefore, their spectral properties are tightly connected.",26623
in simulations that classification effective tasks. is regularization demonstrate Numerical proposed the,Numerical simulations demonstrate that the proposed regularization is effective in classification tasks.,26624
"category-based detection, object temporal tracking. approach object Our appearance object segmentation combines category-independent and","Our approach combines category-based object detection, category-independent object appearance segmentation and temporal object tracking.",26625
on methods data. evaluate paper data-dependent of In performance hashing we binary this,In this paper we evaluate performance of data-dependent hashing methods on binary data.,26626
work. search person We in problem of this investigate wild the the in,We investigate the problem of person search in the wild in this work.,26627
"Will order canceled be or an delivered, withdrawn?","Will an order be delivered, canceled or withdrawn?",26628
"which is for even recognition difficult is vs. sunset task, Sunrise humans. a challenging","Sunrise vs. sunset is a difficult recognition task, which is challenging even for humans.",26629
a date line are words. from segmenting without models searched characters or Different,Different date models are searched from a line without segmenting characters or words.,26630
sub-bands Wavelet to information enhance used gradient gray and decomposition in scale. text are,Wavelet decomposition and gradient sub-bands are used to enhance text information in gray scale.,26631
image and gray MLP based are Tandem features by Binary combined approach.,Binary and gray image features are combined by MLP based Tandem approach.,26632
earliest the diabetic are retinopathy of sign disease. Retinal clinical microaneurysms,Retinal microaneurysms are the earliest clinical sign of diabetic retinopathy disease.,26633
develop basic efficient an approach. study properties for and estimation of our We algorithm,We develop an efficient algorithm for estimation and study basic properties of our approach.,26634
"state-of-the-art. this of Experimental results effectiveness strategy, which demonstrate advanced the has the","Experimental results demonstrate the effectiveness of this strategy, which has advanced the state-of-the-art.",26635
features than markers provide ones better-defined the Fiducial available naturally the scene. in,Fiducial markers provide better-defined features than the ones naturally available in the scene.,26636
"estimation. provides paper, In this a fiducial marker pose STag, that propose stable system we","In this paper, we propose STag, a fiducial marker system that provides stable pose estimation.",26637
localized more be is geometric This utilizing features that repeatably. achieved by can,This is achieved by utilizing geometric features that can be localized more repeatably.,26638
outer detection is homography marker estimation. border and the The square for used of,The outer square border of the marker is used for detection and homography estimation.,26639
refinement novel by using border. step the homography a This inner circular followed is,This is followed by a novel homography refinement step using the inner circular border.,26640
"robustly viewing can and the stably refinement, pose After across estimated conditions. be","After refinement, the pose can be estimated stably and robustly across viewing conditions.",26641
"Preparing slides of each scanning parameters. histopathology multitude and a of steps, consists several with","Preparing and scanning histopathology slides consists of several steps, each with a multitude of parameters.",26642
"a this networks. In systematic solution on propose domain-adversarial we based paper, neural","In this paper, we propose a systematic solution based on domain-adversarial neural networks.",26643
domain model representation leads information better that removing We generalization. hypothesize the to the from,We hypothesize that removing the domain information from the model representation leads to better generalization.,26644
"paper, In reusability models. focus deep the of pre-trained we on convolutional this","In this paper, we focus on the reusability of pre-trained deep convolutional models.",26645
the DDT of method. the proposed Empirical validate effectiveness studies,Empirical studies validate the effectiveness of the proposed DDT method.,26646
"consistently a large On methods datasets, co-localization by existing benchmark state-of-the-art margin. image DDT outperforms","On benchmark image co-localization datasets, DDT consistently outperforms existing state-of-the-art methods by a large margin.",26647
the classical chain. show our comparison that to improvements We achieves significant in method,We show that our method achieves significant improvements in comparison to the classical chain.,26648
"to is The of goal objects, retrieve interior e.g. our engine","The goal of our engine is to retrieve interior objects, e.g.",26649
"with wall and the clocks, similarities that or aesthetic visual share query. furniture","furniture or wall clocks, that share visual and aesthetic similarities with the query.",26650
"to that and return fit well stylistically Additionally, items together. it aesthetically other allows","Additionally, it allows to return other items that aesthetically and stylistically fit well together.",26651
"our using system visual blends goal, and this textual To modalities. the results achieve obtained","To achieve this goal, our system blends the results obtained using textual and visual modalities.",26652
learning of we the To new a scheme. hybrid descriptors address propose binary learning problem,To address the problem of learning binary descriptors we propose a new hybrid learning scheme.,26653
results benchmarks. We show available binary public descriptors and several for methods compare training on,We compare several methods for training binary descriptors and show results on public available benchmarks.,26654
scale is converted using This a weighted to then image scaling. gray,This image is then converted to a gray scale using weighted scaling.,26655
by edges curvelet detail the vessel enhanced are coefficients. The boosting,The vessel edges are enhanced by boosting the detail curvelet coefficients.,26656
fuzzy Optic to classification disk pixels are applying removed avoid C-mean misclassification. the before,Optic disk pixels are removed before applying fuzzy C-mean classification to avoid the misclassification.,26657
component vessels. segmented the connected to are Morphological obtain applied retinal operations analysis and,Morphological operations and connected component analysis are applied to obtain the segmented retinal vessels.,26658
"parameter between pruning Additionally, neuron shown. reduction comparison a is and weight","Additionally, a parameter reduction comparison between neuron and weight pruning is shown.",26659
"of cost region and the evaluating generating However, computational heavy. is proposals","However, the computational cost of generating and evaluating region proposals is heavy.",26660
"the is knowledge, approach so. first to our best To the of this do","To the best of our knowledge, this is the first approach to do so.",26661
efficient semantic neural segmentation. in enables fully end-to-end network convolutional of learning development Recent,Recent development in fully convolutional neural network enables efficient end-to-end learning of semantic segmentation.,26662
experiments segmentation. semantic the are of show conducted Extensive RAN effectiveness in the to,Extensive experiments are conducted to show the effectiveness of the RAN in semantic segmentation.,26663
chemicals How a can of bag micron-sized smart be?,How smart can a micron-sized bag of chemicals be?,26664
inferences artificial make about real How can or an its environment? cell,How can an artificial or real cell make inferences about its environment?,26665
networks chemical can of From reaction sample? distributions probability which kinds,From which kinds of probability distributions can chemical reaction networks sample?,26666
in detection vision. is a fundamental problem computer Face,Face detection is a fundamental problem in computer vision.,26667
"with deep represented faces and coarse features. powerful both can be fine-grained small Specifically, shallow","Specifically, small faces can be represented with both shallow fine-grained and deep powerful coarse features.",26668
classification raw process to the method automatic the maps data that an Our provides results.,Our method provides an automatic process that maps the raw data to the classification results.,26669
"segmentation framework. also classification and object integrates It into localization, one","It also integrates object localization, segmentation and classification into one framework.",26670
the results methods. demonstrate method classification state-of-the-art cloud that the proposed outperforms point Experimental,Experimental results demonstrate that the proposed method outperforms the state-of-the-art point cloud classification methods.,26671
in gives disscussed cvpaper.challenge. challenges The futuristic paper the,The paper gives futuristic challenges disscussed in the cvpaper.challenge.,26672
"transformation orthogonal introduce we linear and transformation. this two transforms, idempotent alternative In paper,","In this paper, we introduce two alternative linear transforms, orthogonal transformation and idempotent transformation.",26673
image different publicly on text scene/video available datasets. tested Our framework been has,Our framework has been tested on different publicly available scene/video text image datasets.,26674
"Devanagari data script, our collected own dataset. For we","For Devanagari script, we collected our own data dataset.",26675
"has However, it revisit time a span and schedule. irregular short","However, it has a short time span and irregular revisit schedule.",26676
and physiography. well regions LSTM with generalizes distinct across climates,LSTM generalizes well across regions with distinct climates and physiography.,26677
progress alignment in has substantial decade. the Face witnessed last,Face alignment has witnessed substantial progress in the last decade.,26678
analysis geometry of role applications. in many pivotal shape shape a Knowledge plays,Knowledge of shape geometry plays a pivotal role in many shape analysis applications.,26679
to known explicit well the evaluation is of probabilities model be infeasible. An,An explicit evaluation of the model probabilities is well known to be infeasible.,26680
"address Olsson consider we strategy et of issue, Gibbs To particle al. the this","To address this issue, we consider the particle Gibbs strategy of Olsson et al.",26681
many Reversibility scientific years. is over of topic intensively disciplines transitions state many in studied,Reversibility of state transitions is intensively studied topic in many scientific disciplines over many years.,26682
their proposed Two is cell of are heterogeneous interrelated and models behavior studied. populations,Two interrelated models of heterogeneous cell populations are proposed and their behavior is studied.,26683
detection and aims at Object accuracy high simultaneously. speed,Object detection aims at high speed and accuracy simultaneously.,26684
"Indicator. the Length this fluctuations (""noise"") Typical example, Fish Large than exhibits random In smaller","In this example, Typical Length exhibits smaller random fluctuations (""noise"") than the Large Fish Indicator.",26685
problems. but approaches solve Heuristic to are simplest necessary the all,Heuristic approaches are necessary to solve all but the simplest problems.,26686
"domain-independent In explore learning. the using this heuristic we work, obtaining machine functions of possibility","In this work, we explore the possibility of obtaining domain-independent heuristic functions using machine learning.",26687
architecture reuse. is efficient highly of  as a computationally feature  result DenseNet The,  The DenseNet architecture is highly computationally efficient as a result of feature reuse.,26688
"the train bottleneck, to now GPU DenseNets. possible it memory Without is deep extremely","Without the GPU memory bottleneck, it is now possible to train extremely deep DenseNets.",26689
"refined DMRI the motion the estimated MC. With through field, reconstruction is","With the estimated motion field, the DMRI reconstruction is refined through MC.",26690
focusing works have directly been on Recent single image. a reconstruction from,Recent works have been focusing on reconstruction directly from a single image.,26691
registration several by We comparing datasets. evaluate method our across quality,We evaluate our method by comparing registration quality across several datasets.,26692
"crowd for neural counting. (CNN) Recently, convolutional promising have achieved models results network","Recently, convolutional neural network (CNN) models have achieved promising results for crowd counting.",26693
"temporal CNN-based our previous both Unlike captures methods, spatial method the dependencies. and fully","Unlike the previous CNN-based methods, our method fully captures both spatial and temporal dependencies.",26694
dataset. state of We line the achieve the text on RIMES art for handwritten,We achieve state of the art for handwritten line text on the RIMES dataset.,26695
"and means of sensing expressions auxiliary are Hence, needed. conveying these","Hence, auxiliary means of sensing and conveying these expressions are needed.",26696
compartment. is to desirable performed The according the of of variants each selection phenotype,The selection of desirable variants is performed according to the phenotype of each compartment.,26697
resolution higher for of monitoring technologies offer Mobile health conditions. opportunities,Mobile technologies offer opportunities for higher resolution monitoring of health conditions.,26698
results trained Cityscapes achieves GridNet on scratch the is from dataset. competitive and,GridNet is trained from scratch and achieves competitive results on the Cityscapes dataset.,26699
"of parts semantic the on this we partially occluded paper, In detecting task address objects.","In this paper, we address the task of detecting semantic parts on partially occluded objects.",26700
cues. Our of local visual the confidence accumulating parts semantic by detects approach,Our approach detects semantic parts by accumulating the confidence of local visual cues.,26701
VehicleSemanticPart voting scheme the part We on dataset evaluate with annotations. our dense,We evaluate our voting scheme on the VehicleSemanticPart dataset with dense part annotations.,26702
these full can the image multi-scale Make of information performance. restoration improve use,Make full use of these multi-scale information can improve the image restoration performance.,26703
our method image results methods. outperforms proposed show that super-resolution state-of-the-art single many Experimental,Experimental results show that our proposed method outperforms many state-of-the-art single image super-resolution methods.,26704
methods. methods patch re-arrangement approaches two The are statistics-based and main,The two main approaches are statistics-based methods and patch re-arrangement methods.,26705
to their methods approaches ideas from hurdles. avoid combine Hybrid to try both,Hybrid methods try to combine ideas from both approaches to avoid their hurdles.,26706
various on impressive of textures. produce kinds They synthesis,They produce impressive synthesis on various kinds of textures.,26707
descriptor patch of efficient We local-patch based multiple-kernel a propose kernels match gradients. on,We propose a multiple-kernel local-patch descriptor based on efficient match kernels of patch gradients.,26708
we theory that new sensorimotor these of observations. propose explains Here learning a,Here we propose a new theory of sensorimotor learning that explains these observations.,26709
method conceptually open and for We multi-class flexible a classification. new set present,We present a conceptually new and flexible method for multi-class open set classification.,26710
propose detector a general purpose uses structure filters. B-COSFIRE brain-inspired the that trainable We curvilinear,We propose a general purpose curvilinear structure detector that uses the brain-inspired trainable B-COSFIRE filters.,26711
"are that and distortion, two-dimensional transformation. noise, robust images synthesize affine adversarial to We","We synthesize two-dimensional adversarial images that are robust to noise, distortion, and affine transformation.",26712
over proposed approach datasets. several effectiveness Experiments demonstrate of the benchmark the,Experiments demonstrate the effectiveness of the proposed approach over several benchmark datasets.,26713
"used only Softmax classification loss. the but is usually it supervision, as penalizes","Softmax is usually used as supervision, but it only penalizes the classification loss.",26714
effectiveness of the different datasets on demonstrate contrastive-center Experiments loss.,Experiments on different datasets demonstrate the effectiveness of contrastive-center loss.,26715
It partitions grid of last first outputs several ConvNet regions. the into convolution,It first partitions the last convolution outputs of ConvNet into several grid regions.,26716
"makes our this and method formulation super-effective. Besides, easy-to-implement","Besides, this formulation makes our method easy-to-implement and super-effective.",26717
approaches. our comparison results of demonstrate approach in the Experimental effectiveness the state-of-the-art with,Experimental results demonstrate the effectiveness of our approach in comparison with the state-of-the-art approaches.,26718
"the to make we In field. this contributions work, two","In this work, we make two contributions to the field.",26719
problem confidence. accurately paper we the of In estimating this predictive consider,In this paper we consider the problem of accurately estimating predictive confidence.,26720
"on propose approaches. based address confidence this we modelling To density novel a measure issue,","To address this issue, we propose a novel confidence measure based on density modelling approaches.",26721
models CS superior performance. and Several have obtained recently proposed been reconstruction,Several CS reconstruction models have been recently proposed and obtained superior performance.,26722
"within two theory. challenges there CS However, exist important still the","However, there still exist two important challenges within the CS theory.",26723
"two this we try paper, problems a to deep with network. In with deal these","In this paper, we try to deal with these two problems with a deep network.",26724
"(CC) to connected text strings. are Next, obtained skeletonized components distinguish the various","Next, the obtained connected components (CC) are skeletonized to distinguish various text strings.",26725
text/non-text these hypotheses HMM-based text classification using verified system. Finally are,Finally these text hypotheses are verified using HMM-based text/non-text classification system.,26726
giving us robust a False text performance. positives eliminated detection are thus,False positives are thus eliminated giving us a robust text detection performance.,26727
show approach obtained existing methods results The text that proposed on surpasses detection.,The results obtained show that proposed approach surpasses existing methods on text detection.,26728
existing spatial correspondence. within regions Most exploit perform local in methods matching local manipulation to,Most existing methods exploit local regions within spatial manipulation to perform matching in local correspondence.,26729
"thus, leads This to pair-wise representations for flexible images.","This thus, leads to flexible representations for pair-wise images.",26730
issue in is Intelligent important Transportation challenging an and recognition scene (ITS). Traffic Systems,Traffic scene recognition is an important and challenging issue in Intelligent Transportation Systems (ITS).,26731
"information, to To CNN the pyramids are exploited spatial features. encode spatial catch","To catch the spatial information, spatial pyramids are exploited to encode CNN features.",26732
full-pol SAR with image well the full-pol reconstructed image. true The agrees,The reconstructed full-pol SAR image agrees well with the true full-pol image.,26733
compact-pol to be reconstruction easily data from of full-pol extended framework data. can This,This framework can be easily extended to reconstruction of full-pol data from compact-pol data.,26734
retrieval for information mining. image sensing Remote is significant great of geological (RS),Remote sensing (RS) image retrieval is of great significant for geological information mining.,26735
"factor, remarkable results we By publicly on available datasets. acquire each optimizing retrieval HRRS","By optimizing each factor, we acquire remarkable retrieval results on publicly available HRRS datasets.",26736
"draw analysis. experimental according in conclusions our and explain the Finally, to phenomenon detail we","Finally, we explain the experimental phenomenon in detail and draw conclusions according to our analysis.",26737
attention the is modeling data different given preprocessing of to in-put Special approaches. and,Special attention is given to preprocessing of in-put data and the different modeling approaches.,26738
literature approach phenotyping. natural to promising language for electronic review a The processing confirms be,The literature review confirms natural language processing to be a promising approach for electronic phenotyping.,26739
"accessibility for texts medical challenge. natural remain lack However, process standards of a and language","However, accessibility and lack of natural language process standards for medical texts remain a challenge.",26740
neural point We two present system powered deep tracking networks. convolutional by a,We present a point tracking system powered by two deep convolutional neural networks.,26741
present spur in Dataset research Here these NREC Agricultural to environments. we the Person-Detection,Here we present the NREC Agricultural Person-Detection Dataset to spur research in these environments.,26742
were known. These two largest subsume that classes classes the previously lifted,These two classes subsume the largest lifted classes that were previously known.,26743
completely this of operator geometry. encodes it property volumetric remarkable that A is,A remarkable property of this operator is that it completely encodes volumetric geometry.,26744
vision. Foreground video classic computer is a segmentation in in sequences topic,Foreground segmentation in video sequences is a classic topic in computer vision.,26745
"jointly reconstruction the CNN, two-stage In are the segmentation loss optimized. loss and","In the two-stage CNN, the reconstruction loss and segmentation loss are jointly optimized.",26746
The in end-to-end simultaneously objects and background way. images foreground an are output,The background images and foreground objects are output simultaneously in an end-to-end way.,26747
people. Falls serious costly and for elderly are,Falls are serious and costly for elderly people.,26748
The respond takes and crucial. a treat person time fallen is to it,The time it takes to respond and treat a fallen person is crucial.,26749
"paper this for people non-invasive system fallen With a detection. new we present ,","With this paper we present a new , non-invasive system for fallen people detection.",26750
passively environment. sensing uses camera Our data the only for stereo approach,Our approach uses only stereo camera data for passively sensing the environment.,26751
"and extensive almost our shows Based miss-classification. no high evaluations, on systems our accuracy","Based on our extensive evaluations, our systems shows high accuracy and almost no miss-classification.",26752
"reproduce the community. to available the To publicly implementation results, scientific our is","To reproduce our results, the implementation is publicly available to the scientific community.",26753
Unsupervised algorithms able detect such and to identify objects. multi-instance are object discovery,Unsupervised multi-instance object discovery algorithms are able to detect and identify such objects.,26754
"results regions traditional compared evaluate, This algorithms. proposal in to to fewer region","This results in fewer regions to evaluate, compared to traditional region proposal algorithms.",26755
network the focuses paper neural training (CNN). on convolutional regularizing of This the,This paper focuses on regularizing the training of the convolutional neural network (CNN).,26756
are structure detect any to configurable preferred They (e.g. line,They are configurable to detect any preferred line structure (e.g.,26757
"usable for automatic representation learning. data ), so","), so usable for automatic data representation learning.",26758
increasingly preferences procedures-based all used particular almost applications. In in are matching,In particular preferences are increasingly used in almost all matching procedures-based applications.,26759
Each of allocates hundreds applicants first year year APB universities. thousands to,Each year APB allocates hundreds of thousands first year applicants to universities.,26760
by to matching is automatically preferences university done seats. This applicants,This is done automatically by matching applicants preferences to university seats.,26761
"problem. In provide a solution we this this to work,","In this work, we provide a solution to this problem.",26762
effectiveness the of demonstrate proposed on benchmark results Experimental datasets the several approach.,Experimental results on several benchmark datasets demonstrate the effectiveness of the proposed approach.,26763
Objective: system. mostly Mammalian composed of Background compact and is Haversian bone,Background and Objective: Mammalian compact bone is composed mostly of Haversian system.,26764
"and Nile geese. sparrows, Materials monitors, Methods: Femora collected were ducks geckos, and from","Materials and Methods: Femora were collected from geckos, Nile monitors, sparrows, ducks and geese.",26765
"Nile geckos vascular. while of was The mostly avascular, bone highly compact monitor was bone","The compact bone of geckos was mostly avascular, while Nile monitor bone was highly vascular.",26766
side showed and vascular one bone Sparrow side the other avascular.,Sparrow bone showed one side vascular and the other side avascular.,26767
"the paradox of The the on ""birthday estimation generalization problem"". based is","The estimation is based on the generalization of the ""birthday paradox problem"".",26768
the homonym. large population majority one least of at has The,The large majority of the population has at least one homonym.,26769
issue in diabetes a is Data performance-degrading overfitting prognosis.,Data overfitting is a performance-degrading issue in diabetes prognosis.,26770
discrete consider We non-negative conditional tests for exponential families.,We consider conditional tests for non-negative discrete exponential families.,26771
MCMC on by The the first sampling based Sturmfels. described algorithm is,The first algorithm is based on the MCMC sampling described by Sturmfels.,26772
between and We link a sampling. establish standard thus permutation algebraic-statistics-based,We thus establish a link between standard permutation and algebraic-statistics-based sampling.,26773
complex can analytically. rarely system The be assessed reliability a industrial of,The reliability of a complex industrial system can rarely be assessed analytically.,26774
of for of a an dynamical this propose class We method adaptation multi-component systems.,We propose an adaptation of this method for a class of multi-component dynamical systems.,26775
and Such in nuclear industry. hydraulic common are systems,Such systems are common in hydraulic and nuclear industry.,26776
"recoloring, edge cut-copy-paste, masking, tonal enhancement. summarization, We and adjustments, demonstrate selection palette","We demonstrate tonal adjustments, selection masking, cut-copy-paste, recoloring, palette summarization, and edge enhancement.",26777
that a We (RankIQA). assessment learns from (NR-IQA) rankings quality propose image no-reference approach,We propose a no-reference image quality assessment (NR-IQA) approach that learns from rankings (RankIQA).,26778
image These human without sets ranked can laborious automatically generated be labeling.,These ranked image sets can be automatically generated without laborious human labeling.,26779
"In present single-shot detector. for object multibox shorthand context-aware paper, CSSD--a we this","In this paper, we present CSSD--a shorthand for context-aware single-shot multibox object detector.",26780
"layers is on of modeling additional SSD, contexts. top built with CSSD multi-scale","CSSD is built on top of SSD, with additional layers modeling multi-scale contexts.",26781
the improves experimental significantly show detection that accuracy. multi-scale The modeling the context results,The experimental results show that the multi-scale context modeling significantly improves the detection accuracy.,26782
"for considered proposed sets Finally, methodology is our censorship. in presence the also of data","Finally, our proposed methodology is also considered for data sets in the presence of censorship.",26783
presented in is recognition languages. of This handwritten for the numerals model Indic,This model is presented for the recognition of handwritten numerals in Indic languages.,26784
for inference probabilistic a in inference. commonly model (PGM) efficient graphical exploit symmetries Lifted algorithms,Lifted inference algorithms commonly exploit symmetries in a probabilistic graphical model (PGM) for efficient inference.,26785
a symmetries These many domain. in miss algorithms,These algorithms miss many symmetries in a domain.,26786
multi-valued different that domain can have also symmetries between find variables methods cardinalities. Our,Our methods can also find symmetries between multi-valued variables that have different domain cardinalities.,26787
clinician versatile These algorithms are complex to enough answer queries. not,These algorithms are not versatile enough to answer complex clinician queries.,26788
construct in the WSIs results information. of from which Current approaches contextual patches smaller loss,Current approaches construct smaller patches from WSIs which results in the loss of contextual information.,26789
recognition of food learning ingredients the tackle multi-label problem We problem. as a,We tackle the problem of food ingredients recognition as a multi-label learning problem.,26790
public new We make this two suitable datasets purpose. for,We make public two new datasets suitable for this purpose.,26791
Deterministic demonstrations. build Deep to upon (DDPG) Policy the algorithm Gradient We use,We build upon the Deep Deterministic Policy Gradient (DDPG) algorithm to use demonstrations.,26792
Normalized Policy and Guided Search Function). Advantage,Guided Policy Search and Normalized Advantage Function).,26793
force-controlled human by Demonstrations by are collected robot kinesthetically a demonstrator. a,Demonstrations are collected by a robot kinesthetically force-controlled by a human demonstrator.,26794
"the caption UIR the bounding of explained with will be best box. the Then, match","Then, the UIR will be explained with the caption of the best match bounding box.",26795
the solutions accuracy approach segmentation state-of-the-art computing and efficiency. terms outperforms in Our of,Our approach outperforms the state-of-the-art solutions in terms of segmentation accuracy and computing efficiency.,26796
"to multiple analyze simultaneously. has been matrices NMF extended Recently,","Recently, NMF has been extended to analyze multiple matrices simultaneously.",26797
"a unified still lacking. is framework However,","However, a unified framework is still lacking.",26798
"update framework. we four to this present algorithms Next, solve","Next, we present four update algorithms to solve this framework.",26799
demerits and explored. these merits are of The systematically algorithms,The merits and demerits of these algorithms are systematically explored.,26800
considered. for attention natural be the to basis is This,This is the natural basis for attention to be considered.,26801
randomly Networks theoretical of models the popular connected most among neuroscience. neurons in are,Networks of randomly connected neurons are among the most popular models in theoretical neuroscience.,26802
"the timescale intrinsic fluctuations. determine to the In regime, chaotic simulations we the performed of","In the chaotic regime, we performed simulations to determine the timescale of the intrinsic fluctuations.",26803
a here proof-of-concept This visually shown dose-saving in new is and study. method cost- qualitative,This cost- and dose-saving new method is shown here visually in a qualitative proof-of-concept study.,26804
during spontaneous Recognizing facial facial is a (AUs) problem. units challenging action displays,Recognizing facial action units (AUs) during spontaneous facial displays is a challenging problem.,26805
"as the a becomes from training especially Such suffers expensive network cost, deeper. training process","Such a training process suffers from expensive training cost, especially as the network becomes deeper.",26806
"modest the Temporal impact on despite lobe caused. network surgery global had disruption efficiency, a","Temporal lobe surgery had a modest impact on global network efficiency, despite the disruption caused.",26807
types different be sources. can adopted for rules These of,These rules can be adopted for different types of sources.,26808
"these is the However, number of when sources are rules large. applicable most of not","However, most of these rules are not applicable when the number of sources is large.",26809
an as the of empty set on The kept will be conflict. the indicator mass,The mass on the empty set will be kept as an indicator of the conflict.,26810
"elicit be the opinion can the to experts. used major among Moreover, it","Moreover, it can be used to elicit the major opinion among the experts.",26811
many OMP the and OLS of application This limits situations. in practical,This limits the application of OMP and OLS in many practical situations.,26812
mitotic and figures task in and Identification a is of diagnostic cells histopathology. standard counting,Identification and counting of cells and mitotic figures is a standard task in diagnostic histopathology.,26813
update of to sequential the inverse We the shrinkage rules estimator approximate matrix. covariance derived,We derived sequential update rules to approximate the inverse shrinkage estimator of the covariance matrix.,26814
learning sequential updates. paves methods that way the improved large-scale approach machine for involve The,The approach paves the way for improved large-scale machine learning methods that involve sequential updates.,26815
word confirms the of handwritten recognition the suitability to HS algorithm problem. holistic This,This confirms the suitability of HS algorithm to the holistic handwritten word recognition problem.,26816
axioms. The logic and ABox equipped is TBox with both,The logic is equipped with both TBox and ABox axioms.,26817
"However, much to remains it how task. contributes context the unclear each","However, it remains unclear how much each context contributes to the task.",26818
computer This a common in is vision. problem,This is a common problem in computer vision.,26819
multiview is unlabeled the $\text{Sym}_m(V_A)$ variety Further studied.,Further the unlabeled multiview variety $\text{Sym}_m(V_A)$ is studied.,26820
Population population typically a fixed selection models size a consider and coefficient. unique genetics,Population genetics models typically consider a fixed population size and a unique selection coefficient.,26821
"Alzheimer's. early such is several diseases, impairment cognitive the stage neurodegenerative Mild of as","Mild cognitive impairment is the early stage of several neurodegenerative diseases, such as Alzheimer's.",26822
"crucial distinction introspection discussed and First, between the reflexion. we","First, we discussed the crucial distinction between introspection and reflexion.",26823
image revolutionize the CNN-s analysis. to medical potential Deep have,Deep CNN-s have the potential to revolutionize medical image analysis.,26824
"However, the slow, is even complete. are learning when rather resulting algorithm the data","However, the resulting learning algorithm is rather slow, even when the data are complete.",26825
on RDF based is is This model fluent of notion graphs. represented in the that,This model is based on the notion of fluent that is represented in RDF graphs.,26826
Image in and an image video editing. role plays important matting,Image matting plays an important role in image and video editing.,26827
"is the ill-posed. image However, formulation inherently of matting","However, the formulation of image matting is inherently ill-posed.",26828
"paper, for a propose deep this real-time matting In automatic we mobile devices. approach","In this paper, we propose a real-time automatic deep matting approach for mobile devices.",26829
state-of-the-art results solvers. achieves with comparable matting The the experiments the proposed approach that show,The experiments show that the proposed approach achieves comparable results with the state-of-the-art matting solvers.,26830
require not target Our novel transfer emotion image image. algorithm a does user-specified proposed,Our proposed novel image emotion transfer algorithm does not require a user-specified target image.,26831
of previous handle the can method than photographs methods. Our more diverse set,Our method can handle more diverse set of photographs than the previous methods.,26832
effectiveness of We the user our a method. showing conducted proposed detailed study,We conducted a detailed user study showing the effectiveness of our proposed method.,26833
"character for In handwritten recognition. methodology paper, a present off-line we this","In this paper, we present a methodology for off-line handwritten character recognition.",26834
are $k$-means feature vectors that a step employed a classification algorithm. then uses These in,These feature vectors are then employed in a classification step that uses a $k$-means algorithm.,26835
database to experiments performed evaluate We proposal. using the our NIST,We performed experiments using the NIST database to evaluate our proposal.,26836
"two resulting maps are feature fed branched Then, the into sub-networks.","Then, the resulting feature maps are fed into two branched sub-networks.",26837
"proposed emotion state-of-the-art TSRG Compared promising recognition with cross-database results. achieves more the methods, recent","Compared with recent state-of-the-art cross-database emotion recognition methods, the proposed TSRG achieves more promising results.",26838
"it to As parameters, compute million of has consumption. huge memory it results to","As it has to compute million of parameters, it results to huge memory consumption.",26839
"conducting the tends and increase finer features complexity. Moreover, extracting to supervised training","Moreover, extracting finer features and conducting supervised training tends to increase the complexity.",26840
detection potential the motion of its impairments. a in Indicating range application age-related of,Indicating its potential application in the detection of a range of age-related motion impairments.,26841
learning We camera based relocalization.  a for deep  propose approach new,  We propose a new deep learning based approach for camera relocalization.,26842
"fundamental, We create biometric biometric study general, databases synthetic principles. to","We create synthetic biometric databases to study general, fundamental, biometric principles.",26843
The from real eye-movement related data for an check database. this biometric validity used was,The real data used for this validity check was from an eye-movement related biometric database.,26844
index coefficient (ICC). the correlation with We temporal persistence intraclass,We index temporal persistence with the intraclass correlation coefficient (ICC).,26845
"also outliers. as data often anomalous known Real contain cases,","Real data often contain anomalous cases, also known as outliers.",26846
resulting spoil information. analysis contain also they but the These may valuable may,These may spoil the resulting analysis but they may also contain valuable information.,26847
"detect such case, to ability is anomalies In either essential. the","In either case, the ability to detect such anomalies is essential.",26848
an We methods present graphical robust the of and resulting detection several tools. outlier overview,We present an overview of several robust methods and the resulting graphical outlier detection tools.,26849
cellwise is introduced. Also new outliers the of topic challenging,Also the challenging new topic of cellwise outliers is introduced.,26850
biochemical new finite reactions. proposes method efficiently computing sensitivities for of difference paper a This,This paper proposes a new finite difference method for efficiently computing sensitivities of biochemical reactions.,26851
reserved simulation of by The rejection-based applying the exactness mechanism. the is,The exactness of the simulation is reserved by applying the rejection-based mechanism.,26852
Our approach rates to allows sensitivities of computing many reaction simultaneously.,Our approach allows to computing sensitivities of many reaction rates simultaneously.,26853
large to applied This especially feature networks. is reaction when useful,This feature is especially useful when applied to large reaction networks.,26854
and applicability its method reaction efficiency. prove We to models our benchmark on biological,We benchmark our method on biological reaction models to prove its applicability and efficiency.,26855
for novel visual between of a objects. relations approach paper pairs This modeling introduces,This paper introduces a novel approach for modeling visual relations between pairs of objects.,26856
are contributions this threefold. of The paper,The contributions of this paper are threefold.,26857
"labels model discriminative clustering weakly-supervised we Second, from learn a image-level relations only. propose to","Second, we propose a weakly-supervised discriminative clustering model to learn relations from image-level labels only.",26858
"of methods to many Owing multi-view the effectiveness are on clustering it. spectral based clustering,","Owing to the effectiveness of spectral clustering, many multi-view clustering methods are based on it.",26859
"novel In a spectral multi-view for propose large-scale we data. work, this clustering method","In this work, we propose a novel multi-view spectral clustering method for large-scale data.",26860
algorithm The the with of data. proposed scales number given linearly,The proposed algorithm scales linearly with the number of given data.,26861
work step close gap. one to this This is,This work is one step to close this gap.,26862
overview A construction for provided. CNN existing is techniques over and comprehensive analysis topology,A comprehensive overview over existing techniques for CNN analysis and topology construction is provided.,26863
classification to novel visualize A with developed. way errors confusion was matrices,A novel way to visualize classification errors with confusion matrices was developed.,26864
"hierarchical described Based and evaluated. on method, are this classifiers","Based on this method, hierarchical classifiers are described and evaluated.",26865
perpetual targets. detection occlusions sensitive are among highly the to methods People the,People detection methods are highly sensitive to the perpetual occlusions among the targets.,26866
walking scenarios of seven-static-camera people. realistic challenging The set-up captures and,The seven-static-camera set-up captures realistic and challenging scenarios of walking people.,26867
"However, low synthesis methods in resolution synthesizing have problems image existing PET images. the","However, existing image synthesis methods have problems in synthesizing the low resolution PET images.",26868
installed in grouping CNNs applicable mobile logarithmic shallow filter in application. is The a proposed,The proposed logarithmic filter grouping is installed in shallow CNNs applicable in a mobile application.,26869
were with CNNs the for shallow tasks. performed Experiments classification,Experiments were performed with the shallow CNNs for classification tasks.,26870
role image (PSF) an in essential Point spread reconstruction. plays function,Point spread function (PSF) plays an essential role in image reconstruction.,26871
spatial containing each trend regions Images intensity of are particular in that a interest. have,Images containing regions that each have a spatial trend in intensity are of particular interest.,26872
"of Finally, a eigenvalue kernalized the KPLS version is solving by achieved the problem.","Finally, the KPLS is achieved by solving a kernalized version of the eigenvalue problem.",26873
rich that object be from information. can of We a believe detectors such learned source,We believe that object detectors can be learned from such a rich source of information.,26874
"patches images. single atoms extracted is learned person of First, using dictionary sparse a from","First, a dictionary of sparse atoms is learned using patches extracted from single person images.",26875
a also foreground flow. simultaneously using We derived calculate estimation depth probability from optical map,We also simultaneously calculate a foreground probability map using depth estimation derived from optical flow.,26876
compared blending results Our improvement great existing show based to handling methods. occlusions in,Our results show great improvement in handling occlusions compared to existing blending based methods.,26877
see our consciousness how intuition arises. is and We how shall understand shall misleading,We shall see how misleading our intuition is and shall understand how consciousness arises.,26878
"them. visual object, relating subject, Understanding predicate identifying a the relationships the and involves","Understanding visual relationships involves identifying the subject, the object, and a predicate relating them.",26879
"statistics to knowledge this, overcome regularize visual linguistic model of To learning. we use","To overcome this, we use knowledge of linguistic statistics to regularize visual model learning.",26880
"Then, distill we into model better to deep knowledge achieve generalization. the a","Then, we distill the knowledge into a deep model to achieve better generalization.",26881
to spatio-temporal of the of goal in video. is this paper location actions determine The,The goal of this paper is to determine the spatio-temporal location of actions in video.,26882
learning weakly-supervised to number of Discriminative has a been tasks. successfully clustering applied,Discriminative clustering has been successfully applied to a number of weakly-supervised learning tasks.,26883
"scalability. drawback limited its One however, discriminative of clustering, is","One drawback of discriminative clustering, however, is its limited scalability.",26884
accounts for This nature as of system a the clade whole. fat-tailed the non-equilibrium between,This between clade non-equilibrium accounts for the fat-tailed nature of the system as a whole.,26885
learning from synthesis. techniques model program The combines deep and,The model combines techniques from deep learning and program synthesis.,26886
network a neural plausible an that convolutional drawing explain proposes image. that We primitives learn,We learn a convolutional neural network that proposes plausible drawing primitives that explain an image.,26887
"programs conditionals. kinds bindings, These or like have of iterative constructs variable simple loops,","These programs have constructs like variable bindings, iterative loops, or simple kinds of conditionals.",26888
joint improved This object generation distribution. the enables from shape,This enables improved generation from the joint object shape distribution.,26889
in quantitative to . notable improvements baselines We comparison achieve existing,We achieve notable quantitative improvements in comparison to existing baselines .,26890
"between the distance We mean consider vectors of two domains. also information, first-order minimize and","We also consider first-order information, and minimize the distance of mean vectors between two domains.",26891
to visual generation concepts. focus represent of symbols We on computational,We focus on computational generation of visual symbols to represent concepts.,26892
"a these causation Incorporating features, definition three given. is actual of","Incorporating these three features, a definition of actual causation is given.",26893
"suggested small. $n$, Past the of studies antigenic be dimensionality may space, have","Past studies have suggested the dimensionality of antigenic space, $n$, may be small.",26894
analysis multidimensional (MDS). principal (PCA) and scaling component including,including principal component analysis (PCA) and multidimensional scaling (MDS).,26895
to fluoroscopic our femur. the segment on evaluate We images strategy X-ray,We evaluate our strategy on fluoroscopic X-ray images to segment the femur.,26896
We five for experiment datasets retrieval. image on standard public,We experiment on five public standard datasets for image retrieval.,26897
dimension related reduction learning The significantly IME layer proposed outperforms and methods methods. manifold,The proposed IME layer significantly outperforms related dimension reduction methods and manifold learning methods.,26898
both results Experimental verify of effectiveness method. data and proposed synthetic real-world the on the,Experimental results on both synthetic and real-world data verify the effectiveness of the proposed method.,26899
can unified be trained The system whole end-to-end a framework. CNN in,The whole system can be trained end-to-end in a unified CNN framework.,26900
source RSA github.com/sciencefans/RSA-for-object-detection. code of The at available is,The source code of RSA is available at github.com/sciencefans/RSA-for-object-detection.,26901
thus charge over mean-field Overbeek-Voorn and and RPA patterns accounts Flory-Huggins theories. has for advantages,RPA accounts for charge patterns and thus has advantages over Flory-Huggins and Overbeek-Voorn mean-field theories.,26902
supervision transfer requires dataset. Learning visual to attributes,Learning to transfer visual attributes requires supervision dataset.,26903
"a applications, capturing This limits is them because often largely their difficult task.","This largely limits their applications, because capturing them is often a difficult task.",26904
"address To issue, propose visual to learn attribute. an the to unsupervised method transfer we","To address the issue, we propose an unsupervised method to learn to transfer visual attribute.",26905
any the The images. transfer function method learn proposed without corresponding can,The proposed method can learn the transfer function without any corresponding images.,26906
"paper, using approach neural deep to propose problem we convolutional this In networks. this","In this paper, we propose to approach this problem using deep convolutional neural networks.",26907
to professional by improve workflow of best the pre-selecting photographers the photos.,to improve the workflow of professional photographers by pre-selecting the best photos.,26908
often sensing Remote images from cover. suffer cloud,Remote sensing images often suffer from cloud cover.,26909
applications required removal remote is Cloud sensing many in of images.,Cloud removal is required in many applications of remote sensing images.,26910
clouds. cope to are with methods Multitemporal-based popular thick and effective,Multitemporal-based methods are popular and effective to cope with thick clouds.,26911
multitemporal-based paper summarization This and comparation methods. of a to existing the contributes experimental,This paper contributes to a summarization and experimental comparation of the existing multitemporal-based methods.,26912
constrained problem problem. We optimization a discretely this model as,We model this problem as a discretely constrained optimization problem.,26913
"factor compressibility defined. trade-off the a to the complexity-performance is control controlling Moreover, as threshold","Moreover, a controlling factor as the compressibility threshold to control the complexity-performance trade-off is defined.",26914
of method. analysis and theoretical preliminary interpretations the provide several We,We provide preliminary theoretical analysis and several interpretations of the method.,26915
of classification localization We human aim zero-shot in actions for and video.,We aim for zero-shot localization and classification of human actions in video.,26916
support and classification contemporary Action proposal. experiments action localization four datasets on video our,Action localization and classification experiments on four contemporary action video datasets support our proposal.,26917
"we address small effective. learning challenge, data the to To showed representation be transfer","To address the small data challenge, we showed transfer representation learning to be effective.",26918
"for an introduce technique embedding We information text documents. FontCode,","We introduce FontCode, an information embedding technique for text documents.",26919
We a Soft model these novel overcome CCA deep to CCA problems. propose,We propose a novel deep CCA model Soft CCA to overcome these problems.,26920
budget. we model advice study action the a under learning transfer of In article this,In this article we study the transfer learning model of action advice under a budget.,26921
The mean. statistic relates the CV corresponding variance to,The CV statistic relates variance to the corresponding mean.,26922
"budget. advice a article the studies learning under Second, distributing for policy","Second, the article studies policy learning for distributing advice under a budget.",26923
(Kou Equi-Energy al. et on sampler Based the,Based on the Equi-Energy sampler (Kou et al.,26924
"analyze is data. to huge work However, hard it extremely such manually","However, it is extremely hard work to manually analyze such huge data.",26925
"the ImageNet), network over-parameterized. likely be fine-tuned is to","ImageNet), the fine-tuned network is likely to be over-parameterized.",26926
also their to demonstrate We of our framework complexity computing analyzed the efficiency.,We also analyzed the computing complexity of our framework to demonstrate their efficiency.,26927
functional uses energy flux-based data-fit measures term. a an combining proposed method regularization and The,The proposed method uses an energy functional combining flux-based data-fit measures and a regularization term.,26928
continuous surface the convex relaxation global scheme A of geometric minima the functional. assures,A continuous convex relaxation scheme assures the global minima of the geometric surface functional.,26929
Vision by of day Computer day specific growing applications. user terms is in,Computer Vision is growing day by day in terms of user specific applications.,26930
is segmenting such image. first of step application The an any,The first step of any such application is segmenting an image.,26931
based on of one defined second class is function. the a measures The similarity distance,The second one is the class of similarity measures defined based on a distance function.,26932
of present pedestrian We for method set on detectors training unannotated images. an new a,We present a new method for training pedestrian detectors on an unannotated set of images.,26933
end-to-end The an training performed fashion. in joint is,The joint training is performed in an end-to-end fashion.,26934
to The of overcomes data regularizer. problem module the drop-out over-fitting scarcity due using arising,The module overcomes the over-fitting problem arising due to scarcity of data using drop-out regularizer.,26935
used to the positives edges. operators small for around post-processing false are Morphological remove,Morphological operators are used for post-processing to remove small false positives around the edges.,26936
number data. on labeled CNN-based of of success methods The a large relies,The success of CNN-based methods relies on a large number of labeled data.,26937
method. novel generation face paper a data This presents,This paper presents a novel face data generation method.,26938
in RGB space. lines Single form have straight colors materials which,Single materials have colors which form straight lines in RGB space.,26939
"an model. forming we the optical that imaging offset via reason for analyze First,","First, we analyze the reason for forming that offset via an optical imaging model.",26940
"the offset. we detect a Second, remove and effective and simple to present way","Second, we present a simple and effective way to detect and remove the offset.",26941
"of of is RGB using interference processing shadows. free ORGB instead image the from Besides,","Besides, image processing using ORGB instead of RGB is free from the interference of shadows.",26942
"untrimmed this problem action temporally in of detection videos. spatio-temporal we address work, In the","In this work, we address the problem of spatio-temporal action detection in temporally untrimmed videos.",26943
results Our from $\pi$-machine the show interpretable efficiently programs traces. can individual data that induce,Our results show that the $\pi$-machine can efficiently induce interpretable programs from individual data traces.,26944
detail. techniques in and explained neural Training are successful architectures network,Training techniques and successful neural network architectures are explained in detail.,26945
errors pixel-wise These the and often ignore are based similarities. structural on measures,These measures are based on pixel-wise errors and often ignore the structural similarities.,26946
representation Tensor space-efficient for a train provides decomposition higher-order (TT) tensors.,Tensor train (TT) decomposition provides a space-efficient representation for higher-order tensors.,26947
"In paper, this we the limitations. address","In this paper, we address the limitations.",26948
temporal is problem. Visual a intrinsically tracking,Visual tracking is intrinsically a temporal problem.,26949
demonstrated visual Correlation for performance have Filters (DCF) high-speed Discriminative object generic excellent tracking.,Discriminative Correlation Filters (DCF) have demonstrated excellent performance for high-speed generic visual object tracking.,26950
Riemannian Stiefel is space. not but a space manifold homogeneous The a symmetric,The Stiefel manifold is a Riemannian homogeneous space but not a symmetric space.,26951
We also weak prove recursive the consistency this estimator. FM of,We also prove the weak consistency of this recursive FM estimator.,26952
complicated powerful flexible in planning learning domains. means Hierarchical offer of a methods behavior reinforcement,Hierarchical reinforcement learning methods offer a powerful means of planning flexible behavior in complicated domains.,26953
"a domain substantial a However, subtasks learning decomposition hierarchical appropriate into of an remains challenge.","However, learning an appropriate hierarchical decomposition of a domain into subtasks remains a substantial challenge.",26954
to as number learning patterns increases of target attractors limit proceeds. The cycle movement corresponding,The number of limit cycle attractors corresponding to target movement patterns increases as learning proceeds.,26955
objects to many The images is edit content desirable materials creators. ability of by in,The ability to edit materials of objects in images is desirable by many content creators.,26956
"Specifically, network given the properties, i.e. predicts intrinsic single image, a first","Specifically, given a single image, the network first predicts intrinsic properties, i.e.",26957
"which a illumination, rendering to are and layer. provided material, then shape,","shape, illumination, and material, which are then provided to a rendering layer.",26958
a And-Or We (ST-AOG). event graph by spatial-temporal the represent,We represent the event by a spatial-temporal And-Or graph (ST-AOG).,26959
using predicted parsing the algorithm. Future temporal are Earley sub-activities and grammar,Future sub-activities are predicted using the temporal grammar and Earley parsing algorithm.,26960
"labels The inferred affordance corresponding object, then and are action, accordingly.","The corresponding action, object, and affordance labels are then inferred accordingly.",26961
element Disease biomedical crucial is a classification research. of,Disease classification is a crucial element of biomedical research.,26962
state. about discloses person's A important face their information affective,A person's face discloses important information about their affective state.,26963
and discarded are as affective Facial occlusions in recognition of noise often treated states.,Facial occlusions are often treated as noise and discarded in recognition of affective states.,26964
"since similar very monocular SLAM approaches. also tackled However, been has with","However, monocular SLAM has since also been tackled with very similar approaches.",26965
"small-scale However, this is memory-intensive representation applicability and reconstructions. to limits","However, this representation is memory-intensive and limits applicability to small-scale reconstructions.",26966
this. have been explored avenues to Several overcome,Several avenues have been explored to overcome this.,26967
detection Pose the is one recognition steps the for human of of fundamental actions.,Pose detection is one of the fundamental steps for the recognition of human actions.,26968
Quick medical accurate crucial for a treatment is diagnosis and the disease. of successful,Quick and accurate medical diagnosis is crucial for the successful treatment of a disease.,26969
"our three has compelling From DenseVoxNet advantages. perspective, the learning","From the learning perspective, our DenseVoxNet has three compelling advantages.",26970
is autonomous of location indispensable a the for Knowledge vehicle driving. about,Knowledge about the location of a vehicle is indispensable for autonomous driving.,26971
prior the The of robustness and quality of localisation. that determine success,The quality and robustness of that prior determine the success of localisation.,26972
the especially We that of outperforms the state low-structure art. environments show Momo in,We show that especially in low-structure environments Momo outperforms the state of the art.,26973
research relations. functionality the objects its on geometrical focuses between relationship Recent on based and,Recent research focuses on the relationship between objects based on its functionality and geometrical relations.,26974
"we framework propose compression move and forward step Here, on based novel a a CNNs.","Here, we move forward a step and propose a novel compression framework based on CNNs.",26975
proposed makes existing also compression standards. Such design coding framework with image compatible a the,Such a design also makes the proposed compression framework compatible with existing image coding standards.,26976
normal t-tests by et.al. transformations samples Inverse the applied Derrick partially overlapping to,Inverse normal transformations applied to the partially overlapping samples t-tests by Derrick et.al.,26977
to are solutions transformation rank comparable The non-parametric when power inverse based normal solutions. using,The power when using inverse normal transformation solutions are comparable to rank based non-parametric solutions.,26978
"This paper tackles different characteristics. problem, a but with similar","This paper tackles a similar problem, but with different characteristics.",26979
"are a metrics shown perceptual the of compression. As JPEG utility, outperform demonstration their to","As a demonstration of their utility, the perceptual metrics are shown to outperform JPEG compression.",26980
"Unlike the learning. machine-learning no uses approaches, outperformance no and statistics,","Unlike machine-learning approaches, the outperformance uses no statistics, and no learning.",26981
set principles The constitute fundamental perceptual of underlying findings similarity. a,The findings constitute a set of fundamental principles underlying perceptual similarity.,26982
"have Since scale in the earliest observed electroencephalography brain. experiments, the mammalian large been oscillations","Since the earliest electroencephalography experiments, large scale oscillations have been observed in the mammalian brain.",26983
handwritten Indic novel word This towards recognition information. paper a zone-wise approach presents using,This paper presents a novel approach towards Indic handwritten word recognition using zone-wise information.,26984
"similar Gurumukhi, scripts). and other Devanagari, Bangla,","Devanagari, Bangla, Gurumukhi, and other similar scripts).",26985
HMM. components zone are middle touching using The mostly recognized characters in are where,The components in middle zone where characters are mostly touching are recognized using HMM.,26986
from studied two defocus Depth sensing stereo are matching depth (DfD) and most passive schemes.,Depth from defocus (DfD) and stereo matching are two most studied passive depth sensing schemes.,26987
possible networks? Is convolutional version noisy it neural its recover an from image using to,Is it possible to recover an image from its noisy version using convolutional neural networks?,26988
state-of-the-art proposed shows when model approaches. The performance denoising the compared to competitive,The proposed model shows competitive denoising performance when compared to the state-of-the-art approaches.,26989
deployed usage. commercial for real-time be can The method,The method can be deployed for real-time commercial usage.,26990
with correlating scores subjective metric The ratings. human are metrics evaluated the by,The metrics are evaluated by correlating metric scores with the human subjective ratings.,26991
existing metric popular all that correlation evaluation the show metrics. proposed results The outperforms other,The correlation results show that the proposed evaluation metric outperforms all other popular existing metrics.,26992
our obtained images. results CT have experiments with Promising clinical been in,Promising results have been obtained in our experiments with clinical CT images.,26993
We different cascades applications for sparse objects. with two segmentation CNN propose,We propose two different CNN cascades for segmentation applications with sparse objects.,26994
Our atrous convolutional the efficiently. is layers model field-of-view on to based filters of increase,Our model is based on atrous convolutional layers to increase the field-of-view of filters efficiently.,26995
reinforcing this and target between embedding by directly data space. accomplish We associations source in,We accomplish this by reinforcing associations between source and target data directly in embedding space.,26996
"object In this we framework. a investigate detection paper, weakly-supervised","In this paper, we investigate a weakly-supervised object detection framework.",26997
object Most detectors. static focus images existing on frameworks learn using to,Most existing frameworks focus on using static images to learn object detectors.,26998
"videos domain generalize to shift. of fail because However, the often existing these to detectors","However, these detectors often fail to generalize to videos because of the existing domain shift.",26999
"investigate these boring we learning detectors Therefore, activities. videos of daily directly from","Therefore, we investigate learning these detectors directly from boring videos of daily activities.",27000
common of of individual each testing. psychological most way characteristics is The studying,The most common way of studying psychological characteristics of each individual is testing.,27001
"and accuracy Real-time scarce. tracking remain high however, algorithms,","Real-time and high accuracy tracking algorithms, however, remain scarce.",27002
for This method presents a visual-inertial odometry. paper novel,This paper presents a novel method for visual-inertial odometry.,27003
will a One hospital hospital acquired suffer twenty-five in patients a from infection. admitted to,One in twenty-five patients admitted to a hospital will suffer from a hospital acquired infection.,27004
"a tracking system for in hospitals. people's we In paper, non-intrusive activity this vision-based propose","In this paper, we propose a non-intrusive vision-based system for tracking people's activity in hospitals.",27005
We method measuring the hand for compliance. evaluate hygiene of our problem,We evaluate our method for the problem of measuring hand hygiene compliance.,27006
"outperforms observational covert solutions studies. as such techniques Empirically, our proximity-based and method existing in-person","Empirically, our method outperforms existing solutions such as proximity-based techniques and covert in-person observational studies.",27007
"fine to post-implantation a probe. allow Furthermore, actuation of and relocation was microdrive the developed","Furthermore, a microdrive was developed to allow fine actuation and post-implantation relocation of the probe.",27008
is prediction learning. unsupervised videos avenue frame in representation a for promising video Future,Future frame prediction in videos is a promising avenue for unsupervised video representation learning.,27009
"pixel focus on hallucinating values, in existing methods predictions. resulting blurry However, directly","However, existing methods focus on directly hallucinating pixel values, resulting in blurry predictions.",27010
extremely challenging nature. inherent image problem is its due to ill-posed an removal Single haze,Single image haze removal is an extremely challenging problem due to its inherent ill-posed nature.,27011
of a manual or new features tedious ones designing can The be selection task.,The manual selection of features or designing new ones can be a tedious task.,27012
scans. proportion A a clinical in patients of follow-up have scanned setting significant,A significant proportion of patients scanned in a clinical setting have follow-up scans.,27013
this task. of The performance assessed then on a classification pre-trained is grading network,The performance of this pre-trained network is then assessed on a grading classification task.,27014
"be low-level information existing However, (e.g. capable models can of the transferring only","However, existing models can only be capable of transferring the low-level information (e.g.",27015
"viewpoints cat$\rightarrow$dog. as bicycle, cow$\rightarrow$sheep, shapes), and motor$\rightarrow$ such","viewpoints and shapes), such as cow$\rightarrow$sheep, motor$\rightarrow$ bicycle, cat$\rightarrow$dog.",27016
characteristics sensing distinctive remote data. The system is in multi-temporal by incorporating designed,The system is designed by incorporating distinctive characteristics in multi-temporal remote sensing data.,27017
sensors depth super-resolution is RGB-D presented. approach for A novel,A novel depth super-resolution approach for RGB-D sensors is presented.,27018
"annoying during ""serration,"" flickering visual But such as it silhouette and presents playback. artifacts, the","But it presents annoying visual artifacts, such as flickering and silhouette ""serration,"" during the playback.",27019
existing problem through object tackle approaches Many mining. proposal this,Many existing approaches tackle this problem through object proposal mining.,27020
"because processed training and detector separately. region are slow object selection is often training Moreover,","Moreover, training is often slow because region selection and object detector training are processed separately.",27021
"be efficiently. model trained the can Therefore, end-to-end entire","Therefore, the entire model can be trained end-to-end efficiently.",27022
quantitatively. both results and We state-of-the-art qualitatively show,We show state-of-the-art results both qualitatively and quantitatively.,27023
invariance in image recognition. Human possesses strong vision,Human vision possesses strong invariance in image recognition.,27024
and objects. both It algorithms good scalability outperforms on stationary with existing dynamic,It outperforms existing algorithms with good scalability on both stationary and dynamic objects.,27025
is usually the similarity coefficient also images. to between inner-product correlation used (PCC) measure Pearson's,Pearson's inner-product correlation coefficient (PCC) is also usually used to measure the similarity between images.,27026
Low-rank modelling. reconstruction visual a method and (LRD) state-of-the-art is data for decomposition,Low-rank decomposition (LRD) is a state-of-the-art method for visual data reconstruction and modelling.,27027
"propose we that this new LRD. embeds To a end, manifold into priors framework","To this end, we propose a new framework that embeds manifold priors into LRD.",27028
capturing of Symmetric positive definite for visual second-order data. statistics are useful matrices (SPD),Symmetric positive definite (SPD) matrices are useful for capturing second-order statistics of visual data.,27029
"datasets, vision computer demonstrating state-of-the-art on performances. experiments eight extensive We showcase","We showcase extensive experiments on eight computer vision datasets, demonstrating state-of-the-art performances.",27030
treatment planning radiotherapy synthesis. MR-to-CT requires accurate MR-only,MR-only radiotherapy treatment planning requires accurate MR-to-CT synthesis.,27031
"could misalignment images However, paired between errors synthesized to images. lead in CT","However, misalignment between paired images could lead to errors in synthesized CT images.",27032
score Noteheads between the written interface music. and are the,Noteheads are the interface between the written score and music.,27033
We classification onsets the intention process as framework. only casting a prediction grasping,We process only the grasping onsets casting intention prediction as a classification framework.,27034
"accuracy suffer the from much they However, drop. non-negligible","However, they suffer much from the non-negligible accuracy drop.",27035
for stochastic low-bit learning (SQ) quantization DNNs. accurate This proposes algorithm paper the,This paper proposes the stochastic quantization (SQ) algorithm for learning accurate low-bit DNNs.,27036
The the to observation. motivation due following is,The motivation is due to the following observation.,27037
with Existing together all iteration. the in algorithms each low-bit elements/filters representation real-valued approximate training,Existing training algorithms approximate the real-valued elements/filters with low-bit representation all together in each iteration.,27038
and are iteration. gradients in The each portions full-precision corresponding updated with separately quantized,The quantized and full-precision portions are updated with corresponding gradients separately in each iteration.,27039
the until is SQ The network is whole quantized. increased gradually ratio,The SQ ratio is gradually increased until the whole network is quantized.,27040
parsing. face coarse-grained We fine-grained and the apply proposed model to,We apply the proposed model to coarse-grained and fine-grained face parsing.,27041
proposes multiview paper for framework This learning end-to-end stereopsis. an,This paper proposes an end-to-end learning framework for multiview stereopsis.,27042
evaluate on We the benchmark. DTU large-scale SurfaceNet,We evaluate SurfaceNet on the large-scale DTU benchmark.,27043
purpose this (RANSAC)-based proposed have RANdom for recent been methods in Consensus SAmple years. Different,Different RANdom SAmple Consensus (RANSAC)-based methods have been proposed for this purpose in recent years.,27044
"us in to allows the solve Problem. fundamental Stochastic turn, Finding Root This, (SRF)","This, in turn, allows us to solve the fundamental Stochastic Root Finding (SRF) Problem.",27045
perform biological roles interactions functions. important on (PPIs) Protein-protein,Protein-protein interactions (PPIs) perform important roles on biological functions.,27046
PPIs. Researches interactions can on mutants understand further of protein,Researches of mutants on protein interactions can further understand PPIs.,27047
"whole, On data further the new PPIs. improve database to the on provides study the","On the whole, the database provides new data to further improve the study on PPIs.",27048
powerful paradigm designers. modeling non-expert shape a tools for Assembly-based provide,Assembly-based tools provide a powerful modeling paradigm for non-expert shape designers.,27049
in the retrieval. Absence indexing database of parts for of the a challenge labels poses,Absence of labels poses a challenge in indexing the database of parts for the retrieval.,27050
"recent a has In of interest. of attracted lot the dictionary years, creation block-structured","In recent years, the creation of block-structured dictionary has attracted a lot of interest.",27051
update. and two such formation step involve Learning process: dictionaries block dictionary,Learning such dictionaries involve two step process: block formation and dictionary update.,27052
are important producing steps an effective Both these in dictionary.,Both these steps are important in producing an effective dictionary.,27053
ecological interaction structure analyses through is of understood networks. interactions commonly of The,The structure of ecological interactions is commonly understood through analyses of interaction networks.,27054
empirically of the These ecological networks. constructed issues accuracy may affect,These issues may affect the accuracy of empirically constructed ecological networks.,27055
designs with in may experiments. these different encountered We sampled sampling networks be field that,We sampled these networks with different sampling designs that may be encountered in field experiments.,27056
"was the of Overall, easiest sampling modules structure design. the nested to of detect, regardless","Overall, the structure of nested modules was the easiest to detect, regardless of sampling design.",27057
identifying conditional the Experimental of our outliers. multivariate in ability results demonstrate approach,Experimental results demonstrate the ability of our approach in identifying multivariate conditional outliers.,27058
multi-Siamese neural We convolutional a new design network. two-stream,We design a new two-stream multi-Siamese convolutional neural network.,27059
MR function. heart the in clinically Segmentation used cardiac of quantify to is cardiac cine,Segmentation of the heart in cardiac cine MR is clinically used to quantify cardiac function.,27060
and were The cross-validation. classification pipeline a stratified evaluated in four-fold segmentation,The segmentation and classification pipeline were evaluated in a four-fold stratified cross-validation.,27061
been using real performed synthetic data. Experiments and have,Experiments have been performed using synthetic and real data.,27062
several The compared has been to method state-of-the-art methods.,The method has been compared to several state-of-the-art methods.,27063
also propose to STD from Field Conditional saliency new Random compute features. We SpatioTemporal (STCRF),We also propose new SpatioTemporal Conditional Random Field (STCRF) to compute saliency from STD features.,27064
at maps saliency different The saliency map by levels. final is computed scale fusing,The final saliency map is computed by fusing saliency maps at different scale levels.,27065
large-scale from an is View-graph to essential (SfM) structure input pipelines. motion,View-graph is an essential input to large-scale structure from motion (SfM) pipelines.,27066
and is SfM Accuracy view-graph. large-scale efficiency dependent input the crucially on of,Accuracy and efficiency of large-scale SfM is crucially dependent on the input view-graph.,27067
edges lead can reconstruction. or or wrong inaccurate inferior Inconsistent to,Inconsistent or inaccurate edges can lead to inferior or wrong reconstruction.,27068
for its efficient formulation solution. network-flow based also We very approximate a construct,We also construct a very efficient network-flow based formulation for its approximate solution.,27069
postures model temporal super-resolution. also The retrieval not of similar only enables but,The model not only enables retrieval of similar postures but also temporal super-resolution.,27070
"next frames can synthesized. recurrent based be on Additionally, a formulation,","Additionally, based on a recurrent formulation, next frames can be synthesized.",27071
neural We attribute for present a multiple approach deep inference structured networks prediction. in,We present a structured inference approach in deep neural networks for multiple attribute prediction.,27072
different node same graph Experiments each object images. show part that represents through the consistently,Experiments show that each graph node consistently represents the same object part through different images.,27073
"present Fully we To Network end, WSVRD. Pairwise Convolutional a Parallel, this for (PPR-FCN) Region-based,","To this end, we present a Parallel, Pairwise Region-based, Fully Convolutional Network (PPR-FCN) for WSVRD.",27074
clinical enabling stabilisation mood of mood an timely episodes goal. important is Early identification,Early identification of mood episodes enabling timely mood stabilisation is an important clinical goal.,27075
"a For a dataset in kitchen particular unlikely. is finding labeled example, large containing instances","For example, finding a large labeled dataset containing instances in a particular kitchen is unlikely.",27076
expensive new Each environment data instances requires with collection new annotation. and,Each new environment with new instances requires expensive data collection and annotation.,27077
`paste' We random backgrounds. automatically them object `cut' and instances on,We automatically `cut' object instances and `paste' them on random backgrounds.,27078
methods. compared properties observed favorable when Several are with conventional,Several favorable properties are observed when compared with conventional methods.,27079
"against the feature robustness improving objective First, invariance scale variance. is of introduced,","First, scale invariance is introduced, improving the robustness of objective against feature variance.",27080
"demonstrated on available has experiments better datasets. three been publicly convergence Third, by","Third, better convergence has been demonstrated by experiments on three publicly available datasets.",27081
augment objects. with real to present images efficient We an procedure virtual,We present an efficient procedure to augment real images with virtual objects.,27082
smoking development. premature causes shown have that heavy studies wrinkles Clinical,Clinical studies have shown that heavy smoking causes premature wrinkles development.,27083
to smoking. is was region analysis effected by the mainly performed analyse statistical Then which,Then the statistical analysis was performed to analyse which region is effected mainly by smoking.,27084
regions inconclusive lack due dataset. to are Other of large scale,Other regions are inconclusive due to lack of large scale dataset.,27085
image effective learning Deep extract to an features. is method,Deep learning is an effective method to extract image features.,27086
"Using a the Google was convolution neural created. framework, library powerful (CNN) Tensorflow network","Using the Google Tensorflow framework, a powerful convolution neural network (CNN) library was created.",27087
completion. recent representation Low underpins of tensor tensor rank much progress in,Low rank tensor representation underpins much of recent progress in tensor completion.,27088
human fundamental is a vision. pose challenging Articulated in estimation yet task computer,Articulated human pose estimation is a fundamental yet challenging task in computer vision.,27089
benchmarks human for pose on investigate estimation. method standard We our two,We investigate our method on two standard benchmarks for human pose estimation.,27090
approach Our on obtains both state-of-the-art results benchmarks.,Our approach obtains state-of-the-art results on both benchmarks.,27091
employ obtain hierarchical the an strategy organization classes. iterative max-margin of clustering to We the,We employ an iterative max-margin clustering strategy to obtain the hierarchical organization of the classes.,27092
state-of-the-arts. our outperforms that datasets benchmark Performance several on validates evaluation method popular existing,Performance evaluation on several popular benchmark datasets validates that our method outperforms existing state-of-the-arts.,27093
estimation a field Depth applications. light problem for is photography fundamental,Depth estimation is a fundamental problem for light field photography applications.,27094
"we (POBR) first superpixel via boundary Specifically, regularization. occluded partially based detect regions","Specifically, we first detect partially occluded boundary regions (POBR) via superpixel based regularization.",27095
semantic videos We an without learning using approach unsupervised representation present labels.,We present an unsupervised representation learning approach using videos without semantic labels.,27096
"to the methods representation. simply techniques skeletal In naive literature, the from transfer off-the-shelf video","In the literature, naive methods simply transfer off-the-shelf techniques from video to the skeletal representation.",27097
"approaches but heavy, complementary, yet performances, drawbacks. exhibit strong Both show they","Both approaches show strong performances, yet they exhibit heavy, but complementary, drawbacks.",27098
autoencoders. banana-shaped for variational for We algorithm proposed distributions fitting training demonstrate and the,We demonstrate the proposed algorithm for fitting banana-shaped distributions and for training variational autoencoders.,27099
"short too predetermined the are and vice receptive a distance, versa. Specifically, fields at small","Specifically, the predetermined receptive fields are too small at a short distance, and vice versa.",27100
DaM to neural fully networks. is convolution convolutional proposed applied The two,The proposed DaM convolution is applied to two fully convolutional neural networks.,27101
common in several tree is Segmenting image structures applications. processing,Segmenting tree structures is common in several image processing applications.,27102
"the the For is the desired output object. task, correction true error","For the error correction task, the desired output is the true object.",27103
net can the accuracy. find that error-detecting high We achieve,We find that the error-detecting net can achieve high accuracy.,27104
protocols experimental Multiple different achieves evaluated on proposed demonstrate the that performance. state-of-the-art experiments method,Multiple experiments evaluated on different experimental protocols demonstrate that the proposed method achieves state-of-the-art performance.,27105
"of meaningful For essential. images are representations search, visual yet large-scale highly compressed","For large-scale visual search, highly compressed yet meaningful representations of images are essential.",27106
models publicly and code make available our We online.,We make our code and models publicly available online.,27107
similarities between measuring descriptions matching sentence images. aims and Textual-visual at,Textual-visual matching aims at measuring similarities between sentence descriptions and images.,27108
tackle Most without problem existing effectively methods annotations. identity-level this utilizing,Most existing methods tackle this problem without effectively utilizing identity-level annotations.,27109
"propose an two-stage framework problem. for paper, matching textual-visual we identity-aware the In this","In this paper, we propose an identity-aware two-stage framework for the textual-visual matching problem.",27110
experiments over The datasets multi-view its validate superiority.,The experiments over multi-view datasets validate its superiority.,27111
"In for we this CNN-based a propose framework online paper, MOT.","In this paper, we propose a CNN-based framework for online MOT.",27112
then is The features. to weight attention spatial the applied map,The spatial attention map is then applied to weight the features.,27113
be as attention It can mechanism. considered temporal,It can be considered as temporal attention mechanism.,27114
delivered convolutional (CNNs) Deep in superior tasks. networks have performance many computer vision neural,Deep convolutional neural networks (CNNs) have delivered superior performance in many computer vision tasks.,27115
be other proposed deep The methods networks. convolutional to can applied also,The proposed methods can also be applied to other deep convolutional networks.,27116
that saliency proposed state-of-the-art Extensive favorably against performs demonstrate our approaches. experiments model,Extensive experiments demonstrate that our proposed saliency model performs favorably against state-of-the-art approaches.,27117
of requires constant. approximation the intractable normalizing an parameter ERGM estimation,ERGM parameter estimation requires the approximation of an intractable normalizing constant.,27118
"When (MPLE). the estimation pseudolikelihood large, one maximum is network is option","When the network is large, one option is maximum pseudolikelihood estimation (MPLE).",27119
"standard and standard generally The underestimates MPLE fast, but is simple errors.","The standard MPLE is simple and fast, but generally underestimates standard errors.",27120
method---the a bootstrap---results coverage We accurate confidence in show that parametric for probabilities resampling intervals.,We show that a resampling method---the parametric bootstrap---results in accurate coverage probabilities for confidence intervals.,27121
"real measure a But, stymied computational large-scale in attempts information has to brains. hurdle integration","But, a computational hurdle has stymied attempts to measure large-scale information integration in real brains.",27122
models. complex cannot the Such efficiently modeled uncertainty deterministic be by,Such complex uncertainty cannot be modeled efficiently by the deterministic models.,27123
"model. paper, existing an to this algorithm propose In ""Sparse prune a CNN Shrink"" we","In this paper, we propose a ""Sparse Shrink"" algorithm to prune an existing CNN model.",27124
resource. model saves resulting directly computational thus pruned The,The resulting pruned model thus directly saves computational resource.,27125
"have results of effectiveness demonstrated These Shrink"" our algorithm. ""Sparse the","These results have demonstrated the effectiveness of our ""Sparse Shrink"" algorithm.",27126
WCNN spectrum. visual low-level available of in images widely face layers with are The trained,The low-level layers of WCNN are trained with widely available face images in visual spectrum.,27127
end-to-end is prior a in by This implemented constraint network. an low-rank,This prior is implemented by a low-rank constraint in an end-to-end network.,27128
representations with become self-supervised computer popular visual has Learning learning in vision.,Learning visual representations with self-supervised learning has become popular in computer vision.,27129
are where design to tasks The idea obtain. free auxiliary to labels is,The idea is to design auxiliary tasks where labels are free to obtain.,27130
brain importance and Knowing of both applications. clinical in great basic connectivity for is research,Knowing brain connectivity is of great importance both in basic research and for clinical applications.,27131
identify that in allows causal neuronal This activity. to reflected are relations population us,This allows us to identify causal relations that are reflected in neuronal population activity.,27132
"also affecting the its activity dynamics, We like explored spectrum. eigenvalue parameters","We also explored parameters affecting its activity dynamics, like the eigenvalue spectrum.",27133
"unobserved this and we In additional factors. nodes confounding study, consider measurement as noise","In this study, we consider measurement noise and unobserved nodes as additional confounding factors.",27134
"amount for investigate the of reliable a we Furthermore, data required estimate.","Furthermore, we investigate the amount of data required for a reliable estimate.",27135
"method apply a dataset. on fMRI proposed we the Additionally,","Additionally, we apply the proposed method on a fMRI dataset.",27136
"we that large connections fraction found inhibitory. Also, a identified of were","Also, we found that a large fraction of identified connections were inhibitory.",27137
VehcileID the revised on experiments conduct dataset. extensive We,We conduct extensive experiments on the revised VehcileID dataset.,27138
embrionary have origin. not These groups do an,These groups do not have an embrionary origin.,27139
"In paper, this investigate is we case. this why the","In this paper, we investigate why this is the case.",27140
marker false for on IDs algorithms positive detection rely marker rejection. fiducial Current,Current fiducial marker detection algorithms rely on marker IDs for false positive rejection.,27141
is eventually be detections rejected as false Time that potential will positives. on wasted,Time is wasted on potential detections that will eventually be rejected as false positives.,27142
accuracy. We how effect detection show and tag viewing size also direction,We also show how tag size and viewing direction effect detection accuracy.,27143
contribution markers applications are used in because (e.g. Our often is fiducial significant real-time,Our contribution is significant because fiducial markers are often used in real-time applications (e.g.,27144
assisted other required where system. is marker navigation) robot the by heavy parts of computation,marker assisted robot navigation) where heavy computation is required by other parts of the system.,27145
solid method optical phantoms The tested with optical predetermined on properties. proposed is,The proposed method is tested on optical solid phantoms with predetermined optical properties.,27146
method The images. on human healthy is also tested skin,The method is also tested on healthy human skin images.,27147
formed curvature or be basis of either touch. representations vision may the of Brain on,Brain representations of curvature may be formed on the basis of either vision or touch.,27148
of is algorithm The ratio the derived. approximation greedy,The approximation ratio of the greedy algorithm is derived.,27149
KNN and algorithms KNN regression performance over improve regression. against state-of-the-art dramatically perform ensembles favorably,KNN regression ensembles perform favorably against state-of-the-art algorithms and dramatically improve performance over KNN regression.,27150
object. tight It corners of on box around a imaginary clicking the involves,It involves clicking on imaginary corners of a tight box around the object.,27151
find. is points and task natural easy are to these This more,This task is more natural and these points are easy to find.,27152
of have logic-based a in developments representations number There measuring knowledge. inconsistency in of been,There have been a number of developments in measuring inconsistency in logic-based representations of knowledge.,27153
"of the development models been In of measures argument computational for limited. has contrast, inconsistency","In contrast, the development of inconsistency measures for computational models of argument has been limited.",27154
"visual discriminative effective we method question generation. present an addition, In for","In addition, we present an effective method for visual discriminative question generation.",27155
baselines through shown quantitative evaluations studies. and results Promising are representative against user,Promising results are shown against representative baselines through quantitative evaluations and user studies.,27156
"systems operating different the on Furthermore, be and can used data platforms.","Furthermore, the data can be used on different operating systems and platforms.",27157
in selection the literature. Feature studied widely been has,Feature selection has been studied widely in the literature.,27158
similarity. Most based are on feature the selection of the existing criteria sample,Most of the existing feature selection criteria are based on the sample similarity.,27159
"size become (HDLSS) for sample high However, measures low dimensional data. the insignificant distance","However, the distance measures become insignificant for high dimensional low sample size (HDLSS) data.",27160
is task. Reinforcement learn to proven a for technique learning a agent an,Reinforcement learning is a proven technique for an agent to learn a task.,27161
to between transfer makes tasks in it environment. skills the harder This same,This makes it harder to transfer skills between tasks in the same environment.,27162
"this training does Furthermore, not reduce new for when task. risk a","Furthermore, this does not reduce risk when training for a new task.",27163
"tags, with description, along Meta images those etc.) (e.g., web title, information","Meta information along with those web images (e.g., title, description, tags, etc.)",27164
sketches greatly arbitrariness high challenging. Recognizing is freehand with,Recognizing freehand sketches with high arbitrariness is greatly challenging.,27165
"though techniques high hardly Consequently, yield performance can recognition they sophisticated employed. learning are even","Consequently, they can hardly yield high recognition performance even though sophisticated learning techniques are employed.",27166
multiple payoff A complementary function is features. and constructed combining cues exploiting by,A payoff function is constructed by exploiting multiple cues and combining complementary features.,27167
"applications, many images Parsing benefits self-driving. scene urban especially","Parsing urban scene images benefits many applications, especially self-driving.",27168
"image considerable Recently, been neural (CNNs) convolutional deep networks very attracting have restoration. in attention","Recently, very deep convolutional neural networks (CNNs) have been attracting considerable attention in image restoration.",27169
of recursive The unit fields. different under learns state representations the multi-level current receptive,The recursive unit learns multi-level representations of the current state under different receptive fields.,27170
"image and apply denosing, JPEG We restoration image three deblocking. tasks, super-resolution MemNet to i.e.,","We apply MemNet to three image restoration tasks, i.e., image denosing, super-resolution and JPEG deblocking.",27171
of a differentiable efficient metric. propose and this We approximation,We propose a differentiable and efficient approximation of this metric.,27172
"by state-of-the-art generative models. generated Here, deep low-level we statistics images of the explore","Here, we explore the low-level statistics of images generated by state-of-the-art deep generative models.",27173
shown in neural (FCNs) labeling outstanding many problems. have networks performance Fully convolutional dense,Fully convolutional neural networks (FCNs) have shown outstanding performance in many dense labeling problems.,27174
of pillar from One successes is layers. mining these features convolutional relevant key in information,One key pillar of these successes is mining relevant information from features in convolutional layers.,27175
"object to is detection multi-level for maps aggregate However, salient underexplored. feature convolutional how better","However, how to better aggregate multi-level convolutional feature maps for salient object detection is underexplored.",27176
"to saliency generate are the map. final fused predicted results the Finally, efficiently","Finally, the predicted results are efficiently fused to generate the final saliency map.",27177
by potential the framework numerical results. is illustrated of The MVIE,The potential of the MVIE framework is illustrated by numerical results.,27178
paper. We propose generation graphical interface user a in based tool groundtruth this,We propose a graphical user interface based groundtruth generation tool in this paper.,27179
"foreground done image on an annotation pixels. input of Here, based the is document","Here, annotation of an input document image is done based on the foreground pixels.",27180
pixels interaction form user grouped are units. together labeling to with Foreground,Foreground pixels are grouped together with user interaction to form labeling units.,27181
with labeled defined the the then These user are by user units labels.,These units are then labeled by the user with the user defined labels.,27182
annotated data different further used document image This in of can be analysis. applications,This annotated data can be further used in different applications of document image analysis.,27183
"Graph significant currently powerful, kernels, dominated by is while limitations. learning graph some suffer which,","Graph learning is currently dominated by graph kernels, which, while powerful, suffer some significant limitations.",27184
"this have many CNNs extensions recently introduced. been To sophisticated challenge, of address","To address this challenge, many sophisticated extensions of CNNs have recently been introduced.",27185
kernels also in is complexity. terms Our time graph approach preferable of to,Our approach is also preferable to graph kernels in terms of time complexity.,27186
brain. the fetal MR of the allow fetus non-invasive images of analysis,MR images of the fetus allow non-invasive analysis of the fetal brain.,27187
for an method segmentation automatic in of images. This presents the paper ICV MR fetal,This paper presents an automatic method for segmentation of the ICV in fetal MR images.,27188
task. is Visual fundamental a tracking time-critical vision and object,Visual object tracking is a fundamental and time-critical vision task.,27189
"of this accuracy. improve learn to speed losing we deep without trackers In the paper,","In this paper, we learn to improve the speed of deep trackers without losing accuracy.",27190
easy for distinct frames This objects. feed-forward cost reduces the or significantly slow-moving with,This significantly reduces the feed-forward cost for easy frames with distinct or slow-moving objects.,27191
anatomical images. in an automated an method localizing We landmark present three-dimensional medical for,We present an automated method for localizing an anatomical landmark in three-dimensional medical images.,27192
landmark accurately in and The neighborhood. second candidate actual the network precisely the localizes,The second network localizes the actual landmark precisely and accurately in the candidate neighborhood.,27193
toward detection and have Face recognition difficult benchmarks environments. shifted more,Face detection and recognition benchmarks have shifted toward more difficult environments.,27194
"contrast, is face attention. currently open-set weak By requires much and recognition more","By contrast, open-set face recognition is currently weak and requires much more attention.",27195
"objects tiny this aim only paper, we to segment datasets. using In","In this paper, we aim to segment objects using only tiny datasets.",27196
is a This detection. shape for descriptor bag-of-words in used approach object,This shape descriptor is used in a bag-of-words approach for object detection.,27197
a an rule discriminative that for We subset efficient variables. solutions of optimal finds describe,We describe an efficient discriminative rule that finds optimal solutions for a subset of variables.,27198
per-instance the technique guarantees quality worst-case provides of Our both the concerning and solution.,Our technique provides both per-instance and worst-case guarantees concerning the quality of the solution.,27199
over Empirical conducted were studies benchmark datasets. several,Empirical studies were conducted over several benchmark datasets.,27200
latent both are structural introduced. The structural latent model the higher-order and model unstructured,The unstructured latent structural model and the higher-order latent structural model are both introduced.,27201
human for Context temporal a Network activities. We present localization of (TCN) precise Temporal,We present a Temporal Context Network (TCN) for precise temporal localization of human activities.,27202
We these a propose for ranking novel representation proposals.,We propose a novel representation for ranking these proposals.,27203
"ranking classification After is applied obtain and proposals, non-maximum suppression is final to detections. performed","After ranking proposals, non-maximum suppression is applied and classification is performed to obtain final detections.",27204
"Despite as re-emerged vaccine high pertussis public coverage, health countries. concern a many in has","Despite high vaccine coverage, pertussis has re-emerged as a public health concern in many countries.",27205
the immunity. is One re-emergence posed waning of for hypothesis,One hypothesis posed for re-emergence is the waning of immunity.,27206
become scene such understanding crucial autonomous as many applications has in driving. Real-time,Real-time scene understanding has become crucial in many applications such as autonomous driving.,27207
reason What way to human is the about right activities?,What is the right way to reason about human activities?,27208
"this current in we the understanding In human work, analyze videos. state activity of","In this work, we analyze the current state of human activity understanding in videos.",27209
state-of-the-art multiple datasets. experiments consider The multiple algorithms and,The experiments consider multiple state-of-the-art algorithms and multiple datasets.,27210
We the aid that RSV results those policies. hope will vaccination in of design,We hope that those results will aid in the design of RSV vaccination policies.,27211
"via we this problem In paper, learning. solve two-phase fundamentally this","In this paper, we solve this problem fundamentally via two-phase learning.",27212
networks are steps. in trained two Our,Our networks are trained in two steps.,27213
demonstrate two-phase learn- Detailed the of effectiveness task. ments our ing each in experi-,Detailed experi- ments demonstrate the effectiveness of our two-phase learn- ing in each task.,27214
the approach. demonstrate our of effectiveness Extensive experiments,Extensive experiments demonstrate the effectiveness of our approach.,27215
"state-of-the-art on results achieve all We challenging i.e. datasets, three","We achieve state-of-the-art results on all three challenging datasets, i.e.",27216
"we our improve of error of reduce quantization influence Thus, robustness the and method. the","Thus, we reduce the influence of quantization error and improve the robustness of our method.",27217
lung for Our to expression applied was adenocarcinomas gene real cancer). method (lung data of,Our method was applied to real data of gene expression for lung adenocarcinomas (lung cancer).,27218
is for studied of commitment widely notion as interaction. The abstraction high-level modeling multiagent a,The notion of commitment is widely studied as a high-level abstraction for modeling multiagent interaction.,27219
decentralized commitment important specifications. of challenge flexible supporting An enactments is,An important challenge is supporting flexible decentralized enactments of commitment specifications.,27220
"information protocols. combine this In advances specifying recent we on commitments and paper,","In this paper, we combine recent advances on specifying commitments and information protocols.",27221
"Tosca, contribute commitment from we a specifications. Specifically, protocols for technique synthesizing information automatically","Specifically, we contribute Tosca, a technique for automatically synthesizing information protocols from commitment specifications.",27222
attention of of It recent gains person increasing because the re-identification advances techniques.,It gains increasing attention because of the recent advances of person re-identification techniques.,27223
task. additional solving the Incorporating vital re-identification spatio-temporal is challenging information for,Incorporating additional spatio-temporal information is vital for solving the challenging re-identification task.,27224
our proposed components. the experiments analysis and effectiveness show and method individual of Extensive,Extensive experiments and analysis show the effectiveness of our proposed method and individual components.,27225
action an computer task vision. is important localization Temporal of,Temporal action localization is an important task of computer vision.,27226
"temporal network However, temporal partially CDC information loses due the to operation. downsampling","However, CDC network loses temporal information partially due to the temporal downsampling operation.",27227
TPC network manner. trained end-to-end can an in be,TPC network can be trained in an end-to-end manner.,27228
illustrate performance to simulations of provide of We a range algorithms. the proposed the also,We also provide a range of simulations to illustrate the performance of the proposed algorithms.,27229
fully are Surveillance data typically for aggregated serving in space. alert epidemic systems,Surveillance data serving for epidemic alert systems are typically fully aggregated in space.,27230
analyses retrospective this by incidence series. in We time unveil classifying,We unveil this in retrospective analyses by classifying incidence time series.,27231
incidence with divided into different regions thus is patterns. The area surveillance,The surveillance area is thus divided into regions with different incidence patterns.,27232
influenza epidemiology impact The was of on spatial heterogeneity complex.,The impact of spatial heterogeneity on influenza epidemiology was complex.,27233
"Third, epidemic were homogeneous space. dynamics the in","Third, the epidemic dynamics were homogeneous in space.",27234
We the proposed the performance model analyze w.r.t. of,We analyze the performance of the proposed model w.r.t.,27235
"of size chosen. the architectural the training input and features data, hyperparameters, diversity and","the architectural hyperparameters, size and diversity of training data, and the input features chosen.",27236
"features, image or more compact. aligned local with better","better aligned with local image features, or more compact.",27237
"has on population processes, In selection. structure substantial natural evolutionary effect a","In evolutionary processes, population structure has a substantial effect on natural selection.",27238
"we affects constant how Here, populations. motion analyze in individuals structured selection of","Here, we analyze how motion of individuals affects constant selection in structured populations.",27239
lattice and the weighted how affects graphs. We investigate evolution further motion square on,We further investigate how motion affects evolution on the square lattice and weighted graphs.,27240
our Experiments demonstrate effective that is method and start-of-the-art outperforms methods. proposed,Experiments demonstrate that our proposed method is effective and outperforms start-of-the-art methods.,27241
state-of-the-art (CNNs) models for classification Networks Convolutional document are Neural tasks. image,Convolutional Neural Networks (CNNs) are state-of-the-art models for document image classification tasks.,27242
"easy However, convenient. solution neither or is","However, neither solution is easy or convenient.",27243
"novel for we this digitizing configurations. chessboard algorithm solve a propose problem, To","To solve this problem, we propose a novel algorithm for digitizing chessboard configurations.",27244
"existing each algorithm we our steps, solutions. that these surpasses For of own designed","For each of these steps, we designed our own algorithm that surpasses existing solutions.",27245
algorithms machine whenever possible. learning our utilizing We support by techniques,We support our algorithms by utilizing machine learning techniques whenever possible.,27246
sensing. coverage remote quality winter vegetation Mapping a problem of is challenge,Mapping winter vegetation quality coverage is a challenge problem of remote sensing.,27247
Neural were classifier deep employed. Two based Network methods Recurrent (RNN),Two deep Recurrent Neural Network (RNN) based classifier methods were employed.,27248
for estimation motion flow can Optical it predicts used kernel since be trajectories.,Optical flow can be used for kernel estimation since it predicts motion trajectories.,27249
"performance that affect and like However problems scale it rotation change. occlusion, from its suffers","However it suffers from problems that affect its performance like occlusion, rotation and scale change.",27250
an in are approaches Most video state-of-the-art pipeline. recognition end-to-end of of existing favor,Most of existing state-of-the-art video recognition approaches are in favor of an end-to-end pipeline.,27251
exception framework DevNet. the is One of,One exception is the framework of DevNet.,27252
the scores (e.g. using Instead directly end-to-end classification of,Instead of directly using the end-to-end classification scores (e.g.,27253
"and effectiveness underestimated. work However, of this line ignored long-term the has been","However, the effectiveness of this line work has long-term been ignored and underestimated.",27254
"we use submission, extensively this strategy. In this","In this submission, we extensively use this strategy.",27255
"usually with in have variation. inter-graph considerable ARGs attributes, hidden soft these subgraphs Common","Common subgraphs hidden in these ARGs usually have soft attributes, with considerable inter-graph variation.",27256
proposed experiments generality demonstrate the method. These of the,These experiments demonstrate the generality of the proposed method.,27257
of proposed the Experimental datasets the effectiveness results on various approach. demonstrate,Experimental results on various datasets demonstrate the effectiveness of the proposed approach.,27258
deep Cell employing from evolved have features techniques. methods hand-crafted to detection learning-based,Cell detection methods have evolved from employing hand-crafted features to deep learning-based techniques.,27259
"this from input CNN compressed regresses Then, vector pixels. the","Then, CNN regresses this compressed vector from the input pixels.",27260
is wild computing a affective Automated in problem challenging vision. setting the computer in,Automated affective computing in the wild setting is a challenging problem in computer vision.,27261
theory. this Grenander's generator We language canonical of express mathematically using the pattern,We mathematically express this using the language of Grenander's canonical pattern generator theory.,27262
"In been applications. adopted have depth consumer-level various recent for cameras years,","In recent years, consumer-level depth cameras have been adopted for various applications.",27263
"are On low-cost, the available. other high-frame-rate cameras video hand,","On the other hand, low-cost, high-frame-rate video cameras are available.",27264
topic represents computer a This in research nascent vision.,This represents a nascent research topic in computer vision.,27265
work of contributions technical three-fold. are this The,The technical contributions of this work are three-fold.,27266
synthetic models in racing from games. many This differs data from prior trained,This differs from many prior models trained from synthetic data in racing games.,27267
construct pairwise network context affinity. branch learnt is pixel-wise semantic edge by the to The,The semantic edge context is learnt by the pairwise network branch to construct pixel-wise affinity.,27268
We Stage (SSH) face the Headless Single introduce detector.,We introduce the Single Stage Headless (SSH) face detector.,27269
make properties These fast SSH and light-weight.,These properties make SSH fast and light-weight.,27270
method from learning unsupervised video for a face sequences. novel identity We present,We present a novel unsupervised method for face identity learning from video sequences.,27271
and characterizing video dynamics. three-dimensional Human captured are appearance sequences in signals motion actions visual,Human actions captured in video sequences are three-dimensional signals characterizing visual appearance and motion dynamics.,27272
"and/or methods adopt learn patterns, and (CNNs action RNNs). Recurrent Convolutional existing To Networks Neural","To learn action patterns, existing methods adopt Convolutional and/or Recurrent Neural Networks (CNNs and RNNs).",27273
"able Memory are (LSTM), temporal to learn RNNs, dynamics. Short-Term motion especially Long","RNNs, especially Long Short-Term Memory (LSTM), are able to learn temporal motion dynamics.",27274
"Additionally, novel introduce a procedure training network. multi-modal our we for training","Additionally, we introduce a novel multi-modal training procedure for training our network.",27275
"the exceed results margins. by wide we state-of-the-art quantitative also the For latter, and qualitative","For the latter, we also exceed the state-of-the-art quantitative and qualitative results by wide margins.",27276
"and this handcrafted paper, from some study features extracted intra-variable In auto-derived we writing. features","In this paper, we study some handcrafted features and auto-derived features extracted from intra-variable writing.",27277
"on from of writer offline work handwriting intra-variability. we identification/verification Here, high Bengali","Here, we work on writer identification/verification from offline Bengali handwriting of high intra-variability.",27278
"the we predictions densely To average extracted or images, CNN patches. classify line page over","To classify page or line images, we average the CNN predictions over densely extracted patches.",27279
"a propose deep image paper, novel using learning medical iterative In segmentation framework. we this","In this paper, we propose a novel medical image segmentation using iterative deep learning framework.",27280
network We for a of deep classification flowers. have different learning developed,We have developed a deep learning network for classification of different flowers.,27281
and The divided segmentation basically is parts; two into classification. Image method,The method is basically divided into two parts; Image segmentation and classification.,27282
part estimation in pose computer semantic tasks segmentation and complementary Human two are vision.,Human pose estimation and semantic part segmentation are two complementary tasks in computer vision.,27283
"the number nodes, increase With grows. of size in batch the an","With an increase in the number of nodes, the batch size grows.",27284
in batch model But with accuracy. large often size lower results the training,But training with large batch size often results in the lower model accuracy.,27285
"propose generation recurrent tracking. we In for methods visual a this paper, filter","In this paper, we propose a recurrent filter generation methods for visual tracking.",27286
method that datasets experiments public algorithms. on demonstrate hand pose Comprehensive proposed state-of-the-art our outperforms,Comprehensive experiments on public hand pose datasets demonstrate that our proposed method outperforms state-of-the-art algorithms.,27287
mobile challenging  a  neural task. on Deploying is networks deep devices,  Deploying deep neural networks on mobile devices is a challenging task.,27288
"Smile database. obtained benchmark of especially deep results, Promising the UvA-NEMO those on features, are","Promising results, especially those of deep features, are obtained on the benchmark UvA-NEMO Smile database.",27289
"two the the (i.e. Furthermore, of fusion features of types","Furthermore, the fusion of the two types of features (i.e.",27290
state-of-the-art significant shallow performance improvements spatio-temporal features compared methods. shows to texture and features) deep,shallow spatio-temporal texture features and deep features) shows significant performance improvements compared to state-of-the-art methods.,27291
one carried asymmetrically from by part Information is another entropy to transfer.,Information is carried asymmetrically from one part to another by entropy transfer.,27292
approaches complexity computational similar. both that the reveals It of is,It reveals that the computational complexity of both approaches is similar.,27293
line-level We identification method. describe novel script a,We describe a novel line-level script identification method.,27294
inefficient to makes harder train to run. it This and,This makes it harder to train and inefficient to run.,27295
"counting be Second, a heuristic to suboptimal model. compared may the learned","Second, the counting heuristic may be suboptimal compared to a learned model.",27296
a encoder line converts a sequence. feature image The into,The encoder converts a line image into a feature sequence.,27297
the the sequence aggregates line. to classify The summarizer,The summarizer aggregates the sequence to classify the line.,27298
with convolutional We summarizers encoders. as various test identical networks inception-style,We test various summarizers with identical inception-style convolutional networks as encoders.,27299
"the purpose. other indispensable and convergence is (BN) hand, residual batch for On normalization denoising","On the other hand, batch normalization (BN) is indispensable for residual denoising and convergence purpose.",27300
"the stacking performance CNN. However, degrades BN ELU of direct of and","However, direct stacking of BN and ELU degrades the performance of CNN.",27301
"applied train a are denoising autoencoders this to model. stacked In paper, deep learning","In this paper, stacked denoising autoencoders are applied to train a deep learning model.",27302
We learning. adversarial priors also improved extend with with latent and variable VCCA,We also extend VCCA with improved latent variable priors and with adversarial learning.,27303
to within information convolutional segmentation network neural (CNN). a integrate an introduce approach We,We introduce an approach to integrate segmentation information within a convolutional neural network (CNN).,27304
these local use embeddings position. a mask neuron to to We relative compute every attention,We use these embeddings to compute a local attention mask relative to every neuron position.,27305
"segmentation-aware baselines. cases, improvements systematic convolution In strong over both yields","In both cases, segmentation-aware convolution yields systematic improvements over strong baselines.",27306
Safety as paramount mobile aerial platforms such is for vehicles. robotic self-driving cars and unmanned,Safety is paramount for mobile robotic platforms such as self-driving cars and unmanned aerial vehicles.,27307
"work, detection. obstacle thin for to In sequences we propose this video use","In this work, we propose to use video sequences for thin obstacle detection.",27308
stereo solution a a We provide both and solution. camera monocular camera,We provide both a monocular camera solution and a stereo camera solution.,27309
from end-to-end training hierarchically to CNN learn able manner. is images abstracted an features in,CNN is able to learn hierarchically abstracted features from images in an end-to-end training manner.,27310
"a with this we propose In paper, ""Learning algorithm. Rethinking""","In this paper, we propose a ""Learning with Rethinking"" algorithm.",27311
"pre-trained Particularly, can it any be boost employed to models.","Particularly, it can be employed to boost any pre-trained models.",27312
CNN mechanism. results models the advantage These demonstrated with of proposed the have feedback training,These results have demonstrated the advantage of training CNN models with the proposed feedback mechanism.,27313
the popularity sharing past during Bike been systems' years. consistently has rising,Bike sharing systems' popularity has consistently been rising during the past years.,27314
are systems. systems emerging indispensable and Managing these of these parts maintaining,Managing and maintaining these emerging systems are indispensable parts of these systems.,27315
discuss improvements also We possible cases. special multiple for,We also discuss multiple possible improvements for special cases.,27316
"clinically lobe enables pathological important regional capability. Automatic pulmonary of analyses a lung segmentation(PPLS) disease,","Automatic pathological pulmonary lobe segmentation(PPLS) enables regional analyses of lung disease, a clinically important capability.",27317
apply learning to first the We are PPLS. deep to,We are the first to apply deep learning to PPLS.,27318
biological regulatory control circuits Genetically function. encoded,Genetically encoded regulatory circuits control biological function.,27319
A system such example is a AC-DC simple particularly the of circuit.,A particularly simple example of such a system is the AC-DC circuit.,27320
(SfM) SLAM incremental from methods. structure-from-motion are visual Many algorithms monocular derived,Many monocular visual SLAM algorithms are derived from incremental structure-from-motion (SfM) methods.,27321
"to contributions visual two we SLAM. present In main particular,","In particular, we present two main contributions to visual SLAM.",27322
Neural impact smooth has great surgery. the a segmentation implementation of local anesthesia on,Neural segmentation has a great impact on the smooth implementation of local anesthesia surgery.,27323
"training structure, longer needs more SegNet has samples. deeper and time, training network it but","SegNet network has deeper structure, but it needs longer training time, and more training samples.",27324
"U-NET for propose segmentation. improved an we In neural this paper, the network","In this paper, we propose an improved U-NET neural network for the segmentation.",27325
through original residual importing network deepens This network. the structure,This network deepens the original structure through importing residual network.,27326
structure has neural U-NET scene improved The network in a application segmentation. good,The improved U-NET network structure has a good application scene in neural segmentation.,27327
"neural an In of paper. binding considers particular, structure earlier the it","In particular, it considers the neural binding structure of an earlier paper.",27328
It in methods areas of new also and behaviour the image simulation. some processing describes,It also describes some new methods in the areas of image processing and behaviour simulation.,27329
"'full with is linking'. a For used structure processing, image grid-like","For image processing, a grid-like structure is used with 'full linking'.",27330
probability trained item are co-occurrences. Two neural the maximize networks of different to convolutional,Two different convolutional neural networks are trained to maximize the probability of item co-occurrences.,27331
differentiate We attributes snow for and of translucency chromatic into estimation. accurate also aberration,We also differentiate snow into attributes of translucency and chromatic aberration for accurate estimation.,27332
and The benefit would indicate vision our applications results computer graphics. involving network,The results indicate our network would benefit applications involving computer vision and graphics.,27333
"semantic this To Deconvolutional Stacked we for problem, Network propose segmentation. a (SDN) address","To address this problem, we propose a Stacked Deconvolutional Network (SDN) for semantic segmentation.",27334
"circumstances, available. However, type only in one is feature pedestrian of many","However, in many circumstances, only one type of pedestrian feature is available.",27335
"and we framework use"" re-ranking re-ID. propose paper, In this for ""Divide a person","In this paper, we propose a ""Divide and use"" re-ranking framework for person re-ID.",27336
"an given is image, into extracted sub-features. feature Specifically, divided the","Specifically, given an image, the extracted feature is divided into sub-features.",27337
feature. information new sub-feature contextual iteratively each into the encoded is of a Then,Then the contextual information of each sub-feature is iteratively encoded into a new feature.,27338
"same vector re-ranking. image the one the Finally, new for from features fused are into","Finally, the new features from the same image are fused into one vector for re-ranking.",27339
effectiveness the proposed demonstrate results person of re-ID benchmarks on two Experimental the framework.,Experimental results on two person re-ID benchmarks demonstrate the effectiveness of the proposed framework.,27340
sequence beyond is video for a Doing this single-handedly imagination.,Doing this for a video sequence single-handedly is beyond imagination.,27341
Our approach stylization problem. a second formulates learning video as,Our second approach formulates video stylization as a learning problem.,27342
the We quantitatively. simpler show clearly and outperform proposed baselines methods both that qualitatively,We show that the proposed methods clearly outperform simpler baselines both qualitatively and quantitatively.,27343
we teach what we What do should teach? and,What do we teach and what should we teach?,27344
annotations are time-consuming to obtain. Pixel-level expensive and,Pixel-level annotations are expensive and time-consuming to obtain.,27345
"segmentation. semantic Hence, significant in image tags supervision weak using impact a could only have","Hence, weak supervision using only image tags could have a significant impact in semantic segmentation.",27346
"existing class. single most are methods designed handle However, background a to","However, most existing methods are designed to handle a single background class.",27347
algorithm scenes a literature. tracking is from evaluated The different of collection on the,The algorithm is evaluated on a collection of different tracking scenes from the literature.,27348
of in situations We address the recognizing problem images.,We address the problem of recognizing situations in images.,27349
"many (nouns). and weapon), values role possible can each attacking take on has","attacking has weapon), and each role can take on many possible values (nouns).",27350
"brain Prominently, case. is the such one","Prominently, the brain is one such case.",27351
"a distribution. avalanches follow power in the law-like long-tailed, configuration, In this network","In this configuration, avalanches in the network follow a long-tailed, power law-like distribution.",27352
"the statistics in empirical at this brain. biological point findings reproduce avalanche Furthermore, the","Furthermore, the avalanche statistics at this point reproduce empirical findings in the biological brain.",27353
"or these imbalanced. In real-world applications, data is available limited the often networks train to","In real-world applications, the data available to train these networks is often limited or imbalanced.",27354
models marginal allow graphical networks Sum-product are (SPNs) of that tractable a inference. probabilistic class,Sum-product networks (SPNs) are a class of probabilistic graphical models that allow tractable marginal inference.,27355
"inference a in (MAP) posteriori However, is the maximum SPNs NP-hard.","However, the maximum a posteriori (MAP) inference in SPNs is NP-hard.",27356
in algorithmic from SPNs inference We investigate perspectives. theoretical both and MAP,We investigate MAP inference in SPNs from both theoretical and algorithmic perspectives.,27357
components? both possible Is to recover that,Is that possible to recover both components?,27358
"first focusing with rounds, The multiple of challenge round the on i.e. comprises gradual learning,","The challenge comprises of multiple rounds, with the first round focusing on gradual learning, i.e.",27359
to problems. the learning to ability already subsequent re-use solve learned for efficiently knowledge,the ability to re-use already learned knowledge for efficiently learning to solve subsequent problems.,27360
is enable FaceBoxes to real-time achieve CPU. to the on The speed designed RDCL,The RDCL is designed to enable FaceBoxes to achieve real-time speed on the CPU.,27361
"FaceBoxes the faces. Moreover, is to of speed the number of invariant","Moreover, the speed of FaceBoxes is invariant to the number of faces.",27362
was binary A pattern variant used of calculating feature vectors. local for,A variant of local binary pattern was used for calculating feature vectors.,27363
or Support feature labels. XGBoost vector vectors their was using trained machine and (SVM) the,Support vector machine (SVM) or XGBoost was trained using the feature vectors and their labels.,27364
used for and TPE optimization random or search of XGBoost. was parameter SVM,TPE or random search was used for parameter optimization of SVM and XGBoost.,27365
cross-validation evaluating our CADx performance system. Leave-one-out and used was the for optimizing of,Leave-one-out cross-validation was used for optimizing and evaluating the performance of our CADx system.,27366
analysis. operating of the area evaluated under characteristic Performance (AUC) curve was using receiver,Performance was evaluated using area under the curve (AUC) of receiver operating characteristic analysis.,27367
"SVM XGBoost classifying better lung conclusion, In nodules. than was for","In conclusion, XGBoost was better than SVM for classifying lung nodules.",27368
random more search efficient was TPE parameter for than optimization.,TPE was more efficient than random search for parameter optimization.,27369
utilizes both neural ResNet The and proposed framework architectures. network Encoder-Decoder,The proposed framework utilizes both ResNet and Encoder-Decoder neural network architectures.,27370
"outperforms existing levels, our At methods. all approach","At all levels, our approach outperforms existing methods.",27371
immune newly prokaryotic is discovered CRISPR a system.,CRISPR is a newly discovered prokaryotic immune system.,27372
for in The viruses problem biology. prokaryots an conditions and of evolutionary interesting is coexistence,The conditions for coexistence of prokaryots and viruses is an interesting problem in evolutionary biology.,27373
of variables increasing. data rapidly of availability numbers sets large The with is,The availability of data sets with large numbers of variables is rapidly increasing.,27374
Monte The to address Markov Carlo current new this adaptive chain paper shortcoming. algorithms proposes,The current paper proposes new adaptive Markov chain Monte Carlo algorithms to address this shortcoming.,27375
outcomes influential interpretable eventual The neurons. to us trace representations to induced back enable,The induced interpretable representations enable us to trace eventual outcomes back to influential neurons.,27376
"research diversified a scope Arabic, has with being a challenges. of potential huge language","Arabic, being a diversified language has a huge scope of research with potential challenges.",27377
"overcome change the vanishing of suitable is problem activation introduced Moreover, gradient. function to in","Moreover, suitable change is introduced in activation function to overcome the problem of vanishing gradient.",27378
system multi-view We present a stereopsis. learnt for,We present a learnt system for multi-view stereopsis.,27379
The the demonstrate the superior results experimental of proposed approach. ability,The experimental results demonstrate the superior ability of the proposed approach.,27380
analytics to Data tactics. basketball create helps teams,Data analytics helps basketball teams to create tactics.,27381
"data and manual However, costly ineffective. and analytics are collection","However, manual data collection and analytics are costly and ineffective.",27382
suitable is problems. for particularly structure series with time dealing Its,Its structure is particularly suitable for dealing with time series problems.,27383
"Combined to generate with outputs. MDN a BLSTMs, multi-modal distribution is of used","Combined with BLSTMs, MDN is used to generate a multi-modal distribution of outputs.",27384
"in arbitrary represent Thus, probability distributions output proposed variables. principle, the model of can, conditional","Thus, the proposed model can, in principle, represent arbitrary conditional probability distributions of output variables.",27385
We SportVu on two our datasets with three-pointer experiments NBA data. tested from model,We tested our model with two experiments on three-pointer datasets from NBA SportVu data.,27386
would shared also on code be web. The the,The code would also be shared on the web.,27387
on the experiment the be would The results codes reproducible; implemented are surely web. shared,The experiment results are surely reproducible; the implemented codes would be shared on the web.,27388
self-training procedure hyperspectral for data. use sequential classification concept the for the of of We,We use the procedure for the concept of self-training for sequential classification of hyperspectral data.,27389
"update incremental the from we re-training model IVM without Moreover, parameters scratch. the in","Moreover, we update the parameters in the incremental IVM model without re-training from scratch.",27390
"with is classifier deal sets. able data the incremental large Therefore, to","Therefore, the incremental classifier is able to deal with large data sets.",27391
IVM the demonstrate classification similar SVM results terms accuracy. in and that of Experimental perform,Experimental results demonstrate that the IVM and SVM perform similar in terms of classification accuracy.,27392
"increase accuracy. can proposed the In self-training the addition, classification strategy","In addition, the proposed self-training strategy can increase the classification accuracy.",27393
have this state-of-the-art developed Many been algorithms for purpose.,Many state-of-the-art algorithms have been developed for this purpose.,27394
influence will Severe and conditions distance. change creepage insulator environment surface,Severe environment conditions will influence insulator surface and change creepage distance.,27395
Standard utilized are filters feature Gabor extraction. deviation for and image,Gabor and Standard deviation filters are utilized for image feature extraction.,27396
problem this shows that research is Our NP-hard.,Our research shows that this problem is NP-hard.,27397
representation-based dictionary classification a approach sparse This presents construction a with novel procedure. paper,This paper presents a sparse representation-based classification approach with a novel dictionary construction procedure.,27398
representation obtained means classification of is by error. the The patch-based,The patch-based classification is obtained by means of the representation error.,27399
on three conducted hyperspectral are well-known datasets. image Experiments,Experiments are conducted on three well-known hyperspectral image datasets.,27400
toric to S-systems. propose solve to use We geometry algebraic,We propose to use toric algebraic geometry to solve S-systems.,27401
Salient seen learning has object deep techniques. detection driven by progress remarkable,Salient object detection has seen remarkable progress driven by deep learning techniques.,27402
features are to introduced these reuse connections long-range and iteratively. short- Dense,Dense short- and long-range connections are introduced to reuse these features iteratively.,27403
molecules endogenous gene small that (miRNAs) regulatory expression modulate MicroRNAs are post-transcriptionally.,MicroRNAs (miRNAs) are small endogenous regulatory molecules that modulate gene expression post-transcriptionally.,27404
basis pairs Many in of the disruption for have we flagged a cancer. identify biological,Many of the flagged pairs we identify have a biological basis for disruption in cancer.,27405
"leverage power by i.e. generative on of the networks, Moreover,","Moreover, by leverage on the power of generative networks, i.e.",27406
key the Occlusions remained of have one challenges.,Occlusions have remained one of the key challenges.,27407
"performance significant the on from benchmarks, We benefits symmetry. demonstrate occlusion-disocclusion standard especially","We demonstrate significant performance benefits on standard benchmarks, especially from the occlusion-disocclusion symmetry.",27408
to we the date. results most KITTI two-frame dataset On the report accurate challenging,On the challenging KITTI dataset we report the most accurate two-frame results to date.,27409
"cues. with two-task paper, integrating In bottom present novel a network this top we and","In this paper, we present a novel two-task network with integrating bottom and top cues.",27410
costly This annotation human effort. with a associated is,This is associated with a costly human annotation effort.,27411
"propose to image particular, colorization. automatic use self-supervised we In","In particular, we propose to use self-supervised automatic image colorization.",27412
"colorization develop we method. In automatic image an a for search fully alternative,","In search for an alternative, we develop a fully automatic image colorization method.",27413
"representation method us self-supervised it learning. a for Additionally, gives","Additionally, it gives us a method for self-supervised representation learning.",27414
This be a challenging turns open out problem. to,This turns out to be a challenging open problem.,27415
is in A* for optimal-cost paths finding best-first a graphs. search algorithm,A* is a best-first search algorithm for finding optimal-cost paths in graphs.,27416
"limited-memory to as approaches parallel, We IDA*. A* survey variants of parallel also such","We also survey approaches to parallel, limited-memory variants of A* such as parallel IDA*.",27417
"to is the architecture However, researchers. deep mystery a neural still the of capacity network","However, the capacity of deep neural network architecture is still a mystery to the researchers.",27418
whose dimensions into a so-called partitioned high-dimensional are domains. A is space space conceptual,A conceptual space is a high-dimensional space whose dimensions are partitioned into so-called domains.,27419
"Within Euclidean used the is distances. compute metric each domain, to","Within each domain, the Euclidean metric is used to compute distances.",27420
represented represented and space are in are this concepts regions. points by as Instances,Instances are represented as points in this space and concepts are represented by regions.,27421
image another in document separators analysis. text-lines one used Line are segregate to from,Line separators are used to segregate text-lines from one another in document image analysis.,27422
"a thrilling could exercise. particular, be the separators text handwritten identifying In in","In particular, identifying the separators in handwritten text could be a thrilling exercise.",27423
an computational would effort for text-line decompressing the a Such document of segmentation. prevent burden,Such an effort would prevent the computational burden of decompressing a document for text-line segmentation.,27424
"approach to both present problems. a Here, we novel","Here, we present a novel approach to both problems.",27425
treatment diagnosis significant and Automated analysis a image medical in has of lesions. value,Automated medical image analysis has a significant value in diagnosis and treatment of lesions.,27426
image modeling. has proposed results The promising recently generative shown training adversarial in,The recently proposed adversarial training has shown promising results in generative image modeling.,27427
tasks. for The both approach results the proposed superior ability of demonstrate the experimental,The experimental results demonstrate the superior ability of the proposed approach for both tasks.,27428
"However, enforcing unnecessary. expensive even sometimes and is PC computationally","However, enforcing PC is computationally expensive and sometimes even unnecessary.",27429
"guaranteeing search for fact, DPC* constraint for is backtrack-free sufficient such networks. In","In fact, DPC* is sufficient for guaranteeing backtrack-free search for such constraint networks.",27430
for show evaluations experimental that constraints. outperforms solving the algorithms significantly state-of-the-art majority-closed Our DPC*,Our experimental evaluations show that DPC* significantly outperforms the state-of-the-art algorithms for solving majority-closed constraints.,27431
"efficient sequence method an We of Eigen introduce Evolution aggregate feature Pooling, a to vectors.","We introduce Eigen Evolution Pooling, an efficient method to aggregate a sequence of feature vectors.",27432
denoising existing image have models many Most CNN layers. for,Most existing CNN models for image denoising have many layers.,27433
"paper, develop residual CNN a Gaussian we this dilated for denoising. In image","In this paper, we develop a dilated residual CNN for Gaussian image denoising.",27434
"formula Moreover, calculate size is dilated we incorporated. receptive present when field a to convolution","Moreover, we present a formula to calculate receptive field size when dilated convolution is incorporated.",27435
"interpreted of the mathematically. Thus, receptive be field change can","Thus, the change of receptive field can be interpreted mathematically.",27436
diagnosis. or inaccurate of cause The these late mortalities major is,The major cause of these mortalities is late or inaccurate diagnosis.,27437
method Manual for the microscopy dominant is as considered diagnostic currently malaria.,Manual microscopy is currently considered as the dominant diagnostic method for malaria.,27438
"consuming However, prone human is errors. time it to and","However, it is time consuming and prone to human errors.",27439
for a diagnostic the system. We hardware have developed software and malaria cost-efficient,We have developed the hardware and software for a cost-efficient malaria diagnostic system.,27440
from take Giemsa-stained motorized A blood is microscope developed to images smears.,A motorized microscope is developed to take images from Giemsa-stained blood smears.,27441
The proposed settings. method different better provides imaging robustness against,The proposed method provides better robustness against different imaging settings.,27442
core model The Component proposed algorithm of Analysis. a of the called Mixture Independent is,The core of the proposed algorithm is a model called Mixture of Independent Component Analysis.,27443
to video using This proposes captions. generation a length variable semantic perform paper network architecture,This paper proposes a network architecture to perform variable length semantic video generation using captions.,27444
domain in do dataset of autonomous We the driving famous the this benchmark on KITTI.,We do this in the domain of autonomous driving on the famous benchmark dataset KITTI.,27445
We environment. the to do map this,We do this to map the environment.,27446
has been this here novel to proposed address A based segmentation approach RANSAC issue.,A novel RANSAC based segmentation approach has been proposed here to address this issue.,27447
of results show potential the Our clinical future method. applications towards,Our results show potential towards future clinical applications of the method.,27448
vision detection numerous operation a applications. Corner is computer vital in,Corner detection is a vital operation in numerous computer vision applications.,27449
"in some detector our that experiment we demonstrate part, However, potential misses often CPDA corners.","However, in our experiment part, we demonstrate that CPDA detector often misses some potential corners.",27450
"the is Moreover, CPDA algorithm detection of costly. computationally","Moreover, the detection algorithm of CPDA is computationally costly.",27451
"as on reducing paper, this focus increasing repeatability. as localization error well In average We","In this paper, We focus on reducing localization error as well as increasing average repeatability.",27452
CPDA. and process preprocessing to of similar proposed steps are The refinements,The preprocessing and refinements steps of proposed process are similar to CPDA.,27453
process effectiveness the results show over CPDA. proposed will of robustness and experimental Our,Our experimental results will show the effectiveness and robustness of proposed process over CPDA.,27454
the are objects for employed single from depth These camera. steps deforming,These are the steps employed for deforming objects from single depth camera.,27455
"for annotations. weak this controlled we impact In different types of experiments, study","In controlled experiments, we study this impact for different types of weak annotations.",27456
are ambiguity. this two for There sources,There are two sources for this ambiguity.,27457
We and state-of-the-art on MALF for face demonstrate results detection FDDB also datasets.,We also demonstrate state-of-the-art results for face detection on FDDB and MALF datasets.,27458
"metric. attempts IQA Therefore, study efficient develop an this effective and to","Therefore, this study attempts to develop an effective and efficient IQA metric.",27459
local the image utilized quality. to two The characterize features proposed model these,The proposed model utilized these two features to characterize the image local quality.,27460
"model with this Furthermore, proposed efficient. compared more competing models, is IQA","Furthermore, compared with competing IQA models, this proposed model is more efficient.",27461
"Unlike we existing the operation. this achieve methods, unraveling forward via pass","Unlike existing methods, we achieve this via unraveling the forward pass operation.",27462
"CNN-Fixations, locations human eye to name analogous We these fixations. loosely","We name these locations CNN-Fixations, loosely analogous to human eye fixations.",27463
and between cue. to focus We psychophysical investigate the experiments foveation performed two relationship,We performed two psychophysical experiments to investigate the relationship between foveation and focus cue.,27464
investigation. the We two from studies results the and discuss suggest further,We discuss the results from the two studies and suggest further investigation.,27465
requires approach for only road time. at training present segmentation annotations We image-level that an,We present an approach for road segmentation that only requires image-level annotations at training time.,27466
"trade-off between cost we the Further, discuss annotation and performance.","Further, we discuss the trade-off between annotation cost and performance.",27467
"introduce technique paper, regularization learning. a we In new this for transfer","In this paper, we introduce a new regularization technique for transfer learning.",27468
"local we scale, On a use low-level geometric simple, invariants.","On a local scale, we use simple, low-level geometric invariants.",27469
"correspondences. covariant scale, constraints compatible a apply On finding for we global","On a global scale, we apply covariant constraints for finding compatible correspondences.",27470
dependencies collecting for We guide voters previous the stages. by on downward voting sampling,We guide the sampling for collecting voters by downward dependencies on previous voting stages.,27471
in procedure. together this accurate results matching All of an,All of this together results in an accurate matching procedure.,27472
"Optical due flow occlusions, estimation to more. areas, motion untextured remains boundaries, challenging and","Optical flow estimation remains challenging due to untextured areas, motion boundaries, occlusions, and more.",27473
"reliable the the across flow is equally estimated image. not Thus,","Thus, the estimated flow is not equally reliable across the image.",27474
that promising and is results show The encouraging. method,The results show that method is promising and encouraging.,27475
in field. in has been surveillance research Frequent occlusion problem major a traffic this,Frequent occlusion in traffic surveillance has been a major problem in this research field.,27476
"large, repositories. but incomplete are graphs knowledge useful, Knowledge","Knowledge graphs are large, useful, but incomplete knowledge repositories.",27477
instances ones. negative contrasting these on relies representations with Learning positive,Learning these representations relies on contrasting positive instances with negative ones.,27478
sampling propose based and additionally negative for We methods. known well sampling embedding methods compare,We compare well known methods for negative sampling and additionally propose embedding based sampling methods.,27479
a method this apply in the Bayesian to then parameters of network. We estimation,We then apply this method to the estimation of parameters in a Bayesian network.,27480
real method demonstrate our challenging images. We synthetic and on,We demonstrate our method on challenging real and synthetic images.,27481
compactly is defined as functions combination The radial a supported basis linear of (CSRBFs). function,The function is defined as a linear combination of compactly supported radial basis functions (CSRBFs).,27482
and Examples are between set operations shapes provided. of reconstruction surface,Examples of surface reconstruction and set operations between shapes are provided.,27483
for This hierarchical image analysis. spatiotemporal orientation novel spacetime a representation presents paper,This paper presents a novel hierarchical spatiotemporal orientation representation for spacetime image analysis.,27484
"layers, is network To manageable novel cross-channel pooling feature proposed. keep the a size across","To keep the network size manageable across layers, a novel cross-channel feature pooling is proposed.",27485
datasets shows Empirical sets standard on a state-of-the-art. multiple evaluation that new it,Empirical evaluation on multiple standard datasets shows that it sets a new state-of-the-art.,27486
representations. transformations to the relate of of We transformations images,We relate transformations of images to transformations of the representations.,27487
Genotypes sequences represented of are $L$ loci. as binary,Genotypes are represented as sequences of $L$ binary loci.,27488
interactions subsets the determine the These genetic neighborhoods model. of or,These subsets or neighborhoods determine the genetic interactions of the model.,27489
"approach learning paper, In a problem. to deep based mitigate this was proposed this","In this paper, a deep learning based approach was proposed to mitigate this problem.",27490
a defining relations by space of achieved descriptor. neighborhood which the is describe feature This,This is achieved by defining feature space relations which describe the neighborhood of a descriptor.,27491
and controlled in experiments our both real-life present scenarios We validate to approach.,We present experiments both in controlled and real-life scenarios to validate our approach.,27492
Contexts important saliency detection task. in the role an play,Contexts play an important role in the saliency detection task.,27493
"contextual the not for However, a helpful is region, information all task. given context final","However, given a context region, not all contextual information is helpful for the final task.",27494
An can aggregating feature be the contextual constructed selectively by information. then attended contextual,An attended contextual feature can then be constructed by selectively aggregating the contextual information.,27495
training. into Both be embedded can fully models joint for are and differentiable CNNs,Both models are fully differentiable and can be embedded into CNNs for joint training.,27496
We objects. models with proposed the to detect also U-Net salient incorporate the architecture,We also incorporate the proposed models with the U-Net architecture to detect salient objects.,27497
improve proposed PiCANets performance. can consistently show Extensive experiments detection saliency the that,Extensive experiments show that the proposed PiCANets can consistently improve saliency detection performance.,27498
"facilitate contrast learning global and The and PiCANets local respectively. homogeneousness, global","The global and local PiCANets facilitate learning global contrast and homogeneousness, respectively.",27499
"projective between describes algebra geometry. contribution relationship fractions, representation, This linear and duality,","This contribution describes relationship between fractions, projective representation, duality, linear algebra and geometry.",27500
of problems Many equations. system to lead linear a,Many problems lead to a system of linear equations.,27501
via final the is clustering simple detected cysts segmentation obtained of locations. of cyst The,The final segmentation of cysts is obtained via simple clustering of the detected cyst locations.,27502
dataset. method proposed two private and one on is public evaluated The datasets,The proposed method is evaluated on two public datasets and one private dataset.,27503
"Overall, proposed all benchmarks. outperforms the system","Overall, the proposed system outperforms all benchmarks.",27504
"as organized hierarchy visual are Imagery texts usually of a elements, several i.e.","Imagery texts are usually organized as a hierarchy of several visual elements, i.e.",27505
"text text blocks. lines words, characters, and","characters, words, text lines and text blocks.",27506
"existing annotated level. are line in word mostly the datasets real Actually, text or","Actually, the existing real text datasets are mostly annotated in word or line level.",27507
achieves benchmarks. state-of-the-art several It performance text scene detection challenging on the,It achieves the state-of-the-art performance on several challenging scene text detection benchmarks.,27508
provided against methods the Comparative previous datasets. analysis on is evaluated competition,Comparative analysis is provided against methods evaluated on the previous competition datasets.,27509
"techniques. associated of software However, the these implementing a there and is inferential lack distributions","However, there is a lack of software implementing these distributions and the associated inferential techniques.",27510
reduced greater predictive may enable or parsimonious analyses and capability. modeling Such,Such analyses may enable reduced or parsimonious modeling and greater predictive capability.,27511
and synthetic and on example. illustrated computational properties Theoretical are considered a,Theoretical and computational properties are considered and illustrated on a synthetic example.,27512
is sufficient realized by uniformly Real-time intensity with image from pixels gradient. regions sampling optimization,Real-time optimization is realized by sampling pixels uniformly from image regions with sufficient intensity gradient.,27513
vision. language Networks in processing computer widely Neural (RNNs) natural been Recurrent used and have,Recurrent Neural Networks (RNNs) have been widely used in natural language processing and computer vision.,27514
algorithms. wavelet at The can be found many discrete heart the of transform image-processing,The discrete wavelet transform can be found at the heart of many image-processing algorithms.,27515
"Consequently, often these form performance bottleneck. a points","Consequently, these points often form a performance bottleneck.",27516
"rearranges and the of inside number steps. approach appropriately thereby Our calculations transform, reduces the","Our approach appropriately rearranges calculations inside the transform, and thereby reduces the number of steps.",27517
"parallel words, to In friendly environments. propose that is a new we scheme other","In other words, we propose a new scheme that is friendly to parallel environments.",27518
"we evaluating multi-core the consistently original CPUs, lifting on scheme. When overcome","When evaluating on multi-core CPUs, we consistently overcome the original lifting scheme.",27519
"a scratch Instead, execution using set from GRP of trained traces. is successful the","Instead, the GRP is trained from scratch using a set of successful execution traces.",27520
the science. of hamper disrupt forward thorough These thus and progress peer-review obstructions,These obstructions hamper thorough peer-review and thus disrupt the forward progress of science.,27521
The and an the notches along nicks trailing unique signature. define individual's edge,The nicks and notches along the trailing edge define an individual's unique signature.,27522
explore ranking two this We methods representation. that use,We explore two ranking methods that use this representation.,27523
of This algorithm to matches curvature. regions learned exploits spatial downweight unstable from also weights,This algorithm also exploits learned spatial weights to downweight matches from regions of unstable curvature.,27524
a the feature interprets second representation The descriptor. as,The second interprets the representation as a feature descriptor.,27525
the of are at keypoints local defined Feature extrema the representation.,Feature keypoints are defined at the local extrema of the representation.,27526
scene understanding problem (SFSU). of addresses the foggy This semantic work,This work addresses the problem of semantic foggy scene understanding (SFSU).,27527
"of study we carefully SFSU. In addition, image the for usefulness dehazing","In addition, we carefully study the usefulness of image dehazing for SFSU.",27528
"code publicly The datasets, models made are available. and","The datasets, models and code are made publicly available.",27529
"considered. The common for standby that two and are active redundancies, allocations, is schemes","The two common schemes for allocations, that is active and standby redundancies, are considered.",27530
"proposed to simple are If lifetimes the approach independent, component leads the manipulations.","If the component lifetimes are independent, the proposed approach leads to simple manipulations.",27531
engineering can practical implemented be This method systems. complex for,This method can be implemented for practical complex engineering systems.,27532
improve algorithms. feature information leverage We pre-calculated this to descriptor,We leverage this pre-calculated information to improve feature descriptor algorithms.,27533
"descriptor This computations. feature to each way, change will significant more without be scale robust","This way, each feature descriptor will be more robust to scale change without significant computations.",27534
"stereo to we In images descriptor construct use vector. addition, the","In addition, we use stereo images to construct the descriptor vector.",27535
has proposed. higher name ranks method improve approaches traditional been of to taxa to Simple,Simple method to improve traditional approaches to name taxa of higher ranks has been proposed.,27536
various detection radiology from or Segmenting images requires objects of contours. pathology accurate images,Segmenting various objects from radiology images or pathology images requires accurate detection of contours.,27537
"insufficiently and However, such are CNNs. in as problems, constraints, some shape studied discontinuity","However, some problems, such as discontinuity and shape constraints, are insufficiently studied in CNNs.",27538
is further to to necessary exploration. challenges clarify encourage the It,It is necessary to clarify the challenges to encourage further exploration.,27539
CNN of on based detection relies contour performance architectures. CNN the The state-of-the-art,The performance of CNN based contour detection relies on the state-of-the-art CNN architectures.,27540
Careful detection. of design motivations to principles and critical investigation contour is beneficial and their,Careful investigation of their design principles and motivations is critical and beneficial to contour detection.,27541
"their strengthens and We detail, weaknesses. in compare those methods clarify","We compare those methods in detail, clarify their strengthens and weaknesses.",27542
on their common style based a develop between Humans items sense attributes. compatibility of,Humans develop a common sense of style compatibility between items based on their attributes.,27543
it style from learn is descriptions. feasible Therefore these compatibility to,Therefore it is feasible to learn style compatibility from these descriptions.,27544
on different powers different imposes smoothing VTV features (e.g. The,The VTV imposes different smoothing powers on different features (e.g.,27545
parameters. based cartoons) regularization on and various edges choosing,edges and cartoons) based on choosing various regularization parameters.,27546
"Thus, remove and preserve the simultaneously edges can noises. model","Thus, the model can simultaneously preserve edges and remove noises.",27547
mainly transfer Traditional data on not focus methods and are applicable. numerical learning,Traditional transfer learning methods mainly focus on numerical data and are not applicable.,27548
target with statistically correlations are to Only domain. transferred high the entities the,Only the entities with statistically high correlations are transferred to the target domain.,27549
real-world of and results demonstrate efficiency on synthetic ACRET. experimental The and effectiveness the datasets,The experimental results on synthetic and real-world datasets demonstrate the effectiveness and efficiency of ACRET.,27550
ACRET for real We apply a intrusion security system enterprise detection. to also,We also apply ACRET to a real enterprise security system for intrusion detection.,27551
(BA) are computer methods either incrementally. for centralized Adjustment Bundle vision in operate Most or,Most methods for Bundle Adjustment (BA) in computer vision are either centralized or operate incrementally.,27552
"must acquisition processing where they and be be scenarios in used cannot Furthermore, distributed. image","Furthermore, they cannot be used in scenarios where image acquisition and processing must be distributed.",27553
We with distributed a algorithm. this BA new address problem,We address this problem with a new distributed BA algorithm.,27554
scales with the The implementation observed linearly our number of points. runtime of,The runtime of our implementation scales linearly with the number of observed points.,27555
determined by behavior and functional Cardiac features. is structural,Cardiac behavior is determined by structural and functional features.,27556
We combine train a to network. tasks then jointly,We then combine tasks to jointly train a network.,27557
"prediction. depth on NYU all PASCAL We detection, methods ImageNet classification, evaluate and VOC","We evaluate all methods on ImageNet classification, PASCAL VOC detection, and NYU depth prediction.",27558
on does not that assume We f equally depend all its arguments.,We assume that f does not depend equally on all its arguments.,27559
study volatility time-varying applications to forecasting We queues. and,We study applications to volatility forecasting and time-varying queues.,27560
for recognition. research human applicable motion directly This technology makes capture,This research makes motion capture technology directly applicable for human recognition.,27561
"motion features important cannot low-level discriminatory estimated. such as optical In consequence, be from flow","In consequence, important discriminatory low-level features from motion such as optical flow cannot be estimated.",27562
"types integrate However, data it to is omic challenging heterogeneous directly.","However, it is challenging to integrate heterogeneous omic data types directly.",27563
use A invariance translation augmented of data. is important of more training factor the,A more important factor of translation invariance is the use of augmented training data.,27564
These show data of results domain-specific importance an augmentation.,These results show an importance of domain-specific data augmentation.,27565
"actions of situations, save are In the limit lives emergency hazards and impact that crucial.","In emergency situations, actions that save lives and limit the impact of hazards are crucial.",27566
"do. In situational what is to needed act, decide order to to awareness","In order to act, situational awareness is needed to decide what to do.",27567
"poses However, big information this data in overflow. and challenges","However, this poses challenges in big data and information overflow.",27568
"excellent support. tools decision vision be can Consequently, computer an","Consequently, computer vision tools can be an excellent decision support.",27569
"Logistic of learning variant is ""soft"" the perceptron learning.","Logistic learning is the ""soft"" variant of perceptron learning.",27570
are evaluated datasets: two SAMM. II The on CASME benchmark coded and experiments FACS,The experiments are evaluated on two benchmark FACS coded datasets: CASME II and SAMM.,27571
uses keypoints keypoint method standard detectors. from The,The method uses keypoints from standard keypoint detectors.,27572
are from ordered distance the keypoints. with the respect to points The,The points are ordered with respect to the distance from the keypoints.,27573
signature the representing vector used keypoints. The constituted as a thus is for,The vector thus constituted is used as a signature for representing the keypoints.,27574
is by The hamming distance. done matching using,The matching is done by using hamming distance.,27575
models review generative Here for and the prospects neuroscience. we of promise network,Here we review the prospects and promise of generative models for network neuroscience.,27576
model is stereo on which algorithm of based works minimization This worldwide matching energy. a,This algorithm works a worldwide matching stereo model which is based on minimization of energy.,27577
"terms, term. term global The and two smoothness comprises firstly energy the the data secondly","The global energy comprises two terms, firstly the data term and secondly the smoothness term.",27578
method superiority. prediction to We the its a on real-fake evaluate demonstrate expression dataset,We evaluate the method on a real-fake expression prediction dataset to demonstrate its superiority.,27579
"input Firstly, map is material the image. the extracted from","Firstly, the material map is extracted from the input image.",27580
"input constrained decomposition. relit is image layer using material Lastly, the","Lastly, the input image is relit using material constrained layer decomposition.",27581
to mode strategies is develop solution of treatment action. alternative A potential with an,A potential solution is to develop treatment strategies with an alternative mode of action.,27582
therapy. consider anti-adhesion one We such strategy:,We consider one such strategy: anti-adhesion therapy.,27583
model two and exudate. The the the of compartments: epithelium physical consists,The model consists of two physical compartments: the epithelium and the exudate.,27584
"by mathematical Lastly, to as model. our predicted improve we efficacy, present ways therapeutic","Lastly, we present ways to improve therapeutic efficacy, as predicted by our mathematical model.",27585
observed have been also MVP-ANM. in Similar properties,Similar properties have also been observed in MVP-ANM.,27586
great MVP-ANM of the potential data. deformation demonstrates analysis resolution This the in of low,This demonstrates the great potential of MVP-ANM in the deformation analysis of low resolution data.,27587
"the conclusion. Finally, ends paper a with","Finally, the paper ends with a conclusion.",27588
many the of datasets. now By perform than better human they subjects on image classification,By now they perform better than human subjects on many of the image classification datasets.,27589
notion the Most classes based on of of are datasets concrete (i.e. these,Most of these datasets are based on the notion of concrete classes (i.e.,27590
are image). classified in object by the images type of the,images are classified by the type of object in the image).,27591
our tasks. natural outstanding real-world We demonstrate performance processing of on framework the language,We demonstrate the outstanding performance of our framework on real-world natural language processing tasks.,27592
"asymmetric metric we for learning RE-ID. the overcome propose unsupervised unsupervised problem, To","To overcome the problem, we propose unsupervised asymmetric metric learning for unsupervised RE-ID.",27593
binarization way effective has to shown network be for an Input acceleration.,Input binarization has shown to be an effective way for network acceleration.,27594
method. approximation Theoretical proposed of analysis guarantee shows error property,Theoretical analysis shows approximation error guarantee property of proposed method.,27595
experimental proposed yields that results while being accelerated. Extensive scheme demonstrate accuracy great recognition the,Extensive experimental results demonstrate that the proposed scheme yields great recognition accuracy while being accelerated.,27596
transportation driving important and Analysis of styles profoundly to calibration. recognition and are intelligent vehicle,Analysis and recognition of driving styles are profoundly important to intelligent transportation and vehicle calibration.,27597
to incorporate in motivates information abundance genotype inference. This strategies phylogenetic,This motivates strategies to incorporate genotype abundance information in phylogenetic inference.,27598
of crucial vision has recognition machine and part become Object recently. computer learning a,Object recognition has become a crucial part of machine learning and computer vision recently.,27599
"make modify interpret CNNs programmatically explicitly their intuitively it Additionally, difficult learned representations. to or","Additionally, CNNs make it difficult to explicitly programmatically modify or intuitively interpret their learned representations.",27600
and two Our each attributes: into color image model decomposes distribution. shape,Our model decomposes each image into two attributes: shape and color distribution.,27601
probabilistically We each a Bayesian criterion to determine use the likelihood category. then of,We then use a Bayesian criterion to probabilistically determine the likelihood of each category.,27602
available. are the on dataset made the and trained models dataset publicly Both,Both the dataset and models trained on the dataset are made publicly available.,27603
online method open-source. The in available is,The method is available online in open-source.,27604
"a is approach, inserted this the In map. the as into ROI net binary","In this approach, the ROI is inserted into the net as a binary map.",27605
"is This particularly real-world true on distorted complex, images.","This is particularly true on complex, real-world distorted images.",27606
"scores. assigned of Moreover, images distributions may subjective in their differ broadly","Moreover, images may broadly differ in their distributions of assigned subjective scores.",27607
or changes. contain level structural often series Time outliers shifts and,Time series often contain outliers and level shifts or structural changes.,27608
approach squares. the ideas robust for least regression The from algorithm FastLTS combines alternating with,The approach combines ideas from the FastLTS algorithm for robust regression with alternating least squares.,27609
testing was results Generalization and evident dataset on one by other. of the training on,Generalization of results was evident by training on one dataset and testing on the other.,27610
as confusion facial humans between expressions General were also exhibited by different trends observed.,General confusion trends between different facial expressions as exhibited by humans were also observed.,27611
"dependencies. exploit are temporal implemented networks, the used to with memory long model, Recursive short-term","Recursive networks, implemented with the long short-term memory model, are used to exploit temporal dependencies.",27612
this on of results videos potential forged show approach. the Preliminary,Preliminary results on forged videos show the potential of this approach.,27613
"The contains and proposed analysis, phases: three merge, method boundary link. local UF-based","The proposed method contains three phases: UF-based local merge, boundary analysis, and link.",27614
"fill gap. is library introduce ChainerCV, a intended software that this to We","We introduce ChainerCV, a software library that is intended to fill this gap.",27615
"good software practices. emphasize implementations simplicity, engineering and flexibility These","These implementations emphasize simplicity, flexibility and good software engineering practices.",27616
and numerous applications related These statistics e.g. fields; methods have in found,These methods have found numerous applications in statistics and related fields; e.g.,27617
"inference in state non-Gaussian in complex static space for and models. models, non-linear","for inference in non-linear non-Gaussian state space models, and in complex static models.",27618
dimerized alpha coiled helical are monomers zippers a Leucine structure. to coil,Leucine zippers are alpha helical monomers dimerized to a coiled coil structure.,27619
Phylogenetic typically trees a bifurcating as history species speciation the tree. represent,Phylogenetic species trees typically represent the speciation history as a bifurcating tree.,27620
"as to resolve the (soft) Moreover, a often is inability shown relationships polytomy.","Moreover, the inability to resolve relationships is often shown as a (soft) polytomy.",27621
"performed group classification process setup. each is within a though, The learning multi-task in","The learning process within each group though, is performed in a multi-task classification setup.",27622
"to strategy least, we a Last improve introduce novel not the image but quality.","Last but not least, we introduce a novel strategy to improve the image quality.",27623
multiple detection. used can obstacle be for such navigation They purposes and as visual,They can be used for multiple purposes such as visual navigation and obstacle detection.,27624
"this can otherwise way, lead to spots we which avoid blind In accidents.","In this way, we avoid blind spots which can otherwise lead to accidents.",27625
"minimize for needed the fisheye perception, cameras surround of utilize number cameras. we To","To minimize the number of cameras needed for surround perception, we utilize fisheye cameras.",27626
"In fisheye of be to supported. has images processing addition,","In addition, processing of fisheye images has to be supported.",27627
project This enable for automated parking cars. valet to self-driving seeks,This project seeks to enable automated valet parking for self-driving cars.,27628
"account must of information for the unitary nature be consciousness. to also However, integrated","However, information must also be integrated to account for the unitary nature of consciousness.",27629
method the Results that recent most techniques. images our outperforms show on various,Results on various images show that our method outperforms the most recent techniques.,27630
"summarization i.e., SumMe, Extensive experiments and on TVSum. conducted three video datasets, are benchmark","Extensive experiments are conducted on three video summarization benchmark datasets, i.e., SumMe, and TVSum.",27631
reports an other important of physicians. radiologists Radiology communication means are between and,Radiology reports are an important means of communication between radiologists and other physicians.,27632
image. the insight normalization is parameters should to adapt input that Our key,Our key insight is that normalization parameters should adapt to the input image.,27633
"in the complicates problem. presence images the addition, of the further In MR inhomogeneity intensity","In addition, the presence of intensity inhomogeneity in the MR images further complicates the problem.",27634
coming sensor dispensable different of is kind cloud) and demanding. great from shows point,point cloud) coming from different kind of sensor is dispensable and shows great demanding.,27635
"transformation. network data be Then, augmentation extend rotation we our robust to use to","Then, we use data augmentation to extend our network be robust to rotation transformation.",27636
control tissue during development and networks Gene (GRNs) and function cellular regulatory homeostasis. making decision,Gene regulatory networks (GRNs) control cellular function and decision making during tissue development and homeostasis.,27637
"of and aid are that complex methods this reason, For simplify necessary. networks exploration","For this reason, methods that simplify and aid exploration of complex networks are necessary.",27638
form applicable projection. we a Zwanzig-Mori develop end broadly this To of the,To this end we develop a broadly applicable form of the Zwanzig-Mori projection.,27639
"revealing of probes near-steady otherwise. accessible easily dynamics, provide These functions memory not state information","These memory functions provide probes of near-steady state dynamics, revealing information not easily accessible otherwise.",27640
recognition This for videos. proposes a in flow-guided convolutional attention action networks paper two-stream,This paper proposes a two-stream flow-guided convolutional attention networks for action recognition in videos.,27641
data. ability be the understood image to perform inference Computer as on can vision,Computer vision can be understood as the ability to perform inference on image data.,27642
in technology advances by Breakthroughs marked inference computer are in vision techniques. often,Breakthroughs in computer vision technology are often marked by advances in inference techniques.,27643
computer schemes proposes inference in thesis applications and This demonstrates novel vision.,This thesis proposes novel inference schemes and demonstrates applications in computer vision.,27644
both discriminative propose vision We for and inference models. generative techniques,We propose inference techniques for both generative and discriminative vision models.,27645
propose for message-passing inference. sampling inference in We techniques and improving MCMC,We propose techniques for improving inference in MCMC sampling and message-passing inference.,27646
of prior design of knowledge. complication the discriminative the inclusion main models A is in,A main complication in the design of discriminative models is the inclusion of prior knowledge.,27647
"the importance unexplored. areas However, or of sparse of the remains gaps, knowledge network,","However, the importance of sparse areas of the network, or knowledge gaps, remains unexplored.",27648
part of is approach network. the CNN-based a The VGG core revised,The core part of the CNN-based approach is a revised VGG network.,27649
"expression facial propose we the CDAAE, adversarial autoencoder, paper, synthesis. difference this conditional In for","In this paper, we propose the conditional difference adversarial autoencoder, CDAAE, for facial expression synthesis.",27650
databases. advantageous is especially training small This with when,This is especially advantageous when training with small databases.,27651
into a a of the population. fixation We type study introduced probability mutant resident when,We study the fixation probability of a mutant type when introduced into a resident population.,27652
model. is a competitive by stochastic This Lotka-Volterra implemented,This is implemented by a stochastic competitive Lotka-Volterra model.,27653
type Higher competing of resident {\phi}. higher when yield with payoffs the values,Higher payoffs when competing with the resident type yield higher values of {\phi}.,27654
amount imbalanced has for research classification. of A effort been to boosting large adapting dedicated,A large amount of research effort has been dedicated to adapting boosting for imbalanced classification.,27655
solution framework on stages. programming The which linear requires lexicographic is a based two,The solution is based on a lexicographic linear programming framework which requires two stages.,27656
"derive framework. also formulation the to corresponding Additionally,we dual proposed the","Additionally,we also derive the dual formulation corresponding to the proposed framework.",27657
"remains challenging generate disparities it high-quality to the However, regions. for inherently ill-posed","However, it remains challenging to generate high-quality disparities for the inherently ill-posed regions.",27658
"CNN To of this we stages. problem, composing architecture cascade novel propose two a tackle","To tackle this problem, we propose a novel cascade CNN architecture composing of two stages.",27659
stages from the of the final The summation outputs gives two disparity. the,The summation of the outputs from the two stages gives the final disparity.,27660
"cascade also training the Moreover, of the it benefits overall network.","Moreover, it also benefits the training of the overall cascade network.",27661
shows scheme our that stereo performance matching residual learning for correspondence. cascade state-of-the-art Experimentation provides,Experimentation shows that our cascade residual learning scheme provides state-of-the-art performance for matching stereo correspondence.,27662
the in is widely most world. the language used Chinese,Chinese is the most widely used language in the world.,27663
of facilitate text images Chinese natural in Algorithms read applications kinds. various that,Algorithms that read Chinese text in natural images facilitate applications of various kinds.,27664
"reading. on focuses competition that Chinese a This RCTW, text new introduces report","This report introduces RCTW, a new competition that focuses on Chinese text reading.",27665
"are set tasks, end-to-end up. text localization namely recognition, and Two","Two tasks, namely text localization and end-to-end recognition, are set up.",27666
"definitions, results This and includes protocols, evaluation report analysis. dataset and description, task summaries","This report includes dataset description, task definitions, evaluation protocols, and results summaries and analysis.",27667
"problem. Chinese future reading research call on text competition, Through this we more for the","Through this competition, we call for more future research on the Chinese text reading problem.",27668
"action methods On our dataset, in large-scale ActivityNet, classification. outperforms video model competing a youtube","On a large-scale youtube video dataset, ActivityNet, our model outperforms competing methods in action classification.",27669
from can descriptive (GANs) images adversarial approaches Recent realistic networks in generative synthesize automatically text.,Recent approaches in generative adversarial networks (GANs) can automatically synthesize realistic images from descriptive text.,27670
Each a face. provides probability posterior for estimated of ages distribution a,Each posterior provides a probability distribution of estimated ages for a face.,27671
paradigm. imbalance information represented by can using learning privileged (LUPI) be This the,This imbalance can be represented by the learning using privileged information (LUPI) paradigm.,27672
effectiveness the facial keypoint. keypoint each framework of presented The is detection for,The effectiveness of the facial keypoint detection framework is presented for each keypoint.,27673
results The results. our proposed method achieved show experimental superior,The experimental results show our proposed method achieved superior results.,27674
Monte conditional on based first Carlo. method is The,The first method is based on conditional Monte Carlo.,27675
estimators We and present them of errors. most four have relative that bounded show,We present four estimators and show that most of them have bounded relative errors.,27676
issues and both of illustrate implementation through procedures the accuracy We studies. numerical discuss,We discuss implementation issues and illustrate the accuracy of both procedures through numerical studies.,27677
"However, some has still capsule robotic endoscopy challenges.","However, robotic capsule endoscopy still has some challenges.",27678
"pig For the a employed. is real stomach evaluations, qualitative","For the qualitative evaluations, a real pig stomach is employed.",27679
task one are restoration Previous designed methods specific HSI only. for,Previous HSI restoration methods are designed for one specific task only.,27680
Hamiltonian Monte estimators. a methodology to Carlo parallelize propose We,We propose a methodology to parallelize Hamiltonian Monte Carlo estimators.,27681
so chains can are that These estimators unbiased. be combined then resulting,These chains can then be combined so that resulting estimators are unbiased.,27682
of We dimensions high coupling toy in on investigate our a scalability the example.,We investigate the scalability of our coupling in high dimensions on a toy example.,27683
the arise in Such routinely as values variable models. in of products latent statistics likelihood,Such products routinely arise in statistics as values of the likelihood in latent variable models.,27684
"ecosystem interactions large-scale vegetation Short-scale in shape yield patterns turn, function that, across landscapes.","Short-scale interactions yield large-scale vegetation patterns that, in turn, shape ecosystem function across landscapes.",27685
in and In event detail. document report result this we some the,In this report we document the event and result in some detail.,27686
intermediary In verification is this multi-level on co-occurrences content studied. based method paper the visual,In this paper the intermediary visual content verification method based on multi-level co-occurrences is studied.,27687
remote paper problem photoplethysmography the in of research (rPPG). reproducible This studies,This paper studies the problem of reproducible research in remote photoplethysmography (rPPG).,27688
implement this network. convolutional encoder-decoder We as idea neural an,We implement this idea as an encoder-decoder convolutional neural network.,27689
this we In address after transfer. of work domain networks compression the,In this work we address the compression of networks after domain transfer.,27690
algorithms We focus on matrix decomposition. low-rank based compression on,We focus on compression algorithms based on low-rank matrix decomposition.,27691
"work, we evaluation obtained of In this emphasize extensive results. on","In this work, we emphasize on extensive evaluation of obtained results.",27692
discriminative Most discover regions to generally better performance. methods learn categorization existing detectors part for,Most existing methods generally learn part detectors to discover discriminative regions for better categorization performance.,27693
could us textual image. So attention in help to attention discover visual,So textual attention could help us to discover visual attention in image.,27694
performance. Residual segmentation multi-scale boost connections in networks employed to the predictions are these and,Residual connections and multi-scale predictions are employed in these networks to boost the segmentation performance.,27695
Visual plays with information interactions daily an in role our indispensable environment.,Visual information plays an indispensable role in our daily interactions with environment.,27696
from how understand single we a image a frame? motion,how we understand a motion from a single image frame?,27697
"a congruence, perception. Leveraging linear set motion transformation between determination of facilitate mapping on shapes","Leveraging on congruence, linear mapping transformation determination between a set of shapes facilitate motion perception.",27698
recent discoveries study scene understanding. Present methodizes for ability cognitive human of,Present study methodizes recent discoveries of human cognitive ability for scene understanding.,27699
find in transformation. linear mapping to graph fashion best-first best The then traversed is,The graph is then traversed in best-first fashion to find best linear mapping transformation.,27700
"method proposed assessed images. is accuracy and of deformed the using The noisy, normal,","The accuracy of the proposed method is assessed using normal, noisy, and deformed images.",27701
in a model works This in coarse-to-fine manner. single essentially a that results,This results in a single model that essentially works in a coarse-to-fine manner.,27702
"aggregates hierarchical which efficiently module multi-scale develop a inception features. Furthermore, inception we","Furthermore, we develop a hierarchical inception module which efficiently aggregates multi-scale inception features.",27703
"Herein, we survey research. most state-of-the-art of comprehensive the DL provide RS","Herein, we provide the most comprehensive survey of state-of-the-art RS DL research.",27704
"for and the community. challenges we Namely, on theories, focus RS tools","Namely, we focus on theories, tools and challenges for the RS community.",27705
theory. and The disintegration notions Bayesian in probability fundamental conditional inversion of are,The notions of disintegration and Bayesian inversion are fundamental in conditional probability theory.,27706
results theory. resulting are abstract in for proving conditional probability The basic descriptions used,The resulting abstract descriptions are used for proving basic results in conditional probability theory.,27707
"of the inversion Finally, disintegration several Bayesian usefulness and in is examples. illustrated","Finally, the usefulness of disintegration and Bayesian inversion is illustrated in several examples.",27708
resonance this to magnetic use we functional imaging idea. Here test (fMRI),Here we use functional magnetic resonance imaging (fMRI) to test this idea.,27709
"the First, facial not detected all of landmarks are under accurate occlusions.","First, not all of the detected facial landmarks are accurate under occlusions.",27710
"enough of not the number be problem. constrain reliable landmarks may to Second, the","Second, the number of reliable landmarks may not be enough to constrain the problem.",27711
propose additional to create points to a hypotheses. pose synthesize (SensiblePoints) method We,We propose a method to synthesize additional points (SensiblePoints) to create pose hypotheses.,27712
of small One imagery. are is sensing in difficulties such objects and remote that crowded,One of such difficulties is that objects are small and crowded in remote sensing imagery.,27713
aggregating factor. local LFE The proposed features solves this decreasing with dilation by module problem,The proposed LFE module solves this problem by aggregating local features with decreasing dilation factor.,27714
applications Economics. regularization data-rich and alpha-norm in has Marketing Sparse many,Sparse alpha-norm regularization has many data-rich applications in Marketing and Economics.,27715
"to solution. Alpha-norm, jumps sparse to and a regularization, contrast in lasso ridge","Alpha-norm, in contrast to lasso and ridge regularization, jumps to a sparse solution.",27716
forecasting. for high-dimensional attractive and is that problems ultra demand This estimation feature in occur,This feature is attractive for ultra high-dimensional problems that occur in demand estimation and forecasting.,27717
"price, and equivalized effects. promotion, flavor, scent, The include volume, demand of brand key predictors","The key predictors of demand include price, equivalized volume, promotion, flavor, scent, and brand effects.",27718
"research. future Finally, conclude we for directions with","Finally, we conclude with directions for future research.",27719
performed without Training is of ground-truth maps. disparity the fashion need in end-to-end an,Training is performed in an end-to-end fashion without the need of ground-truth disparity maps.,27720
Our different well different is unseen as settings. camera self-adaptive imageries as to to network,Our network is self-adaptive to different unseen imageries as well as to different camera settings.,27721
"However, related tasks issues pathological have be digital to considered. some and images","However, digital pathological images and related tasks have some issues to be considered.",27722
scene two There depth. extracting to are parts,There are two parts to extracting scene depth.,27723
"the data. on of success large their labeled However, amount relies a presence of","However, their success relies on the presence of a large amount of labeled data.",27724
The performance tasks. challenging on achieves proposed state-of-the-art model several,The proposed model achieves state-of-the-art performance on several challenging tasks.,27725
"images. degrade could processing the transmission, stereo of Compression, quality and","Compression, transmission, and processing could degrade the quality of stereo images.",27726
"two fusion inspired fused paper, stereo by process, sensory/motor images this are brain together. In","In this paper, inspired by brain sensory/motor fusion process, two stereo images are fused together.",27727
from extracted. Then two image every images synthesized are fused,Then from every fused image two synthesized images are extracted.,27728
synthesized different of are on distributions Effects of distortions the statistical images shown.,Effects of different distortions on statistical distributions of the synthesized images are shown.,27729
"Based images. these from statistical features the extracted changes, on are observed synthesized","Based on the observed statistical changes, features are extracted from these synthesized images.",27730
reveal severity features can type distortions. and of These,These features can reveal type and severity of distortions.,27731
processing. Semantic fundamental sensing is remote in segmentation image a research,Semantic segmentation is a fundamental research in remote sensing image processing.,27732
"complex a the maritime environment, Because sea-land task. of is segmentation challenging the","Because of the complex maritime environment, the sea-land segmentation is a challenging task.",27733
neural named network This convolution proposes a DeepUNet. novel deep paper,This paper proposes a novel deep convolution neural network named DeepUNet.,27734
Plus blocks are two novel bring The that and connections U-connection connection. two new,The two novel blocks bring two new connections that are U-connection and Plus connection.,27735
They are results. get to more promoted precise segmentation,They are promoted to get more precise segmentation results.,27736
common augmentation. to the lack data of approach dataset training is alleviate One,One common approach to alleviate the lack of training data is dataset augmentation.,27737
"obtained of our dataset artificial by ARIGAN. plants propose images, Ax We new a","We propose a new Ax dataset of artificial plants images, obtained by our ARIGAN.",27738
"customers situation, under-charged. a a that are or busy be mistake there overcharged In can","In a busy situation, there can be a mistake that customers are overcharged or under-charged.",27739
Our method consists ellipse network. convol-utional of fitting and neural,Our method consists of ellipse fitting and convol-utional neural network.,27740
and Our problems thus collection. solution is approach annotation to dataset of a the promising,Our approach is thus a promising solution to the problems of annotation and dataset collection.,27741
often problem interpretability is underlying prediction respect The to with unclear. the mechanisms prediction of,The interpretability of prediction mechanisms with respect to the underlying prediction problem is often unclear.,27742
applicability using is evaluated data. data using artificial real-world framework The and demonstrated,The framework applicability is evaluated using artificial data and demonstrated using real-world data.,27743
"network enhance solution to further reconstruction In quality. leverages the a chained our addition,","In addition, our solution leverages a chained network to further enhance the reconstruction quality.",27744
"the work, techniques in gap. we two fill to introduce In this shall","In this work, we shall introduce two techniques to fill in the gap.",27745
for same. metrics ontology formulate We the based,We formulate ontology based metrics for the same.,27746
to  are  networks Convolutional (CNNs) translation. equivariant inherently neural,  Convolutional neural networks (CNNs) are inherently equivariant to translation.,27747
to have forms rotation. Efforts equivariance on solely other of embed concentrated,Efforts to embed other forms of equivariance have concentrated solely on rotation.,27748
We Network the of Transformer in notion CNNs Polar equivariance (PTN). through the expand,We expand the notion of equivariance in CNNs through the Polar Transformer Network (PTN).,27749
from Transformer and PTN Network canonical ideas representations. (STN) Spatial the coordinate combines,PTN combines ideas from the Spatial Transformer Network (STN) and canonical coordinate representations.,27750
both a rotation to The is equivariant result network to translation and scale. and invariant,The result is a network invariant to translation and equivariant to both rotation and scale.,27751
pixels similarity to the score. learn the input image weights the we from final,we learn the weights from the input image pixels to the final similarity score.,27752
"complex. arbitrarily in model limitations be cannot to optimization, Due the","Due to limitations in optimization, the model cannot be arbitrarily complex.",27753
the thesis investigates and stage approach presents This segmentation. a to latter two semantic,This thesis investigates the latter and presents a two stage approach to semantic segmentation.,27754
to The the (e.g. model underlying formulation segmentation agnostic is,The formulation is agnostic to the underlying segmentation model (e.g.,27755
semantic Contextual is segmentation. crucial information for,Contextual information is crucial for semantic segmentation.,27756
"in and current methods, hand-tuned parameters dilation However, state-of-the-art are fixed.","However, in current state-of-the-art methods, dilation parameters are hand-tuned and fixed.",27757
"popular PASCAL benchmarks, report MS COCO, best We results ImageNet. including VOC, the and on","We report the best results on popular benchmarks, including PASCAL VOC, MS COCO, and ImageNet.",27758
conservation. good Our practices development system for the wildlife of data-driven in provides management perspectives,Our system provides good perspectives for the development of data-driven management practices in wildlife conservation.,27759
image strong approaches classification. medical there Currently data-driven in is interest to,Currently there is strong interest in data-driven approaches to medical image classification.,27760
"medical scarce, However, patient fraught legal with and concerns regarding expensive, imaging is data privacy.","However, medical imaging data is scarce, expensive, and fraught with legal concerns regarding patient privacy.",27761
The first an Asteroids-inspired structure. basic domain complex dynamics but is with logical game,The first domain is an Asteroids-inspired game with complex dynamics but basic logical structure.,27762
"logical The simpler the with but complex dynamics Game, Treasure second is structure. more","The second is the Treasure Game, with simpler dynamics but more complex logical structure.",27763
"the class-conditional Optimal projection simplest means. ""Linear ver-sion, The Low-rank"" incorporates (LOL),","The simplest ver-sion, ""Linear Optimal Low-rank"" projection (LOL), incorporates the class-conditional means.",27764
"perform to Graph-cut on is employed Random labeling. Field Multi-label Finally, pixel-wise Markov","Finally, Multi-label Graph-cut on Markov Random Field is employed to perform pixel-wise labeling.",27765
show video proposed outperforms evaluation results challenging the that method methods sequences. state-of-the-art on Extensive,Extensive evaluation results show that the proposed method outperforms state-of-the-art methods on challenging video sequences.,27766
evolving alternative tool an land vehicles Unmanned are (UAV) data. aerial acquire as to tenure,Unmanned aerial vehicles (UAV) are evolving as an alternative tool to acquire land tenure data.,27767
"in is development this developed this of paper. methodological However, goal the towards tool the","However, the methodological development of the tool towards this goal is developed in this paper.",27768
"models exogenous to scale, such are to. hard tune, add However, and variables","However, such models are hard to tune, scale, and add exogenous variables to.",27769
"Bayesian strategy, so-called leading inference. follows to amortized empirical This an design","This design follows an empirical Bayesian strategy, leading to so-called amortized inference.",27770
"in maps, which better DECNN. in results embedding further transformed forward feature are This operation","This embedding operation results in better feature maps, which are further transformed forward in DECNN.",27771
the field. classic is vision computer Face in a alignment problem,Face alignment is a classic problem in the computer vision field.,27772
testing. model Our can real time run at during,Our model can run at real time during testing.,27773
of the the wild setting challenge). in,in the wild setting of the challenge).,27774
"is queries in Consequently, real-user far us unclear. methods how have answering brought these","Consequently, how far have these methods brought us in answering real-user queries is unclear.",27775
answering two questions. pragmatic a This the to paper approach takes,This paper takes a pragmatic approach to answering the two questions.,27776
to different to efforts bundling devoted edge have improve Researchers metrics. according significant,Researchers have devoted significant efforts to improve edge bundling according to different metrics.,27777
fundamental of characterization Collections understanding biological to diversity. of are natural specimens and scientific,Collections of biological specimens are fundamental to scientific understanding and characterization of natural diversity.,27778
"to structure pose systems. challenges vision in and complex many be Furthermore, insects can computer","Furthermore, insects can be complex in structure and pose many challenges to computer vision systems.",27779
knowledge Artificial transfer. such an Intelligence call learning We,We call such learning an Artificial Intelligence knowledge transfer.,27780
"Hamiltonian have and been engineering, implemented Flow Carlo(HFMC) methods chemistry. biology in Monte","Hamiltonian Flow Monte Carlo(HFMC) methods have been implemented in engineering, biology and chemistry.",27781
based gradient the makes rapidly HFMC space. to large state steps explore,HFMC makes large gradient based steps to rapidly explore the state space.,27782
tested barrier the We option case the pricing. of the on algorithm,We tested the algorithm on the case of the barrier option pricing.,27783
learning. within distribution tasks new technique a for on Both rely matching amortized,Both tasks rely on a new technique for distribution matching within amortized learning.,27784
detect part-specific that learn are CNN parts encoded representation. a and bird by Images part-based,Images are encoded by a part-based CNN that detect bird parts and learn part-specific representation.,27785
"code, models data, Our and available. publically are","Our code, data, and models are publically available.",27786
dataset more is to that considered cells. with sparse This dataset a be challenging,This dataset is considered to be more challenging that a dataset with sparse cells.,27787
approached method The unsupervised its of accuracy counterpart. our that supervised of,The accuracy of our unsupervised method approached that of its supervised counterpart.,27788
"However, removed automated be should prior this to border processing.","However, this border should be removed prior to automated processing.",27789
"two I Here, terms. retiring these suggest","Here, I suggest retiring these two terms.",27790
"well-known Yet, whole scalability issue. affected the by is framework the","Yet, the whole framework is affected by the well-known scalability issue.",27791
extensive the the superiority over conducted against demonstrate multi-view datasets experiments state-of-the-arts. real-world The,The extensive experiments conducted against the real-world multi-view datasets demonstrate the superiority over state-of-the-arts.,27792
publicly framework available. The using code Keras source is,The source code using Keras framework is publicly available.,27793
annotated datasets. landmark to partially We techniques two in images from localization present improve,We present two techniques to improve landmark localization in images from partially annotated datasets.,27794
visual for and does vision mean recognition? What computer this,What does this mean for computer vision and visual recognition?,27795
algorithms have classification excellent Current achieved visual accuracy. recognition,Current visual recognition algorithms have achieved excellent classification accuracy.,27796
"cloning, in panorama unique discuss in We i.e. depth two issues","We discuss in depth two unique issues in panorama cloning, i.e.",27797
"like to scheme a we train projector. act a to simple Finally, a propose CNN","Finally, we propose a simple scheme to train a CNN to act like a projector.",27798
"that However, low-quality existing complicated perform on methods poorly attention-based and/or we images. observe","However, we observe that existing attention-based methods perform poorly on complicated and/or low-quality images.",27799
based well-known unified on of face-space representation a concept. the multidimensional is It,It is based on a unified multidimensional representation of the well-known face-space concept.,27800
major steps. of consists three framework proposed The,The proposed framework consists of three major steps.,27801
The properties. architecture SRF-BN appealing has DMT several proposed,The proposed SRF-BN DMT architecture has several appealing properties.,27802
"On detect fail hand, to boundaries. irregular SRF very other accurately object may the","On the other hand, SRF may fail to accurately detect very irregular object boundaries.",27803
understand in to the sensory goal major brain is those A processing variables. extracts how,A major goal in sensory processing is to understand how the brain extracts those variables.,27804
encoding we model linearly. in revisit which are common encoded a Here variables,Here we revisit a common encoding model in which variables are encoded linearly.,27805
single-task on surface label. the a only FCN trained We evaluated (SFCN) first,We first evaluated a single-task FCN (SFCN) trained only on the surface label.,27806
from and is images age real-world importance. estimation face Automatic gaining unconstrained rapidly,Automatic age estimation from real-world and unconstrained face images is rapidly gaining importance.,27807
has significant in this paper This field. three contributions,This paper has three significant contributions in this field.,27808
astrophysicists. This less for than is for true no statisticians,This is no less true for statisticians than for astrophysicists.,27809
navigate. can be to in Their difficult diversity approach,Their diversity in approach can be difficult to navigate.,27810
alleviate common to by this have attempted Some designing interfaces. problem,Some have attempted to alleviate this problem by designing common interfaces.,27811
forest model. by through both approaches fitting illustrate a simple We a example working random,We illustrate both approaches by working through a simple example fitting a random forest model.,27812
"parallel derive algorithm, serial versions. with two we a Beginning","Beginning with a serial algorithm, we derive two parallel versions.",27813
shape realistic Example-based methods deformation mesh powerful tools for are editing.,Example-based mesh deformation methods are powerful tools for realistic shape editing.,27814
complicated user-specified explain model deformation. the a using to overly,using a overly complicated model to explain the user-specified deformation.,27815
is causes world. death the the of Coronary (CAD) disease leading around artery,Coronary artery disease (CAD) is the leading causes of death around the world.,27816
of this common One for the angiography. most imaging is methods X-ray disease diagnosing,One of the most common imaging methods for diagnosing this disease is X-ray angiography.,27817
and of misdiagnosis. to tougher the task cardiologists prone These make diagnoses more challenges,These challenges make the diagnoses task of cardiologists tougher and more prone to misdiagnosis.,27818
segmentation based proposed method on is Our superpixels.,Our proposed segmentation method is based on superpixels.,27819
ridges the arteries extracting use the for image Also we centerlines. coronary,Also we use the image ridges for extracting the coronary arteries centerlines.,27820
shorter that prove evaluation better much our performs in The a method time. results,The evaluation results prove that our method performs better in a much shorter time.,27821
the become can this diffusion a axons abnormality. The as to result anomalous from in,The diffusion in the axons can become to anomalous as a result from this abnormality.,27822
affected. is voltage propagation in this axons the In case,In this case the voltage propagation in axons is affected.,27823
Another along by neurodegenerative hallmark feature different discrete given diseases the is axon. swellings of,Another hallmark feature of different neurodegenerative diseases is given by discrete swellings along the axon.,27824
A obtained the the and is behavior of similar when swellings fractional increase. number effect,A similar behavior is obtained when the number of swellings and the fractional effect increase.,27825
features methods where contrast are those traditionally hand used. is to This crafted in,This is in contrast to those methods where traditionally hand crafted features are used.,27826
calculation features these task. challenging The of and a is selection,The selection and calculation of these features is a challenging task.,27827
challenges are and of highlighted. also The these techniques potential,The challenges and potential of these techniques are also highlighted.,27828
show in an problem can restated the that A-optimality be constrained We artificial as model.,We show that the problem can be restated as constrained A-optimality in an artificial model.,27829
from amount requires Deep a of scratch labeled Neural (DNN) data. Network a large Training,Training a Deep Neural Network (DNN) from scratch requires a large amount of labeled data.,27830
to cameras the pre-calibrated. applications depth of be require Most these,Most of these applications require the depth cameras to be pre-calibrated.,27831
accurate a makeup. policies country's Targeted understanding require socioeconomic an of demographic,Targeted socioeconomic policies require an accurate understanding of a country's demographic makeup.,27832
date. We use detection reported fine-grained construct to the system scale largest this data to,We use this data to construct the largest scale fine-grained detection system reported to date.,27833
"better can nanometer. principle, than resolution achieve In SEMs one","In principle, SEMs can achieve resolution better than one nanometer.",27834
"images characteristics. when works have well Conceptually, the method our similar","Conceptually, our method works well when the images have similar characteristics.",27835
recognition. automated of critical in fingerprint is Minutiae extraction importance,Minutiae extraction is of critical importance in automated fingerprint recognition.,27836
is with fixed to that shallow this pipeline a network demonstrate equivalent We weights.,We demonstrate that this pipeline is equivalent to a shallow network with fixed weights.,27837
camera-laser Calibration for with are scanner panoramic system accuracy costly. extreme equipments,Calibration equipments with extreme accuracy for panoramic camera-laser scanner system are costly.,27838
matrix high Back accuracy. the to regress translation propagation and rotation algorithm utilized is with,Back propagation algorithm is utilized to regress the rotation and translation matrix with high accuracy.,27839
"is The and of really calibration method highly-automated. more it high, this accuracy is","The calibration accuracy of this method is really high, and it is more highly-automated.",27840
the grasping learners underlying are adept at complex sequential Human relationships input. incoming,Human learners are adept at grasping the complex relationships underlying incoming sequential input.,27841
"annotated for and However, guides sources such as provide many classes. images websites field e-commerce","However, sources such as e-commerce websites and field guides provide annotated images for many classes.",27842
encoder-decoders a model conventional feature-based literature. Our and the from approach outperforms,Our model outperforms a feature-based approach and conventional encoder-decoders from the literature.,27843
model compression. potential distillation a solution Knowledge for is,Knowledge distillation is a potential solution for model compression.,27844
framework present resulting We sensitivity analysis mathematical and the system. the for the,We present the mathematical framework and the sensitivity analysis for the resulting system.,27845
Cui informed al. likelihood use (LIS) dimension et We reduction (T.,We use likelihood informed dimension reduction (LIS) (T. Cui et al.,27846
Traditional cues. level visual models on relying few attention saliency predict human the often image,Traditional saliency models often predict the human visual attention relying on few level image cues.,27847
purpose. or very expensive Such research industrial were only was for available often and systems,Such systems were often very expensive and was only available for industrial or research purpose.,27848
"data acquires camera point our First, cloud inexpensive system using a of sensor. person/object","First, our system acquires point cloud data of a person/object using inexpensive camera sensor.",27849
is vessel regions mask WSI A produced. of the wall on,A mask of vessel wall regions on the WSI is produced.,27850
heterogeneous. of very of set segmentations are quality The the is good although slides very,The segmentations are of very good quality although the set of slides is very heterogeneous.,27851
Image increasingly a video on used are scale. and analytics massive being,Image and video analytics are being increasingly used on a massive scale.,27852
computational resources important to It wherever increasingly becoming save possible. is,It is becoming increasingly important to save computational resources wherever possible.,27853
on approximate is HMC performed the then potential.,HMC is then performed on the approximate potential.,27854
spaces. rise approach also constrained to new gives RBHMC a Bayesian inference for on,RBHMC also gives rise to a new approach for Bayesian inference on constrained spaces.,27855
significantly constrained hyperspectral This its image for has applications classification. (HSI),This has significantly constrained its applications for hyperspectral image (HSI) classification.,27856
Three-dimensional is cardiac for of crucial cine-MRI assessing function. importance the,Three-dimensional cine-MRI is of crucial importance for assessing the cardiac function.,27857
of anatomy Features that the describe structures function (e.g. cardiac and,Features that describe the anatomy and function of cardiac structures (e.g.,27858
"precise require structures. cardiac However, segmentation features of these","However, these features require precise segmentation of cardiac structures.",27859
"two-stage selection developed. for is method pipeline Next, a feature LASSO using the","Next, a two-stage pipeline for feature selection using the LASSO method is developed.",27860
"relations. show that spatial accuracy, exploiting the increase the the by results segmentation PPs The","The results show that the PPs increase the segmentation accuracy, by exploiting the spatial relations.",27861
on paper for human estimation. pose This presents a comprehensive review method regression-based,This paper presents a comprehensive review on regression-based method for human pose estimation.,27862
advantages limitation for different their results scenarios. The demonstrate and experimental,The experimental results demonstrate their advantages and limitation for different scenarios.,27863
"an solve optimization the algorithm iterative problem. is described Furthermore, to","Furthermore, an iterative algorithm is described to solve the optimization problem.",27864
other many improvement significant method proposed showed a The over existing methods.,The proposed method showed a significant improvement over many other existing methods.,27865
environment. Many mathematical evolution individuals same the models experience all of that assume,Many mathematical models of evolution assume that all individuals experience the same environment.,27866
"study Moran environments. in process Here, the heterogeneous we","Here, we study the Moran process in heterogeneous environments.",27867
Reproductive both environment. is the by rate type and the determined,Reproductive rate is determined by both the type and the environment.,27868
the heterogeneity environmental the an resident. interesting mutant But and elucidates asymmetry between,But environmental heterogeneity elucidates an interesting asymmetry between the mutant and the resident.,27869
data is Training networks recognition. challenging with in face large-scale triplet,Training triplet networks with large-scale data is challenging in face recognition.,27870
"Combined much found the be triplets OHNM, can with batch hard easier.","Combined with the batch OHNM, hard triplets can be found much easier.",27871
images. quality of that Noise important factor degrades medical an the is,Noise is an important factor that degrades the quality of medical images.,27872
complexity an But essential. accuracy with a having acceptable low hardware algorithm is implementation for,But for hardware implementation having a low complexity algorithm with an acceptable accuracy is essential.,27873
proposed intensities. non-noisy have that method distinguishes The pixels noise-like,The proposed method distinguishes non-noisy pixels that have noise-like intensities.,27874
have hardware All low steps to designed are complexity.,All steps are designed to have low hardware complexity.,27875
proximity. and the representing in standard approach for colors The proposed evaluate relies CIE,The proposed approach relies in the CIE standard for representing colors and evaluate proximity.,27876
problem this use of a to approaches rules. set Other,Other approaches to this problem use a set of rules.,27877
learning quickly become Deep weapon of for lesion brain choice segmentation. has the,Deep learning has quickly become the weapon of choice for brain lesion segmentation.,27878
architecture subspace present neural a for We novel clustering. unsupervised network deep,We present a novel deep neural network architecture for unsupervised subspace clustering.,27879
methods. show experiments Our that unsupervised state-of-the-art the subspace significantly the outperforms method clustering proposed,Our experiments show that the proposed method significantly outperforms the state-of-the-art unsupervised subspace clustering methods.,27880
"disasters still occur frequently. crowd Despite measures, safety","Despite safety measures, crowd disasters still occur frequently.",27881
"dynamics. several In crowd recent models understanding for developed years, researchers","In recent years, researchers developed several models for understanding crowd dynamics.",27882
"We compact for model optical a but flow, effective PWC-Net. CNN called present","We present a compact but effective CNN model for optical flow, called PWC-Net.",27883
the that shape decisions one of customers. trigger Product is of factors preference,Product shape is one of the factors that trigger preference decisions of customers.,27884
"choice defining the without and their parts prototype, relations. of explicitly our","our choice of the prototype, without explicitly defining parts and their relations.",27885
is use entropy. signals we and The that concept basic systems the,The basic signals and systems concept that we use is the entropy.,27886
measure value volume lowest attains a by enclosed for sphere. its Our a,Our measure attains its lowest value for a volume enclosed by a sphere.,27887
"increase hand, deformations the prototype one in the cause an measure. from On","On one hand, deformations from the prototype cause an increase in the measure.",27888
consistent preliminary expectations. with our are experimental Our results,Our preliminary experimental results are consistent with our expectations.,27889
this calculations. Maximum demonstrate such approximating of the Entropy in we optimality methods In paper,In this paper we demonstrate the optimality of Maximum Entropy methods in approximating such calculations.,27890
our of results and verify practices. empirically on a matrices establish We best SparseSuite variety,We empirically verify our results on a variety of SparseSuite matrices and establish best practices.,27891
present by an We clustering system faces full-length and detecting movies. end-to-end for in identity,We present an end-to-end system for detecting and clustering faces by identity in full-length movies.,27892
standard video and multiple state-of-the-art on sets results on present data face also databases. We,We present state-of-the-art results on multiple video data sets and also on standard face databases.,27893
"neural architecture deep is network presented. a this paper, In classification galaxies convolutional for","In this paper, a deep convolutional neural network architecture for galaxies classification is presented.",27894
the is works. compared and testing with accuracy was other comparative related made A result,A comparative result is made and the testing accuracy was compared with other related works.,27895
related outperformed The other of accuracy. works terms in proposed architecture testing,The proposed architecture outperformed other related works in terms of testing accuracy.,27896
"prototypes. number of however, has of a Function comparatively small objects,","Function of objects, however, has a comparatively small number of prototypes.",27897
"for object a objects generic valuable be Therefore, function-based recognition. of could new tool classification","Therefore, function-based classification of new objects could be a valuable tool for generic object recognition.",27898
interactions functions. modeling are helpful object for Hand-object,Hand-object interactions are helpful for modeling object functions.,27899
state The model interaction the alone. quantitative object of the from image infers the,The model infers the quantitative state of the interaction from the object image alone.,27900
demonstrate infer we And hand-object model also can possible that the interactions.,And also we demonstrate that the model can infer possible hand-object interactions.,27901
a auto-encoder dimensionality method method. of type The reduction is,The auto-encoder method is a type of dimensionality reduction method.,27902
auto-encoder propose and transform of We transform parameters. invariant inference a model an,We propose a transform invariant auto-encoder and an inference model of transform parameters.,27903
"a The approach, network faces. emotions by isolated expressed on analyzing works bottom-up neural","The neural network works on a bottom-up approach, analyzing emotions expressed by isolated faces.",27904
scene Bayesian features obtained a emotion integrating global The classifier top-down a descriptor. estimates through,The Bayesian classifier estimates a global emotion integrating top-down features obtained through a scene descriptor.,27905
"system alignment, recognition, Here multi-frame present of holistic superresolution and for we a faces.","Here we present a holistic system for multi-frame recognition, alignment, and superresolution of faces.",27906
popular recognition (YouTube dataset the results for our Faces). video face present We using,We present our results using the popular dataset for video face recognition (YouTube Faces).,27907
task. detection Automated a vision computer is text difficult,Automated text detection is a difficult computer vision task.,27908
"this enables a mind, that proposed regularity-exploiting quantification. technique is With in fast metamodeling uncertainty","With this in mind, a regularity-exploiting metamodeling technique is proposed that enables fast uncertainty quantification.",27909
by obtained indices are coefficients sensitivity directly from expansion Sobol' a post-processing. the mere,Sobol' sensitivity indices are obtained directly from the expansion coefficients by a mere post-processing.,27910
Bayesian accelerated. Carlo chain drastically posterior sampling is Markov Monte via inference,Bayesian inference via Markov chain Monte Carlo posterior sampling is drastically accelerated.,27911
perspective the from disjunction a the We analyze the belief of effect as vector. quantum,We analyze the disjunction effect from the perspective of the belief as a quantum vector.,27912
manifolds. has of this The two paper contribution,The contribution of this paper has two manifolds.,27913
of day data. developed one naturalistic method using The driving is validated then,The developed method is then validated using one day of naturalistic driving data.,27914
of an We model propose operator serves perception. induction that adequate as,We propose that operator induction serves as an adequate model of perception.,27915
universal operator how We explain induction. to reduce models agent to,We explain how to reduce universal agent models to operator induction.,27916
learning data. learning support classification methods sparse for Low-shot image from,Low-shot learning methods for image classification support learning from sparse data.,27917
to semantic image these extend segmentation. dense support techniques We,We extend these techniques to support dense semantic image segmentation.,27918
various Sparse has achieved in tasks. a success image coding great processing,Sparse coding has achieved a great success in various image processing tasks.,27919
the of perspective attempts gap from This fill minimization. to rank work the,This work attempts to fill the gap from the perspective of rank minimization.,27920
please More manuscript.... . see the details,More details please see the manuscript.... .,27921
The of second the is information spatial classification. lack for,The second is the lack of spatial information for classification.,27922
word data offline performance level on character and The reported. is,The performance on character and word level offline data is reported.,27923
"that e.g. to Compared use tracking approaches,","Compared to tracking approaches, that use e.g.",27924
"expression can be only of Moreover, categories facial changed. the","Moreover, only the categories of facial expression can be changed.",27925
the enables continuously intensity low be high. This adjusted architecture to novel expression from to,This novel architecture enables the expression intensity to be continuously adjusted from low to high.,27926
Quantitative effectiveness the ExprGAN. evaluations and used dataset the of demonstrate qualitative on Oulu-CASIA widely,Quantitative and qualitative evaluations on the widely used Oulu-CASIA dataset demonstrate the effectiveness of ExprGAN.,27927
detection. A is learning to employed framework during shape cascaded constraint enhance,A cascaded learning framework is employed to enhance shape constraint during detection.,27928
approach for Bayesian in determinant-free Gaussian inference propose high-dimensional models. a simulation-based We,We propose a determinant-free approach for simulation-based Bayesian inference in high-dimensional Gaussian models.,27929
variables the introduce auxiliary covariance the of to inverse We covariance model. with equal,We introduce auxiliary variables with covariance equal to the inverse covariance of the model.,27930
with scales operations at large be These performed rational approximations. can,These operations can be performed at large scales with rational approximations.,27931
recently proposed. auxiliary More $O(KCHW)$ space requiring approaches memory-efficient been just have,More recently memory-efficient approaches requiring just $O(KCHW)$ auxiliary space have been proposed.,27932
have well method. effectiveness our the of Experiments validated,Experiments have well validated the effectiveness of our method.,27933
adversarial integrates learning feature hallucination This network. end-to-end and framework into an cross-spectral discriminative face,This framework integrates cross-spectral face hallucination and discriminative feature learning into an end-to-end adversarial network.,27934
losses learning These domain-invariant feature removing. independent and modality two enhance noise,These two losses enhance domain-invariant feature learning and modality independent noise removing.,27935
learning networks. with metric shape We cast matching as convolutional,We cast shape matching as metric learning with convolutional networks.,27936
into parts. break process image We end-to-end the two of representation,We break the end-to-end process of image representation into two parts.,27937
"edge established turn the well into images methods to maps. efficient Firstly, chosen are","Firstly, well established efficient methods are chosen to turn the images into edge maps.",27938
models useful species understand disturbances. loss resilience to are to to robustness and community Coextinction,Coextinction models are useful to understand community robustness to species loss and resilience to disturbances.,27939
"Finally, species communities. and beetles honeybees showed were that in several our keystone we","Finally, we showed that honeybees and several beetles were keystone species in our communities.",27940
into on and be recognition summarized Studies sketch generation. roughly can,Studies on sketch can be roughly summarized into recognition and generation.,27941
satisfying image models obtain Existing to failed on sketch performance recognition classification. on,Existing models on image recognition failed to obtain satisfying performance on sketch classification.,27942
"dropped fed the multiple However, when performance categories into model. the were","However, the performance dropped when multiple categories were fed into the model.",27943
are graphs pretty similarity on Multi-view methods popular. based matrixes clustering spectral or,Multi-view spectral clustering methods based on similarity matrixes or graphs are pretty popular.,27944
"original graphs these informative directly data. by utilizing learn Generally, algorithms","Generally, these algorithms learn informative graphs by directly utilizing original data.",27945
"contributions different views different to may data clustering. addition, In have","In addition, different views may have different contributions to data clustering.",27946
algorithm optimization is to this solve problem. developed An iterative,An iterative algorithm is developed to solve this optimization problem.,27947
"and Moreover, MSCAM. real-world demonstrate the of dataset synthetic effectiveness experiments different on datasets a","Moreover, experiments on a synthetic dataset and different real-world datasets demonstrate the effectiveness of MSCAM.",27948
for work recognition focuses methods cost species forest on This systems. reduction,This work focuses on cost reduction methods for forest species recognition systems.,27949
two We between to emerge that the seizures recurrent demonstrate interaction networks. further due,We further demonstrate that recurrent seizures emerge due to the interaction between two networks.,27950
A using a method is to single proposed novel trampoline camera identify herein. video skills,A novel method to identify trampoline skills using a single video camera is proposed herein.,27951
estimates extracted Body pose these angle estimates. and from orientation joint are,Body orientation and joint angle estimates are extracted from these pose estimates.,27952
usage homography estimation a on robotic requires and fast The a algorithm. robust system,The usage on a robotic system requires a fast and robust homography estimation algorithm.,27953
the In a addressed. novel is to the shape approach graph-based present decomposition problem paper,In the present paper a novel graph-based approach to the shape decomposition problem is addressed.,27954
visibility a enriched local shape The neighborhood transformed graph into with information. appropriately is,The shape is appropriately transformed into a visibility graph enriched with local neighborhood information.,27955
different face regions attributes. focus attention on recognizing Humans face when,Humans focus attention on different face regions when recognizing face attributes.,27956
as use input. existing Most classification methods image the whole face attribute,Most existing face attribute classification methods use the whole image as input.,27957
"face parts. of provide defined some to fiducial landmarks Moreover, these on rely methods","Moreover, some of these methods rely on fiducial landmarks to provide defined face parts.",27958
"achieve like sensors cameras several and radars, that, vehicles. are installed LiDARs autonomous To in","To achieve that, several sensors like radars, LiDARs and cameras are installed in autonomous vehicles.",27959
a the general fused sensor data is of The representation to produced surrounding.,The produced sensor data is fused to a general representation of the surrounding.,27960
approach al. thesis grid of et dynamic the Nuss this In map occupancy,In this thesis the dynamic occupancy grid map approach of Nuss et al.,27961
three are is used achieved. while goals,is used while three goals are achieved.,27962
"al. the First, et of Nuss approach","First, the approach of Nuss et al.",27963
for orientation provide moving obstacle. each is to goal second detected information The,The second goal is to provide orientation information for each detected moving obstacle.,27964
"tracking could which map. dynamic based on occupancy are This algorithms, grid improve the","This could improve tracking algorithms, which are based on the dynamic occupancy grid map.",27965
"population evolutionary consider an extended size. process dynamics variable Here, we Wright-Fisher in game with","Here, we consider evolutionary game dynamics in an extended Wright-Fisher process with variable population size.",27966
"that becomes a such the entire scenario, possible population is extinct. In it","In such a scenario, it is possible that the entire population becomes extinct.",27967
which the of depend Survival may in strategy population the prevails on dynamics. game,Survival of the population may depend on which strategy prevails in the game dynamics.,27968
methods takeover We all a of on of bids. demonstrate dataset these real-world,We demonstrate all of these methods on a real-world dataset of takeover bids.,27969
accepted to Makeup is used improve the widely is attractiveness facial public. by well and,Makeup is widely used to improve facial attractiveness and is well accepted by the public.,27970
"makeup appearance facial significant in different will result changes. However, styles","However, different makeup styles will result in significant facial appearance changes.",27971
match images. to face challenging non-makeup and It remains makeup a problem,It remains a challenging problem to match makeup and non-makeup face images.,27972
the two makeup networks gap non-makeup between sensing jointly images. reduce These and,These two networks jointly reduce the sensing gap between makeup and non-makeup images.,27973
"well constrained multiple we Moreover, generator perceptual losses. by make incorporating the","Moreover, we make the generator well constrained by incorporating multiple perceptual losses.",27974
"error estimate expected the positions. Furthermore, systematic the mapped we on","Furthermore, we estimate the expected systematic error on the mapped positions.",27975
"proposed module. network Our end-to-end is trained segmentation begins detections, category-level a with and given","Our proposed network is trained end-to-end given detections, and begins with a category-level segmentation module.",27976
learning We present a learning. novel sets approach deep predict using to for,We present a novel approach for learning to predict sets using deep learning.,27977
"naturally are vectors. sets, real-world described rather Many problems, as however, than","Many real-world problems, however, are naturally described as sets, rather than vectors.",27978
framework modeling We pre-specific this modeling. call,We call this modeling framework pre-specific modeling.,27979
matchers. proposed they The application requirements application-oriented for are as emphasize metrics,The proposed metrics are application-oriented as they emphasize application requirements for matchers.,27980
bivariate copulas arbitrary by more Vine copulas. (conditional) and flexible combining are distributions marginal,Vine copulas are more flexible by combining arbitrary marginal distributions and (conditional) bivariate copulas.,27981
"Yet, effort dimension this accompanied the by as increasing sharply increases. is adaptability computational","Yet, this adaptability is accompanied by sharply increasing computational effort as the dimension increases.",27982
"model. into Finally, we reconcile one joint them","Finally, we reconcile them into one joint model.",27983
all abilities? How learn and high-level does neocortex our develop of the foundations cognitive the,How does the neocortex learn and develop the foundations of all our high-level cognitive abilities?,27984
"introduce sparse regularization deformations. convolutional with along framework, localize helps in this operations, which We","We introduce sparse regularization in this framework, which along with convolutional operations, helps localize deformations.",27985
in state-of-the-art outperforms qualitative methods and both evaluations. quantitative our show that experiments method Extensive,Extensive experiments show that our method outperforms state-of-the-art methods in both qualitative and quantitative evaluations.,27986
"scans. at small CT the (e.g., from abdominal organs pancreas) segmenting aim We","We aim at segmenting small organs (e.g., the pancreas) from abdominal CT scans.",27987
Saliency paper Recurrent Network. presents a This Transformation,This paper presents a Recurrent Saliency Transformation Network.,27988
"input training, dealing networks deep with it joint different over In scales. optimization allows the","In training, it allows joint optimization over the deep networks dealing with different input scales.",27989
"information throughout propagates In segmentation to accuracy. improve iterations it multi-stage testing, visual","In testing, it propagates multi-stage visual information throughout iterations to improve segmentation accuracy.",27990
the monotonicity in the weight orthogonal WOCR makes inherent to function. use parameterize of components,WOCR makes use of the monotonicity inherent in orthogonal components to parameterize the weight function.,27991
computationally of and is determination formulation allows The parameters advantageous. efficient for hence tuning,The formulation allows for efficient determination of tuning parameters and hence is computationally advantageous.,27992
"new for Moreover, better WOCR insights offers variants. deriving","Moreover, WOCR offers insights for deriving new better variants.",27993
"concept. fusion, the an important Information In is conflict","In Information fusion, the conflict is an important concept.",27994
"conflict. Indeed, allows several imperfect combining or sources experts","Indeed, combining several imperfect experts or sources allows conflict.",27995
"been lot. theory this In belief discussed functions, a notion of has the","In the theory of belief functions, this notion has been discussed a lot.",27996
applied efficient binary network for is pedestrian detection. resulting The,The resulting binary network is applied for efficient pedestrian detection.,27997
numerical of features have the on investigate KB completion We performance the impact also KBLRN.,We also investigate the impact numerical features have on the KB completion performance of KBLRN.,27998
"algorithm However, fast DNN is bottleneck in training the level. for current the","However, the current bottleneck for fast DNN training is in the algorithm level.",27999
source code available upon Our request. is,Our source code is available upon request.,28000
binning. four camera The cover datasets spatial resolution using hardware levels,The datasets cover four spatial resolution levels using camera hardware binning.,28001
of implementation Each at one purpose. packages suitable above least offers the one to this,Each one of the above packages offers at least one implementation suitable to this purpose.,28002
native runs and same are the out their carried All on machine on platforms.,All runs are carried out on the same machine and on their native platforms.,28003
"make clustering In kernelized optimization an the model this problem. paper, we for multi-view self-representation","In this paper, we make an optimization model for the kernelized multi-view self-representation clustering problem.",28004
Understanding biodiversity the emergence biology. in in patterns central a problem of nature is,Understanding the emergence of biodiversity patterns in nature is a central problem in biology.,28005
model these speciation. use individual-based of explicit the algorithms a in of illustrate We spatially,We illustrate the use of these algorithms in a spatially explicit individual-based model of speciation.,28006
"neural from very the problem. networks deep However, may suffer overfitting","However, very deep neural networks may suffer from the overfitting problem.",28007
systems day are Conversational famous in lives. AI to becoming day,Conversational AI systems are becoming famous in day to day lives.,28008
how can and in be achieved of it personalization An importance explained detail. is,An importance of personalization and how it can be achieved is explained in detail.,28009
"the Furthermore, applicability extends time ones. single it from to series multiple","Furthermore, it extends the applicability from single time series to multiple ones.",28010
facing the weights. is An users of OWA OWA operators' determination the issue important the,An important issue facing the OWA operators' users is the determination of the OWA weights.,28011
of represented are moments the two These two by truncated dimensions distribution. the first,These two dimensions are represented by the two first moments of the truncated distribution.,28012
of the number We of the criteria study results. the on finally impact,We finally study the impact of the number of criteria on the results.,28013
"network we for proposals. paper, this a propose zoom-out-and-in In generating object","In this paper, we propose a zoom-out-and-in network for generating object proposals.",28014
even more accuracy show profoundly. Recent recognition trends increasing,Recent trends show recognition accuracy increasing even more profoundly.,28015
The of amounts large of number also large memory. parameters require,The large number of parameters also require large amounts of memory.,28016
sizes. and large is increasingly model times This resulting computation long in,This is resulting in increasingly long computation times and large model sizes.,28017
to accuracy Extensive GLAD results competitive experimental compared methods. achieves state-of-the-art show the,Extensive experimental results show GLAD achieves competitive accuracy compared to the state-of-the-art methods.,28018
accelerates Our online Re-ID of without significantly framework loss the procedure accuracy. retrieval,Our retrieval framework significantly accelerates the online Re-ID procedure without loss of accuracy.,28019
"tasks to potential has this work scenarios. person better on in work Therefore, real Re-ID","Therefore, this work has potential to work better on person Re-ID tasks in real scenarios.",28020
identify (NA) of Biological between different species. aims regions molecular alignment similar to networks network,Biological network alignment (NA) aims to identify similar regions between molecular networks of different species.,28021
two pairs networks. PNA aligned between node produces,PNA produces aligned node pairs between two networks.,28022
networks. aligned MNA node than more clusters between two produces,MNA produces aligned node clusters between more than two networks.,28023
We introduce framework a allows that this. for,We introduce a framework that allows for this.,28024
PNA evaluation framework. is perform pairwise better to expected the under,PNA is expected to perform better under the pairwise evaluation framework.,28025
framework. to evaluation multiple under expected better is MNA the perform,MNA is expected to perform better under the multiple evaluation framework.,28026
for propose framework segmentation. a instance-level We object deep learning-based,We propose a deep learning-based framework for instance-level object segmentation.,28027
consists of mainly three method Our steps.,Our method mainly consists of three steps.,28028
"a network. per we spatial learn this refine frame score further propagation map, To","To further refine this per frame score map, we learn a spatial propagation network.",28029
category This framework paper on cognitive solving a proposes based problem for formal theory.,This paper proposes a formal cognitive framework for problem solving based on category theory.,28030
"any two which exactly categories, morphism categories introduce objects. one We with between cognitive are","We introduce cognitive categories, which are categories with exactly one morphism between any two objects.",28031
interpreted categories morphisms states as and as these Objects are transformations in between states.,Objects in these categories are interpreted as states and morphisms as transformations between states.,28032
state (i.e. system) a current the the and goal of,the current state of the system) and a goal (i.e.,28033
and Cognitive system means target evaluators. generators systems by transform of the,Cognitive systems transform the target system by means of generators and evaluators.,28034
grounded examples well. provided methods in AI as basic are Several,Several examples grounded in basic AI methods are provided as well.,28035
"considerable selection production, with humans In word language demands. confronted are","In language production, humans are confronted with considerable word selection demands.",28036
"and findings controllability, TMS effects. function, cognitive These establish network link between a","These findings establish a link between network controllability, cognitive function, and TMS effects.",28037
of composition. by aims at images adjusting aesthetic cropping the their quality improving Image,Image cropping aims at improving the aesthetic quality of images by adjusting their composition.,28038
the Most on supervision) sliding bounding supervised methods box window (without rely weakly mechanism. cropping,Most weakly supervised cropping methods (without bounding box supervision) rely on the sliding window mechanism.,28039
"function proposed reward which an benefits especially cropping. Particularly, aware the aesthetics image method develops","Particularly, the proposed method develops an aesthetics aware reward function which especially benefits image cropping.",28040
We agent end-to-end an architecture the manner. the actor-critic using train in,We train the agent using the actor-critic architecture in an end-to-end manner.,28041
on evaluated datasets. popular The cropping agent is several unseen,The agent is evaluated on several popular unseen cropping datasets.,28042
"split image the first an numerous layer, into watershed algorithm. the using regions In is","In the first layer, an image is split into numerous regions using the watershed algorithm.",28043
result As algorithm increase. a the will speed,As a result the algorithm speed will increase.,28044
methods. algorithm and of the improvement similar to show efficiency the Results its,Results show the efficiency of the algorithm and its improvement to similar methods.,28045
field a wild vision. the task computer the challenging Reading in text in of is,Reading text in the wild is a challenging task in the field of computer vision.,28046
"end-to-end for present In scene an text Attention Network this paper, Convolutional we recognition.","In this paper, we present an end-to-end Attention Convolutional Network for scene text recognition.",28047
"computer vision clutter large-scale image tasks, background challenging many is retrieval. in Severe including","Severe background clutter is challenging in many computer vision tasks, including large-scale image retrieval.",28048
"propose detection In region salient this a method. we work, novel","In this work, we propose a novel salient region detection method.",28049
to with Monte Carlo and simulations application an We soil method our data. demonstrate moisture,We demonstrate our method with Monte Carlo simulations and an application to soil moisture data.,28050
"C++ freely available. The C, codes are data and","The C, C++ codes and data are freely available.",28051
of the the harmony of (consonance). essence the of assumption physical Describes phenomenon,Describes the physical assumption of the essence of the phenomenon of harmony (consonance).,28052
wave for method the color implementation The sRGB article model. also algorithm of provides,The article also provides algorithm of implementation wave method for the sRGB color model.,28053
"of and and RL, RL. in It sub-goals multi-task temporal structural extracts relationships","It extracts temporal and structural relationships of sub-goals in RL, and multi-task RL.",28054
"them. and particular,it sub-goals In relationship among finds","In particular,it finds sub-goals and relationship among them.",28055
segmentation function. cardiac evaluating heart an the is Accurate step important towards of,Accurate segmentation of the heart is an important step towards evaluating cardiac function.,28056
"exact architecture role. only However, a plays network minor the","However, the exact network architecture only plays a minor role.",28057
"optimized However, be in method an this is manner. thus hard is manually-designed, end-to-end to","However, this method is manually-designed, thus is hard to be optimized in an end-to-end manner.",28058
visual outside learning propose cues an from We also objects. by improved version context DeepVoting+,We also propose an improved version DeepVoting+ by learning visual cues from context outside objects.,28059
revolutionizing intelligence artificial reinforcement field. is learning the Deep,Deep reinforcement learning is revolutionizing the artificial intelligence field.,28060
as problems linear be phrased vision Many in computer annotation integer can (ILPs). programs,Many annotation problems in computer vision can be phrased as integer linear programs (ILPs).,28061
"limited. However, extracting visual audio classification very performing and is features jointly and on research","However, research on jointly extracting audio and visual features and performing classification is very limited.",28062
function increases each log-likelihood in algorithm iterative the is step. The and monotonically approximate BSLM,The BSLM algorithm is iterative and monotonically increases the approximate log-likelihood function in each step.,28063
clustering simulations. rule is also and via assessed proposed favorably A model-based,A model-based clustering rule is proposed and also assessed favorably via simulations.,28064
in growing has medicine. treatment become a effects Assessing heterogeneous precision interest advancing,Assessing heterogeneous treatment effects has become a growing interest in advancing precision medicine.,28065
such (ITE) play Individualized treatment an critical role endeavor. effects in a,Individualized treatment effects (ITE) play a critical role in such an endeavor.,28066
traditional `separate regression' in outperforms approach RFIT ITE. the estimating,RFIT outperforms the traditional `separate regression' approach in estimating ITE.,28067
"train, The very easy few requires examples. training framework to is and","The framework is easy to train, and requires very few training examples.",28068
objective. this gradient-based We implement to then learning algorithms practical derive,We then derive practical gradient-based learning algorithms to implement this objective.,28069
Environment results (ALE) show Our in increased and Learning performance the Arcade interpretability.,Our results in the Arcade Learning Environment (ALE) show increased performance and interpretability.,28070
"this the of paper, we address problem recognition. stamp In","In this paper, we address the problem of stamp recognition.",28071
result have we on total different features. As types a of three,As a result on total we have three different types of features.,28072
"shot real-time, object small multi-box Keywords: detection, detector . feature single fusion,","Keywords: small object detection, feature fusion, real-time, single shot multi-box detector .",28073
We image investigate large number the problem of with scalable classification categories. a,We investigate the scalable image classification problem with a large number of categories.,28074
learning propose inter-class hierarchical on a classification method We based novel image structures.,We propose a novel image classification method based on learning hierarchical inter-class structures.,28075
analysis performed via of cloud such is conventional mainly manner satellite images. The,The conventional manner of such cloud analysis is performed mainly via satellite images.,28076
a following problem study We setting in this the contributions. relational and make,We study this problem in a relational setting and make the following contributions.,28077
"we two First, of notions different relational marginals. compare","First, we compare two different notions of relational marginals.",28078
that higher states. rewards achieve learns agent experiments and Our show to the better,Our experiments show that the agent learns to achieve higher rewards and better states.,28079
gut unravel Computational microbes to investigate modeling metabolic mechanistic novel interactions insights. between can systematically,Computational modeling can systematically investigate metabolic interactions between gut microbes to unravel novel mechanistic insights.,28080
"SCFA suggest we patient's Consequently, levels. that each personalized improve treatments dietary could","Consequently, we suggest personalized dietary treatments that could improve each patient's SCFA levels.",28081
Visual few undergone the significant past years. Object in tracking has research improvement,Visual Object tracking research has undergone significant improvement in the past few years.,28082
"networks extensively neural Recently, trackers. convolutional successful deep been most have in used","Recently, deep convolutional neural networks have been extensively used in most successful trackers.",28083
The mask in is random random scheme. used sampling,The random mask is used in random sampling scheme.,28084
"them bounds the with In and paper, different examples. compare we this propose some empirical","In this paper, we propose some different bounds and compare them with the empirical examples.",28085
"challenging to to resolution. scale is spatial high However, it","However, it is challenging to scale to high spatial resolution.",28086
detected multiplexing. is power significantly thanks the to increased The Hadamard,The detected power is significantly increased thanks to the Hadamard multiplexing.,28087
to the field navigation. two monocular novel of algorithm provides The aspects,The algorithm provides two novel aspects to the field of monocular navigation.,28088
region-based methods. achieved great using Face has detection success the,Face detection has achieved great success using the region-based methods.,28089
"fully network. the our (ResNet) backbone adopt Network the we approach, convolutional as In Residual","In our approach, we adopt the fully convolutional Residual Network (ResNet) as the backbone network.",28090
provides choosing more analysis in MIH parameter best application different This freedom scenarios. on for,This analysis provides more freedom on best parameter choosing in MIH for different application scenarios.,28091
This a data amount on training annotated allows manner. weakly of in limited a supervised,This allows training on a limited amount of annotated data in a weakly supervised manner.,28092
data results on microscopy fluorescent real Promising are presented. segmentation,Promising segmentation results on real fluorescent microscopy data are presented.,28093
Those relatively and jointly to are optimized. with other independent hard steps be each are,Those steps are relatively independent with each other and are hard to be jointly optimized.,28094
in issues hinder applications. large-scale These performance the of image model BoWs largely,These issues largely hinder the performance of BoWs model in large-scale image applications.,28095
such Access in information remote directions new up opens imagery. the analysis to of sensing,Access to such information opens up new directions in the analysis of remote sensing imagery.,28096
the introduce Augmented To Memory mitigate challenges we Network (MACN). these Control,To mitigate these challenges we introduce the Memory Augmented Control Network (MACN).,28097
proposed main of parts. network three consists architecture The,The proposed network architecture consists of three main parts.,28098
that plan to network to learns new environments. generalize our We show and can,We show that our network learns to plan and can generalize to new environments.,28099
Person people match across in Re-identification aims camera to public a space. views (re-id) non-overlapping,Person Re-identification (re-id) aims to match people across non-overlapping camera views in a public space.,28100
surveillance challenging because It is in problem clothes. wear people captured a videos many similar,It is a challenging problem because many people captured in surveillance videos wear similar clothes.,28101
"a contrast, is this model novel in deep paper, In learning proposed. multi-scale","In contrast, in this paper, a novel multi-scale deep learning model is proposed.",28102
is explicitly. extracting locations spatial also learned for discriminative different importance of features The,The importance of different spatial locations for extracting discriminative features is also learned explicitly.,28103
for Design generating with framework molecules a Automatic novel is optimized properties. Chemical,Automatic Chemical Design is a framework for generating novel molecules with optimized properties.,28104
"large consuming. such activities time important be of retrieval a from data can However,","However, retrieval of important activities from such a large data can be time consuming.",28105
of the clips. duration shorter activities one Video ways synopsis relatively represent is to using,Video synopsis is one of the ways to represent activities using relatively shorter duration clips.,28106
"trajectories same In situations, the produces multiple object. thus fails, of many tracking","In many situations, tracking fails, thus produces multiple trajectories of the same object.",28107
"discontinuity-preserving strategies. Hence, then several discuss we","Hence, we then discuss several discontinuity-preserving strategies.",28108
this information In approach of two-stage paper in we surveillance the systems. propose video organizing,In this paper we propose the two-stage approach of organizing information in video surveillance systems.,28109
"nontrivial task, is poses facial occlusions. have It expressions, faces may varying or a since","It is a nontrivial task, since faces may have varying facial expressions, poses or occlusions.",28110
of that peristaltic-like realm pumping. find flow We is within more the produced bulk,We find that more bulk flow is produced within the realm of peristaltic-like pumping.,28111
"significantly. have recent road In increased accidents years,","In recent years, road accidents have increased significantly.",28112
"major is driver these fatigue. the for of accidents, reported One reasons as","One of the major reasons for these accidents, as reported is driver fatigue.",28113
"of which installed on Thus, the system a propose we car dashboard. comprises camera a","Thus, we propose a system which comprises of a camera installed on the car dashboard.",28114
fatigue Support are obtained The parameters processed through Vector Classifier level. classifying for the thus,The parameters thus obtained are processed through Support Vector Classifier for classifying the fatigue level.,28115
to After sent classifier the that unit. is alert output,After that classifier output is sent to the alert unit.,28116
"alleviate In how to investigate SSD the framework. this starting we this from problem paper,","In this paper, we investigate how to alleviate this problem starting from the SSD framework.",28117
distribution the model multi-scale representations and StairNet final effectively. unifies semantic detector Our,Our final model StairNet detector unifies the multi-scale representations and semantic distribution effectively.,28118
discuss scheduling implementation practical for strategies. unified and a We retraining framework present,We present a unified framework for retraining scheduling and discuss practical implementation strategies.,28119
convex proposed show estimator We that fast the algorithm. be can by optimization computed a,We show that the proposed estimator can be computed by a fast convex optimization algorithm.,28120
"show over By the the procedures. proposed we simulations, superiority of method extensive comparable","By extensive simulations, we show the superiority of the proposed method over comparable procedures.",28121
on two the technique real apply We datasets.,We apply the technique on two real datasets.,28122
method substancially. been the this for required has time With computational dataset the reduced training,With this method the computational time required for training the dataset has been reduced substancially.,28123
a for extraction the filters feature used ensure performance. Gabor to were better,Gabor filters were used for the feature extraction to ensure a better performance.,28124
performance CGNN three studied The experiments. of throughout is,The performance of CGNN is studied throughout three experiments.,28125
"CGNN v-structures is conditional independences. to the problem identifying Secondly, and of applied","Secondly, CGNN is applied to the problem of identifying v-structures and conditional independences.",28126
depends recognition parts. visual typically from modeling object Fine-grained on subtle difference,Fine-grained visual recognition typically depends on modeling subtle difference from object parts.,28127
"preserve Finally, the discontinuities. depth should it","Finally, it should preserve the depth discontinuities.",28128
proposed demonstrate databases effectiveness the hierarchical on probabilistic of model. results the benchmark Experimental,Experimental results on benchmark databases demonstrate the effectiveness of the proposed hierarchical probabilistic model.,28129
method different structuring to structures. of proposed features elements The combines innate vessel-like detect,The proposed method combines different structuring elements to detect innate features of vessel-like structures.,28130
multi-person problem We of pose study the natural images. in estimation,We study the problem of multi-person pose estimation in natural images.,28131
"A the etc.) knee, (head, and pose identity position estimate describes foot, spatial","A pose estimate describes the spatial position and identity (head, foot, knee, etc.)",28132
body a every part of non-occluded of person.,of every non-occluded body part of a person.,28133
"sensors. set, the different across training and even","the training set, and even across different sensors.",28134
reduction. (PCA) to is component popular perform analysis dimension Principal very,Principal component analysis (PCA) is very popular to perform dimension reduction.,28135
"Bayesian purpose, To principal introduces nonparametric this this analysis a (BNP-PCA). component paper","To this purpose, this paper introduces a Bayesian nonparametric principal component analysis (BNP-PCA).",28136
proposed. of dimension subspace the estimator A new is,A new estimator of the subspace dimension is proposed.,28137
the first on is The behaviour various synthetic studied examples. algorithm of,The behaviour of the algorithm is first studied on various synthetic examples.,28138
nature broadcast systems medium. Inherent the is caused the to the multi-user interference by of,Inherent to multi-user systems is the interference caused by the broadcast nature of the medium.,28139
been the provided illustrate to proposed Simulations high have the scheme. gain of,Simulations have been provided to illustrate the high gain of the proposed scheme.,28140
"example negative, visual randomly A represented concept images. well is generated, positive as as by","A visual concept is represented by randomly generated, positive as well as negative, example images.",28141
"deliberate However, was tests understood, trivially the become easy. would the if concept all","However, if the concept was understood, all the deliberate tests would become trivially easy.",28142
than easier We some did are for concepts DCNN others. find that,We did find that some concepts are easier for DCNN than others.,28143
from active takes samples The role pool. the selecting instance learner an in,The learner takes an active role in selecting samples from the instance pool.,28144
and/or applications Our is observations motivated are expensive. real-world where problem by time-consuming,Our problem is motivated by real-world applications where observations are time-consuming and/or expensive.,28145
the also asymptotically show that We offline is an knowledge policy optimal setting. gradient in,We also show that the knowledge gradient policy is asymptotically optimal in an offline setting.,28146
This knowledge bandits. extends further setting contextual gradient of the work to the,This work further extends the knowledge gradient to the setting of contextual bandits.,28147
a report the its of that of efficiency. results series experiments demonstrate We,We report the results of a series of experiments that demonstrate its efficiency.,28148
"that paper breaks a Beyond limitations. cross-correlator, this proposes cross-correlator kernel linear the (KCC) traditional","Beyond the linear cross-correlator, this paper proposes a kernel cross-correlator (KCC) that breaks traditional limitations.",28149
the The one as in linearized regime problem framed buckling can be elasticity. of of,The problem can be framed as one of buckling in the regime of linearized elasticity.,28150
"extensive the regime. As morphomechanics brain consequence, an of the developing demonstrates post-bifurcation a","As a consequence, the morphomechanics of the developing brain demonstrates an extensive post-bifurcation regime.",28151
"have a unfolded, degree these high treatments reference, ellipsoidal. The of configurations in symmetry--typically,","The reference, unfolded, configurations in these treatments have a high degree of symmetry--typically, ellipsoidal.",28152
"boundary or conditions, folded patterns. symmetric the on the Depending configurations anti-symmetric have","Depending on the boundary conditions, the folded configurations have symmetric or anti-symmetric patterns.",28153
to engineering explain We approach CAISs. herein our,We herein explain our approach to engineering CAISs.,28154
"very this expressivity is of framework needed. high engineering, To a formal accomplish","To accomplish this engineering, a formal framework of very high expressivity is needed.",28155
models three variables. binary random We of statistical compare two,We compare two statistical models of three binary random variables.,28156
larger discuss models. to briefly We extensions,We briefly discuss extensions to larger models.,28157
a a using automatic social pattern for characterization wearable proposes photo-camera. This system paper,This paper proposes a system for automatic social pattern characterization using a wearable photo-camera.,28158
The steps. proposed of consists pipeline major three,The proposed pipeline consists of three major steps.,28159
This for visual paper detection top-down presents an by tasks. classification guided approach saliency,This paper presents an approach for top-down saliency detection guided by visual classification tasks.,28160
"to to Thereafter, are CRFs the used segmentation dense appearance. according refine edema the","Thereafter, dense CRFs are used to refine the segmentation according to the edema appearance.",28161
methods deblurring pleasing may show super-resolution visually off-the-shelf or results. The,The off-the-shelf deblurring or super-resolution methods may show visually pleasing results.,28162
is in overview Fig. system shown A,A system overview is shown in Fig.,28163
"on research super-resolution (SR). much In image conducted has been years, recent","In recent years, much research has been conducted on image super-resolution (SR).",28164
"of with knowledge, methods were concerned however, compressed best few the SR our images. To","To the best of our knowledge, however, few SR methods were concerned with compressed images.",28165
"In to enforcing necessarily. we compile new a determinism without this paper, method propose DNNFs","In this paper, we propose a new method to compile DNNFs without enforcing determinism necessarily.",28166
"neuromodulators the ultimately behaviour. affect populations, neuronal controlling By state of","By controlling the state of neuronal populations, neuromodulators ultimately affect behaviour.",28167
neuromodulation a is candidate of computational networks. large in good studies for therefore affordable It,It is therefore a good candidate for affordable computational studies of neuromodulation in large networks.,28168
"this the study, independent In summarize al. Lanckriet will work G. of et I","In this independent study, I will summarize the work of G. Lanckriet et al.",28169
forms of information investigated. privileged were Different,Different forms of privileged information were investigated.,28170
adaptive mesh-networked agents. D-RHC of control use swarm this for We method,We use this adaptive D-RHC method for control of mesh-networked swarm agents.,28171
our on an the method example validate Using we exploration task. simulator coordinated,Using the simulator we validate our method on an example coordinated exploration task.,28172
community our simulator work. future to We and for the release code,We release our simulator and code to the community for future work.,28173
external Cells signals process internal chemical and through interactions.,Cells process external and internal signals through chemical interactions.,28174
"to is this modeling Kappa, address (BioNetGen, complexity. Simmune, approach PySB) an Rule-based","Rule-based modeling (BioNetGen, Kappa, Simmune, PySB) is an approach to address this complexity.",28175
and able to Our moderately holes. vertices sized method predict missing inpaint is,Our method is able to predict missing vertices and inpaint moderately sized holes.,28176
We textures. patch surface successful show complicated shape that performs completion our of dictionary,We show that our patch dictionary performs successful shape completion of complicated surface textures.,28177
"data In techniques. learning machine functions image learning, training are estimated operator these from using","In image operator learning, these functions are estimated from training data using machine learning techniques.",28178
use convolutional this neural We the of limitation. overcome networks (CNNs) propose to,We propose the use of convolutional neural networks (CNNs) to overcome this limitation.,28179
TRIOSlib site. project The will the be implementations made on available,The implementations will be made available on the TRIOSlib project site.,28180
Stereo These images CMOSIS/AWAIBA. the by with manufactured NanEye are system acquired,These images are acquired with the system NanEye Stereo manufactured by CMOSIS/AWAIBA.,28181
radial the we analysis of distortion comparison on and models. focused particular In,In particular we focused on the analysis and comparison of radial distortion models.,28182
a model concluded parameter We radial the distortion has division that limitations. single with,We concluded that the division model with a single radial distortion parameter has limitations.,28183
hierarchical performing are after grouped clustering tasks on correlation. Individual based their,Individual tasks are grouped after performing hierarchical clustering based on their correlation.,28184
setup. The within learning is classification process performed multi-task in cluster a each,The learning process within each cluster is performed in a multi-task classification setup.,28185
"knowledge, By speed-up we acquired improve and process the performance. the leveraging","By leveraging the acquired knowledge, we speed-up the process and improve performance.",28186
a Segmenting hard leaves images is from problem. natural,Segmenting leaves from natural images is a hard problem.,28187
complex are and of leaves other Cluttered backgrounds: commonplace. often composed,Cluttered and complex backgrounds: often composed of other leaves are commonplace.,28188
"and their appearance perspective. dependent highly Furthermore, illumination viewing upon is","Furthermore, their appearance is highly dependent upon illumination and viewing perspective.",28189
"colour, Background and markers are texture. and using leaf created intensity","Background and leaf markers are created using colour, intensity and texture.",28190
approaches graph-cut results compared. investigated: and and watershed Two are,Two approaches are investigated: watershed and graph-cut and results compared.,28191
removal vein leaf. a protrusion-notch refine to detection extracted are applied and the Primary-secondary,Primary-secondary vein detection and a protrusion-notch removal are applied to refine the extracted leaf.,28192
The demonstrated against approach work. existing is of our efficacy,The efficacy of our approach is demonstrated against existing work.,28193
These discussed by Soares difficulties parallel al. et those,These difficulties parallel those discussed by Soares et al.,28194
relays training full-duplex one-way estimation optimal investigated. sequence are Channel and design for,Channel estimation and optimal training sequence design for full-duplex one-way relays are investigated.,28195
algorithm. is (BFGS) A estimator implemented with maximum likelihood Broyden-Fletcher-Goldfarb-Shanno,A maximum likelihood estimator is implemented with Broyden-Fletcher-Goldfarb-Shanno (BFGS) algorithm.,28196
"RSI, presence channel. overall channel of becomes an source-to-destination the inter-symbol-interference the (ISI) In","In the presence of RSI, the overall source-to-destination channel becomes an inter-symbol-interference (ISI) channel.",28197
minimizing obtained optimal sequence training CRB. The by the is,The optimal training sequence is obtained by minimizing the CRB.,28198
using Simulations our corroborate LTE theoretical results. parameters,Simulations using LTE parameters corroborate our theoretical results.,28199
problem. formulations a by couple of of the We start deriving unconstrained,We start by deriving a couple of unconstrained formulations of the problem.,28200
"faster the a Thus, assumption performed occlusions. is detection with of no","Thus, a faster detection is performed with the assumption of no occlusions.",28201
of the the backbone most local been have of computer vision The problems. descriptors,The local descriptors have been the backbone of most of the computer vision problems.,28202
raw are descriptors input images. the Most the of existing local over generated,Most of the existing local descriptors are generated over the raw input images.,28203
is pattern A decoders. frequency two local proposed (FDLBP) with decoded binary,A frequency decoded local binary pattern (FDLBP) is proposed with two decoders.,28204
high frequency low with and one two works patterns. frequency decoder pattern Each,Each decoder works with one low frequency pattern and two high frequency patterns.,28205
"Finally, to descriptor. are both the form descriptors concatenated decoders single from the","Finally, the descriptors from both decoders are concatenated to form the single descriptor.",28206
"paper, for pipeline visual propose multi-camera system. In multi-target under this a tracking we","In this paper, we propose a pipeline for multi-target visual tracking under multi-camera system.",28207
"the multi-camera algorithms emphasis However, association. camera tracking data across most frame single of on","However, most of the multi-camera tracking algorithms emphasis on single camera across frame data association.",28208
"Local conducted. Maximal for is Occurrence For feature ReID algorithm Representation(LOMO) feature, appearance extraction","For appearance feature, Local Maximal Occurrence Representation(LOMO) feature extraction algorithm for ReID is conducted.",28209
"scene Access critical for RGB-D training RGB-D to is algorithms. datasets understanding diverse large,","Access to large, diverse RGB-D datasets is critical for training RGB-D scene understanding algorithms.",28210
"VQA apply an human-like we attention to attention-based supervision Finally, model.","Finally, we apply human-like attention supervision to an attention-based VQA model.",28211
We a provide using algorithm Rcpp the implementation interface. C++ of,We provide a C++ implementation of the algorithm using Rcpp interface.,28212
recognition a face Most or on works modules the specific focus of idea. research demonstrate,Most of the face recognition works focus on specific modules or demonstrate a research idea.,28213
"- where each model naming we Following e.g. propose this, colour biologically-inspired term a colour","Following this, we propose a biologically-inspired colour naming model where each colour term - e.g.",28214
term by calculated is The for a mechanism. maximum final colour a given pooling pixel,The final colour term for a given pixel is calculated by a maximum pooling mechanism.,28215
a include offers further extension straightforward to It also colour terms.,It also offers a straightforward extension to include further colour terms.,28216
We for handwritten propose a online graphics. of the new framework recognition,We propose a new framework for the recognition of online handwritten graphics.,28217
a We grammar. model as generated graph from graphic a graph a labeled,We model a graphic as a labeled graph generated from a graph grammar.,28218
recognition We applications of and show expressions the mathematical to flowcharts.,We show applications to the recognition of mathematical expressions and flowcharts.,28219
both that accuracy in method shows our Experimentation obtains state-of-the-art applications.,Experimentation shows that our method obtains state-of-the-art accuracy in both applications.,28220
video. from extract A used is a convolutional to neural features frame network level,A convolutional neural network is used to extract frame level features from a video.,28221
frame using a features aggregated memory. long are then convolutional The short-term level,The frame level features are then aggregated using a convolutional long short-term memory.,28222
"apply saliency few DNNs videos. predict works the However, generic that of are to there","However, there are few works that apply DNNs to predict the saliency of generic videos.",28223
"video saliency prediction method. we In a this propose novel paper, DNN-based","In this paper, we propose a novel DNN-based video saliency prediction method.",28224
"show experimental that video in Finally, our saliency results method the advances prediction. the state-of-the-art","Finally, the experimental results show that our method advances the state-of-the-art in video saliency prediction.",28225
becomes models many applications. relevant to analyze in networks able computational Designing evolving,Designing computational models able to analyze evolving networks becomes relevant in many applications.,28226
"techniques. many statisticians inferential and cause developed difficulties the have To of relating address effect,","To address the difficulties of relating cause and effect, statisticians have developed many inferential techniques.",28227
"Then, consistency directions from database, in subjects. we our high different find a across viewing","Then, from our database, we find a high consistency in viewing directions across different subjects.",28228
The benchmark. near-peak Middlebury achieves method proposed the performance on,The proposed method achieves near-peak performance on the Middlebury benchmark.,28229
"performance. optimized implementation is real-time be for simple enough can the Meanwhile, that","Meanwhile, the implementation is simple enough that can be optimized for real-time performance.",28230
development Its statistics the inevitable. modern impact is on of,Its impact on the development of modern statistics is inevitable.,28231
the bipartite task. extracted subgraphs EMKN each from to were Three support (bi-graphs),Three bipartite subgraphs (bi-graphs) were extracted from the EMKN to support each task.,28232
field support was Each the inference. regarded as random Markov to bi-graph a,Each bi-graph was regarded as a Markov random field to support the inference.,28233
Three were measures for evaluation. performance utilized,Three measures were utilized for performance evaluation.,28234
system These methods. the baseline that outperformed results altogether proposed indicate the,These results altogether indicate that the proposed system outperformed the baseline methods.,28235
to medical representation reflect regards The similarity of does distributed entities knowledge relationships level. in,The distributed representation of medical entities does reflect similarity relationships in regards to knowledge level.,28236
"require functions individually. however, energy Different tasks, their designing","Different tasks, however, require designing their energy functions individually.",28237
"tion mechanism we embedding for pact during Besides, representations. com- more atten- assemble achieving","Besides, we assemble atten- tion mechanism during embedding for achieving more com- pact representations.",28238
in possible. resolution Advances sensors made image high CMOS technology have,Advances in CMOS technology have made high resolution image sensors possible.,28239
 science has in issue scrutiny. The under come increased  reproducibility,  The reproducibility issue in science has come under increased scrutiny.,28240
methods use data consistent workflows scripted lies suggestion analysis. or in of the for One,One consistent suggestion lies in the use of scripted methods or workflows for data analysis.,28241
"process been date, is but not successful This to has complete.","This process has been successful to date, but is not complete.",28242
"example method basic component and is each code is discussed, The for shown.","The basic method for each component is discussed, and example code is shown.",28243
class representation different concept analysis models. within a knowledge contains of paper This of object-oriented,This paper contains analysis of concept of a class within different object-oriented knowledge representation models.,28244
"future model the first frame, long-term frames. Given learns generate our to","Given the first frame, our model learns to generate long-term future frames.",28245
generates each The frame. videos stage contents of for first realistic,The first stage generates videos of realistic contents for each frame.,28246
"We large time-lapse and new a scale build this on our dataset. test dataset, approach","We build a large scale time-lapse dataset, and test our approach on this new dataset.",28247
"embolism CT angiography. pulmonary (PE) by currently a disease, assessed Pulmonary is highly mortal","Pulmonary embolism (PE) is a highly mortal disease, currently assessed by pulmonary CT angiography.",28248
second for method proposed facial expression recognition the application. as We a cross-view applied also,We also applied the proposed method for cross-view facial expression recognition as a second application.,28249
the both proposed method effectiveness for experimental of The demonstrate results the applications.,The experimental results demonstrate the effectiveness of the proposed method for both applications.,28250
landmark effective facial Cascade framework for to has be been regression shown detection.,Cascade regression framework has been shown to be effective for facial landmark detection.,28251
"(a.k.a. on spatial question (VQA), different which from answering is relies problem This attention visual","This problem is different from visual question answering (VQA), which relies on spatial attention (a.k.a.",28252
image question from pair. visual estimated and an grounding),visual grounding) estimated from an image and question pair.,28253
"skin reflectance premature, study cross-sectional evaluated selected preferably of A the infants, birth. at","A cross-sectional study evaluated the skin reflectance of selected infants, preferably premature, at birth.",28254
for the first-trimester The age. reference gestational ultrasound was,The first-trimester ultrasound was the reference for gestational age.,28255
"An of pilot sends superimposed (SP), called and method, data a pilot alternative symbols. superposition","An alternative method, called superimposed pilot (SP), sends a superposition of pilot and data symbols.",28256
"longer which, allows pilots This contamination. to use pilot in turn, reduces","This allows to use longer pilots which, in turn, reduces pilot contamination.",28257
convolutional on remarkable image neural (CNNs) networks classification achieve performance tasks. Deep,Deep convolutional neural networks (CNNs) achieve remarkable performance on image classification tasks.,28258
"called is network a proposed. SwGridNet neural new Herein,","Herein, a new neural network called SwGridNet is proposed.",28259
"SwGridNet of in described the paper, As details this architecture are presented. network","As described in this paper, details of the SwGridNet network architecture are presented.",28260
CNNs. SwGridNet results approximates that that The state-of-the-art performance indicate of the deep,The results indicate that the SwGridNet performance approximates that of state-of-the-art deep CNNs.,28261
"sensing, fundamental Automatic in e.g. a classification is remote land problem urban cover","Automatic urban land cover classification is a fundamental problem in remote sensing, e.g.",28262
"variance. challenging, highly high generally inter-class problem low intra-class classes as The have and is","The problem is highly challenging, as classes generally have high inter-class and low intra-class variance.",28263
"test in time, well as training. as at","at test time, as well as in training.",28264
a information. This of useful potentially represents waste,This represents a waste of potentially useful information.,28265
by DSM assuming are images during testing. missing that We missing modalities simulate,We simulate missing modalities by assuming that DSM images are missing during testing.,28266
two the similarity is networks in Quantifying many critical applications. between,Quantifying the similarity between two networks is critical in many applications.,28267
to also applied real was visual brain recognition task. SimiNet a networks obtained during,SimiNet was also applied to real brain networks obtained during a visual recognition task.,28268
outperforms the demonstrate on state-of-the-art Experiment that consistently methods method results benchmarks. our,Experiment results demonstrate that our method consistently outperforms state-of-the-art methods on the benchmarks.,28269
for present suite a perception. benchmark visual We,We present a benchmark suite for visual perception.,28270
all for every tasks Ground-truth is data frame. available for,Ground-truth data for all tasks is available for every frame.,28271
The experiments. via realism is of validated collected data further perceptual the,The realism of the collected data is further validated via perceptual experiments.,28272
IMFs is eliminated de-noising the of soft wavelet by technique. The noise the threshold,The noise of the IMFs is eliminated by the wavelet soft threshold de-noising technique.,28273
(PC) sparse chaos functions. Compressive widely of polynomial for used been approximation has sampling stochastic,Compressive sampling has been widely used for sparse polynomial chaos (PC) approximation of stochastic functions.,28274
The compare defeats systems i.e. are: we non-prioritized,The systems we compare are: non-prioritized defeats i.e.,28275
"reverse defeat. and extended preference-based with attacks, defeats, defeats preference-based","attacks, preference-based defeats, and preference-based defeats extended with reverse defeat.",28276
effectively and vessels visualize in Our reconstruct following system steps. can,Our system can effectively reconstruct and visualize vessels in following steps.,28277
extracting centerlines generating trees. parse by their vessels We segmented and the,We parse the segmented vessels by extracting their centerlines and generating trees.,28278
"After training, output take input, or as answer can then counterpart. the question either iQAN","After training, iQAN can take either question or answer as input, then output the counterpart.",28279
with Antenna MIMO. performance is avoid severe necessary the massive to calibration degradation,Antenna calibration is necessary to avoid the severe performance degradation with massive MIMO.,28280
"interconnection calibration an First, study arbitrary for resulting we the performance strategy.","First, we study the resulting calibration performance for an arbitrary interconnection strategy.",28281
to are simulation theoretical and results Numerical results. analyses included also corroborate our,Numerical simulation results are also included to corroborate our theoretical analyses and results.,28282
detect a or retouched. algorithms original is image researchers have introduced Several to whether face,Several researchers have introduced algorithms to detect whether a face image is original or retouched.,28283
"using are images created packages. retouching software Further, two different retouched","Further, retouched images are created using two different retouching software packages.",28284
generalized task The detection. approach algorithms outperforms of state-of-the-art the for retouching detection existing proposed,The proposed approach outperforms existing state-of-the-art detection algorithms for the task of generalized retouching detection.,28285
fairness for seems irrational. Masses individual experiments which shown preference of have,Masses of experiments have shown individual preference for fairness which seems irrational.,28286
it The research. focus behind reason remains for a,The reason behind it remains a focus for research.,28287
"inhibits promotes sufficient and altruism; spite. The promotes migration selfishness migration fairness, of and existence","The existence of migration inhibits selfishness and fairness, and promotes altruism; sufficient migration promotes spite.",28288
origin into may provide of insights important study Our the fairness. evolutionary,Our study may provide important insights into the evolutionary origin of fairness.,28289
"improvements images. facial detection have There ""in-the-wild"" on landmark tremendous been general for","There have been tremendous improvements for facial landmark detection on general ""in-the-wild"" images.",28290
"In them. existing fact, only one algorithms of the handle usually can","In fact, the existing algorithms usually can only handle one of them.",28291
"predicts the Specifically, the method and landmark landmark iteratively locations. occlusions the","Specifically, the method iteratively predicts the landmark occlusions and the landmark locations.",28292
"comparable images. also is general other It ""in-the-wild"" on methods to","It is also comparable to other methods on general ""in-the-wild"" images.",28293
"usually ignoring interactions. and The task existing methods sequentially, independently each their perform","The existing methods usually perform each task independently and sequentially, ignoring their interactions.",28294
by important the images regions. recognizability of preserving effectively Image resizes retargeting image,Image retargeting effectively resizes images by preserving the recognizability of important image regions.,28295
"traditional exhaustively In user ratings. the addition, evaluation on depends","In addition, the traditional evaluation exhaustively depends on user ratings.",28296
approach for is There results. evaluating a for need a retargeted legitimate methodological,There is a legitimate need for a methodological approach for evaluating retargeted results.,28297
"we for as task. benchmark Second, dataset salient this a object exploit","Second, we exploit salient object dataset as a benchmark for this task.",28298
importance then types We particular different this investigate for of maps problem.,We then investigate different types of importance maps for this particular problem.,28299
"object [i.e. for fact, In single-label classification","In fact, for single-label object classification [i.e.",28300
"pixelwise complex address multi-label problem of we classification. this paper, the In","In this paper, we address the complex problem of multi-label pixelwise classification.",28301
for six scale. pixelwise results classifications This in different each,This results in six pixelwise classifications for each different scale.,28302
into to An single-label. six trained the SVM pixelwise classifications is map a then,An SVM is then trained to map the six pixelwise classifications into a single-label.,28303
results the proposed The reported. and been tested approach extensively has are,The proposed approach has been extensively tested and the results are reported.,28304
correlative is prediction the and roads from transportation by data vehicles. Such real-time analyzing possible,Such prediction is possible by analyzing the real-time transportation data from correlative roads and vehicles.,28305
The approach introduced et in Gon\c{c}alves has been al.,The approach has been introduced in Gon\c{c}alves et al.,28306
"nonmonotonic a represent provides formal argumentation Meanwhile, to logics. way","Meanwhile, formal argumentation provides a way to represent nonmonotonic logics.",28307
"prioritized normative representation this In we paper, by of reasoning a propose argumentation.","In this paper, we propose a representation of prioritized normative reasoning by argumentation.",28308
face is) The Viola-Jones quite detection detector. algorithm was a (and popular face still,The Viola-Jones face detection algorithm was (and still is) a quite popular face detector.,28309
"on two specifically aspects faces improvement, illuminated of faces. badly focus blurred We and","We focus on two aspects of improvement, specifically badly illuminated faces and blurred faces.",28310
blind detector. We want using that methods avoid the pre-processing obstruct may to face,We want to avoid using blind pre-processing methods that may obstruct the face detector.,28311
"we that two experiments. sets of end, To perform","To that end, we perform two sets of experiments.",28312
"compact a environmental of system VPR necessitates representations practical Moreover, features.","Moreover, a practical VPR system necessitates compact representations of environmental features.",28313
image compact descriptors. setting) The (in supervised feature system modified binary a learns codes from,The modified system learns (in a supervised setting) compact binary codes from image feature descriptors.,28314
"We data sets covering performed and have viewpoint on challenging illumination several experiments seasonal, variations.","We have performed experiments on several challenging data sets covering seasonal, illumination and viewpoint variations.",28315
a approach present this novel facial ethnicity from we extracting the images. paper for In,In this paper we present a novel approach for extracting ethnicity from the facial images.,28316
"as we ethnicity problem a Here African-American class and including classification Caucasian. consider three Asian,","Here we consider ethnicity classification as a three class problem including Asian, African-American and Caucasian.",28317
two tasks processing. are fundamental image image in denoising and segmentation Unsupervised,Unsupervised image segmentation and denoising are two fundamental tasks in image processing.,28318
Our both time. approach problems addresses same the at,Our approach addresses both problems at the same time.,28319
by solved standard ILP a The solver. then MIP is off-the-shelf,The ILP is then solved by a standard off-the-shelf MIP solver.,28320
experiments Numerical problem. with are the multicut compared,Numerical experiments are compared with the multicut problem.,28321
the sensing spatiotemporal by specific exploit collected among correlations data designs Existing a modality.,Existing designs exploit the spatiotemporal correlations among data collected by a specific sensing modality.,28322
"monitoring, that correlated. data However, intrinsically heterogeneous as environmental involve are many such applications, collecting","However, many applications, such as environmental monitoring, involve collecting heterogeneous data that are intrinsically correlated.",28323
example to application illustrative making An examined. with practical decision is business,An illustrative example with practical application to business decision making is examined.,28324
representation real-world wide which diversity scenarios. data of a in appears is an Graph important,Graph is an important data representation which appears in a wide diversity of real-world scenarios.,28325
"cost. computation methods analytics graph suffer the most space However, high and","However, most graph analytics methods suffer the high computation and space cost.",28326
an to graph efficient Graph analytics problem. the way effective yet is embedding solve,Graph embedding is an effective yet efficient way to solve the graph analytics problem.,28327
"graph review comprehensive of we literature survey, In this in a conduct the embedding.","In this survey, we conduct a comprehensive review of the literature in graph embedding.",28328
related graph as well introduce We concepts. formal embedding the the of first as definition,We first introduce the formal definition of graph embedding as well as the related concepts.,28329
the Then makeup input attributes seamlessly are the synthesized onto facial recommended image.,Then the recommended makeup attributes are seamlessly synthesized onto the input facial image.,28330
leading the causes of is cancer of Liver death. one cancer,Liver cancer is one of the leading causes of cancer death.,28331
"to presented facilitate regard, of grid technologies RES. have this higher In been penetration smart","In this regard, smart grid technologies have been presented to facilitate higher penetration of RES.",28332
smart Microgrids the of components key the are grids.,Microgrids are the key components of the smart grids.,28333
introduces reliability the RES stochasticity evaluation. to of complexity The,The stochasticity of RES introduces complexity to the reliability evaluation.,28334
method of simulation. takes the into Carlo state through The variability RES account Monte- sampling,The method takes into account the variability of RES through Monte- Carlo state sampling simulation.,28335
"priority improvement highest In particular, has the in indices. the largest the reliability load","In particular, the highest priority load has the largest improvement in the reliability indices.",28336
memetic block a search is Local building algorithms. basic in,Local search is a basic building block in memetic algorithms.,28337
Guided Search of the improve Local efficiency search. (GLS) can local,Guided Local Search (GLS) can improve the efficiency of local search.,28338
use mechanism. to big GLS the penalizing assumption paper This to proposes improve valley structure,This paper proposes to use big valley structure assumption to improve the GLS penalizing mechanism.,28339
Biased called Elite proposed. algorithm is GLS GLS An improved (EB-GLS),An improved GLS algorithm called Elite Biased GLS (EB-GLS) is proposed.,28340
algorithm systematically symmetric tested on We have problem. the salesman traveling the proposed,We have systematically tested the proposed algorithm on the symmetric traveling salesman problem.,28341
that show significantly better is GLS. results than Experimental EB-GLS,Experimental results show that EB-GLS is significantly better than GLS.,28342
and can very ease generalization deep the for models. optimization performance the improve It CNN,It can ease the optimization and improve the generalization performance for very deep CNN models.,28343
inform defenses. work future observations building will These towards effective,These observations will inform future work towards building effective defenses.,28344
paper. part also in of later of the the state model art describe We current,We also describe the current state of the art model in later part of paper.,28345
"is decision use to In it these cases, all advisable i.e. test, robust a","In all these cases, it is advisable to use a robust decision test, i.e.",28346
transmitted appropriately needs reliably Information media. to to over physical be encoded be,Information needs to be appropriately encoded to be reliably transmitted over physical media.,28347
"in information own neurons have to Similarly, their convey the brain. codes","Similarly, neurons have their own codes to convey information in the brain.",28348
critical function is chirp well chirplet describing in Gaussian understood is It that signals.,It is well understood that Gaussian chirplet function is critical in describing chirp signals.,28349
"the describe on paper, we ACT In algorithm ""coarse-refinement"" novel based this scheme. a","In this paper, we describe a novel ACT algorithm based on the ""coarse-refinement"" scheme.",28350
"hardware neuromorphic vision on-sensor implementation In present spatial denoising we of filter. this paper,","In this paper, we present on-sensor neuromorphic vision hardware implementation of denoising spatial filter.",28351
an effect The with blurring increases of size. window increase in,The effect of blurring increases with an increase in window size.,28352
in shift register. the serial-in-parallel-out pixels sequence is for of responses stored neuron's the The,The sequence of the neuron's responses for pixels is stored in the serial-in-parallel-out shift register.,28353
"expanders, motor/generator a of CAES plant consists A caverns, compressors, and set.","A CAES plant consists of compressors, expanders, caverns, and a motor/generator set.",28354
directly models optimization Highly cannot power be utilized in cavern problems. non-linear system,Highly non-linear cavern models cannot be directly utilized in power system optimization problems.,28355
convolutional deep has (BN) de for a training normalization facto become standard networks. Batch,Batch normalization (BN) has become a de facto standard for training deep convolutional networks.,28356
"normalization with weight and normalization propagation translated ReLU. weight normalization,","weight normalization, normalization propagation and weight normalization with translated ReLU.",28357
"could is algorithms BN not replace in if practical, these However, large-scale clear applications. it","However, it is not clear if these algorithms could replace BN in practical, large-scale applications.",28358
lossy design architectures deep compression. image We the of study for,We study the design of deep architectures for lossy image compression.,28359
similarity given involves images looking a Image image. reference similar fetching,Image similarity involves fetching similar looking images given a reference image.,28360
"weather are model the addition, forecasting influential selected. predictors In energy most in","In addition, the most influential weather predictors in energy forecasting model are selected.",28361
translation on enables document also Word a We are now editing. working that to,We are now working on translation to a Word document that also enables editing.,28362
data with Mixed Reality scene. the Medical intraoperative surgical of surgeons video helps to contextualize,Medical Mixed Reality helps surgeons to contextualize intraoperative data with video of the surgical scene.,28363
"fluoroscope example scenario, use C-arm As an augmented we reality an device.","As an example scenario, we use an augmented reality C-arm fluoroscope device.",28364
"occluding objects The scene. can, the surgical become over transparent therefore,","The occluding objects can, therefore, become transparent over the surgical scene.",28365
the also application is first of Mixed for Reality/ Reality. This field Diminished volumetric work,This work is also the first application of volumetric field for Diminished Reality/ Mixed Reality.,28366
"objects, and key subcategories local. of regions discriminative among differences on Reflecting different subtle are","Reflecting on discriminative regions of objects, key differences among different subcategories are subtle and local.",28367
is two important challenging address to these limitations simultaneously. highly It,It is highly challenging to address these two important limitations simultaneously.,28368
Existing one of on focus them. only methods,Existing methods only focus on one of them.,28369
of method types modelled. the coherence concerns of the novelty The,The novelty of the method concerns the types of coherence modelled.,28370
"their adjusted view of Moreover, are adequacy contributions assessed each accordingly. and is","Moreover, adequacy of each view is assessed and their contributions are adjusted accordingly.",28371
events problem addresses This abnormal joint of videos. the recounting in and detection of paper,This paper addresses the problem of joint detection and recounting of abnormal events in videos.,28372
matching person two and Feature are extraction (ReID). crucial components Re-Identification in,Feature extraction and matching are two crucial components in person Re-Identification (ReID).,28373
conditions. simulation The proposed existing the under the are and estimators same compared ones,The proposed estimators and the existing ones are compared under the same simulation conditions.,28374
include The length performance and interval for coverage the estimators confidence measures probability.,The performance measures for the estimators include confidence interval length and coverage probability.,28375
on and datasets proposed validate synthetic Experiments strategy. real out carried both the are to,Experiments on both synthetic and real datasets are carried out to validate the proposed strategy.,28376
and expressions. The neutral faces to facial is registration applied,The registration is applied to neutral faces and facial expressions.,28377
of is (UC) operation power important the problems. one most Unit commitment system,Unit commitment (UC) is one of the most important power system operation problems.,28378
considered. of minimum between CAES processes charging is discharging The switch and time the,The minimum switch time between the charging and discharging processes of CAES is considered.,28379
used are model. piecewise whole to and McCormick relaxation the then linearization linearize,McCormick relaxation and piecewise linearization are then used to linearize the whole model.,28380
the new These of results feature detectors. insights behavior provide into,These results provide new insights into the behavior of feature detectors.,28381
"years, been or object developed. detectors have near recent real-time several In real-time","In recent years, several real-time or near real-time object detectors have been developed.",28382
The modified detector several on to RCNN is aerial Faster compared imagery datasets.,The modified detector is compared to Faster RCNN on several aerial imagery datasets.,28383
has evaluated biometric various databases. been available huge The over model publicly,The model has been evaluated over various huge publicly available biometric databases.,28384
treatment paper transition considers effects of conditional This identification probabilities. on the,This paper considers the identification of treatment effects on conditional transition probabilities.,28385
average on Instead these effects. we bounds treatment derive informative,Instead we derive informative bounds on these average treatment effects.,28386
"impose do restrictions, hazards. bounds example, (semi)parametric Our proportional for not","Our bounds do not impose (semi)parametric restrictions, for example, proportional hazards.",28387
proposed network. the to train algorithms also learning Appropriate are,Appropriate learning algorithms are also proposed to train the network.,28388
"tensor-based against dimensionality are Furthermore, reduction. in evaluated their classfiers proposed the capabilities","Furthermore, the proposed tensor-based classfiers are evaluated against their capabilities in dimensionality reduction.",28389
are analysis of highly correlated. LH subbands coefficients suggests that the wavelet and HL Our,Our analysis suggests that the wavelet coefficients of HL and LH subbands are highly correlated.,28390
"the we coefficients. Hence, sparsify decomposition packet further Mallat to decorrelate wavelet","Hence, we decorrelate Mallat wavelet packet decomposition to further sparsify the coefficients.",28391
Weights approach cluster to a Adaptive presents (AWC). Clustering non-parametric This paper called new analysis,This paper presents a new approach to non-parametric cluster analysis called Adaptive Weights Clustering (AWC).,28392
method and scalable is The computationally feasible.,The method is scalable and computationally feasible.,28393
inhomogeneity. of optimal AWC study to theoretical local states sensitivity Our,Our theoretical study states optimal sensitivity of AWC to local inhomogeneity.,28394
"the to be model this misspecified. When mismatched or said is happens,","When this happens, the model is said to be mismatched or misspecified.",28395
"Therefore, for effective decomposition. registration an approach LRS multi-view proposes paper by the this weighted","Therefore, this paper proposes an effective approach for multi-view registration by the weighted LRS decomposition.",28396
the can of reduced of LRS robustness improve sparsity The reconstructed matrix decomposition.,The reduced sparsity of reconstructed matrix can improve the robustness of LRS decomposition.,28397
The be whole end-to-end. network can trained,The whole network can be trained end-to-end.,28398
"avoids the It accelerates processing intermediate not accumulation, error only speed. but also","It not only avoids intermediate error accumulation, but also accelerates the processing speed.",28399
effectiveness our proposed experiments the show Extensive and of approach. efficiency,Extensive experiments show the effectiveness and efficiency of our proposed approach.,28400
This samples. smoke synthetic detection for video a smoke using method proposes paper,This paper proposes a method for video smoke detection using synthetic smoke samples.,28401
precise rich The samples. offer and data annotated virtual automatically can,The virtual data can automatically offer precise and rich annotated samples.,28402
The extracted from the researches annotated on samples work samples. original adaptation to existed mainly,The existed researches mainly work on the adaptation to samples extracted from original annotated samples.,28403
as two take and These object methods detection adaptation parts. the independent domain,These methods take the object detection and domain adaptation as two independent parts.,28404
The procedure is stage. end-to-end an training,The training procedure is an end-to-end stage.,28405
"classification, and are location the learning. The adaptation combined in","The classification, location and adaptation are combined in the learning.",28406
The experiments. the proposed baseline original surpasses the model of in performance our,The performance of the proposed model surpasses the original baseline in our experiments.,28407
"detector is Appendix domain for in A. described the two-stage adaptation Moreover,","Moreover, the domain adaptation for two-stage detector is described in Appendix A.",28408
as been way traffic relieve has a congestion. recently traffic Predicting conditions explored to,Predicting traffic conditions has been recently explored as a way to relieve traffic congestion.,28409
"In spatial introduced attention is information. align addition, temporal mechanism to and","In addition, attention mechanism is introduced to align spatial and temporal information.",28410
outperforms traditional and deep statistical a methods method. state-of-the-art It significantly learning,It significantly outperforms traditional statistical methods and a state-of-the-art deep learning method.,28411
on performance potential the point of optimizing out mobile the ways findings Our devices.,Our findings point out the potential ways of optimizing the performance on mobile devices.,28412
"resource requirements different Second, CNN of we the the model computations.","Second, we model the resource requirements of the different CNN computations.",28413
stress feature deep a is The characterized drought network with aggregation. based,The drought stress is characterized with a deep network based feature aggregation.,28414
aspect an learning. important reinforcement Reward engineering of is,Reward engineering is an important aspect of reinforcement learning.,28415
to This world. exploration systems physical the with expensive interact be requires operation when can,This operation can be expensive when exploration requires systems to interact with the physical world.,28416
proposed to conducted are the simulated of approach. evaluate A experiments set,A set of simulated experiments are conducted to evaluate the proposed approach.,28417
is demanding a very area of research. Face still recognition,Face recognition is still a very demanding area of research.,28418
task. for are Local used this descriptors widely,Local descriptors are widely used for this task.,28419
descriptors information mainly to increased based dimensionality. the suffer local The due wider,The wider local information based descriptors mainly suffer due to the increased dimensionality.,28420
inherently illumination invariant. descriptor The uniform is,The descriptor is inherently uniform illumination invariant.,28421
descriptor. to ability further adapted of boost the also multi-scale is The mechanism discriminative the,The multi-scale mechanism is also adapted to further boost the discriminative ability of the descriptor.,28422
proposed image retrieval databases. face descriptor under over framework evaluated is the The,The proposed descriptor is evaluated under the image retrieval framework over face databases.,28423
"face DLib deep scenarios. it the also Moreover, in outperforms many learning based descriptor","Moreover, it also outperforms the deep learning based DLib face descriptor in many scenarios.",28424
era. major become Surveillance has Vision in on necessity based Computer current a,Surveillance based on Computer Vision has become a major necessity in current era.,28425
mass determining of critical both in that show variation. are Results factors patterns seed,Results show that both factors are critical in determining patterns of seed mass variation.,28426
this approach presented multifocus is in paper. new A fusion image,A new multifocus image fusion approach is presented in this paper.,28427
different transform the contourlet used to the First is source images components. into decompose,First the contourlet transform is used to decompose the source images into different components.,28428
"extracted some are from Then, salient components. features","Then, some salient features are extracted from components.",28429
"extract spatial features, is used. frequency to In salient order","In order to extract salient features, spatial frequency is used.",28430
"selected maximum by rule. best are the Subsequently, coefficients the selection components from the","Subsequently, the best coefficients from the components are selected by the maximum selection rule.",28431
"coefficients. transform to applied the inverse the is Finally, contourlet selected","Finally, the inverse contourlet transform is applied to the selected coefficients.",28432
show the method. the of Experiments superiority proposed,Experiments show the superiority of the proposed method.,28433
most century. even code remains popular the Braille this reading in tactile,Braille remains the most popular tactile reading code even in this century.,28434
Braille. numerous amount are locked in There of literature up,There are numerous amount of literature locked up in Braille.,28435
supports only colour. either white works Existing documents in dull bright or in,Existing works supports only documents in white either bright or dull in colour.,28436
any Hardly on printed documents hand in work Braille. could traced be ordinary,Hardly any work could be traced on hand printed ordinary documents in Braille.,28437
of form treat it We control function and unspecified unknown. the the leave functional as,We leave the functional form of the control function unspecified and treat it as unknown.,28438
"retrieval used In approach to retrieve this images. classification stage paper, and similar two we","In this paper, we used two stage classification and retrieval approach to retrieve similar images.",28439
"fusion this multi-index propose we for new scheme a image retrieval. In paper,","In this paper, we propose a new multi-index fusion scheme for image retrieval.",28440
"indexes build visual various our first we Specially, multiple from representations.","Specially, we first build our multiple indexes from various visual representations.",28441
ISPs. the differential revenue pricing how effects study then of We,We then study how differential pricing effects the revenue of ISPs.,28442
of compact data. for time-series a paper This for modeling Markov reduced-order representation presents technique,This paper presents a technique for reduced-order Markov modeling for compact representation of time-series data.,28443
selected criteria. final model using is information-theoretic The,The final model is selected using information-theoretic criteria.,28444
is sets on concept different examples. two as The proposed validated data elucidated and,The proposed concept is elucidated and validated on two different data sets as examples.,28445
"is significant challenging patients. However, task to different this appearances of across lesion variations due","However, this task is challenging due to significant variations of lesion appearances across different patients.",28446
a amount further large image This challenge exacerbated dealing with of is data. when,This challenge is further exacerbated when dealing with a large amount of image data.,28447
at fitting theoretical study for neurons. aiming present model a sensory We,We present a theoretical study aiming at model fitting for sensory neurons.,28448
end in presented form Results of the text. tabular this at are,Results are presented in tabular form at the end of this text.,28449
"signature proposed. method this verification is writer-dependent paper, In a","In this paper, a writer-dependent signature verification method is proposed.",28450
are the In train the to only proposed one-class genuine used signatures SVMs. method,In the proposed method only genuine signatures are used to train the one-class SVMs.,28451
proposed other systems in literature. outperforms system the The existing,The proposed system outperforms other existing systems in the literature.,28452
module modelling We Gaussian software within the framework. process introduce developed UQLab the (GP),We introduce the Gaussian process (GP) modelling module developed within the UQLab software framework.,28453
mathematical tools (a.k.a. the key first GP the of briefly at We modelling present basis,We first briefly present the key mathematical tools at the basis of GP modelling (a.k.a.,28454
"and computational theoretical as well Kriging), associated framework. the as","Kriging), as well as the associated theoretical and computational framework.",28455
applications. is of segmentation biomedical most segmentation Multi-atlas in widely-used techniques image approach the one,Multi-atlas segmentation approach is one of the most widely-used image segmentation techniques in biomedical applications.,28456
features are a for utilized defining deep learned atlas further in selection. measure The similarity,The learned deep features are further utilized in defining a similarity measure for atlas selection.,28457
"framework used how we real-world discuss can scenarios. in be the Nevertheless,","Nevertheless, we discuss how the framework can be used in real-world scenarios.",28458
"allow particular, self-exciting as for spatio-temporal models well spatial temporal as dependence In self-excitation.","In particular, self-exciting spatio-temporal models allow for spatial dependence as well as temporal self-excitation.",28459
"an to spatial the likelihood. however, intractable leads model large regions, temporal or For","For large spatial or temporal regions, however, the model leads to an intractable likelihood.",28460
as This be easily and it applied implemented. convenient quickly is method is can,This method is convenient as it can easily be applied and is quickly implemented.,28461
how models. We will demonstrate in for Bayesian to this do a setting Spatio-Temporal Self-Exciting,We will demonstrate how to do this in a Bayesian setting for Self-Exciting Spatio-Temporal models.,28462
both simulation data. performance LA illustrated The method of and real-world is extended the with,The performance of the extended LA method is illustrated with both simulation and real-world data.,28463
discovery approach hit provides rate for identification statistical The control. principled algorithm false a and,The algorithm provides a principled statistical approach for hit identification and false discovery rate control.,28464
of present We in state contributions the CSC. to art two the,We present two contributions to the state of the art in CSC.,28465
model Expectation-Maximization accomplished method. The training is through an entire,The entire model training is accomplished through an Expectation-Maximization method.,28466
estimation is in (re-ID) important system. component an Label re-identification person an unsupervised,Label estimation is an important component in an unsupervised person re-identification (re-ID) system.,28467
ushered has age an discordance. in of Phylogenomics,Phylogenomics has ushered in an age of discordance.,28468
left often trying sense Researchers to such incongruences. are of make,Researchers are often left trying to make sense of such incongruences.,28469
of amount some and quantify DiscoVista helps of potential causes. discordance the its,DiscoVista helps quantify the amount of discordance and some of its potential causes.,28470
is processing. image substantial also method robust noise particularly without to Our any requiring,Our method is also particularly robust to noise without requiring any substantial image processing.,28471
"in the of coefficient an a of interested distribution PDE. diffusion we elliptic are Specifically,","Specifically, we are interested in the distribution of a diffusion coefficient of an elliptic PDE.",28472
discretisation resolution PDE of The is the fixed.,The resolution of the PDE discretisation is fixed.,28473
"contrast, of In a computational employs discretisations our PDE cost. the to hierarchy estimator decrease","In contrast, our estimator employs a hierarchy of PDE discretisations to decrease the computational cost.",28474
the idea on and sampler This multi-resolution generalises proposed in [P.S. builds,This idea builds on and generalises the multi-resolution sampler proposed in [P.S.,28475
"fully our and between is Importantly, tempering choice adaptive. bridging","Importantly, our choice between tempering and bridging is fully adaptive.",28476
learnt image is suitable the for We also that representation retrieval. demonstrate,We also demonstrate that the learnt representation is suitable for image retrieval.,28477
transferring painterly images. has in impressive to style Recent success work shown,Recent work has shown impressive success in transferring painterly style to images.,28478
"style fall however, transfer. approaches, of short photorealistic These","These approaches, however, fall short of photorealistic style transfer.",28479
to exhibit are Our artifacts less finer prone details the than results state-of-the-art. and,Our results exhibit finer details and are less prone to artifacts than the state-of-the-art.,28480
Vision in applications. sports detection is player important based,Vision based player detection is important in sports applications.,28481
image method Our a (player/non-player) trains labeled network from classification binary patches. first,Our method first trains a binary (player/non-player) classification network from labeled image patches.,28482
"method in image to a Then, efficiently the our testing. whole network applies","Then, our method efficiently applies the network to a whole image in testing.",28483
We soccer games. and basketball conducted experiments on,We conducted experiments on basketball and soccer games.,28484
solution natural A presented. illumination shape-from-shading numerical is to under,A numerical solution to shape-from-shading under natural illumination is presented.,28485
"completion refinement shading-aware depth applications to map Real-world and are denoising, presented.","Real-world applications to shading-aware depth map denoising, refinement and completion are presented.",28486
to Bayes Our various predictor study. competitors well empirical Monte a in compared performs Carlo,Our empirical Bayes predictor performs well compared to various competitors in a Monte Carlo study.,28487
is sufficient a not prominent meaningful and curves these information to convey shape. Extracting about,Extracting these curves is not sufficient to convey prominent and meaningful information about a shape.,28488
organizations. a Process is Management central Business (BPM) of today element,Business Process Management (BPM) is a central element of today organizations.,28489
is important task re-identification vision. computer (ReID) Person an in,Person re-identification (ReID) is an important task in computer vision.,28490
"metric deep a become with learning for common learning framework has a Recently, loss ReID.","Recently, deep learning with a metric learning loss has become a common framework for ReID.",28491
from consider and nonlinearities. images periodic the reconstructing signals of problem We,We consider the problem of reconstructing signals and images from periodic nonlinearities.,28492
approach. data demonstrate numerical experiments Several our real the of on effectiveness,Several numerical experiments on real data demonstrate the effectiveness of our approach.,28493
is a fundamental in the geological one of rocks of tasks study. Classification,Classification of rocks is one of the fundamental tasks in a geological study.,28494
The examine section a thin images a under to sampled microscope. process expert requires human,The process requires a human expert to examine sampled thin section images under a microscope.,28495
Each is structure. number into form finite grid image of partitioned a a cells that,Each image is partitioned into a finite number of cells that form a grid structure.,28496
classification. inside colour determine and each its pixels Edge cell profile of,Edge and colour profile of pixels inside each cell determine its classification.,28497
cells section determine a classification The the voting scheme. then majority image via thin individual,The individual cells then determine the thin section image classification via a majority voting scheme.,28498
"the embrace and (CoPaNet). Network Pathway propose work, this In this Competitive observation we","In this work, we embrace this observation and propose the Competitive Pathway Network (CoPaNet).",28499
the This by a variety features capability subnetworks. learning enhances in model of mechanism,This mechanism enhances the model capability by learning a variety of features in subnetworks.,28500
"feature CoPaNet be Moreover, the added can to the cross-block reuse. encourage to shortcut","Moreover, the cross-block shortcut can be added to the CoPaNet to encourage feature reuse.",28501
amounts similar comparable of state-of-the-art results using or the parameters. obtained CoPaNet,CoPaNet obtained the state-of-the-art or comparable results using similar amounts of parameters.,28502
"In (BS) this number technology, is of equipped antennas. Station large with a Base","In this technology, Base Station (BS) is equipped with a large number of antennas.",28503
"However, provided beam lead of to inaccuracies information the misalignment.","However, inaccuracies of the provided information lead to beam misalignment.",28504
beam the Some in design suggested parameters effect. are this to literature combat,Some beam design parameters are suggested in the literature to combat this effect.,28505
But how these tuned? should parameters be,But how these parameters should be tuned?,28506
the sea-floor. images Side-look (SAS) very produce aperture can high synthetic sonar quality of,Side-look synthetic aperture sonar (SAS) can produce very high quality images of the sea-floor.,28507
approaches. on alternative are and several segmentation scenes shown compared Results SAS to,Results are shown on several SAS scenes and compared to alternative segmentation approaches.,28508
task suffering inter- variability. is organs both and and intra-observer Marking tumors challenging from a,Marking tumors and organs is a challenging task suffering from both inter- and intra-observer variability.,28509
"the hence, Fast required similar for are, facilitate to search cases. algorithms","Fast algorithms are, hence, required to facilitate the search for similar cases.",28510
"have developed somewhat new approach is computer-intensive. We albeit a that to straightforward apply, (HARVEST)","We have developed a new approach (HARVEST) that is straightforward to apply, albeit somewhat computer-intensive.",28511
"we have statistical relevance. derived this idea, feature for a Motivated test by simple","Motivated by this idea, we have derived a simple statistical test for feature relevance.",28512
"is expensive work. Obviously, a long and this","Obviously, this is a long and expensive work.",28513
occurs) a samples (i.e. highlight when negative and,when a highlight occurs) and negative samples (i.e.,28514
quality device. low images case captured especially for is the with heavy haze a This,This is especially the case for heavy haze images captured with a low quality device.,28515
method is neural networks. based convolutional Our on,Our method is based on convolutional neural networks.,28516
and of method evaluate real our both using We examples. world effectiveness synthetic proposed the,We evaluate the effectiveness of our proposed method using both synthetic and real world examples.,28517
a large behaviors. of Periodic nonlinear oscillators dynamic leads number of to forcing,Periodic forcing of nonlinear oscillators leads to a large number of dynamic behaviors.,28518
"with the variability. In circadian this of deletion prediction, clock reduces agreement","In agreement with this prediction, deletion of the circadian clock reduces variability.",28519
in provide of surface shape the most The techniques information normals. terms common,The most common techniques provide the shape information in terms of surface normals.,28520
contributions main our are paper of of nature. The theoretical,The main contributions of our paper are of theoretical nature.,28521
technical is in differential usage A calculus. our the of matrix framework novelty,A technical novelty in our framework is the usage of matrix differential calculus.,28522
method. we At of properties of hand evaluation discuss important the an experimental,At hand of an experimental evaluation we discuss important properties of the method.,28523
"strategy approaches. results our achieves more than accurate Overall, competing","Overall, our strategy achieves more accurate results than competing approaches.",28524
"of lengths is depth information scene the estimated. blurs,","lengths of the blurs, scene depth information is estimated.",28525
"we grade. to AlphaGo Based belongs conclude on that the this, third","Based on this, we conclude that AlphaGo belongs to the third grade.",28526
where Colombia are populations in is the malaria What endemic? regions and,What are the regions and populations in Colombia where malaria is endemic?,28527
associations regions exist in Colombia? outbreaks epidemic between What between,What associations exist between epidemic outbreaks between regions in Colombia?,28528
act and as epidemic regions. between key Some bridges endemic locations,Some key locations act as bridges between epidemic and endemic regions.,28529
"A translucency reproducing interpretation. is for open by Thus, encoded","Thus, reproducing translucency encoded by A is open for interpretation.",28530
by commercial measurements be spectrophotometers graphic conducted Such can in arts. used,Such measurements can be conducted by commercial spectrophotometers used in graphic arts.,28531
"real material often Gaussian. in distribution the not is However, of applications, a","However, in real applications, the distribution of a material is often not Gaussian.",28532
"endmember Gaussian this mixture use to models the represent variability. paper, (GMM) we In","In this paper, we use Gaussian mixture models (GMM) to represent the endmember variability.",28533
"PCANet, noticeable network, histogram for employs shallow representation as the feature one pooling.","PCANet, as one noticeable shallow network, employs the histogram representation for feature pooling.",28534
"about there three However, method. kind are pooling this main problems of","However, there are three main problems about this kind of pooling method.",28535
"exponentially output Third, used. the of of the histogram-based number feature maps with dimensionality grows","Third, the dimensionality of histogram-based output grows exponentially with the number of feature maps used.",28536
"network, method combine we order with PCANet. Thus pooling second the shallow the statistical i.e.,","Thus we combine the second order statistical pooling method with the shallow network, i.e., PCANet.",28537
Experiments of our method. the robustness and effectiveness PCANet-II proposed demonstrate,Experiments demonstrate the effectiveness and robustness of our proposed PCANet-II method.,28538
using geometric realized learning-based where decoding is process a approach The calibration is no needed.,The decoding process is realized using a learning-based approach where no geometric calibration is needed.,28539
with techniques evaluations were recent comparisons and quantitative Both conducted. also,Both quantitative evaluations and comparisons with recent techniques were also conducted.,28540
an the We fast-flow analytical propose description in limit.,We propose an analytical description in the fast-flow limit.,28541
with in motion. modelling It social systems provides biological for tool or a individuals,It provides a modelling tool for biological or social systems with individuals in motion.,28542
"Pyramidal network the is model channels based gradually Firstly, RoR designed. on RoR, increasing with","Firstly, based on RoR, the Pyramidal RoR network model with channels gradually increasing is designed.",28543
multiple series. interrelated present new We a of time systems method for forecasting,We present a new method for forecasting systems of multiple interrelated time series.,28544
a has learning cooperation recently. multi-agent reinforcement been for topic and competition Deep hot,Deep reinforcement learning for multi-agent cooperation and competition has been a hot topic recently.,28545
problem on methods on paper multi-agent settings. based local This under actor-critic observations cooperative focuses,This paper focuses on cooperative multi-agent problem based on actor-critic methods under local observations settings.,28546
"stresses Given and tissue. deforming in expanding its elastic an emerge properties,","Given its elastic properties, stresses emerge in an expanding and deforming tissue.",28547
"however, elastic time mainly tissues short scales, have properties. On","On short time scales, however, tissues have mainly elastic properties.",28548
agents Dynamic with classical problem decentralized with information. a contracts asymmetric is decision-making multiple,Dynamic contracts with multiple agents is a classical decentralized decision-making problem with asymmetric information.,28549
to approach generalization. The domain extended is,The approach is extended to domain generalization.,28550
experiments promising show both very For the applications results.,For both applications the experiments show very promising results.,28551
Sparse randomly undersampled techniques alleviate limitation acquisitions. can recovery via this,Sparse recovery techniques can alleviate this limitation via randomly undersampled acquisitions.,28552
"is method sampling proposed for this a MRI. To statistically-segregated multiple-acquisition address problem,","To address this problem, a statistically-segregated sampling method is proposed for multiple-acquisition MRI.",28553
Fourier datasets. improved both quality significantly achieves multiple-acquisition in of Segregated compressed-sensing and sampling reconstructions,Segregated sampling achieves significantly improved quality in both Fourier and compressed-sensing reconstructions of multiple-acquisition datasets.,28554
programmes the investigate systematically effect for heterogeneity We search workers. of job unemployed,We systematically investigate the effect heterogeneity of job search programmes for unemployed workers.,28555
"we empirical non-experimental To effects, causal investigate heterogeneous with models Lasso-type possibly combine estimators. employment","To investigate possibly heterogeneous employment effects, we combine non-experimental causal empirical models with Lasso-type estimators.",28556
Swiss analyses records. are on data empirical security The administrative based rich from social,The empirical analyses are based on rich administrative data from Swiss social security records.,28557
six training. the after only months start considerable find the first heterogeneities during We of,We find considerable heterogeneities only during the first six months after the start of training.,28558
"document heterogeneous status. also we Furthermore, by residence employment effects","Furthermore, we also document heterogeneous employment effects by residence status.",28559
to region-of-interest a frame. inside locate used is the Gaze,Gaze is used to locate the region-of-interest inside a frame.,28560
"combinations BRISGAZE-ACTIONS. i.e., results a present parameters) on the several dataset, two We (of using","We present several results using combinations (of the two parameters) on a dataset, i.e., BRISGAZE-ACTIONS.",28561
depicting egocentric videos dataset activities. contains daily-living The several,The dataset contains egocentric videos depicting several daily-living activities.,28562
is temporal implementing by The entropy two of the further quality cuts improved measures.,The quality of the temporal cuts is further improved by implementing two entropy measures.,28563
neural Convolutional have networks areas (CNNs) compute been over widely in many vision. used,Convolutional neural networks (CNNs) have been widely used over many areas in compute vision.,28564
Unlike vocabulary pipelines use SfM that (Voc.) modern,Unlike modern SfM pipelines that use vocabulary (Voc.),28565
"algorithm. similar priors to GraphMatch combines iterative scheme these two the PatchMatch an ""sample-and-propagate"" into","GraphMatch combines these two priors into an iterative ""sample-and-propagate"" scheme similar to the PatchMatch algorithm.",28566
paper image approach introduces This to denoising. new a means non-local,This paper introduces a new approach to non-local means image denoising.,28567
non-local This robust means. is method called,This method is called robust non-local means.,28568
Novelty RNNs in way CNNs and trained. the those are lies,Novelty lies in the way those CNNs and RNNs are trained.,28569
"Computational prevents ""CNN+RNN"" training the of complexity systems. end-to-end","Computational complexity prevents the end-to-end training of ""CNN+RNN"" systems.",28570
"CNNs trained independently first, RNNs. the usually from Therefore, are","Therefore, CNNs are usually trained first, independently from the RNNs.",28571
"accomplish our for neural matching. this, uses To representations algorithm","To accomplish this, our algorithm uses neural representations for matching.",28572
"Additionally, be color the transfer spatially should variant and globally coherent.","Additionally, the color transfer should be spatially variant and globally coherent.",28573
"color transfer, a and jointly approach strategy. matching proposed optimizes coarse-to-fine Our adopting","Our proposed approach jointly optimizes matching and color transfer, adopting a coarse-to-fine strategy.",28574
extended one-to-many color be from one-to-one can method The successfully to proposed transfer.,The proposed method can be successfully extended from one-to-one to one-to-many color transfer.,28575
addresses image. latter problem further input of of the The the elements mismatching,The latter further addresses the problem of mismatching elements of the input image.,28576
it We variety on our method validate proposed image a content. by of testing large,We validate our proposed method by testing it on a large variety of image content.,28577
recognize basic-level each is hundreds Fine-grained of in subcategories classification image category. to,Fine-grained image classification is to recognize hundreds of subcategories in each basic-level category.,28578
localization Existing key discriminative methods subcategories. to the find employ among distinctions,Existing methods employ discriminative localization to find the key distinctions among subcategories.,28579
major Imbalance In-phase/quadrature-phase performance-limiting considered in transceivers. impairment a is (IQI) direct-conversion,In-phase/quadrature-phase Imbalance (IQI) is considered a major performance-limiting impairment in direct-conversion transceivers.,28580
"for different the SER modulation this, we derive Capitalizing schemes. expressions the analytic on of","Capitalizing on this, we derive analytic expressions for the SER of the different modulation schemes.",28581
stable discriminative convergence. with is learned training and with end-to-end COCO bundled,COCO is bundled with discriminative training and learned end-to-end with stable convergence.,28582
the parameters spectrum conditions the are associated power for recovery analyzed. Perfect and,Perfect recovery conditions for the associated parameters and the power spectrum are analyzed.,28583
address to allows parametrization optimization us propose to problem This diffusion. functional inverse a,This parametrization allows us to propose a functional optimization problem to address inverse diffusion.,28584
"visual for handle problem, graph To learn this we a representation patch-based tracking.","To handle this problem, we learn a patch-based graph representation for visual tracking.",28585
This in applied tracking and and updating. learned graph dynamically is object model,This graph is dynamically learned and applied in object tracking and model updating.,28586
"each tracking During three main the frame. in steps algorithm performs the process, proposed","During the tracking process, the proposed algorithm performs three main steps in each frame.",28587
"the the first unsupervised into video by is video each segmentation. In supervoxels decomposed step,","In the first step, each video is decomposed into supervoxels by the unsupervised video segmentation.",28588
"presents, examines approaches paper skin and segmentation. to This compares two different lesion","This paper presents, examines and compares two different approaches to skin lesion segmentation.",28589
The a based step. and first approach equalization introduces histogram uses U-Nets preprocessing,The first approach uses U-Nets and introduces a histogram equalization based preprocessing step.,28590
"preprocessing using segmentation histogram Moreover, step equalization improved significantly the the results. for U-Net","Moreover, using the histogram equalization for preprocessing step significantly improved the U-Net segmentation results.",28591
"(USDL), Sparse approach, Unified Learning constitutes Dynamics is steps. which Our two named of","Our approach, which is named Unified Sparse Dynamics Learning (USDL), constitutes of two steps.",28592
"an formulation. the of through atemporal the First, system is equations of derived application weak","First, an atemporal system of equations is derived through the application of the weak formulation.",28593
"recovery Additionally, significantly accuracy such metrics the coefficient. with USDL's correlates theoretical exact as","Additionally, USDL's accuracy significantly correlates with theoretical metrics such as the exact recovery coefficient.",28594
be machine models. can mutual learning characterized by contamination problems Many,Many machine learning problems can be characterized by mutual contamination models.,28595
"AR across held the constant order Furthermore, (e.g. voxels is","Furthermore, the AR order is held constant across voxels (e.g.",28596
"multi-scale limiting integration tradition connected, layers adjacent only information. allow in of However, they","However, in tradition they only allow adjacent layers connected, limiting integration of multi-scale information.",28597
"present shortcut convolutional To of improve framework performance, a concatenating neural further we their networks.","To further improve their performance, we present a concatenating framework of shortcut convolutional neural networks.",28598
causes Myocardial of leading Infarction death the worldwide. is one of,Myocardial Infarction is one of the leading causes of death worldwide.,28599
example are and evidence included. Simulation empirical an,Simulation evidence and an empirical example are included.,28600
"across dataset without the proposed also addition, better method In additional training. has ability generalization","In addition, the proposed method also has better generalization ability across dataset without additional training.",28601
"configurations. proposed More architecture on considering specifically, extensive the an analysis we different provide","More specifically, we provide an extensive analysis on the proposed architecture considering different configurations.",28602
"However, IA still explicit procedures. of and lacks design straightforward","However, IA still lacks of explicit and straightforward design procedures.",28603
four is out. comparison carried A the also algorithms between,A comparison between the four algorithms is also carried out.,28604
The to proposed promising. be seem approaches cooperative,The cooperative proposed approaches seem to be promising.,28605
hyperparameters likelihood on estimating inference. maximum estimation) kernel using model,estimating kernel hyperparameters using maximum likelihood estimation) on model inference.,28606
"we learning video In online transform work, apply this denoising. to","In this work, we apply online transform learning to video denoising.",28607
"denoising efficient processing. offers proposed little requires memory, online and video The","The proposed online video denoising requires little memory, and offers efficient processing.",28608
assorted learning classification have in Machine very tasks. efficient techniques proven,Machine learning techniques have proven very efficient in assorted classification tasks.,28609
"we recognition direct this into the convert detection process end, problem. To bit pattern a","To this end, we convert the direct bit detection process into a pattern recognition problem.",28610
learning CU partition. enables on the deep This,This enables deep learning on the CU partition.,28611
"hierarchical learning HCPM. early-terminated CNN for to predict Then, propose we the an (ETH-CNN)","Then, we propose an early-terminated hierarchical CNN (ETH-CNN) for learning to predict the HCPM.",28612
to penetrating ground introduced The (GPR) signal processing. method radar is VMD,The VMD method is introduced to ground penetrating radar (GPR) signal processing.,28613
from method engineering practical case. a subsequently to The is used analyze signal a GPR,The method is subsequently used to analyze a GPR signal from a practical engineering case.,28614
"work, sequences. ABC derive QMC In (quasi- on algorithms we Carlo) this Monte based","In this work, we derive ABC algorithms based on QMC (quasi- Monte Carlo) sequences.",28615
literature. We from through QMC taken our the ABC examples approach illustrate several,We illustrate our QMC approach through several examples taken from the ABC literature.,28616
"histograms LBP BoVW via features, deep the validation. classify leave-one-out investigate We images to and","We investigate deep features, LBP histograms and BoVW to classify the images via leave-one-out validation.",28617
empirical state-of-the-art common datasets benchmark on in-the-wild show We pose which tests results. present,We present empirical tests on common in-the-wild pose benchmark datasets which show state-of-the-art results.,28618
well testing models. We pre-trained training our as release open-source as and code our,We open-source our training and testing code as well as release our pre-trained models.,28619
encoding used have recently for Autoencoders images. been medical,Autoencoders have been recently used for encoding medical images.,28620
learning makes architecture which faster. the of integration rather training The promotes shallow a MLP,The integration of MLP promotes a rather shallow learning architecture which makes the training faster.,28621
"UDN. in However, severe the major is obstacle interference","However, severe interference is the major obstacle in UDN.",28622
of theory. by the deal game on cooperative relying with the Most contributions interference,Most of the contributions deal with the interference by relying on cooperative game theory.,28623
"signal can CoMP dense invoked transmission. centralized be processing for Under supporting C-RAN,","Under dense C-RAN, centralized signal processing can be invoked for supporting CoMP transmission.",28624
dense the summarize C-RANs. main in user-centric challenges We,We summarize the main challenges in dense user-centric C-RANs.,28625
"namely, investigate CSI. inter-cluster We relying this on on CSI, requirement large-scale by partial only","We investigate this requirement by only relying on partial CSI, namely, on inter-cluster large-scale CSI.",28626
quantified sample foveal using sets methods. are and The different two verified sampling,The foveal sample sets are verified and quantified using two different sampling methods.,28627
"CNN, represents part. certain In conv-layer interpretable filter in a an a high each object","In an interpretable CNN, each filter in a high conv-layer represents a certain object part.",28628
types applied can of method different different structures. with Our to CNNs be,Our method can be applied to different types of CNNs with different structures.,28629
"residual signal-noise nonlinearities signal-signal from the compensated, interactions. arises When limitation are performance","When signal-signal nonlinearities are compensated, the performance limitation arises from residual signal-noise interactions.",28630
and signal interactions co-propagating of consist transceiver and ASE noise. the between beating These nonlinear,These interactions consist of nonlinear beating between the signal and co-propagating ASE and transceiver noise.,28631
simulations. are confirmed by numerical The findings theoretical,The theoretical findings are confirmed by numerical simulations.,28632
large generate of a chronic disease events. amounts can affected A patient by,A patient affected by a chronic disease can generate large amounts of events.,28633
network fast Convolu- realize results. can tional processing while neural achieving rable compa-,Convolu- tional neural network can realize fast processing while achieving compa- rable results.,28634
network and consists of The a ReconNet. fully-connected layer new,The new network consists of a fully-connected layer and ReconNet.,28635
layer as which has fully-connected output low-dimension The acts measurement.,The fully-connected layer which has low-dimension output acts as measurement.,28636
the layer and and obtain We adaptive measurement. simultaneously train fully-connected ReconNet,We train the fully-connected layer and ReconNet simultaneously and obtain adaptive measurement.,28637
network new original the outperforms the one. that show Experiments,Experiments show that the new network outperforms the original one.,28638
results detect how the way. can experimental edge indicate SVM efficient The in,The experimental results indicate how the SVM can detect edge in efficient way.,28639
that Several based high-level works perform attempt on to landmarks. face-related exist analysis tasks,Several works exist that attempt to perform high-level face-related analysis tasks based on landmarks.,28640
of Program. a We large-scale the solving problem Quadratically Quadratic Constrained consider,We consider the problem of solving a large-scale Quadratically Constrained Quadratic Program.,28641
in problems naturally and many scientific occur applications. Such web,Such problems occur naturally in many scientific and web applications.,28642
"there tackle this they not which are mostly methods scalable. efficient problem, Although are","Although there are efficient methods which tackle this problem, they are mostly not scalable.",28643
by be transformed then state-of-the-art solved applying can programming any problem solvers. large-scale quadratic The,The transformed problem can then be solved by applying any state-of-the-art large-scale quadratic programming solvers.,28644
contain sounds about vital pathology. information   Lung pulmonary,  Lung sounds contain vital information about pulmonary pathology.,28645
"this of short-term lung paper, associated characteristics use we In recognize spectral sounds diseases. to","In this paper, we use short-term spectral characteristics of lung sounds to recognize associated diseases.",28646
"Also, kernels. multiple was SVM tested performance the with","Also, the performance was tested with multiple SVM kernels.",28647
sounds. conduct two evaluation include performance which adventitious and extensive using databases We normal experiments,We conduct extensive performance evaluation experiments using two databases which include normal and adventitious sounds.,28648
than The performs kernel better other considerably kernels. Bhattacharyya,The Bhattacharyya kernel performs considerably better than other kernels.,28649
"feature we optimize the proposed algorithm. of configuration Further, the extraction","Further, we optimize the configuration of the proposed feature extraction algorithm.",28650
in of Neuromodulation fundamental acquisition role a new plays behaviours. the,Neuromodulation plays a fundamental role in the acquisition of new behaviours.,28651
"Here, the of we on behaviour. cholinergic depression characterize further effects","Here, we further characterize the effects of cholinergic depression on behaviour.",28652
"contributions. we second part, focus algorithmic our this In on","In this second part, we focus on our algorithmic contributions.",28653
"removing sequence. ""noise"" video a problem the from denoising refers to of Video","Video denoising refers to the problem of removing ""noise"" from a video sequence.",28654
"We extensive denoising our show, using the experiments, that denoising algorithms. state-of-the-art approach outperforms","We show, using extensive experiments, that our denoising approach outperforms the state-of-the-art denoising algorithms.",28655
reconstruction super-resolution. high-quality single recently image have demonstrated networks Convolutional for neural,Convolutional neural networks have recently demonstrated high-quality reconstruction for single image super-resolution.,28656
sub-band progressively the multiple of reconstructs at proposed high-resolution network The images residuals levels. pyramid,The proposed network progressively reconstructs the sub-band residuals of high-resolution images at multiple pyramid levels.,28657
Purpose: Neural have reconstruction networks received undersampled interest acquisitions. for of MR recent,Purpose: Neural networks have received recent interest for reconstruction of undersampled MR acquisitions.,28658
to in trained testing were networks compared domain. networks the directly Domain-transferred,Domain-transferred networks were compared to networks trained directly in the testing domain.,28659
difficult. We appear make scientific that wide range of review to challenges success a then,We then review a wide range of challenges that appear to make scientific success difficult.,28660
rests ideas. proposed The main two framework on,The proposed framework rests on two main ideas.,28661
"problems novel multi-label We framework propose continuous arbitrary, on discretisation the of a for domains.","We propose a novel framework for the discretisation of multi-label problems on arbitrary, continuous domains.",28662
single problems. AL focus learning Most methods on instance,Most AL methods focus on single instance learning problems.,28663
"method, The hierarchically the in called instance aggregative proposed data sampling}, \textit{cluster-based other space. clusters","The other proposed method, called \textit{cluster-based aggregative sampling}, clusters data hierarchically in the instance space.",28664
and features continuously is LTE new refined released.,LTE is continuously refined and new features released.,28665
channel RF radiated over-the-air It modes. supports emulated and,It supports RF channel emulated and over-the-air radiated modes.,28666
configured. be and testbed accessed can The remotely,The testbed can be remotely accessed and configured.,28667
"In terms of they desirable reduction. value, gave conditions for a","In terms of value, they gave conditions for a desirable reduction.",28668
"of factor paper, definition this the Q we nonlinear In for oscillators. discuss","In this paper, we discuss the definition of Q factor for nonlinear oscillators.",28669
computed numerically It oscillator from can be rigorously and equations. formulated,It can be formulated rigorously and computed numerically from oscillator equations.,28670
"a has Image computer progress captioning, topic popular substantial in in achieved vision, recent years.","Image captioning, a popular topic in computer vision, has achieved substantial progress in recent years.",28671
"distinctiveness work. overlooked previous of natural often in the is However, descriptions","However, the distinctiveness of natural descriptions is often overlooked in previous work.",28672
"learning we In this Learning new captioning. work, propose method, image for (CL), Contrastive a","In this work, we propose a new learning method, Contrastive Learning (CL), for image captioning.",28673
be can which displays. information retargeted fields angular depth to Light multi-panel preserve,Light fields preserve angular information which can be retargeted to multi-panel depth displays.,28674
and position aligned the of to viewer. dynamically views synthesized are the according Displayed,Displayed views are synthesized and aligned dynamically according to the position of the viewer.,28675
"on The verified experimentally is three-panel simulated display. technique proposed aerial outlined, a and","The proposed technique is outlined, simulated and verified experimentally on a three-panel aerial display.",28676
"content. scene resolution with effective varies the However, significantly","However, the effective resolution varies significantly with scene content.",28677
of Recent algorithms SLAM demonstrated impressive odometry visual have levels direct precision. and,Recent direct visual odometry and SLAM algorithms have demonstrated impressive levels of precision.,28678
"to a in camera order However, achieve results. calibration competitive require they photometric","However, they require a photometric camera calibration in order to achieve competitive results.",28679
a are of not fixed. individuals population organisms of in Phenotypes,Phenotypes of individuals in a population of organisms are not fixed.,28680
provide adaptation Phenotypic basis and selection. target fluctuations the a of can be for,Phenotypic fluctuations can provide a basis for adaptation and be the target of selection.,28681
preconditioner and imaging parallel efficient reconstructions. Design Purpose: a of and sensing compressed for fast,Purpose: Design of a preconditioner for fast and efficient parallel imaging and compressed sensing reconstructions.,28682
systems be efficiently linear solved techniques. Such using can preconditioning effective,Such linear systems can be solved efficiently using effective preconditioning techniques.,28683
compressed Key parallel sensing; Bregman; imaging preconditioning; words: Split .,Key words: preconditioning; compressed sensing; Split Bregman; parallel imaging .,28684
"amount this, be of a grader there Because fair can of variability.","Because of this, there can be a fair amount of grader variability.",28685
stored hospitals' Communication and over collected been Archiving Picture several and in have decades Systems.,have been collected over several decades and stored in hospitals' Picture Archiving and Communication Systems.,28686
science dataset available (under publicly the made will This open initiative). be,This dataset will be made publicly available (under the open science initiative).,28687
linear We mixed graphs. consider with associated that equation models are structural,We consider linear structural equation models that are associated with mixed graphs.,28688
most Deliberate environments. to in effective for and RL succeed agents necessary exploration is,Deliberate and effective exploration is necessary for RL agents to succeed in most environments.,28689
observed that uncertain. frequently deemed are States features more less have,States that have less frequently observed features are deemed more uncertain.,28690
when Sensory be rapidly more they stimuli are recognized expected. can,Sensory stimuli can be recognized more rapidly when they are expected.,28691
depends sensory expectation affecting information. phenomenon of the cortical on processing This,This phenomenon depends on expectation affecting the cortical processing of sensory information.,28692
coupling a to investigate waveguide the coplanar resonator waveguide of coplanar analytically We feedline. a,We investigate analytically the coupling of a coplanar waveguide resonator to a coplanar waveguide feedline.,28693
calculated. order the shift for are terms and external Leading frequency quality factor,Leading order terms for the external quality factor and frequency shift are calculated.,28694
"estimates. error"" trajectories, simulation introducing trajectory a Probabilistic finite to necessarily number of generates ""simulation","Probabilistic simulation necessarily generates a finite number of trajectories, introducing ""simulation error"" to trajectory estimates.",28695
calculates that our dataset. show ground-truth from probabilities the method a accurately transition We phantom,We show that our method accurately calculates the ground-truth transition probabilities from a phantom dataset.,28696
"the challenging performance directly to their structures. complicated Meanwhile, predicting due is internal","Meanwhile, directly predicting their performance is challenging due to the complicated internal structures.",28697
new a is high-throughput A strategy proposed. based (HTS) on screening system energy process optimization,A new energy system optimization strategy based on a high-throughput screening (HTS) process is proposed.,28698
games industry. of has the caused The paradigm in a mobile shift video-game emergence,The emergence of mobile games has caused a paradigm shift in the video-game industry.,28699
"on approach we this an In abandon survival based present ensembles. article, to predicting game","In this article, we present an approach to predicting game abandon based on survival ensembles.",28700
research. data driven community a the This large research for barrier in creates entry of,This creates a large barrier of entry in the research community for data driven research.,28701
Effect constellation the properties considered. is of anchors' and noise of geometric,Effect of noise and geometric properties of the anchors' constellation is considered.,28702
extracted connICA traits. main task-sensitive The hybrid hybrid two,The hybrid connICA extracted two main task-sensitive hybrid traits.,28703
"between second, visual, attentional, subcortical mainly the networks. and connectivity encompassing DMN The","The second, mainly encompassing the connectivity between visual, attentional, DMN and subcortical networks.",28704
cope allows This us the aspect during with tracking. flexibly variation ratio,This allows us cope with the aspect ratio variation flexibly during tracking.,28705
"we of IBCCF direction our an multipliers. develop alternating method optimize model, To","To optimize our IBCCF model, we develop an alternating direction method of multipliers.",28706
game science. Understanding player is behavior fundamental in data,Understanding player behavior is fundamental in game data science.,28707
"in-game of events. developers game evaluate need impact beforehand In particular, to the","In particular, game developers need to evaluate beforehand the impact of in-game events.",28708
Simulation is to of crucial monetization. engagement player and events optimization maximize these increase,Simulation optimization of these events is crucial to increase player engagement and maximize monetization.,28709
Our development a of to the step data-driven games. towards purpose ultimate is take,Our ultimate purpose is to take a step towards the data-driven development of games.,28710
of planar novel models algorithm for unsupervised from depth-data. propose a We extraction piecewise,We propose a novel algorithm for unsupervised extraction of piecewise planar models from depth-data.,28711
and to effectively in to perceive their three navigate surroundings dimensions.,to effectively perceive their surroundings and to navigate in three dimensions.,28712
"discovered perspective. In insights plant from addition, reveal interesting science groupings a our","In addition, our discovered groupings reveal interesting insights from a plant science perspective.",28713
This mucus may presence consider layer the a the epithelial wall. of model also coating,This model may also consider the presence of a mucus layer coating the epithelial wall.,28714
to performed JawNO of great peripheral Simulations that were the changes. caliber showed sensitivity airways,Simulations were performed that showed the great sensitivity of JawNO to peripheral airways caliber changes.,28715
"Moreover, on location FENO the showed change. distinct behaviors depending caliber of the","Moreover, FENO showed distinct behaviors depending on the location of the caliber change.",28716
"industries. faced challenge by a broad churn, several is","churn, is a broad challenge faced by several industries.",28717
"churn is player to In decisive retention rise social and games, increase decreasing revenues. mobile","In mobile social games, decreasing churn is decisive to increase player retention and rise revenues.",28718
"regressions, left the have only when could game. methods, applied all Classical be players like","Classical methods, like regressions, could be applied only when all players have left the game.",28719
the is data and of called This churn. in a is nature censored problem,This is called a censored data problem and is in the nature of churn.,28720
"risk survival the times. assess we factors player that predicted Additionally, the explain","Additionally, we assess the risk factors that explain the predicted player survival times.",28721
synthetic outperforms data real Extensive experiments state-of-the-art. our on method and show that,Extensive experiments on synthetic and real data show that our method outperforms state-of-the-art.,28722
"the seamless included way. in be a predictions pose can contextual information to Consequently,","Consequently, contextual information can be included to the pose predictions in a seamless way.",28723
"more than sketch recognition For example are rarer images, in and which photos. distinctly abstract","For example recognition in sketch images, which are distinctly more abstract and rarer than photos.",28724
"dataset covering domains. DG Secondly, benchmark develop photo, painting and a cartoon we sketch,","Secondly, we develop a DG benchmark dataset covering photo, sketch, cartoon and painting domains.",28725
"is benchmarks. more existing relevant, domain and both This than practically shift) harder (bigger","This is both more practically relevant, and harder (bigger domain shift) than existing benchmarks.",28726
causes Reasoning engineering effects the about naturally systems. and safety-critical of arises in,Reasoning about causes and effects naturally arises in the engineering of safety-critical systems.,28727
reasoning relationships. requires fault also systems causality design of The on management,The design of fault management systems also requires reasoning on causality relationships.,28728
"timely and to of faults, a ensure particular, identification i.e. In needs system fail-operational detection","In particular, a fail-operational system needs to ensure timely detection and identification of faults, i.e.",28729
through the the observable system. of run-time their faults effects occurrence on recognize,recognize the occurrence of run-time faults through their observable effects on the system.,28730
"we to for In fault a management propose systems. approach this formal complex work,","In this work, we propose a formal approach to fault management for complex systems.",28731
notions tree introduce of sets. minimal the first cut fault and We,We first introduce the notions of fault tree and minimal cut sets.,28732
property symmetry Discrete exploits ZT Transformation Fourier the conjugate (DFT). The of proposed,The proposed ZT exploits the conjugate symmetry property of Discrete Fourier Transformation (DFT).,28733
proposed interlacing configurations: is ZT. the using two transformation and The implemented different concatenating,The proposed transformation is implemented using two different configurations: the interlacing and concatenating ZT.,28734
trainable parts three an end-to-end These compose network.,These three parts compose an end-to-end trainable network.,28735
method based The \textit{forward} is a message-passing recursive on algorithm.,The method is based on a \textit{forward} recursive message-passing algorithm.,28736
"these from methods are unstable. However, changepoints the detected","However, the detected changepoints from these methods are unstable.",28737
developed. with regimes is of associated also Estimation parameters,Estimation of parameters associated with regimes is also developed.,28738
"machines visual can tactile can think auditory, forms principally. and in language, so of People","People can think in auditory, visual and tactile forms of language, so can machines principally.",28739
to them possible is think radio in But for language? it,But is it possible for them to think in radio language?,28740
"for According general a intelligence, presented i.e. first principle to","According to a first principle presented for general intelligence, i.e.",28741
white and zones The the the colored patterns. classical not Turing follow does distribution of,The distribution of the white and colored zones does not follow the classical Turing patterns.,28742
consist two immobile a factors factors factor. diffusing and cell-autonomous of transcription These,These factors consist of two diffusing factors and a cell-autonomous immobile transcription factor.,28743
distinguishing model Globally the four situations. different to allowed,Globally the model allowed to distinguishing four different situations.,28744
"initial activation factor in Moreover, lines. of the produces straight slope transcription an the","Moreover, an initial slope in the activation of the transcription factor produces straight lines.",28745
"results The tones. situation color not to in lead does third different patterns, but homogeneous","The third situation does not lead to patterns, but results in different homogeneous color tones.",28746
the also We under the rehabilitation. analyze a motion patient to apply method of,We also apply the method to analyze the motion of a patient under rehabilitation.,28747
requirements often output thinning a application. vary of with procedure The,The output requirements of a thinning procedure often vary with application.,28748
of obtained the results The the validity prove procedure.,The results obtained prove the validity of the procedure.,28749
marker events. is Coronary atherosclerosis calcium and of artery cardiovascular significant (CAC) a,Coronary artery calcium (CAC) is a significant marker of atherosclerosis and cardiovascular events.,28750
"i.e. algorithm, system learning supervised proposed The uses a deep","The proposed system uses a supervised deep learning algorithm, i.e.",28751
"StarCraft, long which Macro-management has a an time. important problem in been is studied for","Macro-management is an important problem in StarCraft, which has been studied for a long time.",28752
last Various years. in methods datasets have with assorted the together proposed few been,Various datasets together with assorted methods have been proposed in the last few years.,28753
"each final of feature pre-defined and match. actions of high-level vectors, result MSC well-designed consists","MSC consists of well-designed feature vectors, pre-defined high-level actions and final result of each match.",28754
"concealed which and or the emotions. expressions, involuntary facial are suppressed rapid Micro-expressions indicate","Micro-expressions are rapid and involuntary facial expressions, which indicate the suppressed or concealed emotions.",28755
"obtains the automatic spotting (ME) Recently, increasing research micro-expression attention. on","Recently, the research on automatic micro-expression (ME) spotting obtains increasing attention.",28756
step spotting prior crucial ME tasks. analysis is further a ME to,ME spotting is a crucial step prior to further ME analysis tasks.,28757
"this a To spotting end, we benchmark (MESB). micro-expression construct","To this end, we construct a micro-expression spotting benchmark (MESB).",28758
"framework. we up multi-scale evaluation window Firstly, sliding set a based","Firstly, we set up a sliding window based multi-scale evaluation framework.",28759
"a series we introduce Secondly, protocols. of","Secondly, we introduce a series of protocols.",28760
"of we results methods. Thirdly, also provide popular baseline","Thirdly, we also provide baseline results of popular methods.",28761
most one of the of tasks is challenging identification Image computer in vision. areas different,Image identification is one of the most challenging tasks in different areas of computer vision.,28762
the the each for deformation. matched evaluate We also difference image orientation of distribution keypoint,We also evaluate the distribution of the matched keypoint orientation difference for each image deformation.,28763
"methods situations, be In beneficial. processes related likelihood point such and may","In such situations, point processes and related likelihood methods may be beneficial.",28764
supported study in be theoretical is and by will it The simulations. nature,The study is theoretical in nature and it will be supported by simulations.,28765
detection employing programming. is semi-definite by a secured The then,The detection is then secured by employing a semi-definite programming.,28766
propose when modeling of topics algorithms topic the is number for new We unknown.,We propose new algorithms for topic modeling when the number of topics is unknown.,28767
"they one of techniques. several fastest among parametric the of are state Moreover, art the","Moreover, they are one of the fastest among several state of the art parametric techniques.",28768
under Statistical estimator conditions. our of is some established consistency,Statistical consistency of our estimator is established under some conditions.,28769
of parts social lives nowadays personal share media. large People their through,People nowadays share large parts of their personal lives through social media.,28770
"changes backward great appearance. in poses) and viewpoints, unusual","backward viewpoints, unusual poses) and great changes in appearance.",28771
"existing apart evaluation, propose a ranking metric. linguistic new metrics, For we from","For evaluation, apart from existing linguistic metrics, we propose a new ranking metric.",28772
procedure. validity show of results The the experimental the,The experimental results show the validity of the procedure.,28773
was improved disappeared the hemorrhage in cases. was of vitreous Macular edema most and,Macular edema was improved and vitreous hemorrhage was disappeared in most of the cases.,28774
"their recommended These been administrative economic have historically and for ethical, benefits. designs","These designs have historically been recommended for their ethical, administrative and economic benefits.",28775
"graphically. compared performance several designs can demonstrate be how this, of the we Following","Following this, we demonstrate how the performance of several designs can be compared graphically.",28776
process. actual generate the documents to toolkit structured a production synthetic document We emulating propose,We propose a toolkit to generate structured synthetic documents emulating the actual document production process.,28777
train perform can document Synthetic used analysis to systems to tasks. be documents,Synthetic documents can be used to train systems to perform document analysis tasks.,28778
"as formulated framework for weighting and Bayesian the classifier combination. a is ensemble Furthermore,","Furthermore, the ensemble is formulated as a Bayesian framework for classifier weighting and combination.",28779
The receiver a signals of containing is initial preamble. synchronization for different applicable time proposed,The proposed receiver is applicable for initial time synchronization of different signals containing a preamble.,28780
shifts. large can be handle modified further to Doppler It,It can further be modified to handle large Doppler shifts.,28781
"to of samples. This and leads processing test reduced time, computation faster recognition performance, improved","This leads to improved recognition performance, reduced computation time, and faster processing of test samples.",28782
gender classification and This on on research ethnicity focuses iris images. performing,This research focuses on performing ethnicity and gender classification on iris images.,28783
classification. for two ethnicity model each and is The on datasets proposed gender evaluated,The proposed model is evaluated on two datasets each for ethnicity and gender classification.,28784
of nature experimental the The challenging algorithms with problem. comparison and existing showcase the results,The experimental results and comparison with existing algorithms showcase the challenging nature of the problem.,28785
"gesture level. three the of consists posture, levels: and the The locomotion, the descriptor sub-action","The sub-action descriptor consists of three levels: the posture, the locomotion, and the gesture level.",28786
give three one the sub-action to address categories The three action levels representation for problem.,The three levels give three sub-action categories for one action to address the representation problem.,28787
are be thought to involved memory Spindle consolidation. in,Spindle are thought to be involved in memory consolidation.,28788
"junctions undergo long-term Moreover, activity-dependent gap plasticity.","Moreover, gap junctions undergo activity-dependent long-term plasticity.",28789
that We gap hypothesized modulate junction plasticity can oscillations. spindle,We hypothesized that gap junction plasticity can modulate spindle oscillations.,28790
medical important Radon techniques transform its in are and inverse operation tasks. imaging,Radon transform and its inverse operation are important techniques in medical imaging tasks.,28791
global performance its Our show improvement in retrieval version. versus by LRD using results significant,Our results show significant improvement in retrieval performance by using LRD versus its global version.,28792
users The eliminate and interferences successive remaining the through self-interference interference cancellation.,The users eliminate the remaining interferences through self-interference and successive interference cancellation.,28793
in formalization conditionals novel counterfactual quantified present We of a logic. modal a,We present a novel formalization of counterfactual conditionals in a quantified modal logic.,28794
and play moral a role ethical reasoning. vital in Counterfactual conditionals,Counterfactual conditionals play a vital role in ethical and moral reasoning.,28795
validate two model with our projects. We,We validate our model with two projects.,28796
shows can project be feasible. This that also model computationally our,This project also shows that our model can be computationally feasible.,28797
"problems agents environments. Reinforcement algorithms that solve interesting in can complex, train learning","Reinforcement learning algorithms can train agents that solve problems in complex, interesting environments.",28798
a highly training. This for environment complex that capable suggests requires a agent,This suggests that a highly capable agent requires a complex environment for training.,28799
the Karhunen-Lo\'eve transform (KLT) used The is in step. first,The Karhunen-Lo\'eve transform (KLT) is used in the first step.,28800
transform images a size. Multiple Saak cascaded larger to are of transforms,Multiple Saak transforms are cascaded to transform images of a larger size.,28801
backpropagation labels Neither used in data determination. is kernel nor,Neither data labels nor backpropagation is used in kernel determination.,28802
DRL applications cyber-physical technique have The validated. effectiveness the in three been of these,The effectiveness of the DRL technique in these three cyber-physical applications have been validated.,28803
"detection salient for bottom-up paper, propose images. In novel a this framework object panoramic we","In this paper, we propose a novel bottom-up salient object detection framework for panoramic images.",28804
accuracy. classifier Pretrained augmentation and networks increased data,Pretrained networks and data augmentation increased classifier accuracy.,28805
the these processes. in each play of Our stability electrophysiological analysis indicated of networks at,Our electrophysiological analysis indicated stability of the networks at play in each of these processes.,28806
estimations and also loss accuracy. high the complex mode-dependent extract amplitude channel We with,We also extract the complex amplitude channel estimations and mode-dependent loss with high accuracy.,28807
of locomotion models second estimating state considers part space input--output using legged The data.,The second part considers estimating state space models of legged locomotion using input--output data.,28808
identification state Our of basis for locomotion. results space a provide towards models legged,Our results provide a basis towards identification of state space models for legged locomotion.,28809
available examples of (i.e. the increase number As I/O,As the number of available I/O examples increase (i.e.,28810
"more), portfolio becomes or best induction the program thousand adapted a approach.","a thousand or more), portfolio adapted program induction becomes the best approach.",28811
strongly nodes how properties depend brain their are functional The on networks of chosen.,The properties of functional brain networks strongly depend on how their nodes are chosen.,28812
"Commonly, of Regions fMRI measurement of voxels. (ROIs), by defined are groupings nodes Interest pre-determined","Commonly, nodes are defined by Regions of Interest (ROIs), pre-determined groupings of fMRI measurement voxels.",28813
"Here, dynamic how networks. as ROIs nodes ask behave of brain we","Here, we ask how ROIs behave as nodes of dynamic brain networks.",28814
"there we voxel-level is rich that correlation ROIs. structure Finally, reveal inside","Finally, we reveal that there is rich voxel-level correlation structure inside ROIs.",28815
"from neuroscience node strategies Therefore, greatly definition tailored dynamical for networks. would network benefit","Therefore, network neuroscience would greatly benefit from node definition strategies tailored for dynamical networks.",28816
"error rate the is of network the average bit Consequently, calculated.","Consequently, the average bit error rate of the network is calculated.",28817
the outcomes is out Simulation carried paper. of to also verify the,Simulation is also carried out to verify the outcomes of the paper.,28818
ultrasound masses in and (BUS) breast for We localization classification of propose images. a framework,We propose a framework for localization and classification of masses in breast ultrasound (BUS) images.,28819
propose with training loss and a semi-supervised systematic training weakly selection. scenario appropriate We,We propose a systematic weakly and semi-supervised training scenario with appropriate training loss selection.,28820
node causes uncertainty Such the to locations. sensitivity blockages in interfering active,Such sensitivity to blockages causes uncertainty in the active interfering node locations.,28821
"blockage specifications. model In mmWave paper, first we propose this considering a","In this paper, we first propose a blockage model considering mmWave specifications.",28822
node configuration. and model randomness spatial spectral both The considers in proposed interference,The proposed interference model considers both spatial and spectral randomness in node configuration.,28823
validated analytical are results The via simulations. Monte-Carlo,The analytical results are validated via Monte-Carlo simulations.,28824
approach panel and web on generalizable showed crawled is promising images. Our solar results,Our approach is generalizable and showed promising results on web crawled solar panel images.,28825
"based we fast Finally algorithm, on optimization propose gradient a a proximal method. ,","Finally , we propose a fast optimization algorithm, based on a proximal gradient method.",28826
(CNNs) Convolutional have state-of-the-art achieved automatic networks medical for image performance segmentation. neural,Convolutional neural networks (CNNs) have achieved state-of-the-art performance for automatic medical image segmentation.,28827
"have clinical demonstrated robust and results they not However, sufficiently use. for accurate","However, they have not demonstrated sufficiently accurate and robust results for clinical use.",28828
a considering weighted propose network the function uncertainty fine-tuning. We and for interaction-based also loss,We also propose a weighted loss function considering network and interaction-based uncertainty for the fine-tuning.,28829
neural computer reverberating net of of simulation. study We means by dynamics a,We study dynamics of a reverberating neural net by means of computer simulation.,28830
network stimulated and freely. goes then different The with is initially stimuli,The network is initially stimulated with different stimuli and then goes freely.,28831
regimes firing the The formation. repetitive memory could periodic underpin in,The repetitive firing in the periodic regimes could underpin memory formation.,28832
provide Semantic of in We technologies proof existing Web to an concept. experimental order apply,We apply existing Semantic Web technologies in order to provide an experimental proof of concept.,28833
this methods tackle been proposed have through proposals. mining Numerous object to problem,Numerous methods have been proposed to tackle this problem through mining object proposals.,28834
to often approaches model sensitive to local solutions. initialization are undesirable Such and minimum converge,Such approaches are sensitive to model initialization and often converge to undesirable local minimum solutions.,28835
using and We proposals segmentation further multiple refine instance learning these cues.,We further refine these proposals using multiple instance learning and segmentation cues.,28836
on PASCAL ILSVRC detailed We and experimental validation VOC present datasets. the,We present detailed experimental validation on the PASCAL VOC and ILSVRC datasets.,28837
state-of-the-art that progressive results algorithm Experimental against methods. demonstrate performs the representation our adaptation favorably,Experimental results demonstrate that our progressive representation adaptation algorithm performs favorably against the state-of-the-art methods.,28838
"existing been detection depth has explored saliency in well information models. However, not","However, depth information has not been well explored in existing saliency detection models.",28839
"method stereoscopic proposed. In images a detection for this letter, novel is saliency","In this letter, a novel saliency detection method for stereoscopic images is proposed.",28840
community. emerging in It topic is newly a vision computer,It is a newly emerging topic in computer vision community.,28841
demonstrate of co-saliency the datasets RGBD proposed model. two on our Experiments effectiveness,Experiments on two RGBD co-saliency datasets demonstrate the effectiveness of our proposed model.,28842
"hard empowerment in especially to is prohibitively compute, nonlinear also continuous But spaces.","But empowerment is also prohibitively hard to compute, especially in nonlinear continuous spaces.",28843
"introduce policies. efficient, empowerment-maximising method learning We amortised an for","We introduce an efficient, amortised method for learning empowerment-maximising policies.",28844
"share Hence, networks noise usually large to sources. need neurons in","Hence, neurons in large networks usually need to share noise sources.",28845
"of the to that we transfers neurons. Finally, same show networks functional spiking design","Finally, we show that the same design transfers to functional networks of spiking neurons.",28846
in architecture deep have (Resnets) prominent learning. networks become a Residual,Residual networks (Resnets) have become a prominent architecture in deep learning.,28847
"of is still research. a ongoing understanding However, comprehensive a Resnets of topic","However, a comprehensive understanding of Resnets is still a topic of ongoing research.",28848
features. iterative refinement perform of recent view A that Resnets argues,A recent view argues that Resnets perform iterative refinement of features.,28849
expose of We attempt further properties this aspect. to,We attempt to further expose properties of this aspect.,28850
"To and empirically. end, Resnets this study both we analytically","To this end, we study Resnets both analytically and empirically.",28851
computationally efficient enables The formulation time-causal recognition. time-recursive,The time-recursive formulation enables computationally efficient time-causal recognition.,28852
experimental The demonstrates to compared state-of-the-art. competitive evaluation performance,The experimental evaluation demonstrates competitive performance compared to state-of-the-art.,28853
for models every domain. fundamental application Networks data are in practically used,Networks are fundamental models for data used in practically every application domain.,28854
addresses network several We a challenges. model task-focused selection methodology propose which key,We propose a task-focused network model selection methodology which addresses several key challenges.,28855
can representations. applied across and arbitrary be tasks avoid and sparse) to overfitting models,sparse) models to avoid overfitting and can be applied across arbitrary tasks and representations.,28856
"in and significance methodology. our We stability, sensitivity, show testing","We show stability, sensitivity, and significance testing in our methodology.",28857
"be and However, using spoofed lens. eyeballs can fabricated contact still it","However, it can still be spoofed using fabricated eyeballs and contact lens.",28858
available results architecture are algorithms. and state-of-the-art proposed promising The outperforms quite the lens detection,The proposed architecture results are quite promising and outperforms the available state-of-the-art lens detection algorithms.,28859
by reduce Images longer to wavelength captured introduced are effects a clouds. of the,Images captured by a longer wavelength are introduced to reduce the effects of clouds.,28860
is such (SAR) Aperture the clouds exist. Radar example that Synthetic improves visibility even an,Synthetic Aperture Radar (SAR) is such an example that improves visibility even the clouds exist.,28861
"the the On decreases the hand, increases. as spatial resolution other wavelength","On the other hand, the spatial resolution decreases as the wavelength increases.",28862
"is often However, order or pre-determined. cannot of variables be available not the","However, the order of variables is often not available or cannot be pre-determined.",28863
established the is of consistent some conditions. proposed under property regularity weak estimate The,The consistent property of the proposed estimate is established under some weak regularity conditions.,28864
method data The into of classification. proposed also for is discriminant linear applied real analysis,The proposed method is also applied into linear discriminant analysis of real data for classification.,28865
"landmarks blurred especially facilitates more accurate and for analysis, automated This small or pathologies. image","This facilitates more accurate automated image analysis, especially for small or blurred landmarks and pathologies.",28866
Carlo the past signal Monte in during methods popular (MC) very have decades. processing become,Monte Carlo (MC) methods have become very popular in signal processing during the past decades.,28867
"pdf proposal is it However, becomes more the computationally updated. demanding each sampling time from","However, sampling from the proposal pdf becomes more computationally demanding each time it is updated.",28868
"the faster approach. algorithm ARS Thus, the resulting standard than is","Thus, the resulting algorithm is faster than the standard ARS approach.",28869
neural that can solve learning-based recent task. convolutional We deep networks show this,We show that recent deep learning-based convolutional neural networks can solve this task.,28870
developing findings research motivate more robust future networks. deep into These,These findings motivate future research into developing more robust deep networks.,28871
"different In approach. apply will we a Learning step, Deep a next","In a next step, we will apply a different Deep Learning approach.",28872
"evaluation, and wage curricula determination, Applications credits risk assessment and include loans. for automatic","Applications include automatic curricula evaluation, wage determination, and risk assessment for credits and loans.",28873
"been more objective be Instead, efficient. shown has to function fairness the including in","Instead, including fairness in the objective function has been shown to be more efficient.",28874
term. independence the Both using criterion the on fairness Hilbert as Schmidt methods rely,Both methods rely on using the Hilbert Schmidt independence criterion as the fairness term.,28875
methods functions deal nonlinear by with the to formulation Replacing linear problems. kernel the allows,Replacing the linear formulation by kernel functions allows the methods to deal with nonlinear problems.,28876
"LEDs, orthogonal. large different of channel number UTs For vectors of asymptotically become a","For a large number of LEDs, channel vectors of different UTs become asymptotically orthogonal.",28877
"Four left organs abdominal kidneys) (liver, spleen, right were and evaluated.","Four abdominal organs (liver, spleen, left and right kidneys) were evaluated.",28878
the of the accuracy researched. on weight was Effect,Effect of the weight on the accuracy was researched.,28879
feasible promising. show method results The and our is,The results show our method is feasible and promising.,28880
"exhibit repetitive patterns independent translational two in regular are which wallpapers, directions. symmetry","wallpapers, are regular repetitive patterns which exhibit translational symmetry in two independent directions.",28881
disease based connectome the classification data. brain on This of paper problem considers,This paper considers the problem of brain disease classification based on connectome data.,28882
brain. representation is of A network a human connectome a,A connectome is a network representation of a human brain.,28883
approximate The features further classification. used are for eigenvalues obtained as,The obtained approximate eigenvalues are further used as features for classification.,28884
game pose tree significant branching for factors Games search challenge with a algorithms. large,Games with large branching factors pose a significant challenge for game tree search algorithms.,28885
crucial The inhibition for neuronal and balance is between excitation computation.,The balance between excitation and inhibition is crucial for neuronal computation.,28886
"networks areas. in many be neuronal have to inhomogeneous cortical been However, found","However, neuronal networks have been found to be inhomogeneous in many cortical areas.",28887
imaging Depth has properties. focused intrinsics largely on sensor and,Depth imaging has largely focused on sensor and intrinsics properties.,28888
"of largely acquire the accuracy is on capture. However, dependent the pixel","However, the accuracy of acquire pixel is largely dependent on the capture.",28889
in (DR) the world. prominent a Diabetic is Retinopathy blindness of cause,Diabetic Retinopathy (DR) is a prominent cause of blindness in the world.,28890
automated be system ophthalmologists. detection can for microaneurysm a helpful An,An automated microaneurysm detection can be a helpful system for ophthalmologists.,28891
"distance, However at person still a problem. challenging remains identification a","However person identification at a distance, still remains a challenging problem.",28892
identification. as Gait an feature for can recognition biometric be human seen and essential,Gait can be seen as an essential biometric feature for human recognition and identification.,28893
and tested We publicly results. achieved have have CASIA-B available state-of-the-art network database this over,We have tested this network over CASIA-B publicly available database and have achieved state-of-the-art results.,28894
with measures independent features learn implicitly. optimize We adversarial such which objectives to propose,We propose to learn independent features with adversarial objectives which optimize such measures implicitly.,28895
"various (airborne, transmission direct etc. contact, routes via The","The various transmission routes (airborne, via direct contact, etc.",28896
determines new hosts; pathogens fitness. infect their To success must the survive,To survive pathogens must infect new hosts; the success determines their fitness.,28897
"pathogens? can of concluded what evolutionary viral here, about From be strategies the","From here, what can be concluded about the evolutionary strategies of viral pathogens?",28898
the this addressed This question is paper. biological in,This is the biological question addressed in this paper.,28899
pathogens maxima out These fitness the as landscape. local of turn,These pathogens turn out as local maxima of the fitness landscape.,28900
"The agent-based equations, models networks, integral and simulation, involve differential- and probability.","The models involve differential- and integral equations, agent-based simulation, networks, and probability.",28901
applications has imaging interest been for early detection of hyperspectral There symptoms. increasing disease in,There has been increasing interest in hyperspectral imaging applications for early detection of disease symptoms.,28902
An a significant face reliable remains in method thus robust challenge preprocessing analysis. illumination,An illumination robust preprocessing method thus remains a significant challenge in reliable face analysis.,28903
transitions. of study effective An phase allows model the,An effective model allows the study of phase transitions.,28904
"model, generalized. our proposed two schemes: b) a) we (LoS) of In line consider sight","In our proposed model, we consider two schemes: a) line of sight (LoS) b) generalized.",28905
"ABSs the see LoS the LoS. should as the all users In IoT scheme,","In the LoS scheme, all the ABSs should see the IoT users as LoS.",28906
neurodegenerative Alzheimer's a Disease (AD) is disease. progressive,Alzheimer's Disease (AD) is a progressive neurodegenerative disease.,28907
achieves model state-of-the-art results. that our evaluation shows experimental Our,Our experimental evaluation shows that our model achieves state-of-the-art results.,28908
present causal for framework a set. We learning Bayesian a rule,We present a Bayesian framework for learning a causal rule set.,28909
We and search strategies efficiency. heuristics bounding grounded provide theoretically improve to,We provide theoretically grounded heuristics and bounding strategies to improve search efficiency.,28910
systems. disk Optic prerequisite is step segmentation in retinal automatic a screening,Optic disk segmentation is a prerequisite step in automatic retinal screening systems.,28911
"thresholding, robust various conditions. Adaptive to makes illumination and variations our acquisition algorithm image","Adaptive thresholding, makes our algorithm robust to illumination variations and various image acquisition conditions.",28912
"important. and real-time are more efficiency applications, In robustness","In real-time applications, efficiency and robustness are more important.",28913
"cast we abilities, clustering promote a two into these task. To multi-view the registration","To promote these two abilities, we cast the multi-view registration into a clustering task.",28914
approaches. on and datasets several with compared benchmark proposed approach We tested state-of-the-art our it,We tested our proposed approach on several benchmark datasets and compared it with state-of-the-art approaches.,28915
validate robustness sets. registration the its multi-view and for of results efficiency point Experimental,Experimental results validate its efficiency and robustness for the registration of multi-view point sets.,28916
are Traditional typically modeling. face based methods on,Traditional methods are typically based on face modeling.,28917
We Adversarial Network. an transfer based method on Generative end-to-end face propose,We propose an end-to-end face transfer method based on Generative Adversarial Network.,28918
Person cameras. the different across same Re-identification person identify is (ReID) to,Person Re-identification (ReID) is to identify the same person across different cameras.,28919
in of widely used series is forecasting multitude domains. Time a,Time series forecasting is widely used in a multitude of domains.,28920
lesser unable model requires but is generalized the The outperform martingale. assumptions linear to,The generalized linear model requires lesser assumptions but is unable to outperform the martingale.,28921
vehicle paper Wave processing (FM-CW) signals This Frequency classification. investigates Modulated-Continuos of radar for the,This paper investigates the processing of Frequency Modulated-Continuos Wave (FM-CW) radar signals for vehicle classification.,28922
model to predict good obtain using performance. When category we the vehicle the trained,When using the trained model to predict the vehicle category we obtain good performance.,28923
"these, All a data-driven learning deep fully approach. however, inhibit of reconstruction","All of these, however, inhibit a fully data-driven reconstruction deep learning approach.",28924
"In are the optimal chosen filters addition, an not sense. heuristic in","In addition, the heuristic filters are not chosen in an optimal sense.",28925
"directly tackle a to issue, the we this To propose formulation reconstruction learn filter.","To tackle this issue, we propose a formulation to directly learn the reconstruction filter.",28926
"allow will fully Furthermore, deep trainable learning data-driven approaches. this reconstruction","Furthermore, this will allow fully trainable data-driven reconstruction deep learning approaches.",28927
"controls. The monocular from directly images ""intentions"" a robot and single to camera maps intention-net","The intention-net maps images from a single monocular camera and ""intentions"" directly to robot controls.",28928
intention-net. the intentions provides planned The to path,The planned path provides intentions to the intention-net.,28929
"dependent the fully of signals. ratio was of Previously, the PR on compression specifications","Previously, the ratio of compression was fully dependent on the specifications of PR signals.",28930
induce aliasing the (ADC). analog-to-digital at intentional main idea signal The is converter to,The main idea is to induce intentional signal aliasing at the analog-to-digital converter (ADC).,28931
process detection the have We on dataset. KITTI proposed the evaluated extensively,We have extensively evaluated the proposed detection process on the KITTI dataset.,28932
that LiDAR provide auxiliary data information for CNN-based can approaches. indicates This,This indicates that LiDAR data can provide auxiliary information for CNN-based approaches.,28933
on approach. a non-rigid novel reports reconstruction template-free monocular paper This surface,This paper reports on a novel template-free monocular non-rigid surface reconstruction approach.,28934
thresholding and gradient. value proximal singular We camera by estimate pose,We estimate camera pose by singular value thresholding and proximal gradient.,28935
"matrix. space the the smoothing In proposed by measurement trajectory are constraints integrated of SMSR,","In the proposed SMSR, trajectory space constraints are integrated by smoothing of the measurement matrix.",28936
escaping trapped random under forces. to particles of process problem is the analogous The,The process is analogous to the escaping problem of trapped particles under random forces.,28937
extinction the strongly fixed depends nature on mean the point(s). of The deterministic time corresponding,The mean extinction time strongly depends on the nature of the corresponding deterministic fixed point(s).,28938
"later Fokker-Planck In simpler case, approach the approximation appropriate. is the also","In the later case, the simpler Fokker-Planck approximation approach is also appropriate.",28939
"segmentations. Particularly error-free provide regions, automatic tissue fail methods still deep to in","Particularly in deep tissue regions, automatic methods still fail to provide error-free segmentations.",28940
scatterers. surface due This the by to small-scale fading caused rough is,This is due to the small-scale fading caused by rough surface scatterers.,28941
for regime. outage closed-form for ratio derive the probability the further We expressions signal-to-noise high,We further derive closed-form expressions for the outage probability for the high signal-to-noise ratio regime.,28942
scratch. shows against quite competitive from pre-trained are The networks that result training,The result shows that pre-trained networks are quite competitive against training from scratch.,28943
the samples of of for number each is objective GAI The to minimize process.,The objective of GAI is to minimize the number of samples for each process.,28944
whose We algorithm develop lower also sample matches an complexity the almost bound.,We also develop an algorithm whose sample complexity almost matches the lower bound.,28945
learning real-world issue machine a is outliers Robustness central to in applications.,Robustness to outliers is a central issue in real-world machine learning applications.,28946
"limited color under At night, illumination. distortion occurs","At night, color distortion occurs under limited illumination.",28947
VPGNet available. will model publicly and benchmark the be The,The benchmark and the VPGNet model will be publicly available.,28948
"to further feature and fusion types Moreover, input strategies investigate we improve accuracy. detection into","Moreover, we investigate into input feature types and fusion strategies to further improve detection accuracy.",28949
data in two extremely theoretic are useful relationships sets. quantities Information or discovering more between,Information theoretic quantities are extremely useful in discovering relationships between two or more data sets.,28950
the neighbour quantities One is systems---for nearest estimating for these estimators. popular method---particularly continuous,One popular method---particularly for continuous systems---for estimating these quantities is the nearest neighbour estimators.,28951
results are the the scheme. merits of provided corroborate Numerical proposed to,Numerical results are provided to corroborate the merits of the proposed scheme.,28952
depths multi-focused different of images cameras. captured with of are focus These,These multi-focused images are captured with different depths of focus of cameras.,28953
of the which easily DCT block calculate developed Laplacian four domain. We in matrices,We developed four matrices which calculate the Laplacian of block easily in DCT domain.,28954
unsuitable to block reduce selection. greatly works error Our due,Our works greatly reduce error due to unsuitable block selection.,28955
enables information of individual. validate This an hypotheses to observation-based driving about analysts contextual the,This contextual information enables analysts to validate observation-based hypotheses about the driving of an individual.,28956
"real-world We the examples. interesting for derives contexts, how also through different characteristics demonstrate framework","We also demonstrate how the framework derives interesting characteristics for different contexts, through real-world examples.",28957
an included. is atomistic between reversible formation is solvent Bond and particles explicit,Bond formation between particles is reversible and an explicit atomistic solvent is included.,28958
no clusters. incorrectly assembled There are also,There are also no incorrectly assembled clusters.,28959
parameter spaces. Scaling allows asymptotic high-dimensional exploration of,Scaling allows asymptotic exploration of high-dimensional parameter spaces.,28960
central the report metabolism differences the clusters. carbon key of some We two in,We report some key differences in the central carbon metabolism of the two clusters.,28961
"the efficiency for we some modifications tests. several of Furthermore, show","Furthermore, we show the efficiency of several modifications for some tests.",28962
characterized has for evidence accumulation environments. of work Previous static temporal decision-making in,Previous work has characterized temporal accumulation of evidence for decision-making in static environments.,28963
"environments with real-world However, time. over statistics that decision-making change involves","However, real-world decision-making involves environments with statistics that change over time.",28964
"been studied However, from a memory spaces barely perspective. maps, cognitive theoretical unlike have","However, unlike cognitive maps, memory spaces have been barely studied from a theoretical perspective.",28965
"time-consuming sample multiple markers, tracking, impossible becomes and for for large sizes. especially Manual","Manual tracking, especially for multiple markers, becomes time-consuming and impossible for large sample sizes.",28966
"these for need in recent segmentation the Therefore, years. has automatic of grown markers","Therefore, the need for automatic segmentation of these markers has grown in recent years.",28967
each used team corresponding by and techniques We the the report performances.,We report the techniques used by each team and the corresponding performances.,28968
Convolutional achieve currently accuracy Networks Neural in (CNNs) classification. image state-of-the-art,Convolutional Neural Networks (CNNs) currently achieve state-of-the-art accuracy in image classification.,28969
"hierarchical a follow confusion Interestingly, over class structure patterns the classes. the","Interestingly, the class confusion patterns follow a hierarchical structure over the classes.",28970
is numerically. this striking of importance The relationship demonstrated,The striking importance of this relationship is demonstrated numerically.,28971
Numerical this the indicate weaker. is but results illustrate relationship,Numerical results illustrate this but indicate the relationship is weaker.,28972
"network for model recognition. present convolutional we In neural a this pose-based paper, gait new","In this paper, we present a new pose-based convolutional neural network model for gait recognition.",28973
"flow motion frames. To the optical estimate extract consecutive between we information,","To extract motion information, we estimate the optical flow between consecutive frames.",28974
We gait pose-based descriptors. model convolutional deep computes a that propose,We propose a deep convolutional model that computes pose-based gait descriptors.,28975
outperforms show The approach our methods. of that results state-of-the-art these experiments,The results of these experiments show that our approach outperforms state-of-the-art methods.,28976
"accurately show parameters. we Under the can that the mild statistical estimate assumptions, algorithm","Under mild statistical assumptions, we show that the algorithm can accurately estimate the parameters.",28977
results. preliminary to experiments Promising the theoretical support are presented,Promising preliminary experiments are presented to support the theoretical results.,28978
signals important in Filtering  is processing. and an images issue ,  Filtering is an important issue in signals and images processing.,28979
transform videos images Many and compressed discrete (DCT). using are cosine,Many images and videos are compressed using discrete cosine transform (DCT).,28980
"detection precise estimate To model essential. is parameters the measurements, noise impulse BG from a","To estimate the BG model parameters from noise measurements, a precise impulse detection is essential.",28981
operating is coupling positioning principle The of system of resonators. on based inductive the,The operating principle of the positioning system is based on inductive coupling of resonators.,28982
"educational information interfaces use include interactive Possible cases retrieval and scenarios. games,","Possible use cases include interactive games, information retrieval interfaces and educational scenarios.",28983
Meehan investigation by a recent et We al. extend,We extend a recent investigation by Meehan et al.,28984
latter the decision as treated The state. is,The latter is treated as the decision state.,28985
guaranteeing finding such ${\cal is constraints about under study R}$ the Mathematically stabilization. on problem,Mathematically the problem under study is about finding constraints on ${\cal R}$ guaranteeing such stabilization.,28986
this form partial sufficient (in the problem We a of solution conditions). found of,We found a partial solution of this problem (in the form of sufficient conditions).,28987
also other environments. on classes We mental comment of,We also comment on other classes of mental environments.,28988
presented. correct algorithm identify is such inaccuracies to An and,An algorithm to identify and correct such inaccuracies is presented.,28989
"tasks. Besides, can method improve performance vision of the our","Besides, our method can improve the performance of vision tasks.",28990
points. global and enforces the semantics on fine-grained provides the consistency Then FC-CRF,Then the FC-CRF enforces global consistency and provides fine-grained semantics on the points.,28991
Recurrent NN the as joint optimization. differentiable We latter implement a allow to,We implement the latter as a differentiable Recurrent NN to allow joint optimization.,28992
historical on analyze technique initialization documents. We the,We analyze the initialization technique on historical documents.,28993
the the detection for Sampling investigates for This Dropout first object of time. paper utility,This paper investigates the utility of Dropout Sampling for object detection for the first time.,28994
search a large-scale hashing method tasks. widely-used is in retrieval for Similarity-preserving nearest neighbour image,Similarity-preserving hashing is a widely-used method for nearest neighbour search in large-scale image retrieval tasks.,28995
via deep-network-based efficient There been hashing research generating representation on considerable has the image methods.,There has been considerable research on generating efficient image representation via the deep-network-based hashing methods.,28996
"the largely unsolved. representation deep searching issue space of remains However, in efficient the","However, the issue of efficient searching in the deep representation space remains largely unsolved.",28997
challenge. learning indicate task machine results significant a that the Preliminary poses,Preliminary results indicate that the task poses a significant machine learning challenge.,28998
the This decision inducing when on trees. privacy pat-terns focuses paper preserving sensitive of,This paper focuses on preserving the privacy of sensitive pat-terns when inducing decision trees.,28999
rules binary for in approach sensitive classification hiding record a We adopt aug-mentation datasets.,We adopt a record aug-mentation approach for hiding sensitive classification rules in binary datasets.,29000
role Electric plays infrastructure in rapid (EVs). of Providing adequate charging Vehicles a proliferation momentous,Providing adequate charging infrastructure plays a momentous role in rapid proliferation of Electric Vehicles (EVs).,29001
infrastructure EV Easy access would range. various limited mobility obstacles regarding to such remove,Easy access to such infrastructure would remove various obstacles regarding limited EV mobility range.,29002
BSS the investigates various of This advantages from the building perspectives. paper,This paper investigates the advantages of building the BSS from various perspectives.,29003
"fed Finally, scheme. these into processed are features tandem system, GMM-HMM in a being a","Finally, these features are processed in a tandem system, being fed into a GMM-HMM scheme.",29004
decision acts point after path combining log-likelihoods. the by The this fusion Viterbi,The decision fusion acts after this point by combining the Viterbi path log-likelihoods.,29005
We complex this by two stages. challenge process address the conditional decomposing generative into,We address this challenge by decomposing the complex generative process into two conditional stages.,29006
guide is this the semantic of effective map. formulated constraint segmentation spatial An to generation,An effective spatial constraint is formulated to guide the generation of this semantic segmentation map.,29007
through both demonstrate the evaluations. approach of our and effectiveness We qualitative quantitative,We demonstrate the effectiveness of our approach through both quantitative and qualitative evaluations.,29008
learning Block the features. are histograms as unsupervised then extracted,Block histograms are then extracted as the unsupervised learning features.,29009
"system are features speech Finally, for in tandem the a GMM-HMM obtained recognition. used","Finally, the obtained features are used in a tandem GMM-HMM system for speech recognition.",29010
The proposed via validated and simulations both is technique experimentally.,The proposed technique is validated both via simulations and experimentally.,29011
"efficient recognition not of is free However, action difficulty.","However, efficient action recognition is not free of difficulty.",29012
recognition challenges desired attain to insufficient existing make performance. methods data-driven These,These challenges make existing data-driven recognition methods insufficient to attain desired performance.,29013
"fully on-line Evaluation interfaces. a accessible Platform framework, RRC Annotation through The and modular is","The RRC Annotation and Evaluation Platform is a modular framework, fully accessible through on-line interfaces.",29014
Framework for use. Annotation available aspects All and RRC research are of Evaluation the,All aspects of the RRC Annotation and Evaluation Framework are available for research use.,29015
spaces Deep their encode complex input feature capacity the have data. of transformations to,Deep feature spaces have the capacity to encode complex transformations of their input data.,29016
"feature-space is understanding between relationship the encoded images relative However, two transformed difficult.","However, understanding the relative feature-space relationship between two transformed encoded images is difficult.",29017
"instance, rotated two relative what is feature between For relationship space the images?","For instance, what is the relative feature space relationship between two rotated images?",29018
when decoded we space? is interpolate What in feature,What is decoded when we interpolate in feature space?,29019
Disentangling very they nonlinear interact difficult these because is ways. in,Disentangling these is difficult because they interact in very nonlinear ways.,29020
recent proposed been extensions approaches the years. manifold Supervised in several learning have of,Supervised extensions of several manifold learning approaches have been proposed in the recent years.,29021
these frequentist distributions. with Bayesian can combined distributions be,Bayesian distributions can be combined with these frequentist distributions.,29022
system work This and content. video clothing fashion from retrieving for products proposes a,This work proposes a system for retrieving clothing and fashion products from video content.,29023
relies spatial on localization of information pipeline Our semantic markings. purely and road,Our pipeline purely relies on spatial localization and semantic information of road markings.,29024
"occlusions shadows. is from and conditions, algorithm vehicles, Furthermore, weather our to robust various","Furthermore, our algorithm is robust to various weather conditions, occlusions from vehicles, and shadows.",29025
medical imaging Hybrid imaging potential applications. in promises large,Hybrid imaging promises large potential in medical imaging applications.,29026
task. to for yield generators suitable proved this All results,All generators proved to yield suitable results for this task.,29027
by then obtained Instance-level are algorithms. labels conventional clustering,Instance-level labels are then obtained by conventional clustering algorithms.,29028
"accuracy real-world additionally training boosts on Furthermore, synthesized data. validation data on","Furthermore, additionally training on synthesized data boosts validation accuracy on real-world data.",29029
source Our be synthesized data will open-sourced. code and,Our source code and synthesized data will be open-sourced.,29030
is predictions close of obtained. games model with A actual between fit the,A close fit between predictions of the model with actual games is obtained.,29031
considerable matrix insight The the models covariance understanding for inverse multivariate statistical setting. in provides,The inverse covariance matrix provides considerable insight for understanding statistical models in the multivariate setting.,29032
minimax norm. estimation for estimating spectral bounds We precision under the matrices banded establish,We establish minimax estimation bounds for estimating banded precision matrices under the spectral norm.,29033
are theoretical sharpness our experiments bounds. the Our complemented demonstrating results of by,Our theoretical results are complemented by experiments demonstrating the sharpness of our bounds.,29034
stable several are The years. predictions over,The predictions are stable over several years.,29035
reviewing been has a time reducing some methods using the intelligent need challenge. for The,The need for reducing the reviewing time using some intelligent methods has been a challenge.,29036
WCE frames for The possibility stage of detection provide polyps. the early,The WCE frames provide the early stage possibility for detection of polyps.,29037
for the highest Different are sensitivity detection superpixel of find examined numbers polyps. SLIC to,Different SLIC superpixel numbers are examined to find the highest sensitivity for detection of polyps.,29038
promising is studies. superpixel improve previous results The the SLIC to segmentation of,The SLIC superpixel segmentation is promising to improve the results of previous studies.,29039
intention Anticipating actions applications. many observing one's human by has,Anticipating human intention by observing one's actions has many applications.,29040
loss. gradient whole network cross-entropy using policy jointly trained the and We,We jointly trained the whole network using policy gradient and cross-entropy loss.,29041
detector under untrained worked landmark a two-stage designed We that successfully scenarios.,We designed a two-stage landmark detector that successfully worked under untrained scenarios.,29042
"other in works literature, descriptor specially more in outperformed CNN-based CNN the scenarios. Our difficult","Our CNN descriptor outperformed other CNN-based works in the literature, specially in more difficult scenarios.",29043
"simulators, at large however, processing to wood. volume tend of These be slow","These simulators, however, tend to be slow at processing large volume of wood.",29044
are random algorithms walks methods. sampling from derived point Both interior,Both random walks are sampling algorithms derived from interior point methods.,29045
over Vaidya speed-up the in examples. walk The walk numerical illustrated of the are Dikin,The speed-up of the Vaidya walk over the Dikin walk are illustrated in numerical examples.,29046
\textit{all} This $\mathbb{R}^n$. measures for on result holds probability,This result holds for \textit{all} probability measures on $\mathbb{R}^n$.,29047
"a exceeding the large IJB-A, IJB-C, IJB-B and by previous margin. state-of-the-art","IJB-A, IJB-B and IJB-C, exceeding the previous state-of-the-art by a large margin.",29048
"paper, In we MR perform estimation-theoretic design for to fingerprinting. framework present this an experiment","In this paper, we present an estimation-theoretic framework to perform experiment design for MR fingerprinting.",29049
parameter time data optimized demonstrate experiments the acquisition and/or reduce improve substantially estimation. that We,We demonstrate that the optimized experiments substantially reduce data acquisition time and/or improve parameter estimation.,29050
"is segmentation the to considered Manual ""gold-standard"". be","Manual segmentation is considered to be the ""gold-standard"".",29051
evaluated generated standard gold CNN masks. and the silver We by models,We evaluated CNN models generated by the silver and gold standard masks.,29052
"in Stan, is the implemented For automatic. procedure estimation models","For models implemented in Stan, the estimation procedure is automatic.",29053
We examples. of the illustrate with the package functionality three,We illustrate the functionality of the package with three examples.,29054
user may participation their The in grids applications. limit smart increasing,The increasing user participation in smart grids may limit their applications.,29055
"for distributed In emerging paper, proposed. a in grids is load solution smart management this","In this paper, a distributed solution for load management in emerging smart grids is proposed.",29056
demonstrate effectiveness. its various systems test with results Simulation,Simulation results with various test systems demonstrate its effectiveness.,29057
division. Spindles structures chromosomes segregate during that cell are microtubule-based self-organized,Spindles are self-organized microtubule-based structures that segregate chromosomes during cell division.,29058
nucleation. the the is spindle controlled turnover and between The by mass microtubule balance of,The mass of the spindle is controlled by the balance between microtubule turnover and nucleation.,29059
regulation the poorly nucleation spatial of microtubule remain mechanisms that understood. control The,The mechanisms that control the spatial regulation of microtubule nucleation remain poorly understood.,29060
and of experiments we to modeling Here this mathematical issue. investigate use a combination,Here we use a combination of experiments and mathematical modeling to investigate this issue.,29061
microtubules and the around We and concentration in measure spindle. the tubulin of,We measure the concentration of tubulin and microtubules in and around the spindle.,29062
"and the localized generate. nucleators microtubule microtubules they activated Thus, are by both","Thus, microtubule nucleators are both localized and activated by the microtubules they generate.",29063
the molecules structural to encapsulated These hexamer. stability constrained impart and,These encapsulated and constrained molecules impart structural stability to the hexamer.,29064
reveal Simulations of the results of waters these cavity. removal that in a collapse,Simulations reveal that removal of these waters results in a collapse of the cavity.,29065
The combining multi-fidelity PCE. PCE is by constructed correction the and low-fidelity,The multi-fidelity PCE is constructed by combining the low-fidelity and correction PCE.,29066
"computed indices Sobol this Following using step, are this the PCE. combined","Following this step, the Sobol indices are computed using this combined PCE.",29067
"samples variance. to minimize splits the spatial each tree training, During the greedily","During training, each tree greedily splits the samples to minimize the spatial variance.",29068
"little, by is bright Intelligence (AI) building. revealing that future Artificial the Little are newspapers","Little by little, newspapers are revealing the bright future that Artificial Intelligence (AI) is building.",29069
antidote an crisis This inside. has job market possible,This possible job market crisis has an antidote inside.,29070
"their own digging By jobs, doing graves. passionately are workers their these","By passionately doing their jobs, these workers are digging their own graves.",29071
the restoration methods from clean to underlying image Image recover corrupted observations. aim,Image restoration methods aim to recover the underlying clean image from corrupted observations.,29072
"EPLL we to paper, this algorithm. the propose In approximations original three","In this paper, we propose three approximations to the original EPLL algorithm.",29073
time of in takes is hypotheses. shortcut but quadratic Hommel's exact the number,Hommel's shortcut is exact but takes quadratic time in the number of hypotheses.,29074
"is but linearithmic conservative. Hochberg's time, shortcut takes only","Hochberg's shortcut takes only linearithmic time, but is conservative.",29075
"other importantly, models. consumption deep that less than is much of the learning memory More","More importantly, the memory consumption is much less than that of other deep learning models.",29076
"adversarial knowledge some necessitate etc.) parameters, Most (architecture, existing crafting approaches examples for","Most existing approaches for crafting adversarial examples necessitate some knowledge (architecture, parameters, etc.)",29077
layer-wise present updates. of feedback-prop iterative on We based two residual and variants,We present two variants of feedback-prop based on layer-wise and residual iterative updates.,29078
interesting results unreported CNNs. deep but property previously of Our dynamic unveil a,Our results unveil a previously unreported but interesting dynamic property of deep CNNs.,29079
performance transformation. confirmed the Both superior elliptical results the of proposed,Both results confirmed the superior performance of the proposed elliptical transformation.,29080
local sensitive electrical activity. is in method Such fluctuations to a less,Such a method is less sensitive to local fluctuations in electrical activity.,29081
This two procedure. the in contains steps scheme encoding,This scheme contains two steps in the encoding procedure.,29082
"aim indoor from scenes RGBD In paper, this one we to interpret image.","In this paper, we aim to interpret indoor scenes from one RGBD image.",29083
"images. is means it directly lifting extend to for non-local denoising of Unfortunately, difficult","Unfortunately, it is difficult to directly extend lifting for non-local means denoising of images.",29084
"and horizontal image. separable the processing produces However, stripes the vertical in often","However, the separable processing often produces vertical and horizontal stripes in the image.",29085
one The the performing novelty for that searches. is observation in can lifting two-dimensional use,The novelty is in the observation that one can use lifting for performing two-dimensional searches.,29086
noise model fluctuating a generic selection. system for and a consider we demographic with Here,Here we consider a generic model for a system with demographic noise and fluctuating selection.,29087
"time parameterized their turn, correlation their fluctuations, amplitude are $\gamma$ $\delta$. in These by and","These fluctuations, in turn, are parameterized by their amplitude $\gamma$ and their correlation time $\delta$.",29088
"Hence, Sum of (EWSR). metric Expected typical optimization is choice Rate a the Weighted the","Hence, a typical choice of the optimization metric is the Expected Weighted Sum Rate (EWSR).",29089
"the presence daunting However, the the operator a of expectation optimization makes task.","However, the presence of the expectation operator makes the optimization a daunting task.",29090
for this approximation. derived They then constant bound a,They then derived a constant bound for this approximation.,29091
One Frequency candidate Division (GFDM). Multiplexing is Generalised important,One important candidate is Generalised Frequency Division Multiplexing (GFDM).,29092
"however, computations These require complex bit-level are using they highly if studies, simulators. performed","These studies, however, require highly complex computations if they are performed using bit-level simulators.",29093
channel reveal data. this results realistic the accuracy The model of using,The results reveal the accuracy of this model using realistic channel data.,29094
"gestures noisy. because human are sensors them remains by captured programming challenging However,","However, programming them remains challenging because human gestures captured by sensors are noisy.",29095
"is teacher training used samples a data. pre-trained from to the network observe large Specifically,","Specifically, a large pre-trained teacher network is used to observe samples from the training data.",29096
well better performs compared Loop on to descriptors. binary evaluated all as or datasets previous,Loop performs as well or better on all datasets evaluated compared to previous binary descriptors.,29097
constraints formulated been The of have application the specific problem. as requirements optimization additional,The application specific requirements have been formulated as additional constraints of the optimization problem.,29098
solvers optimization the formulated used mixed-integer been Recently solving in have problem. developed,Recently developed mixed-integer optimization solvers have been used in solving the formulated problem.,29099
apply meta-parameter for the random to technique the forests. We selection study of,We apply the technique to the study of meta-parameter selection for random forests.,29100
"we Based this, a fast propose conversion. on model","Based on this, we propose a fast model conversion.",29101
fundamental segmentation is neuroimage in a step first Brain analysis.,Brain segmentation is a fundamental first step in neuroimage analysis.,29102
regression a We generic provide LSTM based structure node. each for first,We first provide a generic LSTM based regression structure for each node.,29103
Extended comparison. (DEKF) also Filtering for Distributed based a training Kalman We algorithm introduce,We also introduce a Distributed Extended Kalman Filtering (DEKF) based training algorithm for comparison.,29104
for used detection gesture Machine and Vector classification. Support (SVM) was,Support Vector Machine (SVM) was used for gesture detection and classification.,29105
moments representations. We pseudo-Zernike image based sparse classification propose via radar,We propose radar image classification via pseudo-Zernike moments based sparse representations.,29106
methods publicly on prove MSTAR proposed the We validity the available of our dataset.,We prove the validity of our proposed methods on the publicly available MSTAR dataset.,29107
approach that issues. topological auxiliary limits an introduce impact parameter the of We,We introduce an auxiliary parameter approach that limits the impact of topological issues.,29108
 most osteoarthritis is common Knee (OA) the musculoskeletal disorder. ,  Knee osteoarthritis (OA) is the most common musculoskeletal disorder.,29109
"predictions level multiple becomes a capsule higher agree, When active.","When multiple predictions agree, a higher level capsule becomes active.",29110
performance model We our demonstrate of generalized Bayesian algorithm the problem. in a linear,We demonstrate the performance of our algorithm in a Bayesian generalized linear model problem.,29111
an problem. Image is matting vision important,Image matting is an important vision problem.,29112
it methods sampling-based combine propagation-based methods. stream and methods for main The,The main stream methods for it combine sampling-based methods and propagation-based methods.,29113
is especially trimap coarse. beneficial the This when is,This is especially beneficial when the trimap is coarse.,29114
be effects retrograde anti-aging effect adverse may responses motivated against side The circumstances. of,The anti-aging effect may be side effects of retrograde responses motivated against adverse circumstances.,29115
be metabolism than aging. ROS rather correlated closely may with more,ROS may be more closely correlated with metabolism rather than aging.,29116
We flaws to aim representation dataset caused discover potential bias. by,We aim to discover representation flaws caused by potential dataset bias.,29117
Experiments have demonstrated of method. the effectiveness our,Experiments have demonstrated the effectiveness of our method.,29118
are experimental The article. results in shown Cross-Art this dataset on,The experimental results on Cross-Art dataset are shown in this article.,29119
"that modalities cover visual, We on variety motion a and audio. of focused","We focused on a variety of modalities that cover visual, motion and audio.",29120
"better understand visualized how the we to each aggregation takes process Also, modality effect.","Also, we visualized the aggregation process to better understand how each modality takes effect.",29121
have Influenza studied. widely outbreaks been Objectives,Objectives Influenza outbreaks have been widely studied.,29122
"influenza between unexplored. remained and However, patterns festivals religious the","However, the patterns between influenza and religious festivals remained unexplored.",29123
"collected for dates were countries' and and Secondary of the Hajj Hanukkah. population, data the","Secondary data were collected for the countries' population, and the dates of Hajj and Hanukkah.",29124
further dates explored influenza patterns Weekly period. across study were and A the religious festival,Weekly influenza A patterns and religious festival dates were further explored across the study period.,29125
Detection is problem. of under a signal signal processing classical noise a,Detection of a signal under noise is a classical signal processing problem.,29126
localization has object of problem vision. become problems of mainstream the The of one,The problem of object localization has become one of the mainstream problems of vision.,29127
Henderson's stable equation. mixed and This investigates a to model article solve fast method,This article investigates a fast and stable method to solve Henderson's mixed model equation.,29128
tactic on operations through the enabled is design performed row This matrix.,This tactic is enabled through row operations performed on the design matrix.,29129
illumination response with in spikes sensor The scene sub-millisecond generates resolution to changes.,The sensor generates spikes with sub-millisecond resolution in response to scene illumination changes.,29130
also well of level-classes the method proposed as grows. scales number the We show,We also show the proposed method scales well as the number of level-classes grows.,29131
"can this for Even assumption though some be invalid. scenarios, particular","Even though for some particular scenarios, this assumption can be invalid.",29132
pattern The recognition. outperform that method results promising the demonstrates methods state-of-the-art on,The method demonstrates promising results that outperform the state-of-the-art methods on pattern recognition.,29133
critically robust role fitting computer vision. in important plays estimation problems a in consensus Maximum,Maximum consensus estimation plays a critically important role in robust fitting problems in computer vision.,29134
maximization Our work with linear begins constraints. by reformulating complementarity consensus,Our work begins by reformulating consensus maximization with linear complementarity constraints.,29135
subproblems the efficiently optimization. solve Both algorithms to perform convex,Both algorithms solve convex subproblems to efficiently perform the optimization.,29136
"Compared sizes. realistic on input exact the more approach to practical our is algorithms, much","Compared to the exact algorithms, our approach is much more practical on realistic input sizes.",29137
"our applicable to estimation residuals is approach . Further, geometric naturally problems with","Further, our approach is naturally applicable to estimation problems with geometric residuals .",29138
have as actively solution been tasks. a compression Bottleneck autoencoders researched image to,Bottleneck autoencoders have been actively researched as a solution to image compression tasks.,29139
"reconstructed images. produce we However, bottleneck that subjectively low quality autoencoders observed","However, we observed that bottleneck autoencoders produce subjectively low quality reconstructed images.",29140
switches circuit rearrange flicking by techniques to connections. implemented are specific These,These techniques are implemented by flicking specific switches to rearrange circuit connections.,29141
can by automation. in signals The be switches order to realize controlled,The switches can be controlled by signals in order to realize automation.,29142
global obtained a local by is count patches. of final the The sum,The final global count is obtained by a sum of the local patches.,29143
of rate conventional highest the outflow nasally. is Purpose: The aqueous humor,Purpose: The rate of conventional aqueous humor outflow is the highest nasally.,29144
"Except cleared BABB the entire structures, Results: pigmented for eye.","Results: Except for pigmented structures, BABB cleared the entire eye.",29145
of from drainage were Networks the downstream lumina TM interconnected structures. to traced,Networks of interconnected lumina were traced from the TM to downstream drainage structures.,29146
"outflow Conclusions: analysis volumetric of of cleared eyes enabled the RSCM tract. high-resolution,","Conclusions: RSCM of cleared eyes enabled high-resolution, volumetric analysis of the outflow tract.",29147
in multi-person The task scenes estimation quite is natural human of pose challenging.,The task of multi-person human pose estimation in natural scenes is quite challenging.,29148
approaches. include methods and both Existing top-down bottom-up,Existing methods include both top-down and bottom-up approaches.,29149
are regression dependent two naturally tasks each on These other.,These two regression tasks are naturally dependent on each other.,29150
many Synthesizing face real sketches inverse and have photos applications. its from,Synthesizing face sketches from real photos and its inverse have many applications.,29151
be presents that a work fully classifier designed is This new to interpretable. specifically,This work presents a new classifier that is specifically designed to be fully interpretable.,29152
poses is caused attenuation the challenge rain. most serious by the EHF The band,The most serious challenge the EHF band poses is the attenuation caused by rain.,29153
measured The used prediction. rain is for statistics this,The measured rain statistics is used for this prediction.,29154
variability of rain Bangladesh. also of different investigated over is attenuation The sessions the,The variability of the rain attenuation is also investigated over different sessions of Bangladesh.,29155
The conditions information is mathematically characterized. under false attack different,The false information attack under different conditions is mathematically characterized.,29156
"a be To is tracking used concrete, example an paper. system target in the as","To be concrete, a target tracking system is used as an example in the paper.",29157
crucial knowledge domain any is task. for Representing,Representing domain knowledge is crucial for any task.,29158
"analyzed categorized classifiers. The are descriptors by works fundamental parts, and i.e., their","The analyzed works are categorized by their fundamental parts, i.e., descriptors and classifiers.",29159
of image Quantitative classification analysis accurate often pixels on process. a depends through segmentation,Quantitative image analysis often depends on accurate classification of pixels through a segmentation process.,29160
"model. NCS paper, formulation alternative a we an to In propose learn this","In this paper, we propose an alternative formulation to learn a NCS model.",29161
"text. of it and top features horizontal On texts, conventional multi-oriented curved-oriented the","On top of the conventional horizontal and multi-oriented texts, it features curved-oriented text.",29162
"metrics evaluation relevance and Furthermore, alternative clinical we in propose their and practice. discuss","Furthermore, we propose and discuss alternative evaluation metrics and their relevance in clinical practice.",29163
of mitigating attack presentation paramount that is importance. proposed results The experimental synthetic the illustrate,The experimental results illustrate that mitigating the proposed synthetic presentation attack is of paramount importance.,29164
What contexts? in data-to-text mean vagueness does,What does vagueness mean in data-to-text contexts?,29165
ways sets what fuzzy to theory improve systems? data-to-text can contribute In,In what ways can fuzzy sets theory contribute to improve data-to-text systems?,29166
 spiked the data for We classifiers (SSE) high-dimensional eigenvalue consider model.  under strongly,  We consider classifiers for high-dimensional data under the strongly spiked eigenvalue (SSE) model.,29167
high-dimensional We have first often that data the SSE model. show,We first show that high-dimensional data often have the SSE model.,29168
eigenstructures using We model. SSE classifier for a consider the distance-based,We consider a distance-based classifier using eigenstructures for the SSE model.,29169
give performance the classifier. new the studies simulation and of We discuss,We give simulation studies and discuss the performance of the new classifier.,29170
"sets. microarray new using data classifier Finally, demonstrate we by the","Finally, we demonstrate the new classifier by using microarray data sets.",29171
a for technique regression new global we optimality. that symbolic this In guarantees introduce study,In this study we introduce a new technique for symbolic regression that guarantees global optimality.,29172
improving Littering for cities. quantification of important is an step cleanliness,Littering quantification is an important step for improving cleanliness of cities.,29173
types Collected of images different wastes. containing,Collected images containing different types of wastes.,29174
and the developed images training then for system. annotated are benchmarking These,These images are then annotated for training and benchmarking the developed system.,29175
Our results variant backgrounds. on of on accurate detection show real case scenarios littering,Our results on real case scenarios show accurate detection of littering on variant backgrounds.,29176
problem vision. a is retrieval in challenging Clothing computer,Clothing retrieval is a challenging problem in computer vision.,29177
also have approach such demonstrate face the in that We can benefits privacy-friendly localization.,We also demonstrate the benefits that such approach can have in privacy-friendly face localization.,29178
inherit It from to design both distances. distances new interesting therefore is the that best,It is therefore interesting to design new distances that inherit the best from both distances.,29179
represent the the collective motion also model propose We a motion of to crowd.,We also propose a motion model to represent the collective motion of the crowd.,29180
The is to trivially algorithm activity. also recognize individual extended,The algorithm is also extended trivially to recognize individual activity.,29181
set closed scenarios. Current focuses research retrieval on logo,Current logo retrieval research focuses on closed set scenarios.,29182
information class-specific topological in descriptors. is show well Results that reflected,Results show that class-specific information is reflected well in topological descriptors.,29183
information. complementary capture investigated can descriptors directly non-topological with The and compete descriptors,The investigated descriptors can directly compete with non-topological descriptors and capture complementary information.,29184
a As with descriptors. they non-topological state-of-the-art improve consequence when combined the,As a consequence they improve the state-of-the-art when combined with non-topological descriptors.,29185
various Imaging Magnetic for important diagnostic precise of (MRI) Resonance is tool pathologies. an detection,Magnetic Resonance Imaging (MRI) is an important diagnostic tool for precise detection of various pathologies.,29186
in be was The accurate method to tumor. robust proposed and found fairly detecting,The proposed method was found to be fairly accurate and robust in detecting tumor.,29187
show association structure group this of and estimation detection. hidden corresponds track We that to,We show that estimation of this hidden structure corresponds to track association and group detection.,29188
programming We formulation. structure estimate hidden a under this linear,We estimate this hidden structure under a linear programming formulation.,29189
solved a is under problem SVM framework. This structured,This problem is solved under a structured SVM framework.,29190
"considered Polysomnography, is for gold the task. standard such","Polysomnography, is considered the gold standard for such task.",29191
"not it unwieldy and very However, analysis. suitable long-term for therefore is","However, it is very unwieldy and therefore not suitable for long-term analysis.",29192
under can the beneficial that behaviour conditions to Here cheating be certain population. we argue,Here we argue that under certain conditions cheating behaviour can be beneficial to the population.,29193
which cooperating in a pathogen. system Here models we an antagonizes immune study,Here we study models in which an immune system antagonizes a cooperating pathogen.,29194
concentration we bounds. question In used to order this of analyze important measure,In order to analyze this important question we used concentration of measure bounds.,29195
obtained we the applying of When estimator's distribution. tight bounds the bounds these tail on,When applying these bounds we obtained tight bounds on the tail of the estimator's distribution.,29196
tail provide We the probability on norm. bounds error's estimation,We provide probability tail bounds on the estimation error's norm.,29197
estimation simple and analysis is $L_{\infty}$ error. Our on type simple uses method bounds the,Our analysis method is simple and uses simple $L_{\infty}$ type bounds on the estimation error.,29198
The is tightness the of simulation. through tested bounds,The tightness of the bounds is tested through simulation.,29199
The this models novel. general sample squares noise of finite least analysis model is with,The finite sample analysis of least squares models with this general noise model is novel.,29200
"proposed the evaluate We two problems, method on of the efficiency i.e.","We evaluate the efficiency of the proposed method on two problems, i.e.",29201
and limit stock prediction on order based data. price book analysis image facial,facial image analysis and stock price prediction based on limit order book data.,29202
"the turn in This space. probability sampling algorithms in latent and improves clustering distributions,","This in turn improves probability distributions, sampling algorithms and clustering in the latent space.",29203
"interaction CR of and vastly unexplored. remained However, has the control","However, the interaction of CR and control has remained vastly unexplored.",29204
"be separation violated. to is the shown principle Consequently,","Consequently, the separation principle is shown to be violated.",29205
also conditions Several are stability of discussed. stochastic,Several conditions of stochastic stability are also discussed.,29206
results. effectiveness Illustrative provided are the the numerical of show examples to,Illustrative numerical examples are provided to show the effectiveness of the results.,29207
"data difficult. detection training hyperspectral target problems, is many accurately In acquiring labeled","In many hyperspectral target detection problems, acquiring accurately labeled training data is difficult.",29208
"can be After detector test based signature on a then target applied data. learning signatures,","After learning target signatures, a signature based detector can then be applied on test data.",29209
"is to an proposed condition is exhibit example Besides, sharp. the","Besides, an example is proposed to exhibit the condition is sharp.",29210
the during modelling words examine unseen the then embeddings We of training. in capacity,We then examine the capacity of the embeddings in modelling words unseen during training.,29211
"circumstances. However, certain be may assumption in not the true","However, the assumption may not be true in certain circumstances.",29212
"possibility we genuine the of samples. Therefore, only explore CAS using implementing","Therefore, we explore the possibility of implementing CAS using only genuine samples.",29213
We form in quantization loss propose a new measured triplets. of therefore,We therefore propose a new form of quantization loss measured in triplets.,29214
to optimization localization. a of functions provide joint time-frequency solution the Slepian problem,Slepian functions provide a solution to the optimization problem of joint time-frequency localization.,29215
"steps. of make paper, we significant improvements two this In the both on","In this paper, we make significant improvements on both of the two steps.",29216
"the of idea apply steps. we in extensively Besides, show pre-training both its importance and","Besides, we apply the idea of pre-training extensively and show its importance in both steps.",29217
"the To design reduce propose codebook a design. symmetric complexity, we also","To reduce the design complexity, we also propose a symmetric codebook design.",29218
"minimum maximized. the distance to For aims approaches, design proposed the all be Euclidean","For all the proposed design approaches, the minimum Euclidean distance aims to be maximized.",29219
design symmetric complexity and design codebook reduce can simultaneously. Our detection,Our symmetric codebook design can reduce design and detection complexity simultaneously.,29220
further improve of the results proposed confirm these the that The descriptor. techniques performance,The results confirm that these techniques further improve the performance of the proposed descriptor.,29221
of Neuroscience. one Codifying modern the memories problems of is fundamental,Codifying memories is one of the fundamental problems of modern Neuroscience.,29222
functional behind mechanisms remain unknown. this largely phenomenon The,The functional mechanisms behind this phenomenon remain largely unknown.,29223
phenomena. and Theorems on of concentration measure the based Separation Stochastic argument is The,The argument is based on Stochastic Separation Theorems and the concentration of measure phenomena.,29224
a organization for neurons. of single in basis ensembles results memories These constitute complex of,These results constitute a basis for organization of complex memories in ensembles of single neurons.,29225
"produced. such biologically pose predictions implausible circumstances, Under be may","Under such circumstances, biologically implausible pose predictions may be produced.",29226
typically such Explicit is challenging. constraints of learning,Explicit learning of such constraints is typically challenging.,29227
follows conditional of the Generative Adversarial Training (GANs). of network strategy Networks the,Training of the network follows the strategy of conditional Generative Adversarial Networks (GANs).,29228
"may data around to Some meaningful their information value. explanations questions subjects, offer raising little","Some explanations may offer little meaningful information to data subjects, raising questions around their value.",29229
in of support goals these We assess the each GDPR. how finds,We assess how each of these goals finds support in the GDPR.,29230
"to approaches network improve problem, propose To simple-yet-effective we three the this training. mitigate","To mitigate this problem, we propose three simple-yet-effective approaches to improve the network training.",29231
"find First, two-stage to use good we a to local progressively propose optimization strategy minima.","First, we propose to use a two-stage optimization strategy to progressively find good local minima.",29232
"propose quantized net quantized optimize we and a weights first Specifically, activations. then with to","Specifically, we propose to first optimize a net with quantized weights and then quantized activations.",29233
the is in simultaneously. methods to them contrast This traditional which optimize,This is in contrast to the traditional methods which optimize them simultaneously.,29234
"the low-precision the By to model model provides hints training. doing so, guide full-precision","By doing so, the full-precision model provides hints to guide the low-precision model training.",29235
"we convex typically iterative solve require problem, this optimization To solvers.","To solve this problem, we typically require iterative convex optimization solvers.",29236
then We based graph graph. on apply this filters low-pass,We then apply low-pass graph filters based on this graph.,29237
"However, input->output the on transform. describing only literature conventional is usually focused too the","However, the conventional literature is usually too focused on only describing the input->output transform.",29238
analytic decoding. and The implies for inverse to response the ensure the conditions model,The analytic inverse implies conditions for the response and the model to ensure decoding.,29239
"a further developed is evaluated, Various weighting scheme to accuracy. distance metrics and are boost","Various distance metrics are evaluated, and a weighting scheme is developed to further boost accuracy.",29240
Convolutional (CNNs) approach have successful vision-related most networks the neural in become many domains.,Convolutional neural networks (CNNs) have become the most successful approach in many vision-related domains.,29241
"domains data they However, are to limited where is abundant.","However, they are limited to domains where data is abundant.",29242
depth This extracting various allows tasks. sacrificing without relevant features at to task-specific features all,This allows extracting task-specific features at various depth without sacrificing features relevant to all tasks.,29243
image datasets There noisy leverage trend emerging visual an in many is tasks. recognition to,There is an emerging trend to leverage noisy image datasets in many visual recognition tasks.,29244
"the among noise deep} datasets approaches. label However, \mbox{performance the of severely the degenerates learning","However, the label noise among the datasets severely degenerates the \mbox{performance of deep} learning approaches.",29245
"still be the is applied able to labels At the training. time, high-quality same for","At the same time, the high-quality labels is still able to be applied for training.",29246
have Comprehensive demonstrated learning state-of-the-art results CAN the approaches. outperforms deep,Comprehensive results have demonstrated CAN outperforms the state-of-the-art deep learning approaches.,29247
"network a precoding. develop For for we MIMO recurrent highly energy-efficient systems, massive neural parallel","For massive MIMO systems, we develop a highly parallel recurrent neural network for energy-efficient precoding.",29248
in vision. of the computer stereo one is Computational problems classical,Computational stereo is one of the classical problems in computer vision.,29249
"the a Furthermore, learning propose simple hugely correlation we problem. aggregation that space simplifies","Furthermore, we propose a simple space aggregation that hugely simplifies the correlation learning problem.",29250
eyes-open cases were for and head-tilt eyes-closed investigated each condition. The,The eyes-open and eyes-closed cases were investigated for each head-tilt condition.,29251
lengths suffice not does experimental demonstrate findings. changing muscle to our explain We that,We demonstrate that changing muscle lengths does not suffice to explain our experimental findings.,29252
"reflex and stretch be the to tremor physiological of decoupled. First, it amplitudes permits","First, it permits the amplitudes of physiological tremor and stretch reflex to be decoupled.",29253
"and joint treaties, for has result the implications design The robots. contracts, of ventures,","The result has implications for the design of contracts, treaties, joint ventures, and robots.",29254
a in (RL). \textit{reinforcement methods} We present combines techniques framework \textit{formal that with learning},We present a framework that combines techniques in \textit{formal methods} with \textit{reinforcement learning} (RL).,29255
prediction DTI the for approach. RBM our Results: method compare well-known We against,Results: We compare our method for DTI prediction against the well-known RBM approach.,29256
"model Our categories. with huge, on a be can infinite, problems number potentially of used","Our model can be used on problems with a huge, potentially infinite, number of categories.",29257
techniques demonstrated performance problems. tackling have important state-of-the-art optimisation Bayesian Information-theoretic global optimisation in,Information-theoretic Bayesian optimisation techniques have demonstrated state-of-the-art performance in tackling important global optimisation problems.,29258
"data. multi-view of this the deep based paper, on autoencoder investigate methods performance we In","In this paper, we investigate the performance of deep autoencoder based methods on multi-view data.",29259
graph also model. To the of we thermodynamic demonstrate consistency a developed version bond,To demonstrate thermodynamic consistency we also developed a bond graph version of the model.,29260
"image Prior have aesthetics. interestingness, actively popularity, high-level attributes studied and including memorability, works","Prior works have actively studied high-level image attributes including interestingness, memorability, popularity, and aesthetics.",29261
"human studied none of ever has them shot. action However, ""attractiveness"" of the","However, none of them has ever studied the ""attractiveness"" of human action shot.",29262
did accuracy. comparison This the that not evidence provides harm strong a speed,This comparison provides a strong evidence that speed did not harm the accuracy.,29263
inference methods to causal biases. have mitigate these developed been Many,Many causal inference methods have been developed to mitigate these biases.,29264
"observed the synth-validation, with In generative known to estimate distributions treatment effects. data we use","In synth-validation, we use the observed data to estimate generative distributions with known treatment effects.",29265
delays loss like pollutions. cause heavy crossing economics huge problems Border and environmental,Border crossing delays cause problems like huge economics loss and heavy environmental pollutions.,29266
the center head It their curve. toward that a tilt known of car is drivers,It is known that car drivers tilt their head toward the center of a curve.,29267
"the to susceptible In carsickness than are less are generally passengers. drivers addition,","In addition, drivers are generally less susceptible to carsickness than are the passengers.",29268
examples intentionally pick to the their informative show Teachers students. most,Teachers intentionally pick the most informative examples to show their students.,29269
methods nuisance important is such An issue biological for separating variation. from signal known relevant,An important known issue for such methods is separating relevant biological signal from nuisance variation.,29270
"combination and saliency a Next, of map background foreground performed. is","Next, a combination of foreground and background saliency map is performed.",29271
evaluations against other the demonstrate Extensive outstanding our method proposed experimental of methods. effectiveness,Extensive experimental evaluations demonstrate the effectiveness of our proposed method against other outstanding methods.,29272
to train Those and expensive models difficult are to parameterize.,Those models are expensive to train and difficult to parameterize.,29273
"other data test most the augmentation are influential. input resolution and the factors, Of","Of the other factors, test data augmentation and input resolution are the most influential.",29274
"when data, also combined, models, with Deeper extra help.","Deeper models, when combined, with extra data, also help.",29275
"absence critical, learning is experiment: penalties. huge its transfer Transfer performance brings --","-- transfer experiment: Transfer learning is critical, its absence brings huge performance penalties.",29276
automated curating on analysis public Conclusions Advancing skin datasets. lesion requires Significance: research and larger,Conclusions and Significance: Advancing research on automated skin lesion analysis requires curating larger public datasets.,29277
"enhancement. low-light system, a for image fusion we multi-exposure framework design by Inspired human visual","Inspired by human visual system, we design a multi-exposure fusion framework for low-light image enhancement.",29278
"first estimation fusion for illumination the weight image techniques. Specifically, design matrix we using","Specifically, we first design the weight matrix for image fusion using illumination estimation techniques.",29279
our synthesize images. introduce Then camera response to multi-exposure we model,Then we introduce our camera response model to synthesize multi-exposure images.,29280
difficult. of is amplitude The phase and reconstruction,The reconstruction of phase and amplitude is difficult.,29281
MATLAB The algorithms implemented a as are Toolbox.,The algorithms are implemented as a MATLAB Toolbox.,29282
The tests. demonstrated the of algorithms efficiency simulation is by,The efficiency of the algorithms is demonstrated by simulation tests.,29283
We sparse recurrent a convolutional model. propose auto-encoder,We propose a convolutional recurrent sparse auto-encoder model.,29284
image answer boxes an from We the answer to extract use technique sheets. registration,We use an image registration technique to extract the answer boxes from answer sheets.,29285
approach. a two straight-forward strategies classification: We two-stage classifier of evaluate and approach a,We evaluate two strategies of classification: a straight-forward approach and a two-stage classifier approach.,29286
two We feature neural network. methods handcrafted a test and convolutional,We test two handcrafted feature methods and a convolutional neural network.,29287
"easy-to-use the In the proposed graphical present an end, we interface system. of user","In the end, we present an easy-to-use graphical user interface of the proposed system.",29288
"updating strategies and propose, weights. several evaluate We for empirically analyze, the","We propose, analyze, and empirically evaluate several strategies for updating the weights.",29289
proposed We compare recommendations. strategies our and the conclude with statistically,We statistically compare the proposed strategies and conclude with our recommendations.,29290
"data paper, augmentation generative propose method adversarial networks In a this using we (GAN).","In this paper, we propose a data augmentation method using generative adversarial networks (GAN).",29291
data the find complement complete It and neighboring manifold between can classes. better margins and,It can complement and complete the data manifold and find better margins between neighboring classes.,29292
"we least-squared adversarial gradient as the vanishing employ order avoid In loss to problem, loss.","In order to avoid gradient vanishing problem, we employ the least-squared loss as adversarial loss.",29293
evaluation validate three to also datasets benchmark propose performance. methods several GAN's We on,We also propose several evaluation methods on three benchmark datasets to validate GAN's performance.,29294
sequential data more Modeling become in has and important practice. more,Modeling sequential data has become more and more important in practice.,29295
"Some forecasting. sensors driving, weather and are virtual autonomous applications","Some applications are autonomous driving, virtual sensors and weather forecasting.",29296
To model so recurrent such systems used. are models called,To model such systems so called recurrent models are used.,29297
variational that resulting the can an lower exists. bound for We distribution optimal show,We can show that for the resulting lower bound an optimal variational distribution exists.,29298
through Training bound. optimizing variational lower realized the is,Training is realized through optimizing the variational lower bound.,29299
"the complexity. Variational computational reduce Using (DVI), Distributed we can Inference","Using Distributed Variational Inference (DVI), we can reduce the computational complexity.",29300
"easily can learning, e.g. for unsupervised be adapted Furthermore, method our","Furthermore, our method can easily be adapted for unsupervised learning, e.g.",29301
its and the variable deep version. model latent,the latent variable model and its deep version.,29302
"detection video-based is prerequisite pervasive for accurate, robust eye-tracking. essential an Real-time, pupil and","Real-time, accurate, and robust pupil detection is an essential prerequisite for pervasive video-based eye-tracking.",29303
"sufficient CLE detection quality automatic was shown. Recently, images the of feasibility carcinoma for of","Recently, the feasibility of automatic carcinoma detection for CLE images of sufficient quality was shown.",29304
types artifact deteriorations. motion-induced prevalent most the image are Amongst,Amongst the most prevalent artifact types are motion-induced image deteriorations.,29305
"of detection. important Hence, clinical work provides applicability automatic this step towards an carcinoma","Hence, this work provides an important step towards clinical applicability of automatic carcinoma detection.",29306
"conventional machine and assessed. learning Both, learning-based deep novel, were approaches","Both, conventional machine learning and novel, deep learning-based approaches were assessed.",29307
"this paper, the we classification. problem video address In multi-label of","In this paper, we address the problem of multi-label video classification.",29308
"classification. external into knowledge the to propose gap, incorporate graphs narrowing video we Towards","Towards narrowing the gap, we propose to incorporate external knowledge graphs into video classification.",29309
"are available. publicly Moreover, methods only few","Moreover, only few methods are publicly available.",29310
"near to can target reducing accuracy. generated features ambiguous being boundaries, lead This classification class","This can lead to ambiguous features being generated near class boundaries, reducing target classification accuracy.",29311
multi-spectral we a paper In registration for and fully approach fusion. image introduce this end-to-end,In this paper we introduce a fully end-to-end approach for multi-spectral image registration and fusion.,29312
grayscale stream The of frames. is a sensor image standard global-shutter sequence APS,The APS stream is a sequence of standard grayscale global-shutter image sensor frames.,29313
"information on RGBD image depth ignore RGB and Moreover, they the for images. mainly focus","Moreover, they mainly focus on RGB image and ignore the depth information for RGBD images.",29314
to scheme consistent homogeneous co-saliency proposed is and obtain more The map. iteration,The iteration scheme is proposed to obtain more homogeneous and consistent co-saliency map.,29315
The cosaliency RGBD demonstrate two datasets proposed effectiveness the of our on experiments framework.,The experiments on two RGBD cosaliency datasets demonstrate the effectiveness of our proposed framework.,29316
exploit meet to performance. demands integral block the as histogram We building real-time of,We exploit integral histogram as building block to meet the demands of real-time performance.,29317
"strategies. is, between There gap a two significant however, these","There is, however, a significant gap between these two strategies.",29318
"the rendering methods unreliable. altogether, results throw Modern geometric often deep constraints away more","Modern deep methods more often throw away geometric constraints altogether, rendering the results unreliable.",29319
bring disparate make two In paper we to strategies these an together. effort this seemingly,In this paper we make an effort to bring these two seemingly disparate strategies together.,29320
a sensing. problem hyperspectral processing is remote in Spectral (SU) data unmixing,Spectral unmixing (SU) is a data processing problem in hyperspectral remote sensing.,29321
"algorithm, been the including employed. clusters has network single-node proposed a In","In the proposed algorithm, a network including single-node clusters has been employed.",29322
this pixel hyperspectral in network. as in images a considered node Each,Each pixel in hyperspectral images considered as a node in this network.,29323
of has by effectiveness supported studies. practically theoretically and regularization The been several sparse,The effectiveness of sparse regularization has been supported practically and theoretically by several studies.,29324
"Interpretable ""Independently new (IILasso). regularization propose In Lasso"" method, paper, we this a","In this paper, we propose a new regularization method, ""Independently Interpretable Lasso"" (IILasso).",29325
"intuitively also can Hence, improve performance overfitting. by the regression we avoiding coefficients interpret and","Hence, we can interpret regression coefficients intuitively and also improve the performance by avoiding overfitting.",29326
analyses data and Synthetic also the real of IILasso. indicate effectiveness,Synthetic and real data analyses also indicate the effectiveness of IILasso.,29327
"five were different into shapes. designed Firstly, the markers","Firstly, the markers were designed into five different shapes.",29328
"TensorFlow on-line. trained data, are models The codes available and","The data, trained models and TensorFlow codes are available on-line.",29329
"effects information. other, had so on effect These that each population offset total little activity","These effects offset each other, so that population activity had little effect on total information.",29330
Semantic as on leap already Web. The has internet next the started,The next leap on the internet has already started as Semantic Web.,29331
discusses of which Semantic industries currently eight This reap benefits different the paper Web.,This paper discusses eight different industries which currently reap the benefits of Semantic Web.,29332
empirically our We analysis analytically. perform an both of also extensive and attention module,We also perform an extensive analysis of our attention module both empirically and analytically.,29333
resolution the in The on are effects deteriorating manner. of diffraction this quantified,The deteriorating effects of diffraction on the resolution are quantified in this manner.,29334
different three-dimensional Two by are recorded modes. focused scenes imaging real/virtual and,Two different three-dimensional scenes are recorded by real/virtual and focused imaging modes.,29335
the datasets. benchmark Caltech PASCAL and on are conducted Extensive the VOC experiments,Extensive experiments are conducted on the PASCAL VOC and the Caltech benchmark datasets.,29336
analyzed. each effectiveness of also The component novel is,The effectiveness of each novel component is also analyzed.,29337
single-neuron in of been studied computational neuroscience. dynamics have Mechanistic extensively models,Mechanistic models of single-neuron dynamics have been extensively studied in computational neuroscience.,29338
"reproduce models empirically measured has quantitatively data identifying challenging. can which been However,","However, identifying which models can quantitatively reproduce empirically measured data has been challenging.",29339
"distributions our On efficiently posterior and synthetic approach recovers data, ground-truth estimates parameters.","On synthetic data, our approach efficiently estimates posterior distributions and recovers ground-truth parameters.",29340
in of is topics one key mind-brain-environment the relations psychology. Understanding,Understanding mind-brain-environment relations is one of the key topics in psychology.,29341
"spectral to simple problem reduces, conventional thanks a The optimization relaxation, to multidimensional this search.","The conventional multidimensional optimization problem reduces, thanks to this relaxation, to a simple spectral search.",29342
Our available two achieves datasets. superior on method performance publicly,Our method achieves superior performance on two publicly available datasets.,29343
"for extracting We sequential approach dynamics from multiple, low-dimensional propose recordings. an","We propose an approach for extracting low-dimensional dynamics from multiple, sequential recordings.",29344
"we spatial-temporal For a attention adaptive aggregation, propose novel mechanism.","For adaptive aggregation, we propose a novel spatial-temporal attention mechanism.",29345
substantially. have proposed clearly that the related PersonRank outperforms and methods We demonstrated,We have demonstrated that the proposed PersonRank outperforms related methods clearly and substantially.,29346
learn cognition? How be populations used to can neuronal about uniquely,How can neuronal populations be uniquely used to learn about cognition?,29347
graph. time We a evolves of consider that path-signal a problem over on a localizing,We consider a problem of localizing a path-signal that evolves over time on a graph.,29348
"are to these classifier. and presented Then, features concatenated a latent","Then, these latent features are concatenated and presented to a classifier.",29349
"outperforms Furthermore, method state-of-the-art existing methods. the","Furthermore, the method outperforms existing state-of-the-art methods.",29350
"considers a fixed lifetime the Forest amend parameter. we that Mondrian First, original algorithm,","First, we amend the original Mondrian Forest algorithm, that considers a fixed lifetime parameter.",29351
"provide analysis a establishing for we consistency. conditions Second, simple theoretical","Second, we provide a theoretical analysis establishing simple conditions for consistency.",29352
from used learned whitening PCA commonly descriptor discriminatively the training outperforms whitening. data same CNN,CNN descriptor whitening discriminatively learned from the same training data outperforms commonly used PCA whitening.,29353
various Convolutional years. tasks accuracy networks recent on neural in demonstrated have high,Convolutional neural networks have demonstrated high accuracy on various tasks in recent years.,29354
"examples. they However, are extremely adversarial to vulnerable","However, they are extremely vulnerable to adversarial examples.",29355
"cause perturbations to For added clean can fail. neural images example, to imperceptible networks convolutional","For example, imperceptible perturbations added to clean images can cause convolutional neural networks to fail.",29356
"effects. to paper, mitigate to utilize at inference time randomization adversarial this In we propose","In this paper, we propose to utilize randomization at inference time to mitigate adversarial effects.",29357
discussed. Implications future directions and are research,Implications and future research directions are discussed.,29358
access user. reach per optical is demand It networks predicted that multi-Gb/s in will,It is predicted that demand in optical access networks will reach multi-Gb/s per user.,29359
challenging. robust always recognizers Arabic for has been Building,Building robust recognizers for Arabic has always been challenging.,29360
datasets text video two - ACTIV. available publicly previous outperform ALIF We and state-of-the-art on,We outperform previous state-of-the-art on two publicly available video text datasets - ALIF and ACTIV.,29361
"model to sequence follows sequence segmentation-free, The a transcription approach.","The model follows a segmentation-free, sequence to sequence transcription approach.",29362
"of dependencies results. to the model yields recognition contextual RNNs superior Further, ability","Further, the ability of RNNs to model contextual dependencies yields superior recognition results.",29363
the output for also The the LSTM transformer. spatial computing parameters,The LSTM also output the parameters for computing the spatial transformer.,29364
grouped to these shape to segments. step identity The assign next contour a is,The next step is to assign a shape identity to these grouped contour segments.,29365
algorithm image datasets. experimentation comprehensive two on crystal a our We provide of,We provide a comprehensive experimentation of our algorithm on two crystal image datasets.,29366
"against algorithm proposed baselines. test the outperforming both two our baselines, We algorithm with our","We test our algorithm against two baselines, with our proposed algorithm outperforming both the baselines.",29367
"account compression paper, for process. in propose this explicitly the training In to we","In this paper, we propose to explicitly account for compression in the training process.",29368
building study of disentangle problem variation. the of independent factors We models that,We study the problem of building models that disentangle independent factors of variation.,29369
As set. weakly we labeled a data use training,As data we use a weakly labeled training set.,29370
it without interest is as labeling readily costs. available particular annotation This of may be,This labeling is of particular interest as it may be readily available without annotation costs.,29371
reference ambiguity. this We call issue the,We call this issue the reference ambiguity.,29372
"show and Moreover, role training. the dimensionality feature of the adversarial we","Moreover, we show the role of the feature dimensionality and adversarial training.",29373
EEG (Electroencephalogram)-based image system. propose We an annotation,We propose an EEG (Electroencephalogram)-based image annotation system.,29374
"adaptation, addition, i.e. has approach the of high speed In a","In addition, the approach has a high speed of adaptation, i.e.",29375
input generalization mapping This the outside support allows of the the parameterization of measure.,This parameterization allows generalization of the mapping outside the support of the input measure.,29376
domain approach two adaptation generative on proposed our We applications: showcase and modeling.,We showcase our proposed approach on two applications: domain adaptation and generative modeling.,29377
Quantitative has (QSAR) Structure-Activity in medicinal invaluable tool chemistry. an Relationship proved,Quantitative Structure-Activity Relationship (QSAR) has proved an invaluable tool in medicinal chemistry.,29378
We on Kramers transceiver detection. schemes based Kronig (KK) propose two,We propose two transceiver schemes based on Kramers Kronig (KK) detection.,29379
three The of parts. mainly TFNet consists,The TFNet mainly consists of three parts.,29380
skeletons. a network estimated curve refine propose prediction Then to dense the we,Then we propose a dense prediction network to refine the estimated curve skeletons.,29381
"the of synchronization rest This, clearly, allows the simple with network.","This, clearly, allows simple synchronization with the rest of the network.",29382
accepted. We paper publicly will is available this dataset make after the,We will make the dataset publicly available after this paper is accepted.,29383
topic gesture long Hand hot interaction. computer in recognition been has a human,Hand gesture recognition has long been a hot topic in human computer interaction.,29384
camera-based gesture systems Traditional properly hand cannot circumstances. under work dark recognition,Traditional camera-based hand gesture recognition systems cannot work properly under dark circumstances.,29385
hand received are time-frequency gesture with signals analysis. processed then The,The received hand gesture signals are then processed with time-frequency analysis.,29386
to gestures. are used neural different classify Convolutional networks,Convolutional neural networks are used to classify different gestures.,29387
"related distance and scale such are as gesture Besides, recognition investigated. factors","Besides, related factors such as recognition distance and gesture scale are investigated.",29388
"driver isolators, digital proposed of complementary gate and MOSFETs, consists Si Our GaN HEMTs.","Our proposed gate driver consists of digital isolators, complementary Si MOSFETs, and GaN HEMTs.",29389
network The dealing with neural is sequences. recurrent appropriate (RNN) temporal for,The recurrent neural network (RNN) is appropriate for dealing with temporal sequences.,29390
"in with the existing RNN are the system. Compared innovations proposed models, involved three","Compared with the existing RNN models, three innovations are involved in the proposed system.",29391
"for hidden RNN proposed information layer a for temporal new learning better. is function First,","First, a new hidden layer function for RNN is proposed for learning temporal information better.",29392
Pool it we Memory call (MPU). Unit,we call it Memory Pool Unit (MPU).,29393
architecture. simple a The has proposed MPU,The proposed MPU has a simple architecture.,29394
the parameter dimension. calculating at iteration hybrid-parameter has changes proposed temporal when RNN The,The proposed hybrid-parameter RNN has parameter changes when calculating the iteration at temporal dimension.,29395
for the states states increasing layer capacity. all hidden layer Stacked hidden combine expression the,Stacked hidden layer states combine all the hidden layer states for increasing the expression capacity.,29396
unobtrusive rate the (PPG) (HR). Wrist photoplethysmography allows heart of monitoring,Wrist photoplethysmography (PPG) allows unobtrusive monitoring of the heart rate (HR).,29397
conditional from simulations. cash transfer illustrated using with data These findings program a and are,These findings are illustrated using data from a conditional cash transfer program and with simulations.,29398
"significant problems, monitoring quality has recently increasing air pollution index received (AQI) attention. air Given","Given significant air pollution problems, air quality index (AQI) monitoring has recently received increasing attention.",29399
"puts quality. Thus, solution termination the with condition odds at learning efficiency","Thus, the termination condition puts learning efficiency at odds with solution quality.",29400
"its up holds show validate claims. that empirically, and We motivating our to it algorithm","We validate our algorithm empirically, and show that it holds up to its motivating claims.",29401
Bayesian optimization GPflowOpt framework known novel for as Python introduced. A is,A novel Python framework for Bayesian optimization known as GPflowOpt is introduced.,29402
"and is framework well and documented, provides The scalability. thoroughly tested","The framework is thoroughly tested and well documented, and provides scalability.",29403
"easy of use permits strategies custom it in implemented modeling GPflow. Finally,","Finally, it permits easy use of custom modeling strategies implemented in GPflow.",29404
"We the applications. realisations neural-symbolic then and to systems, of proceed computation, describe","We then proceed to describe the realisations of neural-symbolic computation, systems, and applications.",29405
facing research. Finally and avenues for further challenges the the area present we,Finally we present the challenges facing the area and avenues for further research.,29406
"paper, we representation for introduce In framework. depth estimation surface unsupervised normal this a","In this paper, we introduce a surface normal representation for unsupervised depth estimation framework.",29407
system real-world that ignores person detected coming from the sequentially. scenario This are images multi-cameras,This ignores the real-world scenario that person images detected from multi-cameras system are coming sequentially.,29408
data the information. within-class sketch matrix) from covariance,within-class covariance matrix) from the sketch data information.,29409
theoretical covariance approximate of effect analysis on the provide estimated matrix. within-class We the,We provide theoretical analysis on the effect of the estimated approximate within-class covariance matrix.,29410
"In the derive (i.e. bounds Fisher particular, discriminant on we upper lower score and","In particular, we derive upper and lower bounds on the Fisher discriminant score (i.e.",29411
"user typically be the to location they identify on. manually the focused to require However,","However, they typically require the user to manually identify the location to be focused on.",29412
"as critical, generally problems of often feature Roguelike element games repetitive gameplay. a exploration yet","Roguelike games generally feature exploration problems as a critical, yet often repetitive element of gameplay.",29413
dungeon to paper of roguelike approach This presents an environments. algorithmic exploration,This paper presents an algorithmic approach to exploration of roguelike dungeon environments.,29414
for the a through optimized further parameterization comprehensive data investigate algorithm We analysis.,We further investigate optimized parameterization for the algorithm through a comprehensive data analysis.,29415
"Domain environments. for critical in adaptation success new, is unseen","Domain adaptation is critical for success in new, unseen environments.",29416
Adversarial model. novel discriminatively-trained We Adaptation Domain Cycle-Consistent propose a,We propose a novel discriminatively-trained Cycle-Consistent Adversarial Domain Adaptation model.,29417
model of visual can be in prediction a recognition settings. and applied Our variety,Our model can be applied in a variety of visual recognition and prediction settings.,29418
"and bimanual detection. suture anastomosis, which is task such automated requires needle handling One","One such task is automated anastomosis, which requires bimanual needle handling and suture detection.",29419
demonstrate two Experiments different on types accuracy sutures proposed the framework. the of of,Experiments on two different types of sutures demonstrate the accuracy of the proposed framework.,29420
scene for proposed this in segmentation method adaptation urban domain A is work.,A domain adaptation method for urban scene segmentation is proposed in this work.,29421
"study, sampling. In utilization we of trans-dimensional this beyond RJMCMC propose","In this study, we propose utilization of RJMCMC beyond trans-dimensional sampling.",29422
"from consistency location, cues: scores four contrast, continuity. derived are appearance These and egocentric position","These scores are derived from four egocentric cues: contrast, location, position consistency and appearance continuity.",29423
a high-resolution in conducted sensing proposed of dataset. algorithms remote were the Evaluation,Evaluation of the proposed algorithms were conducted in a high-resolution remote sensing dataset.,29424
corn in more the the than in is live systematic Price discovery market. cattle,Price discovery is more systematic in the corn than in the live cattle market.,29425
large-scale be antenna arrays. addressed The pilot contamination with cannot problem,The pilot contamination problem cannot be addressed with large-scale antenna arrays.,29426
advantage of diversity by The mismatch. intentional sampling scheme takes inducing timing,The scheme takes advantage of sampling diversity by inducing intentional timing mismatch.,29427
"square derive ZF estimator. the (MSE) bound we error upper minimum the of Moreover,","Moreover, we derive the minimum square error (MSE) upper bound of the ZF estimator.",29428
"systems, for ability The future intelligent e.g. the predicting is important of","The ability of predicting the future is important for intelligent systems, e.g.",29429
robots vehicles early to and accordingly. plan and autonomous decisions make,autonomous vehicles and robots to plan early and make decisions accordingly.,29430
"dynamics and illicit through from treatment. Key addiction addiction prescription, sources, considered include","Key dynamics considered include addiction through prescription, addiction from illicit sources, and treatment.",29431
this presented signature is method for paper. offline verification A in,A method for offline signature verification is presented in this paper.,29432
entertainment and in experience learning experience Scrabble. This the paper explores,This paper explores the entertainment experience and learning experience in Scrabble.,29433
"network, we is a Furthermore, show Frangi-Net as neural trainable. that,","Furthermore, we show that, as a neural network, Frangi-Net is trainable.",29434
important is cortical from understanding in in networks. fMRI Measuring connectivity processing functional,Measuring functional connectivity from fMRI is important in understanding processing in cortical networks.,29435
that introduce a method We derivative new estimating for here functional connectivity. uses,We introduce here a new method that uses derivative for estimating functional connectivity.,29436
"we with simulations, commonly Using method benchmarked used other our methods.","Using simulations, we benchmarked our method with other commonly used methods.",29437
Our network achieves better method complex in simulations. results,Our method achieves better results in complex network simulations.,29438
new provides estimate connectivity. functional This method way alternative an to,This new method provides an alternative way to estimate functional connectivity.,29439
"video the fast-forward need videos. Finally, particular technique for a of egocentric stabilization we discuss","Finally, we discuss the need of a particular video stabilization technique for fast-forward egocentric videos.",29440
the achieve better pre-processing technique A performances to of proposed algorithm. is also,A pre-processing technique is also proposed to achieve better performances of the algorithm.,29441
"surveillance. search intelligent classification and as video important highly wide Video is such applications, with","Video classification is highly important with wide applications, such as video search and intelligent surveillance.",29442
and repertoires their by functional incoming areas' connections. structural shaped Brain are outgoing,Brain areas' functional repertoires are shaped by their incoming and outgoing structural connections.,29443
"empirically most energetic connections measured spatial and networks, short, In are constraints. reflecting","In empirically measured networks, most connections are short, reflecting spatial and energetic constraints.",29444
"a that However, unlikely. this hypothesis of long-distance contend non-specificity is connections we implies","However, this hypothesis implies a non-specificity of long-distance connections that we contend is unlikely.",29445
a high-resolution techniques two-dimensional image radar target. a to SAR/ISAR are generate of,SAR/ISAR are radar techniques to generate a two-dimensional high-resolution image of a target.,29446
of the The of utility these on practical procedures schemes. coupling development good is dependent,The practical utility of these procedures is dependent on the development of good coupling schemes.,29447
"an and features capture from propose models, We spatial which videos. audio of several ensemble","We propose an ensemble of several models, which capture spatial and audio features from videos.",29448
"by convolutional datasets. recognition pretrained Spatial networks, on neural captured features are face large","Spatial features are captured by convolutional neural networks, pretrained on large face recognition datasets.",29449
"imaging image to of degradation However, due polarimetry is a noise. challenge major","However, a major challenge of imaging polarimetry is image degradation due to noise.",29450
modular Bayesian scientific Python. is a highly for ABCpy written library in (ABC) Approximate Computation,ABCpy is a highly modular scientific library for Approximate Bayesian Computation (ABC) written in Python.,29451
ABCpy. These modularity from benefits come of the mainly,These benefits come mainly from the modularity of ABCpy.,29452
algorithms. the in some the of us ABC inherent imbalance towards This points,This points us towards the inherent imbalance in some of the ABC algorithms.,29453
common distortion source is intensity images. of Spatially-varying in noise medical a,Spatially-varying intensity noise is a common source of distortion in medical images.,29454
"feature-maps. use algorithm hierarchical the register simple to Then, multi-resolution a we average","Then, we use a simple hierarchical multi-resolution algorithm to register the average feature-maps.",29455
"for in account this existed fails the the to variations However, data. methodology","However, this methodology fails to account for the variations existed in the data.",29456
outperforms Extensive LCCF demonstrate state-of-the-art methods. the that experiments,Extensive experiments demonstrate that LCCF outperforms the state-of-the-art methods.,29457
source code will be publicly The available.,The source code will be publicly available.,29458
in Simulation transient stability effectiveness results its demonstrate assessment.,Simulation results demonstrate its effectiveness in transient stability assessment.,29459
a implemented is stored zipped R (codeSubmit.zip). examples in the of code archive,R code of the implemented examples is stored in a zipped archive (codeSubmit.zip).,29460
and documented. of monitor well The application climate UAV geography population the is to,The application of UAV to monitor the population and climate geography is well documented.,29461
to iterative piecewise avoid in CRF inference. computationally Effective applied is training order the expensive,Effective piecewise training is applied in order to avoid the computationally expensive iterative CRF inference.,29462
"masks. we the that deconvolution Furthermore, deep a introduce improves segmentation network","Furthermore, we introduce a deep deconvolution network that improves the segmentation masks.",29463
under scenario. and PTZ for are the that Results required both confirm tracking speed robustness,Results confirm that both speed and robustness are required for tracking under the PTZ scenario.,29464
averaging by Then results. obtained patch-level the the output is,Then the output is obtained by averaging the patch-level results.,29465
where cases few mainly also to label. patches contribute the image It fails in,It also fails in cases where few patches mainly contribute to the image label.,29466
"combat To of develop representations. we the hyper-image scenarios, these notion","To combat these scenarios, we develop the notion of hyper-image representations.",29467
patches. is trained stage first on The,The first stage is trained on patches.,29468
gender which hand features human information. reveal can distinctive The possesses,The human hand possesses distinctive features which can reveal gender information.,29469
"on CNN we this, design gender a the recognition tackle two-stream Based to problem.","Based on this, we design a two-stream CNN to tackle the gender recognition problem.",29470
"example, own and replicate. For their proteins viruses cells, they express where infect","For example, viruses infect cells, where they express their own proteins and replicate.",29471
overcome the is method to types proposed shortcoming classification previous This of methods. in,This method is proposed to overcome the shortcoming of classification types in previous methods.,29472
concepts combine to this parts. visual In several semantic detect model,In this model several visual concepts combine to detect semantic parts.,29473
shown tracking in based have favorable Convolutional (CNN) neural performance recent approaches networks benchmarks.,Convolutional neural networks (CNN) based tracking approaches have shown favorable performance in recent benchmarks.,29474
"The results maintaining tracking proposed superior real-time approach while performance, in speed.","The proposed approach results in superior tracking performance, while maintaining real-time speed.",29475
"method the solve such complex steps. into To several is problem, divided presented a","To solve such a complex problem, the presented method is divided into several steps.",29476
user's full-body data are The recorded and locomotion the simultaneously. IMU's,The user's full-body locomotion and the IMU's data are recorded simultaneously.,29477
"data Then, be way handled the more such a is efficiently. would in that preprocessed","Then, the data is preprocessed in such a way that would be handled more efficiently.",29478
"where (higher level). the method the the predicts motion belongs Firstly, input phase hierarchical","Firstly, the method predicts the phase where the input motion belongs (higher hierarchical level).",29479
with studied. of direction arrivals help of of estimation is $TV$-minimization The,The estimation of direction of arrivals with help of $TV$-minimization is studied.,29480
guarantee for the a discuss concrete antenna of few the impact We recovery designs.,We discuss the impact of the recovery guarantee for a few concrete antenna designs.,29481
"Additionally, part the are of numerical the findings theoretical performed. supporting simulations","Additionally, numerical simulations supporting the findings of the theoretical part are performed.",29482
to generally a noise decreases sensory Adding performance. signal human,Adding noise to a sensory signal generally decreases human performance.,29483
"improve a performance However to resonance can process too, stochastic due noise called (SR).","However noise can improve performance too, due to a process called stochastic resonance (SR).",29484
"Therefore, psychophysical provides a explanation efficiency reduced rate for lapse EN in paradigms.","Therefore, lapse rate provides a psychophysical explanation for reduced efficiency in EN paradigms.",29485
solve propose We this Message problem. Passing (MP-SVGD) SVGD to,We propose Message Passing SVGD (MP-SVGD) to solve this problem.,29486
"descriptors deeply nor the matrix learned. the kernel is neither Nevertheless,","Nevertheless, neither the descriptors nor the kernel matrix is deeply learned.",29487
"Worse, pursuit an SPD the representation. they of are hindering optimal separately, considered","Worse, they are considered separately, hindering the pursuit of an optimal SPD representation.",29488
of the estimator this the proposed then Weibull and cases. is An for improvement Pareto,An improvement of this estimator is then proposed for the Pareto and the Weibull cases.,29489
"numerical these experiments. the is via compared various efficiency estimators of Finally,","Finally, the efficiency of these estimators is compared via various numerical experiments.",29490
is irregular because often This texts in are (e.g. scene,This is because scene texts are often in irregular (e.g.,29491
be network trained only can annotations. images end-to-end word-level The and by whole using,The whole network can be trained end-to-end by using only images and word-level annotations.,29492
infeasible encountered. cy-cles often Thermodynamically ranges causing are flux metabolic (TICs) unbounded,Thermodynamically infeasible cy-cles (TICs) causing unbounded metabolic flux ranges are often encountered.,29493
violate and satisfy of balance directionality mass law second constraints TICs the but the thermodynamics.,TICs satisfy the mass balance and directionality constraints but violate the second law of thermodynamics.,29494
"Next, SR we the making efficiency the how improve explore by to network slimmer.","Next, we explore how to improve the SR efficiency by making the network slimmer.",29495
"the Two trimming, and trimming one-shot cascade are proposed. methodologies, the","Two methodologies, the one-shot trimming and the cascade trimming, are proposed.",29496
"of applications Such the are task. explanations of small, human-understandable interest potential particular for","Such small, human-understandable explanations are of particular interest for potential applications of the task.",29497
a Leap gesture proposes recognition with system Motion paper Controller. This novel hand,This paper proposes a novel hand gesture recognition system with Leap Motion Controller.,29498
Our is more model show much that results accurate work. our than previous,Our results show that our model is much more accurate than previous work.,29499
challenging computer problem is in text Scene a vision. detection,Scene text detection is a challenging problem in computer vision.,29500
"framework problem the be perspective. Markov from can MAB game solved the Also, also","Also, the MAB problem can also be solved from the Markov game framework perspective.",29501
"(TS) the also performance. to Thompson evaluate proposed is as approach Sampling used benchmark Meanwhile,","Meanwhile, Thompson Sampling (TS) is also used as benchmark to evaluate the proposed approach performance.",29502
Centrality attributes. are as indices visualized node,Centrality indices are visualized as node attributes.,29503
methods. three positioning markers nine studies experiment Our and visual,Our experiment studies nine visual markers and three positioning methods.,29504
"deep (DNNs). method learning article, this neural networks propose In for we transfer a","In this article, we propose a transfer learning method for deep neural networks (DNNs).",29505
many applications. used has widely Deep been learning in,Deep learning has been widely used in many applications.,29506
the transfer methods for solving of is DNNs. learning problem conventional One for this,One of the conventional methods for solving this problem is transfer learning for DNNs.,29507
for that proposed The methods conventional method DNNs. learning our transfer outperforms suggest results,The results suggest that our proposed method outperforms conventional transfer learning methods for DNNs.,29508
of estimate automatically in The task images. crowd crowd counting to is number the pedestrian,The task of crowd counting is to automatically estimate the pedestrian number in crowd images.,29509
fields scales. different corresponding receptive (heads) to have columns Multiple of pedestrians different,Multiple columns have different receptive fields corresponding to pedestrians (heads) of different scales.,29510
people The the computed number of density is integrating map. by,The number of people is computed by integrating the density map.,29511
of results The improvements state-of-the-art. SaCNN over significant the demonstrate,The results demonstrate significant improvements of SaCNN over the state-of-the-art.,29512
results promising of the translation. report qualitative We for face-to-cartoon task,We report promising qualitative results for the task of face-to-cartoon translation.,29513
be identity person. changed can without attribute The of the the changing glasses of wearing,The attribute of wearing glasses can be changed without changing the identity of the person.,29514
refer specific to image manipulation. synthesis as attribute We process this,We refer to this specific synthesis process as image attribute manipulation.,29515
"has practice, and multi-fold In it implications. profound","In practice, it has profound and multi-fold implications.",29516
exists tree Deciding projection a whether NP-hard. is,Deciding whether a tree projection exists is NP-hard.,29517
the results also K the for of returning Tractability established solutions. best are problem,Tractability results are also established for the problem of returning the best K solutions.,29518
"are proposed such optimization and analyzed. algorithms problems parallel for Finally,","Finally, parallel algorithms for such optimization problems are proposed and analyzed.",29519
the issues Scalability real-world realization. major Vehicle-to-Vehicle of one is for network,Scalability is one of the major issues for real-world Vehicle-to-Vehicle network realization.,29520
to time-series. driver/vehicle jointly framework through forecasting dynamical is behavior vehicle able model the This,This framework is able to jointly model driver/vehicle behavior through forecasting the vehicle dynamical time-series.,29521
operator widely RoIPooling to propose operator used AOGParsing the an R-CNN. in substitute We,We propose an AOGParsing operator to substitute the RoIPooling operator widely used in R-CNN.,29522
a convolutional train method folding-unfolding AOG the We propose and to networks end-to-end.,We propose a folding-unfolding method to train the AOG and convolutional networks end-to-end.,29523
structures promising hurting We the that the latent can method without performance. unfold show,We show that the method can unfold promising latent structures without hurting the performance.,29524
image This classification. paper training real-time document for and approach presents an testing for,This paper presents an approach for real-time training and testing for document image classification.,29525
"(time-)efficient and production environments, accurate it to In crucial is perform training.","In production environments, it is crucial to perform accurate and (time-)efficient training.",29526
"Motivated Vision, approach. Computer from a two-stage propose we","Motivated from Computer Vision, we propose a two-stage approach.",29527
"for this As suitable makes real-time document learning-based applications. approach classification large-scale novel such, deep","As such, this novel approach makes deep learning-based document classification suitable for large-scale real-time applications.",29528
show more estimation We channel algorithm based to DCD also that robust the is errors.,We also show that the DCD based algorithm is more robust to channel estimation errors.,29529
"on method Flying shows and datasets. Chairs, MPI-Sintel results KITTI benchmark Our promising","Our method shows promising results on Flying Chairs, MPI-Sintel and KITTI benchmark datasets.",29530
reality. This MRARC gives capability complex handling the in noises various rise to of in,This gives rise to the capability of MRARC in handling various complex noises in reality.,29531
validate for face experiments MRARC data efficacy the recognition. real-world of on The robust,The experiments on real-world data validate the efficacy of MRARC for robust face recognition.,29532
"end, system an additional Diagnosis Sequential oracle measurements. approaches To this for ask","To this end, Sequential Diagnosis approaches ask an oracle for additional system measurements.",29533
diagnosis. work model-based (optimal) in presents strategies selection measurement for This sequential,This work presents strategies for (optimal) measurement selection in model-based sequential diagnosis.,29534
is deficits. persistent dyslexia and by characterized spelling Developmental reading,Developmental dyslexia is characterized by persistent reading and spelling deficits.,29535
"schemes In exist. we universally consistent prove vertex no nomination addition, that","In addition, we prove that no universally consistent vertex nomination schemes exist.",29536
are graphs intelligence many artificial useful for Knowledge (AI) tasks.,Knowledge graphs are useful for many artificial intelligence (AI) tasks.,29537
"knowledge missing facts. graphs However, have often","However, knowledge graphs often have missing facts.",29538
"graph knowledge the been graphs, To have populate embedding developed. models","To populate the graphs, knowledge graph embedding models have been developed.",29539
entity employs embeddings that principle their relations. represent It the between differences the,It employs the principle that the differences between entity embeddings represent their relations.,29540
"However, regularization. TransE has with problem a its","However, TransE has a problem with its regularization.",29541
sphere the in TransE be a embeddings embedding forces to space. vector entity on,TransE forces entity embeddings to be on a sphere in the embedding vector space.,29542
accuracies adversely regularization the the affects also of link The predictions.,The regularization also affects adversely the accuracies of the link predictions.,29543
"model, novel TorusE, a proposes regularization embedding solve This to paper problem. the","This paper proposes a novel embedding model, TorusE, to solve the regularization problem.",29544
can of The be defined group. principle on any TransE Lie,The principle of TransE can be defined on any Lie group.,29545
a literature. Detecting been in that mutex addressed planning the problem is frequently has pairs,Detecting mutex pairs is a problem that has been addressed frequently in the planning literature.,29546
deep foundation performance Neural architectures networks for (DNNs). the are of neural improving,Neural architectures are the foundation for improving performance of deep neural networks (DNNs).,29547
unified framework DNNs. provides in developed a the state-of-the-art It practices for best,It provides a unified framework for the best practices developed in state-of-the-art DNNs.,29548
obtains best model using dissection. the AOGNet also score interpretability network,AOGNet also obtains the best model interpretability score using network dissection.,29549
AOGNet shows adversarial potential better defense. in further,AOGNet further shows better potential in adversarial defense.,29550
"ResNet the performance in ResNeXt R-CNN. backbones MS-COCO, than and better Mask AOGNet In obtains","In MS-COCO, AOGNet obtains better performance than the ResNet and ResNeXt backbones in Mask R-CNN.",29551
and stack is One continuous on storage data. the of a of probabilistic interpretation action,One interpretation of a continuous stack is the probabilistic storage of and action on data.,29552
"fully with from machine CNNs. current representations However, not benefit skeleton learning do","However, current skeleton representations do not fully benefit from machine learning with CNNs.",29553
are a limited to single Existing hand. typically datasets,Existing datasets are typically limited to a single hand.,29554
made datasets will and Source be available. publicly,Source and datasets will be made publicly available.,29555
"this pattern determines then speed process The formation i.e. of velocity,","The speed of this process determines then pattern formation velocity, i.e.",29556
an a developed from noisy initial pattern. situation final the time elapsed to,the elapsed time from an initial noisy situation to a final developed pattern.,29557
extends to value-at-risk(CVaR) the MDP algorithm also objective. conditional with a The problem,The algorithm also extends to the MDP problem with a conditional value-at-risk(CVaR) objective.,29558
Encoder-Decoder The style. based is architecture on,The architecture is based on Encoder-Decoder style.,29559
two knowledge. to causal There types contrasting are discovering of approaches,There are two contrasting types of approaches to discovering causal knowledge.,29560
"world open However, a remains recognition face challenge. still","However, open world face recognition still remains a challenge.",29561
"problem. our the open recognition face efficacy Finally, for of method the we world show","Finally, we show the efficacy of our method for the open world face recognition problem.",29562
training for We whole-image algorithm develop mammograms. on an based breast cancer end-to-end diagnosis,We develop an end-to-end training algorithm for whole-image breast cancer diagnosis based on mammograms.,29563
the annotations first at stage of only It lesion training. requires,It requires lesion annotations only at the first stage of training.,29564
"level can labels. classifier using be image After only trained a whole image that,","After that, a whole image classifier can be trained using only image level labels.",29565
reliance the greatly on This lesion annotations. reduced,This greatly reduced the reliance on lesion annotations.,29566
"our with DDSM models compare Using current favorably the benchmark, as state-of-the-art.","Using DDSM as benchmark, our models compare favorably with the current state-of-the-art.",29567
Disentangling representation a has variation very factors become of challenging problem on learning.,Disentangling factors of variation has become a very challenging problem on representation learning.,29568
datasets. rigorously descriptor SDASS the four is tested of popular The performance on,The performance of the SDASS descriptor is rigorously tested on four popular datasets.,29569
accurate The our effectiveness result confirms method. the further SDASS of,The accurate result further confirms the effectiveness of our SDASS method.,29570
learn We develop a to good to find wide-baseline deep for stereo. correspondences architecture,We develop a deep architecture to learn to find good correspondences for wide-baseline stereo.,29571
the defend perturbations. the first dedicated to effectively against present We networks framework such,We present the first dedicated framework to effectively defend the networks against such perturbations.,29572
through detector. A image passed and the by query first is verified PRN the,A query image is first passed through the PRN and verified by the detector.,29573
literature input is a probe only data on limited vehicles. using from methods There,There is only a limited literature on methods using data input from probe vehicles.,29574
to learn parameters. We propose model an efficient algorithm,We propose an efficient algorithm to learn model parameters.,29575
We algorithm arbitrary times travel of predicting an durations. trips for on propose,We propose an algorithm for predicting travel times on trips of arbitrary durations.,29576
from images pre-processing step. a extraction Brain common is,Brain extraction from images is a common pre-processing step.,29577
all four better competing methods on model performs these than Our datasets.,Our model performs better than these competing methods on all four datasets.,29578
spaces It parameter can explore guided high-dimensional flows. by efficiently Hamiltonian simulated,It can efficiently explore high-dimensional parameter spaces guided by simulated Hamiltonian flows.,29579
easy supervised train. method is a objective and which Our to uses fast training,Our method uses a supervised training objective which is fast and easy to train.,29580
covariance field The statistics. the kernels ubiquitous spatial of in is of use,The use of covariance kernels is ubiquitous in the field of spatial statistics.,29581
detection thesis facilitate suggests in relations capacity. underlying contextual would The stronger that greater improvements,The underlying thesis suggests that stronger contextual relations would facilitate greater improvements in detection capacity.,29582
"video this introduced. efficient yet paper, first-person is a recognition activity method for In simple","In this paper, a simple yet efficient activity recognition method for first-person video is introduced.",29583
relations between scene can motions. them The and represent the dynamics local,The relations between them can represent the scene dynamics and local motions.,29584
reduce to introduced grouping cost. computational strategy The helps the considerably,The introduced grouping strategy helps to considerably reduce the computational cost.,29585
in which classified. results linearly be highly features representation method discriminative The can,The representation method results in highly discriminative features which can be linearly classified.,29586
subject the classification of many has in of studies Spectral-spatial images recent years. been hyperspectral,Spectral-spatial classification of hyperspectral images has been the subject of many studies in recent years.,29587
"In challenging. pixels, of few becomes only very presence this the task labeled","In the presence of only very few labeled pixels, this task becomes challenging.",29588
On our outperforms method images considered significantly baselines. these,On these images our method significantly outperforms considered baselines.,29589
references an understand of to use better meaning to We the how symbolic show ad.,We show how to use symbolic references to better understand the meaning of an ad.,29590
applications. attributes with wide range identity-preserving Generating has a faces high fidelity facial of different,Generating high fidelity identity-preserving faces with different facial attributes has a wide range of applications.,29591
"make two further extensions based Then, on our method. AEZSL we","Then, we further make two extensions based on our AEZSL method.",29592
be strengths either and divalent can predicted metals then of The binding systematically. trivalent,The binding strengths of either divalent and trivalent metals can then be predicted systematically.,29593
training relationship difficult Obtaining expensive. categories sufficient is and samples possible for all,Obtaining sufficient training samples for all possible relationship categories is difficult and expensive.,29594
"tackle this propose to In a problem. this guided work, language we framework natural","In this work, we propose a natural language guided framework to tackle this problem.",29595
LBP of LBP improved and cases. in Deep multiscale all the performance traditional,Deep LBP improved the performance of traditional and multiscale LBP in all cases.,29596
Generative al. [Goodfellow Networks Adversarial et (GANs),Generative Adversarial Networks (GANs) [Goodfellow et al.,29597
et experiments We conduct extensive on CelebA al. [Liu,We conduct extensive experiments on CelebA [Liu et al.,29598
An person challenge annotation (re-ID) the is of difficulty. re-identification intrinsic,An intrinsic challenge of person re-identification (re-ID) is the annotation difficulty.,29599
"explainable. the making process recurrent, the mechanisms are Furthermore, and visualizable attention referring","Furthermore, the attention mechanisms are recurrent, making the referring process visualizable and explainable.",29600
dual The combined to information from attended are object. these reason the referred about sources,The attended information from these dual sources are combined to reason about the referred object.,29601
"transformation. in assumption is this no true However, typeface longer","However, this assumption is no longer true in typeface transformation.",29602
hierarchical and sub-networks: discriminator. transfer of a proposed a The HAN adversarial two consists network,The proposed HAN consists of two sub-networks: a transfer network and a hierarchical adversarial discriminator.,29603
another. to maps characters transfer from The network typeface one,The transfer network maps characters from one typeface to another.,29604
ingredient The use computational essential resources of is an intelligence. limited efficient of,The efficient use of limited computational resources is an essential ingredient of intelligence.,29605
"intractable. metareasoning computations to rational according this computationally Selecting but this, would optimally achieve is","Selecting computations optimally according to rational metareasoning would achieve this, but this is computationally intractable.",29606
captioning. and Deep many models state-of-the-art are video recognition tasks for video vision action including,Deep models are state-of-the-art for many vision tasks including video action recognition and video captioning.,29607
"spatiotemporal relatively content - lacking studies videos. models for such are visual of However,","However, such studies are relatively lacking for models of spatiotemporal visual content - videos.",29608
spectra. by and have smooth local filters enforcing are symmetry Resulting localized,Resulting filters have local symmetry and are localized by enforcing smooth spectra.,29609
broad years. drawn Neural has transfer style attention in recent,Neural style transfer has drawn broad attention in recent years.,29610
"For validation, problem. proposed we the Chinese the algorithm applied Typeface to transfer","For validation, we applied the proposed algorithm to the Chinese Typeface transfer problem.",29611
experiment and the generation our on method. robustness have effectiveness character results demonstrated Extensive of,Extensive experiment results on character generation have demonstrated the effectiveness and robustness of our method.,29612
geometric to deep perspective This of learning. a the provides generalizability,This provides a geometric perspective to the generalizability of deep learning.,29613
"step hampered or in be by pace can counting Nonetheless, walking fluctuations accuracy demeanor.","Nonetheless, step counting accuracy can be hampered by fluctuations in walking pace or demeanor.",29614
events. can analysis toe-off steps wrist cycle illustrates proximal devices recurrently to that detect Gait,Gait cycle analysis illustrates that wrist devices can recurrently detect steps proximal to toe-off events.,29615
"and Such solving deeper their capabilities, in knowledge e.g. broader require applications problem","Such applications require deeper and broader knowledge in their problem solving capabilities, e.g.",29616
EKG physiology findings. knowledge integrating and results anatomy patient other with and,integrating anatomy and physiology knowledge with EKG results and other patient findings.,29617
are Target repositories these for smart not human. readers machines;,Target readers for these repositories are smart machines; not human.,29618
"deep, basic, knowledge important, equally is important. and is but Data commonsense","Data is important, but knowledge deep, basic, and commonsense is equally important.",29619
scenes. graph structured the producing We of visual investigate problem representations of,We investigate the problem of producing structured graph representations of visual scenes.,29620
appearing Our substructures work of regularly analyzes in role graphs. motifs: scene the,Our work analyzes the role of motifs: regularly appearing substructures in scene graphs.,29621
new dataset. Genome quantitative such Visual insights in on the repeated structures present We,We present new quantitative insights on such repeated structures in the Visual Genome dataset.,29622
that of are object vice-versa. Our analysis but predictive labels relation labels not highly shows,Our analysis shows that object labels are highly predictive of relation labels but not vice-versa.,29623
demonstrate The conventional public ensemble benchmarks. trackers on scheme experimental results outperforms the proposed that,The experimental results demonstrate that the proposed scheme outperforms conventional ensemble trackers on public benchmarks.,29624
"methods only produce to GPU But memory low-resolution outputs. due can limitations, these","But due to GPU memory limitations, these methods can only produce low-resolution outputs.",29625
We methods state-of-the-art several deep re-evaluate carefully hashing setting. with a designed experimental,We re-evaluate several state-of-the-art deep hashing methods with a carefully designed experimental setting.,29626
"deep the all re-examined. papers conclusions Thus, hashing the in should carefully be","Thus, the conclusions in all the deep hashing papers should be carefully re-examined.",29627
Brazilian sign the specially on focuses paper language (LIBRAS). This popular,This paper specially focuses on the popular Brazilian sign language (LIBRAS).,29628
deep a a idea proposed customized architecture. through is The network implemented having neural,The proposed idea is implemented through a deep neural network having a customized architecture.,29629
theory. of Generating in colorings computational sets building complete Ramsey is important block an,Generating complete sets of colorings is an important building block in computational Ramsey theory.,29630
"piece-wise function. a this we design To end, loss","To this end, we design a piece-wise loss function.",29631
We compute semantic fusing background segmentations. and ground structures corresponding a by representation,We compute a ground representation by fusing background structures and corresponding semantic segmentations.,29632
robust is objects. and handles Our temporarily occlusion stationary to method,Our method is robust to occlusion and handles temporarily stationary objects.,29633
using show drone results We imagery. qualitative,We show qualitative results using drone imagery.,29634
the We dataset. perform presented quantitative this approach a of using evaluation,We perform a quantitative evaluation of the presented approach using this dataset.,29635
determined a for experimental proteins class and measures. representative The of large parameters are using,The parameters for a large and representative class of proteins are determined using experimental measures.,29636
"inpainting. this we conditioned Adversarial a semantically present paper, Generative (GAN) for sequence In Network","In this paper, we present a semantically conditioned Generative Adversarial Network (GAN) for sequence inpainting.",29637
This in inpainting performance. more appealing achieving helps,This helps in achieving more appealing inpainting performance.,29638
"Though for faces. generic, targeted our inpainting algorithm was on","Though generic, our algorithm was targeted for inpainting on faces.",29639
are non-line-of-sight of Active interest systems imaging for applications. growing diverse,Active non-line-of-sight imaging systems are of growing interest for diverse applications.,29640
"must be ultrafast specialized, calibrated. typically carefully expensive, requires lasers and that detectors This","This typically requires expensive, specialized, ultrafast lasers and detectors that must be carefully calibrated.",29641
"state-of-the-art approaches, method. or by achieved to Compared performances can be comparable our improved","Compared to state-of-the-art approaches, comparable or improved performances can be achieved by our method.",29642
"not supervised this free-space. issue, designed most are Although addresses segmentation for methods weakly","Although weakly supervised segmentation addresses this issue, most methods are not designed for free-space.",29643
image Computer Vision in large-scale has leveraging achieved by Significant datasets. been progress,Significant progress has been achieved in Computer Vision by leveraging large-scale image datasets.,29644
"However, for still limited. Computer beyond complex Vision are large-scale classification tasks datasets","However, large-scale datasets for complex Computer Vision tasks beyond classification are still limited.",29645
gap concepts. bridge and between images semantic annotations high-level low-level rich These the,These rich annotations bridge the semantic gap between low-level images and high-level concepts.,29646
proposed is methods. different benchmark effective an dataset improve The computational to evaluate and,The proposed dataset is an effective benchmark to evaluate and improve different computational methods.,29647
benefits of LDMNet in demonstrate experiments. We the two,We demonstrate two benefits of LDMNet in the experiments.,29648
GI new the entire Capsule to relatively trace. Wireless is a technology examine (WCE) Endoscopy,Wireless Capsule Endoscopy (WCE) is relatively a new technology to examine the entire GI trace.,29649
is time-consuming images and human prone Reviewing these error. all to,Reviewing all these images is time-consuming and prone to human error.,29650
physicians assisting has to review to intelligent been a methods frames. challenge the develop It,It has been a challenge to develop intelligent methods assisting physicians to review the frames.,29651
"faster QS. than However, dramatically was SLIC","However, SLIC was dramatically faster than QS.",29652
images as proposes recognition input. facial fusion-based This paper method a which gender uses,This paper proposes a fusion-based gender recognition method which uses facial images as input.,29653
"four recognition are state-of-the-art inspired systems. Thereafter, gender which by frameworks different are proposed","Thereafter, four different frameworks are proposed which are inspired by state-of-the-art gender recognition systems.",29654
"uses second and Vector Machine filters, framework kernel (SVM). PCA, Support Gabor The","The second framework uses Gabor filters, PCA, and kernel Support Vector Machine (SVM).",29655
"the of four decisions voting. fused are frameworks using Finally, weighted","Finally, the four decisions of frameworks are fused using weighted voting.",29656
and of Experimental show results proposed the power the effectiveness method.,Experimental results show the power and effectiveness of the proposed method.,29657
test this difficult for from the We depth-estimation endoscopy. approach notoriously task of,We test this approach for the notoriously difficult task of depth-estimation from endoscopy.,29658
"the end-to-end called trained is method, from scratch. Network (RN), Our Relation","Our method, called the Relation Network (RN), is trained end-to-end from scratch.",29659
"few-shot our providing improved on extended easily to zero-shot framework learning. is performance Besides learning,","Besides providing improved performance on few-shot learning, our framework is easily extended to zero-shot learning.",29660
models random is many assumption in for Stationarity a processes. key statistical,Stationarity is a key assumption in many statistical models for random processes.,29661
"extensive we of classification. visual method, demonstrate experiments our on effectiveness the To conduct","To demonstrate the effectiveness of our method, we conduct extensive experiments on visual classification.",29662
notably state-of-the-art methods. Experiment that our show results outperforms method,Experiment results show that our method notably outperforms state-of-the-art methods.,29663
evaluated restoration measure Image distortion algorithms typically are by (e.g. some,Image restoration algorithms are typically evaluated by some distortion measure (e.g.,29664
"PSNR, perceptual by perceived quality. or opinion that quantify VIF) IFC, SSIM, human scores","PSNR, SSIM, IFC, VIF) or by human opinion scores that quantify perceived perceptual quality.",29665
(GANs) approach bound. the provide way principled We that show perception-distortion a to also generative-adversarial-nets,We also show that generative-adversarial-nets (GANs) provide a principled way to approach the perception-distortion bound.,29666
vision theoretical support in their to This tasks. success constitutes observed low-level,This constitutes theoretical support to their observed success in low-level vision tasks.,29667
representations performance. retrieval crucial data is for Learning the cross-modal appropriate multi-modal for,Learning appropriate representations for multi-modal data is crucial for the cross-modal retrieval performance.,29668
and We an the such of benefits challenges discuss approach.,We discuss the challenges and benefits of such an approach.,29669
majorly on images. viewpoint in input from focus novel interpolation synthesis Researches multi-view,Researches in novel viewpoint synthesis majorly focus on interpolation from multi-view input images.,29670
"we deep achieve learning-based goal, novel To technique. a this propose","To achieve this goal, we propose a novel deep learning-based technique.",29671
a recent Networks become of years. (NN) Compression studied in highly has topic Neural,Compression of Neural Networks (NN) has become a highly studied topic in recent years.,29672
video consecutive synthesise or Frame attempts interpolation more one frames frames. to given,Frame interpolation attempts to synthesise frames given one or more consecutive video frames.,29673
"These and problems, tackle often algorithm two efficiency namely reconstruction techniques quality.","These techniques often tackle two problems, namely algorithm efficiency and reconstruction quality.",29674
"this generative we adversarial (\mbox{FIGAN}). for In interpolation network paper, a present frame multi-scale","In this paper, we present a multi-scale generative adversarial network for frame interpolation (\mbox{FIGAN}).",29675
that achieves the experimental demonstrate network The state-of-the-art our performances. results,The experimental results demonstrate that our network achieves the state-of-the-art performances.,29676
information rate has utmost the systems. ever-growing optical of become communication demands for Meeting importance,Meeting the ever-growing information rate demands has become of utmost importance for optical communication systems.,29677
the of research disruptive been flow Quantitative reports. has gathering subjective into nature by hampered,Quantitative research into flow has been hampered by the disruptive nature of gathering subjective reports.,29678
We neural a encoding symmetry brain a segmentation tumor network brain task. into explore for,We explore encoding brain symmetry into a neural network for a brain tumor segmentation task.,29679
in loss whole to an multi-task the The train way. network enables end-to-end function us,The multi-task loss function enables us to train the whole network in an end-to-end way.,29680
variation proposed. a model total is functions kernel based In on paper this novel,In this paper a novel total variation model based on kernel functions is proposed.,29681
"(e.g. $a^n b^n$, and x^R$), context-sensitive $x","$a^n b^n$, $x x^R$), and context-sensitive (e.g.",29682
"we research combine venues, problem. the color quantization on work, focusing both In this","In this work, we combine both research venues, focusing on the color quantization problem.",29683
"Algorithm, problems. Information-retrieval-related successfully Our Genetic in apparatus soft-computing strategy utilized employs optimization a","Our strategy employs Genetic Algorithm, a soft-computing apparatus successfully utilized in Information-retrieval-related optimization problems.",29684
approach effectiveness tested baselines that on in the all indicate outperformed Results scenarios. focused representation,Results indicate that the approach focused on representation effectiveness outperformed baselines in all tested scenarios.,29685
AMS load joint enables the communication training the process. dramatically reduces thus and arduous,AMS thus dramatically reduces the communication load and enables the arduous joint training process.,29686
prediction are ground-truth results indistinguishable Its nearly to from humans.,Its prediction results are nearly indistinguishable from ground-truth to humans.,29687
rail network. to concept track been index monitor complete has A proposed health also,A track health index concept has also been proposed to monitor complete rail network.,29688
"than lower ranked to behave players. shown However, have professional differently been players","However, professional players have been shown to behave differently than lower ranked players.",29689
random to different variations splits. measures' times the estimate repeated is This to due multiple,This is repeated multiple times to estimate the measures' variations due to different random splits.,29690
This high and variable to prediction with external power. leads subsets internal,This leads to variable subsets with high internal and external prediction power.,29691
scalability and limits in real data applicability scarcity models segmentation semantic of Such critically applications.,Such data scarcity critically limits scalability and applicability of semantic segmentation models in real applications.,29692
scene. across actions in Human several often the objects involve inter-related interactions complex,Human actions often involve complex interactions across several inter-related objects in the scene.,29693
ActivityNet The and on large-scale datasets: Kinetics is Captions. two validated proposed method,The proposed method is validated on two large-scale datasets: Kinetics and ActivityNet Captions.,29694
image important applications. low-level vision with many Single task is an dehazing,Single image dehazing is an important low-level vision task with many applications.,29695
different of kinds have investigated address priors visual problem. to this researches Early,Early researches have investigated different kinds of visual priors to address this problem.,29696
"they not on assumptions are when their fail However, specific images. valid may","However, they may fail when their assumptions are not valid on specific images.",29697
relatively also good achieve networks Recent in this task. performance deep,Recent deep networks also achieve relatively good performance in this task.,29698
propagation network. proposed is framework A train learning lightweight to our,A lightweight learning framework is proposed to train our propagation network.,29699
caption (CVAEs). This image generation conditional paper explores auto-encoders variational using,This paper explores image caption generation using conditional variational auto-encoders (CVAEs).,29700
too little descriptions CVAEs Gaussian fixed yield a prior with with variability. Standard,Standard CVAEs with a fixed Gaussian prior yield descriptions with too little variability.,29701
excellent recently computational Discriminative based performance correlation with great efficiency. trackers filter have (DCF) achieved,Discriminative correlation filter (DCF) based trackers have recently achieved excellent performance with great computational efficiency.,29702
motion. according We to target the of prediction the area change search,We change the search area according to the prediction of target motion.,29703
"choosing optimal nodes sample of However, to is NP-hard. subset an","However, choosing an optimal subset of nodes to sample is NP-hard.",29704
"numerical various approach we the proposed the experiments. performance demonstrate Finally, through of","Finally, we demonstrate the performance of the proposed approach through various numerical experiments.",29705
"it and some However, extent. imprecise to well-known that tags user-provided are is incomplete","However, it is well-known that user-provided tags are incomplete and imprecise to some extent.",29706
"and image stable is model highly flexible our sets. large-scale handle Therefore, to","Therefore, our model is highly flexible and stable to handle large-scale image sets.",29707
rounding error. In digital computers the the generates of this propagation,In digital computers this generates the propagation of the rounding error.,29708
"in key has the studied. a training, size, mini-batch well not But factor been","But mini-batch size, a key factor in the training, has not been well studied.",29709
"the CelebA MNIST, on demonstrate effectiveness datasets. We approach Sprites, our of the and","We demonstrate the effectiveness of our approach on the MNIST, Sprites, and CelebA datasets.",29710
image classification models These thousands on of applications hundreds can rely produce that (e.g.,These applications rely on image classification models that can produce hundreds of thousands (e.g.,29711
"image medical In are images. anatomical interactive reconstructed from volumetric structures extracted segmentation,","In interactive medical image segmentation, anatomical structures are extracted from reconstructed volumetric images.",29712
also flexibility with The tracker proposed provides trackers. couple the advanced correlation to filter more,The proposed tracker also provides flexibility to couple with the more advanced correlation filter trackers.,29713
"Metropolis-Hastings section, investigate a first In we on theoretical framework algorithm. briefly","In first section, we briefly investigate a theoretical framework on Metropolis-Hastings algorithm.",29714
be This building plugged computer vision block many architectures. can into,This building block can be plugged into many computer vision architectures.,29715
"attempts models are at after selected proposed the framework, enhanced and refutation. In computational","In the proposed computational framework, models are selected and enhanced after attempts at refutation.",29716
for and multimodal methodology introduce a simultaneously. textual explanations visual also We generating,We also introduce a multimodal methodology for generating visual and textual explanations simultaneously.,29717
smartphone stacks. focal in cameras by using refocusing capture Post achievable effect is,Post capture refocusing effect in smartphone cameras is achievable by using focal stacks.,29718
streams. online a face identity method present video from unsupervised for We novel learning,We present a novel online unsupervised method for face identity learning from video streams.,29719
paper validation method This procedures. proposes error a model incorporates numerical that to due,This paper proposes a model validation method that incorporates error due to numerical procedures.,29720
models have been for Two Duffing-Ueda systems investigated. Sine identified Circuit Map and,Two identified models for Sine Map and Duffing-Ueda Circuit systems have been investigated.,29721
been indexes applied. RMSE The and MAPE have,The indexes RMSE and MAPE have been applied.,29722
technique matrix representations factorization (NMF) Non-negative latent a finding of is data. for,Non-negative matrix factorization (NMF) is a technique for finding latent representations of data.,29723
models. has The topic been applied construct to corpora method to,The method has been applied to corpora to construct topic models.,29724
"real document by NMF which However, likelihood often violated has are assumptions corpora.","However, NMF has likelihood assumptions which are often violated by real document corpora.",29725
It - on fundamental capabilities resides of hidden the underlying structure. productive premise the endowments,It resides on the premise of hidden capabilities - fundamental endowments underlying the productive structure.,29726
"we nonparametric add a to In work, component models. this such clustering","In this work, we add a nonparametric clustering component to such models.",29727
"RFB RFB of further assemble SSD, the to Net detector. constructing We top the","We further assemble RFB to the top of SSD, constructing the RFB Net detector.",29728
than accuracy We methods that without demonstrate approach undue yields complexity. higher our state-of-the-art,We demonstrate that our approach yields higher accuracy than state-of-the-art methods without undue complexity.,29729
"and multiple Zika, as diseases, the such of Aedes Chikungunya. is aegypti main Dengue, vector","Aedes aegypti is the main vector of multiple diseases, such as Dengue, Zika, and Chikungunya.",29730
"modifications geographical continuously range evolving. weather Due patterns, is to in its","Due to modifications in weather patterns, its geographical range is continuously evolving.",29731
population model We a dynamics Ae. for develop the of,We develop a model for the population dynamics of Ae.,29732
daily An extinction enhance probability. in will rainfall its increase variability likewise,An increase in daily rainfall variability will likewise enhance its extinction probability.,29733
"duration season a obtain we and extinction between probability. relationship dry Finally, nonlinear","Finally, we obtain a nonlinear relationship between dry season duration and extinction probability.",29734
can to impact outbreaks. have on findings These ability significant predict a our disease,These findings can have a significant impact on our ability to predict disease outbreaks.,29735
the are task easing methodologies for There ontology and building. techniques many of,There are many methodologies and techniques for easing the task of ontology building.,29736
a array cross series This synchronization as (e.g. diverse has of modal such applications time,This has a diverse array of applications such as cross modal time series synchronization (e.g.,29737
and discretized images. curves of video) in alignment MOCAP to,MOCAP to video) and alignment of discretized curves in images.,29738
CVXR disciplined problem's verify applies the then convexity. programming convex signed (DCP) to,CVXR then applies signed disciplined convex programming (DCP) to verify the problem's convexity.,29739
We several demonstrate CVXR's applications. framework modeling with,We demonstrate CVXR's modeling framework with several applications.,29740
kinetics. and in stochastic chemical biochemical short This approximations is a of common review two,This is a short review of two common approximations in stochastic chemical and biochemical kinetics.,29741
in refer this article All chapters in to the aforementioned chapter references book.,All chapter references in this article refer to chapters in the aforementioned book.,29742
recently in area. object shown performance (DCF) excellent have Discriminative tracking visual filters correlation,Discriminative correlation filters (DCF) have recently shown excellent performance in visual object tracking area.,29743
any (ZSL) data. recognize aims without objects training learning to novel Zero-shot unseen from classes,Zero-shot learning (ZSL) aims to recognize objects from novel unseen classes without any training data.,29744
the as space We problem. this call problem shift,We call this problem as the space shift problem.,29745
the feature The image space. semantic bridges embedding and SRG space,The SRG bridges the image feature space and semantic embedding space.,29746
setting. method easily the learning extended to generalized Our zero-shot be can,Our method can be easily extended to the generalized zero-shot learning setting.,29747
on three other method methods our that outperforms all datasets. popular Experiments datasets on show,Experiments on three popular datasets show that our method outperforms other methods on all datasets.,29748
"our can of number state-of-the-art with samples, small achieve a performance. method the Even training","Even with a small number of training samples, our method can achieve the state-of-the-art performance.",29749
"made code, Our have models data, pretrained and available. been publicly","Our data, code, and pretrained models have been made publicly available.",29750
method architecture. and loss for training new suitable the functions a also We introduce,We also introduce suitable loss functions and a training method for the new architecture.,29751
"introduce architecture, the on proposed different variants. two we Based","Based on the proposed architecture, we introduce two different variants.",29752
"More our algorithm specifically, stages: GlobalNet includes RefineNet. two and","More specifically, our algorithm includes two stages: GlobalNet and RefineNet.",29753
Deep have models real-world in success applications. many great gained learning,Deep learning models have gained great success in many real-world applications.,29754
novel We bounds describe MWSP problem. this algorithm a efficiency on that with combines provable,We describe a novel algorithm that combines efficiency with provable bounds on this MWSP problem.,29755
multi-image estimate correspondences method matching semantic This images. across proposes work a to multiple,This work proposes a multi-image matching method to estimate semantic correspondences across multiple images.,29756
"denoising a Non-Local-Means set neighbor the patch, reference filter denoise To a processes patches. of","To denoise a reference patch, the Non-Local-Means denoising filter processes a set of neighbor patches.",29757
the the computational (NN) to algorithm. Nearest used of limit Few are Neighbors burden,Few Nearest Neighbors (NN) are used to limit the computational burden of the algorithm.,29758
higher pesticides. has the for intensive use demanded need productivity The of agricultural,The need for higher agricultural productivity has demanded the intensive use of pesticides.,29759
is images tomography (CT) scan computed in a problem. location challenging Relative prediction,Relative location prediction in computed tomography (CT) scan images is a challenging problem.,29760
"Hierarchical -- of (Sutton as action et. known extended temporally abstractions, also options type a","Hierarchical abstractions, also known as options -- a type of temporally extended action (Sutton et.",29761
"principle, these defence vigor, even In include risk-averseness. behaviors or can","In principle, these behaviors can include vigor, defence or even risk-averseness.",29762
"options domain reusable in (i.e., RoboCup soccer We scenarios winning/losing). in a learn different","We learn reusable options in different scenarios in a RoboCup soccer domain (i.e., winning/losing).",29763
"are extracted the a architecture, features spectrum-based on learned center by Based classifier. classified the","Based on the learned architecture, the extracted spectrum-based features are classified by a center classifier.",29764
re-identification another. often models one domain to fail (re-ID) trained well on generalize to Person,Person re-identification (re-ID) models trained on one domain often fail to generalize well to another.,29765
"framework. translation"" attempt, a we via present In ""learning our","In our attempt, we present a ""learning via translation"" framework.",29766
images the translated supervised We then re-ID methods. by train with models,We then train re-ID models with the translated images by supervised methods.,29767
"model. consists local framework attention global model a segmentation two submodels, The a and of","The framework consists of two submodels, a global attention model and a local segmentation model.",29768
object pixel-wise time tracking framework challenging a dataset demonstrate our VOT . We on real,We demonstrate our real time pixel-wise object tracking framework on a challenging VOT dataset .,29769
that shape. depends the the wound on This of rate wound closure shows initial,This shows that the rate of wound closure depends on the initial wound shape.,29770
"is because important result never is tested. This fundamental invoked, assumption but this often","This fundamental result is important because this assumption is often invoked, but never tested.",29771
"still occlusion due problem. is to the However, sunglasses, and a mask challenging issue","However, the occlusion issue due to mask and sunglasses, is still a challenging problem.",29772
code reproduction. be released will for The,The code will be released for reproduction.,29773
an We problem-salient paper. segmentation consider instance this interesting in,We consider an interesting problem-salient instance segmentation in this paper.,29774
"high-quality boxes, Other outputs also network segments. producing our bounding instance-level than","Other than producing bounding boxes, our network also outputs high-quality instance-level segments.",29775
"an deblurring. end-to-end present for learned method We DeblurGAN, motion","We present DeblurGAN, an end-to-end learned method for motion deblurring.",29776
content The . a GAN the on based conditional learning is and loss,The learning is based on a conditional GAN and the content loss .,29777
art appearance. DeblurGAN structural state-of-the achieves performance visual in both the and measure similarity,DeblurGAN achieves state-of-the art performance both in the structural similarity measure and visual appearance.,29778
users' without resulting model approach raw data. to sensitive is The able make accessing prediction,The resulting approach is able to make model prediction without accessing users' sensitive raw data.,29779
scripts English different Telugu script beautiful The however very German. like and is from Germanic,The beautiful Telugu script however is very different from Germanic scripts like English and German.,29780
important algorithm the tools one in of field The many of is statistics. EM,The EM algorithm is one of many important tools in the field of statistics.,29781
test scenes. dataset photorealistic and non-photorealistic The includes both,The test dataset includes both photorealistic and non-photorealistic scenes.,29782
"a framework proposed consequence, consumer in is for deployment As and suitable the industrial applications.","As a consequence, the proposed framework is suitable for deployment in consumer and industrial applications.",29783
a High-quality shadow problem in shadow challenging is anti-aliasing mapping.,High-quality shadow anti-aliasing is a challenging problem in shadow mapping.,29784
'survival When of this fastest?' the favor selection does,When does selection favor this 'survival of the fastest?',29785
network perspective compels the distortion-invariant This estimation. perform to,This compels the network to perform perspective distortion-invariant estimation.,29786
Algorithms of (e.g. framework which sensor types data cases are special for this different,Algorithms which are special cases of this framework for different sensor data types (e.g.,29787
a by applications. many is obscured challenge significant objects for occluders Imaging,Imaging objects obscured by occluders is a significant challenge for many applications.,29788
"found on generalization we However, max-pooling improvement is an intriguing used. have when","However, we have found an intriguing improvement on generalization when max-pooling is used.",29789
"widely-used Oulu-CASIA. on We CK+ experiments two datasets, FER conducted and","We conducted experiments on two widely-used FER datasets, CK+ and Oulu-CASIA.",29790
sensitivity approaches. time-dependent global history-aware processes analysis of requires The,The global sensitivity analysis of time-dependent processes requires history-aware approaches.,29791
"the complex. is However, of the for direct application problem BP detection multivariable","However, the direct application of BP for the multivariable detection problem is complex.",29792
include We structures in participants interesting to ways. scene specific developed audience have,We have developed specific scene structures to include audience participants in interesting ways.,29793
refinement further clothing blurry a initial enhance with We the area network.,We further enhance the initial blurry clothing area with a refinement network.,29794
an initialization procedure tracking method used Our as methods. be with combining when can,Our method can be used as an initialization procedure when combining with tracking methods.,29795
"interference signal result, is ratio a degraded. As the (SIR) to","As a result, the signal to interference ratio (SIR) is degraded.",29796
of to treatment is considerable optimal The precision interest estimation of medicine. regimes,The estimation of optimal treatment regimes is of considerable interest to precision medicine.,29797
"Although properties. theoretical the method simple, it is nice possesses","Although the method is simple, it possesses nice theoretical properties.",29798
consistent. is regime the that We $k$-nearest neighbor show universally causal,We show that the causal $k$-nearest neighbor regime is universally consistent.,29799
"causal of the the However, dimensionality, suffer neighbor from $k$-nearest regime may i.e. curse","However, the causal $k$-nearest neighbor regime may suffer from the curse of dimensionality, i.e.",29800
learning networks various space These scheme pings. and visual corresponding can map- the handle,These networks and the corresponding learning scheme can handle various visual space map- pings.,29801
way Our offers GAN for novel search a ranking specific to through images object patterns.,Our ranking GAN offers a novel way to search through images for object specific patterns.,29802
analyzed significant response The between variability species. data salt and within Results: in revealed,Results: The analyzed data revealed significant variability in salt response within and between species.,29803
All reduced progressively NaCl by concentrations. parameters were increased growth,All growth parameters were progressively reduced by increased NaCl concentrations.,29804
"level. reduction with their tolerance the accordance among in species varied growth However,","However, the growth reduction varied among species in accordance with their tolerance level.",29805
"error the of high a genome challenge generating rates technology accurate However, pose while assemblies.","However, high error rates of the technology pose a challenge while generating accurate genome assemblies.",29806
coding model a layers. We hierarchical using define problem our sparse as,We define our model as a sparse coding problem using hierarchical layers.,29807
layer that a new variable introduce models depths. We convolution temporal kernel temporal,We introduce a new temporal layer that models variable temporal convolution kernel depths.,29808
a Residual and stream model. consists Each Networks(ResNets) a of recurrent,Each stream consists of a Residual Networks(ResNets) and a recurrent model.,29809
empirical is data. field mathematical to fit ISS a,ISS is a mathematical fit to empirical field data.,29810
ISS Gibbs/Shannon that to demonstrated proportional is the Entropy. is It,It is demonstrated that ISS is proportional to the Gibbs/Shannon Entropy.,29811
"Keywords: Co-morbidity, Injury Score, Force Blunt Trauma, Entropy. Severity","Keywords: Blunt Force Trauma, Injury Severity Score, Co-morbidity, Entropy.",29812
based are map estimation representation. methods on heat pose State-of-the-art human,State-of-the-art human pose estimation methods are based on heat map representation.,29813
"It is heat efficient, any compatible and with methods. based differentiable, map","It is differentiable, efficient, and compatible with any heat map based methods.",29814
hierarchical simultaneous propose a for penalty lag Granger and causality We estimation. also,We also propose a hierarchical penalty for simultaneous Granger causality and lag estimation.,29815
by this image we guided paper sketch. investigate hand In generation,In this paper we investigate image generation guided by hand sketch.,29816
single of in brain Recordings simple resolution with activity neuron whole are organisms. feasible now,Recordings of whole brain activity with single neuron resolution are now feasible in simple organisms.,29817
"between consistently In activation differs worms. contrast, individual neurons of","In contrast, activation of individual neurons differs consistently between worms.",29818
"ingredient: However, we approach on an autonomy. favor one that centers additional","However, we favor an approach that centers on one additional ingredient: autonomy.",29819
"turbo EP-based review we implementations. equalization In previous addition,","In addition, we review previous EP-based turbo equalization implementations.",29820
Instead decoder. of priors exploit we the default outputs considering the uniform of,Instead of considering default uniform priors we exploit the outputs of the decoder.,29821
is correction appealing an but Expressive voice problem. challenging singing,Expressive singing voice correction is an appealing but challenging problem.,29822
which recordings A two time-warping robust singing solution. a provide algorithm can synchronizes promising,A robust time-warping algorithm which synchronizes two singing recordings can provide a promising solution.,29823
"Finally, practical, demonstrate applicability the in the scenario. proposed real-world of method a we","Finally, we demonstrate the applicability of the proposed method in a practical, real-world scenario.",29824
This of videos. for objects new in paper segmentation proposes framework a semantic,This paper proposes a new framework for semantic segmentation of objects in videos.,29825
approaches online under We and supervision different levels. offline propose,We propose offline and online approaches under different supervision levels.,29826
improvement over previous methods. In model and original method state-of-the-art achieved the experiments our great,In experiments our method achieved great improvement over the original model and previous state-of-the-art methods.,29827
"network a using is (CNN). multiscale segmented LV the neural myocardium First, convolutional","First, the LV myocardium is segmented using a multiscale convolutional neural network (CNN).",29828
"LV To using subsequently the unsupervised myocardium, segmented autoencoder is encoded convolutional characterize it (CAE).","To characterize the segmented LV myocardium, it is subsequently encoded using unsupervised convolutional autoencoder (CAE).",29829
learn Causal Generative functional Neural Networks data. (CGNNs) observational to We causal from models present,We present Causal Generative Neural Networks (CGNNs) to learn functional causal models from observational data.,29830
multivariate distributional asymmetries causal structures. CGNNs and conditional discover to and independencies leverage bivariate,CGNNs leverage conditional independencies and distributional asymmetries to discover bivariate and multivariate causal structures.,29831
different Objects in and different ratios. sub-regions aspect for usually have detection characteristics distinct,Objects for detection usually have distinct characteristics in different sub-regions and different aspect ratios.,29832
novel to answers of learn the demonstrate system (i.e. produce completely capability Experiments the to,Experiments demonstrate the capability of the system to learn to produce completely novel answers (i.e.,29833
test training) from during seen provided at examples time. never,never seen during training) from examples provided at test time.,29834
and image-to-image learn integrated features transformation The both for registration. can matrix image framework simultaneously,The image-to-image integrated framework can simultaneously learn both image features and transformation matrix for registration.,29835
and trained implement to proposed efficiently. is network can The easy be,The proposed network is easy to implement and can be trained efficiently.,29836
"trained better be achieve convenience. to Thirdly, can end-to-end CMCGAN","Thirdly, CMCGAN can be trained end-to-end to achieve better convenience.",29837
the propose three information. of strategies deep resonance maximize to fusion visual-audio multimodal We benefits,We propose three multimodal deep fusion strategies to maximize the benefits of visual-audio resonance information.,29838
high explores order. low from on first The one feature the fusion to cross-modalities impact,The first one explores the impact on cross-modalities feature fusion from low to high order.,29839
second dependency the sharing The establishes of networks. short-term corresponding by weights visual-audio front-end,The second establishes the visual-audio short-term dependency by sharing weights of corresponding front-end networks.,29840
"understood. the remain inputs encode neurons external mechanisms underlying Despite intensive how research, poorly","Despite intensive research, the mechanisms underlying how neurons encode external inputs remain poorly understood.",29841
characteristic way. analysis expose patterns an empowers model's learning in to interpretable us Such,Such analysis empowers us to expose model's characteristic learning patterns in an interpretable way.,29842
usage. paper theory and underlying The the present showcases introduces its,The present paper introduces the underlying theory and showcases its usage.,29843
"tracking, large time-consuming for impossible sizes. and Manual multiple markers, becomes sample for especially","Manual tracking, especially for multiple markers, becomes time-consuming and impossible for large sample sizes.",29844
"Therefore, segmentation in the has years. for need grown of markers recent these automatic","Therefore, the need for automatic segmentation of these markers has grown in recent years.",29845
"best Finally, the find probabilistic a among match used function superpixels. to is","Finally, a probabilistic function is used to find the best match among superpixels.",29846
similar visually pixels important for information scene disambiguating Contextual in cues provides segmentation.,Contextual information provides important cues for disambiguating visually similar pixels in scene segmentation.,29847
regions. is infer different predictor trained dependency between image to dependencies The implicitly contextual,The dependency predictor is implicitly trained to infer contextual dependencies between different image regions.,29848
"dependencies, data-driven The feature inference facilitates mechanism learning. enables and of proposed contextual context-aware","The proposed mechanism enables data-driven inference of contextual dependencies, and facilitates context-aware feature learning.",29849
on spatial information. convolutions aggregate networks Neural rely to,Neural networks rely on convolutions to aggregate spatial information.,29850
"Despite annotations also the scale, large cleanliness. relatively possess its high","Despite its relatively large scale, the annotations also possess high cleanliness.",29851
"In the problem this streams. aerial of comparison cross-modal of data address work, we","In this work, we address the problem of cross-modal comparison of aerial data streams.",29852
"first but studies the about many few Currently, about are second the step, step. there","Currently, there are many studies about the first step, but few about the second step.",29853
"same modified U-net. a network, modules share The the backbone two","The two modules share the same backbone network, a modified U-net.",29854
The been publicly made has code available.,The code has been made publicly available.,29855
with Calls models well-founded to build learning. express a arms discomfort to interpretable machine,Calls to arms to build interpretable models express a well-founded discomfort with machine learning.,29856
"that!"" claims accept However, won't that ""just stakeholders","However, claims that stakeholders ""just won't accept that!""",29857
not study. a do proposed sufficient provide field of for a foundation,do not provide a sufficient foundation for a proposed field of study.,29858
(UoS). union focus subspaces on is theory detection this for of of paper The,The focus of this paper is on detection theory for union of subspaces (UoS).,29859
a generalization of et As Santoro al.,As a generalization of Santoro et al.,29860
"perceptrons, relational and networks. multi-layer we limitations Pretty-CLEVR, of probe relational Using recurrent the","Using Pretty-CLEVR, we probe the limitations of multi-layer perceptrons, relational and recurrent relational networks.",29861
for effective is learning. loss utilized more further The focal end-to-end,The focal loss is further utilized for more effective end-to-end learning.,29862
is motion one research of Human most of the activities. recognition human-centered important branches,Human motion recognition is one of the most important branches of human-centered research activities.,29863
"attention. recent attracted years, RGB-D recognition on much In motion has based data","In recent years, motion recognition based on RGB-D data has attracted much attention.",29864
"this a paper, detailed presented. is recent In of advances motion in RGB-D-based recognition overview","In this paper, a detailed overview of recent advances in RGB-D-based motion recognition is presented.",29865
require domains. vision across different images matching problems of instances object Many,Many vision problems require matching images of object instances across different domains.,29866
and These Person include Re-identification (person image (FG-SBIR) fine-grained retrieval ReID). sketch-based,These include fine-grained sketch-based image retrieval (FG-SBIR) and Person Re-identification (person ReID).,29867
"exist the in DNN. Importantly, earlier layers already of mid-level features","Importantly, mid-level features already exist in earlier layers of the DNN.",29868
"need be final extracted, represented, and They fused with layer. properly the just to","They just need to be extracted, represented, and fused properly with the final layer.",29869
"for unified idea, framework this but propose we simple CDIM. a on powerful Based","Based on this simple but powerful idea, we propose a unified framework for CDIM.",29870
training for that available new domains. cannot be results testing leveraged in data This effectively,This results in that available training data cannot be effectively leveraged for new testing domains.,29871
the show PTGAN. that could experiments be the domain by Comprehensive gap narrowed-down substantially,Comprehensive experiments show that the domain gap could be substantially narrowed-down by the PTGAN.,29872
hypothesis closely related We two problems. nonparametric testing investigate,We investigate two closely related nonparametric hypothesis testing problems.,29873
or use generating Most saliency attention low-level features models existing predictions. when descriptions task,Most existing saliency models use low-level features or task descriptions when generating attention predictions.,29874
"However, characteristics investigated. gaze patterns rarely the observer between is and link","However, the link between observer characteristics and gaze patterns is rarely investigated.",29875
to trade-offs. improve Efficient vs CNN were ResNets proposed accuracy like and designs DenseNet efficiency,Efficient CNN designs like ResNets and DenseNet were proposed to improve accuracy vs efficiency trade-offs.,29876
"across layers. connectivity, information increased allowing efficient flow They essentially the","They essentially increased the connectivity, allowing efficient information flow across layers.",29877
the expressive in power well connectedness while CNNs. of can efficiency results the preserve Sparsity,Sparsity results in efficiency while well connectedness can preserve the expressive power of the CNNs.,29878
trade-offs the original X-Nets DenseNet-BC and ResNet than architectures. give better performance,X-Nets give better performance trade-offs than the original ResNet and DenseNet-BC architectures.,29879
function of resolution expression The proposed quantization in is bits. a,The proposed expression is a function of quantization resolution in bits.,29880
non-uniform and our quantizers in simulation. employed Uniform are,Uniform and non-uniform quantizers are employed in our simulation.,29881
and We images. for one-step method blurry propose fast effective very face restoring a,We propose a very fast and effective one-step restoring method for blurry face images.,29882
they face in Also small images handle size. cannot,Also they cannot handle face images in small size.,29883
one sharp Network. Neural using method face restores step Convolutional in images Our proposed directly,Our proposed method restores sharp face images directly in one step using Convolutional Neural Network.,29884
made with embracing Adversarial Generative progress been (GANs). Networks much Image-to-image translation has,Image-to-image translation has been made much progress with embracing Generative Adversarial Networks (GANs).,29885
"In Networks paper, Region Adversarial for image-to-image we translation. (DRPAN) high-quality this present Proposal Discriminative","In this paper, we present Discriminative Region Proposal Adversarial Networks (DRPAN) for high-quality image-to-image translation.",29886
"on mobile where neural limited. networks are computational resources used Deep increasingly are devices,","Deep neural networks are increasingly used on mobile devices, where computational resources are limited.",29887
"paper with novel In this unprecedented CondenseNet, a efficiency. architecture network we develop","In this paper we develop CondenseNet, a novel network architecture with unprecedented efficiency.",29888
a It called module learned connectivity with convolution. combines group novel dense,It combines dense connectivity with a novel module called learned group convolution.,29889
based respiration signal non-contact emerging technology. strength is alternative Received rate as an monitoring,Received signal strength based respiration rate monitoring is emerging as an alternative non-contact technology.,29890
"introduced. monitoring for a paper, rate received model signal strength respiration this is In","In this paper, a received signal strength model for respiration rate monitoring is introduced.",29891
background the foreground both feature and representations. the,both the foreground and the background feature representations.,29892
values the can significantly improve filters the Suitably rate. designed to applied convergence state,Suitably designed filters applied to the state values can significantly improve the convergence rate.,29893
approximation. for Simulation support results provide the,Simulation results provide support for the approximation.,29894
cohort that characteristics common share subpopulations (e.g. Epidemiology identifying of at participants aims,Epidemiology aims at identifying subpopulations of cohort participants that share common characteristics (e.g.,29895
data. to alcohol consumption) in risk study diseases explain cohort factors of,alcohol consumption) to explain risk factors of diseases in cohort study data.,29896
can large cohort Subspace datasets. in subpopulations study heterogeneous leveraged be find clustering to and,Subspace clustering can be leveraged to find subpopulations in large and heterogeneous cohort study datasets.,29897
subspace epidemiologists to S-ADVIsED validate findings explore derived enables and from clustering.,S-ADVIsED enables epidemiologists to explore and validate findings derived from subspace clustering.,29898
can criteria subspace clusters by considering assess quality the via different of visualization. Users,Users can assess the quality of subspace clusters by considering different criteria via visualization.,29899
"subspace for involved intervals a cluster variables adjusted. in Furthermore, be can","Furthermore, intervals for variables involved in a subspace cluster can be adjusted.",29900
assumption common are is machine learning i.i.d. A that data in training,A common assumption in machine learning is that training data are i.i.d.,29901
"the can data. informative, data require less more learning is Because","Because the data is more informative, learning can require less data.",29902
helpful a learning---when comes that assumes this evidence a We teacher. learner pedagogical from call,We call this pedagogical learning---when a learner assumes that evidence comes from a helpful teacher.,29903
"how we this machine ask work might learning work, pedagogical algorithms. for In learning","In this work, we ask how pedagogical learning might work for machine learning algorithms.",29904
powerful that We teachers' examples learning. that can clustering found facilitate structure greatly contain,We found that teachers' examples contain powerful clustering structure that can greatly facilitate learning.,29905
procedure rounding The the reduce errors. present to acts,The present procedure acts to reduce the rounding errors.,29906
Runge-Kutta discretization three approach using We method on our schemes tested based .,We tested our approach using three discretization schemes based on Runge-Kutta method .,29907
"allow algorithms. evaluation contours Additionally, the processing basic of image partial object occlusion with","Additionally, object contours allow the evaluation of partial occlusion with basic image processing algorithms.",29908
Mass has been encoding Distance denominated (DCME). of technique to Encoding This Center,This encoding technique has been denominated Distance to Center of Mass Encoding (DCME).,29909
target as validate both We using variants segmentation task. medical instrument,We validate both variants using medical instrument segmentation as target task.,29910
"to are Learning-based i.e. well believed gaze for methods work estimation, unconstrained","Learning-based methods are believed to work well for unconstrained gaze estimation, i.e.",29911
"a regarding gaze or camera estimation camera. from assumptions without user, RGB environment, monocular","gaze estimation from a monocular RGB camera without assumptions regarding user, environment, or camera.",29912
makes work limitations. three these contributions addressing towards Our,Our work makes three contributions towards addressing these limitations.,29913
"challenges including illumination target facial key and We appearance variation. conditions, gaze study range,","We study key challenges including target gaze range, illumination conditions, and facial appearance variation.",29914
"appearance-based GazeNet, deep gaze we the estimation method. propose first Finally,","Finally, we propose GazeNet, the first deep appearance-based gaze estimation method.",29915
a other the components. to well-defined API component Each uses of with communicate system,Each component of the system uses a well-defined API to communicate with other components.,29916
experiments. model received substantial and Selective psychological The Tuning support neurophysiological (ST) through,The Selective Tuning (ST) model received substantial support through psychological and neurophysiological experiments.,29917
practical framework to it visual applies STAR ST tasks. The expands and,The STAR framework expands ST and applies it to practical visual tasks.,29918
execution. all communication task elements among turn in these the enable CPs for visual,CPs in turn enable the communication among all these elements for visual task execution.,29919
We flexibility are that correlated. negatively and modularity show highly,We show that flexibility and modularity are highly negatively correlated.,29920
stably WiFi-offloading functions. and reordering our Experimental the with works results demonstrate that design,Experimental results demonstrate that our design works stably with the WiFi-offloading and reordering functions.,29921
domains. two remarkable have image-to-image shown in translation Recent for success studies,Recent studies have shown remarkable success in image-to-image translation for two domains.,29922
"the target. most Finally, is tracked locate we candidate to that similar the","Finally, we locate the candidate that is most similar to the tracked target.",29923
"in multiplicative the literature, and have Several fuzzy matrices. approaches such proposed interval as been","Several approaches have been proposed in the literature, such as multiplicative and fuzzy interval matrices.",29924
"Cartesian additive, a unique system. multiplicative, and coordinate fuzzy, on","multiplicative, additive, and fuzzy, on a unique Cartesian coordinate system.",29925
study available. in are models The and publicly codes pretrained used this,The codes and pretrained models used in this study are publicly available.,29926
in central computer curriculum. Artificial is a science topic the Intelligence,Artificial Intelligence is a central topic in the computer science curriculum.,29927
"to jointly. of is multiple learning, information use human In it sources common","In human learning, it is common to use multiple sources of information jointly.",29928
"feature task. from learn existing single approaches most a only However, learning","However, most existing feature learning approaches learn from only a single task.",29929
subtype CTC cancer breast metastasizes Inflammatory via (IBC) a aggressive highly chiefly that clusters. is,Inflammatory breast cancer (IBC) is a highly aggressive subtype that chiefly metastasizes via CTC clusters.,29930
recognition Identity from community. ear of an biometric active within is images the research field,Identity recognition from ear images is an active field of research within the biometric community.,29931
entire system We early for detection melanoma. the investigate imaging mobile of of an design,We investigate the design of an entire mobile imaging system for early detection of melanoma.,29932
"smartphone-captured on visible focus images. previous we from light work, Different","Different from previous work, we focus on smartphone-captured visible light images.",29933
"on and Second, memory a constraints. is computation to smartphone stringent performed subject processing","Second, processing performed on a smartphone is subject to stringent computation and memory constraints.",29934
per-input convolutional layers We this basis. learning observation by skip to on exploit a,We exploit this observation by learning to skip convolutional layers on a per-input basis.,29935
part should discovery located. part that well A be prerequisite each of is,A prerequisite of part discovery is that each part should be well located.,29936
"Specifically, learning features contributions. discriminative and retrieval person part-informed we two for target make at","Specifically, we target at learning discriminative part-informed features for person retrieval and make two contributions.",29937
(i) Convolutional A network Baseline Part-based (PCB). named,(i) A network named Part-based Convolutional Baseline (PCB).,29938
"of an several convolutional part-level a it image features. consisting descriptor Given outputs input,","Given an image input, it outputs a convolutional descriptor consisting of several part-level features.",29939
pooling (ii) method. A (RPP) part refined,(ii) A refined part pooling (RPP) method.,29940
Experiment boost. to confirms gain round PCB that of another performance RPP allows,Experiment confirms that RPP allows PCB to gain another round of performance boost.,29941
network the layers depths by increase architectures blocks. usually or state-of-the-art cascading The building convolutional,The state-of-the-art network architectures usually increase the depths by cascading convolutional layers or building blocks.,29942
"increase present to alternative method the paper, an In we this depth.","In this paper, we present an alternative method to increase the depth.",29943
feature fundamental is videos in in learning computer vision. a problem Spatiotemporal,Spatiotemporal feature learning in videos is a fundamental problem in computer vision.,29944
cancer. an metabolism of hallmark is emerging Abnormal,Abnormal metabolism is an emerging hallmark of cancer.,29945
results design. genes to driver These identify drug and cancer for be may used,These results may be used to identify cancer driver genes and for drug design.,29946
"show current Moreover, proposed we image-to-image unsupervised methods. translation that state-of-the-art the outperforms method","Moreover, we show that the proposed method outperforms current state-of-the-art unsupervised image-to-image translation methods.",29947
are Neural Networks (CNNs) very effective tasks. Convolutional for recognition pattern many,Convolutional Neural Networks (CNNs) are very effective for many pattern recognition tasks.,29948
"However, deep needs and CNNs large computation extensive training training data.","However, training deep CNNs needs extensive computation and large training data.",29949
BFT. Through of this of effectiveness we idea simulations the demonstrate,Through simulations we demonstrate the effectiveness of this idea of BFT.,29950
"Faster However, (such can classes. R-CNN) object handle as only state-of-the-art the detectors pre-specified current","However, the current state-of-the-art object detectors (such as Faster R-CNN) can only handle pre-specified classes.",29951
"and addition, are samples large required positive In of visual for negative amounts training.","In addition, large amounts of positive and negative visual samples are required for training.",29952
Its challenges due image variability are and to descriptions. the of possible ambiguity,Its challenges are due to the variability and ambiguity of possible image descriptions.,29953
"success, their technique. this develop captioning Inspired convolutional we image a paper, by in","Inspired by their success, in this paper, we develop a convolutional image captioning technique.",29954
adversarial networks advances in significant to generative has Image completion due achieved progress (GANs).,Image completion has achieved significant progress due to advances in generative adversarial networks (GANs).,29955
and low-level because loss exists there loss. This gap reconstruction between is a high-level adversarial,This is because there exists a gap between low-level reconstruction loss and high-level adversarial loss.,29956
"outperforms and on methods. method Paris StreetView proposed Face CelebA our Evaluated significantly existing dataset,","Evaluated on CelebA Face and Paris StreetView dataset, our proposed method significantly outperforms existing methods.",29957
Correlational propose Long-term a FuCoLoT Tracker. Fully -- We,We propose FuCoLoT -- a Fully Correlational Long-term Tracker.,29958
based estimation. failure response tracking for correlation is mechanism used the novel on A,A novel mechanism based on the correlation response is used for tracking failure estimation.,29959
confidence a and fold based voting. training combination The of uses cross method,The method uses a combination of cross fold training and confidence based voting.,29960
vote directly scheme. The pixel-wise estimations into be a casting can translated,The pixel-wise estimations can be directly translated into a vote casting scheme.,29961
accurate. efficient is and highly method Our,Our method is efficient and highly accurate.,29962
a is of decoys. based initial growing found algorithm pool local in The similarities on,The algorithm is based on growing local similarities found in a pool of initial decoys.,29963
"examples. they to been vulnerable adversarial have to also shown However, be","However, they have also been shown to be vulnerable to adversarial examples.",29964
understanding defending in and will examples. efforts future aid observations Our adversarial against,Our observations will aid future efforts in understanding and defending against adversarial examples.,29965
used recognition. human aiding method in Pretraining been unsupervised The has action widely,The unsupervised Pretraining method has been widely used in aiding human action recognition.,29966
strategy feature aggregate features. to is Then map sequence a pooling adopted,Then a feature map pooling strategy is adopted to aggregate sequence features.,29967
"to similar a Siamese network Subsequently, recognition. network designed to perform is is which","Subsequently, a network which is similar to Siamese network is designed to perform recognition.",29968
conducted large on Cross-view recognition gait OU-ISIR dataset. are experiments population,Cross-view gait recognition experiments are conducted on OU-ISIR large population dataset.,29969
demonstrate can extract features our aggregate network silhouette results that effectively. and sequence The from,The results demonstrate that our network can extract and aggregate features from silhouette sequence effectively.,29970
"is with reduced devised. symbol-level a In work, algorithm complexity optimization this","In this work, a symbol-level optimization algorithm with reduced complexity is devised.",29971
"that On outperforms result model baselines. datasets, demonstrates other state-of-the-art a diagram available publicly our","On publicly available diagram datasets, our model demonstrates a state-of-the-art result that outperforms other baselines.",29972
"the proposed various of shows applications. answering on method for further question potentials experiments Moreover,","Moreover, further experiments on question answering shows potentials of the proposed method for various applications.",29973
Learning and very data generate from problem. represent to is challenging videos unlabeled a,Learning to represent and generate videos from unlabeled data is a very challenging problem.,29974
according these of be divided generation The intrinsic difficulties. to video process should,The process of video generation should be divided according to these intrinsic difficulties.,29975
"videos. to we optical motion information Thus, generate as employ flow","Thus, we employ optical flow as motion information to generate videos.",29976
plausible realistic and appearance brings with consistency. hierarchical motion videos This approach more,This hierarchical approach brings more realistic videos with plausible motion and appearance consistency.,29977
react element key and becomes prevent Anticipation timely accidents. order a to in,Anticipation becomes a key element in order to react timely and prevent accidents.,29978
enhancement register rely optical video to frames flow Many in on a video sequence. algorithms,Many video enhancement algorithms rely on optical flow to register frames in a video sequence.,29979
matching and receives solving jointly applying few of attention. graph way and the However cuts,However the way of jointly applying and solving graph cuts and matching receives few attention.,29980
and cuts their graph correspondence i.e. establishing,graph cuts and establishing their correspondence i.e.,29981
"By flip, feature can convolution layer the using produce three maps.","By using flip, the convolution layer can produce three feature maps.",29982
learning propose in first contribution (RL). Our reinforcement learning is attribute twofold: we,Our contribution is twofold: first we propose attribute learning in reinforcement learning (RL).,29983
"learning. do we way the using to different propose classification, similarity a this In paper,","In this paper, we propose a different way to do the classification, using similarity learning.",29984
"effective. The is simple, approach scalable and","The approach is simple, scalable and effective.",29985
our adaptation performance domain unsupervised that model show We different scenarios. state-of-the-art achieves in,We show that our model achieves state-of-the-art performance in different unsupervised domain adaptation scenarios.,29986
given model. to illustrative the the application effectiveness An is proposed demonstrate of,An illustrative application is given to demonstrate the effectiveness of the proposed model.,29987
the to method Our provide particle detect resampling object accordingly. can candidates more by,Our method can provide more candidates by particle resampling to detect the object accordingly.,29988
variations huge intra-class these visual similarities. dermoscopic lesions has of The and high inter-class appearance,The dermoscopic appearance of these lesions has huge intra-class variations and high inter-class visual similarities.,29989
of segmentation on research is current of Most skin single-class classes lesions. focusing irrespective,Most current research is focusing on single-class segmentation irrespective of classes of skin lesions.,29990
"the and loss hybrid of To performance learning are improve function transfer used. a FCNs,","To improve the performance of FCNs, transfer learning and a hybrid loss function are used.",29991
"learning is trivial utilizing it when not video a CNN Nevertheless, for spatio-temporal representation.","Nevertheless, it is not trivial when utilizing a CNN for learning spatio-temporal video representation.",29992
approach evaluate dataset. on study controlled a We our the COCO in,We evaluate our approach in a controlled study on the COCO dataset.,29993
and us framework which a develop is allows to robust. discriminative This classification texture,This allows us to develop a texture classification framework which is discriminative and robust.,29994
and combinations of on All transforms linear methods are histogram operations. based the,All the methods are based on combinations of linear transforms and histogram operations.,29995
trainable a The new is end-to-end neural prepended network to pipeline.,The neural network is prepended to a new end-to-end trainable pipeline.,29996
"training, robust and exhibits outstanding Our generalization in is capabilities. highly efficient, system accurate,","Our system is efficient, highly accurate, robust in training, and exhibits outstanding generalization capabilities.",29997
outdoor exceeds and datasets. indoor It on state-of-the-art consistently,It exceeds state-of-the-art consistently on indoor and outdoor datasets.,29998
technique used. bit-stealing a display was the resolution luminance of To increase the,To increase the luminance resolution of the display a bit-stealing technique was used.,29999
This paper a demonstrates (i.e. that representation mesh,This paper demonstrates that a mesh representation (i.e.,30000
data also clouds. to similar however mesh A point is unstructured,A mesh however is also unstructured data similar to point clouds.,30001
real-world synthetic with promising and our results. evaluate We on very datasets method,We evaluate our method on synthetic and real-world datasets with very promising results.,30002
Cobb-Douglas paper of This in heterogeneous production the the presents function. elasticities identification,This paper presents the identification of heterogeneous elasticities in the Cobb-Douglas production function.,30003
noise the input. auxiliary by learned the style can Also be,Also the style can be learned by the auxiliary noise input.,30004
"two the In vision, computer in are separate traditionally problems solved tracks.","In computer vision, the two problems are traditionally solved in separate tracks.",30005
"a emulate human use to perception. as we stacks input focal pair Specifically, of","Specifically, we use a pair of focal stacks as input to emulate human perception.",30006
focal rendering. We first training field synthesized a stack light depth-guided dataset comprehensive by construct,We first construct a comprehensive focal stack training dataset synthesized by depth-guided light field rendering.,30007
show high-quality them unified maps. to a to We depth integrate obtain how into BDfF-Net,We show how to integrate them into a unified BDfF-Net to obtain high-quality depth maps.,30008
"PSO. classical walking helps converge particle This random to the then quicker constrained, non-uniform movement","This constrained, non-uniform movement helps the random walking particle to converge quicker then classical PSO.",30009
"effective error error-prone; methods estimation is hence, detection depth desirable. are Stereo","Stereo depth estimation is error-prone; hence, effective error detection methods are desirable.",30010
the dataset method. show the our of effectiveness on Middlebury Experiments,Experiments on the Middlebury dataset show the effectiveness of our method.,30011
The this a art descriptor by implicitly addresses camera-invariant subspace. problem learning,The art implicitly addresses this problem by learning a camera-invariant descriptor subspace.,30012
"consider this introducing In adaptation. explicitly camera by paper, style (CamStyle) this we challenge","In this paper, we explicitly consider this challenge by introducing camera style (CamStyle) adaptation.",30013
data camera can augmentation the style as that smooths a CamStyle serve disparities. approach,CamStyle can serve as a data augmentation approach that smooths the camera style disparities.,30014
"considerable of method, over-fitting, noise. while a increasing against incurs diversity data also level This","This method, while increasing data diversity against over-fitting, also incurs a considerable level of noise.",30015
"over-fitting. extent improvement systems demonstrate we all With LSR, the of consistent of regardless in","With LSR, we demonstrate consistent improvement in all systems regardless of the extent of over-fitting.",30016
compared also with accuracy We art. of competitive state the report the,We also report competitive accuracy compared with the state of the art.,30017
regions. these salient CAR-Net successfully attends to,CAR-Net successfully attends to these salient regions.,30018
"CAR-Net on accuracy forecasting Drone the reaches Stanford standard (SDD). state-of-the-art trajectory Additionally, Dataset benchmark,","Additionally, CAR-Net reaches state-of-the-art accuracy on the standard trajectory forecasting benchmark, Stanford Drone Dataset (SDD).",30019
"generalize to ability we CAR-Net's Finally, show scenes. unseen to","Finally, we show CAR-Net's ability to generalize to unseen scenes.",30020
The of the affects molecular rate recombination mode of evolution.,The rate of recombination affects the mode of molecular evolution.,30021
architecture. this selection argue that variation genome generates on We,We argue that this variation generates selection on genome architecture.,30022
"calculated. For noise DCT a variance robust of estimate is coefficients, the","For the DCT coefficients, a robust estimate of noise variance is calculated.",30023
"implementations applied of sparse networks such ""dense"" are convolutional Standard inefficient when very on data.","Standard ""dense"" implementations of convolutional networks are very inefficient when applied on such sparse data.",30024
problem. super-resolution Face Super-Resolution a is domain-specific (SR),Face Super-Resolution (SR) is a domain-specific super-resolution problem.,30025
facial knowledge better face for prior leveraged The be could images. super-resolving specific,The specific facial prior knowledge could be leveraged for better super-resolving face images.,30026
"coarse network recover to a (HR) high-resolution coarse a image. Specifically, construct we SR first","Specifically, we first construct a coarse SR network to recover a coarse high-resolution (HR) image.",30027
be made available upon publication. Code will,Code will be made available upon publication.,30028
deterministic geometric fast. and performed purely reduction are which is The operations using,The reduction is performed using purely geometric operations which are deterministic and fast.,30029
"preserved. true the removed, outliers Further, globally are optimal since solution is only","Further, since only true outliers are removed, the globally optimal solution is preserved.",30030
"effectiveness various experiments, preprocessing the of demonstrate we and On method. data our synthetic real","On various synthetic and real data experiments, we demonstrate the effectiveness of our preprocessing method.",30031
material. available supplementary as is Demo code,Demo code is available as supplementary material.,30032
framework. false detection into rates detection probabilities missed Background are the incorporated and,Background false detection rates and missed detection probabilities are incorporated into the framework.,30033
for metric as of the used Information geometries. is Fisher then optimal design a agent,Fisher Information is then used as a metric for the design of optimal agent geometries.,30034
"results presented of algorithm. are Finally, effectiveness the simulation the to demonstrate","Finally, simulation results are presented to demonstrate the effectiveness of the algorithm.",30035
addressing lack is training of through the model generative a helpful data. in,through a generative model is helpful in addressing the lack of training data.,30036
show are . arithmetic floating-point We to that difference these related,We show that these difference are related to floating-point arithmetic .,30037
and is architecture has residual of similar U-Net. units with The to that built network,The network is built with residual units and has similar architecture to that of U-Net.,30038
"two-fold: of first, units deep networks. training model of ease is this benefits The residual","The benefits of this model is two-fold: first, residual units ease training of deep networks.",30039
"available. of sake detection trained we our the For code reproducibility, publicly and make networks","For the sake of reproducibility, we make our trained networks and detection code publicly available.",30040
"first The by occluded since problem not given. the raindrops are intractable, is regions","The problem is intractable, since first the regions occluded by raindrops are not given.",30041
"generative network an resolve the adversarial attentive training. apply problem, we using To","To resolve the problem, we apply an attentive generative network using adversarial training.",30042
and visual Our discriminative into is the attention idea both networks. main generative to inject,Our main idea is to inject visual attention into both the generative and discriminative networks.,30043
"our the raindrop about surroundings. attention their regions learns During and training, visual","During the training, our visual attention learns about raindrop regions and their surroundings.",30044
"of building work al. Finally, recent et the Babaeizadeh on","Finally, building on the recent work of Babaeizadeh et al.",30045
"Best of (BLADE). RAISR, Linear We generalization name Adaptive we a Enhancement describe which","We describe a generalization of RAISR, which we name Best Linear Adaptive Enhancement (BLADE).",30046
re-search wide learning. interest of adversarial gotten deep field Generative (GAN) network has in the,Generative adversarial network (GAN) has gotten wide re-search interest in the field of deep learning.,30047
tasks. have competitive results specific on achieved GAN of Variations,Variations of GAN have achieved competitive results on specific tasks.,30048
"instances of and are diversity However, stability worth still training of generated the studying further.","However, the stability of training and diversity of generated instances are still worth studying further.",30049
key ill-posed depth ambiguity the challenge A nature. is,A key challenge is the ill-posed depth ambiguity nature.,30050
novel This representation skeleton map presents for regression. a paper intermediate feature named,This paper presents a novel intermediate feature representation named skeleton map for regression.,30051
image irrelavant structural context properties e.g. from distills of It RGB,It distills structural context from irrelavant properties of RGB image e.g.,30052
"easily can clean via generated be network. It and deconvolution simple, is","It is simple, clean and can be easily generated via deconvolution network.",30053
investigated Both and feature prediction been tackle the modeling to problem. learning have,Both feature learning and prediction modeling have been investigated to tackle the problem.,30054
"single poses. a estimation Though of yield existing methods most target effective, discriminative deterministic","Though effective, most existing discriminative methods yield a single deterministic estimation of target poses.",30055
of trainable. functions a framework whole naturally The differentiable with is two mixture end-to-end density,The whole framework with a mixture of two differentiable density functions is naturally end-to-end trainable.,30056
fixed feed-forward image use segmentation State-of-the-art systems with costs. computational pipelines for semantic,State-of-the-art systems for semantic image segmentation use feed-forward pipelines with fixed computational costs.,30057
suited find efficiently that architecture segmenting is uniquely for this We videos.,We find that this architecture is uniquely suited for efficiently segmenting videos.,30058
"face two progression, underlying i.e. requirements of age The","The two underlying requirements of face age progression, i.e.",30059
"are and the identity aging studied well not permanence, accuracy in literature.","aging accuracy and identity permanence, are not well studied in the literature.",30060
"approach. network paper, generative a novel we based adversarial In present this","In this paper, we present a novel generative adversarial network based approach.",30061
the approach quantitative fidelity visual Both the that advances show evaluations and state-of-the-art.,Both visual fidelity and quantitative evaluations show that the approach advances the state-of-the-art.,30062
"to context method. the improve our was of used performance further Furthermore, infor-mation","Furthermore, context infor-mation was used to further improve the performance of our method.",30063
"unsorted lack like basically annotations and are However, semantic location. they type and","However, they are basically unsorted and lack semantic annotations like type and location.",30064
"comprehensive A task. for is this and large-scale introduced DeepLesion, dataset,","A large-scale and comprehensive dataset, DeepLesion, is introduced for this task.",30065
the attributes annotation little of useful lesions. effort describe manual but require They,They require little manual annotation effort but describe useful attributes of the lesions.,30066
"retrieval, qualitative lesion quantitative clustering, promising classification. results and and on show Experiments","Experiments show promising qualitative and quantitative results on lesion retrieval, clustering, and classification.",30067
for propose algorithms annotation intra-patient mining. We and lesion missing matching,We propose algorithms for intra-patient lesion matching and missing annotation mining.,30068
problem challenging computer an in important detection and Object vision. is,Object detection is an important and challenging problem in computer vision.,30069
applications real represents are well and DOTA Experiments Vision Earth that quite demonstrate challenging.,Experiments demonstrate that DOTA well represents real Earth Vision applications and are quite challenging.,30070
recall. either low methods or precision low Current from suffer,Current methods suffer from either low precision or low recall.,30071
"a difficulty we method. problem, classification-based adopt of this the decrease To","To decrease the difficulty of this problem, we adopt a classification-based method.",30072
adopted network and for K-means A classifier classification. a clustering convolutional module neural are,A convolutional neural network classifier and a K-means clustering module are adopted for classification.,30073
naturally diversity this entails in Limiting choices loss. welfare ensuring agents' for manner some,Limiting agents' choices for ensuring diversity in this manner naturally entails some welfare loss.,30074
to results than likely others. are be Some empirical more published,Some empirical results are more likely to be published than others.,30075
inference. Such estimates and leads selective to publication distorted biased,Such selective publication leads to biased estimates and distorted inference.,30076
architectures. experimented have We several with state-of-the-art,We have experimented with several state-of-the-art architectures.,30077
resolution We also these networks. investigated limits of have,We have also investigated resolution limits of these networks.,30078
the strengthens local discriminative Integrating representation. ability information the contextual of,Integrating the contextual information strengthens the discriminative ability of local representation.,30079
and between complementary leverage local feature. information global We the also,We also leverage the complementary information between local and global feature.,30080
model interpret prediction. significant allow the for Our to constructed physicians is features to,Our model is constructed to allow physicians to interpret the significant features for prediction.,30081
"videos, On each worse tracker current of best these repetitive on the most loop. performs","On most of these repetitive videos, the best current tracker performs worse on each loop.",30082
between optimization tracking. illustrates versus difference long-term the for This short-term,This illustrates the difference between optimization for short-term versus long-term tracking.,30083
"employs also tracker guided assessment. the cautious importantly, proposed self-quality updating, Most by","Most importantly, the proposed tracker also employs cautious updating, guided by self-quality assessment.",30084
in nearest many search great neighbor achieved success (ANN) Approximate has tasks.,Approximate nearest neighbor (ANN) search has achieved great success in many tasks.,30085
derive performance guarantees loss PQ of our model. a online the We which bound,We derive a loss bound which guarantees the performance of our online PQ model.,30086
(FSO) Space of Optics the is systems Free one wireless evolving technologies.,Free Space Optics systems (FSO) is one of the evolving wireless technologies.,30087
"of input link factor computed. attenuation Impact power and been has range, rate, data","Impact of data rate, link range, input power and attenuation factor has been computed.",30088
These issues this discussed in are comment.,These issues are discussed in this comment.,30089
analysis the of attention the AttnGAN. performed also A layers visualizing detailed by is,A detailed analysis is also performed by visualizing the attention layers of the AttnGAN.,30090
"densely feature points, interest features. any sampled around TS trajectories The encodes appearance only without","The TS feature encodes only trajectories around densely sampled interest points, without any appearance features.",30091
"images, of We which collected foot an contain DFU different from extensive patients. dataset","We collected an extensive dataset of foot images, which contain DFU from different patients.",30092
"classification. used we in Furthermore, Convolutional the (CNNs) for Neural time Networks DFU first","Furthermore, we used Convolutional Neural Networks (CNNs) for the first time in DFU classification.",30093
we learning This both classifiers deep and machine have learning the outperformed tested.,This outperformed both the machine learning and deep learning classifiers we have tested.,30094
potential approach This a paradigm diabetic novel has care. foot deliver shift in the to,This novel approach has the potential to deliver a paradigm shift in diabetic foot care.,30095
be Adversarial Networks proved kinds generation of image efficient various to tasks. are Generative on,Generative Adversarial Networks are proved to be efficient on various kinds of image generation tasks.,30096
"we is want images precisely. still it a challenge However, generate to if","However, it is still a challenge if we want to generate images precisely.",30097
how with images attribute. generate to researchers Many focus on one,Many researchers focus on how to generate images with one attribute.,30098
is work. image But generation tough still under a multiple attributes,But image generation under multiple attributes is still a tough work.,30099
"in-the-wild SDAP synthesizing allows faces. efficiently aging Finally,","Finally, SDAP allows efficiently synthesizing in-the-wild aging faces.",30100
"Wild and (AGFW), in (CACD). Faces MORPH, Cross-Age the Celebrity AginG Dataset FG-NET,","FG-NET, MORPH, AginG Faces in the Wild (AGFW), and Cross-Age Celebrity Dataset (CACD).",30101
Experimental of our on both and algorithm. results scenes the benefits indoor outdoor demonstrate,Experimental results on both indoor and outdoor scenes demonstrate the benefits of our algorithm.,30102
expression for sequences. is in Spatio-temporal encoding dynamics video encoding essential feature facial,Spatio-temporal feature encoding is essential for encoding facial expression dynamics in video sequences.,30103
expression. in predicting delay the causes This a,This causes a delay in predicting the expression.,30104
"on-the-fly. prediction As the be could performed such,","As such, the prediction could be performed on-the-fly.",30105
a The maximization seen as can be problem. likelihood problem,The problem can be seen as a likelihood maximization problem.,30106
sources It of returns the same as set algorithm. the used widely FastICA,It returns the same set of sources as the widely used FastICA algorithm.,30107
same to the compared proposed better approach performance other The that proposals classified dataset. showed,The proposed approach showed better performance compared to other proposals that classified the same dataset.,30108
from throughout the experience Lessons highlighted paper. are also learned the,Lessons learned from the experience are also highlighted throughout the paper.,30109
such to dramatically recently deep has of due accuracy models increased The learning.,The accuracy of such models has dramatically increased recently due to deep learning.,30110
saliency by significant fixation predicted over algorithms. improvement is sequences a This state-of-the-art,This is a significant improvement over fixation sequences predicted by state-of-the-art saliency algorithms.,30111
"as of dependency sparsity."" this to We ""region refer form","We refer to this form of dependency as ""region sparsity.""",30112
weights encourages prior The region-sparse to bases two be different resulting simultaneously. in,The resulting prior encourages weights to be region-sparse in two different bases simultaneously.,30113
a for unconstrained recognition. face structure new deep We propose network,We propose a new deep network structure for unconstrained face recognition.,30114
an computer face is important vision. Human analysis in task,Human face analysis is an important task in computer vision.,30115
"face facial analysis. for studies, According provide crucial could dynamics cognitive-psychological cues to","According to cognitive-psychological studies, facial dynamics could provide crucial cues for face analysis.",30116
"method has experiments, proposed verified. been By comparative of the the effectiveness","By comparative experiments, the effectiveness of the proposed method has been verified.",30117
"the max-pooling literature. in used operators of the most Currently, is commonly one computational","Currently, max-pooling is one of the most commonly used operators in the computational literature.",30118
of the We mechanism shortcomings of max-pooling. some that address hypothesise this contrast-variant-pooling can,We hypothesise that this contrast-variant-pooling mechanism can address some of the shortcomings of max-pooling.,30119
"increase have recently computed based To gained much attention. techniques performance, retrieval distances on re-ranking","To increase retrieval performance, re-ranking techniques based on computed distances have recently gained much attention.",30120
re-ranking re-ranking achieves that state-of-the-art performance. new a unsupervised framework We automatic and propose,We propose a new unsupervised and automatic re-ranking framework that achieves state-of-the-art re-ranking performance.,30121
"ImageNet the pre-trained we used Specifically, model the dataset. on GoogleNet","Specifically, we used the GoogleNet model pre-trained on the ImageNet dataset.",30122
recent There detection object  in progresses has image years. for significant  been,  There has been significant progresses for image object detection in recent years.,30123
human plays a in representation action role recognition vital Motion videos. in,Motion representation plays a vital role in human action recognition in videos.,30124
The using for difference derivation also frames. two between theoretical provides the support,The derivation also provides theoretical support for using the difference between two frames.,30125
"enables extract frames simultaneously. the to information, It temporal CNN between spatiotemporal the especially information","It enables the CNN to extract spatiotemporal information, especially the temporal information between frames simultaneously.",30126
simple but is validated powerful by This idea results. experimental,This simple but powerful idea is validated by experimental results.,30127
predicts observed people videos. We of locations future new task first-person in that a present,We present a new task that predicts future locations of people observed in first-person videos.,30128
a continuously first-person recorded video a by wearable stream camera. Consider,Consider a first-person video stream continuously recorded by a wearable camera.,30129
We framework a into these a three observations architecture. convolution-deconvolution multi-stream incorporate prediction with,We incorporate these three observations into a prediction framework with a multi-stream convolution-deconvolution architecture.,30130
of is both performance The numerically. demonstrated FPS-SFT and theoretically,The performance of FPS-SFT is demonstrated both theoretically and numerically.,30131
"study, with network network recurrent neural In on we this use a attention convolutional features.","In this study, we use a recurrent neural network with attention on convolutional network features.",30132
multi-step time We framework series a for regression. probabilistic general propose,We propose a framework for general probabilistic multi-step time series regression.,30133
"nature temporal and (e.g. we Specifically, expressiveness Neural exploit Networks the Sequence-to-Sequence of","Specifically, we exploit the expressiveness and temporal nature of Sequence-to-Sequence Neural Networks (e.g.",30134
"is nets and scheme, *forking-sequences*, boost training stability designed performance. to sequential new A for","A new training scheme, *forking-sequences*, is designed for sequential nets to boost stability and performance.",30135
Object military is operations. automatic critical for Detection,Object Detection is critical for automatic military operations.,30136
results many challenging Recent in promising in learning deep tasks. advances achieved have,Recent advances in deep learning have achieved promising results in many challenging tasks.,30137
significantly the public object detection datasets. CNN-based performance generic on These methods detection several improve,These CNN-based methods improve the detection performance significantly on several public generic object detection datasets.,30138
of Lesions identified were interest radiologists. by,Lesions of interest were identified by radiologists.,30139
"were experiments and GoogleNet, Network in utilized our VGG, CIFAR. architectures","Network architectures utilized in our experiments were GoogleNet, VGG, and CIFAR.",30140
"fully off-the-shelf connected the features from the For performed the layer extracted best. the approach,","For the off-the-shelf approach, the features extracted from the fully connected layer performed the best.",30141
radiogenomic discovering play in a in Conclusion: may Deep cancer. role associations learning breast,Conclusion: Deep learning may play a role in discovering radiogenomic associations in breast cancer.,30142
"techniques tempting. approaches MICE, Beneath such Learning classical the is of as Machine application","Beneath classical approaches such as MICE, the application of Machine Learning techniques is tempting.",30143
"classification on for is a and core, random respectively. its based In forest it regression,","In its core, it is based on a random forest for classification and regression, respectively.",30144
"random the protocol particular, In resampling other forest within strategies suggested. are","In particular, other resampling strategies within the random forest protocol are suggested.",30145
"on is and conducted. focusing data empirical credit analysis an information Facebook Finally,","Finally, an empirical analysis focusing on credit information and Facebook data is conducted.",30146
Stein Discrepancy. also Riemannian Stein's Kernelized and discover Riemannian Identity We,We also discover Riemannian Stein's Identity and Riemannian Kernelized Stein Discrepancy.,30147
"named N-P we criterion. criterion, develop a multi-hypothesis novel test generalized the Then,","Then, we develop a novel test criterion, named the generalized multi-hypothesis N-P criterion.",30148
monitoring. method for non-invasive cardiac potential Seismocardiography a (SCG) offers,Seismocardiography (SCG) offers a potential non-invasive method for cardiac monitoring.,30149
"effect is signal on SCG the of study, this investigated. the morphology In the respiration","In this study, the effect of the respiration on the SCG signal morphology is investigated.",30150
SCG have slightly that different tended morphologies. to two Results showed events,Results showed that SCG events tended to have two slightly different morphologies.,30151
Pulmonary diagnosing is Background: diseases. common a various respiratory tool auscultation for,Background: Pulmonary auscultation is a common tool for diagnosing various respiratory diseases.,30152
have details in studies humans. pulmonary sounds Previous documented of many,Previous studies have documented many details of pulmonary sounds in humans.,30153
"pressure on sound inside generation is and However, airways information animal scarce. loss","However, information on sound generation and pressure loss inside animal airways is scarce.",30154
"hand-crafted low-level videos Therefore, are represent needed methods. high-level semantic features rather than to","Therefore, high-level semantic features to represent videos are needed rather than low-level hand-crafted methods.",30155
inverse to solve uses deep learning that techniques new problems. method a We propose the,We propose a new method that uses deep learning techniques to solve the inverse problems.,30156
"of compensate wrong extra large needed Thus, for are rewards. these number to a iterations","Thus, a large number of extra iterations are needed to compensate for these wrong rewards.",30157
"be rewards way, instantly. wrong corrected the In this can","In this way, the wrong rewards can be corrected instantly.",30158
"algorithms. the estimator Competitive proposed overcomes drawbacks of Double Hence, the existing Algorithm","Hence, the proposed Double Competitive Algorithm overcomes the drawbacks of existing estimator algorithms.",30159
the $\epsilon-optimality$ to of the show A analysis scheme. proposed refined is presented,A refined analysis is presented to show the $\epsilon-optimality$ of the proposed scheme.,30160
"scales calibrate relative and end, first To cameras. between each camera we estimate this","To this end, we first calibrate each camera and estimate relative scales between cameras.",30161
equipped The users with directional (SUs) antennas. are steerable secondary,The secondary users (SUs) are equipped with steerable directional antennas.,30162
fed a is neuron with Poisson A process.,A neuron is fed with a Poisson process.,30163
applied neuron delay a finite the itself impulse to output after Each $\Delta$. is,Each output impulse is applied to the neuron itself after a finite delay $\Delta$.,30164
inhibitory Cl-type a as fast delivered being synapse. This impulse acts through,This impulse acts as being delivered through a fast Cl-type inhibitory synapse.,30165
between relations moments Similar corresponding are derived.,Similar relations between corresponding moments are derived.,30166
is initial That $p(t)$ well. segment is model-independent of the as,That is the initial segment of $p(t)$ is model-independent as well.,30167
checked are of expressions simulation. by Monte means obtained The Carlo,The obtained expressions are checked by means of Monte Carlo simulation.,30168
success learning of progressed Recent research computer further. deep in such have vision,Recent success of deep learning in computer vision have progressed such research further.,30169
employ slices image for to informative We the training. select most entropy,We employ image entropy to select the most informative slices for training.,30170
vibrations signals induced Seismocardiographic cardiac activity. are by (SCG) chest surface,Seismocardiographic (SCG) signals are chest surface vibrations induced by cardiac activity.,30171
function. a These diagnosing may offer monitoring for method signals heart and,These signals may offer a method for diagnosing and monitoring heart function.,30172
characteristics. signal time-frequency approach features One determining estimate to is of its,One approach of determining signal features is to estimate its time-frequency characteristics.,30173
signals than synthetic had lower for considered. WVD NRMSE STFT,STFT had lower NRMSE than WVD for synthetic signals considered.,30174
also discrimination PCT between components. provided SPWVD better and frequency signal,PCT and SPWVD also provided better discrimination between signal frequency components.,30175
"prior the neural we networks. this paper, within of consider knowledge use In","In this paper, we consider the use of prior knowledge within neural networks.",30176
change maximal known transforms error of to that We bounds. able is use demonstrate,We demonstrate that use of known transforms is able to change maximal error bounds.,30177
"salient and detection. model In an for discriminative we paper, efficient propose object this","In this paper, we propose an efficient and discriminative model for salient object detection.",30178
impedance. well described behavior by This and neuronal as is known the is resonance,This behavior is known as resonance and is well described by the neuronal impedance.,30179
($G_h^{Der}$ to factors the contributes $\tau_h$) of appearance The and increment of greatly resonance. both,The increment of both factors ($G_h^{Der}$ and $\tau_h$) greatly contributes to the appearance of resonance.,30180
due $G_h^{Der}$. dependent is the of to resonance We voltage dependence also demonstrate voltage that,We also demonstrate that resonance is voltage dependent due to the voltage dependence of $G_h^{Der}$.,30181
Generative modeling in machine learning. a high-dimensional problem is of data key,Generative modeling of high-dimensional data is a key problem in machine learning.,30182
and latent Successful models. models approaches include autoregressive variable,Successful approaches include latent variable models and autoregressive models.,30183
current focus Embedding is spaces of vector a graphs research. knowledge (KGs) into continuous,Embedding knowledge graphs (KGs) into continuous vector spaces is a focus of current research.,30184
such embedding logic attention. increasing recently model Combining an with has attracted rules,Combining such an embedding model with logic rules has recently attracted increasing attention.,30185
RUGE and We prediction YAGO. in link evaluate Freebase on,We evaluate RUGE in link prediction on Freebase and YAGO.,30186
the The sensor direction. direction is the fixed of from moving,The direction of the sensor is fixed from the moving direction.,30187
"are on Typically, sensors mobile mounted vehicles.","Typically, mobile sensors are mounted on vehicles.",30188
the measures it Each sensor continuously the object. from target to distance,Each sensor continuously measures the distance from it to the target object.,30189
detection. impressive large performance improvement advances object The of detectors for scale recent show convolutional,The recent advances of convolutional detectors show impressive performance improvement for large scale object detection.,30190
tasks. strength a We extensively of and semantic demonstrate classification DAUs segmentation on the,We extensively demonstrate the strength of DAUs on a classification and semantic segmentation tasks.,30191
"novel deep us from allow perspectives. to Furthermore, DAUs study networks","Furthermore, DAUs allow us to study deep networks from novel perspectives.",30192
comparably performs find that the hybrid computational We approach providing while benefits.,We find that the hybrid approach performs comparably while providing computational benefits.,30193
module. object work proposes This relation an,This work proposes an object relation module.,30194
is require and easy It to networks. embed in additional supervision existing not does,It does not require additional supervision and is easy to embed in existing networks.,30195
detection. object efficacy the relations modeling It in CNN of based verifies,It verifies the efficacy of modeling object relations in CNN based detection.,30196
It gives rise to end-to-end the fully object detector. first,It gives rise to the first fully end-to-end object detector.,30197
"challenge. significant missing occlusions parts However, dealing still a is and with","However, dealing with occlusions and missing parts is still a significant challenge.",30198
"partial this method completion novel learning-based the a In work, shapes. we of for propose","In this work, we propose a novel learning-based method for the completion of partial shapes.",30199
on appearance part. exhibits realistic shape completed The the a unknown,The completed shape exhibits a realistic appearance on the unknown part.,30200
images using We input flow architecture. start optical bi-directional between U-Net computing the a by,We start by computing bi-directional optical flow between the input images using a U-Net architecture.,30201
"warped images Finally, each are fused intermediate form frame. and input the two to linearly","Finally, the two input images are warped and linearly fused to form each intermediate frame.",30202
statistics a notation. captures approach note raga patterns of A repetitive the based dictionary within,A dictionary based approach captures the statistics of repetitive note patterns within a raga notation.,30203
The L-system simulate inference problem to is to an plant. given find a L-system,The L-system inference problem is to find an L-system to simulate a given plant.,30204
able existing is than PMIT to complex more approaches. systems infer,PMIT is able to infer more complex systems than existing approaches.,30205
been application Classically problem of adjustment solved has (BA). this bundle through the,Classically this problem has been solved through the application of bundle adjustment (BA).,30206
"able sufficiently these of However, are structures to models local point not capture the clouds.","However, these models are not able to sufficiently capture the local structures of point clouds.",30207
The to learn progressively the tree vectors representation designed method is structure. along,The method is designed to learn representation vectors progressively along the tree structure.,30208
proposed benchmarks on provides show challenging that Experiments features. set discriminative point the model,Experiments on challenging benchmarks show that the proposed model provides discriminative point set features.,30209
Qualitative high-accuracy imagery undistortions diverse demonstrate on rectifications. results and,Qualitative results on diverse imagery demonstrate high-accuracy undistortions and rectifications.,30210
document performance well. of analysis field as in have the achieved state-of-the-art They,They have achieved state-of-the-art performance in the field of document analysis as well.,30211
"data of However, annotated effort. require CNNs hence, training and, manual a amount large great","However, CNNs require a large amount of annotated training data and, hence, great manual effort.",30212
"connected. far, utilized fully-connected all approaches So networks, existing processors in implicitly which are have","So far, existing approaches have implicitly utilized fully-connected networks, in which all processors are connected.",30213
facial chosen twelve The facial extracted features. from distances are,The facial distances chosen are extracted from twelve facial features.,30214
Then Support distances using the facial are and Neural Machine (SVM) Network Vector (NN). trained,Then the facial distances are trained using Support Vector Machine (SVM) and Neural Network (NN).,30215
has years. models data increasing attracted deep recent to learn attention Exploiting in synthetic,Exploiting synthetic data to learn deep models has attracted increasing attention in recent years.,30216
"create complex By themselves, hierarchies can that behavior. we composing recursively display skills with","By recursively composing skills with themselves, we can create hierarchies that display complex behavior.",30217
greatly the the performance predictive can of methods. choices influence These,These choices can greatly influence the predictive performance of the methods.,30218
input the analysis Gaussian does not The method signal. require,The analysis method does not require the Gaussian input signal.,30219
theoretical the various our of demonstrate in conditions analysis. effectiveness results Simulation,Simulation results in various conditions demonstrate the effectiveness of our theoretical analysis.,30220
cases Many of recent framework. works as our specific seen be can general,Many recent works can be seen as specific cases of our general framework.,30221
datasets. these We each art of on state the of performance demonstrate,We demonstrate state of the art performance on each of these datasets.,30222
system challenging footage. hand outperforms tracking current the on state-of-the-art RGB-only our that demonstrate We,We demonstrate that our hand tracking system outperforms the current state-of-the-art on challenging RGB-only footage.,30223
initial reference. as crash Florida The used Tesla an is being,The Florida Tesla crash is being used as an initial reference.,30224
designed SfSNet rendering a is to lambertian reflect physical model.,SfSNet is designed to reflect a physical lambertian rendering model.,30225
real synthetic a SfSNet labeled world unlabeled learns mixture and images. of from,SfSNet learns from a mixture of labeled synthetic and unlabeled real world images.,30226
predict is used This along lighting. original image to with the,This is used along with the original image to predict lighting.,30227
in central controller The strategy. centralized is used,The central controller is used in centralized strategy.,30228
three used This is into measure categories pixels to categorize i.e.,This measure is used to categorize pixels into three categories i.e.,30229
"depending which dominate. and the double-bounce on mechanisms volume, scattering above of odd-bounce,","odd-bounce, double-bounce and volume, depending on which of the above scattering mechanisms dominate.",30230
is based Wishart category. for the used on iteratively complex each distribution,based on the complex Wishart distribution is iteratively used for each category.,30231
thus are scattering preserved in mechanisms classification Dominant scheme. this,Dominant scattering mechanisms are thus preserved in this classification scheme.,30232
"low object detector, e.g. IoU with trained An threshold,","An object detector, trained with low IoU threshold, e.g.",30233
"degrade performance However, the detection to IoU with increasing thresholds. tends","However, detection performance tends to degrade with increasing the IoU thresholds.",30234
"Cascade object to detection proposed is address R-CNN, these architecture, the A multi-stage problems.","A multi-stage object detection architecture, the Cascade R-CNN, is proposed to address these problems.",30235
"module Feature main Reuse), our (Gated GFR call We advantages. which two exhibits","We call our module GFR (Gated Feature Reuse), which exhibits two main advantages.",30236
"result, of effective sizes detecting a model our objects. As diverse in is more","As a result, our model is more effective in detecting diverse sizes of objects.",30237
face of Recent witnessed detection results deep promising learning. using have years,Recent years have witnessed promising results of face detection using deep learning.,30238
train the to propose model. Loss effectively further Hierarchical FANet a We,We further propose a Hierarchical Loss to effectively train the FANet model.,30239
VGA-resolution time on for Our images real can in run GPU. detector,Our detector can run in real time for VGA-resolution images on GPU.,30240
"The problems significant especially have of daunting, for which complexity. is uncertainty under decision-making task","The task of decision-making under uncertainty is daunting, especially for problems which have significant complexity.",30241
is modality diagnosis for SD-OCT used macular non-invasive of a defects. imaging cross-sectional,SD-OCT is a non-invasive cross-sectional imaging modality used for diagnosis of macular defects.,30242
cancer in special cells received The attention years. of plasticity recent has phenotypic,The phenotypic plasticity of cancer cells has received special attention in recent years.,30243
simulations of model Extensive the our and demonstrate validity algorithm.,Extensive simulations demonstrate the validity of our model and algorithm.,30244
"a images introduce this In fruits. we dataset new, of containing high-quality, paper","In this paper we introduce a new, high-quality, dataset of images containing fruits.",30245
the We typographic transfer (ex. to both seek stylization,We seek to transfer both the typographic stylization (ex.,30246
as ears) textual serifs as the and (ex. well stylization,serifs and ears) as well as the textual stylization (ex.,30247
each may these with inconsistent be classifiers Predictions other. of,Predictions of these classifiers may be inconsistent with each other.,30248
"effectively for the D-Net augmenting it data hard-to-predict Hence, cases. training with is","Hence, it is effectively augmenting the training data for D-Net with hard-to-predict cases.",30249
"harder it larger those becomes to implement derive model. However, of a and scale","However, it becomes harder to derive and implement those of a larger scale model.",30250
"require methods but not do are calibrated accurate. training they Statistics-based images, less","Statistics-based methods do not require calibrated training images, but they are less accurate.",30251
In many learning-based methods. accuracy terms statistics-based method the proposed outperforms and all of,In terms of accuracy the proposed method outperforms all statistics-based and many learning-based methods.,30252
inter-camera enables for unsupervised This constancy. color effectively learning,This effectively enables inter-camera unsupervised learning for color constancy.,30253
semantic the embedding learning Most methods consider problem as a zero-shot visual existing one.,Most existing zero-shot learning methods consider the problem as a visual semantic embedding one.,30254
"added a learning zero-shot converted is data, problem. classification pseudo traditional With naturally to","With added pseudo data, zero-shot learning is naturally converted to a traditional classification problem.",30255
propose nine-layer datasets. leaf for a Flavia CNN identification Foliage using and the famous We,We propose a nine-layer CNN for leaf identification using the famous Flavia and Foliage datasets.,30256
CNNs of Usually requires datasets for the supervised huge training. deep learning,Usually the supervised learning of deep CNNs requires huge datasets for training.,30257
"datasets plant the few only However, species. a per contain examples used","However, the used datasets contain only a few examples per plant species.",30258
"transfer we to data and augmentation our learning overfitting. Therefore, apply from network prevent","Therefore, we apply data augmentation and transfer learning to prevent our network from overfitting.",30259
for concept are A cycle-consistent transfer generative adversarial well-known models. domain unsupervised,A well-known concept for unsupervised domain transfer are cycle-consistent generative adversarial models.,30260
"the generators Unfortunately, resulting change often the of the scene. structure","Unfortunately, the resulting generators often change the structure of the scene.",30261
the in undesirable change semantics. This causes an,This causes an undesirable change in the semantics.,30262
three to ways on We the with depending this type of propose degradation. problem cope,We propose three ways to cope with this problem depending on the type of degradation.,30263
synthesis Face facial on a conditioned photo-sketch given at sketch/photo a generating photo/sketch. aims,Face photo-sketch synthesis aims at generating a facial sketch/photo conditioned on a given photo/sketch.,30264
law of including is applications and enforcement. It wide digital entertainment,It is of wide applications including digital entertainment and law enforcement.,30265
"Specially, for novel face generative a synthesis. photo-sketch adversarial propose (CA-GAN) network we composition-aided","Specially, we propose a novel composition-aided generative adversarial network (CA-GAN) for face photo-sketch synthesis.",30266
"Finally, compelling add and stacked defects (SCA-GAN) we to details. rectify CA-GANs further use","Finally, we use stacked CA-GANs (SCA-GAN) to further rectify defects and add compelling details.",30267
"Besides, is the proposed method considerable ability. that we demonstrate of generalization","Besides, we demonstrate that the proposed method is of considerable generalization ability.",30268
is and Identifying variants connecting phenotype. for Motivation: essential genotype step an genomic,Motivation: Identifying genomic variants is an essential step for connecting genotype and phenotype.,30269
approach of consists The of usual reads. from inference variants alignments statistical sequencing of,The usual approach consists of statistical inference of variants from alignments of sequencing reads.,30270
different can with callers wide variant a range of variant types resolve high State-of-the-art accuracy.,State-of-the-art variant callers can resolve a wide range of different variant types with high accuracy.,30271
"calling. interest for there variant methods Therefore, in is developing resource-efficient","Therefore, there is interest in developing methods for resource-efficient variant calling.",30272
provides alternative memory-efficient calling. usual variant a the fast and Ococo to,Ococo provides a fast and memory-efficient alternative to the usual variant calling.,30273
"of the aim two to this the the is best exploit worlds. In paper,","In this paper, the aim is to exploit the best of the two worlds.",30274
The evaluation the that results outperforms existing model new methods. demonstrate,The evaluation results demonstrate that the new model outperforms existing methods.,30275
proposal This leveraging avoids the tube of pixel-wise by generation segmentation. approach effectively bottom-up annotations,This bottom-up approach effectively avoids tube proposal generation by leveraging the pixel-wise annotations of segmentation.,30276
shot propose named Enriched a detection novel (DES). We with network Semantics Detection single object,We propose a novel single shot object detection network named Detection with Enriched Semantics (DES).,30277
"annotation is no The weak segmentation segmentation supervised branch by required. i.e., extra is ground-truth,","The segmentation branch is supervised by weak segmentation ground-truth, i.e., no extra annotation is required.",30278
common detection compared with favorably performs methods. STIP description and model Our,Our model performs favorably compared with common STIP detection and description methods.,30279
"), video. to we network a directly deep the compressed propose train on","), we propose to train a deep network directly on the compressed video.",30280
"we density, higher to training be the has a information representation found easier. and This","This representation has a higher information density, and we found the training to be easier.",30281
"information. addition, albeit signals In free, provide video motion a noisy, compressed the in","In addition, the signals in a compressed video provide free, albeit noisy, motion information.",30282
techniques them to We effectively. propose use novel,We propose novel techniques to use them effectively.,30283
Success communication adapt for and to to requires parties learn other. engage each both,Success communication requires both parties to engage and learn to adapt for each other.,30284
"a training, are the interactive from common function. guidance During reinforced reward by agents both","During interactive training, both agents are reinforced by the guidance from a common reward function.",30285
"grounding the rate. task, visual success significantly improve and task","visual grounding task, and significantly improve the task success rate.",30286
neighbor paper the search. compact studies approximate to approach nearest coding This,This paper studies the compact coding approach to approximate nearest neighbor search.,30287
quantization. The approach is composite near-orthogonal called resulting,The resulting approach is called near-orthogonal composite quantization.,30288
approach the datasets. efficacy proposed the benchmark of We show several empirically over,We empirically show the efficacy of the proposed approach over several benchmark datasets.,30289
of the problem in unconstrained videos. moving objects We study segmenting,We study the problem of segmenting moving objects in unconstrained videos.,30290
an explicit memory two-stream The is with a module. framework neural network,The framework is a two-stream neural network with an explicit memory module.,30291
"SegTrack. and dataset DAVIS, on segmentation our extensively motion We Freiburg-Berkeley three method evaluate benchmarks,","We evaluate our method extensively on three benchmarks, DAVIS, Freiburg-Berkeley motion segmentation dataset and SegTrack.",30292
"have and However, approximate not real rank tensor. of they a properly limitations the may","However, they have limitations and may not properly approximate the real rank of a tensor.",30293
"optimization. use Besides, the do low-rank they in not property explicitly","Besides, they do not explicitly use the low-rank property in optimization.",30294
"our efficacy Beneficial the from TNN, of the of completion. tensor approach rankness low improves","Beneficial from the low rankness of TNN, our approach improves the efficacy of tensor completion.",30295
"detections optical segmentation, person etc. flow, keypoint","person segmentation, optical flow, keypoint detections etc.",30296
local susceptible models Optimization are to minima.,Optimization models are susceptible to local minima.,30297
"learning based work, camera input. a motion for this model propose In we capture single","In this work, we propose a learning based motion capture model for single camera input.",30298
of image-to-image framework conditional the generative SSNet based designed adversarial (cGAN). was on networks,SSNet was designed based on the framework of image-to-image conditional generative adversarial networks (cGAN).,30299
"Philosophy title was full Topics and The essay's in Computer ""Some Science"".","The essay's full title was ""Some Topics in Philosophy and Computer Science"".",30300
"combine Often natural events, two to jointly during modalities affect correlated perception. these human","Often correlated during natural events, these two modalities combine to jointly affect human perception.",30301
"the generating we pose task sound paper, input. In visual given this of","In this paper, we pose the task of generating sound given visual input.",30302
"to we generated. models the Finally, discuss measures of quality predict","Finally, we discuss measures to predict the quality of models generated.",30303
are intense under importance found their of scientific amyloids clinical Because research.,Because of their clinical importance amyloids are found under intense scientific research.,30304
sequences. amyloidogenic studies the focus on Most on of the of amyloids role,Most of the studies on amyloids focus on the role of amyloidogenic sequences.,30305
the lower Interval error and limit were used. analysis,Interval analysis and the lower error limit were used.,30306
show of efficiency examples the numerical Four proposal. the,Four numerical examples show the efficiency of the proposal.,30307
"designed model on based a AlexNet. system,and sorting gangue convolutional built network neural a We","We designed a gangue sorting system,and built a convolutional neural network model based on AlexNet.",30308
zooming and Interpolation image based a speed complexity. computational provide methods execution high low,Interpolation based image zooming methods provide a high execution speed and low computational complexity.,30309
"the quality zoomed is the cases. unsatisfactory However, images in many of","However, the quality of the zoomed images is unsatisfactory in many cases.",30310
of methods resolution create the super- image. is to to details main The challenge new,The main challenge of super- resolution methods is to create new details to the image.,30311
algorithm a a This strategy. zoom-out-zoom-in details paper new to create using new proposes,This paper proposes a new algorithm to create new details using a zoom-out-zoom-in strategy.,30312
the This to the estimated final adding image. strategy by permits reducing effects error blurring,This strategy permits reducing blurring effects by adding the estimated error to the final image.,30313
images confirm for pleasing results. Experimental algorithm's the natural to visually create ability results,Experimental results for natural images confirm the algorithm's ability to create visually pleasing results.,30314
"clear the However, of is consequences lacking. of a understanding population-level co-infection","However, a clear understanding of the population-level consequences of co-infection is lacking.",30315
argues fate. paper explainable risks that AI similar This a,This paper argues that explainable AI risks a similar fate.,30316
into the boosting A process. training is integrated network framework during proposed our,A boosting framework is integrated into our proposed network during the training process.,30317
in is precision point simulation of presented. IEEE standard An on function recursive floating improvement,An improvement on precision of recursive function simulation in IEEE floating point standard is presented.,30318
to are presented studies Numerical the method. illustrate,Numerical studies are presented to illustrate the method.,30319
performance. convolutional excellent up-sampling achieve networks Deep neural image,Deep convolutional neural networks achieve excellent image up-sampling performance.,30320
"methods high-resolution (e.g. depending to CNN-based traditional However, restore on highly interpolations tend results","However, CNN-based methods tend to restore high-resolution results highly depending on traditional interpolations (e.g.",30321
"is Moreover, with image compatible to existing DSN improve compression. codecs image","Moreover, DSN is compatible with existing image codecs to improve image compression.",30322
Image high-level orientation scene understanding. detection requires,Image orientation detection requires high-level scene understanding.,30323
use contextual to information and Humans recognition orient images. correctly scene object,Humans use object recognition and contextual scene information to correctly orient images.,30324
for topic new by survival approach modeling We supervised a introduce is analysis. that,We introduce a new approach for topic modeling that is supervised by survival analysis.,30325
"flying frames. a consecutive image track To we between utilize information the small residual drone,","To track a small flying drone, we utilize the residual information between consecutive image frames.",30326
directed core of At meta-layer designed method is the a our loop-embedded structure. with graph,At the core of our method is a meta-layer designed with loop-embedded directed graph structure.,30327
"ImageNet, large-scale performance competitive we superior to achieve existing counterparts. On","On large-scale ImageNet, we achieve competitive performance superior to existing counterparts.",30328
"paper, we In to a steps of all this architecture matching. propose incorporate stereo network","In this paper, we propose a network architecture to incorporate all steps of stereo matching.",30329
The first features. the calculates multi-scale part shared,The first part calculates the multi-scale shared features.,30330
has KITTI evaluated datasets. proposed the on Scene Flow been method and The,The proposed method has been evaluated on the Scene Flow and KITTI datasets.,30331
novo challenge. is of protein scientific folding De open an prediction,De novo prediction of protein folding is an open scientific challenge.,30332
a transformation. poorly applying that were tuned linear proteins of preserve by native HSs structures,HSs that poorly preserve native structures of proteins were tuned by applying a linear transformation.,30333
visual awareness correlated argue attention that with fixations We of always or not objects. are,We argue that fixations are not always correlated with attention or awareness of visual objects.,30334
during recall fixations to during qualitatively the The only exploration. fixations of related are locations,The locations of fixations during recall are only qualitatively related to the fixations during exploration.,30335
"for measure importance sense, elements they In of provide human of a this observers. visual","In this sense, they provide a measure of importance of visual elements for human observers.",30336
Lifestyle of Vlogs We surprisingly large interaction use data. as Internet source the diverse and,We use Internet Lifestyle Vlogs as the source of surprisingly large and diverse interaction data.,30337
"data. biases gathered Additionally, data built exposes common our into explicitly","Additionally, our data exposes biases built into common explicitly gathered data.",30338
our of hands. component We of by central interaction data the sense -- analyzing make,We make sense of our data by analyzing the central component of interaction -- hands.,30339
prediction We hands. demonstrate additionally future of,We additionally demonstrate future prediction of hands.,30340
for of location-based development services (LBS). the localization The technology indoor is important,The localization technology is important for the development of indoor location-based services (LBS).,30341
radio fingerprint-based one is The frequency (RF) approaches. the localization most of promising,The radio frequency (RF) fingerprint-based localization is one of the most promising approaches.,30342
"approach. localization we efficient robust these indoor an challenges, To and two address propose","To address these two challenges, we propose an efficient and robust indoor localization approach.",30343
recent years. great progress detection been made Scene has in text,Scene text detection has been made great progress in recent years.,30344
to The quadrangle. rectangle rectangle evolving further manners rotated and to detection axis-aligned are from,The detection manners are evolving from axis-aligned rectangle to rotated rectangle and further to quadrangle.,30345
cue-specific end-to-end. model learns language-vision We a that network develop novel embedding reciprocity this,We develop a novel cue-specific language-vision embedding network that learns this reciprocity model end-to-end.,30346
"we the In this approximation package. the \textbf{sinib} implemented in saddlepoint paper,","In this paper, we implemented the saddlepoint approximation in the \textbf{sinib} package.",30347
to two provide usage. We its examples illustrate,We provide two examples to illustrate its usage.,30348
example real-world One simulated uses while healthcare other the uses data. data,One example uses simulated data while the other uses real-world healthcare data.,30349
jointly camera. taken a a present denoising technique bursts We of for images from handheld,We present a technique for jointly denoising bursts of images taken from a handheld camera.,30350
the the defining error. of the bivariate glyphs of We introduce by cause corresponding inefficiency,We introduce the cause of the inefficiency of bivariate glyphs by defining the corresponding error.,30351
be done instance-level at feature-level. Our and method can,Our method can be done at instance-level and feature-level.,30352
series of tasks learning in Deep vision. computer a achieved substantial success in has,Deep learning has achieved substantial success in a series of tasks in computer vision.,30353
"of enable in To the is deep and context standardization interoperability coding, feature urgent important.","To enable interoperability in the context of deep feature coding, standardization is urgent and important.",30354
"pixel truth. document This pixel multi-labeled and at tool evaluates ground support segmentation level,","This tool evaluates document segmentation at pixel level, and support multi-labeled pixel ground truth.",30355
"gender images lot Recently, gained of facial from importance. has of a recognition","Recently, recognition of gender from facial images has gained a lot of importance.",30356
"observed proposed method Through existing outperforms experiments, approaches. we the have that,","Through experiments, we have observed that, the proposed method outperforms existing approaches.",30357
"and In inefficient. computational complex contrast, state-of-the-art models saliency are many","In contrast, many state-of-the-art computational saliency models are complex and inefficient.",30358
grayscale preserve Previous that low-resolution color information. images and high-resolution saliency showed studies,Previous studies showed that low-resolution color and high-resolution grayscale images preserve saliency information.,30359
such that limitation as is expect approaches observed sequences inputs. One action,One limitation is that such approaches expect observed action sequences as inputs.,30360
of distributions. input action such observed be sequences will then The,The input will then be sequences of such observed action distributions.,30361
"large research handling remains open significant motions an effort, problem. Despite","Despite significant research effort, handling large motions remains an open problem.",30362
through The and kernel value (SVD). a decomposition then image singular recovered can be,The kernel and image can then be recovered through a singular value decomposition (SVD).,30363
This quantitative measures. and is both confirmed visually through,This is confirmed both visually and through quantitative measures.,30364
of curvilinear applications. structures is problem an practical with important Delineation multiple Computer in Vision,Delineation of curvilinear structures is an important problem in Computer Vision with multiple practical applications.,30365
people. multiple activity forecasting an human propose of We approach complex video for involving,We propose an approach for forecasting video of complex human activity involving multiple people.,30366
the prediction handle too complex Direct in variability pixel-level activities. appearance simple to is,Direct pixel-level prediction is too simple to handle the appearance variability in complex activities.,30367
among interactions a captures hierarchical adopting group-based mechanism. dynamic model by interaction Our people,Our hierarchical model captures interactions among people by adopting a dynamic group-based interaction mechanism.,30368
"neural Finally, complete the networks placed to rendering. these are in filters encoder-decoder","Finally, these filters are placed in encoder-decoder neural networks to complete the rendering.",30369
"continuous present DeepCache, design principled We mobile in learning a for inference cache deep vision.","We present DeepCache, a principled cache design for deep learning inference in continuous mobile vision.",30370
temporal locality streams. input video model efficiency by in benefits exploiting execution DeepCache,DeepCache benefits model execution efficiency by exploiting temporal locality in input video streams.,30371
loss placing the layers. %We different study also embedding on network,%We also study placing the embedding loss on different network layers.,30372
Our publicly is GitHub. code on available,Our code is publicly available on GitHub.,30373
"scenarios. an Detection autonomous AVOD, We for present Aggregate network driving View Object","We present AVOD, an Aggregate View Object Detection network for autonomous driving scenarios.",30374
The from characters segmenting a or candidate are words. searched keywords line without,The candidate keywords are searched from a line without segmenting characters or words.,30375
Oriented Histogram Pyramid in word our Gradient of used spotting has (PHOG) feature been framework.,Pyramid Histogram of Oriented Gradient (PHOG) feature has been used in our word spotting framework.,30376
information correlation channels. can different obtain among They color,They can obtain correlation information among different color channels.,30377
"is performance However, different conditions. in unstable their","However, their performance is unstable in different conditions.",30378
"likelihood African population glaucoma, and American orbital disease. increased inflammation, of thyroid has","African American population has increased likelihood of glaucoma, orbital inflammation, and thyroid disease.",30379
filter is via least the The Wiener mean solution. filter often (LMS) derived squares,The least mean squares (LMS) filter is often derived via the Wiener filter solution.,30380
formulate on LMS Based Knowledge-Aided filter. a Knowledge-Aided we Kaczmarz this,Based on this Knowledge-Aided Kaczmarz we formulate a Knowledge-Aided LMS filter.,30381
their the algorithms and on analyze show of We the simulation convergence results performance.,We analyze the convergence of the algorithms and show simulation results on their performance.,30382
"rule, As a data are mandatory companies. for analytics general now","As a general rule, data analytics are now mandatory for companies.",30383
combined frame-level Super-event are CNNs per-segment per-frame or annotations. representations provide with to,Super-event representations are combined with per-frame or per-segment CNNs to provide frame-level annotations.,30384
Translation spectacularly achieves developments Image-to-Image Unsupervised advanced nowadays.,Unsupervised Image-to-Image Translation achieves spectacularly advanced developments nowadays.,30385
corresponding representations of by domains' decoders. combination different is Linear fusing also the obtained,Linear combination of different domains' representations is also obtained by fusing the corresponding decoders.,30386
for This of geometry. homotopy continuation-based methods the allows use from algebraic numerical,This allows for the use of homotopy continuation-based methods from numerical algebraic geometry.,30387
discriminant and together. integrates This learning sparsity model,This learning model integrates discriminant and sparsity together.,30388
"we called on sparse model, this a (LDSR). propose discriminant then representation locality-based Based classifier","Based on this model, we then propose a classifier called locality-based discriminant sparse representation (LDSR).",30389
end-to-end an strongly training and enables learning time. reduces the This,This enables an end-to-end learning and strongly reduces the training time.,30390
"This ill-posed from often benefit leads can inverse strategy regularization to strategies. which problems, spatial","This strategy often leads to ill-posed inverse problems, which can benefit from spatial regularization strategies.",30391
"proposed. multiscale In sparse paper, novel a regularization approach is spatial for unmixing this","In this paper, a novel multiscale spatial regularization approach for sparse unmixing is proposed.",30392
evaluates problems. A proposed cross-modality suite domain that data new adaptation comprehensively is,A new data suite is proposed that comprehensively evaluates cross-modality domain adaptation problems.,30393
on benchmark state two datasets. Deep art performance achieves of the Magnet,Deep Magnet achieves state of the art performance on two benchmark datasets.,30394
"these fine-grained not classification, For position-sensitive are needed. filters","For fine-grained classification, these position-sensitive filters are not needed.",30395
occlusion. degrees of moderate to robust are Recent RGB-D-based methods,Recent RGB-D-based methods are robust to moderate degrees of occlusion.,30396
"inputs, for objects. partly method well works occluded For no previous RGB","For RGB inputs, no previous method works well for partly occluded objects.",30397
arising The paper practical present work. some this explains insights from,The present paper explains some practical insights arising from this work.,30398
"sparse study a worked factorization Poisson example. Finally, exchangeable matrix we as","Finally, we study sparse exchangeable Poisson matrix factorization as a worked example.",30399
blendshapes can learn for powerful very It also behaviour. be parametrising to facial used,It can also be used to learn very powerful blendshapes for parametrising facial behaviour.,30400
The purposes. will publicly available made for database research be,The database will be made publicly available for research purposes.,30401
"are Furthermore, test the analytical several functions employed to CNN metamodel. also high-dimensional","Furthermore, several high-dimensional analytical functions are also employed to test the CNN metamodel.",30402
"is (IGA)-based successfully. the applied also to Moreover, optimization CNN metamodel proposed IsoGeometric Analysis","Moreover, the proposed CNN metamodel is also applied to IsoGeometric Analysis (IGA)-based optimization successfully.",30403
and data methods in can remove scenarios. State-of-the-art on controlled synthetic reflections,State-of-the-art methods can remove reflections on synthetic data and in controlled scenarios.,30404
"not generalize on based well real-world they strong and are images. do assumptions However, to","However, they are based on strong assumptions and do not generalize well to real-world images.",30405
"used. common Contrary real-world images challenging are information is misconception, when to even polarization a","Contrary to a common misconception, real-world images are challenging even when polarization information is used.",30406
"challenging. However,the large acquisition supervised dataset often of is a","However,the acquisition of a large supervised dataset is often challenging.",30407
"traffic sign also for applications, recognition. intelligent This transportation i.e., true is","This is also true for intelligent transportation applications, i.e., traffic sign recognition.",30408
"images. unlike other official Traffic signs, have objects,","Traffic signs, unlike other objects, have official images.",30409
real using relationship synthetic domains. co-domain embedding perform and from a We quadruple,We perform co-domain embedding using a quadruple relationship from real and synthetic domains.,30410
from Our explicit different relationships domains. among similarity pairwise quadruplet the fully samples utilizes network,Our quadruplet network fully utilizes the explicit pairwise similarity relationships among samples from different domains.,30411
step problems. is an important classification Feature for preprocessing selection,Feature selection is an important preprocessing step for classification problems.,30412
optimal It the dataset. features in with deals selecting original near,It deals with selecting near optimal features in the original dataset.,30413
"methods. can so problem, efficient NP-hard is an than exact meta-heuristics more be selection Feature","Feature selection is an NP-hard problem, so meta-heuristics can be more efficient than exact methods.",30414
shutter with rolling budget Modern are equipped vehicles often automotive commercial camera. a,Modern automotive vehicles are often equipped with a budget commercial rolling shutter camera.,30415
efficiently motion minimal rolling shutter camera The the proposed and estimates solver accurately.,The proposed minimal solver estimates the rolling shutter camera motion efficiently and accurately.,30416
"available be the may of not retraining. time previously training samples seen at Also,","Also, previously seen training samples may not be available at the time of retraining.",30417
several applications. We proposed recognition on the evaluated scheme,We evaluated the proposed scheme on several recognition applications.,30418
The of paper main this contribution twofold. is,The main contribution of this paper is twofold.,30419
"matching framework of proposed leads consistent candidates. them to movements walkers, The random and synchronizes","The proposed framework synchronizes movements of random walkers, and leads them to consistent matching candidates.",30420
(BO) is for model-based gradient-free black-box a Bayesian function approach optimization. optimization,Bayesian optimization (BO) is a model-based approach for gradient-free black-box function optimization.,30421
"definitions and exact often maps, application-specific different are The scenarios for of (e.g. hand-crafted however,","The exact definitions of maps, however, are often application-specific and hand-crafted for different scenarios (e.g.",30422
"challenging In sign this paper, of robustness we the under investigate conditions. algorithms recognition traffic","In this paper, we investigate the robustness of traffic sign recognition algorithms under challenging conditions.",30423
"tracking skeleton depth real-time stream-based proposed. generic, is method a accurate for Finally, human multiple","Finally, a generic, multiple depth stream-based method for accurate real-time human skeleton tracking is proposed.",30424
classification is rapidly extremely which of high capable and identifying performing samples test accuracy.,which is capable of identifying test samples extremely rapidly and performing high classification accuracy.,30425
"those practice not linearly However, be in separable. may samples","However, in practice those samples may not be linearly separable.",30426
linearly are can non-separable. results given especially promising achieve samples the KBTC when,KBTC can achieve promising results especially when the given samples are linearly non-separable.,30427
"SICs, we which develop procedures cross By do estimation require not parameter using validation. any","By using SICs, we develop parameter estimation procedures which do not require any cross validation.",30428
a introducing algorithm. regularising This Class (CRL) Rectification possible by Loss made is,This is made possible by introducing a Class Rectification Loss (CRL) regularising algorithm.,30429
"work, method In present for adaptation. a domain unsupervised this we","In this work, we present a method for unsupervised domain adaptation.",30430
"features boundaries. can class Therefore, ambiguous generator generate trained a near","Therefore, a trained generator can generate ambiguous features near class boundaries.",30431
feature features generator to target the support the to generate near A minimize discrepancy. learns,A feature generator learns to generate target features near the support to minimize the discrepancy.,30432
image and other on datasets Our of segmentation. outperforms methods several semantic classification method,Our method outperforms other methods on several datasets of image classification and semantic segmentation.,30433
morphogenesis Understanding fate cell remains a patterning embryo the challenge. in formidable mammalian and,Understanding cell fate patterning and morphogenesis in the mammalian embryo remains a formidable challenge.,30434
"We quantification, CAC. intermediate for circumventing an method CAC automatic segmentation propose of","We propose an automatic method for CAC quantification, circumventing intermediate segmentation of CAC.",30435
The the determines ConvNet box a method using for a bounding heart around localization.,The method determines a bounding box around the heart using a ConvNet for localization.,30436
"section, our Introduction, design. first In experimental present we the","In the first section, Introduction, we present our experimental design.",30437
We a defense denoiser classification. high-level image propose representation guided as (HGD) for,We propose high-level representation guided denoiser (HGD) as a defense for image classification.,30438
"guiding to than defend the can HGD be one models transferred it. Third, other","Third, HGD can be transferred to defend models other than the one guiding it.",30439
resolution face both improves facial landmarks. the detects and,both improves face resolution and detects the facial landmarks.,30440
by We in super-resolution proposing new a residual-based (c) the state-of-the-art improve architecture. face upon,(c) We improve upon the state-of-the-art in face super-resolution by proposing a new residual-based architecture.,30441
"improvement show we state-of-the-art over face Quantitatively, for the alignment. both and (d) super-resolution large","(d) Quantitatively, we show large improvement over the state-of-the-art for both face super-resolution and alignment.",30442
"good real-world low results we on Qualitatively, time resolution images. first the (e) for show","(e) Qualitatively, we show for the first time good results on real-world low resolution images.",30443
possesses CIMM structure features that two are task. for The the noise distinctive important removal,The CIMM structure possesses two distinctive features that are important for the noise removal task.,30444
for and conditioned a of question. series open refrigerators) actions on plan the,open refrigerators) and plan for a series of actions conditioned on the question.,30445
is items possible. to as quickly goal identify good The as N,The goal is to identify N good items as quickly as possible.,30446
"found. and good to group until the next move continue are exact Then, items N","Then, move to the next group and continue until exact N good items are found.",30447
"solution In finite under A is provided. policy optimal N note, for this an","In this note, an optimal solution under policy A for finite N is provided.",30448
Dynamic Shur-convexity programming; problem; . Optimal Partition design; Keywords:,Keywords: Dynamic programming; Optimal design; Partition problem; Shur-convexity .,30449
similarity We property include functions. our objective in also preserving a,We also include a similarity preserving property in our objective functions.,30450
"difficult optimizations are and binary, involving independence, constraints solve. balance these resulting to The","The resulting optimizations involving these binary, independence, and balance constraints are difficult to solve.",30451
data business information. economic and Web are of a search valuable source,Web search data are a valuable source of business and economic information.,30452
economic Previous data search utilized web Google Trends forecasting. have studies for,Previous studies have utilized Google Trends web search data for economic forecasting.,30453
Python software Experiments in conducted an written open-source are and using C++.,Experiments are conducted using an open-source software written in Python and C++.,30454
curvature}. this \emph{signed-control We the notion Ricci call,We call this notion the \emph{signed-control Ricci curvature}.,30455
quality intelligence artificial the aspects strongly including Safety systems of require explainability. critical,Safety critical systems strongly require the quality aspects of artificial intelligence including explainability.,30456
have image functions wide-spread in computer processing vision. Gabor applications and,Gabor functions have wide-spread applications in image processing and computer vision.,30457
"loss we a the regularize better to transitive achieve introduce network. To this, consistency","To achieve this, we introduce a transitive consistency loss to better regularize the network.",30458
results Experimental state-of-the-art method performs have shown against that favorably methods. our,Experimental results have shown that our method performs favorably against state-of-the-art methods.,30459
"involve problems, general MCMC across For chains. state-of-the-art convergence diagnostics comparisons multiple","For general MCMC problems, state-of-the-art convergence diagnostics involve comparisons across multiple chains.",30460
"discrete-space Thus, here specifically tailored MCMC. single-chain convergence diagnostic we to propose a","Thus, we propose here a single-chain convergence diagnostic specifically tailored to discrete-space MCMC.",30461
"determined. density, were proliferation, cell parameters viability, and cell Extracellular topology concentration, ultrastructure matrix","Extracellular matrix concentration, cell density, proliferation, cell viability, topology and ultrastructure parameters were determined.",30462
"monolayer less than docetaxel showed and cultures. cell spheroids clearly that epirubicin, sensitive were cisplatin,","cisplatin, docetaxel and epirubicin, showed that spheroids were clearly less sensitive than monolayer cell cultures.",30463
For use a a we with SDA. detection particle coupled LoG filter,For particle detection we use a LoG filter coupled with a SDA.,30464
of rational be in described systems can orthonormal terms basis Most functions.,Most rational systems can be described in terms of orthonormal basis functions.,30465
on to transformation module local alleviate and focuses The feature regionlet selection local learning variations.,The regionlet learning module focuses on local feature selection and transformation to alleviate local variations.,30466
additional Deep Regionlets is efforts. The without trained framework end-to-end,The Deep Regionlets framework is trained end-to-end without additional efforts.,30467
"address paper, the problem In we synthetic this shape-from-shading deep networks training images. with by","In this paper, we address the shape-from-shading problem by training deep networks with synthetic images.",30468
our approach that achieves shape-from-shading performance We on benchmark. state-of-the-art show a,We show that our approach achieves state-of-the-art performance on a shape-from-shading benchmark.,30469
cascaded We detection. regressor eye new center a propose for,We propose a new cascaded regressor for eye center detection.,30470
"accurately the detect alignment Instead, eyes method. more an we using existing feature facial","Instead, we detect the eyes more accurately using an existing facial feature alignment method.",30471
regression powerful of features improve using machinery. by We both robustness localization and advanced the,We improve the robustness of localization by using both advanced features and powerful regression machinery.,30472
automatically capsule wardrobes. We propose to create,We propose to automatically create capsule wardrobes.,30473
We task pose as a selection subset problem. the,We pose the task as a subset selection problem.,30474
often students' accurately. population that responses Results and portrayed sample suggest,Results suggest that students' responses often portrayed sample and population accurately.,30475
"for network super-resolution. present a paper, frequency In we neural image domain this","In this paper, we present a frequency domain neural network for image super-resolution.",30476
show super-resolution the against and on results literature. compare elsewhere alternatives We in,We show results on super-resolution and compare against alternatives elsewhere in the literature.,30477
allows We to state-of-the-art datasets. that it us both show the outperform on,We show that it allows us to outperform the state-of-the-art on both datasets.,30478
"project, adversarial training this (i.e. we improvements In explored for various","In this project, we explored various improvements for adversarial training (i.e.",30479
"publications However, environments. in learning driving mainly recent unusual consider","However, recent publications mainly consider learning in unusual driving environments.",30480
"of four of the verify model, the intelligence. swarm To feasibility we simulate types","To verify the feasibility of the model, we simulate four types of swarm intelligence.",30481
and deformations blurs. image both Turbulence-degraded distorted by are frames space-time-varying turbulent,Turbulence-degraded image frames are distorted by both turbulent deformations and space-time-varying blurs.,30482
Different and are explored. of fidelity regularization choices terms,Different choices of fidelity and regularization terms are explored.,30483
"model. efficacy of out, the carried which demonstrate been have experiments Extensive proposed our","Extensive experiments have been carried out, which demonstrate the efficacy of our proposed model.",30484
"range, camera affected etc. pose and conditions, by Because weather of","Because of affected by weather conditions, camera pose and range, etc.",30485
"Demonstrated benchmarks. and on by FDDB Our experiments, FACE promising achieves network performance WIDER","Demonstrated by experiments, Our network achieves promising performance on WIDER FACE and FDDB benchmarks.",30486
has in recognition. It used face been animation and widely,It has been widely used in face animation and recognition.,30487
model cost Different functions are training explored. for,Different cost functions for training model are explored.,30488
"Segmentation of (before imagery brain CSF and after into i.e. and tissue brain surgery,","Segmentation of brain imagery into brain tissue and CSF (before and after surgery, i.e.",30489
role postop) crucial evaluating a surgical pre-op vs. plays in treatment.,pre-op vs. postop) plays a crucial role in evaluating surgical treatment.,30490
been often pre-op Segmentation and relatively of images is researched. has problem straightforward a well,Segmentation of pre-op images is often a relatively straightforward problem and has been well researched.,30491
CT labeled set scans training employed. pixel identities with of is a,a training set of CT scans with labeled pixel identities is employed.,30492
images. a and procedure segmenting training post-op hydrocephalic CT test customized for,a customized training and test procedure for segmenting post-op hydrocephalic CT images.,30493
in has displayed impressive area progress RGBD work input. the Recent using,Recent work in the area has displayed impressive progress using RGBD input.,30494
"are what need the be And, to challenges that next tackled?","And, what are the next challenges that need to be tackled?",30495
for We input a epistemic novel and uncertainty. predictive derive aleatoric analysis sensitivity of variables,We derive a novel sensitivity analysis of input variables for predictive epistemic and aleatoric uncertainty.,30496
black-box probabilistic models. Our interpretability increases of the method complex,Our method increases the interpretability of complex black-box probabilistic models.,30497
is reconstruction a vision. in task single expression image computer challenging a from Face,Face expression reconstruction from a single image is a challenging task in computer vision.,30498
of from performance training state by the a Initially establish the scratch. art CNN we,Initially we establish the state of the art performance by training a CNN from scratch.,30499
"the the neural domain not adaptation efficient is transfer. Surprisingly, style most","Surprisingly, the most efficient domain adaptation is not the neural style transfer.",30500
"and levels. variations, frequency shifts signal includes This phase","This includes phase variations, frequency shifts and signal levels.",30501
for We practical challenges the opportunities and also real-world discuss scenarios. potential,We also discuss the potential opportunities and practical challenges for real-world scenarios.,30502
"(and based for convolutional) feedforward methods on (CNNs) deep regression. pose popular networks became Recently,","Recently, methods based on deep (and convolutional) feedforward networks (CNNs) became popular for pose regression.",30503
study is The three-fold. of this aim,The aim of this study is three-fold.,30504
discretely a is observed in is X time. that process Suppose multivariate diffusion,Suppose X is a multivariate diffusion process that is observed discretely in time.,30505
"the of consistent consists recovering of problem process, path smoothing the observations. with the The","The smoothing problem consists of recovering the path of the process, consistent with the observations.",30506
the Filtering called The is algorithm algorithm. Guiding (BFFG) Forward resulting Backward,The resulting algorithm is called the Backward Filtering Forward Guiding (BFFG) algorithm.,30507
algorithm the estimation. to extend We parameter include,We extend the algorithm to include parameter estimation.,30508
relies al. proposed Schauer guided The method et on in introduced proposals,The proposed method relies on guided proposals introduced in Schauer et al.,30509
We efficiency in a of challenging problems. illustrate its number,We illustrate its efficiency in a number of challenging problems.,30510
the constructs video target are image. sequence acquired of spatial generated The a from,The spatial constructs of a generated video sequence are acquired from the target image.,30511
feature. is dynamics suppressed using achieved That appearance the proposed,That is achieved using the proposed appearance suppressed dynamics feature.,30512
about and the ecology literature reviewed behavior. We and is synthesized what known their,We reviewed the literature and synthesized what is known about their ecology and behavior.,30513
video on video our multiple validate thousands of and idea popular clips We codecs.,We validate our idea on thousands of video clips and multiple popular video codecs.,30514
"cross loss weighted the make entropy to a is designed effective. more training Moreover,","Moreover, a weighted cross entropy loss is designed to make the training more effective.",30515
any expressions world the previous knowledge. than changed of Computation has more,Computation has changed the world more than any previous expressions of knowledge.,30516
"with important, Pixelwise applications. is challenging, an yet semantic many labeling image task","Pixelwise semantic image labeling is an important, yet challenging, task with many applications.",30517
both quantitative generated provide the a of analysis We also and results. qualitative,We also provide both a quantitative and qualitative analysis of the generated results.,30518
been consciousness as Problem Hard illusion. of has dismissed an The,The Hard Problem of consciousness has been dismissed as an illusion.,30519
underwater vehicles have exploration. (AUVs) been Autonomous deployed for underwater,Autonomous underwater vehicles (AUVs) have been deployed for underwater exploration.,30520
"battery its potential limited its However, storage on-board data energy is and capacity. by confined","However, its potential is confined by its limited on-board battery energy and data storage capacity.",30521
"docking systems. In a vision we these propose for based following work, underwater framework this","In this work, we propose a vision based framework for underwater docking following these systems.",30522
"algorithm accurate integrated For perspective-n-point is into pose a our estimation, framework.","For accurate pose estimation, a perspective-n-point algorithm is integrated into our framework.",30523
"the of knowledge, To first UDID is docking underwater best publicly available dataset. our the","To the best of our knowledge, UDID is the first publicly available underwater docking dataset.",30524
"in we examine by module Then, our the water underwater estimation experimental experiments pool. pose","Then, we examine the pose estimation module by underwater experiments in our experimental water pool.",30525
decoder attention with by Then model. coverage-based GRU-RNN implemented the the a also is,Then the decoder is also implemented by the GRU-RNN with a coverage-based attention model.,30526
"the seven of power the On approach. datasets, we demonstrate","On seven datasets, we demonstrate the power of the approach.",30527
"example, For even includes this all link functions.","For example, this includes all even link functions.",30528
alternative and an by recovery We this proposing convex analyzing resolve issue method.,We resolve this issue by proposing and analyzing an alternative convex recovery method.,30529
"existing step. extraction feature detection a include algorithms Most preprocessing and step,","Most existing algorithms include a preprocessing step, feature extraction and detection step.",30530
The frictions latter distortions is semi-parametrically to to used due macroeconomic in estimate variables.,The latter is used to semi-parametrically estimate distortions due to frictions in macroeconomic variables.,30531
"these models. propose we estimates, test complete for on a novel Based","Based on these estimates, we propose a novel test for complete models.",30532
"than model facts. wrong theoretical magnitude five orders less of about the the Currently, gave","Currently, the wrong theoretical model gave about the five orders of magnitude less than facts.",30533
novel utilize design can Our network these fully advantages.,Our novel network design can fully utilize these advantages.,30534
what But if nodes characteristics? the of change some their,But what if some of the nodes change their characteristics?,30535
also analyzed hypotheses. The different system the of between perspective is decision from,The system is also analyzed from the perspective of decision between different hypotheses.,30536
pathogens throughout must The recognize system immune life. countless,The immune system must recognize countless pathogens throughout life.,30537
effective How the this receptors? does but repertoire system organize finite of,How does this system organize the finite but effective repertoire of receptors?,30538
interacting receptors cross-reactive T-cell more/less High/low amino acids. have strongly,High/low cross-reactive T-cell receptors have more/less strongly interacting amino acids.,30539
Hashing is storage and image retrieval to large-scale to widely retrieval efficiency. due applied the,Hashing is widely applied to large-scale image retrieval due to the storage and retrieval efficiency.,30540
at Most decisions approaches this human traditional require step.,Most traditional approaches require human decisions at this step.,30541
"model the example, removing obtained. or user For until a is a satisfactory adding variable","For example, the user adding or removing a variable until a satisfactory model is obtained.",30542
constraint approximate also efficient to t-test. an for constraint coefficient We the the propose,We also propose an efficient constraint to approximate the constraint for the coefficient t-test.,30543
the We effect. analog no to McGurk found,We found no analog to the McGurk effect.,30544
"is incomplete. to a such self-occlusion, due map Unfortunately, always UV","Unfortunately, due to self-occlusion, such a UV map is always incomplete.",30545
in of skeletal development the performance the method We stages. further of evaluate context,We further evaluate performance of the method in the context of skeletal development stages.,30546
for methods bone assessment. outperforms common age approach other Our,Our approach outperforms other common methods for bone age assessment.,30547
demand supervised training data trained of extensive approaches Existing pathologists. from amount state-of-the-art,Existing state-of-the-art approaches demand extensive amount of supervised training data from trained pathologists.,30548
cancer from test nucleus method types. images segmentation our using We for four,We test our method for nucleus segmentation using images from four cancer types.,30549
paper we ensemble an for In object classification. of propose this deep and features local,In this paper we propose an ensemble of local and deep features for object classification.,30550
out from We distinctive information. also deep characteristic convolutional various find that features encode networks,We also find out that features from various deep convolutional networks encode distinctive characteristic information.,30551
the that this empirically performances. positively system demonstrate impacts We,We empirically demonstrate that this positively impacts the system performances.,30552
"for ImageNet. our estimation, depth supervised even monocular with outperforms Moreover, pre-training pre-training unsupervised method","Moreover, for monocular depth estimation, our unsupervised pre-training method even outperforms supervised pre-training with ImageNet.",30553
"Neural as Networks, exceptionally learning perform well many in machine ConvNets tasks. complex Convolution known","Convolution Neural Networks, known as ConvNets exceptionally perform well in many complex machine learning tasks.",30554
"complexity. main spatial model/computation and exist representation, feature challenges representation, including temporal (image) information Three","Three main challenges exist including spatial (image) feature representation, temporal information representation, and model/computation complexity.",30555
generalizes Our conclusion properties. to different datasets with very,Our conclusion generalizes to datasets with very different properties.,30556
"MaskLab interest, Within region semantic combining foreground/background direction of segmentation and performs prediction. by each","Within each region of interest, MaskLab performs foreground/background segmentation by combining semantic and direction prediction.",30557
demonstrate correlated is agreement that with the highly ensemble We the segmentation errors.,We demonstrate that the ensemble agreement is highly correlated with the segmentation errors.,30558
"our method measures provides that Therefore, corrections. can local user guide","Therefore, our method provides measures that can guide local user corrections.",30559
"(GWRR) model ridge we a leaf-nodes' Finally, generalized the regression weighted regressors. for present","Finally, we present a generalized weighted ridge regression (GWRR) model for the leaf-nodes' regressors.",30560
how a convolutional estimate can be bump We show deep such used encoder-decoder to maps.,We show how a deep convolutional encoder-decoder can be used to estimate such bump maps.,30561
in (e.g. Symbol handwritten detection graphics techniques online,Symbol detection techniques in online handwritten graphics (e.g.,30562
a expressions) of single methods diagrams consist type. mathematical specifically graphic for designed and,diagrams and mathematical expressions) consist of methods specifically designed for a single graphic type.,30563
recognition. can models neural for Complicated high achieve and image accuracy deep network,Complicated and deep neural network models can achieve high accuracy for image recognition.,30564
"computational of can the cost reduce number was which and proposed, Therefore, dramatically. parameters MobileNet","Therefore, MobileNet was proposed, which can reduce the number of parameters and computational cost dramatically.",30565
use is idea MobileNet The to a convolution. of depthwise separable main,The main idea of MobileNet is to use a depthwise separable convolution.,30566
"to we propose MobileNet. architecture paper, In this improve the new a","In this paper, we propose a new architecture to improve the MobileNet.",30567
intrinsic from databases. trajectory point of a framework for discovery interest presents This paper,This paper presents a framework for intrinsic point of interest discovery from trajectory databases.,30568
quasi-species high dimensions. equation dynamics related study replicator-mutator the in closely like and We evolutionary,We study quasi-species and closely related evolutionary dynamics like the replicator-mutator equation in high dimensions.,30569
for consequences study the the We equation. of our replicator-mutator discuss,We discuss the consequences of our study for the replicator-mutator equation.,30570
"variant filter a correlation predict is motion Moreover, the translational of used individually. to","Moreover, a variant of correlation filter is used to predict the translational motion individually.",30571
components. aims pixels Face generate semantically completion facial new missing for to,Face completion aims to generate semantically new pixels for missing facial components.,30572
large variations generative task of face due to is challenging appearance. It a,It is a challenging generative task due to large variations of face appearance.,30573
face under paper structured generative occlusions. studies This completion,This paper studies generative face completion under structured occlusions.,30574
three domains. between and as formulated are the fusing processes disentangling The transformations the,The disentangling and fusing processes are formulated as the transformations between the three domains.,30575
is method evaluated Meshface problem. verification on proposed The,The proposed method is evaluated on Meshface verification problem.,30576
of communication. daily Emojis popular part a digital very become have,Emojis have become a very popular part of daily digital communication.,30577
"in been kinds two evaluation. of Meanwhile, this datasets used have","Meanwhile, two kinds of datasets have been used in this evaluation.",30578
"ability the opens horizons to vision advanced Moreover, text for improving detect computer many problems.","Moreover, ability to detect text opens the horizons for improving many advanced computer vision problems.",30579
are regions. These fine-tuned for neighbor regions the non-textual further rejecting,These regions are further fine-tuned for rejecting the non-textual neighbor regions.,30580
"remote high potentially hyperspectral feature very In dimensional. sensing, can data become","In hyperspectral remote sensing, feature data can potentially become very high dimensional.",30581
"of limited. training data the amount oftentimes is However,","However, the amount of training data is oftentimes limited.",30582
hyperspectral show efficacy We approach the the proposed of datasets. on two,We show the efficacy of the proposed approach on two hyperspectral datasets.,30583
One transformations more traditional strategies is data of augmentations successful the the above. mentioned,One of the more successful data augmentations strategies is the traditional transformations mentioned above.,30584
generate with We of styles. to experiment images also different GANs,We also experiment with GANs to generate images of different styles.,30585
the We on method discuss this various shortcomings of and datasets. successes,We discuss the successes and shortcomings of this method on various datasets.,30586
"In work, hand demonstrate this on mobile detection HMDs. egocentric localization we gesture and real-time","In this work, we demonstrate real-time egocentric hand gesture detection and localization on mobile HMDs.",30587
using multipliers. alternating of solve problem resultant optimization the direction convex method the We,We solve the resultant convex optimization problem using the alternating direction method of multipliers.,30588
"is Blizzard The from Warcraft Wars variant a Tower environment Entertainment. Line of III,","The environment is a variant of Tower Line Wars from Warcraft III, Blizzard Entertainment.",30589
for regularization keypoint Such provide prediction can instances. effective on view consistency unlabeled,Such view consistency can provide effective regularization for keypoint prediction on unlabeled instances.,30590
"regularize we a predictions alignment geometric In to addition, the term in introduce target domain.","In addition, we introduce a geometric alignment term to regularize predictions in the target domain.",30591
minimization. resulting loss optimized The function be effectively via alternating can,The resulting loss function can be effectively optimized via alternating minimization.,30592
"can to proposals Furthermore, segmentation. ground-truth generate used it be weakly for supervised","Furthermore, it can be used to generate ground-truth proposals for weakly supervised segmentation.",30593
Hough popular transform algorithm. computer low-level is a vision,Hough transform is a popular low-level computer vision algorithm.,30594
"subset dyadic patterns. special Because are their of structure, these called","Because of their special structure, these subset are called dyadic patterns.",30595
In paper investigated. patterns dyadic this of various properties are,In this paper various properties of dyadic patterns are investigated.,30596
approximation error upper bounds on are Exact derived.,Exact upper bounds on approximation error are derived.,30597
"paper, proposed novel In problems. tackle a to these recurrent we this network","In this paper, we proposed a novel recurrent network to tackle these problems.",30598
adoption It the also Short-Term Long video (LSTM) enables frames. between units of Memory,It also enables the adoption of Long Short-Term Memory (LSTM) units between video frames.,30599
frames. in imposing We geometric effective found very among RNN memory augmented is consistency such,We found such memory augmented RNN is very effective in imposing geometric consistency among frames.,30600
successfully videos input degradation well It outputs. quality sequential also in handles while the stabilizes,It also well handles input quality degradation in videos while successfully stabilizes the sequential outputs.,30601
"promise, daunting however, Realizing this raises challenges.","Realizing this promise, however, raises daunting challenges.",30602
human behavior. preferences infer or Inverse observed reinforcement attempts learning to (IRL) rewards from,Inverse reinforcement learning (IRL) attempts to infer human rewards or preferences from observed behavior.,30603
"from observations. `normative' need cannot address exclusively assumptions, be simple we which this, To deduced","To address this, we need simple `normative' assumptions, which cannot be deduced exclusively from observations.",30604
compared detection to image is object challenging more object detection. Video,Video object detection is more challenging compared to image object detection.,30605
"Visual causing clues motion corresponding by frames. weakened defocus on failure get and blur,","Visual clues get weakened by defocus and motion blur, causing failure on corresponding frames.",30606
"propagation effective in sacrifice methods accuracy. based but Feature speed, they improving proved the the","Feature propagation based methods proved effective in improving the speed, but they sacrifice the accuracy.",30607
and it So to is improve performance speed possible simultaneously?,So is it possible to improve speed and performance simultaneously?,30608
"established sparsely iteratively features. extracted our is absorbing framework, impression an frame feature In by","In our framework, an impression feature is established by iteratively absorbing sparsely extracted frame features.",30609
perspective can on feature new a hope enhancement. video Network We Impression provide,We hope Impression Network can provide a new perspective on video feature enhancement.,30610
methods ultrametric. clustering guarantees quite and for provide Our are performance general spectral any with,Our methods are quite general and provide performance guarantees for spectral clustering with any ultrametric.,30611
SISR the task The a attractive been research over decades. has two topic last very,The SISR task has been a very attractive research topic over the last two decades.,30612
is breeding. grapevine the fruit in berry The most size traits of important one,The berry size is one of the most important fruit traits in grapevine breeding.,30613
"to Then, network neural use we artifacts. a post-processing remove these","Then, we use a post-processing neural network to remove these artifacts.",30614
"algorithm train for to an is advanced deep Meanwhile, neural learning our proposed networks compression.","Meanwhile, an advanced learning algorithm is proposed to train our deep neural networks for compression.",30615
It formulation non-parametric upon the IRL in Expectation-Maximization is the setting. clustering built and,It is built upon the Expectation-Maximization formulation and non-parametric clustering in the IRL setting.,30616
viral prevalent disease. Dengue most the is arthropod-borne,Dengue is the most prevalent arthropod-borne viral disease.,30617
"or However, few infections no most cause DENV symptoms.","However, most DENV infections cause few or no symptoms.",30618
for linear a such modeling convenient provide way models Matrix data.,Matrix linear models provide a convenient way for modeling such data.,30619
"many of sparse models desired. is In these estimation situations,","In many situations, sparse estimation of these models is desired.",30620
"to linear general models sparse matrix data. present fitting high-throughput fast, methods We structured for","We present fast, general methods for fitting sparse matrix linear models to structured high-throughput data.",30621
biomolecular properties multiple in functionally important contexts. period and amplitude are The of oscillators,The period and amplitude of biomolecular oscillators are functionally important properties in multiple contexts.,30622
"conjunction acceleration algorithms. be the most can Thus, with structure used in inpainting exemplar-based","Thus, the acceleration structure can be used in conjunction with most exemplar-based inpainting algorithms.",30623
imbalance paths. and between mismatch by caused I/Q I/Q phase amplitude is,I/Q imbalance is caused by amplitude and phase mismatch between I/Q paths.,30624
"I/Q compensation digital in is the imbalance critical. Thus, domain","Thus, I/Q imbalance compensation in the digital domain is critical.",30625
this paper. two There contributions are main in,There are two main contributions in this paper.,30626
based control a methods of to used `Indifference' agents. reward class to refers,`Indifference' refers to a class of methods used to control reward based agents.,30627
"censuses modeling. population Typically, is and using statistical obtained data","Typically, population data is obtained using censuses and statistical modeling.",30628
"recently, emerged, capable methods areas. identifying have data remotely-sensed using of urbanized effectively More","More recently, methods using remotely-sensed data have emerged, capable of effectively identifying urbanized areas.",30629
define this (GoR) on Restoration phenomenon. Gain We of based further,We further define Gain of Restoration (GoR) based on this phenomenon.,30630
usually Convolutional are layer-by-layer. stacking built operations networks by (CNNs) neural convolutional,Convolutional neural networks (CNNs) are usually built by stacking convolutional operations layer-by-layer.,30631
"deal model in scalability single lack a multiple Moreover, they degradations. learning with non-blindly to","Moreover, they lack scalability in learning a single model to non-blindly deal with multiple degradations.",30632
can How experimental the literature? on a of stay top researcher,How can a researcher stay on top of the experimental literature?,30633
compared experimental computational data? data efficiently How can be modelling to,How can computational modelling data be efficiently compared to experimental data?,30634
effectively modelling can computational inform experimentalists? most How,How can computational modelling most effectively inform experimentalists?,30635
Here exploit to this the framework challenges above. for in research psychiatry mentioned we address,Here we exploit this framework for research in psychiatry to address the challenges mentioned above.,30636
of multiple consist Most steps. and existing deep works several networks pre-processing neural,Most existing works consist of multiple deep neural networks and several pre-processing steps.,30637
"avoidance algorithms have been by a multiple developed. obstacle webcam, detection taken and Using simple","Using multiple taken by a simple webcam, obstacle detection and avoidance algorithms have been developed.",30638
Spatial-Temporal Networks for video detection. Memory object introduce We,We introduce Spatial-Temporal Memory Networks for video object detection.,30639
cardiovascular segmentation has assessing pathologies. numerous retinal in disease ophthalmologic The of and morphology applications,The segmentation of retinal morphology has numerous applications in assessing ophthalmologic and cardiovascular disease pathologies.,30640
an in Augmented domains. technology emerging reality many is application,Augmented reality is an emerging technology in many application domains.,30641
"hair augmentation. paper, the In we address this problem live color of","In this paper, we address the problem of live hair color augmentation.",30642
"achieve this quickly hair segmented goal, be and To accurately. to needs","To achieve this goal, hair needs to be segmented quickly and accurately.",30643
"are to simpler and the While noisy there is such data obtain, segmentations coarse. much","While such data is much simpler to obtain, the segmentations there are noisy and coarse.",30644
whole in end-to-end trained an is The manner. model,The whole model is trained in an end-to-end manner.,30645
This methods. on that with state-of-the-art par achieves are results,This achieves results that are on par with state-of-the-art methods.,30646
"achieve To multi-stage a our goal, convolutional network. introduce scale-aware we neural","To achieve our goal, we introduce a scale-aware multi-stage convolutional neural network.",30647
the method show methods. proposed that several state-of-the-art outperformed results Experiment,Experiment results show that the proposed method outperformed several state-of-the-art methods.,30648
intelligent intelligence to build artificial in There growing interest socially is robots.,There is growing interest in artificial intelligence to build socially intelligent robots.,30649
"many in In and most with interactions the stamps. video grounded attributes time are addition,","In addition, most interactions and many attributes are grounded in the video with time stamps.",30650
"reason We understanding. propose interaction ordering, and understanding via also for methods","We also propose methods for interaction understanding via ordering, and reason understanding.",30651
"the size. not but greatly LFSP The maintains light also reduces redundancy, only graph field","The LFSP not only maintains light field redundancy, but also greatly reduces the graph size.",30652
advantages These to segmentation useful make LFSP improve efficiency.,These advantages make LFSP useful to improve segmentation efficiency.,30653
"graph using we an segmentation algorithm structure, field efficient Based graph-cuts. on light LFSP present","Based on LFSP graph structure, we present an efficient light field segmentation algorithm using graph-cuts.",30654
to extract Many summaries have such algorithms been automatically. proposed,Many algorithms have been proposed to extract such summaries automatically.,30655
question. How is best outputs an important evaluate but to these little-discussed,How best to evaluate these outputs is an important but little-discussed question.,30656
"al. on algorithm the we et Based greedy to Kannappan results, recommend due a matching","Based on the results, we recommend a greedy matching algorithm due to Kannappan et al.",30657
"conference earlier extends et an paper, Owens al. This paper","This paper extends an earlier conference paper, Owens et al.",30658
"challenging and However, is face complex visibility. various background, orientations because of appearance it","However, it is challenging because of complex background, various orientations and face appearance visibility.",30659
"and of Therefore, are representation images critical. mapping to a descriptive face poses it","Therefore, a descriptive representation of face images and mapping it to poses are critical.",30660
"improved. computed this way, multiple model is In the mapping with tasks","In this way, the computed mapping model with multiple tasks is improved.",30661
how show process We embeddings. related this to imprinting is proxy-based,We show how this imprinting process is related to proxy-based embeddings.,30662
plant as These affect guilds preferences pollinator as coexistence within well changes in guilds. within,These changes in preferences affect coexistence within pollinator guilds as well as within plant guilds.,30663
"PointNet promising achieved Among point has sets. results existing directly works, on by learning","Among existing works, PointNet has achieved promising results by directly learning on point sets.",30664
generation column novel testing. software for a combinatorial paper This proposes framework,This paper proposes a novel column generation framework for combinatorial software testing.,30665
generic heterogeneous column is and Our arrays generation coverage can alphabets. over mixed accommodate algorithm,Our column generation algorithm is generic and can accommodate mixed coverage arrays over heterogeneous alphabets.,30666
market. systems with of growth demand the increasing smartphone The been biometric has for the,The demand for biometric systems has been increasing with the growth of the smartphone market.,30667
"the to manufacturers, security is According guaranteed be level high. to smartphone","According to smartphone manufacturers, the security level is guaranteed to be high.",30668
"the In temporal in applications, and crowd domains. real-world crowd spatial counting densities vary greatly","In real-world crowd counting applications, the crowd densities vary greatly in spatial and temporal domains.",30669
counting on our Experimental state-of-the-art three show method datasets. results performance that challenging achieves crowd,Experimental results show that our method achieves state-of-the-art performance on three challenging crowd counting datasets.,30670
vs. days) parallelize because they better. hours,hours vs. days) because they parallelize better.,30671
"to ES. and with contain unknown it such is how encourage exploration local optima),","contain local optima), and it is unknown how to encourage such exploration with ES.",30672
cannot studies comparison use such larger But feasibly surveys. user,But larger comparison studies cannot feasibly use such user surveys.,30673
subsequently truth when ground propose formal comparing protocol available. for is summaries We a,We subsequently propose a formal protocol for comparing summaries when ground truth is available.,30674
We the data mini the smart mine propose extract to card activities. to,We propose to mine the smart card data to extract the mini activities.,30675
"trip our in We the of France. method on different settings Nancy, collection test passenger","We test our method in different settings on the passenger trip collection of Nancy, France.",30676
reduction. important We very results experimental a demonstrating mismatch report,We report experimental results demonstrating a very important mismatch reduction.,30677
Protein a detailed description dihedral conformation. of protein local provide angles,Protein dihedral angles provide a detailed description of protein local conformation.,30678
"angle from prediction challenging. is direct alone However, sequence","However, direct angle prediction from sequence alone is challenging.",30679
"clusters a mixture the prediction with output predicted their of by we Finally, real-valued probabilities.","Finally, we output real-valued prediction by a mixture of the clusters with their predicted probabilities.",30680
"by can error be is, the real well approximated bounds. our estimated That prediction","That is, the real prediction error can be well approximated by our estimated bounds.",30681
civil communities. increased mode construction gained and attentions Image-based vibration in identification,Image-based vibration mode identification gained increased attentions in civil and construction communities.,30682
video-based structure visualize method motion recent small to measure motions. was and A magnification developed,A recent video-based motion magnification method was developed to measure and visualize small structure motions.,30683
This new low-cost and shape potential presents approach vibration mode measurement a for identification.,This new approach presents a potential for low-cost vibration measurement and mode shape identification.,30684
was Pilot on simple body this rigid reported. approach using structures studies,Pilot studies using this approach on simple rigid body structures was reported.,30685
outdoor have structures investigated. been complex on validity Its not,Its validity on complex outdoor structures have not been investigated.,30686
This method for in and both tested was outdoor environments indoor validation.,This method was tested in both indoor and outdoor environments for validation.,30687
multiple in collaboration agents. Many the tasks of artificial intelligence require,Many tasks in artificial intelligence require the collaboration of multiple agents.,30688
multi-agent for deep domains. learning We exam reinforcement,We exam deep reinforcement learning for multi-agent domains.,30689
a Such leverages hierarchical naturally from advantages another. one structure,Such a hierarchical structure naturally leverages advantages from one another.,30690
"independent reasoning. composed action representation, learnable communication and","composed action representation, learnable communication and independent reasoning.",30691
"an tasks, to these image correspond multiple of most input In outputs. may","In most of these tasks, an input image may correspond to multiple outputs.",30692
"show approaches current of However, minor diversity only very existing outputs. the","However, current existing approaches only show very minor diversity of the outputs.",30693
a problem. cast nonlinear regression estimation from is Age typically images as facial,Age estimation from facial images is typically cast as a nonlinear regression problem.,30694
is space of problem challenge The the this w.r.t. feature main facial,The main challenge of this problem is the facial feature space w.r.t.,30695
"Deep an propose Forests for (DRFs), model, In we end-to-end paper, this estimation. Regression age","In this paper, we propose Deep Regression Forests (DRFs), an end-to-end model, for age estimation.",30696
sequence paper re-IDentification database establishes This for Video a mainly Long large-scale person (LVreID).,This paper mainly establishes a large-scale Long sequence Video database for person re-IDentification (LVreID).,30697
streak tomography. paper angle This addresses in limited reduction,This paper addresses streak reduction in limited angle tomography.,30698
"purpose artifacts reduce is Hence, streak to the scales. this of main various at paper","Hence, the main purpose of this paper is to reduce streak artifacts at various scales.",30699
the (ssaTV) propose scale-space anisotropic We total implementations. variation two different in algorithm,We propose the scale-space anisotropic total variation (ssaTV) algorithm in two different implementations.,30700
to objects localize propose structure We a recurrent novel jointly. and attentional recognize,We propose a novel recurrent attentional structure to localize and recognize objects jointly.,30701
"are complete fused observations and detection Meanwhile, those tasks. to classification","Meanwhile, those observations are fused to complete detection and classification tasks.",30702
results can method effective context by mine proposed several Experimental that the local indicate observations.,Experimental results indicate that the proposed method can mine effective context by several local observations.,30703
"changing steps. Moreover, precision the easily the speed and are improved number recurrent by of","Moreover, the precision and speed are easily improved by changing the number of recurrent steps.",30704
"code our of source will the approach. we proposed open Finally,","Finally, we will open the source code of our proposed approach.",30705
"transportation an to data urban crucial prediction is growth Hence, system. network","Hence, transportation network data is crucial to an urban growth prediction system.",30706
"highly However, under constrained. alone reprojection the model leaves the loss","However, the reprojection loss alone leaves the model highly under constrained.",30707
person. runs Our the model bounding a containing given in real-time box,Our model runs in real-time given a bounding box containing the person.,30708
a novel images. fusing static present learning architecture deep for We multi-exposure,We present a novel deep learning architecture for fusing static multi-exposure images.,30709
sequence. use approaches fusion features hand-crafted to Current (MEF) input multi-exposure fuse,Current multi-exposure fusion (MEF) approaches use hand-crafted features to fuse input sequence.,30710
"not weak robust are the hand-crafted conditions. input to However, representations varying","However, the weak hand-crafted representations are not robust to varying input conditions.",30711
"image for Moreover, perform pairs. poorly exposure extreme they","Moreover, they perform poorly for extreme exposure image pairs.",30712
is a in deployment for requirement its also real-world It scenarios.,It is also a requirement for its deployment in real-world scenarios.,30713
for a efficient multi-task paper proposes framework novel This reinforcement learning.,This paper proposes a novel framework for efficient multi-task reinforcement learning.,30714
to This new of during acquire training. agents skills stages enables continually different,This enables agents to continually acquire new skills during different stages of training.,30715
task human corresponds a learned description. to Each language,Each learned task corresponds to a human language description.,30716
method same then will any abnormalities detect extended on Track. be Railway the to The,The same method will then be extended to detect any abnormalities on the Railway Track.,30717
of that in contrast these this We find is predictions. behaviour entropy measure to the,We find that the behaviour of this entropy measure is in contrast to these predictions.,30718
was of the Inference application MCMC performed. also through,Inference through the application of MCMC was also performed.,30719
"fitting made. were evaluated methods further were and recommendations Finally, novel","Finally, novel fitting methods were evaluated and further recommendations were made.",30720
equations space solve wave computationally making approaches them and expensive typically slow operate. to Current,Current approaches typically solve wave space equations making them computationally expensive and slow to operate.,30721
simulation. We intravascular (IVUS) framework to apply ultrasound the,We apply the framework to intravascular ultrasound (IVUS) simulation.,30722
specific intensities. stage them speckle refines subsequently tissue preserve to The I GAN,The stage I GAN subsequently refines them to preserve tissue specific speckle intensities.,30723
"largely of however, self and unknown. operation The of formation the are, mechanisms","The mechanisms of operation and formation of the self are, however, largely unknown.",30724
"prereflective, even inputs. visual self initially independent of is prelinguistic, perhaps and This largely","This self is prereflective, prelinguistic, and initially perhaps even largely independent of visual inputs.",30725
"developmental perspectives We of will neuroscience. investigate and the psychology, phenomenology, this from","We will investigate this from the perspectives of phenomenology, developmental psychology, and neuroscience.",30726
term into prior framework. integrate shape optimisation variational We a,We integrate a shape prior term into variational optimisation framework.,30727
"methods, insights with into optimisation are heterogeneous Together platforms for implementation details the profound provided.","Together with the profound insights into optimisation methods, implementation details for heterogeneous platforms are provided.",30728
have made by predictions experiments real-life model. the Novel been in also mimicking,Novel predictions have also been made by mimicking real-life experiments in the model.,30729
"in than In pixels human more is regions recognition. of of visual important information system,","In human visual system, information of regions is more important than of pixels in recognition.",30730
"approach simulates our Consequently, recognition. closely human proposed","Consequently, our proposed approach closely simulates human recognition.",30731
"the model features are method use identify to error. refractive with that ""attention"" Our correlated","Our model use the ""attention"" method to identify features that are correlated with refractive error.",30732
indicate systems. Results ensuring to the elevated compared aforementioned high an while reliability SE a,Results indicate a high SE while ensuring an elevated reliability compared to the aforementioned systems.,30733
a properties system The key and model efficiently its channelizer time-varying captures operation. of digital,The model efficiently captures key properties of a digital channelizer system and its time-varying operation.,30734
video mask This framework hashing for retrieval. paper with end-to-end fast deep proposes an category,This paper proposes an end-to-end deep hashing framework with category mask for fast video retrieval.,30735
"provide of resources over very channel especially fading AMC channels. Thus, use can efficient","Thus, AMC can provide very efficient use of channel resources especially over fading channels.",30736
digital modulation an Modulation (QAM) and Quadrature is employed ficient widely Amplitude ef- technique.,Quadrature Amplitude Modulation (QAM) is an ef- ficient and widely employed digital modulation technique.,30737
a rectangular It employs constellation. signal typically,It typically employs a rectangular signal constellation.,30738
are of two-dimensional the the space. signal constellation the regions decision partitions Therefore of square,Therefore the decision regions of the constellation are square partitions of the two-dimensional signal space.,30739
using based software-defined efficiency evaluated a (SDR) are and The feasibility prototype. radio,The feasibility and efficiency are evaluated using a software-defined radio (SDR) based prototype.,30740
We public with different health three consider conditions. regions,We consider three regions with different public health conditions.,30741
"""buffer the regions. second the region a between two extremal Therefore, zone"" is","Therefore, the second region is a ""buffer zone"" between the two extremal regions.",30742
actions agent's their taken surrounding the to the when effects Affordances objects. relate on,Affordances relate the agent's actions to their effects when taken on the surrounding objects.,30743
classic a and Our competitive collaborative are both game scenarios. soccer in models involving evaluated,Our models are evaluated in a classic soccer game involving both competitive and collaborative scenarios.,30744
Conditions are given. existence its uniqueness for and,Conditions for its existence and uniqueness are given.,30745
"for we proposed parabolic the approach use construct spline. a constructing the earlier To curve,","To construct the curve, we use the approach proposed earlier for constructing a parabolic spline.",30746
to additional idea function. values The is use of unknown some with points,The idea is to use additional points with unknown values of some function.,30747
Examples of given. constructing curve are spline a,Examples of constructing a spline curve are given.,30748
significant Catastrophic forgetting impact in a learning. has reinforcement negative,Catastrophic forgetting has a significant negative impact in reinforcement learning.,30749
pseudorehearsal agent different a approaches. and in tested pole balancing task We compared,We tested agent in a pole balancing task and compared different pseudorehearsal approaches.,30750
can found that and We assist learning forgetting. decrease pseudorehearsal have,We have found that pseudorehearsal can assist learning and decrease forgetting.,30751
"by changing disorders is routines, compounded aging to dietary obesity. and poor Susceptibility lifestyles, such","Susceptibility to such disorders is compounded by changing lifestyles, poor dietary routines, aging and obesity.",30752
"predicted were Similarly, wearable age recording samples from gait sensors. on-body groups and by","Similarly, age groups were predicted by recording gait samples from on-body and wearable sensors.",30753
profiling Such limit human convenient body age of indices. specialized methods and and mass active,Such specialized methods limit active and convenient profiling of human age and body mass indices.,30754
"gait recording smartphone by these sensors means as issues, a We signals. introducing address for","We address these issues, by introducing smartphone sensors as a means for recording gait signals.",30755
is one segmentation. lack based key generalizability learning limitation deep of A of,A lack of generalizability is one key limitation of deep learning based segmentation.,30756
"independently a images. segmentation trained efforts synthetic network using Then, these","Then, these efforts trained a segmentation network independently using synthetic images.",30757
"not the two and segmentation. synthesis independent However, use stages between information complementary did these","However, these two independent stages did not use the complementary information between synthesis and segmentation.",30758
The states parameters. as from by variational itself distinguishes using approaches latent variational approach previous,The approach distinguishes itself from previous variational approaches by using latent states as variational parameters.,30759
cognition last during as psychology two discipline important decades. Quantum the mathematical of emerged an,Quantum cognition emerged as an important discipline of mathematical psychology during the last two decades.,30760
approach improves of predictions Our replicates models music and perception. on based also symbolic,Our approach replicates and also improves predictions based on symbolic models of music perception.,30761
important with models variables recent role econometric High-dimensional literature. endogenous in increasingly an play linear,High-dimensional linear models with endogenous variables play an increasingly important role in recent econometric literature.,30762
regions estimators confidence that that. propose Our to contribution achieve is would and main,Our main contribution is to propose estimators and confidence regions that would achieve that.,30763
"procedures. is of new high-dimension out parameters carried nuisance via Moreover, pivotal estimation","Moreover, estimation of high-dimension nuisance parameters is carried out via new pivotal procedures.",30764
on templates amino-acylated nucleic-acid which adapters propose We assembled. short,We propose short nucleic-acid templates on which amino-acylated adapters assembled.,30765
peptides. localization activated peptide to ligation precursors drives Spatial from generate phosphodiester-bond-catalytic,Spatial localization drives peptide ligation from activated precursors to generate phosphodiester-bond-catalytic peptides.,30766
for bounding Annotation box introduce We Intelligent Dialogs annotation.,We introduce Intelligent Annotation Dialogs for bounding box annotation.,30767
share features. healthy stages of continue to disease geometric vs. several,healthy vs. stages of disease continue to share several geometric features.,30768
"load reduced. classification computation this can in much image be the patch-level way, In based","In this way, the computation load in patch-level based image classification can be much reduced.",30769
Experimental benefits state alternatives. over confirm art the that of ALSF results can offer,Experimental results confirm that ALSF can offer benefits over state of the art alternatives.,30770
"development artificial recent have extensive attention. of sensing remote satellite, attracted images Inspired by","Inspired by recent development of artificial satellite, remote sensing images have attracted extensive attention.",30771
assignment has high would and by pathologists. This relevance generally require assessment vast clinical microscopic,This assignment has high clinical relevance and would generally require vast microscopic assessment by pathologists.,30772
forecast boundary region model through the We algorithm. overlapped patch-based error of eliminate our,We eliminate the boundary error of patch-based model through our overlapped region forecast algorithm.,30773
datasets. that begun as a proxy to virtual have For of function datasets real,For that virtual datasets have begun to function as a proxy of real datasets.,30774
environmental The be in can completely. conditions controlled artificial scenes,The environmental conditions in artificial scenes can be controlled completely.,30775
The the kernel function of mean is bandwidth a algorithm. in a shift crucial parameter,The bandwidth of a kernel function is a crucial parameter in the mean shift algorithm.,30776
contributions. proposes adaptive three bandwidth contains main strategy paper This which a novel,This paper proposes a novel adaptive bandwidth strategy which contains three main contributions.,30777
"such robust are placed this setting less that e.g. cameras In sensors, strategically","In this setting cameras are strategically placed such that less robust sensors, e.g.",30778
"(FOVs) activity, of field cameras. that located within monitor geophones of are the views seismic","geophones that monitor seismic activity, are located within the field of views (FOVs) of cameras.",30779
a on scheduling distributed promising initial We experimental show simulated problem. results,We show promising initial experimental results on a simulated distributed scheduling problem.,30780
and add of visualization further explainability Our to results. the analysis,Our visualization and analysis further add to the explainability of results.,30781
from information an stored the at of image contour study locations. We problem reconstructing,We study the problem of reconstructing an image from information stored at contour locations.,30782
"binary a attribute proposed representation. were predict which AttributeSVMs Recently,","Recently, AttributeSVMs were proposed which predict a binary attribute representation.",30783
"time, method At in word defined influential the this state-of-the-art spotting. their segmentation-based","At their time, this influential method defined the state-of-the-art in segmentation-based word spotting.",30784
"designed different architectures, we for specifically addition, two In spotting. word CNN propose","In addition, we propose two different CNN architectures, specifically designed for word spotting.",30785
end-to-end in able to These fashion. be architectures an trained are,These architectures are able to be trained in an end-to-end fashion.,30786
identification The of is writer local and for features encoding essential writer part an retrieval.,The encoding of local features is an essential part for writer identification and writer retrieval.,30787
"triangulation VLAD encoding established work, In the we this compare embedding. with","In this work, we compare the established VLAD encoding with triangulation embedding.",30788
"Interference (IPI) call problem Perturbation this we this paper, the In problem. Instance","In this paper, we call this problem the Instance Perturbation Interference (IPI) problem.",30789
"propose network to adversarial generative paper, aforementioned In we improved the overcome limitations. an this","In this paper, we propose an improved generative adversarial network to overcome the aforementioned limitations.",30790
employing manner Network (RNN) Recurrent Attention is fully-differentiable. and recurrent Neural a in handled is,Attention is handled in a recurrent manner employing Recurrent Neural Network (RNN) and is fully-differentiable.,30791
"using extracted distributions human are information: Instead, articulated pose. attention external","Instead, attention distributions are extracted using external information: human articulated pose.",30792
present segmentation efficient associates We mask-based by segmentation. reported tracker masks the an which pixel-precise,We present an efficient segmentation mask-based tracker which associates pixel-precise masks reported by the segmentation.,30793
"explored. However, need to factors resolution further for be model the","However, the factors for model resolution need to be further explored.",30794
We method our tree apply classify also regions. to,We also apply our method to classify tree regions.,30795
images places Image neighboring refers missing inpainting to in pixels. using filling,Image inpainting refers to filling missing places in images using neighboring pixels.,30796
also processing. It in tasks different has many image of applications,It also has many applications in different tasks of image processing.,30797
in processing require complexities results in turn computational which changes considerable remarkable These time.,These changes require considerable computational complexities which in turn results in remarkable processing time.,30798
online the of accelerate the video. would in speed execution frame and capability This inpainting,This would accelerate the execution speed and the capability of online frame inpainting in video.,30799
cost-function neighboring spatial and combination all applied is features a in statistical of pixels. The,The applied cost-function is a combination of statistical and spatial features in all neighboring pixels.,30800
prediction. The classification exact are skill features and score for then used skill,The features are then used for skill classification and exact skill score prediction.,30801
Using the time RMIS in for feedback score real training. allow framework can potentially presented,Using the framework presented can potentially allow for real time score feedback in RMIS training.,30802
parameter in online simulation of is study. efficiency illustrated this scheme The a estimation,The efficiency of this online parameter estimation scheme is illustrated in a simulation study.,30803
data research. code We publicly and further will facilitate share our to,We will publicly share our code and data to facilitate further research.,30804
"propose also structure Besides a new modification, we with hierarchical classifier dice loss. a the","Besides the structure modification, we also propose a new classifier with a hierarchical dice loss.",30805
their Competitive they are figures more effective predecessors. prove than,Competitive figures prove they are more effective than their predecessors.,30806
such idea is the deterministic Quasi-Monte behind as methods Carlo This (QMC). sampling,This is the idea behind deterministic sampling methods such as Quasi-Monte Carlo (QMC).,30807
is classification from images a fine-grained problem. classification Food,Food classification from images is a fine-grained classification problem.,30808
"scalability food time is curation images Manual cost, of prohibitive. and","Manual curation of food images is cost, time and scalability prohibitive.",30809
"web the but is other hand, data contains noise. On freely available","On the other hand, web data is available freely but contains noise.",30810
"classifying problem images minimal the this with curation. we paper, of data food In address","In this paper, we address the problem of classifying food images with minimal data curation.",30811
key step creation selection space. further a the prototypes of The for of is the,The selection of prototypes is a key step for the further creation of the space.,30812
on prototypes. is matching of based labels criterion their supervised closest counting The objects and,The supervised criterion is based on counting matching labels of objects and their closest prototypes.,30813
"are Specifically, modeled jointly. different modalities optimized but separately","Specifically, different modalities are modeled separately but optimized jointly.",30814
is Another modalities. of it is more proposed easily property the to that extended approach,Another property of the proposed approach is that it is easily extended to more modalities.,30815
synthesis. recurrent variational a propose auto-encoder for We texture,We propose a recurrent variational auto-encoder for texture synthesis.,30816
"for texture loss is novel used synthesizer. A the function, training FLTBNK,","A novel loss function, FLTBNK, is used for training the texture synthesizer.",30817
loss It rotational function. invariant is and color partially,It is rotational and partially color invariant loss function.,30818
"software addition, may imply hardware abrupt signal. and updates in In changes","In addition, hardware and software updates may imply abrupt changes in signal.",30819
longitudinal particularly in automatic combined for are image methods studies. effects The used challenging analysis,The combined effects are particularly challenging for automatic image analysis methods used in longitudinal studies.,30820
increased a estimations implication variation risk measurement The and (e.g. bias is the of in,The implication is increased measurement variation and a risk of bias in the estimations (e.g.,30821
volume change for a in the structure).,in the volume change for a structure).,30822
"applicability. shows our model good general Furthermore, training","Furthermore, our training model shows good general applicability.",30823
driving scenarios. relevant interesting that successfully discovers to approach We objects show this,We show that this approach successfully discovers interesting objects relevant to driving scenarios.,30824
enabling learning direct has approach driving. for for relevance Our large-scale autonomous object,Our approach has direct relevance for enabling large-scale object learning for autonomous driving.,30825
"this for CommSense communication-sensing, call short. We","We call this communication-sensing, CommSense for short.",30826
"leveraging a by In this network generative study, above adversarial (GAN). address issue we the","In this study, we address the above issue by leveraging a generative adversarial network (GAN).",30827
as enables us the in-domain to comparison. change detection treat task This an cross-domain image,This enables us to treat the cross-domain change detection task as an in-domain image comparison.,30828
This the detectors. in-domain change on of leverage generic allows large to body literature us,This allows us to leverage the large body of literature on in-domain generic change detectors.,30829
proposed efficacy of Experiments approach. the validate,Experiments validate efficacy of the proposed approach.,30830
often produce on scene and trained datasets results. models are parsing quality ideal Outdoor,Outdoor scene parsing models are often trained on ideal datasets and produce quality results.,30831
"to However, real a this leads when to applied world. the discrepancy","However, this leads to a discrepancy when applied to the real world.",30832
big up approaches The power discover various drug-drug to new is data interactions. of opening,The power of big data is opening up new approaches to discover various drug-drug interactions.,30833
"to we interactions. this framework, In PRD, propose paper, drug-drug called a predict novel","In this paper, we propose a novel framework, called PRD, to predict drug-drug interactions.",30834
"First, drug different a is graph knowledge from large-scale generated sources.","First, a large-scale drug knowledge graph is generated from different sources.",30835
"on To the were effectiveness real-world datasets. the framework, validate experiments extensive proposed conducted of","To validate the effectiveness of the proposed framework, extensive experiments were conducted on real-world datasets.",30836
diseases affect the is one coffee that Coffee worldwide. plantations of rust main,Coffee rust is one of the main diseases that affect coffee plantations worldwide.,30837
"characterized attractors. have multiple dynamics often However, networks by gene","However, gene networks often have dynamics characterized by multiple attractors.",30838
is dynamics also stochastic molecular approach frameworks. and Our other types networks of generalizable to,Our approach is also generalizable to other types of molecular networks and stochastic dynamics frameworks.,30839
to recently. Many adaptation domain been proposed have the solve methods problem,Many methods have been proposed to solve the domain adaptation problem recently.,30840
"is effect assumption If may negative adaptation. of transfer satisfied, the the degrade domain not","If the assumption is not satisfied, the effect of negative transfer may degrade domain adaptation.",30841
"into Firstly, common and parts the disentangled features learned are specific parts.","Firstly, the learned features are disentangled into common parts and specific parts.",30842
"specific the of each individual Conversely, the style domain. parts characterize unique","Conversely, the specific parts characterize the unique style of each individual domain.",30843
"trained potential also high network style-transferred Besides, to images. the generate demonstrates","Besides, the trained network also demonstrates high potential to generate style-transferred images.",30844
"on its network first dataset. a aerial of hyperspectral trained is The kind AeroCampus,","The network is trained on AeroCampus, a first of its kind aerial hyperspectral dataset.",30845
(HSI). consists spatial high images images AeroCampus resolution color and hyperspectral resolution spatial of low,AeroCampus consists of high spatial resolution color images and low spatial resolution hyperspectral images (HSI).,30846
"aesthetic challenging. Despite is progress, visual computational recent still","Despite recent progress, computational visual aesthetic is still challenging.",30847
We results for and psychophysics minimal data show for interpretation. modeling their images,We show psychophysics data for minimal images and modeling results for their interpretation.,30848
effectively a This regression applied forecasting. is of for based quantile extrapolation nonlinear form,This effectively is a form of extrapolation based nonlinear quantile regression applied for forecasting.,30849
representing components contains respectively. and luminance color chrominance the intensity A color information image and,A color image contains luminance and chrominance components representing the intensity and color information respectively.,30850
"segmentation detection border However, their of state-of-the-art have deficiencies phase. methods many the in","However, many of the state-of-the-art segmentation methods have deficiencies in their border detection phase.",30851
an HBCR would automatic This improvement practical help for in system applications. building,This improvement would help in building an automatic HBCR system for practical applications.,30852
"out gain. carried is the obtain parameter the Additionally, optimization to maximal array","Additionally, the parameter optimization is carried out to obtain the maximal array gain.",30853
paper. in control called resulting optimal scheme is (OPARC) array The precise and response this,The resulting scheme is called optimal and precise array response control (OPARC) in this paper.,30854
"to OPARC. and presented superiority the proposed verify the simulation are effectiveness Finally, of results","Finally, simulation results are presented to verify the effectiveness and superiority of the proposed OPARC.",30855
"on the Subsequently, based called refines score FineNet, minutiae map. network, candidate another locations","Subsequently, another network, called FineNet, refines the candidate minutiae locations based on score map.",30856
effectiveness the networks. of with We using deep together the fingerprint domain the demonstrate knowledge,We demonstrate the effectiveness of using the fingerprint domain knowledge together with the deep networks.,30857
following detrend can Removing reduce points the step positives. false and time,Removing time points and the following detrend step can reduce false positives.,30858
"on the However, complicated data. effects regression filtering covariates has and","However, covariates regression and filtering has complicated effects on the data.",30859
"subject false declined analysis, consistently filtering. for positives after that single Note","Note that for single subject analysis, false positives declined consistently after filtering.",30860
cross-entropy the using trained models are typically Captioning loss.,Captioning models are typically trained using the cross-entropy loss.,30861
"is to assessments. However, with metrics better designed other human performance on evaluated their correlate","However, their performance is evaluated on other metrics designed to better correlate with human assessments.",30862
approach optimize fast to through We of a algorithm. REINFORCE the one's propose interest objective,We propose a fast approach to optimize one's objective of interest through the REINFORCE algorithm.,30863
complete We call Consensus-based Sequence Training (CST). proposal the,We call the complete proposal Consensus-based Sequence Training (CST).,30864
"approach proposed the algorithm, synthesis By an pattern is applying effective devised.","By applying the proposed algorithm, an effective pattern synthesis approach is devised.",30865
Simulation proposed demonstrate of results provided to algorithm. performance the are the,Simulation results are provided to demonstrate the performance of the proposed algorithm.,30866
"algorithm the are In of array OPARC applications signal to addition, multi-point processing the studied.","In addition, the applications of the multi-point OPARC algorithm to array signal processing are studied.",30867
to use networks This regular that on grids. operate neural representation convolutional allows us,This representation allows us to use convolutional neural networks that operate on regular grids.,30868
advance versatile in tool biotechnology. This used applications been prokaryotic also has to,This versatile prokaryotic tool has also been used to advance applications in biotechnology.,30869
We addressed the can be offer future by that community. directions physics,We offer future directions that can be addressed by the physics community.,30870
to importance retrieval image pairwise great similarity image is Learning an task. of,Learning pairwise image similarity is of great importance to an image retrieval task.,30871
"At efficient focus the organized training into for training units optimization. are ranking stage, samples","At the training stage, training samples are organized into focus ranking units for efficient optimization.",30872
over results embedding proposed model superiority existing Experimental the models. of the show metric,Experimental results show the superiority of the proposed model over existing metric embedding models.,30873
an gives this to affirmative letter question. answer presented The,The presented letter gives an affirmative answer to this question.,30874
two adapt imaging to we to tasks. radar to CV-CNN Then modifications introduce it the,Then we introduce two modifications to the CV-CNN to adapt it to radar imaging tasks.,30875
"to details Subsequently, are and the shown implementation method data presented. is generate training some","Subsequently, the method to generate training data is shown and some implementation details are presented.",30876
"components labeling Thus, is connected improved. the of efficiency","Thus, the efficiency of connected components labeling is improved.",30877
This under a from license. GitHub dataset free is available,This dataset is available from GitHub under a free license.,30878
and analysis matches of made. these appraisal is An initial also,An initial analysis and appraisal of these matches is also made.,30879
computer tasks. of important vision for features deep is many useful extraction The,The extraction of useful deep features is important for many computer vision tasks.,30880
from those tasks. Deep extracted features classification networks in perform have to well proved,Deep features extracted from classification networks have proved to perform well in those tasks.,30881
We price elasticities scanner by our groceries of method data. based estimating on demonstrate,We demonstrate our method by estimating price elasticities of groceries based on scanner data.,30882
understanding are humans as beings stakes of huge. informational The,The stakes of understanding humans as informational beings are huge.,30883
"in functionality vanish changes, crucial Without generation. human a will drastic","Without drastic changes, crucial human functionality will vanish in a generation.",30884
"and the cheap properly. Fortunately, cures easy if are promoted","Fortunately, the cures are cheap and easy if promoted properly.",30885
"Five current hypotheses and the problem-space, of solution humanity's possible five span space. a","Five of the hypotheses span humanity's current problem-space, and five a possible solution space.",30886
"\emph{et double-cut-and-join} by al. Yancopoulos (DCJ) {\em introduced operation, The","The {\em double-cut-and-join} (DCJ) operation, introduced by Yancopoulos \emph{et al.",30887
DCJ Our operations is the circular visualization tool program for first using layout.,Our program is the first visualization tool for DCJ operations using circular layout.,30888
"Since can labels unsupervised. our are fully be not required, method","Since labels are not required, our method can be fully unsupervised.",30889
"this multi-camera well-annotated makes DukeMTMC great benchmark tracking large-scale, in which field. a progress is","DukeMTMC is a large-scale, well-annotated multi-camera tracking benchmark which makes great progress in this field.",30890
theories of computational we with integrate reinforcement framework learning Here analogy the (RL). of,Here we integrate theories of analogy with the computational framework of reinforcement learning (RL).,30891
support of this results the approach. Simulation power,Simulation results support the power of this approach.,30892
support development multifocal to of Mathematical developed of have models these been tumors claims.,Mathematical models of development of multifocal tumors have been developed to support these claims.,30893
"we to and matching method this this propose In a problem. fast accurate address paper,","In this paper, we propose a fast and accurate matching method to address this problem.",30894
study. this to fill We this attempt gap in,We attempt to fill this gap in this study.,30895
"to center the change been itself. has the calculated Next, respect fractional with","Next, the fractional change has been calculated with respect to the center itself.",30896
compared of are change to generate pattern. a bit values two fractional next binary The,The two values of fractional change are next compared to generate a binary bit pattern.,30897
image four popular for on method of has tested The databases. retrieval the been performance,The performance of the method has been tested for image retrieval on four popular databases.,30898
improvement methods. many over method existing has The significant proposed other a shown,The proposed method has shown a significant improvement over many other existing methods.,30899
of We surface set function. to bandlimited a the zero-level be assume the,We assume the surface to be the zero-level set of a bandlimited function.,30900
used create dataset. FIJI/ImageJ tools ground truth to are the NIH,NIH FIJI/ImageJ tools are used to create the ground truth dataset.,30901
digital of the computers. in field challenging Optical has been Character a Recognition advent,Optical Character Recognition has been a challenging field in the advent of digital computers.,30902
humans to both be information to is needed where machines. readable It is and,It is needed where information is to be readable both to humans and machines.,30903
involved the one This paper of process i.e. the deals steps in OCR with pre-processing,This paper deals with one of the pre-processing steps involved in the OCR process i.e.,30904
simple for optimized and and skew-detection efficient The been correction. algorithm kept has,The algorithm has been kept simple and optimized for efficient skew-detection and correction.,30905
The the been testing performance aptly analysis demonstrated. algorithm after of has,The performance analysis of the algorithm after testing has been aptly demonstrated.,30906
We skin either of the benign. lesions as or consider specific classifying case malignant,We consider the specific case of classifying skin lesions as either malignant or benign.,30907
"approach. using this on a most two-stage to order cascade focus drawback, In strategies alleviate","In order to alleviate this drawback, most strategies focus on using a two-stage cascade approach.",30908
from the work is Dubey Our done by et work inspired al.,Our work is inspired from the work done by Dubey et al.,30909
than and and texture better color information texture descriptors. individually existing color perform,color and texture information individually perform better than existing texture and color descriptors.,30910
method the experiments. The from provided results satisfactory proposed,The proposed method provided satisfactory results from the experiments.,30911
significant a documents Script videos. in identification and analysing plays role,Script identification plays a significant role in analysing documents and videos.,30912
"CNN-LSTM into First, and framework. into we patches convert feed them images the a","First, we convert the images into patches and feed them into a CNN-LSTM framework.",30913
calculated applying weights after LSTM. patch layer Attention-based are softmax,Attention-based patch weights are calculated applying softmax layer after LSTM.,30914
"to features. CNN local do multiplication Next, corresponding with these yield weights of patch-wise we","Next, we do patch-wise multiplication of these weights with corresponding CNN to yield local features.",30915
of extracted are cell state last Global features from also LSTM.,Global features are also extracted from last cell state of LSTM.,30916
to framework results methods. The superior in achieves proposed conventional comparison,The proposed framework achieves superior results in comparison to conventional methods.,30917
size here mathematical of formulations cell of We review some of the the regulation. problem,We review here some of the mathematical formulations of the problem of cell size regulation.,30918
generate. they stochastic models focus and on the statistics We coarse-grained,We focus on coarse-grained stochastic models and the statistics they generate.,30919
these biologically the relevant from insights review obtained We models.,We review the biologically relevant insights obtained from these models.,30920
thigh solution for MRI This presents segmentation. an paper end-to-end quadriceps,This paper presents an end-to-end solution for MRI thigh quadriceps segmentation.,30921
used widely techniques fields published and are many in Approximation already. methods have many been,Approximation methods are widely used in many fields and many techniques have been published already.,30922
is high scattered approach convenient generally data The RBF for dimensional sets.,The RBF approach is generally convenient for high dimensional scattered data sets.,30923
data are finding of method The scattered. a subset points needs LOWESS if nearest,The LOWESS method needs finding a subset of nearest points if data are scattered.,30924
loss. chronic irreversible eye vision disease leads a Glaucoma is to that,Glaucoma is a chronic eye disease that leads to irreversible vision loss.,30925
"feature hand-crafted segment them methods fundus visual separately, images. Most rely on existing from and","Most existing methods segment them separately, and rely on hand-crafted visual feature from fundus images.",30926
multiple constructs to receptive multi-scale level achieve field The pyramid an layer sizes. input image,The multi-scale input layer constructs an image pyramid to achieve multiple level receptive field sizes.,30927
"final loss the multi-label proposed generate segmentation to Finally, is a function map.","Finally, a multi-label loss function is proposed to generate the final segmentation map.",30928
"two models. of these mathematical connection between classes letter, derive this In we a","In this letter, we derive a mathematical connection between these two classes of models.",30929
"version task-driven broader, computationally-oriented and science, Data as however, understood statistics. of often a is","Data science, however, is often understood as a broader, task-driven and computationally-oriented version of statistics.",30930
"role of communication and learning, the machine reproducibility, computation, theory). analysis, exploratory","exploratory analysis, machine learning, reproducibility, computation, communication and the role of theory).",30931
segmentation. object present our In system paper we video this human-in-the-loop for,In this paper we present our system for human-in-the-loop video object segmentation.,30932
video method backbone for segmentation. our of The a system one-shot object is,The backbone of our system is a method for one-shot video object segmentation.,30933
with in run-time. comparison algorithm shows significant proposed the a A others improvement of,A comparison of the proposed algorithm with others shows a significant improvement in run-time.,30934
results shown. algorithms also for known selected Experimental are,Experimental results for selected known algorithms are also shown.,30935
time other. beside are aligning obtained by each descriptions of set multi-channel series A,A set of multi-channel time series are obtained by aligning descriptions beside each other.,30936
past has sampling The nor bias to the ScreenerNet requires history. remember neither learning,The ScreenerNet neither has sampling bias nor requires to remember the past learning history.,30937
task We propose we and name (PS). panoptic study segmentation a,We propose and study a task we name panoptic segmentation (PS).,30938
methods. The template the proposed emotion recognition outperforms existing matching technique,The proposed emotion recognition technique outperforms the existing template matching methods.,30939
in could propose context an have detection. that role We image detection essential saliency,We propose that context detection could have an essential role in image saliency detection.,30940
of features. high extraction This level requires,This requires extraction of high level features.,30941
contrast map with and maps. information saliency Saliency is from semantic fused based color,Saliency map from semantic information is fused with color and contrast based saliency maps.,30942
map is final then saliency The generated.,The final saliency map is then generated.,30943
segmentation our method performance. have The experiment results obtained demonstrated superior that,The experiment results have demonstrated that our method obtained superior segmentation performance.,30944
against algorithm given. clipping convex is A polyhedron line for new,A new algorithm for line clipping against convex polyhedron is given.,30945
has The and in the case expected complexity complexity. O(sqrt(N))) N worst suggested O(N) algorithm,The suggested algorithm has O(N) complexity in the worst N case and expected O(sqrt(N))) complexity.,30946
speed of triangles. because is 'known up of achieved order' The,The speed up is achieved because of 'known order' of triangles.,30947
is In parallelization order allow strategy novel multi-thread a introduced. also implementations to,In order to allow multi-thread implementations a novel parallelization strategy is also introduced.,30948
"mathematical simulation presents FNS This based. on paper routines and software which models, implementation is","This paper presents mathematical models, software implementation and simulation routines on which FNS is based.",30949
event-driven brain based to aims This personalised future groundwork models. lay for the work,This work aims to lay the groundwork for future event-driven based personalised brain models.,30950
algorithms. benchmarks several We for include learning,We include benchmarks for several learning algorithms.,30951
contracting motion. within engaged suggests in most The directed that cells are model lattices,The model suggests that most cells within contracting lattices are engaged in directed motion.,30952
vehicle presents new navigation. information integrating This semantic for approach paper a for vision-based,This paper presents a new approach for integrating semantic information for vision-based vehicle navigation.,30953
Image a is numerous which restoration applications. has process inpainting,Image inpainting is a restoration process which has numerous applications.,30954
inpainting of Different have been implementation used algorithms. for approaches,Different approaches have been used for implementation of inpainting algorithms.,30955
only consider purpose. direction Interpolation for approaches this one,Interpolation approaches only consider one direction for this purpose.,30956
inpainting. present paper a image to this new In perspective we,In this paper we present a new perspective to image inpainting.,30957
bicubic consider We two-dimensional and interpolations. both one-dimensional and directions multiple apply,We consider multiple directions and apply both one-dimensional and two-dimensional bicubic interpolations.,30958
pixels preserve in pixels. a formation hyperbolic selected corner to are better Neighboring,Neighboring pixels are selected in a hyperbolic formation to better preserve corner pixels.,30959
results. our We work recent our superior with approaches inpainting show to compare,We compare our work with recent inpainting approaches to show our superior results.,30960
used architecture early-merging backbone is as the network. An,An early-merging architecture is used as the backbone network.,30961
"very annotation. cheap Our using while robust fast accurate, is method and","Our method is accurate, fast and robust while using very cheap annotation.",30962
such information. a graphical barcode of representation Persistence is,Persistence barcode is a graphical representation of such information.,30963
experiments human results. Proof-of-concept suggested the further validated on subjects,Proof-of-concept experiments on human subjects further validated the suggested results.,30964
a attacks of threat Adversarial serious success learning pose to the in practice. deep,Adversarial attacks pose a serious threat to the success of deep learning in practice.,30965
fact to This has lead contributions a large influx direction. of this in,This fact has lead to a large influx of contributions in this direction.,30966
"literature the Finally, of draw direction. to a we on the research provide broader outlook","Finally, we draw on the literature to provide a broader outlook of the research direction.",30967
"the the phase sustains oscillations pathway, realistic with indirect relationships. Unlike pathway hyperdirect","Unlike the hyperdirect pathway, the indirect pathway sustains oscillations with realistic phase relationships.",30968
transient with stimuli basal data. in Changes experimental responses accord ganglia to,Changes in basal ganglia responses to transient stimuli accord with experimental data.,30969
indirect results. and separation direct pathways is necessary these of the not Strict for,Strict separation of the direct and indirect pathways is not necessary for these results.,30970
disjoint (\textit{re-id}) across to refers views. Person pedestrians matching yet re-identification camera non-overlapping,Person re-identification (\textit{re-id}) refers to matching pedestrians across disjoint yet non-overlapping camera views.,30971
co-occurrence learned well-aligned representations latent reflect images. of are to the patterns paired The,The learned latent representations are well-aligned to reflect the co-occurrence patterns of paired images.,30972
expression considered as interpretable seen A smaller a model an reliable model. be can decision,A smaller expression can be seen as an interpretable model considered a reliable decision model.,30973
with often performed This Programming as solution which represents expression is Genetic trees. their,This is often performed with Genetic Programming which represents their solution as expression trees.,30974
These find computational constraints. a as challenge under pose optimal the to solution,These pose as a challenge to find the optimal solution under computational constraints.,30975
"heuristic it called data SymTree. an was test to introduced also structure, this In order","In order to test this data structure, it was also introduced an heuristic called SymTree.",30976
hypothesis is A new representation where component. Bernoulli new local creates measurement a presented each,A new local hypothesis representation is presented where each measurement creates a new Bernoulli component.,30977
expensive are experiments make. to often difficult and Field,Field experiments are often difficult and expensive to make.,30978
"computational codes. issues, bypass developed To industrial companies have these","To bypass these issues, industrial companies have developed computational codes.",30979
code intends the to system. as physical to as be The close possible,The code intends to be as close as possible to the physical system.,30980
second concerns physical gap The system. the the and code the between,The second concerns the gap between the code and the physical system.,30981
These lead codes. time to improvements consuming,These improvements lead to time consuming codes.,30982
"meet would with difficulties, They substantial e.g.","They would meet with substantial difficulties, e.g.",30983
"demand excessive computational to and memory applied problems. when massive cost,","excessive memory demand and computational cost, when applied to massive problems.",30984
method present to We new tackle this problem. a,We present a new method to tackle this problem.,30985
CNNs. translation hypotheses invariance the of There regarding in multiple are source,There are multiple hypotheses regarding the source of translation invariance in CNNs.,30986
that is operation. to Another pooling invariance due is the possibility,Another possibility is that invariance is due to the pooling operation.,30987
a fundamental connections. is feature of Directionality network,Directionality is a fundamental feature of network connections.,30988
"nodes is neglected. when considerably differ However, hub directionality","However, hub nodes differ considerably when directionality is neglected.",30989
with Changes to firing in dopamine are compared due modeled experiment. rates loss,Changes in firing rates due to modeled dopamine loss are compared with experiment.,30990
and A the combined dynamics oscillations of paper discusses system. this companion,A companion paper discusses the dynamics and oscillations of this combined system.,30991
negative recognition of the effect has fingerprint systems. on Elastic fingerprints of performance a distortion,Elastic distortion of fingerprints has a negative effect on the performance of fingerprint recognition systems.,30992
effect This to in brings applications. negative authentication inconvenience users,This negative effect brings inconvenience to users in authentication applications.,30993
this methods aimed at addressing problem still Current have limitations.,Current methods aimed at addressing this problem still have limitations.,30994
"not requiring time to and they computation are rectify Secondly, significant efficient samples.","Secondly, they are not efficient and requiring significant computation time to rectify samples.",30995
loops. visual of framework simultaneous closing (SLAM) We surface localization mapping present a and,We present a visual simultaneous localization and mapping (SLAM) framework of closing surface loops.,30996
combines dense alignment. It matching surface both sparse and feature,It combines both sparse feature matching and dense surface alignment.,30997
object problem We the i.e. of referring investigate (OR),We investigate the problem of object referring (OR) i.e.,30998
localize to visual scene target language object a a description. with a in coming,to localize a target object in a visual scene coming with a language description.,30999
when object a the Humans gaze expression. issue referring at also they,Humans also gaze at the object when they issue a referring expression.,31000
videos and gaze. human in OR addresses with language This paper,This paper addresses OR in videos with language and human gaze.,31001
"context. gaze, effectively cues, results human method spatio-temporal that Experimental show our utilizes motion and","Experimental results show that our method effectively utilizes motion cues, human gaze, and spatio-temporal context.",31002
real-world large-scale vein of heavily sets. The data recognition finger algorithms on depends development,The development of finger vein recognition algorithms heavily depends on large-scale real-world data sets.,31003
"The interpolation is using SCO filter. resampling, compensated by then a time-domain","The SCO is then compensated by resampling, using a time-domain interpolation filter.",31004
fundamental a inference challenge. graphical function $Z$ is partition Computing discrete a the model of,Computing the partition function $Z$ of a discrete graphical model is a fundamental inference challenge.,31005
"used is are computationally this practice. in intractable, Since variational often approximations","Since this is computationally intractable, variational approximations are often used in practice.",31006
"lower variational transformations used to $Z$. so-called Recently, improve gauge were on bounds","Recently, so-called gauge transformations were used to improve variational lower bounds on $Z$.",31007
"nonsymmetric results demonstrate experimental of the for effectiveness generic, Our models. WMBE-G even","Our experimental results demonstrate the effectiveness of WMBE-G even for generic, nonsymmetric models.",31008
networks uses propose that novel extraction. (CNNs) a We feature neural for convolutional method,We propose a novel method that uses convolutional neural networks (CNNs) for feature extraction.,31009
select then to different used These to CNNs features. are train,These are then used to train different CNNs to select features.,31010
resting-state previously sex were We that novel not connectivity differences in reported. detected,We detected novel sex differences in resting-state connectivity that were not previously reported.,31011
additionally configuration the interact and with sexes. age These self-reports between basal divergently,These self-reports additionally divergently interact with age and the basal configuration between sexes.,31012
A detailed our evaluation of compares approach performance quality rendering and to lines. rasterization-based,A detailed performance and quality evaluation compares our approach to rasterization-based rendering of lines.,31013
The the IoT growth rapid mobile services. of of era future shaping is,The rapid growth of IoT era is shaping the future of mobile services.,31014
everything. information devices to heterogeneous connectivity a enables Advanced broadcast mobile where communication technology,Advanced communication technology enables a heterogeneous connectivity where mobile devices broadcast information to everything.,31015
to challenges. the We tackle these into intelligence IoT artificial integrate,We integrate the artificial intelligence into IoT to tackle these challenges.,31016
by networks. learn recurrent environment dynamics is idea to basic The,The basic idea is to learn environment dynamics by recurrent networks.,31017
of brings level method IoT a Our new intelligence.,Our method brings a new level of IoT intelligence.,31018
correct diagnosis important cancer is of Early a very and treatment. aspect,Early and correct diagnosis is a very important aspect of cancer treatment.,31019
then to of further be malignity Tumour determine the analysis computed can used tumour.,Tumour analysis further computed can then be used to determine malignity of the tumour.,31020
Semantic computer is vision. most tasks of challenged the one image in segmentation,Semantic image segmentation is one of the most challenged tasks in computer vision.,31021
"these pre-outputs get final concatenate Finally, we the to output.","Finally, we concatenate these pre-outputs to get the final output.",31022
"can is loss loss reduction way, the we this back the propagated. when reduce In","In this way, we can reduce the loss reduction when the loss is back propagated.",31023
"In up linearly way, to scale approach our can large this datasets.","In this way, our approach can linearly scale up to large datasets.",31024
square least-mean TS equalization. is a for This algorithm adaptive in phase-dependent decision-directed employed later,This TS is later employed in a phase-dependent decision-directed least-mean square algorithm for adaptive equalization.,31025
"are In and trees this in presence technology, other buildings. things detected of like vehicles","In this technology, vehicles are detected in presence of other things like trees and buildings.",31026
"Viola based vehicles on detected this In paper, Jones. boosting by the technique","In this paper, vehicles detected based on the boosting technique by Viola Jones.",31027
"network compact efficient, Our is light-weight. and","Our network is efficient, compact and light-weight.",31028
may to extra due and parameters selection variability have fitting. These model,These parameters may have extra variability due to model selection and fitting.,31029
"and beta reproducible. Of most powers were theta, band alpha, these,","Of these, theta, alpha, and beta band powers were most reproducible.",31030
"we access same ""reality"". not the At a priviledged to require do time,","At the same time, we do not require a priviledged access to ""reality"".",31031
"actions, to and e.g., ideas of animals. can even be systems, value also Our extended","Our ideas can also be extended to value systems, and even actions, e.g., of animals.",31032
"large a However, tors. efficiency of fac- number accuracy architecture ConvNet and on depend","However, ConvNet architecture efficiency and accuracy depend on a large number of fac- tors.",31033
rameters Bayesian of Gaussian Optimization determined by The Process prior with algorithm. hyperpa-,The hyperpa- rameters of determined by Bayesian Optimization with Gaussian Process prior algorithm.,31034
function layers. after activation non-linear the implemented is convolutional ReLu,ReLu non-linear activation function is implemented after the convolutional layers.,31035
in layers. is out downsampling points pooling to pooling Max the eration carried op- data,Max pooling op- eration is carried out to downsampling the data points in pooling layers.,31036
for optimizer backpropagation. with Adam is descent gradient used Mini-batch algorithm,Mini-batch gradient descent with Adam optimizer algorithm is used for backpropagation.,31037
classification report. Developed and architecture is validated with confusion matrix,Developed architecture is validated with confusion matrix and classification report.,31038
"The in updating for accounted model equation"". the term ""model discrepancy through formulation is our","The model discrepancy term is accounted for in our formulation through the ""model updating equation"".",31039
CDF candidate is selected. similar to signal most the The with received the,The candidate with the most similar CDF to the received signal is selected.,31040
between distance The average measure these is similarity the CDFs. of minimum,The measure of similarity is the minimum average distance between these CDFs.,31041
amplitude used commonly employed. digital Five coherent systems optical quadrature are in formats modulation,Five commonly used quadrature amplitude modulation formats in digital coherent optical systems are employed.,31042
signal-to-noise Results interest. ratios algorithm that proposed of at accurate the classification show optical achieves,Results show that the proposed algorithm achieves accurate classification at optical signal-to-noise ratios of interest.,31043
"Furthermore, carrier require not recovery. it does","Furthermore, it does not require carrier recovery.",31044
restore proportions. are the morphed combined order facial and further These then original to in,These are then combined and further morphed in order to restore the original facial proportions.,31045
"Therefore, segmentation to is this needed instance address problem.","Therefore, instance segmentation is needed to address this problem.",31046
instance the segmented are together. instances within by linking out Text pixels same first,Text instances are first segmented out by linking pixels within the same instance together.,31047
without directly from then bounding segmentation extracted location the regression. are boxes Text result,Text bounding boxes are then extracted directly from the segmentation result without location regression.,31048
several This and conception insights life yields conjectures. of new,This conception of life yields several new insights and conjectures.,31049
"deterioration leaky DLMS performance may in algorithm the However, sparse the from system. suffer","However, the leaky DLMS algorithm may suffer from performance deterioration in the sparse system.",31050
"state proposed we the the for theoretical implement analysis algorithms. Moreover, steady","Moreover, we implement the steady state theoretical analysis for the proposed algorithms.",31051
the TS is also arrangement carrier used offset. the The to of frequency estimate,The arrangement of the TS is also used to estimate the carrier frequency offset.,31052
for networks. demonstrate Here theoretically we frequency-independent synthesizing nonreciprocal framework a multi-port,Here we demonstrate a framework for synthesizing theoretically frequency-independent multi-port nonreciprocal networks.,31053
"three architectures: intelligence genres fact, there of In (e.g. exist logics","In fact, there exist three genres of intelligence architectures: logics (e.g.",31054
"of Bayes, HMM}), other. are incompatible which \textit{Naive to each all","\textit{Naive Bayes, HMM}), all of which are incompatible to each other.",31055
"with model iGraph, By principle. of a design paradigm the semantic recommendation we","By the paradigm of iGraph, we design a recommendation model with semantic principle.",31056
"the Second, manner the probabilistic probabilities. features, infers in graph distributions of the of","Second, the probabilistic graph infers the distributions of features, in the manner of probabilities.",31057
"Experimentally, the state-of-the-art we conclusions. and baselines beat our verify","Experimentally, we beat the state-of-the-art baselines and verify our conclusions.",31058
noise cancel to filters communication received use in systems Conventional a signal.,Conventional communication systems use filters to cancel noise in a received signal.,31059
"provided. discussed comprehensive To a are also overview, denoising the of establish comparison techniques performance","To establish comprehensive overview, a performance comparison of the discussed denoising techniques are also provided.",31060
"results achieved deep very Recently, promising visual object learning in has tracking.","Recently, deep learning has achieved very promising results in visual object tracking.",31061
"two In update appearance dictionaries online of addition, to changes objects. handle we target","In addition, we online update two dictionaries to handle appearance changes of target objects.",31062
uncertainty spherical channel imperfect a is assume the that We and CSI region. bounded by,We assume imperfect CSI and that the channel uncertainty is bounded by a spherical region.,31063
lacks that Alzheimer's (AD) options. a treatment neurodegenerative Disease effective disorder is,Alzheimer's Disease (AD) is a neurodegenerative disorder that lacks effective treatment options.,31064
Two proposed mutations of binding. stabilized out four,Two out of four proposed mutations stabilized binding.,31065
the minimization (ADMM) direction method. proposed method is solve employed to The alternative,The alternative direction minimization method (ADMM) is employed to solve the proposed method.,31066
and mouse simulations evaluate method. are proposed numerical the perform Both studies to real,Both numerical simulations and real mouse studies are perform to evaluate the proposed method.,31067
stochastic intensive process approximating tool is computationally computer for powerful emulation a Gaussian models.,Gaussian stochastic process emulation is a powerful tool for approximating computationally intensive computer models.,31068
"a is estimation However, task. challenging of in parameters the GaSP emulator","However, estimation of parameters in the GaSP emulator is a challenging task.",31069
improvements inter-layer This style that yields in methods. demonstrates visible paper transfer correlations controlling,This paper demonstrates that controlling inter-layer correlations yields visible improvements in style transfer methods.,31070
"gram than achieve control this We rather by computing matrices. within-layer, cross-layer,","We achieve this control by computing cross-layer, rather than within-layer, gram matrices.",31071
find to We cross-layer are gram (a) that statistics. matrices within-layer sufficient control,We find that (a) cross-layer gram matrices are sufficient to control within-layer statistics.,31072
improves style transfer Inter-layer correlations texture synthesis. and,Inter-layer correlations improves style transfer and texture synthesis.,31073
"(e.g. paper examples style shows The problems numerous ""hard"" on real transfer","The paper shows numerous examples on ""hard"" real style transfer problems (e.g.",31074
been convergence rotations in adaptations properties. different have resulting literature the in Several proposed with,Several rotations have been proposed in the literature resulting in adaptations with different convergence properties.,31075
in estimating special symbol of training utilized The is the the also CFO. structure,The special structure of the training symbol is also utilized in estimating the CFO.,31076
to tailored junctions disambiguate specifically without is Our algorithm sacrificing quality.,Our algorithm is tailored specifically to disambiguate junctions without sacrificing quality.,31077
percutaneous X-ray for orthopaedic a guidance surgical cornerstone is Fluoroscopic procedures.,Fluoroscopic X-ray guidance is a cornerstone for percutaneous orthopaedic surgical procedures.,31078
"of anatomy from observations effects of simplification. However, the suffer the two-dimensional three-dimensional projective","However, two-dimensional observations of the three-dimensional anatomy suffer from the effects of projective simplification.",31079
complex decision comparisons into breaks Evaluating ones. down tractable problems such pairwise,Evaluating pairwise comparisons breaks down such complex decision problems into tractable ones.,31080
"there PCMs. significant However, using drawbacks are of two","However, there are two significant drawbacks of using PCMs.",31081
"First, evaluate in humans an manner. PCMs inconsistent","First, humans evaluate PCMs in an inconsistent manner.",31082
show `consistent' to PCMs MERWs. balanced detailed only We that correspond,We show that only `consistent' PCMs correspond to detailed balanced MERWs.,31083
filled We in use an incompletely also to propose AHP. approach PCMs,We also propose an approach to use incompletely filled PCMs in AHP.,31084
as avenues are discussed Potential well. future,Potential future avenues are discussed as well.,31085
"analytic keywords: maximum chains, hierarchy entropy markov process, .","keywords: analytic hierarchy process, markov chains, maximum entropy .",31086
Both but affect interactions were digital. that physical now once,Both affect interactions that were once physical but now digital.,31087
neural image the define (CNNs) for current state-of-the-art Convolutional networks recognition.,Convolutional neural networks (CNNs) define the current state-of-the-art for image recognition.,31088
learning This humans trust models. deep will and between increase,This will increase trust between humans and deep learning models.,31089
Most tasks. detection existing separate treat as methods recognition and text,Most existing methods treat text detection and recognition as separate tasks.,31090
"is share and introduced Specially, convolutional RoIRotate between recognition. features to detection","Specially, RoIRotate is introduced to share convolutional features between detection and recognition.",31091
quantity estimation. This uncertainty as us treat a during allows latent macroeconomic to,This allows us to treat macroeconomic uncertainty as a latent quantity during estimation.,31092
"and short-term increase. equity exports to levels prices, tend while interest decline, Moreover, rates unemployment","Moreover, equity prices, short-term interest rates and exports tend to decline, while unemployment levels increase.",31093
this have Several proposed domain. methods been in,Several methods have been proposed in this domain.,31094
"pre-trained network, We i.e. a convolutional adapt","We adapt a pre-trained convolutional network, i.e.",31095
network samples. only We by train this end-to-end a few using training,We train this network end-to-end by using only a few training samples.,31096
search engine deployed Our commercially Web-based application. is available through a and,Our search engine is commercially deployed and available through a Web-based application.,31097
consists and A sentence of corresponding a a Schema question. Winograd,A Winograd Schema consists of a sentence and a corresponding question.,31098
"answer and use successfully the these reasoning. requires questions, commonsense To of knowledge one","To successfully answer these questions, one requires the use of commonsense knowledge and reasoning.",31099
"answering templates, stored in the used to be are later for questions. WSC These","These are stored in templates, to be later used for answering the WSC questions.",31100
to This WSC approach well tasks. works subset respect a with of,This approach works well with respect to a subset of WSC tasks.,31101
"physician. to immediately Many the images, however, diagnostic simply informative lack features","Many images, however, simply lack diagnostic features immediately informative to the physician.",31102
non-diagnostic. images classify automatically diagnostic or We as to sought,We sought to automatically classify images as diagnostic or non-diagnostic.,31103
is the the by images provided all truth pathologist. for The ground,The ground truth for all the images is provided by the pathologist.,31104
and us train enabled This cost larger that models predictive lower performance. better have to,This enabled us to train larger models that have lower cost and better predictive performance.,31105
The of function represents random the characteristic complete distribution the of variable. characterization,The characteristic function represents complete characterization of the distribution of the random variable.,31106
image Single haze and challenging problem. a ill-posed very is removal,Single image haze removal is a very challenging and ill-posed problem.,31107
are image hazy input the haze-free to the from image. These used compute,These are used to compute the haze-free image from the hazy input image.,31108
"unresolved potential yet neuroscience, challenges In the two of breakthroughs. field rapidly-evolving are network","In the rapidly-evolving field of network neuroscience, two yet unresolved challenges are potential breakthroughs.",31109
"and First, be noninvasive techniques. estimated use functional neuroimaging brain should networks easy to from","First, functional brain networks should be estimated from noninvasive and easy to use neuroimaging techniques.",31110
field. the community that the to contributions describe brought this SP major We research,We describe the major contributions that the SP community brought to this research field.,31111
"in also brain both normal and results report obtained pathological in We states. concrete applications,","We also report results obtained in concrete applications, in both normal and pathological brain states.",31112
. in eventually and of applications processing methods signal term are Future directions provided,Future directions in term of signal processing methods and applications are eventually provided .,31113
"for and translation, rotation Advanced invariance. models provide size,","Advanced models provide for translation, size, and rotation invariance.",31114
"the size, mechanism translation, provided Further, and perceptual for rotation invariance.","Further, the perceptual mechanism provided for translation, size, and rotation invariance.",31115
"comparable GN the the literature are Furthermore, published models against model. benchmarked ISRS in","Furthermore, comparable models published in the literature are benchmarked against the ISRS GN model.",31116
"CCL first we CCLM. CC, relationship this of To CCC, explain the end, and","To this end, we first explain the relationship of CC, CCC, CCL and CCLM.",31117
events between differentiation and results difficulty in events. Preliminary automatic the demonstrated independent seizure slowing,Preliminary results demonstrated the difficulty in automatic differentiation between seizure events and independent slowing events.,31118
aspects. sub from network Each extract different discriminative features could,Each sub network could extract discriminative features from different aspects.,31119
information. which both temporal and subnet a could spatial explore is network first two-stream Our,Our first subnet is a two-stream network which could explore both temporal and spatial information.,31120
a subnets the wellperformance and guaranteed diversity. between differences Each subnet basenet shares,Each subnet shares a wellperformance basenet and differences between subnets guaranteed the diversity.,31121
made the deep progress Edge help of Convolutional with Networks has (ConvNet). significant detection,Edge detection has made significant progress with the help of deep Convolutional Networks (ConvNet).,31122
human have level approached edge performance based ConvNet detectors These standard on benchmarks.,These ConvNet based edge detectors have approached human level performance on standard benchmarks.,31123
detectors' outputs. study We of systematical provide a these,We provide a systematical study of these detectors' outputs.,31124
be taxes and can to carryover strategy accommodate lottery adjusted pools. The,The strategy can be adjusted to accommodate lottery taxes and carryover pools.,31125
assume We independently numbers bettors small other that of the tickets. buy,We assume that the other bettors independently buy small numbers of tickets.,31126
in geometry triangle great is importance Information graphics transfer between computer of and processing. meshes,Information transfer between triangle meshes is of great importance in computer graphics and geometry processing.,31127
"it receive information the is in method input, both as Our general e.g. can","Our method is general both in the information it can receive as input, e.g.",31128
"always function is it but This necessarily not exists, fitness unique.","This fitness function always exists, but it is not necessarily unique.",31129
numerical for conclusions. We support present extensive our,We present extensive numerical support for our conclusions.,31130
other is than non-maximum No efficient suppression involved. post-processing an,No post-processing other than an efficient non-maximum suppression is involved.,31131
public the have on evaluated proposed We datasets. TextBoxes++ four,We have evaluated the proposed TextBoxes++ on four public datasets.,31132
"In text outperforms of TextBoxes++ competing experiments, in runtime. all methods accuracy localization and terms","In all experiments, TextBoxes++ outperforms competing methods in terms of text localization accuracy and runtime.",31133
"either exact rely photographs. Existing edge or on of maps, need retrieval approaches existing","Existing approaches either need exact edge maps, or rely on retrieval of existing photographs.",31134
We the of comparisons traditional to using Gibbs et samplers dataset provide al. Efron diabetes,We provide comparisons to traditional Gibbs samplers using the diabetes dataset of Efron et al.,31135
classification has and and identification an been so task. far important Plant difficult,Plant classification and identification has so far been an important and difficult task.,31136
extraction. in for prepared be five is to input feature image steps The first pre-processed,The input image is first pre-processed in five steps to be prepared for feature extraction.,31137
their photorealistic analysis. Recovering facial aids perception artistic and from human the latent portraits faces,Recovering the latent photorealistic faces from their artistic portraits aids human perception and facial analysis.,31138
method Network Discriminative two (DN). of and (SRN) IFRP consists Network Style components: Removal Our,Our IFRP method consists of two components: Style Removal Network (SRN) and Discriminative Network (DN).,31139
power generation has important and for side electricity scheduling. management price become Short-term demand forecasting,Short-term electricity price forecasting has become important for demand side management and power generation scheduling.,31140
least localization In of the frequently linear solution field is the used. square,In the field of localization the linear least square solution is frequently used.,31141
fails solution Gaussian Unfortunately to this it comes if non noise.,Unfortunately this solution fails if it comes to non Gaussian noise.,31142
"video frame multi-scale in a segmented superpixels. Initially, a is into foreground mask","Initially, a foreground mask in a video frame is segmented into multi-scale superpixels.",31143
"performance degrades visible faces. their occluded and Hence, for partially","Hence, their performance degrades for partially visible and occluded faces.",31144
and We the datasets standard use facial attribute LFWA evaluations. for CelebA,We use the CelebA and LFWA facial attribute datasets for standard evaluations.,31145
that recent methods shows for Our outperforms faces. other evaluation significantly SPLITFACE especially partial,Our evaluation shows that SPLITFACE significantly outperforms other recent methods especially for partial faces.,31146
framework novel handle a We this to denoising problem. propose,We propose a novel denoising framework to handle this problem.,31147
signal supervisory Our unknown training. from leveraging allows during poses approach observations as multi-view,Our approach allows leveraging multi-view observations from unknown poses as supervisory signal during training.,31148
and (CAD) Diagnosis is for Breast BUS systems. image ultrasound challenging Computer-Aided (BUS) segmentation critical,Breast ultrasound (BUS) image segmentation is challenging and critical for BUS Computer-Aided Diagnosis (CAD) systems.,31149
provides representing of conceptual way a of highly framework influential spaces knowledge. The geometric,The highly influential framework of conceptual spaces provides a geometric way of representing knowledge.,31150
"vary supervised, These from unsupervised based. graph-theory and to methods to statistics from","These methods vary from unsupervised to supervised, and from statistics to graph-theory based.",31151
"proposed learning. to generation achieve selection sample and better feature are manifold Furthermore,","Furthermore, feature generation and sample selection are proposed to achieve better manifold learning.",31152
of a of different are consciousness heterogeneous Disorders or diseases mixture injuries.,Disorders of consciousness are a heterogeneous mixture of different diseases or injuries.,31153
"accurate, and novel interpretable. We robust, that model suggest therefore prognostic this is","We therefore suggest that this novel prognostic model is accurate, robust, and interpretable.",31154
"states were simulator. memory the the of with this, states associated (RAM) For the","For this, the states were associated with the (RAM) memory states of the simulator.",31155
"but using instead. problem In the work, the this screen same we consider planning","In this work, we consider the same planning problem but using the screen instead.",31156
missing Tensor is of elements tensors. technique filling completion of a data incomplete the,Tensor completion is a technique of filling missing elements of the incomplete data tensors.,31157
on optimization as the It such nuclear-norm minimization. studied convex based scheme being actively,It being actively studied based on the convex optimization scheme such as nuclear-norm minimization.,31158
acoustic demonstrates lines. first based switched work low-loss non-reciprocal This on network the delay,This work demonstrates the first non-reciprocal network based on switched low-loss acoustic delay lines.,31159
results verify Numerical effectiveness the efficiency approach. of the and,Numerical results verify the effectiveness and efficiency of the approach.,31160
"multispectral paper, image we compression propose method In this cameras. single-sensor an for efficient","In this paper, we propose an efficient image compression method for single-sensor multispectral cameras.",31161
multispectral data before encodes proposed the The captured method interpolation.,The proposed method encodes the captured multispectral data before interpolation.,31162
We spectral images. a the transform compression also method new for multispectral propose mosaicked of,We also propose a new spectral transform method for the compression of mosaicked multispectral images.,31163
our analysis. theoretical experimental Our results corroborate,Our experimental results corroborate our theoretical analysis.,31164
"called we neural network this (Badrinarayanan In state-of-the-art work, convolutional architecture, a SegNet employ et.","In this work, we employ a state-of-the-art convolutional neural network architecture, called SegNet (Badrinarayanan et.",31165
and both images sparsity Diffusion-weighted transform Methods: exhibit low-rankness.,Methods: Diffusion-weighted images exhibit both transform sparsity and low-rankness.,31166
have neural stereo networks performance excellent Deep for matching. shown,Deep neural networks have shown excellent performance for stereo matching.,31167
analysis functions includes analysis for survival package as univariate The also tools. complementary,The package also includes functions for univariate survival analysis as complementary analysis tools.,31168
"the As of this replay. implementation in-GPU know, experience far first is I as","As far as I know, this is the first in-GPU implementation of experience replay.",31169
visual when Even these not properties are (e.g.,Even when these properties are not visual (e.g.,31170
"property), their infer a is we visually. softness may still physical presence","softness is a physical property), we may still infer their presence visually.",31171
We attributes. material to as refer visual material such properties,We refer to such material properties as visual material attributes.,31172
"however, at fashion scale, is such a annotations in challenging. Obtaining consistent","Obtaining such annotations in a consistent fashion at scale, however, is challenging.",31173
estimation Discrete are Models literature. structural Dynamic the in (DDCMs) Choice important,Dynamic Discrete Choice Models (DDCMs) are important in the structural estimation literature.,31174
our baselines. simple experimental Our results the PQT algorithm show significantly outperforms that,Our experimental results show that our simple PQT algorithm significantly outperforms the baselines.,31175
of Contrast ultrasound by speckle. adversely of excessive quality affected images presence the are and,Contrast and quality of ultrasound images are adversely affected by the excessive presence of speckle.,31176
"inherent being However, characterization property, helps imaging speckle and tissue an in tracking.","However, being an inherent imaging property, speckle helps in tissue characterization and tracking.",31177
"despeckling any images oversmoothing. requires without of the speckle reduction ultrasound Thus, the of extent","Thus, despeckling of the ultrasound images requires the reduction of speckle extent without any oversmoothing.",31178
Experimental is able approaches. that despeckling DRNN state-of-the-art show evaluations outperform the proposed to the,Experimental evaluations show that the proposed DRNN is able to outperform the state-of-the-art despeckling approaches.,31179
"publicly-available over significant on We improvements datasets, evaluate showing state-of-the-art. two the","We evaluate on two publicly-available datasets, showing significant improvements over the state-of-the-art.",31180
Deep achieved solving strides great reinforcement challenging learning in motion has control tasks.,Deep reinforcement learning has achieved great strides in solving challenging motion control tasks.,31181
on several success have detection object achieved detection benchmarks. remarkable Deep systems large-scale object CNN-based,Deep CNN-based object detection systems have achieved remarkable success on several large-scale object detection benchmarks.,31182
addresses into this detectors. by classifiers work image-level Previous transforming issue object,Previous work addresses this issue by transforming image-level classifiers into object detectors.,31183
this utilized We scale-invariant property descriptor. construct to a have,We have utilized this property to construct a scale-invariant descriptor.,31184
development and management for RATs. has requirement the of emphasized control diverse This unified,This development has emphasized the requirement for unified control and management of diverse RATs.,31185
the overall of to utilization may suboptimal resources. network lead This,This may lead to suboptimal utilization of the overall network resources.,31186
Our a in applications agnostic provides framework deployment of also for a RAT architecture fashion.,Our architecture also provides a framework for deployment of applications in a RAT agnostic fashion.,31187
Experimental our provided architecture. the results the platform benefits obtained demonstrate using by,Experimental results obtained using the platform demonstrate the benefits provided by our architecture.,31188
It system also results in improved performance.,It also results in improved system performance.,31189
We formalization second categorical and computational the present of Kant's a implementation formulation of imperative.,We present a formalization and computational implementation of the second formulation of Kant's categorical imperative.,31190
causally this how actions. terms in Here affected persons by we principle are of interpret,Here we interpret this principle in terms of how persons are causally affected by actions.,31191
used rich Observable Processes in (POMDPs) learning. often machine are environments Decision Markov Partially,Partially Observable Markov Decision Processes (POMDPs) are rich environments often used in machine learning.,31192
issue POMDPs studied. of causal has in information and little been But relatively structures the,But the issue of information and causal structures in POMDPs has been relatively little studied.,31193
"and uncertainty, learning. allows This a understanding better of POMDP information,","This allows a better understanding of POMDP uncertainty, information, and learning.",31194
allow a distributions set distributions. Compound construction rich of of,Compound distributions allow construction of a rich set of distributions.,31195
a to use the Here to that compound we approximation quadrature family. quadrature define integral,Here we use a quadrature approximation to that integral to define the quadrature compound family.,31196
(Poisson discrete technique distributions. applied LogNormal) continuous is and to This,This technique is applied to discrete (Poisson LogNormal) and continuous distributions.,31197
accuracy and such properties We performance in extensive as estimation forecasting. to simulations run investigate,We run extensive simulations to investigate properties such as estimation accuracy and performance in forecasting.,31198
We direct alone causes find spatially-smoothed coupling. without spatial that pattern even noise formation,We find that spatially-smoothed noise alone causes pattern formation even without direct spatial coupling.,31199
of Gaussian training are using in sets patches. hyperspectral correlated inferred Processes clusters spatio-spectrally The,The Gaussian Processes are inferred in sets using clusters of spatio-spectrally correlated hyperspectral training patches.,31200
RGB the image. spectral quantization is set of Each transformed test match the to,Each set is transformed to match the spectral quantization of the test RGB image.,31201
paper single consider depth this image estimation. In of we problem monocular the,In this paper we consider the problem of single monocular image depth estimation.,31202
belongs method interactive families: to learning-based two efforts and method. Previous roughly,Previous efforts belongs roughly to two families: learning-based method and interactive method.,31203
unfamiliar generalization ability perform for low But poorly typically they and scenes. suffer,But they suffer low generalization ability and typically perform poorly for unfamiliar scenes.,31204
"expensive nature time-consuming. makes Besides, such and method data aquisition data-hungry for","Besides, data-hungry nature for such method makes data aquisition expensive and time-consuming.",31205
"Interactive of depth which, requires is variance. errorneous and large method of however, annotation human","Interactive method requires human annotation of depth which, however, is errorneous and of large variance.",31206
object require of sparse than size for raw Our depth. real-world label method rather,Our method require sparse label for real-world size of object rather than raw depth.,31207
map geometric to inferred following size according depth relationships Coarse A labels. then is,A Coarse depth map is then inferred following geometric relationships according to size labels.,31208
on map conditional random refine optimization depth doing energy function Then the we field(CRF). by,Then we refine the depth map by doing energy function optimization on conditional random field(CRF).,31209
a tool image became powerful deep model for learning extraction. feature Recently,Recently deep learning model became a powerful tool for image feature extraction.,31210
"we salient multi-scale network (MSDNN) this object a for paper, detection. propose In deep neural","In this paper, we propose a multi-scale deep neural network (MSDNN) for salient object detection.",31211
evaluated extensively datasets. four object proposed is salient detection on model benchmark The,The proposed model is extensively evaluated on four salient object detection benchmark datasets.,31212
"considered we currents increased beam the strategies. of for Additionally, each","Additionally, we considered increased beam currents for each of the strategies.",31213
a training the learner learners. previous reweighted sample receives Each from,Each learner receives a reweighted training sample from the previous learners.,31214
"two Further, the in functions which our loss diversity we increase propose ensemble.","Further, we propose two loss functions which increase the diversity in our ensemble.",31215
be initialization applied for can These training. functions weight during or loss either,These loss functions can be applied either for weight initialization or during training.,31216
and to this provide led that demonstrate that an this design. improves intuition performance We,We demonstrate that this improves performance and provide an intuition that led to this design.,31217
"object COCO Imagenet on performance our image detection, classification, measure segmentation. VOC We","We measure our performance on Imagenet classification, COCO object detection, VOC image segmentation.",31218
"regularization near more For and also entropic plausible less meant permutations images. OT,","For OT, less entropic regularization also meant near permutations and more plausible images.",31219
"network. a to a random method, statistical results namely convolutional are compared Synthesis","Synthesis results are compared to a statistical method, namely a random convolutional network.",31220
considered. is The data in augmentation of space feature problem,The problem of data augmentation in feature space is considered.,31221
manifold and parametrization the terms of pose in This of exploits architecture appearance. pose a,This architecture exploits a parametrization of the pose manifold in terms of pose and appearance.,31222
"data feature using transfer space (w.r.t. for notably, by augmentation Most","Most notably, by using feature space transfer for data augmentation (w.r.t.",31223
method as signals. spatially-evolving the set image The of histopathological multidimensional considers proposed a,The proposed method considers the histopathological image as a set of multidimensional spatially-evolving signals.,31224
accurate samples. aid in pathologist biopsy system and of diagnosis This reliable can the,This system can aid the pathologist in accurate and reliable diagnosis of biopsy samples.,31225
each contains same We number the posit site individuals. that of,We posit that each site contains the same number of individuals.,31226
"the actuating. thing of processing, ability and/or A sensing, smart communicating has","A smart thing has the ability of sensing, processing, communicating and/or actuating.",31227
an agent. of embodied called is agent This commonly type,This type of agent is commonly called an embodied agent.,31228
variety are realistic a anomalies. able to Surveillance videos of capture,Surveillance videos are able to capture a variety of realistic anomalies.,31229
"by both and exploiting to we videos. anomalous learn anomalies this propose normal paper, In","In this paper, we propose to learn anomalies by exploiting both normal and anomalous videos.",31230
video-level the at labels instead of are (anomalous or training normal) clip-level.,the training labels (anomalous or normal) are at video-level instead of clip-level.,31231
used two This dataset tasks. be for can,This dataset can be used for two tasks.,31232
results We anomalous baselines on learning several of recent activity deep the recognition. provide,We provide the results of several recent deep learning baselines on anomalous activity recognition.,31233
grouped heterogeneity pattern. individual using a We model,We model individual heterogeneity using a grouped pattern.,31234
"the common each allow For coefficients. breaks structural in we group,","For each group, we allow common structural breaks in the coefficients.",31235
"differ timing, number, can groups. the breaks across these However, of size and","However, the number, timing, and size of these breaks can differ across groups.",31236
good finite proposed in method results the Carlo Monte performance demonstrate samples. of the,Monte Carlo results demonstrate the good performance of the proposed method in finite samples.,31237
"networks are their poorly understood. Despite deep prevalence,","Despite their prevalence, deep networks are poorly understood.",31238
"at This to due, parameterized their part, in nature. is least highly","This is due, at least in part, to their highly parameterized nature.",31239
We complexity. across data this varying analysis sets of four apply,We apply this analysis across four data sets of varying complexity.,31240
ATP. force motive synthases synthesize ATP proton to a utilize,ATP synthases utilize a proton motive force to synthesize ATP.,31241
"membrane. enzymes membrane-embedded hydrolyze pump ATP protons also can the over these reverse, to In","In reverse, these membrane-embedded enzymes can also hydrolyze ATP to pump protons over the membrane.",31242
in clinical daily one most examinations radiological the X-rays of are common Chest routines.,Chest X-rays are one of the most common radiological examinations in daily clinical routines.,31243
diseases chest X-rays is an task trainees. entry-level for often radiologist thorax Reporting using,Reporting thorax diseases using chest X-rays is often an entry-level task for radiologist trainees.,31244
"TieNet system. the chest transform a we Furthermore, X-ray reporting into","Furthermore, we transform the TieNet into a chest X-ray reporting system.",31245
output reporting disease report simulates together. and and the It classification can a preliminary process,It simulates the reporting process and can output disease classification and a preliminary report together.,31246
"in model errors Subsequently, rectify auxiliary pre-trained fine-tuned is manual labels. on labels to the","Subsequently, the pre-trained model is fine-tuned on manual labels to rectify errors in auxiliary labels.",31247
of techniques. an essential Arithmetic coding class is coding,Arithmetic coding is an essential class of coding techniques.,31248
"However, increases exponentially of the complexity of with LUT the length context.","However, the complexity of LUT increases exponentially with the length of context.",31249
"to performance. such compression modeling the context, which inevitably restricts Thus, are solutions limited large","Thus, such solutions are limited to modeling large context, which inevitably restricts the compression performance.",31250
the extraction Functional modular commonly used is to brain's explore subnetwork structure.,Functional subnetwork extraction is commonly used to explore the brain's modular structure.,31251
is network neurons. composed linear non-saturating The threshold of,The network is composed of non-saturating linear threshold neurons.,31252
The constraint by of neurons. the direction exploration is steered,The direction of exploration is steered by the constraint neurons.,31253
then The reconstruct is to remaining images. the sub-aperture used embedding super-resolved,The super-resolved embedding is then used to reconstruct the remaining sub-aperture images.,31254
post-processing using improved step. be a back-projection This performance can iterative further as,This performance can be further improved using iterative back-projection as a post-processing step.,31255
with compare our data several three state-of-the-art sets. method benchmark methods We on,We compare our method with several state-of-the-art methods on three benchmark data sets.,31256
problem. address and (DMAS) Multiply Delay introduced was this Sum to,Delay Multiply and Sum (DMAS) was introduced to address this problem.,31257
"and numerically is (DS-DMAS), method, introduced The namely Stage Double DMAS evaluated experimentally.","The introduced method, namely Double Stage DMAS (DS-DMAS), is evaluated numerically and experimentally.",31258
of as processing as requires well image natural language It both processing. expertise,It requires expertise of both image processing as well as natural language processing.,31259
paper discuss This image task. available for different captioning models about,This paper discuss about different available models for image captioning task.,31260
how In be that can model we discussed addition have to this implemented.,In addition to that we have discussed how this model can be implemented.,31261
"also the model In we matrices. end, the performance of have standard evaluation evaluated using","In the end, we have also evaluated the performance of model using standard evaluation matrices.",31262
"very errors. is synchronization to performance However, systems sensitive of the OFDM","However, the performance of OFDM systems is very sensitive to synchronization errors.",31263
components from made typically of simple regardless are complexity. Systems their,Systems are typically made from simple components regardless of their complexity.,31264
in farms. remains unstructured of as challenging environments Accurate highly such crop localisation,Accurate localisation of crop remains highly challenging in unstructured environments such as farms.,31265
"tracking. of consists detection, This parts: system estimation, quality three and","This system consists of three parts: detection, quality estimation, and tracking.",31266
using achieved detection the Efficient FasterRCNN is framework.,Efficient detection is achieved using the FasterRCNN framework.,31267
"of the evaluate authors do map. focus not the efficiency However,","However, the authors do not evaluate the efficiency of focus map.",31268
for paper evaluating purposed a This method it.,This paper purposed a method for evaluating it.,31269
different of several keywords. questions kinds test We generate to,We generate several kinds of questions to test different keywords.,31270
"is there not dataset. benchmark However, a","However, there is not a benchmark dataset.",31271
labels evaluation. attribute nine sequence the has based attribute for Each,Each sequence has nine attribute labels for the attribute based evaluation.,31272
research. for provide guidelines of The findings experiments these some future,The findings of these experiments provide some guidelines for future research.,31273
"images. we In introduce cross-spectral align to novel method deep-learning paper, this a","In this paper, we introduce a novel deep-learning method to align cross-spectral images.",31274
to on spectra. descriptor different approach a Our learned relies which invariant is,Our approach relies on a learned descriptor which is invariant to different spectra.,31275
Most aspects. three insufficient all models existing are in,Most existing models are insufficient in all three aspects.,31276
sampling fidelity. diverse of perform images FusedGAN high very controllable with can The,The FusedGAN can perform controllable sampling of diverse images with very high fidelity.,31277
argue achieved generation stages. that can the disentangling controllability process into by We be various,We argue that controllability can be achieved by disentangling the generation process into various stages.,31278
half The detection period the of is signal. input period the,The detection period is the half period of the input signal.,31279
method of the sequence. three-phase signal's independent and is This phase unbalance,This method is independent of the three-phase signal's unbalance and phase sequence.,31280
or was signals. the The through proposed method experiments unbalanced verified single three-phase for,The proposed method was verified through experiments for the single or three-phase unbalanced signals.,31281
semantic for a by inferring approach novel hierarchical synthesis layout. text-to-image We propose,We propose a novel hierarchical approach for text-to-image synthesis by inferring semantic layout.,31282
"dynamic over the long Despite term. maintained chromatin certain marks nature, must be their","Despite their dynamic nature, certain chromatin marks must be maintained over the long term.",31283
of of context in is concept the innateness intelligence. discussed rarely The artificial,The concept of innateness is rarely discussed in the context of artificial intelligence.,31284
"that multiple these for reasons. are argue I overstated, claims","I argue that these claims are overstated, for multiple reasons.",31285
that bridges two. present We an the approach,We present an approach that bridges the two.,31286
critical facilitate very networks is of in training aggregation deep end-to-end Such to manner. an,Such aggregation is critical to facilitate training of very deep networks in an end-to-end manner.,31287
While aggregation works investigate alternative (e.g. operations subsequent,While subsequent works investigate alternative aggregation operations (e.g.,31288
is abating Neurostimulation epileptic therapy a promising for seizures.,Neurostimulation is a promising therapy for abating epileptic seizures.,31289
"stimulation extremely However, is difficult experimentally. patterns optimal it identify to","However, it is extremely difficult to identify optimal stimulation patterns experimentally.",31290
activity this seizure-like in-silico Spontaneous network. in induced is reconstructed neuronal,Spontaneous seizure-like activity is induced in-silico in this reconstructed neuronal network.,31291
used any to not periodic Commonly seizures abate were trains frequency. able permanently at,Commonly used periodic trains were not able to permanently abate seizures at any frequency.,31292
effector the upper lower for cells bounds the We find population. and immune,We find the upper and lower bounds for the effector immune cells population.,31293
of considered implications are Biological our results .,Biological implications of our results are considered .,31294
(PA). nonlinear distortion in wireless transmitters is primary source The amplifier the power of,The primary source of nonlinear distortion in wireless transmitters is the power amplifier (PA).,31295
We this to present a low-shot uses approach novel idea. that learning,We present a novel approach to low-shot learning that uses this idea.,31296
"be placed performance. increased integrated seamlessly for Moreover, manually framework landmarks in can the","Moreover, manually placed landmarks can be seamlessly integrated in the framework for increased performance.",31297
deep classification. solution feature of one-class problem in the propose learning-based for a We learning,We propose a deep learning-based solution for the problem of feature learning in one-class classification.,31298
is testing template framework facilitate matching-based the to process. A introduced,A template matching-based framework is introduced to facilitate the testing process.,31299
employ hand-crafted recognition features. of for existing methods Most the periocular,Most of the existing methods for periocular recognition employ hand-crafted features.,31300
"biochemical, cells by Metabolic governed economic physiological, are in physical, principles. and fluxes","Metabolic fluxes in cells are governed by physical, biochemical, physiological, and economic principles.",31301
"additive, usually cost"". are flows due costs two an ""compromise The of additional not to","The costs of two flows are usually not additive, due to an additional ""compromise cost"".",31302
"where directions, enzymatic cost jump. shows fluxes polytopes, change their flux Between a the","Between flux polytopes, where fluxes change their directions, the enzymatic cost shows a jump.",31303
"of polytope, flux The the i.e. solutions are distributions in flux corners","The solutions are flux distributions in corners of the flux polytope, i.e.",31304
adaptive margin proposed angular An to face the features. of capability is enhance further discrimination,An adaptive angular margin is further proposed to enhance the discrimination capability of face features.,31305
vs. vs. or normal non-seizure seizure ictal.,seizure vs. non-seizure or normal vs. ictal.,31306
not well ternary classifying case do perform methods These e.g. when,These methods do not perform well when classifying ternary case e.g.,31307
"number the CNN large the learnable In of model, is a parameters. bottleneck","In a CNN model, the bottleneck is the large number of learnable parameters.",31308
Deep the impressive learning-based for of methods images. results medical achieved segmentation,Deep learning-based methods achieved impressive results for the segmentation of medical images.,31309
"we the influence initial different learning Furthermore, rates. compared of","Furthermore, we compared the influence of different initial learning rates.",31310
role multivariate The in expansion statistics. functional of a plays ANOVA mapping fundamental a,The functional ANOVA expansion of a multivariate mapping plays a fundamental role in statistics.,31311
unique expansion The to assigned the is is distribution once covariates. unique a,The expansion is unique once a unique distribution is assigned to the covariates.,31312
variance mixtures. and under obtain for decomposition dimension results new We distribution,We obtain new results for variance decomposition and dimension distribution under mixtures.,31313
the Implications global computer analysis of experiments are discussed. for also sensitivity,Implications for the global sensitivity analysis of computer experiments are also discussed.,31314
Supervised in used widely vehicle. learning training autonomous is driving,Supervised learning is widely used in training autonomous driving vehicle.,31315
"labeled amount it is large supervised trained However, data. of with","However, it is trained with large amount of supervised labeled data.",31316
"with virtual relatively Nevertheless, performance easier. environment is in much training an agent good","Nevertheless, training an agent with good performance in virtual environment is relatively much easier.",31317
"simulator. in racing car agent a trained The TORCS, is","The agent is trained in TORCS, a car racing simulator.",31318
"situation, a requests an where from start decision support. decision individual We","We start from a decision situation, where an individual requests decision support.",31319
explain deliberated be empirically. validated how judgment of We can models,We explain how models of deliberated judgment can be validated empirically.,31320
aiding the framework for proposed of our decision then We significance practice. explore,We then explore the significance of our proposed framework for decision aiding practice.,31321
of task. in aerial location broken challenging insulators a images The is,The location of broken insulators in aerial images is a challenging task.,31322
"glass on proposes the focusing paper, solution. This deep self-blast insulator, learning a","This paper, focusing on the self-blast glass insulator, proposes a deep learning solution.",31323
with a results. modality Comparison single promising shows approach,Comparison with a single modality approach shows promising results.,31324
task computer segmentation vision. image is Pixel-wise in demanding,Pixel-wise image segmentation is demanding task in computer vision.,31325
"construct of Within (REX). randomized we class algorithm simple, a exchange methods, this","Within this class of methods, we construct a simple, randomized exchange algorithm (REX).",31326
"the converges proposed D-optimality, For prove that algorithm we optimum. the to","For D-optimality, we prove that the proposed algorithm converges to the optimum.",31327
A-optimal and These REX to for I-optimal enable one computing use designs. formulas,These formulas enable one to use REX for computing A-optimal and I-optimal designs.,31328
Temporal applied recurrent is then learning the to temporal-variant features. decoupled,Temporal recurrent learning is then applied to the decoupled temporal-variant features.,31329
computer due Scale-space in theoretical been representation popular foundation. to vision community its has,Scale-space representation has been popular in computer vision community due to its theoretical foundation.,31330
"simple geometrically this conceptually interpretable objective a In and propose paper, function, we i.e.","In this paper, we propose a conceptually simple and geometrically interpretable objective function, i.e.",31331
"verification. for margin deep additive Softmax face (AM-Softmax),","additive margin Softmax (AM-Softmax), for deep face verification.",31332
emphasize also the feature in normalization importance and discuss paper. of the We,We also emphasize and discuss the importance of feature normalization in the paper.,31333
SAIC and collected dataset. on newly Experiments a public are the Udacity conducted dataset,Experiments are conducted on the public Udacity dataset and a newly collected SAIC dataset.,31334
angles predicts speed the proposed Results model show values and that accurately. steering,Results show that the proposed model predicts steering angles and speed values accurately.,31335
evolved dispersal selection. and of control natural of Migration has animals under,Migration and dispersal of animals has evolved under control of natural selection.,31336
"increase In the strategy a simple Darwinian fitness. should dispersal formalisation, of","In a simple formalisation, the strategy of dispersal should increase Darwinian fitness.",31337
of on introduce diffusion We kinesis new dependent coefficient with models purposeful fitness.,We introduce new models of purposeful kinesis with diffusion coefficient dependent on fitness.,31338
"Darwinian evaluation used, the coefficient. of reproduction The fitness local and instant is","The local and instant evaluation of Darwinian fitness is used, the reproduction coefficient.",31339
series are a experiments. of The numerical explored models properties of by,The properties of models are explored by a series of numerical experiments.,31340
positive that kinesis proven steady of cannot states. modify It is stability homogeneous,It is proven that kinesis cannot modify stability of positive homogeneous steady states.,31341
(e.g. representations information) high-level and low-level the framework features Our (pixel-wise exploits jointly,Our framework jointly exploits the low-level features (pixel-wise information) and high-level representations (e.g.,31342
"like brain tumors, structure gray brain matter, etc.)","brain tumors, brain structure like gray matter, etc.)",31343
which brain are important complexity for challenging cross resolving in between modalities the structures.,between cross modalities which are important for resolving the challenging complexity in brain structures.,31344
proposed five MRI datasets. our framework Experiments the on state-of-the-art brain that demonstrate advances,Experiments demonstrate that our proposed framework advances the state-of-the-art on five brain MRI datasets.,31345
"registration, brain segmentation, cross-modality, . MRI Keywords: image-to-image,","Keywords: image-to-image, cross-modality, registration, segmentation, brain MRI .",31346
"low-precision beamforming with phase they can control. RF with Further, be implemented","Further, they can be implemented with RF beamforming with low-precision phase control.",31347
we two this make work In contributions.,In this work we make two contributions.,31348
"cause variational performance First, this a of identify over-pruning. gap, we","First, we identify a cause of this performance gap, variational over-pruning.",31349
"explanation we for phenomenon. a Second, theoretically this introduce grounded","Second, we introduce a theoretically grounded explanation for this phenomenon.",31350
design used for priors Gaussian are in aerospace process performing commonly Bayesian optimization.,Gaussian process priors are commonly used in aerospace design for performing Bayesian optimization.,31351
"and aerospace we context Student's-T optimization. of discuss Here, the their in advantages describe processes,","Here, we describe Student's-T processes, and discuss their advantages in the context of aerospace optimization.",31352
for shown various very neural (DNNs) restoration image (IR) networks results Deep have tasks. promising,Deep neural networks (DNNs) have shown very promising results for various image restoration (IR) tasks.,31353
"challenging achieving for further network design architectures major of However, remains improvements. the a","However, the design of network architectures remains a major challenging for achieving further improvements.",31354
"denoisers end-to-end Through BP optimized. the be modules both jointly and training, the can","Through end-to-end training, both the denoisers and the BP modules can be jointly optimized.",31355
"hampered techniques deep property ""$n<<p$"" outcome for has disease application This of learning classification.","This ""$n<<p$"" property has hampered application of deep learning techniques for disease outcome classification.",31356
The is to layers sparse method able prevent between overfitting. achieve to network connection,The method is able to achieve sparse connection between network layers to prevent overfitting.,31357
"image a ignored issue is However, quality in pattern largely recognition traditional literature.","However, image quality is a largely ignored issue in traditional pattern recognition literature.",31358
nets nets with has Incorporating generation adopted been in widely image tasks. encoding-decoding adversarial,Incorporating encoding-decoding nets with adversarial nets has been widely adopted in image generation tasks.,31359
"the robustness, generality and proposed Experimental method. results demonstrate of effectiveness, the","Experimental results demonstrate the effectiveness, robustness, and generality of the proposed method.",31360
to is train models. training learned data normally deeply Sufficient required,Sufficient training data normally is required to train deeply learned models.,31361
"have data the not usually annotation does labels. However, generated","However, the generated data usually does not have annotation labels.",31362
weaknesses Our focuses architectural approaches. literature-like on understanding and strengths study of,Our study focuses on understanding architectural strengths and weaknesses of literature-like approaches.,31363
is and images. task localizing forgery Image the of detection in parts tampered forged detecting,Image forgery detection is the task of detecting and localizing forged parts in tampered images.,31364
to SCNN chroma designed information and the utilize is of saturation.,SCNN is designed to utilize the information of chroma and saturation.,31365
over we in models the particle observe both improvement significant standard a filter. For performance,For both models we observe a significant improvement in performance over the standard particle filter.,31366
by which activity neuronal behavior The robust question ubiquitous yields neuroscience. in is means a,The means by which neuronal activity yields robust behavior is a ubiquitous question in neuroscience.,31367
and respiratory features acoustic reproduces The between of activity observed song. model neuronal and correlations,The model reproduces observed correlations between neuronal and respiratory activity and acoustic features of song.,31368
"framework an execution of the the time components is Additionally, of analysis for presented.","Additionally, an analysis of execution time for the components of the framework is presented.",31369
Modeling an ill-posed role in essential plays regularity statistical problems. image processing,Modeling statistical regularity plays an essential role in ill-posed image processing problems.,31370
very embedded for on running technologies systems. vision attractive are practical Computer applications,Computer vision technologies are very attractive for practical applications running on embedded systems.,31371
"between transition, KCFs a filter. we potential drifts To individual minimize any particle developed","To minimize any potential drifts between individual KCFs transition, we developed a particle filter.",31372
with probability low outage Numerical that achieve schemes the reduced show results backhaul load. proposed,Numerical results show that the proposed schemes achieve low outage probability with reduced backhaul load.,31373
"Vehicles (UAVs), surveillance Aerial Aerial i.e. from Unmanned","Aerial surveillance from Unmanned Aerial Vehicles (UAVs), i.e.",31374
"well for police moving interest as monitoring. growing disaster cameras, area with of is as","with moving cameras, is of growing interest for police as well as disaster area monitoring.",31375
the images resolutions are more ground detailed camera For increasing. steadily,For more detailed ground images the camera resolutions are steadily increasing.,31376
"video data the too. significantly, of amount transmit increasing to is Simultaneously","Simultaneously the amount of video data to transmit is increasing significantly, too.",31377
bit We distribution explain present the the to detailed of analysis gains. a within frames,We present a detailed analysis of bit distribution within the frames to explain the gains.,31378
"incorporating did ImageNet not An showed earlier Hybrid-CNN, help much. that attempt,","An earlier attempt, Hybrid-CNN, showed that incorporating ImageNet did not help much.",31379
"to extractor scale feature Thus, each adapting particular the (i.e.","Thus, adapting the feature extractor to each particular scale (i.e.",31380
"preferred. these linear applications, is more a model assumption In piecewise","In these applications, a piecewise linear model assumption is more preferred.",31381
This the greatly can cost. computational reduce,This can greatly reduce the computational cost.,31382
with various piecewise constant model apply the We classical method assumption. to our filters,We apply our method to various classical filters with the piecewise constant model assumption.,31383
proposed method. show the the Experimental of effectiveness several results of applications,Experimental results of several applications show the effectiveness of the proposed method.,31384
"capturing limited due sensor data However, resolution light poses to resolution spatio-angular field inherently trade-off.","However, due to limited sensor resolution capturing light field data inherently poses spatio-angular resolution trade-off.",31385
"tackle solution this trade-off. a resolution to the paper, propose learning In deep we based","In this paper, we propose a deep learning based solution to tackle the resolution trade-off.",31386
"field resolution light sensor coded we from single a reconstruct full image. Specifically,","Specifically, we reconstruct full sensor resolution light field from a single coded image.",31387
We for stages. these neural networks three propose,We propose three neural networks for these stages.,31388
demonstrate image. of the Our coded recovery from parallax better results,Our results demonstrate better recovery of parallax from the coded image.,31389
"results based dictionary better than quatitatively. learning approaches Also, both and get qualitatively we","Also, we get better results than dictionary learning based approaches both qualitatively and quatitatively.",31390
"we intractable, the problem inner formulation convex a approximation. propose original computationally Since is","Since the original problem formulation is computationally intractable, we propose a convex inner approximation.",31391
"very these directly from to the However, features data. methods few learn have attempted","However, very few methods have attempted to learn these features directly from the data.",31392
labelling a video role in pixel plays surveillance. (FG) vital Foreground,Foreground (FG) pixel labelling plays a vital role in video surveillance.,31393
"then, specific exemplary desired is, database FG It with trained predict to segmentations objects.","It is, then, trained with database specific exemplary segmentations to predict desired FG objects.",31394
channel of map rain network. dark extract image feature a We in as,We extract dark channel of rain image as a feature map in network.,31395
the All of by are back-propagation. optimized parameters,All of the parameters are optimized by back-propagation.,31396
"At receiver, estimators the are of two corresponding kinds presented.","At the corresponding receiver, two kinds of estimators are presented.",31397
"Interferometry e.g. or Persistent pixel-cluster-based Scatterer (PSI), methods,","Persistent Scatterer Interferometry (PSI), or pixel-cluster-based methods, e.g.",31398
of of InSAR inherent exploitation the object This properties an level. on enables phase stacks,This enables the exploitation of inherent properties of InSAR phase stacks on an object level.,31399
"thirty a ten by PSI, of settings. factor typical in to","PSI, by a factor of ten to thirty in typical settings.",31400
"is effective with against proposed The unmodeled phases. particularly as pixels such outliers, method","The proposed method is particularly effective against outliers, such as pixels with unmodeled phases.",31401
method stacks. the demonstrated performance of using TerraSAR-X high-resolution is The image proposed promising,The promising performance of the proposed method is demonstrated using high-resolution TerraSAR-X image stacks.,31402
results systems. in relevance and physical biological have Our,Our results have relevance in biological and physical systems.,31403
for connectomics. a model is investigated We navigation parsimonious routing whether,We investigated whether navigation is a parsimonious routing model for connectomics.,31404
"video research ways. we two in saliency this work, to contribute In","In this work, we contribute to video saliency research in two ways.",31405
is multipreference respect semantics. sound We closure provide multipreference mechanism the which with to a,We provide a multipreference closure mechanism which is sound with respect to the multipreference semantics.,31406
"a much Despite from general VQA progress, is far recent problem. solved","Despite much recent progress, general VQA is far from a solved problem.",31407
Staff Recognition. Optical crucial step in pre-processing a Music is removal line,Staff line removal is a crucial pre-processing step in Optical Music Recognition.,31408
at The Generator images output. to staff-less produce intends the,The Generator intends to produce staff-less images at the output.,31409
high textures Adversarial generated in quality to images. more loss retrieve helps,Adversarial loss helps to retrieve more high quality textures in generated images.,31410
Our other achieves approaches. method in comparison to superior performance conventional,Our method achieves superior performance in comparison to other conventional approaches.,31411
fill We transform spatial gap. propose numerical to this to our (DSNT) differentiable,We propose our differentiable spatial to numerical transform (DSNT) to fill this gap.,31412
"The parameters, exhibits differentiable, spatial trainable adds DSNT generalization. fully is no layer and good","The DSNT layer adds no trainable parameters, is fully differentiable, and exhibits good spatial generalization.",31413
emotion research has in a decade. past Automatic the recognition become topic trending,Automatic emotion recognition has become a trending research topic in the past decade.,31414
research a the survey present boost comprehensive new to hoping We field. in,We present a new comprehensive survey hoping to boost research in the field.,31415
define complete We a emotional body gesture recognition. automatic for then framework,We then define a complete framework for automatic emotional body gesture recognition.,31416
"community three research for its dataset use. representations) release to this the Lastly, (and we","Lastly, we release this dataset (and its three representations) for the research community to use.",31417
theoretical literature. We interesting highlight empirical an between gap and,We highlight an interesting gap between empirical and theoretical literature.,31418
the motivating Several examples methods. application illustrate the of,Several motivating examples illustrate the application of the methods.,31419
"CT. tapered method propose novel Second, we evalua-tion eval incorporates that into a","Second, we propose a novel evalua-tion method that incorporates tapered eval into CT.",31420
"adding version improved was by then most n-tuple The trained features, above additional importantly features.","The above trained version was then improved by adding additional features, most importantly n-tuple features.",31421
"learning Second-order pooling, visual proven deep for based a.k.a.~bilinear pooling, has effective recognition.","Second-order pooling, a.k.a.~bilinear pooling, has proven effective for deep learning based visual recognition.",31422
in many widely vehicles aerial is in detection The images of applications. applied,The detection of vehicles in aerial images is widely applied in many applications.,31423
"paper, network focal convolutional propose neural we loss framework this double In (DFL-CNN). novel a","In this paper, we propose a novel double focal loss convolutional neural network framework (DFL-CNN).",31424
show outperforms the our DFL-CNN results experimental detection. that on The vehicle baselines,The experimental results show that our DFL-CNN outperforms the baselines on vehicle detection.,31425
learning This these presents speed and to algorithm paper CNNs. up simplify a,This paper presents a learning algorithm to simplify and speed up these CNNs.,31426
model an Our of encoder-decoder consists employs which LSTM to framework sequence module. proposed sequence,Our proposed framework employs sequence to sequence model which consists of an encoder-decoder LSTM module.,31427
"characters Toolkit Tamil, LIPI are used and of for Devanagari our dataset experiments. Telugu","Tamil, Telugu and Devanagari characters of LIPI Toolkit dataset are used for our experiments.",31428
method the approaches. performance has compared conventional to other superior achieved Our proposed,Our proposed method has achieved superior performance compared to the other conventional approaches.,31429
dataset data of multisensory new analysis. opens This opportunities,This dataset opens new opportunities of multisensory data analysis.,31430
"same neighboring agent at node the no empty, is stays site. If each","If no neighboring node is empty, each agent stays at the same site.",31431
world's the epidemic. in largest is Tuberculosis(TB) India TB,Tuberculosis(TB) in India is the world's largest TB epidemic.,31432
India the makes worse. the with in combined resistant problem of emergence drug This bacteria,This combined with the emergence of drug resistant bacteria in India makes the problem worse.,31433
of dynamically EdgeConv layer network. each computed on acts the in graphs,EdgeConv acts on graphs dynamically computed in each layer of the network.,31434
existing be is can into architectures. plugged and differentiable It,It is differentiable and can be plugged into existing architectures.,31435
power allocations derived for expressions per subcarrier. and Closed-form are bit,Closed-form expressions are derived for bit and power allocations per subcarrier.,31436
"modeling. UAV challenges in exist channel several However,","However, several challenges exist in UAV channel modeling.",31437
"not airframe yet been has shadowing rotary small for size investigated Also, UAVs.","Also, airframe shadowing has not yet been investigated for small size rotary UAVs.",31438
"imaging light-field into paradigm the Herein time, laryngoscopy. first for the of we introduce","Herein for the first time, we introduce the paradigm of light-field imaging into laryngoscopy.",31439
the derived subcarrier. are Closed-form per optimal distributions expressions power and bit for,Closed-form expressions are derived for the optimal bit and power distributions per subcarrier.,31440
performance simulations. of the is algorithm proposed extensive through investigated The,The performance of the proposed algorithm is investigated through extensive simulations.,31441
action recognition. of convey significant information Dynamics for skeletons body human human,Dynamics of human body skeletons convey significant information for human action recognition.,31442
formulation greater but generalization power expressive stronger not only also leads to capability. This,This formulation not only leads to greater expressive power but also stronger generalization capability.,31443
"substantial NTU-RGBD, large two achieves over it Kinetics and datasets, On mainstream improvements methods.","On two large datasets, Kinetics and NTU-RGBD, it achieves substantial improvements over mainstream methods.",31444
"this model. a In employing an propose image based energy paper, inpainting method structured we","In this paper, we propose a structured image inpainting method employing an energy based model.",31445
erative Gen- propose an recognition using Networks. Adversarial We framework action,We propose an action recognition framework using Gen- erative Adversarial Networks.,31446
"contribution, analyses of computational, communication the are and careful presented. memory, As part","As part of the contribution, careful computational, memory, and communication analyses are presented.",31447
the multivariate error suitable in then DEM. A gridded is method estimating developed for,A suitable multivariate method is then developed for estimating the error in gridded DEM.,31448
another font useful Conversion applications. in life real font of is one very to,Conversion of one font to another font is very useful in real life applications.,31449
network images. end-to-end word trained the is complete for The,The network is trained end-to-end for the complete word images.,31450
"eliminates segmentations. character necessary steps, Thus the it like pre-processing","Thus it eliminates the necessary pre-processing steps, like character segmentations.",31451
learn model our to that conditional helps We mapping to setting one-to-many extend function.,We extend our model to conditional setting that helps to learn one-to-many mapping function.,31452
earlier on are square to translation image related works performed Most the of images.,Most of the earlier works related to image translation are performed on square images.,31453
Our first is architecture images widths. can of proposed handle the work varying which,Our proposed architecture is the first work which can handle images of varying widths.,31454
Word images on present. width the of characters generally depending have varying number,Word images generally have varying width depending on the number of characters present.,31455
"font synthetically model a generated we our on dataset. Hence, test","Hence, we test our model on a synthetically generated font dataset.",31456
the image some compare our translation. We method state-of-the-art of for with methods,We compare our method with some of the state-of-the-art methods for image translation.,31457
artifacts. number remove a proposed Recently of CNN-based image compression were techniques to,Recently a number of CNN-based techniques were proposed to remove image compression artifacts.,31458
by other details or image subtle destroyed restore are missed that algorithms. It can otherwise,It can restore subtle image details that are otherwise destroyed or missed by other algorithms.,31459
"created ear have database infrared we wave based Thus, band. long on","Thus, we have created ear database based on long wave infrared band.",31460
have to distinctive the reveal different proposed We use features. algorithms to,We have proposed to use different algorithms to reveal the distinctive features.,31461
"the of have using reduced subspace Then, dimensions number we methods.","Then, we have reduced the number of dimensions using subspace methods.",31462
"data classifier accordance methods. in the with the of is reduced dimension Finally,","Finally, the dimension of data is reduced in accordance with the classifier methods.",31463
the results fusion was successful. technique have The that showed,The results have showed that the fusion technique was successful.,31464
"for identification, This example important for criminal investigation evaluation forensics, especially etc. corpse is","This evaluation is important for especially forensics, for example corpse identification, criminal investigation etc.",31465
Network. memory four and based on by proposed were Most external components Memory approaches,Most approaches were based on external memory and four components proposed by Memory Network.,31466
accuracy. is predictive This with to high trees function crucial obtain,This function is crucial to obtain trees with high predictive accuracy.,31467
splits of compares bias a Some suffer it different when can from arities. functions gain,Some gain functions can suffer from a bias when it compares splits of different arities.,31468
neurons. is paper membrane This to the equation potential an introduction for,This paper is an introduction to the membrane potential equation for neurons.,31469
"well are Its sample applications. as described, properties as","Its properties are described, as well as sample applications.",31470
excitable neurons. from model of a stochastic Simulated network discrete-time connected fully derived time-series were,Simulated time-series were derived from a network model of fully connected discrete-time stochastic excitable neurons.,31471
and neuronal multifractal Monofractal to applied time-series. analyses gain were,Monofractal and multifractal analyses were applied to neuronal gain time-series.,31472
"indices ratio. additional fractal including of average scaling criticality, closely to Peak corresponded branching","Peak fractal scaling corresponded closely to additional indices of criticality, including average branching ratio.",31473
"Networks scaling relationships. structure, exhibited multifractal multiple or","Networks exhibited multifractal structure, or multiple scaling relationships.",31474
between localizing objects their the There compromise estimating viewpoints. and exists a,There exists a compromise between localizing the objects and estimating their viewpoints.,31475
solving simultaneously. models deep both problems This is an using exploration work learning of for,This work is an exploration of using deep learning models for solving both problems simultaneously.,31476
achieve state-of-the-art in the performance both proposed datasets. the we With models,With the proposed models we achieve the state-of-the-art performance in both datasets.,31477
tumor the The motion of mass-effect from expansion growing the models mass. network,The expansion network models the mass-effect from the growing motion of tumor mass.,31478
more. technology becomes crucial biometrics development of The,The development of technology biometrics becomes crucial more.,31479
and is with speech recognition system images a needed. ear the combine Therefore of robust,Therefore a robust recognition system with the combine of speech and ear images is needed.,31480
algorithms are recognition used. and MFCC DTW For speech,For speech recognition MFCC and DTW algorithms are used.,31481
"identification. and ear algorithms Gabor used filter are DCVA Also, for","Also, Gabor filter and DCVA algorithms are used for ear identification.",31482
level rate is increased by matching score fusion. success making Multi-biometrics,Multi-biometrics success rate is increased by making matching score level fusion.,31483
ear and think be can that We speech complementary.,We think that speech and ear can be complementary.,31484
"human result multi-biometrics and identifications. is it based Therefore, that speech effective for ear is","Therefore, it is result that multi-biometrics based speech and ear is effective for human identifications.",31485
segmentation a Image in analysis. medical problem fundamental image is,Image segmentation is a fundamental problem in medical image analysis.,31486
"sometimes unavailable. big medical on expensive annotations or datasets are expert However, tedious,","However, expert annotations on big medical datasets are tedious, expensive or sometimes unavailable.",31487
in problem segmentation Instance vision. significance is computer a of,Instance segmentation is a problem of significance in computer vision.,31488
"data preparing this is costly. time-consuming annotated for and task However, extremely","However, preparing annotated data for this task is extremely time-consuming and costly.",31489
"can pipeline scanned, well-suited i.e. that scenes to be Our is","Our pipeline is well-suited to scenes that can be scanned, i.e.",31490
code to is generalizing multi-class the common method classification of (ECC). binary A error correcting,A common method of generalizing binary to multi-class classification is the error correcting code (ECC).,31491
"for of in a instance optimized number ECCs be them making ways, orthogonal. by may","ECCs may be optimized in a number of ways, for instance by making them orthogonal.",31492
Orthogonal than ECCs predicted ECCs by random accurate recent more literature. always are as,Orthogonal ECCs are always more accurate than random ECCs as predicted by recent literature.,31493
is in worth increase the speed depend on application. will penalty the accuracy the Whether,Whether the speed increase is worth the penalty in accuracy will depend on the application.,31494
are the libraries. using The GMM of parameters estimated distributions spectral,The parameters of the GMM distributions are estimated using spectral libraries.,31495
parameters. are estimated on Abundances distribution based the,Abundances are estimated based on the distribution parameters.,31496
"Meanwhile, the teacher's to a the robot children class. co-learn with as in acts assistant","Meanwhile, the robot acts as a teacher's assistant to co-learn with children in the class.",31497
for helpful Experimental that with children. the results robot learning is show and disadvantaged below-basic,Experimental results show that learning with the robot is helpful for disadvantaged and below-basic children.,31498
"data may some However, the appropriate sets. assumptions errors distributional on not be for these","However, these distributional assumptions on the errors may not be appropriate for some data sets.",31499
"methods analysis. regression carry In be to may nonparametric on these case, the considered","In these case, nonparametric methods may be considered to carry on the regression analysis.",31500
Empirical one is (EL) nonparametric of these methods. method likelihood,Empirical likelihood (EL) method is one of these nonparametric methods.,31501
"consider with paper, method constraints. this the EL In robustifyed we","In this paper, we consider the EL method with robustifyed constraints.",31502
a new clustering stationary unsupervised stochastic processes. introduce problem: ergodic wide-sense We learning,We introduce a new unsupervised learning problem: clustering wide-sense stationary ergodic stochastic processes.,31503
as data real-world are synthetic examples Clustering and of applications. data provided,Clustering synthetic data and real-world data are provided as examples of applications.,31504
"for marked NPOLE-MMHP, an of Furthermore, algorithm, multivariate Hawkes design processes. we estimation","Furthermore, we design an algorithm, NPOLE-MMHP, for estimation of multivariate marked Hawkes processes.",31505
important in long East of an Handwriting has Asia. Chinese skill been,Handwriting of Chinese has long been an important skill in East Asia.,31506
generate Chinese further characters. We handwritten to propose CycleGAN DenseNet,We further propose DenseNet CycleGAN to generate Chinese handwritten characters.,31507
interpretation auscultation sounds Cardiac stethoscope. in involves abnormalities expert of heart using,Cardiac auscultation involves expert interpretation of abnormalities in heart sounds using stethoscope.,31508
"auscultation In solution. propose this Recurrent based automated cardiac Neural a work, we Networks (RNNs)","In this work, we propose a Recurrent Neural Networks (RNNs) based automated cardiac auscultation solution.",31509
"the ConvNet pre-trained on is a natural First, dataset. image","First, the ConvNet is pre-trained on a natural image dataset.",31510
and smear LBC datasets. method on The Pap proposed both is evaluated,The proposed method is evaluated on both Pap smear and LBC datasets.,31511
(H&E LBC) are manual on superior also HEMLBC performances stained the achieved Similar dataset.,Similar superior performances are also achieved on the HEMLBC (H&E stained manual LBC) dataset.,31512
Our for cervical method screening. primary automation-assisted of the is in development reading promising systems,Our method is promising for the development of automation-assisted reading systems in primary cervical screening.,31513
"the Thus, be this employed can in loss representation. cosine non-categorical","Thus, the cosine loss can be employed in this non-categorical representation.",31514
results improves point-wise integration localizations coarse minimal The costs. further at annotation of the,The integration of coarse point-wise localizations further improves the results at minimal annotation costs.,31515
"systems These or metabolic cascades processes, signaling represent complex-breakage biomolecular networks. example oligomerization reactions, for","These systems represent for example biomolecular oligomerization processes, complex-breakage reactions, signaling cascades or metabolic networks.",31516
"(DPN) In this adapt paper, problem. network pyramid we a this deep tackle to","In this paper, we adapt a deep pyramid network (DPN) to tackle this problem.",31517
eyes. snowy human mountains) distinguish are difficult to that perceptually by,snowy mountains) that are perceptually difficult to distinguish by human eyes.,31518
of large for tractability deconvolution the of images. results the Simulation our model confirm,Simulation results confirm the tractability of our model for the deconvolution of large images.,31519
"propose instead. we this to study end, To distraction pedestrian","To this end, we propose to study pedestrian distraction instead.",31520
related in and Seizure each other are manner. complex to synchronization,Seizure and synchronization are related to each other in complex manner.,31521
implicated consciousness during of seizures. Altered in synchrony has been loss partial,Altered synchrony has been implicated in loss of consciousness during partial seizures.,31522
"studied altered However, been consciousness of has the termination not of following seizures well. mechanism","However, the mechanism of altered consciousness following termination of seizures has not been studied well.",31523
"signals MI to than have Next, more component two principal by we method. extended","Next, we have extended MI to more than two signals by principal component method.",31524
"were alpha, beta The and theta, analyses in delta, out carried gamma bands.","The analyses were carried out in delta, theta, alpha, beta and gamma bands.",31525
"general, in difficulties, high especially leads in Performing computational such dimensions. to still transformations,","Performing such transformations, in general, still leads to computational difficulties, especially in high dimensions.",31526
"imagery. CNN propose In SAR this report, novel for end-to-end from we ATR a","In this report, we propose a novel CNN for end-to-end ATR from SAR imagery.",31527
named CNN performs stages network SAR of The end-to-end. ATR three support (VersNet) verification all,The CNN named verification support network (VersNet) performs all three stages of SAR ATR end-to-end.,31528
"areas, in preserve residual the most details To the learns of age-attribute-irrelevant generator face. our","To preserve most of the details in age-attribute-irrelevant areas, our generator learns the residual face.",31529
preserve pixel the detailed A adopted also to information face. input of is facial loss,A pixel loss is also adopted to preserve detailed facial information of the input face.,31530
by method the with state-of-the-art. results proposed the show performance Experimental appealing of comparing,Experimental results show appealing performance of the proposed method by comparing with the state-of-the-art.,31531
image of due phenomenon. haze quality common of Degradation is very a the to presence,Degradation of image quality due to the presence of haze is a very common phenomenon.,31532
"problem methods distortion have gloomy environment. these of illumination) (poor the However, color in","However, these methods have the problem of color distortion in gloomy (poor illumination) environment.",31533
"output all constantcomposition With composition. matching sequences have conventional distribution (CCDM), identical","With conventional constantcomposition distribution matching (CCDM), all output sequences have identical composition.",31534
is treatment reliable Accurate and essential to growth segmentation tumor analysis tumor selection. and,Accurate and reliable tumor segmentation is essential to tumor growth analysis and treatment selection.,31535
operations. morphological UNet The (GMM) segmentation Gaussian and refined by is Mixture Model then,The UNet segmentation is then refined by Gaussian Mixture Model (GMM) and morphological operations.,31536
maps. is the UNet each node determined the The graph of by probability cost for,The cost for each node of the graph is determined by the UNet probability maps.,31537
descriptors dimensional processing in chosen image nature. high generally are involved in manually and Feature,Feature descriptors involved in image processing are generally manually chosen and high dimensional in nature.,31538
"Consortium prohibiting become probably However, to not is WWW industrial a SWRL go-ing acceptance. standard,","However, SWRL is probably not go-ing to become a WWW Consortium standard, prohibiting industrial acceptance.",31539
determination sensitivity analysis. preliminary for provides probabilities also three of This approaches in work,This preliminary work also provides three approaches of determination in probabilities for sensitivity analysis.,31540
systems. evaluated IEEE test are risk The using proposed indices,The proposed risk indices are evaluated using IEEE test systems.,31541
"as in the one environment. nanoparticles Microorganisms, targets such bacteria, of the of are first","Microorganisms, such as bacteria, are one of the first targets of nanoparticles in the environment.",31542
modulation the key the process of choice best is One this of aspect format.,One key aspect of this process is the choice of the best modulation format.,31543
deploy from soccer this videos. calibration method We new annotate to data pan-tilt-zoom,We deploy this new calibration method to annotate pan-tilt-zoom data from soccer videos.,31544
are as data collected new The used images. references for,The collected data are used as references for new images.,31545
data our soccer two system and We datasets. synthetic on real demonstrate,We demonstrate our system on synthetic data and two real soccer datasets.,31546
the approach state-of-the-art superior performance over achieves Our two-point method.,Our two-point approach achieves superior performance over the state-of-the-art method.,31547
how systems Control goals. user-defined to certain the reach state theorists consider influence of to,Control theorists consider how to influence the state of systems to reach certain user-defined goals.,31548
"of come with conceptual However, confusion. opportunities risk the these also","However, these opportunities also come with the risk of conceptual confusion.",31549
"insights. novel to suggest I distinct inquiry produce of field continues and that Here, each","Here, I suggest that each field of inquiry continues to produce novel and distinct insights.",31550
learning challenging the problem unsupervised from we this of work videos. In address,In this work we address the challenging problem of unsupervised learning from videos.,31551
"tasks. the on validate various where extensive experimental An presented, proposed models is we evaluation","An extensive experimental evaluation is presented, where we validate the proposed models on various tasks.",31552
follow. rigid These not to do calculation procedure provide a,These do not provide a rigid calculation procedure to follow.,31553
of target. This a investigates update estimating paper the agent's variance temporal-difference learning,This paper investigates estimating the variance of a temporal-difference learning agent's update target.,31554
These estimated of Monte using values straightforwardly be methods. can returns batches averaging Carlo by,These values can be straightforwardly estimated by averaging batches of returns using Monte Carlo methods.,31555
higher-order could moments to estimate Naturally attempt the {\lambda}-return. of one,Naturally one could attempt to estimate higher-order moments of the {\lambda}-return.,31556
is about estimating the the This {\lambda}-return. of paper variance,This paper is about estimating the variance of the {\lambda}-return.,31557
"planning models Thus, the directly process. planning stochastic to into incorporate attempts probabilistic","Thus, probabilistic planning attempts to incorporate stochastic models directly into the planning process.",31558
compared obtain better models. to state-of-the-art We accuracy prediction the,We obtain better prediction accuracy compared to the state-of-the-art models.,31559
shown have good tracking deep various learning challenges. based on trackers Recent performance,Recent deep learning based trackers have shown good performance on various tracking challenges.,31560
"not techniques performance Moreover, been evaluated of noisy in tracking has environments.","Moreover, performance of tracking techniques has not been evaluated in noisy environments.",31561
structure of the (approximately) study subspaces. data from generated independent underlying of We a union,We study the underlying structure of data (approximately) generated from a union of independent subspaces.,31562
"an to propose this facial we behaviors generate In agent. method for a paper,","In this paper, we propose a method to generate facial behaviors for an agent.",31563
terms exact of the confluent test obtain distributions We functions. statistic of in hypergeometric,We obtain exact distributions of the test statistic in terms of confluent hypergeometric functions.,31564
"of accuracy. classification Consequently, the runtime without accelerated tasks the degrading the classification our method","Consequently, our method accelerated the runtime of the classification tasks without degrading the classification accuracy.",31565
Receiver for used evaluating diagnostic widely systems. analysis characteristic operating is (ROC),Receiver operating characteristic (ROC) analysis is widely used for evaluating diagnostic systems.,31566
shown cross-validation this leave-pair-out been to bias. correct The has (LPO),The leave-pair-out (LPO) cross-validation has been shown to correct this bias.,31567
"(TLPO) new this propose a leave-pair-out In study, method tournament called cross-validation. we","In this study, we propose a new method called tournament leave-pair-out (TLPO) cross-validation.",31568
"work, attempt systematically In to methods for characterize estimation. density we this","In this work, we attempt to systematically characterize methods for density estimation.",31569
non-Markovian proposed to For model we example transformations based RNN dependencies.,For example we proposed RNN based transformations to model non-Markovian dependencies.,31570
detection and illustrate We models of image in use the modeling. outlier our,We illustrate the use of our models in outlier detection and image modeling.,31571
of for framework a novel introduce distributions. family data learning driven a we Finally,Finally we introduce a novel data driven framework for learning a family of distributions.,31572
"we standard domains. the extend planning of to In such work, generalized formulation this","In this work, we extend the standard formulation of generalized planning to such domains.",31573
value policies discussed. representations also learning generalized is of for general The,The value of generalized representations for learning general policies is also discussed.,31574
in yet dynamics Low-dimensional rich brain. the often emerge,Low-dimensional yet rich dynamics often emerge in the brain.,31575
"sleep, epilepsy, and dynamics oscillations voluntary Examples include movement. chaotic during and","Examples include oscillations and chaotic dynamics during sleep, epilepsy, and voluntary movement.",31576
"dynamics of low mechanism emergence general remains a However, dimensional the elusive. for","However, a general mechanism for the emergence of low dimensional dynamics remains elusive.",31577
"chaos, dynamics include mixed-mode mixed-mode oscillations, and The chaotic synchronization.","The dynamics include mixed-mode oscillations, mixed-mode chaos, and chaotic synchronization.",31578
"the cases, of approximated. point analytically the homeostatic or determined all In stability set is","In all cases, the stability of the homeostatic set point is analytically determined or approximated.",31579
intelligence is general The be inevitable. development to many artificial by considered of,The development of artificial general intelligence is considered by many to be inevitable.,31580
is becoming aware not so after such What certain. intelligence does,What such intelligence does after becoming aware is not so certain.,31581
"intelligence meaningful is of a timely and topic. containment research artificial general Thus,","Thus, containment of artificial general intelligence is a timely and meaningful research topic.",31582
"diagrams Further, ontology constructed to we demonstrate relationships. intended","Further, we constructed ontology diagrams to demonstrate intended relationships.",31583
network dream of transformer spatial makes true. Emergence come,Emergence of spatial transformer network makes dream come true.,31584
yet and is Fine-grained recognition action task. segmentation an challenging important,Fine-grained action segmentation and recognition is an important yet challenging task.,31585
Our show better results than methods. method par proposed on performs state-of-the-art or that the,Our results show that the proposed method performs on par or better than state-of-the-art methods.,31586
"paper, obtaining this classification. we In high linking quality for on results object better focus","In this paper, we focus on obtaining high quality object linking results for better classification.",31587
sensors. increasingly equipped Modern are with vehicles complex,Modern vehicles are equipped with increasingly complex sensors.,31588
generate volumes modeling for analysis. These that sensors provide opportunities large of data and,These sensors generate large volumes of data that provide opportunities for modeling and analysis.,31589
"state has i.e. set observations, Each a of","Each state has a set of observations, i.e.",31590
inference for split-merge carried algorithm. using HDP sampling distributions The the is out topic,The inference for the topic distributions is carried out using HDP split-merge sampling algorithm.,31591
a disadvantages the and of thorough technique. proposed advantages the also of We analysis provide,We also provide a thorough analysis of the advantages and disadvantages of the proposed technique.,31592
existing comparison that methods. show state-of-the-art Experimental in with proposed the performance achieves method results,Experimental results show that the proposed method achieves state-of-the-art performance in comparison with existing methods.,31593
training This paper improving achieves both GANs. simultaneously of the by stability aims,This paper achieves both aims simultaneously by improving the stability of training GANs.,31594
sensitivity distortions in the to environments) are in certain database. under-represented,sensitivity to distortions in certain environments) are under-represented in the database.,31595
this the specific In a above study we work example statement. of,In this work we study a specific example of the above statement.,31596
Here problem solution we for stimuli. a using this artificial outline,Here we outline a solution for this problem using artificial stimuli.,31597
methods detection focus text. text quadrangle Traditional mostly on,Traditional text detection methods mostly focus on quadrangle text.,31598
a vision. in is Activity computer problem fundamental detection,Activity detection is a fundamental problem in computer vision.,31599
activities particularly of is temporal challenging. scales different Detecting,Detecting activities of different temporal scales is particularly challenging.,31600
information is activity better for Temporal into contextual classifiers fused recognition.,Temporal contextual information is fused into activity classifiers for better recognition.,31601
"at the be end-to-end. importantly, entire levels model More all trained can","More importantly, the entire model at all levels can be trained end-to-end.",31602
"requires according expert selection/extraction careful to feature knowledge. Conventionally, the SVM-MVPA","Conventionally, the SVM-MVPA requires careful feature selection/extraction according to expert knowledge.",31603
for then We such a using vectors. recognizing actions manipulation SVM employ linear,We then employ a linear SVM for recognizing manipulation actions using such vectors.,31604
the on task considers This X-ray disease images. thorax paper of chest classification,This paper considers the task of thorax disease classification on chest X-ray images.,31605
Existing image the generally as global methods use for input network learning.,Existing methods generally use the global image as input for network learning.,31606
limited strategy a Such aspects. two is in,Such a strategy is limited in two aspects.,31607
irrelevant (excessive) be noisy by the global CNNs image may Training areas. using affected,Training CNNs using global image may be affected by the (excessive) irrelevant noisy areas.,31608
"images. using branch global global Specifically, first we a CNN learn","Specifically, we first learn a global CNN branch using global images.",31609
training region for branch. CNN local a The local used is,The local region is used for training a local CNN branch.,31610
"noise signals. are In to the existing the traditional used filters systems, cancel from received","In existing systems, traditional filters are used to cancel the noise from the received signals.",31611
white noise. nonlinear and random additive noise Gaussian are with Simulations performed,Simulations are performed with additive white Gaussian noise and random nonlinear noise.,31612
an dataset. on is videos system The traffic four of urban validated,The system is validated on four videos of an urban traffic dataset.,31613
of these at This boundary provides areas. integrative the review paper an,This paper provides an integrative review at the boundary of these areas.,31614
"The of featured. and visualization, are topics prediction segmentation, class","The topics of visualization, segmentation, and class prediction are featured.",31615
"for described. data"" and innovations large, complex modern, ""big Recent sets are","Recent innovations for modern, large, and complex ""big data"" sets are described.",31616
trainable An recognition end-to-end multi-language scene (fully is for proposed. differentiable) text and method localization,An end-to-end trainable (fully differentiable) method for multi-language scene text localization and recognition is proposed.,31617
a show multi-language The multi-script obtaining is experiments accurate challenging annotations problem. that,The experiments show that obtaining accurate multi-language multi-script annotations is a challenging problem.,31618
an in important cognitive process sensing Spectrum is radio.,Spectrum sensing is an important process in cognitive radio.,31619
achieve goal. sparse recovery is suitable algorithm this a to of Selection necessary,Selection of a suitable sparse recovery algorithm is necessary to achieve this goal.,31620
of sparse been A proposed. recovery number algorithms have,A number of sparse recovery algorithms have been proposed.,31621
"into and compares This categories, them performances. paper surveys sparse classify recovery the their algorithms,","This paper surveys the sparse recovery algorithms, classify them into categories, and compares their performances.",31622
the recordings. initial the acoustics of was room analysis An on,An initial analysis was on the room acoustics of the recordings.,31623
"parameters interest of pressure spectral and were sound The centroid. autocorrelation level,","The parameters of interest were sound pressure level, autocorrelation and spectral centroid.",31624
"analyze The models perceptual resulting auditory from used were variables, to echolocation i.e. the","The auditory models were used to analyze echolocation resulting from the perceptual variables, i.e.",31625
simulate chosen each models auditory Relevant to were variable.,Relevant auditory models were chosen to simulate each variable.,31626
"non-parametric used for sharpness. distance, A determine thresholds and loudness, was method pitch to","A non-parametric method was used to determine thresholds for distance, loudness, pitch and sharpness.",31627
"the We auditory binaural discuss information, model and approach. also movements","We also discuss binaural information, movements and the auditory model approach.",31628
deep become networks have methods neural trend. a using Colorization recent,Colorization methods using deep neural networks have become a recent trend.,31629
large problems for critical is strategy For update inference the on-line scale performance.,For large scale on-line inference problems the update strategy is critical for performance.,31630
small Sound in vibrations cause waves objects. nearby,Sound waves cause small vibrations in nearby objects.,31631
A the extract video. can techniques exist few that from sound literature in,A few techniques exist in the literature that can extract sound from video.,31632
this image locations. paper at we vibration patterns study In different local,In this paper we study local vibration patterns at different image locations.,31633
differently. that locations show We vibrate in image the different,We show that different locations in the image vibrate differently.,31634
We a local improves quality that and produce state-of-the-art. vibrations carefully sound aggregate,We carefully aggregate local vibrations and produce a sound quality that improves state-of-the-art.,31635
this use to phenomenon estimate direction. sound We,We use this phenomenon to estimate sound direction.,31636
to been ASL have improvement. but provide methods proposed only MRI post-processing moderate Various denoise,Various post-processing methods have been proposed to denoise ASL MRI but only provide moderate improvement.,31637
"abstract might indicate facilitate advances using Recent facts high-level, inputs the reasoning. from extracted that","Recent advances indicate that using high-level, abstract facts extracted from the inputs might facilitate reasoning.",31638
research become image has in processing. segmentation important problem an Superpixel,Superpixel segmentation has become an important research problem in image processing.,31639
correspond those in at The structurally superpixels trees to seeds. ISF spanning rooted,The superpixels in ISF structurally correspond to spanning trees rooted at those seeds.,31640
methods choices We illustrate five present different its ISF to of components.,We present five ISF methods to illustrate different choices of its components.,31641
efficiency. methods from effectiveness compared and the in approaches are state-of-the-art with These,These methods are compared with approaches from the state-of-the-art in effectiveness and efficiency.,31642
"led algorithms. to decade, In the advancement reconstruction image regularization past sparsity-driven of has","In the past decade, sparsity-driven regularization has led to advancement of image reconstruction algorithms.",31643
"on of rely such models (e.g. regularizers sparsity Traditionally, analytical","Traditionally, such regularizers rely on analytical models of sparsity (e.g.",31644
"are around by arguments centered deep recent methods increasingly learning. data-driven more However, inspired","However, more recent methods are increasingly centered around data-driven arguments inspired by deep learning.",31645
"face central The face recognition, and involves discrimination. identification, feature including face of verification task","The central task of face recognition, including face verification and identification, involves face feature discrimination.",31646
"lacks power deep of discrimination. softmax usually the the loss of CNNs However, traditional","However, the traditional softmax loss of deep CNNs usually lacks the power of discrimination.",31647
maximizing and idea: minimizing inter-class same the share improved losses these variance. variance intra-class All,All these improved losses share the same idea: maximizing inter-class variance and minimizing intra-class variance.,31648
CosFace. LMCL as to with trained We our refer model,We refer to our model trained with LMCL as CosFace.,31649
"attention alignment. our layers knowledge approach First, all convolutional transfers through in the","First, our approach transfers knowledge in all the convolutional layers through attention alignment.",31650
"representations, the (FC) fully activations connected previous align high-level Most methods e.g., layers. of","Most previous methods align high-level representations, e.g., activations of the fully connected (FC) layers.",31651
"label we the for Second, the unlabeled posterior estimate distribution of network data target training.","Second, we estimate the posterior label distribution of the unlabeled data for target network training.",31652
principles of those find organizing the the parts. assembly to It remains,It remains to find the principles organizing the assembly of those parts.,31653
in inpainting for image plausible networks potential (CNNs) results. producing have their convolutional exhibited Deep,Deep convolutional networks (CNNs) have exhibited their potential in image inpainting for producing plausible results.,31654
is algorithm An end-to-end Shift-Net. train developed the learning further to,An end-to-end learning algorithm is further developed to train the Shift-Net.,31655
motion one at puppets time. as characters can that This a treats method perform only,This method treats characters as puppets that can perform only one motion at a time.,31656
to time particular of the propose We automatically a strategy the step method. adapt,We propose a particular strategy to automatically adapt the time step of the method.,31657
ubiquitous neural medical currently (DCNN) in are imaging. networks convolutional Deep,Deep convolutional neural networks (DCNN) are currently ubiquitous in medical imaging.,31658
also promise has large-scale medical data in for It great retrieval. improving accuracies,It has also great promise for improving accuracies in large-scale medical data retrieval.,31659
a originating Noise in the environment several sources of from degrades RF systems. performance communication,Noise originating from several sources in a RF environment degrades the performance of communication systems.,31660
was The error. square performance mean and two done bit metrics: comparison error using rate,The performance comparison was done using two metrics: bit error rate and mean square error.,31661
Gaussian The nonlinear LMS and both noise. PSO under that outperforms show results random,The results show that PSO outperforms LMS under both Gaussian and nonlinear random noise.,31662
to Any learning choose optimization gradient a requires descent rate.,Any gradient descent optimization requires to choose a learning rate.,31663
"descent gradient This any can for optimized. learning algorithm machine be way,","This way, gradient descent for any machine learning algorithm can be optimized.",31664
"surfaces manually image, each five graders expert each two annotated retinal For (i.e.","For each image, two expert graders each manually annotated five retinal surfaces (i.e.",31665
the segmentation of automatic state art techniques. proposed compare method against We the five,We compare the proposed method against five state of the art automatic segmentation techniques.,31666
consider variants problem. of this We different,We consider different variants of this problem.,31667
episodes. minimize The number of such goal to the is,The goal is to minimize the number of such episodes.,31668
of the goal minimize is number events. to reticulation The,The goal is to minimize the number of reticulation events.,31669
"high Real-time remain however, algorithms, and accuracy tracking scarce.","Real-time and high accuracy tracking algorithms, however, remain scarce.",31670
important is component segmentation Image systems. image understanding many an of,Image segmentation is an important component of many image understanding systems.,31671
aims spatially group It a perceptually in coherent to manner. and pixels,It aims to group pixels in a spatially and perceptually coherent manner.,31672
"the have collection these of parameters that Typically, control over-segmentation of a algorithms degree produced.","Typically, these algorithms have a collection of parameters that control the degree of over-segmentation produced.",31673
such human-like properly select perceptual remains challenge parameters still grouping. a for to It,It still remains a challenge to properly select such parameters for human-like perceptual grouping.,31674
"In the by of choices segments this different we produced work, of exploit diversity parameters.","In this work, we exploit the diversity of segments produced by different choices of parameters.",31675
"state-of-the-art with our we can performance compare achieve improved showing that We algorithm's results. algorithms,","We compare our algorithm's performance with state-of-the-art algorithms, showing that we can achieve improved results.",31676
facial of to points face localization landmarks images. on refer the Facial fundamental,Facial landmarks refer to the localization of fundamental facial points on face images.,31677
has been used CelebA BEGAN database. from generator train framework face a The to,The BEGAN framework has been used to train a face generator from CelebA database.,31678
"deep is learning especially prerequisite, a usually Adequate when adopted. data is","Adequate data is usually a prerequisite, especially when deep learning is adopted.",31679
neural representation extract to region-level for prediction. spatio-temporal is designed network A effectively,A neural network is designed to effectively extract region-level representation for spatio-temporal prediction.,31680
"we of flow crowd the prediction Using a verify experiment, citywide demonstration as RegionTrans. effectiveness","Using citywide crowd flow prediction as a demonstration experiment, we verify the effectiveness of RegionTrans.",31681
"approximate In to has methods. one to such scenarios, resort","In such scenarios, one has to resort to approximate methods.",31682
"approximation both application. Finally, evaluate a we the data and synthetic with real-world","Finally, we evaluate the approximation with both synthetic data and a real-world application.",31683
"issue this scarcity hyperspectral we the dataset with paper, In classification. the image address","In this paper, we address the dataset scarcity issue with the hyperspectral image classification.",31684
on Experimental results algorithm dataset state-of-the-art the outperforms the video-story the show that approach. proposed,Experimental results on the video-story dataset show that the proposed algorithm outperforms the state-of-the-art approach.,31685
"neural captioning achieved promising have in Recently, image models network deep results task.","Recently, deep neural network models have achieved promising results in image captioning task.",31686
the have proposed of demonstrated method. Experiments effectiveness the,Experiments have demonstrated the effectiveness of the proposed method.,31687
series modeling in extremely complex successful data. proved time LSTMs),LSTMs) proved extremely successful in modeling complex time series data.,31688
"train, however, are Fully SSMs, for problems. smaller to probabilistic even often hard found","Fully probabilistic SSMs, however, are often found hard to train, even for smaller problems.",31689
These correlations training. are key robust the to,These correlations are the key to robust training.,31690
and on robustness Scalability high dimensional demonstrated are problem. a,Scalability and robustness are demonstrated on a high dimensional problem.,31691
"achieved several Recently, have Adversarial (GANs) success. Generative Nets methods great conditional based","Recently, several conditional Generative Adversarial Nets (GANs) based methods have achieved great success.",31692
images generate age fitting group. face individual distributions They conditioned each real can the on,They can generate images fitting the real face distributions conditioned on each individual age group.,31693
two The of discriminative network and consists C-GANs transformation conditional a networks.,The C-GANs consists of a conditional transformation network and two discriminative networks.,31694
blocks. residual conditional transformation specially with network imitates designed The several procedure aging the,The conditional transformation network imitates the aging procedure with several specially designed residual blocks.,31695
the network to distribution. The real age guides the fit synthesized discriminative conditional face,The age discriminative network guides the synthesized face to fit the real conditional distribution.,31696
face gain cross-age verification. observe We also for performance,We also observe performance gain for cross-age face verification.,31697
decoder. encoder a recurrent and (RNNs) employs proposed an networks neural The both TRAN as,The proposed TRAN employs recurrent neural networks (RNNs) as both an encoder and a decoder.,31698
the histology liver reptiles. is the of of known Little about,Little is known about the histology of the liver of reptiles.,31699
"lipid RER hepatocytes and and numerous abundant glycogen granules well-developed mitochondria. droplets, contained The","The hepatocytes contained numerous lipid droplets, abundant glycogen granules and well-developed RER and mitochondria.",31700
space the Disse hepatocytes many separating sinusoids contained The between recesses. of and,The space of Disse separating between the hepatocytes and sinusoids contained many recesses.,31701
hyper-rational the to generalize game theory. We and the expand choice implement,We implement the hyper-rational choice to generalize and expand the game theory.,31702
how decisions. Hyper-rationality makers understand behave helps interactive us human in decision,Hyper-rationality helps us understand how human decision makers behave in interactive decisions.,31703
"segmentation non-trivial Unfortunately, into is inevitably errors and propagates temporal steps. subsequent itself","Unfortunately, temporal segmentation itself is non-trivial and inevitably propagates errors into subsequent steps.",31704
out large scale carried Experiments datasets. are on two,Experiments are carried out on two large scale datasets.,31705
results framework. the Experimental demonstrate effectiveness of the proposed,Experimental results demonstrate the effectiveness of the proposed framework.,31706
"affects in Therefore, iris overall performance. recognition segmentation the failure","Therefore, failure in iris segmentation affects the overall recognition performance.",31707
"SegDenseNet, The deep DenseNets. termed algorithm as algorithm, learning is proposed on based a","The proposed algorithm, termed as SegDenseNet, is a deep learning algorithm based on DenseNets.",31708
"are little which attention. image has we received generating sentiment-conveying particular, In descriptions, in interested","In particular, we are interested in generating sentiment-conveying image descriptions, which has received little attention.",31709
distance different brain approach states lacking. pairwise A functional to systematic at is measure,A systematic approach to measure pairwise functional distance at different brain states is lacking.,31710
lifted states. multiset a on for Bayesian based model present recursive exact We filtering,We present a model for exact recursive Bayesian filtering based on lifted multiset states.,31711
with These results requirements aim Saussurean humanists. fulfill to elaborated,These results aim to fulfill requirements elaborated with Saussurean humanists.,31712
"include others. facial gaze usage Potential emotion many eye scenarios recognition, recognition, tracking and","Potential usage scenarios include facial recognition, emotion recognition, eye gaze tracking and many others.",31713
"learning, with inductive When process enhance perform humans knowledge. often the background they","When humans perform inductive learning, they often enhance the process with background knowledge.",31714
classifier original then the The a resulting is used as for problem. new feature,The resulting classifier is then used as a new feature for the original problem.,31715
have the features performance learning improve existing We generated of that the significantly shown algorithms.,We have shown that the generated features significantly improve the performance of existing learning algorithms.,31716
of (OWA) Averaging a this new Weighted Ordered we family In operators. paper develop,In this paper we develop a new family of Ordered Weighted Averaging (OWA) operators.,31717
from vector desired Weight orness a of is obtained operator. the,Weight vector is obtained from a desired orness of the operator.,31718
"line relation. the of follow remaining exception straight one the weight, a With","With the exception of one weight, the remaining follow a straight line relation.",31719
"and a a result, As is fast developed. algorithm robust","As a result, a fast and robust algorithm is developed.",31720
"a of cancer highest the Colorectal death, in cancer-related one causes is of men. especially","Colorectal cancer is a one of the highest causes of cancer-related death, especially in men.",31721
a this based segmentation proposed neural polyp In we on convolutional network. method paper,In this paper we proposed a polyp segmentation method based on convolutional neural network.,31722
of is method enhanced strategies. Performance the two by,Performance of the method is enhanced by two strategies.,31723
"Yet, this explanation is no of there success. theoretical","Yet, there is no theoretical explanation of this success.",31724
"with are generated First, by representations a order. joints previous fixed chaining skeleton","First, previous skeleton representations are generated by chaining joints with a fixed order.",31725
semantic is corresponding joints the is meaning unclear and structural The lost. the information among,The corresponding semantic meaning is unclear and the structural information among the joints is lost.,31726
"previous an not focus on joints. Second, ability have models to do informative","Second, previous models do not have an ability to focus on informative joints.",31727
is first base with A the simplest attention introduced. model structure,A base attention model with the simplest structure is first introduced.,31728
experiment state-of-the-art. Kinetic NTU SBU the RGB+D results on Interaction Our and outperforms,Our experiment results on NTU RGB+D and SBU Kinetic Interaction outperforms the state-of-the-art.,31729
"women's concerns decision social affect this we to how compete. investigate paper, In image","In this paper, we investigate how social image concerns affect women's decision to compete.",31730
predictions. our consistent results are fairly the theoretical experiment with of The,The results of the experiment are fairly consistent with our theoretical predictions.,31731
in incidence prostate requirement diagnosis. disease for high a poses early rate The of detection,The high incidence rate of prostate disease poses a requirement in early detection for diagnosis.,31732
"Jaccard and functions ConNet. adopt loss DSC train DDSP loss Meanwhile, our we to","Meanwhile, we adopt DSC loss and Jaccard loss functions to train our DDSP ConNet.",31733
"the using data bivariate challenging, Causal case. in especially is observational inference","Causal inference using observational data is challenging, especially in the bivariate case.",31734
static cells. model The in available stem outperforms counting cell of publicly a dataset,The model outperforms static cell counting in a publicly available dataset of stem cells.,31735
"ConvNet-LSTM limitations are working of and the discussed. advantages, conditions The method","The advantages, working conditions and limitations of the ConvNet-LSTM method are discussed.",31736
"an existing ANN the (ANNC) model was introduced. works, by center-loss In supervised","In the existing works, an ANN model supervised by center-loss (ANNC) was introduced.",31737
temporal from obtained RGB only flow. Many focus on term optical short information methods,Many RGB methods focus only on short term temporal information obtained from optical flow.,31738
Light has recent years. photography field thoroughly in studied been,Light field photography has been studied thoroughly in recent years.,31739
for drawbacks One in the the of is need imaging. multi-lens its,One of its drawbacks is the need for multi-lens in the imaging.,31740
"multi-lens system. obtains lens, compressed It version regular a by one only the of","It obtains by only one lens, a compressed version of the regular multi-lens system.",31741
approach existing recovery complexity. quality solutions and computational of Our outperforms terms in,Our approach outperforms existing solutions in terms of recovery quality and computational complexity.,31742
"the years, greatly for Recent sensing has improved learning (CS) technology. of application compressive deep","Recent years, compressive sensing (CS) has improved greatly for the application of deep learning technology.",31743
"input For and block. is the usually by measured block image reconstructed convenience,","For convenience, the input image is usually measured and reconstructed block by block.",31744
in causes images. usually This reconstructed block effect,This usually causes block effect in reconstructed images.,31745
"present novel network this paper, we solve In this CNN-based a problem. to","In this paper, we present a novel CNN-based network to solve this problem.",31746
proposed framework function. We by the error square mean loss (MSE) train,We train the proposed framework by mean square error (MSE) loss function.,31747
no is effect all the show in proposed that block there Experiments method. at,Experiments show that there is no block effect at all in the proposed method.,31748
are performed Extensive our experiments ver- comparisons ify to method. and,Extensive experiments and comparisons are performed to ver- ify our method.,31749
pixel-level annotated build images a of dataset of HOC. consisting We,We build a dataset consisting of pixel-level annotated images of HOC.,31750
and the dataset The published code will paper. with be,The dataset and code will be published with the paper.,31751
"such blocking Though from suffer artifacts. methods serious efficient,","Though efficient, such methods suffer from serious blocking artifacts.",31752
"the is convolutional Instead measurements. whole the linear reconstructed image of reconstructing individual blocks, from","Instead of reconstructing individual blocks, the whole image is reconstructed from the linear convolutional measurements.",31753
"the training, can filters sensing end-to-end the network jointly and Through be optimized. reconstruction","Through end-to-end training, the sensing filters and the reconstruction network can be jointly optimized.",31754
(CS) sensing Compressive measurements scene at recover the to and works sub-Nyquist acquire rate images.,Compressive sensing (CS) works to acquire measurements at sub-Nyquist rate and recover the scene images.,31755
the in level. pixel scene methods Existing CS images recover always,Existing CS methods always recover the scene images in pixel level.,31756
level. Our focuses task pixel no on longer,Our task no longer focuses on pixel level.,31757
"better Instead, effect. visual make a to we work","Instead, we work to make a better visual effect.",31758
are 'Casablanca'. on presented and TV the feature series Results 'Sherlock' film the,Results are presented on the TV series 'Sherlock' and the feature film 'Casablanca'.,31759
"we artificial discuss intelligence. prospective in trends explainable Finally,","Finally, we discuss prospective trends in explainable artificial intelligence.",31760
"of validating present transduction simulating signal a building, for We networks. protocol models and","We present a protocol for building, validating and simulating models of signal transduction networks.",31761
"networks. of validate transduction models Here, present to and method large signal build we a","Here, we present a method to build and validate large models of signal transduction networks.",31762
"is workflow the rxncon, based on reaction-contingency language. The","The workflow is based on rxncon, the reaction-contingency language.",31763
"of problem paper, an we address this In learning. zero-shot open","In this paper, we address an open problem of zero-shot learning.",31764
on that vectors Its a i.e. mapping principle based from learning feature associates extracted is,Its principle is based on learning a mapping that associates feature vectors extracted from i.e.,31765
and and/or describe interest. of vectors objects attribute scenes images that,images and attribute vectors that describe objects and/or scenes of interest.,31766
assumption The when deemed this fails. phenomenon contextual is,The phenomenon is deemed contextual when this assumption fails.,31767
in important quantum Contextuality an issue is physics.,Contextuality is an important issue in quantum physics.,31768
"is Specific syntax of allow proposed contexts"". to ""measurement the specification","Specific syntax is proposed to allow the specification of ""measurement contexts"".",31769
is translated into a probabilistic modular way. a The program in hypergraph,The probabilistic program is translated into a hypergraph in a modular way.,31770
"In of paper, (CS) matrices deterministic in considered. this measurement construction Compressive is Sensing","In this paper, deterministic construction of measurement matrices in Compressive Sensing (CS) is considered.",31771
"over sensing linear these constructed. theorem an codes, matrices existing by applying are deterministic Then,","Then, by applying an existing theorem over these linear codes, deterministic sensing matrices are constructed.",31772
"evaluate constructed are matrices this presented. of examples sensing procedure, To two","To evaluate this procedure, two examples of constructed sensing matrices are presented.",31773
"Welch optimal both Based on the bound, asymptotically results. examples achieve","Based on the Welch bound, both examples asymptotically achieve optimal results.",31774
"a a penalized for is generator not mode. missing Specifically,","Specifically, a generator is not penalized for missing a mode.",31775
full We susceptible show these still distribution. the to capturing data that are therefore not,We show that these are therefore still susceptible to not capturing the full data distribution.,31776
critical are support Explainability two decision and systems. interpretability of aspects,Explainability and interpretability are two critical aspects of decision support systems.,31777
"researchers are only to it these is starting recently their explore Despite importance, that aspects.","Despite their importance, it is only recently that researchers are starting to explore these aspects.",31778
"those study analysis. in the first impressions and mechanisms review context of Specifically, we","Specifically, we review and study those mechanisms in the context of first impressions analysis.",31779
"this knowledge, direction. this the To is in our effort first of best the","To the best of our knowledge, this is the first effort in this direction.",31780
"in describe first on video. organized Additionally, we impressions we challenge a analysis from explainability","Additionally, we describe a challenge we organized on explainability in first impressions analysis from video.",31781
proposed methods artistic portraits style styles Numerous transfer which date. have produce to been of,Numerous style transfer methods which produce artistic styles of portraits have been proposed to date.,31782
"Style of develop Removal We layers. composed fully-connected deconvolutional convolutional, a and Network","We develop a Style Removal Network composed of convolutional, fully-connected and deconvolutional layers.",31783
stylized components facial The designed are layers to from face extract images. convolutional,The convolutional layers are designed to extract facial components from stylized face images.,31784
learned apply synthesizability synthesizable detecting regions to the We videos. finally in,We finally apply the learned synthesizability to detecting synthesizable regions in videos.,31785
"learning within network become acceleration the hot a Thus, deep community. topic has","Thus, network acceleration has become a hot topic within the deep learning community.",31786
"few we a Finally, directions. and possible future discuss introduce will","Finally, we will introduce and discuss a few possible future directions.",31787
the particles for difficult micrographs contrast. low small from Selecting is especially with particles,Selecting particles from the micrographs is difficult especially for small particles with low contrast.,31788
"somewhat time-consuming. particle still addition, is semi-automated picking In","In addition, semi-automated particle picking is still somewhat time-consuming.",31789
"by was inspired it approach our is matching, template template-free. completely While","While our approach was inspired by template matching, it is completely template-free.",31790
stripping) neuroimaging. problem (skull is extraction in a challenging Brain,Brain extraction (skull stripping) is a challenging problem in neuroimaging.,31791
been (CNN) tasks. for has network widely used vision-based Convolutional neural,Convolutional neural network (CNN) has been widely used for vision-based tasks.,31792
for energy-constrained The and hardware-friendly re-source-limited needed designs embed-ded are devices.,The hardware-friendly designs are needed for re-source-limited and energy-constrained embed-ded devices.,31793
has shrunk size existing of the BNN. work further No,No existing work has further shrunk the size of BNN.,31794
The to bit-sliced dimensional converted for-mat. data high input a is,The input data is converted to a high dimensional bit-sliced for-mat.,31795
"stage, bit the we post-training im-pact In of analyze slices the different accuracy. to","In post-training stage, we analyze the im-pact of different bit slices to the accuracy.",31796
the interplay among emerged to quantify bands was rhythms. The compared structure across frequency,The emerged structure was compared across frequency bands to quantify the interplay among rhythms.,31797
is Remote cartography. sensing extensively in used,Remote sensing is extensively used in cartography.,31798
provide satellites topographical Synthetic Aperture maps. can Radar resolution high,Synthetic Aperture Radar satellites can provide high resolution topographical maps.,31799
SAR evaluation This road for Fully-Convolutional Networks images. an presents Neural segmentation letter in of,This letter presents an evaluation of Fully-Convolutional Neural Networks for road segmentation in SAR images.,31800
"in promising roads Our of test dataset. shows results, models the successfully extracting most our","Our models shows promising results, successfully extracting most of the roads in our test dataset.",31801
"of segmentation conduct interest. we region of the First,","First, we conduct segmentation of the region of interest.",31802
"model. using network (i.e., a neural pouches) deep","(i.e., pouches) using a deep neural network model.",31803
"and Second, obtain we image an align pouches sequences. across transformation optimal the","Second, we obtain an optimal transformation and align pouches across the image sequences.",31804
that references (CPS). in data issues concern sensor It mining systems physical cyber,It references issues that concern mining sensor data in cyber physical systems (CPS).,31805
"hybrid real-time includes and This systems. networks, sensor","This includes sensor networks, real-time and hybrid systems.",31806
sensor Such literature are of not data. mining problems addressing sufficiently covered by,Such problems are not covered sufficiently by literature addressing mining of sensor data.,31807
link there knowledge causality - be discovery. - this based no Without without can physics,Without this link - without causality - there can be no physics based knowledge discovery.,31808
"kinds explanation human-interpretable exactly understood. are truly remains poorly However, what of","However, exactly what kinds of explanation are truly human-interpretable remains poorly understood.",31809
"to HRF uncertainties Moreover, method proposed the cope the related with can efficiently modeling.","Moreover, the proposed method can efficiently cope with uncertainties related to the HRF modeling.",31810
"the Hence, available information that from using parameters atlases. be can brain tuned related is","Hence, the related parameters can be tuned using information that is available from brain atlases.",31811
"evaluated including techniques, method proposed is several against other The GLM. popular","The proposed method is evaluated against several other popular techniques, including GLM.",31812
the of influence removing on the illumination of to image scene. colors the,to removing the influence of illumination on the colors of the image scene.,31813
Some are estimation the known of illumination called methods methods. best statistics-based so the,Some of the best known illumination estimation methods are the so called statistics-based methods.,31814
After corrected best the known datasets. given are the that for benchmark results,After that the corrected results are given for the best known benchmark datasets.,31815
and results are presented The discussed. experimental,The experimental results are presented and discussed.,31816
"recognition using for on-body human-labeled However, sequential from activity data are lacking. human attributes sensors,","However, for human activity recognition using sequential data from on-body sensors, human-labeled attributes are lacking.",31817
"centered predicting design, an presents and for three IMU attributes. temporal-convolutions including architectures, It deep","It presents three deep architectures, including temporal-convolutions and an IMU centered design, for predicting attributes.",31818
"the applicability explanations such such, of and suspect. to remains interactions human-robot human-AI As","As such, the applicability of such explanations to human-AI and human-robot interactions remains suspect.",31819
cell are cartilage Chondrocytes different types. as population one described in,Chondrocytes are described as one cell population in different cartilage types.,31820
mentioned species. not other were Lipochondrocytes in,Lipochondrocytes were not mentioned in other species.,31821
the rabbits. current explore type cell in this to of The existence study aimed,The current study aimed to explore the existence of this cell type in rabbits.,31822
result. this electron microscopy scanning The confirmed,The scanning electron microscopy confirmed this result.,31823
transplant. result should cartilage be This during considered,This result should be considered during cartilage transplant.,31824
"systems, current and in the machine learning reasoning incompatible. are modules However, perception","However, in current machine learning systems, the perception and reasoning modules are incompatible.",31825
learns abductive and with trial-and-error the reasoning help The perception process. a framework of simultaneously,The framework learns perception and reasoning simultaneously with the help of a trial-and-error abductive process.,31826
present framework. Machine an novel as learning implementation this of Neural-Logical the We,We present the Neural-Logical Machine as an implementation of this novel learning framework.,31827
direction for learning explores ability. The framework approaching abductive human-level a learning new,The abductive learning framework explores a new direction for approaching human-level learning ability.,31828
its proposed existing and Simulation method of techniques. over verify results the the effectiveness superiority,Simulation results verify the effectiveness of the proposed method and its superiority over existing techniques.,31829
We mode-seeking a on cast task as hypergraphs. representative of problem geometric the model fitting,We cast the task of geometric model fitting as a representative mode-seeking problem on hypergraphs.,31830
These recordings athletes' can to used be performance. the assess quantitatively,These recordings can be used to quantitatively assess the athletes' performance.,31831
"frame. annotations far, in parts so manual each body video The quantitative evaluation, of requires","The quantitative evaluation, so far, requires manual annotations of body parts in each video frame.",31832
novel training this We to approach difficulty. a address propose,We propose a novel training approach to address this difficulty.,31833
"labels image dense we propagate the guessed labelings. labelings, to cheaply-obtained sparse sparse produce Given","Given cheaply-obtained sparse image labelings, we propagate the sparse labels to produce guessed dense labelings.",31834
segmentation these CNN-based mimic standard trained A labelings. is network to,A standard CNN-based segmentation network is trained to mimic these labelings.,31835
Experiments outperforms segmentation naive in approach. that show also way network this training a the,Experiments also show that training a segmentation network in this way outperforms the naive approach.,31836
the railway. maintenance of for all pivotal of functioning assets Regular proper is,Regular maintenance of all the assets is pivotal for proper functioning of railway.,31837
for and cumbersome Manual leave room maintenance be very errors. can,Manual maintenance can be very cumbersome and leave room for errors.,31838
health. a Hence rail continuously is to need track monitor there,Hence there is a need to continuously monitor rail track health.,31839
for kink simulated overgrowth. sun images vegetation Training are and,Training images are simulated for sun kink and vegetation overgrowth.,31840
problem remote the high-resolution sensing to imagery. Scene a is classification understand fundamental,Scene classification is a fundamental problem to understand the high-resolution remote sensing imagery.,31841
"on ConvNet In based a paper, aggregation. we context novel present representation with this a","In this paper, we present a novel representation based on a ConvNet with context aggregation.",31842
"to Argentina? Messi Dybala, in Paulo heir it the Is of the team national","Is it Paulo Dybala, the heir to Messi in the national team of Argentina?",31843
someone Or else? the is completely Messi alike most player to,Or is the most alike player to Messi someone completely else?,31844
stressful is job. Pilot tedious very and Train a,Train Pilot is a very tedious and stressful job.,31845
The an is increased outcome pilots. safety of train,The outcome is an increased safety of train pilots.,31846
Dirichlet to weights have are The distribution. prior assumed a,The weights are assumed to have a Dirichlet prior distribution.,31847
conclusions. instability We approach suffers produce an that that from this show can misleading,We show that this approach suffers from an instability that can produce misleading conclusions.,31848
We how mean is instability this remedied by discuss field more approximations. accurate,We discuss how this instability is remedied by more accurate mean field approximations.,31849
"methods. several successful of this we and review In popular these most the chapter, of","In this chapter, we review several of the most popular and successful of these methods.",31850
"learning. been to this has reinforcement ""model-based"" Recently ""model-free"" dichotomy linked and","Recently this dichotomy has been linked to ""model-based"" and ""model-free"" reinforcement learning.",31851
one common the in cancers the States. Colorectal is of cancer United,Colorectal cancer is one of the common cancers in the United States.,31852
"some have However, limitations. application specific domains","However, some application domains have specific limitations.",31853
"which problem, incremental this forgetting. classifier learning the from we address catastrophic In suffers paper,","In this paper, we address the incremental classifier learning problem, which suffers from catastrophic forgetting.",31854
"with these are there However, main approaches. four problems","However, there are four main problems with these approaches.",31855
"is loss not efficient classification. the function First, for","First, the loss function is not efficient for classification.",31856
"problem new classes. there between is the past Second, and unbalance","Second, there is unbalance problem between the past and new classes.",31857
with present potentially important results. Byrne and a paper on topic a colleagues timely,Byrne and colleagues present a paper on a timely topic with potentially important results.,31858
"networks. end-to-end compression an framework we with propose neural convolutional image mixed-resolution paper, this In","In this paper, we propose an end-to-end mixed-resolution image compression framework with convolutional neural networks.",31859
skin of cancer. deadliest is form the Melanoma,Melanoma is the deadliest form of skin cancer.,31860
"in can but Computer practice. not melanoma in assist systems detection, widespread clinical are","Computer systems can assist in melanoma detection, but are not widespread in clinical practice.",31861
had ranking. Choice great measure performance impact on the of,Choice of performance measure had great impact on the ranking.,31862
resemblance improve diagnostic substantially. semi-automatic/manual segmentation will to accuracy not,resemblance to semi-automatic/manual segmentation will not improve diagnostic accuracy substantially.,31863
worldwide. is cancer the cancer main one of causes Breast of death,Breast cancer is one of the main causes of cancer death worldwide.,31864
diagnosis the diagnostic Computer-aided improving showed potential for accuracy. systems,Computer-aided diagnosis systems showed potential for improving the diagnostic accuracy.,31865
neural approach deep Our architectures network utilizes trees several gradient and boosted classifier.,Our approach utilizes several deep neural network architectures and gradient boosted trees classifier.,31866
"To classification. this in our methods approach image automated other histopathological knowledge, common outperforms","To our knowledge, this approach outperforms other common methods in automated histopathological image classification.",31867
analysis a for mixed-frequency structural develop new data. We model with VAR,We develop a new VAR model for structural analysis with mixed-frequency data.,31868
homogeneous It test a representations data mixed-frequency versus provides general frequency-based to also models. framework,It also provides a general framework to test homogeneous frequency-based representations versus mixed-frequency data models.,31869
synthesized images. These high-spatial-frequency methods can in in loss of turn suffer from information,These methods can in turn suffer from loss of high-spatial-frequency information in synthesized images.,31870
utilized Information cross-sections to synthesis are neighboring further from quality. improved,Information from neighboring cross-sections are utilized to further improved synthesis quality.,31871
a paper Decomposition Empirical model Mode of we present mathematical (EMD). the this In,In this paper we present a mathematical model of the Empirical Mode Decomposition (EMD).,31872
gaining hybrid These intelligence approaches hence are traction.,These hybrid intelligence approaches are hence gaining traction.,31873
"becomes Manual time-consuming multiple tracking, markers, for especially and impossible.","Manual tracking, especially for multiple markers, becomes time-consuming and impossible.",31874
on structural We evaluate the neural the two network recurrent Volleyball of variants Dataset.,We evaluate two variants of the structural recurrent neural network on the Volleyball Dataset.,31875
"the of types cancers among deadliest the cancers, are most common Skin them. being melanoma","Skin cancers are the most common types of cancers, melanoma being the deadliest among them.",31876
The of for morphology used diagnosis these are clinicians and melanomas. by occurrence of patterns,The occurrence and morphology of these patterns are used by clinicians for diagnosis of melanomas.,31877
by analysis al. Dmochowski component proposed et as Correlated,Correlated component analysis as proposed by Dmochowski et al.,31878
spatial are assumption identified are that identical. the involved Correlated under components the networks,Correlated components are identified under the assumption that the involved spatial networks are identical.,31879
"network deep Overfitting neural is problem the in in latest even architectures. networks, crucial a","Overfitting is a crucial problem in deep neural networks, even in the latest network architectures.",31880
training. is to stability key important An of achieve,An important key is to achieve stability of training.,31881
"the which well. demonstrate Through conditions various experiments ShakeDrop we conditions, under under works","Through experiments under various conditions, we demonstrate the conditions under which ShakeDrop works well.",31882
"""coarse-to-fine"" single i.e. deblurring, the scheme, In image","In single image deblurring, the ""coarse-to-fine"" scheme, i.e.",31883
complex motion. datasets We our large-scale evaluate on method deblurring with,We evaluate our method on large-scale deblurring datasets with complex motion.,31884
"focus Image-Based efforts Retrieval on single-modal mainly (IBIR). Existing retrieval Image task, especially","Existing efforts mainly focus on single-modal retrieval task, especially Image-Based Image Retrieval (IBIR).",31885
errors. (AI) All Intelligence make artificial systems,All artificial Intelligence (AI) systems make errors.,31886
"mistakes the unexpected, from errors differ often human errors). typical These (""non-human"" are and","These errors are unexpected, and differ often from the typical human mistakes (""non-human"" errors).",31887
"used landmark facial to variational Specifically, is auto-encoder a a embedding. learn","Specifically, a variational auto-encoder is used to learn a facial landmark embedding.",31888
"are Finally, these translated sequences face into videos. landmark","Finally, these landmark sequences are translated into face videos.",31889
"volume, composite In values, estimate at instead the every we a tensor of voxel. intensity","In the composite volume, instead of intensity values, we estimate a tensor at every voxel.",31890
of data vivo in approach. on the results human show the present efficacy We to,We present results on in vivo human data to show the efficacy of the approach.,31891
the learning for desired focus objective a robot. on function We,We focus on learning the desired objective function for a robot.,31892
"it a topic, developments many this problem. remains challenging previous on Despite","Despite many previous developments on this topic, it remains a challenging problem.",31893
is using Evaluation real performed and synthetic data.,Evaluation is performed using synthetic and real data.,31894
"is an Moreover, of evaluation Blensor realistic carried figures using obtained out.","Moreover, an evaluation of realistic figures obtained using Blensor is carried out.",31895
isolation to primary certain (e.g. be prezygotic those holding For of cases,For those holding certain cases of prezygotic isolation to be primary (e.g.,31896
"inviability cases Likewise, to (e.g. primary upon onus holding the hybrid is be of those","Likewise, the onus is upon those holding cases of hybrid inviability to be primary (e.g.",31897
"causing reproductive be only sympatric The barrier a latter, when can as acting isolation, primary.","The latter, when acting as a sympatric barrier causing reproductive isolation, can only be primary.",31898
"and key analyze of training the a series efficiency. Then, performance we problems concerning","Then, we analyze a series of key problems concerning the training performance and efficiency.",31899
natural Automatically research captioning an language images sentences is topic. important with,Automatically captioning images with natural language sentences is an important research topic.,31900
human-like art sentences. State are produce able the to models of,State of the art models are able to produce human-like sentences.,31901
"given important scene. to these describe attributes require marketing of a companies However,","However, marketing companies require to describe these important attributes of a given scene.",31902
"third output produces modality, image which real-valued add ratings. simultaneously a We also","We also add a third output modality, which simultaneously produces real-valued image ratings.",31903
"scaling in role reduction dimensionality central Multi-dimensional visualization. and plays data-exploration, a (MDS)","Multi-dimensional scaling (MDS) plays a central role in data-exploration, dimensionality reduction and visualization.",31904
capabilities into difficult approaches. integrate Such to regression-based are classic,Such capabilities are difficult to integrate into classic regression-based approaches.,31905
"be Perhaps this surprisingly, that can learned we in show still classifiers setting.","Perhaps surprisingly, we show that classifiers can still be learned in this setting.",31906
arbitrary be those of different inferred latent of an view views. features using Thus can,Thus latent features of an arbitrary view can be inferred using those of different views.,31907
enables characteristic to limited viewing RLS. us from data further This in augment,This characteristic further enables us to augment data from limited viewing in RLS.,31908
hybrid indicator-based research algorithms. This proposes that combines a approach novel approximate and evolutionary exact,This research proposes a novel indicator-based hybrid evolutionary approach that combines approximate and exact algorithms.,31909
"pixels than non-crack are pixels. much Typically, crack fewer","Typically, crack pixels are much fewer than non-crack pixels.",31910
five with tested method two and The methods. public on databases existing compared is,The method is tested on two public databases and compared with five existing methods.,31911
methods. outperforms Experimental that the show other it results,Experimental results show that it outperforms the other methods.,31912
challenging. Reinforcement learning many in pairs is environments action-state with,Reinforcement learning in environments with many action-state pairs is challenging.,31913
policy the number is episodes thoroughly space. At the to search needed issue of,At issue is the number of episodes needed to thoroughly search the policy space.,31914
problem stochastic this manner. address heuristics Most in conventional search a,Most conventional heuristics address this search problem in a stochastic manner.,31915
leave training policy the can large stages. of This the space during portions early unvisited,This can leave large portions of the policy space unvisited during the early training stages.,31916
problems Pedestrian in emerging car one detection the key is of self-driving industry.,Pedestrian detection is one of the key problems in emerging self-driving car industry.,31917
has proven algorithm good for accuracy And pedestrian detection. provide to HOG,And HOG algorithm has proven to provide good accuracy for pedestrian detection.,31918
in SVM-based realizing a system. pedestrian prediction The detection integrated accelerator with is,The accelerator is integrated with SVM-based prediction in realizing a pedestrian detection system.,31919
whole the existing the the consumption power is comparable with system of best implementations. And,And the power consumption of the whole system is comparable with the best existing implementations.,31920
multi-frequency from phase (power) intensity considered. The is retrieval observations,The phase retrieval from multi-frequency intensity (power) observations is considered.,31921
is object to The complex-valued. reconstructed be,The object to be reconstructed is complex-valued.,31922
algorithm is alternating between variables. projections The the object and with iterative measurement,The algorithm is iterative with alternating projections between the object and measurement variables.,31923
"self-supervised methods this of complicating explore tasks representation In for learning. paper, we","In this paper, we explore methods of complicating self-supervised tasks for representation learning.",31924
"we That to to do and data them. is, encourage network damage recover severe a","That is, we do severe damage to data and encourage a network to recover them.",31925
"First, colorization. jigsaw self-supervised inpainting, each three task we powerful and candidates: complicate of puzzle,","First, we complicate each of three powerful self-supervised task candidates: jigsaw puzzle, inpainting, and colorization.",31926
the representations. aforementioned pushes general-purpose robust damage obtain recovery the network to The and of,The recovery of the aforementioned damage pushes the network to obtain robust and general-purpose representations.,31927
learning Our achieves in performance classification on and transfer approach semantic segmentation. PASCAL state-of-the-art,Our approach achieves state-of-the-art performance in transfer learning on PASCAL classification and semantic segmentation.,31928
data sensors observation temporal or resolution. deliver Earth (EO) weekly daily with,Earth observation (EO) sensors deliver data with daily or weekly temporal resolution.,31929
"land approaches, use Most (LULC) however, and and mono-temporal expect cover cloud-free land observations.","Most land use and land cover (LULC) approaches, however, expect cloud-free and mono-temporal observations.",31930
Neural Convolution power. increasingly scale Networks Large deep (CNNs) demands the computing,Large scale deep Convolution Neural Networks (CNNs) increasingly demands the computing power.,31931
to trainable We become further network allow layer extend to the CDCT the (i.e.,We further extend the network to allow the CDCT layer to become trainable (i.e.,31932
"a varying Today, exist. good number scopes of solutions with","Today, a number of good solutions with varying scopes exist.",31933
capable of setting annotations SlideRunner (for object like e.g. is centers,SlideRunner is capable of setting annotations like object centers (for e.g.,31934
boundaries object as (e.g. as well cells),cells) as well as object boundaries (e.g.,31935
directly propose segmentation. We to piecewise motion intermediate affine new method estimate a without fields,We propose a new method to estimate piecewise affine motion fields directly without intermediate segmentation.,31936
framework key A is estimator our of signals. piecewise-affine for vector-valued one-dimensional an component efficient,A key component of our framework is an efficient one-dimensional piecewise-affine estimator for vector-valued signals.,31937
advantage The of first of initialization. approach methods absence segmentation-based is its our over,The first advantage of our approach over segmentation-based methods is its absence of initialization.,31938
handling in the material performed Lifting is common task a workplaces. manual,Lifting is a common manual material handling task performed in the workplaces.,31939
as Musculoskeletal risk Disorders. one is main of Work-related factors for It considered the,It is considered as one of the main risk factors for Work-related Musculoskeletal Disorders.,31940
"study, to aforementioned In based a approach this multi-view limitations. perceptron address propose deep we","In this study, we propose a multi-view based deep perceptron approach to address aforementioned limitations.",31941
"former our achieves performance that i.e. marker-based The methods, results approach prove comparable with","The results prove that our approach achieves comparable performance with former marker-based methods, i.e.",31942
"used different Recently, discipline. several been from variety to of problems solve have clustering algorithms","Recently, several clustering algorithms have been used to solve variety of problems from different discipline.",31943
any Probability are the density assumed not for of functions sources.,Probability density functions are not assumed for any of the sources.,31944
studies Viterbi the estimating path framework. article methods The Bayesian different for the in,The article studies different methods for estimating the Viterbi path in the Bayesian framework.,31945
simulated the performance to the methods. of compare used with are Examples data,Examples with simulated data are used to compare the performance of the methods.,31946
across information for The that suggest the entire cortex. sentence widespread semantic results is generations,The results suggest that semantic information for sentence generations is widespread across the entire cortex.,31947
search. addresses This of work the problem neighbor billion-scale nearest,This work addresses the problem of billion-scale nearest neighbor search.,31948
"recognition illumination The pose, of variation makes face still problem. and a expression challenging","The variation of pose, illumination and expression makes face recognition still a challenging problem.",31949
"a aligned are by pre-processing faces approaches, usually As eyes. holistic in","As a pre-processing in holistic approaches, faces are usually aligned by eyes.",31950
eye-aligned results the improvement using proposed a to Experimental great recognition. show method comparison in,Experimental results show a great improvement using the proposed method in comparison to eye-aligned recognition.,31951
results. are the event-stream assess to SR Two proposed metrics,Two metrics are proposed to assess the event-stream SR results.,31952
potential work of many applications enables This event-based vision.,This work enables many potential applications of event-based vision.,31953
unconstrained specialized detection is images facial from challenging a and problem. Smile,Smile detection from unconstrained facial images is a specialized and challenging problem.,31954
that improved approach proposed performance. show achieves Experiments state-of-the-art the,Experiments show that the proposed approach achieves improved state-of-the-art performance.,31955
Robustness this and paper. the also blur in model of noise evaluated to is artifacts,Robustness of the model to noise and blur artifacts is also evaluated in this paper.,31956
people motivates realworld to CNNs This deploy applications. to,This motivates people to deploy CNNs to realworld applications.,31957
datasets Experiments three of our demonstrate on widely-used effectiveness DRLIH proposed the approach.,Experiments on three widely-used datasets demonstrate the effectiveness of our proposed DRLIH approach.,31958
"very approaches are these expensive in However, computation. terms of learning deep","However, these deep learning approaches are very expensive in terms of computation.",31959
"used Units are (GPUs) Graphic in scale In implementations. for most Processing large cases,","In most cases, Graphic Processing Units (GPUs) are in used for large scale implementations.",31960
They frame show for results and analysis video sentiment promising both prediction.,They show promising results for both sentiment analysis and video frame prediction.,31961
"decide automatically of i.e. The number HDP clusters, the model","The HDP model automatically decide the number of clusters, i.e.",31962
of and the categories activities atomic interactions.,the categories of atomic activities and interactions.,31963
the (HDP) of Our the benefits (GP). model couples discriminant generative the framework with model,Our framework couples the benefits of the generative model (HDP) with the discriminant model (GP).,31964
two ability thoughts ago. years million approximately recursively together The evolved to chain,The ability to recursively chain thoughts together evolved approximately two million years ago.,31965
at second dates to Century Science least the fiction AD.,Science fiction dates to at least the second Century AD.,31966
is that thriving paradigm. Event technology digital Representation change processing could image a Address,Address Event Representation is a thriving technology that could change digital image processing paradigm.,31967
associated enhances connectivity based histogram that descriptor is A with two new spikes oriented representations.,A new descriptor that enhances spikes connectivity is associated with two oriented histogram based representations.,31968
"have been and developing Consequently, on made efficient algorithms. efforts filling effective hole","Consequently, efforts have been made on developing effective and efficient hole filling algorithms.",31969
"especially may complex not of scenarios However, the hold. texture, many assumption in","However, in many scenarios especially of complex texture, the assumption may not hold.",31970
"Generative (GANs), particular significant models, received attention adversarial in have networks recently. generative","Generative models, in particular generative adversarial networks (GANs), have received significant attention recently.",31971
have and have variants applications. in been of A utilized number GAN many proposed been,A number of GAN variants have been proposed and have been utilized in many applications.,31972
take We generalized a towards Convolutional using Neural this Networks step system (CNNs).,We take a step towards this generalized system using Convolutional Neural Networks (CNNs).,31973
narrow tasks. performs This distinct two intelligence artificial,This narrow artificial intelligence performs two distinct tasks.,31974
"it mobile of using agent's First, provides real-time the the position estimation algorithm. forward","First, it provides real-time estimation of the mobile agent's position using the forward algorithm.",31975
graphical efficiency We our method. present results and statistical illustrate of the to,We present statistical and graphical results to illustrate the efficiency of our method.,31976
Single attention super-resolution significant learning gained via deep the literature. has (SR) recently image in,Single image super-resolution (SR) via deep learning has recently gained significant attention in the literature.,31977
networks quality when for Most existing SR data training deep abundant. results is produce high,Most existing deep networks for SR produce high quality results when training data is abundant.,31978
"However, training when sharply limited. performance is their degrades","However, their performance degrades sharply when training is limited.",31979
recently tracking comprehensive survey of comprises algorithms. proposed first work part of a this The,The first part of this work comprises a comprehensive survey of recently proposed tracking algorithms.,31980
helps the cost step to and both improve A performance. reduce computational fine-tuning,A fine-tuning step helps to reduce both the computational cost and improve performance.,31981
"First, new into deep embedding neural networks. watermarks formulate a we problem:","First, we formulate a new problem: embedding watermarks into deep neural networks.",31982
"situations, and embedding We in on networks. neural watermarking define types attack requirements, also deep","We also define requirements, embedding situations, and attack types on watermarking in deep neural networks.",31983
"Currently, is very images bitewing of challenging a task. Segmentation radiograpy","Currently, Segmentation of bitewing radiograpy images is a very challenging task.",31984
hybrid for semantic deep This a image proposes (G-SHDL) network paper generative learning ScatterNet segmentation.,This paper proposes a generative ScatterNet hybrid deep learning (G-SHDL) network for semantic image segmentation.,31985
real noise. the by always environments corrupted In measurements are,In real environments the measurements are always corrupted by noise.,31986
optimization measurement noise algorithms. effects the non-linear analytical solution more than This,This measurement noise effects the analytical solution more than non-linear optimization algorithms.,31987
reduced how noisy This article this environments. in significantly be can approach presents risk an,This article presents an approach how this risk can be significantly reduced in noisy environments.,31988
"the both In work, this propose we advantages combine to methods. from","In this work, we propose to combine the advantages from both methods.",31989
encoder are and decoder schemes main leverage the concepts that introduced. Practical,Practical encoder and decoder schemes that leverage the main concepts are introduced.,31990
An developed approach the system. for is iterative optimizing,An iterative approach is developed for optimizing the system.,31991
of significant focuses optimal on HMS quantization a extension It paradigm. the also,It also focuses on a significant extension of the optimal HMS quantization paradigm.,31992
"exhibiting properties is of fondamental (birefringence, optics of The diattenuation) design polarization straightforward. filters","The design of filters exhibiting fondamental properties of polarization optics (birefringence, diattenuation) is straightforward.",31993
learning Honey memory. popular and bees are model a for,Honey bees are a popular model for learning and memory.,31994
Previous future interactions. has experience and social shown been modulate to affect,Previous experience has been shown to affect and modulate future social interactions.,31995
"colony. of for have far, all observations bees So reported no a history been lifetime","So far, no lifetime history observations have been reported for all bees of a colony.",31996
the join intriguing collective system. scientific animal investigate to We researchers invite effort this to,We invite researchers to join the collective scientific effort to investigate this intriguing animal system.,31997
open-source. our components of All are system,All components of our system are open-source.,31998
maintaining show preliminary visual Our efficiency increase while in results coding an quality. satisfactory,Our preliminary results show an increase in coding efficiency while maintaining satisfactory visual quality.,31999
evaluated on The datasets. proposed is benchmark several framework,The proposed framework is evaluated on several benchmark datasets.,32000
be preferred over one conditions should another? which technique under,under which conditions should one technique be preferred over another?,32001
images does information results? the image more or prior adding registration improve,does adding more images or prior image information improve the registration results?,32002
image. construct for of Scene an graphs a capturing provide these aspects formal,Scene graphs provide a formal construct for capturing these aspects of an image.,32003
"is complete the from Then, scene subgraphs. constructed graph these","Then, the complete scene graph is constructed from these subgraphs.",32004
"localization i.e. in address problem We cooperative the distributed networks, wireless of","We address the problem of distributed cooperative localization in wireless networks, i.e.",32005
without own determine to knowledge their position (agents) positions. wish nodes prior,nodes without prior position knowledge (agents) wish to determine their own positions.,32006
"as cooperative considered other is agents well. in from However, positioning, information","However, in cooperative positioning, information from other agents is considered as well.",32007
encoding positions. uncertainty Cooperative the agents' of requires positioning of,Cooperative positioning requires encoding of the uncertainty of agents' positions.,32008
agents. impose possible to the of We polygons on use positions convex constraints,We use convex polygons to impose constraints on the possible positions of agents.,32009
object based tracking. correlation considerable have trackers success in visual shown (DCF) Discrete filter,Discrete correlation filter (DCF) based trackers have shown considerable success in visual object tracking.,32010
"an proposed unsupervised which need transfer typography pairing. doesn't this In method paper, we","In this paper, we proposed an unsupervised typography transfer method which doesn't need pairing.",32011
per-agent method detailed modeling Lagrangian enables the behavior of in Our a formulation.,Our method enables the detailed modeling of per-agent behavior in a Lagrangian formulation.,32012
dense model and short-range simulate avoidance We collision crowds. long-range to sparse both and,We model short-range and long-range collision avoidance to simulate both sparse and dense crowds.,32013
use in suitable Our games. new for interactive is method,Our new method is suitable for use in interactive games.,32014
studies relaxed paper with a This coding hierarchical layered framework a structure.,This paper studies a layered coding framework with a relaxed hierarchical structure.,32015
"inherent its coding Alternatively, due loss scalable structure. to has conventional","Alternatively, conventional scalable coding has inherent loss due to its structure.",32016
between extracts penalty. common performance quality the levels It information with negligible different,It extracts the common information between different quality levels with negligible performance penalty.,32017
"results confirm work. the important for Simulation the superiority of practically sources,","Simulation results for practically important sources, confirm the superiority of the work.",32018
"global However, is more context using overfitting. prone to","However, using more global context is prone to overfitting.",32019
"forgeries cloning, and a image methods. combination other removal involve splicing, of resampling, region Realistic","Realistic image forgeries involve a combination of splicing, resampling, cloning, region removal and other methods.",32020
tracking (MTT). model robust issue multiple a target in is Designing affinity key the,Designing a robust affinity model is the key issue in multiple target tracking (MTT).,32021
"chloroplasts. has cluster Surprisingly, been found, for eight structure","Surprisingly, eight cluster structure has been found, for chloroplasts.",32022
"schedules. and work, often habit, to other creatures Consumers tied of shopping are periodic,","Consumers are creatures of habit, often periodic, tied to work, shopping and other schedules.",32023
documents. line method for text two-stage This a detection work presents historical,This work presents a two-stage text line detection method for historical documents.,32024
line is baseline. Each represented by text its detected,Each detected text line is represented by its baseline.,32025
line. class The separator end beginning each text of and marks,The separator class marks beginning and end of each text line.,32026
This augmentation data strategies. utilizing is by achieved,This is achieved by utilizing data augmentation strategies.,32027
open and the ARU-Net source. framework is train The to run,The framework to train and run the ARU-Net is open source.,32028
"However, to annotate daunting ImageNet-like task laboriously dateset. a is an it","However, it is a daunting task to laboriously annotate an ImageNet-like dateset.",32029
Both of the some points. errors strategies inevitably for data incur,Both strategies inevitably incur errors for some of the data points.,32030
"optimum. a local After mislead network a neural wrong labels to all, could bad the","After all, the wrong labels could mislead a neural network to a bad local optimum.",32031
We labels. two-stage from the framework suggest learning a noisy for,We suggest a two-stage framework for the learning from noisy labels.,32032
other the noisy images The are ignored. labels of,The noisy labels of the other images are ignored.,32033
"train we a network a in semi-supervised the second deep stage, manner. In neural","In the second stage, we train a deep neural network in a semi-supervised manner.",32034
from We data the learning consider of then incomplete sets. structure problem,We then consider the problem of structure learning from incomplete data sets.,32035
"by which This computationally structural EM, demanding. be however can very is addressed","This can be addressed by structural EM, which however is computationally very demanding.",32036
"relationships network, a nodes coopetition. and between co-existence is competitive cooperative called different of Within","Within a network, co-existence of cooperative and competitive relationships between different nodes is called coopetition.",32037
proposal represent of The possible temporal tubes streams objects. the the all region,The temporal proposal streams represent all the possible region tubes of the objects.,32038
algorithms. comparing other co-segmentation framework proposed evaluate by video We it our to,We evaluate our proposed framework by comparing it to other video co-segmentation algorithms.,32039
performance achieves on state-of-the-art method datasets. benchmark improved Our,Our method achieves improved performance on state-of-the-art benchmark datasets.,32040
Supervised by cross-modal hashing information. incorporating achieved progress have considerable side semantic methods,Supervised cross-modal hashing methods have achieved considerable progress by incorporating semantic side information.,32041
decision as in MLP are in APLOCO. used also variables The the criteria included independent,The independent variables included in MLP are also used as the decision criteria in APLOCO.,32042
"the show method in I examples, outperforms simulation BLP. contraction In mapping the algorithm","In simulation examples, I show the method outperforms the contraction mapping algorithm in BLP.",32043
shares. find the models robust near-zero I characteristics also market pure in method with remains,I also find the method remains robust in pure characteristics models with near-zero market shares.,32044
"cost the in accuracy. usually the labeling expense at Nevertheless, of is model save","Nevertheless, the save in labeling expense is usually at the cost of model accuracy.",32045
We synthetic in that and it simulations. show benefits performance demonstrate its consistent is,We show that it is consistent and demonstrate its performance benefits in synthetic simulations.,32046
"use orientation ICP the seed data to we this and registration regularize technique. particular, In","In particular, we use this orientation data to seed and regularize the ICP registration technique.",32047
on is GPU. This filter efficiently the implemented,This filter is implemented efficiently on the GPU.,32048
essential for cognitive an radio. is spectrum technology Wideband (WSS) sensing,Wideband spectrum sensing (WSS) is an essential technology for cognitive radio.,32049
"However, a of rate WSS. still sampling bottleneck the is","However, the sampling rate is still a bottleneck of WSS.",32050
proposed. Several have methods sub-Nyquist sensing been,Several sub-Nyquist sensing methods have been proposed.,32051
"this sampling. method paper, In we a on WSS novel sub-Nyquist (MC) Multi-coset based propose","In this paper, we propose a novel sub-Nyquist WSS method based on Multi-coset (MC) sampling.",32052
SNR-robust detection and design We multiband simple algorithm. signal a low-complexity,We design a simple SNR-robust and low-complexity multiband signal detection algorithm.,32053
bandwidths is related subsignals. the the relations of of scalar the to The,The scalar of the relations is related to the bandwidths of the subsignals.,32054
outperforms in ED results detector the low show regime sub-Nyquist the Experimental in especially SNRs.,Experimental results show the detector outperforms ED in the sub-Nyquist regime especially in low SNRs.,32055
"cost loops. domain, are interactions each the forming by entropy In attractive of balanced","In each domain, attractive interactions are balanced by the entropy cost of forming loops.",32056
using on-device images. for entire We an urban system design large-scale the present of localization,We present the design of an entire on-device system for large-scale urban localization using images.,32057
connection. agnostic require does network GPS and is Our not design,Our design is GPS agnostic and does not require network connection.,32058
"a pose estimation. we one-many In addition, RANSAC accurate propose for new","In addition, we propose a new one-many RANSAC for accurate pose estimation.",32059
RANSAC structures addresses (e.g. the new challenge The of building one-many repetitive,The new one-many RANSAC addresses the challenge of repetitive building structures (e.g.,32060
visual perceiving system while attention pays to The an image. salient regions human,The human visual system pays attention to salient regions while perceiving an image.,32061
We of reference the utility SED stereoscopic (FRSIQA). assessment image demonstrate full quality in,We demonstrate the utility of SED in full reference stereoscopic image quality assessment (FRSIQA).,32062
maps We consider inter-gradient magnitude predicting gradient and structural for similarity.,We consider gradient magnitude and inter-gradient maps for predicting structural similarity.,32063
maps using refine We depth quality. for SED quality evaluating this,We refine this quality using SED maps for evaluating depth quality.,32064
"quality. image overall Finally, we stereo depth this combine and obtain to an quality luminance","Finally, we combine this luminance and depth quality to obtain an overall stereo image quality.",32065
field Neural achieved object of have (R-CNNs) in detection. Networks the Region-based success Convolutional great,Region-based Convolutional Neural Networks (R-CNNs) have achieved great success in the field of object detection.,32066
the newly We feature designed into integrate the architecture. encoder R-CNN Faster,We integrate the newly designed feature encoder into the Faster R-CNN architecture.,32067
Validation conducted and MS experiments COCO VOC PASCAL are datasets. on both,Validation experiments are conducted on both PASCAL VOC and MS COCO datasets.,32068
fictional Westeros longitudinal kingdom Setting: Design cohort study of the in and A Essos. and,Design and Setting: A longitudinal cohort study in the fictional kingdom of Westeros and Essos.,32069
rate characters Conclusions: high in mortality is a among the GoT. There,Conclusions: There is a high mortality rate among the characters in GoT.,32070
than house trended Allegiance Stark to house to Targaryen. be safer,Allegiance to house Stark trended to be safer than house Targaryen.,32071
screening safe for We metric study learning.,We study safe screening for metric learning.,32072
"is However, number possible the even triplets dataset. of quite a huge for small","However, the number of possible triplets is quite huge even for a small dataset.",32073
"of screening relationships. variants and their several derive We analyze rules,","We derive several variants of screening rules, and analyze their relationships.",32074
benchmark effectiveness demonstrate triplet of screening. Numerical the on safe datasets experiments,Numerical experiments on benchmark datasets demonstrate the effectiveness of safe triplet screening.,32075
also and It that joint denoising is CNN model demosaicing. the demonstrated can perform,It is also demonstrated that the CNN model can perform joint denoising and demosaicing.,32076
artifacts methods. Visual visual than confirm the reduce methods comparisons existing more that proposed,Visual comparisons confirm that the proposed methods reduce more visual artifacts than existing methods.,32077
"has forecasting the to occurrence methods it events. expanded Lately, of for","Lately, it has expanded to methods for forecasting the occurrence of events.",32078
"rigorous in theory sensing. Thereby, the advance of eigenvalue-based an makes letter spectrum the","Thereby, the letter makes an advance in the rigorous theory of eigenvalue-based spectrum sensing.",32079
ideas Hamiltonian is framework symplectic systems on based and from Our dynamical integration.,Our framework is based on ideas from Hamiltonian dynamical systems and symplectic integration.,32080
make to biopsy when to bronchoscopy. decision needs a Doctors timely patients performing whether the,Doctors needs to make a decision whether to biopsy the patients timely when performing bronchoscopy.,32081
is based system transfer developed learning. The on,The system is developed based on transfer learning.,32082
We a transfer novel fine-tuning method: sentential propose learning .,We propose a novel transfer learning method: sentential fine-tuning .,32083
"methods methods, our fine-tuning Compared performance. traditional best to achieves the","Compared to traditional fine-tuning methods, our methods achieves the best performance.",32084
stationary for always outgoing the plan edge. vertex chooses every same A,A stationary plan for every vertex always chooses the same outgoing edge.,32085
"fixed stationary or sufficient not plans stopping-time fixed for distribution, For are optimality. horizon","For fixed horizon or fixed stopping-time distribution, stationary plans are not sufficient for optimality.",32086
applications in satellite-denied geomatics (e.g. many information valuable is Positioning,Positioning information is valuable in many satellite-denied geomatics applications (e.g.,32087
method blob-like from is segmenting structures capable images. noisy tubular Our of and,Our method is capable of segmenting tubular and blob-like structures from noisy images.,32088
"some directions methods are current Fourth, and highlighted. future potential deficiencies of several","Fourth, some potential deficiencies of current methods and several future directions are highlighted.",32089
methods they solutions. and optimize The criterion differ the constraints placed the on the on,The methods differ on the criterion they optimize and the constraints placed on the solutions.,32090
correlation further note solution constraints. inter-set without this This (ISC) maximizes that shows,This note shows that this solution maximizes inter-set correlation (ISC) without further constraints.,32091
note to this. aims This remedy short,This short note aims to remedy this.,32092
"and architectures, state-of-the-art created for ensembles CNN DenseNet. ResNet experiments using our We two","We created ensembles for our experiments using two state-of-the-art CNN architectures, ResNet and DenseNet.",32093
"our reliability. and has that results capacity, show safety Experimental method high","Experimental results show that our method has high capacity, safety and reliability.",32094
is photoacoustic algorithm in the (PA) image Delay-and-Sum (DAS) common formation. most used,Delay-and-Sum (DAS) is the most common algorithm used in photoacoustic (PA) image formation.,32095
aggressive key to idea an framework. is Our MobileNet applying strategy downsampling,Our key idea is applying an aggressive downsampling strategy to MobileNet framework.,32096
the reduces (i) three remarkably design computational brings This advantages: It cost.,This design brings three advantages: (i) It remarkably reduces the computational cost.,32097
significant performance the capacity information and increases achieves (ii) It improvements.,(ii) It increases the information capacity and achieves significant performance improvements.,32098
fast speed. It is (iii) actual and inference engineering-friendly provides,(iii) It is engineering-friendly and provides fast actual inference speed.,32099
the We formulation RL. present for first model DR a in learning,We first present a formulation for learning the DR model in RL.,32100
and optimal. estimators that strongly are MRDR prove consistent the asymptotically We,We prove that the MRDR estimators are strongly consistent and asymptotically optimal.,32101
the compared with and proposed method show our also experimentally of GPC GPC+. We usefulness,We also experimentally show the usefulness of our proposed method compared with GPC and GPC+.,32102
both is expensive labor-intensive. road networks and Mapping currently,Mapping road networks is currently both expensive and labor-intensive.,32103
imagery High-resolution a to network. a infer aerial automatically promising provides avenue road,High-resolution aerial imagery provides a promising avenue to automatically infer a road network.,32104
"involves communicating mobile termed collaborative This data the the feature and between intelligence, cloud. paradigm,","This paradigm, termed collaborative intelligence, involves communicating feature data between the mobile and the cloud.",32105
for We improving a strategy feature also propose compression. lossy the accuracy under,We also propose a strategy for improving the accuracy under lossy feature compression.,32106
a the from novel problem to of sampler sampling truncated this Gaussian We distribution. apply,We apply this novel sampler to the problem of sampling from a truncated Gaussian distribution.,32107
"can well processed be equally that way. images not all However,","However, not all images can be equally well processed that way.",32108
and of crowdsourcing varying typically in exhibit degrees The questions task difficulty subjectivity. a,The questions in a crowdsourcing task typically exhibit varying degrees of difficulty and subjectivity.,32109
"of we Moreover, a a develop of subjectivity measure quantitative the question.","Moreover, we develop a quantitative measure of the subjectivity of a question.",32110
by removing difficult questions highly inappropriately questions. subjective or,by removing highly subjective questions or inappropriately difficult questions.,32111
train CNN Such data necessary to state-of-the-art is architectures.,Such data is necessary to train state-of-the-art CNN architectures.,32112
Such training is viable an ours. only sets such as with large approach,Such an approach is viable only with large training sets such as ours.,32113
real to images on demonstrates that images. CNNs trained This generalize artificial well,This demonstrates that CNNs trained on artificial images generalize well to real images.,32114
"experiments and Our deblurring heavy on results under images show noise. large blurs, even promising","Our experiments show promising deblurring results on images even under large blurs, and heavy noise.",32115
"called an image term, quantile universal regularization and sparse structure-preserving introduces (QuaSI) paper This prior.","This paper introduces an universal and structure-preserving regularization term, called quantile sparse image (QuaSI) prior.",32116
various modalities. prior for from denoising images suitable The imaging medical is,The prior is suitable for denoising images from various medical imaging modalities.,32117
but speckle OCT retina is the human inherently scans noise. impaired by offers of high-resolution,OCT offers high-resolution scans of the human retina but is inherently impaired by speckle noise.,32118
on a has high-frequency other lower and the CT noise. shows resolution hand,CT on the other hand has a lower resolution and shows high-frequency noise.,32119
Experiments the multiple emphasize of on excellent method. datasets the performance proposed,Experiments on multiple datasets emphasize the excellent performance of the proposed method.,32120
"we short-time LCT short-time the Moreover, the generalize OLCT. to","Moreover, we generalize the short-time LCT to the short-time OLCT.",32121
key lightweight of component The dependency the is local module. RSNet a,The key component of the RSNet is a lightweight local dependency module.,32122
"all RSNets benchmarks. these its methods on surpass In simplest state-of-the-art previous form,","In its simplest form, RSNets surpass all previous state-of-the-art methods on these benchmarks.",32123
method a problems. classification novel This paper solving proposes for one-class,This paper proposes a novel method for solving one-class classification problems.,32124
"optimal enclosing hypersphere is space, feature the class In that the determined. target then","In that feature space, the optimal hypersphere enclosing the target class is then determined.",32125
non-linear mappings provide for linear the We and both method. proposed,We provide both linear and non-linear mappings for the proposed method.,32126
and of distinct multi-dimensional degree complexity implications. technological technologies have business Both,Both technologies have distinct degree of technological complexity and multi-dimensional business implications.,32127
"loopholes of to unfortunately, It free less likely thus, is and flaws. be","It is thus, unfortunately, less likely to be free of loopholes and flaws.",32128
alliance possibilities. AI of and numerous blockchain create to is The expected,The alliance of AI and blockchain is expected to create numerous possibilities.,32129
of single constraints method. thermal These major limitations image the traditional the are,These limitations are the major constraints of the traditional single thermal image method.,32130
evaluated Signal-to-Noise result Ratio. The was by quantitatively using,The result was quantitatively evaluated by using Signal-to-Noise Ratio.,32131
using compared the to observed was performance image methods Favorable time-series single approach.,Favorable performance was observed using time-series methods compared to the single image approach.,32132
The unearths underlying of structure also the about the information chain physical process.,The structure of the chain also unearths information about the underlying physical process.,32133
"which paraphraser and by is convolutional two translator. done are called a a modules, This","This is done by two convolutional modules, which are called a paraphraser and a translator.",32134
create greedily the on Current ThreadTone the draw to circle. approaches paintings chords,Current approaches to create ThreadTone paintings greedily draw the chords on the circle.,32135
"Our pictures. high produce on approach cartoon well sketches results portraits, as as can quality","Our approach can produce high quality results on portraits, sketches as well as cartoon pictures.",32136
"collection points frame and represented RGB-D which is a are features, as of planes. An","An RGB-D frame is represented as a collection of features, which are points and planes.",32137
objects in show shapes. from results qualitative scenes indoor various We with,We show qualitative results from indoor scenes with objects in various shapes.,32138
for object robotic used to be on-the-fly technique generation manipulation. The enables model,The technique enables on-the-fly object model generation to be used for robotic manipulation.,32139
the of superparaboloids. we family with Superquadrics to deal superellipsoids refer and a shapes: here,Superquadrics refer to a family of shapes: here we deal with the superellipsoids and superparaboloids.,32140
"each for sampled are point. provide to able normals Additionally, we","Additionally, we are able to provide normals for each sampled point.",32141
issues among the one major Osteoarthritis population. of the health is elderly (OA),Osteoarthritis (OA) is one of the major health issues among the elderly population.,32142
and technology OA of the popular progress most course. to MRI the is observe evaluate,MRI is the most popular technology to observe and evaluate the progress of OA course.,32143
"the inefficient MRI the extreme makes labor analysis However, process of cost expensive. and","However, the extreme labor cost of MRI analysis makes the process inefficient and expensive.",32144
"methods have extensively. researchers In MRI knee segmentation the decades, past investigated automatic/semi-automatic","In the past decades, researchers have investigated automatic/semi-automatic knee MRI segmentation methods extensively.",32145
information Neurons transforming process activity. barrages of into synaptic by inputs spiking,Neurons process information by transforming barrages of synaptic inputs into spiking activity.,32146
novel (IA). We Potassium the identify a current for A-type role,We identify a novel role for the A-type Potassium current (IA).,32147
"of In and paper, this we recent modeling advances review in facial kinship. aging","In this paper, we review recent advances in modeling of facial aging and kinship.",32148
"propagation, largely has overlooked. EP) single-sweep schemes expectation been","single-sweep expectation propagation, EP) schemes has been largely overlooked.",32149
inference space the methodology methods. $\mathcal{O}(n)$ present state efficient with existing We means of combining,We present means of combining the efficient $\mathcal{O}(n)$ state space methodology with existing inference methods.,32150
"provide approaches. extend existing implementing and all code methods, We unifying","We extend existing methods, and provide unifying code implementing all approaches.",32151
services. capabilities these with IoT processing to embedded objects fulfill be to need,IoT objects need to be embedded with processing capabilities to fulfill these services.,32152
complex critical Modeling conditional distributions of variety a in is settings.,Modeling complex conditional distributions is critical in a variety of settings.,32153
"estimators and complexity trade-off between overfitting. must modeling functional distributional without heteroscedasticity These complexity,","These estimators must trade-off between modeling distributional complexity, functional complexity and heteroscedasticity without overfitting.",32154
way provides traveling booming and environment-friendly an all Bike sharing for world. the over is,Bike sharing provides an environment-friendly way for traveling and is booming all over the world.,32155
well applied also when It to areas. transfers unseen,It also transfers well when applied to unseen areas.,32156
determinant classic of thus for the reproduced. the identity is The gradient,The classic identity for the gradient of the determinant is thus reproduced.,32157
does Random the switch. and away operates probabilistically this with Switching (RS) notion,Random Switching (RS) does away with this notion and probabilistically operates the switch.,32158
affects the The probability DC amplitude pulse total PSD. and,The pulse amplitude probability affects the DC and total PSD.,32159
acts harmonic minimum length filter. pulse like The selective a,The minimum pulse length acts like a selective harmonic filter.,32160
shaped The filtering without by be can probabilities. external PSD changing these therefore,The PSD can therefore be shaped without external filtering by changing these probabilities.,32161
good several consider reasons for are high There RS performance applications. to,There are several good reasons to consider RS for high performance applications.,32162
based in highly been sign to time. influential have epistasis on Approaches gene recent interactions,Approaches to gene interactions based on sign epistasis have been highly influential in recent time.,32163
propose combines We both a that new aspects framework of approaches.,We propose a new framework that combines aspects of both approaches.,32164
numerical results. experiments provide to support the theoretical We,We provide numerical experiments to support the theoretical results.,32165
"the challenge, method the proposed in suggesting achieved were is the the highest state-of-the-art. These","These were the highest achieved in the challenge, suggesting the proposed method is the state-of-the-art.",32166
of the scanners system protocols also investigated. and adaptability different to The is,The adaptability of the system to different scanners and protocols is also investigated.,32167
presented test effect ensemble to is further of study size. A quantitative the,A quantitative study is further presented to test the effect of ensemble size.,32168
"software made available. publicly and are Additionally, method of models our","Additionally, software and models of our method are made publicly available.",32169
Wood-composite materials related today humidity they as deformations. used are homogenize directional widely,Wood-composite materials are widely used today as they homogenize humidity related directional deformations.,32170
process. that arise typically conditioning in the,that typically arise in the conditioning process.,32171
"In risk of the bound. likelihood associated explore we the maximum framework stochastic estimation,","In the framework of stochastic maximum likelihood estimation, we explore the associated risk bound.",32172
component algorithm principal In new streaming for paper this analysis. propose a we,In this paper we propose a new algorithm for streaming principal component analysis.,32173
"in limited all memory, high-dimensional samples store the cannot the regime. devices With small","With limited memory, small devices cannot store all the samples in the high-dimensional regime.",32174
useful. contained could the data previously streamed be information However in,However the information contained in previously streamed data could be useful.,32175
"demonstrated Cycle-GAN Cycle-Consistency Adversarial Generative state-of-the-art, Networks of the power with Loss. The","The state-of-the-art, Cycle-GAN demonstrated the power of Generative Adversarial Networks with Cycle-Consistency Loss.",32176
"are is training the promising, there its for process. scope in results optimization While","While its results are promising, there is scope for optimization in the training process.",32177
single-perspective into can of perceptually image be stitching multiple-perspective. divided and Results,Results of image stitching can be perceptually divided into single-perspective and multiple-perspective.,32178
"paper, natural this image In we propose stitching. warps for two single-perspective","In this paper, we propose two single-perspective warps for natural image stitching.",32179
intelligence tasks find artificial dynamics. knowledge about to require to models in able be Several,Several tasks in artificial intelligence require to be able to find models about knowledge dynamics.,32180
"include They belief abduction. fusion belief revision, and and merging,","They include belief revision, fusion and belief merging, and abduction.",32181
open paper This generating framework adventure from games a for presents data.,This paper presents a framework for generating adventure games from open data.,32182
"has feedforward, connections. recurrent feedback, and It","It has feedforward, feedback, and recurrent connections.",32183
"multiple updating, used recursive image After classification. of the cycles representation is for","After multiple cycles of recursive updating, the representation is used for image classification.",32184
over computation performance given improve time. Its to more cycles of tended,Its performance tended to improve given more cycles of computation over time.,32185
"reuses run PCN and recursively a architecture bottom-up In processes. top-down to short, single","In short, PCN reuses a single architecture to recursively run bottom-up and top-down processes.",32186
artificial is important mid-level in arguably intelligence. of behavior one the components understanding Human most,Human behavior understanding is arguably one of the most important mid-level components in artificial intelligence.,32187
"specific multi-task learning datasets and cumbersome. can such be relies on datasets However, task constructing","However, multi-task learning relies on task specific datasets and constructing such datasets can be cumbersome.",32188
"consideration labeling efforts, It huge requires amounts of etc. data, statistical","It requires huge amounts of data, labeling efforts, statistical consideration etc.",32189
Song the of recent built al. work upon We et,We built upon the recent work of Song et al.,32190
"the channels SSCnet. investigate the RGB improve this colour work, In we potential of to","In this work, we investigate the potential of the RGB colour channels to improve SSCnet.",32191
"tasks, performance. which generalization prediction multi-task related We simultaneously learning, learns improve to consider","We consider multi-task learning, which simultaneously learns related prediction tasks, to improve generalization performance.",32192
"bound Moreover, method. the provide of proposed the performance we","Moreover, we provide the performance bound of the proposed method.",32193
real-world proposed validated and both The of for datasets. effectiveness the method synthetic is,The effectiveness of the proposed method is validated for both synthetic and real-world datasets.,32194
saliency optimal level is work. highlighted OPCs (OSLO) An for this in,An optimal saliency level for OPCs (OSLO) is highlighted in this work.,32195
"effective Here, internal watershed performed is markers. with efficiently segmentation","Here, watershed segmentation is performed efficiently with effective internal markers.",32196
existing of method show Experiments accuracy. that methods terms our in outperforms,Experiments show that our method outperforms existing methods in terms of accuracy.,32197
"in we paper, Microsoft visual deployed In this a search web-scale general Bing. system introduce","In this paper, we introduce a web-scale general visual search system deployed in Microsoft Bing.",32198
is itself. due not the apparently the machinery to CNN The learning failure computational method,The failure is apparently due to the learning method not the CNN computational machinery itself.,32199
imaging satellite the high-resolution Current of technology shooting ground. enables pictures,Current satellite imaging technology enables shooting high-resolution pictures of the ground.,32200
"images, pictures easily overhead forged. digital be any other kind can As also of","As any other kind of digital images, overhead pictures can also be easily forged.",32201
"and paper, detection for image localization. an forgery we satellite algorithm this In propose","In this paper, we propose an algorithm for satellite image forgery detection and localization.",32202
algorithm training. available works are forged the for under no that assumption Our images,Our algorithm works under the assumption that no forged images are available for training.,32203
"satellite adversarial of a network images. feature we a representation generative learn Using pristine (GAN),","Using a generative adversarial network (GAN), we learn a feature representation of pristine satellite images.",32204
machine one-class is to features vector A on support their determine trained distribution. these (SVM),A one-class support vector machine (SVM) is trained on these features to determine their distribution.,32205
"forgeries Finally, are anomalies. detected image as","Finally, image forgeries are detected as anomalies.",32206
Bees may memories acquire throughout their lives. unique,Bees may acquire unique memories throughout their lives.,32207
social time even These interactions over experiences affect large frames.,These experiences affect social interactions even over large time frames.,32208
approach similar in this recent that show literature. systems proposed outperforms We,We show that this approach outperforms similar systems proposed in recent literature.,32209
"moieties the steady explain model states, drift from generalising non-unique previous results and studies. Conserved","Conserved moieties explain model drift and non-unique steady states, generalising the results from previous studies.",32210
real has performance environment. physical in agent demonstrate zero-shot this that a We good,We demonstrate that this agent has good zero-shot performance in a real physical environment.,32211
"robust propose paper, method surveillance. we for visual In intelligent this change detection a","In this paper, we propose a robust change detection method for intelligent visual surveillance.",32212
"learning three independent, features are feasible. conditionally making approximately multi-source The","The three features are approximately conditionally independent, making multi-source learning feasible.",32213
Bayes are with foreground rule. then estimated posteriors Pixel-wise,Pixel-wise foreground posteriors are then estimated with Bayes rule.,32214
shows with of existing approaches the comparison proposed scheme. the effectiveness A,A comparison with existing approaches shows the effectiveness of the proposed scheme.,32215
the a with conclude approach. of critique We,We conclude with a critique of the approach.,32216
real compared validated data method with state-of-the-art is The and proposed and methods. on synthetic,The proposed method is validated on synthetic and real data and compared with state-of-the-art methods.,32217
by due tree forest surrounding and propagation patches shrink grow grasslands. to fires Thus in,Thus forest patches grow due to tree propagation and shrink by fires in surrounding grasslands.,32218
We tropical by forest-grassland scaling model of introducing ecotones. explain this a fire propagation heterogeneous,We explain this scaling by introducing a heterogeneous fire propagation model of tropical forest-grassland ecotones.,32219
determine Deviations patches. stability individual relation the from perimeter-area forest this of,Deviations from this perimeter-area relation determine the stability of individual forest patches.,32220
of of in source tensor challenges variability one the separation. Modeling is methods decomposition,Modeling variability in tensor decomposition methods is one of the challenges of source separation.,32221
"However, has been on with the imposing so mode possible. variability constraints far not","However, so far imposing constraints on the mode with variability has not been possible.",32222
"and cope lane can ego-lanes, not with changes.","ego-lanes, and can not cope with lane changes.",32223
achieve and competitive tuSimple dataset the verify We method our results. on,We verify our method on the tuSimple dataset and achieve competitive results.,32224
"In the we uncertainty in consider locations to them. distribution particular, assigning prior by a","In particular, we consider uncertainty in the locations by assigning a prior distribution to them.",32225
dataset. rainfall application are a illustrated methods with an to The,The methods are illustrated with an application to a rainfall dataset.,32226
"volumes with to scan acquisition leads times. resolution However, long of high the spatial MR","However, the acquisition of MR volumes with high spatial resolution leads to long scan times.",32227
"super-resolution this retrospectively. MRI forests To (VSRF) volumetric end, propose resolution to we enhance","To this end, we propose volumetric super-resolution forests (VSRF) to enhance MRI resolution retrospectively.",32228
and evidences soft virtual considered. are Both,Both soft and virtual evidences are considered.,32229
"ATS is by drug on notably campaign most abuse fought drugs. against countries, all The","The campaign against drug abuse is fought by all countries, most notably on ATS drugs.",32230
"Granger causal these models to estimation may In of interactions. cases, using lead inconsistent linear","In these cases, using linear models may lead to inconsistent estimation of Granger causal interactions.",32231
capture interactions illustrate in nonlinear dataset. detecting motion methods in human our a likewise We,We likewise illustrate our methods in detecting nonlinear interactions in a human motion capture dataset.,32232
"we new of scenes. as In paper, way a representing scenarios introduce this","In this paper, we introduce scenarios as a new way of representing scenes.",32233
effectiveness experimentally. We propose its heuristics for and tasks demonstrate efficient computation classification its in,We propose efficient heuristics for its computation and demonstrate its effectiveness in classification tasks experimentally.,32234
Recent deep learning increased in advances systems. performance significantly of recognition the face have,Recent advances in deep learning have significantly increased the performance of face recognition systems.,32235
"experiments system benefits face our synthetic recognition Thus, any in using images. face from","Thus, any face recognition system in our experiments benefits from using synthetic face images.",32236
"The well as data synthetic experiments, available. all generator, are as publicly","The synthetic data generator, as well as all experiments, are publicly available.",32237
Such to supervised setting a poorly. methods causes perform,Such a setting causes supervised methods to perform poorly.,32238
detecting image these a an We face for meets requirements. anomalous describe method that,We describe a method for detecting an anomalous face image that meets these requirements.,32239
autoencoders. by 'peeking' in Obvious residuals) (autoencoder behavior constructions defeated codes; are autoencoder a,Obvious constructions (autoencoder codes; autoencoder residuals) are defeated by a 'peeking' behavior in autoencoders.,32240
sequence been Image has or generation as successfully generation an cast transformation problem. autoregressive,Image generation has been successfully cast as an autoregressive sequence generation or transformation problem.,32241
self-attention is Recent work sequences. textual has shown effective an modeling that of way,Recent work has shown that self-attention is an effective way of modeling textual sequences.,32242
"network of image EPVS. input, quantification outputs score as and The takes a an","The network takes an image as input, and outputs a quantification score of EPVS.",32243
"EPVS Furthermore, correlates scoring. automatic to the as age visual similarly scoring","Furthermore, the automatic EPVS scoring correlates similarly to age as visual scoring.",32244
and phylogenetic family-tree China. provides model Our historical a context against study detailed in,Our study provides a detailed phylogenetic and historical context against family-tree model in China.,32245
of Our work door guided opens to attentional the normatively allocation. studies,Our work opens the door to normatively guided studies of attentional allocation.,32246
build whispered This and visual recognition to systems silent for introduces speech. need the speech,This introduces the need to build visual speech recognition systems for silent and whispered speech.,32247
"almost proposed only. have the However, data been on trained recently systems vocalised all","However, almost all the recently proposed systems have been trained on vocalised data only.",32248
suggested. pointed algorithmic are flaws and are out Several remedies,Several algorithmic flaws are pointed out and remedies are suggested.,32249
constraints instances results with multi-vehicle on also are new Experimental distance reported.,Experimental results on new multi-vehicle instances with distance constraints are also reported.,32250
"overhead. the protocol optimize it is the beam-alignment Therefore, to to signaling minimize imperative","Therefore, it is imperative to optimize the beam-alignment protocol to minimize the signaling overhead.",32251
It a is search algorithm is bisection proved optimal. that,It is proved that a bisection search algorithm is optimal.,32252
on superpixel step. image pattern recognition rely tasks generation a Several as fundamental,Several image pattern recognition tasks rely on superpixel generation as a fundamental step.,32253
a is very All computational at low this achieved cost.,All this is achieved at a very low computational cost.,32254
(GGP) a learning. reinforcement testbed General for Game Playing provides good,General Game Playing (GGP) provides a good testbed for reinforcement learning.,32255
"a specification rules GGP, given. is of games In","In GGP, a specification of games rules is given.",32256
solved can by be problems learning. GGP reinforcement,GGP problems can be solved by reinforcement learning.,32257
"converges, As although Q-learning MCTS. much than expected, slower","As expected, Q-learning converges, although much slower than MCTS.",32258
"idea with MCTS, Q-learning Borrowing Carlo an give Monte we from Search, enhance to QM-learning.","Borrowing an idea from MCTS, we enhance Q-learning with Monte Carlo Search, to give QM-learning.",32259
improves This Q-learning. enhancement performance the of pure,This enhancement improves the performance of pure Q-learning.,32260
"in expanding population qualitative frequently preventing heterogeneity Attempts effectively dimensionality, analysis. include to system result","Attempts to include population heterogeneity frequently result in expanding system dimensionality, effectively preventing qualitative analysis.",32261
"competitive recent outperforms Bayesian baselines statistical with It displays significance, with approaches. strong performance and","It outperforms baselines with strong statistical significance, and displays competitive performance with recent Bayesian approaches.",32262
"general, in promotes find traits. We nature of that the coexistent environment","We find that in general, nature promotes the environment of coexistent traits.",32263
research results environmental and future for These monitoring. suggest microbial directions,These results suggest directions for future microbial research and environmental monitoring.,32264
word for images. for efficient holistic a learning representation present an handwritten framework We,We present a framework for learning an efficient holistic representation for handwritten word images.,32265
The traditional network with a neural deep method classification loss. convolutional uses proposed,The proposed method uses a deep convolutional neural network with traditional classification loss.,32266
of biochemical consisting circuits a interactions capable ring of sustained of Ring are oscillations. oscillators,Ring oscillators are biochemical circuits consisting of a ring of interactions capable of sustained oscillations.,32267
"of genes. this or of holds results Strikingly, independently used number the regulatory functions","Strikingly, this results holds independently of the regulatory functions used or number of genes.",32268
were Functional reconstructed bands. source networks computed in EEG using brain different connectivity frequency,Functional brain networks were reconstructed using EEG source connectivity computed in different frequency bands.,32269
theoretical performed assess between analyses groups. both differences Graph were,Graph theoretical analyses were performed assess differences between both groups.,32270
their optimizing This applications. practical training in limits and usage for real-world,This limits their practical usage in training and optimizing for real-world applications.,32271
"architectures, light-weight are this On such address the contrary, to proposed as issue. being SqueezeNet,","On the contrary, light-weight architectures, such as SqueezeNet, are being proposed to address this issue.",32272
These procedure. mostly from designing stem following an ad-hoc inefficiencies,These inefficiencies mostly stem from following an ad-hoc designing procedure.,32273
"we SimpNet}. principles, simple {\it propose on a such called architecture Based","Based on such principles, we propose a simple architecture called {\it SimpNet}.",32274
most all gene events This evolutionary of is comprehensive histories. analysis genome-wide in the,This is the most comprehensive genome-wide analysis of all events in gene evolutionary histories.,32275
"constraints learn the However, correspondences set-level such instance-level cannot (e.g.","However, such set-level constraints cannot learn the instance-level correspondences (e.g.",32276
parts task). in aligned semantic configuration object,aligned semantic parts in object configuration task).,32277
(e.g. in results often limitation positives This false,This limitation often results in false positives (e.g.,32278
"mode collapse to or semantic leads geometric further and problem. artifacts),","geometric or semantic artifacts), and further leads to mode collapse problem.",32279
"on be set-level the exploited both instance-level. and constraints could Therefore,","Therefore, the constraints could be exploited on both set-level and instance-level.",32280
The the the specific independent algorithm collect method used data. experimental is to of,The algorithm is independent of the specific experimental method used to collect the data.,32281
and model the inference. constraint through stochastic parameters variational distribution over We characterize then posterior,We then characterize the posterior distribution over model and constraint parameters through stochastic variational inference.,32282
and present approach a depth monocular learning from ego-motion of unsupervised We for video. novel,We present a novel approach for unsupervised learning of depth and ego-motion from monocular video.,32283
masks to also areas validity We in information incorporate avoid penalizing which exists. no useful,We also incorporate validity masks to avoid penalizing areas in which no useful information exists.,32284
AI systems rely plans produce of for may them upon achieving to objectives. Users desired,Users of AI systems may rely upon them to produce plans for achieving desired objectives.,32285
present results We problems. our analysis theoretical on of complexity the,We present theoretical results on the complexity analysis of our problems.,32286
Versatile for demanded mobile systems. has camera filtering fast but been,Versatile but fast filtering has been demanded for mobile camera systems.,32287
devices. allows to real-time multiscale applications filtering We an which present approach low-powered on,We present an approach to multiscale filtering which allows real-time applications on low-powered devices.,32288
We estimating the Hough a employs use in angle. technique which transform kernel,We use a technique which employs Hough transform in estimating the kernel angle.,32289
"transform is estimate accurately kernel the introduced. length, To using approach a novel the cepstral","To accurately estimate the kernel length, a novel approach using the cepstral transform is introduced.",32290
making paper introduced SHELP. system a The called decision computer-assisted medical,The paper introduced a computer-assisted medical decision making system called SHELP.,32291
"models very end-to-end research However, is limited. on audiovisual","However, research on end-to-end audiovisual models is very limited.",32292
"can which perform is counting. proposed, a application object-independent In vision product machine this paper,","In this paper, a machine vision application is proposed, which can perform object-independent product counting.",32293
in one system. used Basically the camera is,Basically one camera is used in the system.,32294
"be in moment of may practice. fourth-order always their not (kurtosis) satisfied However, assumption","However, their assumption of fourth-order moment (kurtosis) may not always be satisfied in practice.",32295
"construct to personas In intelligent we short, use artificially synthetic playtesters.","In short, we use artificially intelligent personas to construct synthetic playtesters.",32296
across accuracy restore algorithms which domain. whole Then provide we and alternate efficiency the,Then we provide alternate algorithms which restore accuracy and efficiency across the whole domain.,32297
to verify expressions. are MATLAB simulations derived provided,MATLAB simulations are provided to verify derived expressions.,32298
particularly suitable Both long-range are structures communications. for,Both structures are particularly suitable for long-range communications.,32299
"applications recent cognitive macroeconomics, matrices imaging, and Such vision, appear computer in in genetics. neuroscience,","Such matrices appear in recent applications in cognitive neuroscience, imaging, computer vision, macroeconomics, and genetics.",32300
"quite scenes. image under task complex especially vision challenging, this However, is","However, this vision task is quite challenging, especially under complex image scenes.",32301
refined improvements. performance effectively for by structural coarse The information these are results prediction,The coarse prediction results are effectively refined by these structural information for performance improvements.,32302
"we propose to address a issues. paper, method In these this","In this paper, we propose a method to address these issues.",32303
a stylization and consists method smoothing a proposed of step The step.,The proposed method consists of a stylization step and a smoothing step.,32304
efficiently. closed-form steps computed of Each and be can the a has solution,Each of the steps has a closed-form solution and can be computed efficiently.,32305
efficiency over approach standard offers substantial a implementation. the gains find resulting that in We,We find that the resulting approach offers substantial gains in efficiency over a standard implementation.,32306
present concept Douglas analysis a formal of Hofstadter's \emph{superrationality}. We of,We present a formal analysis of Douglas Hofstadter's concept of \emph{superrationality}.,32307
"symmetric by We study and superrationally defining in them games. justifiable start actions,","We start by defining superrationally justifiable actions, and study them in symmetric games.",32308
captured theory. epistemic formal beliefs game These from of \emph{type} in the are notion drawn,These beliefs are captured in the formal notion of \emph{type} drawn from epistemic game theory.,32309
superrational We that find conditions outcomes. guarantee,We find conditions that guarantee superrational outcomes.,32310
"functional regimes Finally, are networks. in different communication found different specialized","Finally, different communication regimes are found in different specialized functional networks.",32311
problem. from emotion recognition challenging Automated in the remains images facial a wild,Automated emotion recognition in the wild from facial images remains a challenging problem.,32312
This frequency response. narrowband a (FSS) a bandpass frequency with selective surface article presents,This article presents a bandpass frequency selective surface (FSS) with a narrowband frequency response.,32313
made is unit miniaturized FSS designed of cell. The elements,The designed FSS is made of miniaturized elements unit cell.,32314
new different synthesis from images. viewpoints images to of view aims synthesize given Novel,Novel view synthesis aims to synthesize new images from different viewpoints of given images.,32315
features trained are values. The labeled CNNs from inverse depth obtained with depth sparse,The inverse depth features are obtained from CNNs trained with sparse labeled depth values.,32316
viewpoints. framework multiple fuse from images can This easily different,This framework can easily fuse multiple images from different viewpoints.,32317
is dataset. Our evaluated KITTI method on the,Our method is evaluated on the KITTI dataset.,32318
of results method can distortion. generate scene show The our that novel without views large-scale,The results show that our method can generate novel views of large-scale scene without distortion.,32319
evaluation. our effectiveness and approach The demonstrated through qualitative quantitative is of,The effectiveness of our approach is demonstrated through qualitative and quantitative evaluation.,32320
using node localization paper measurements. considers in range-only sensor static networks This,This paper considers node localization in static sensor networks using range-only measurements.,32321
"Matlab's with comparisons and simulations, Using implementation. provide DILOC we MDS","Using simulations, we provide comparisons with DILOC and Matlab's MDS implementation.",32322
ordinary bands images. multispectral have spectral images than far more Hyperspectral,Hyperspectral images have far more spectral bands than ordinary multispectral images.,32323
Rich band favorable conditions more information the for tremendous applications. provides,Rich band information provides more favorable conditions for the tremendous applications.,32324
This development surveys of paper comprehensively clustering. the trajectory,This paper comprehensively surveys the development of trajectory clustering.,32325
"clustering can categories: supervised grouped three unsupervised, Existing methods and into be semi-supervised trajectory algorithms.","Existing trajectory clustering methods can be grouped into three categories: unsupervised, supervised and semi-supervised algorithms.",32326
introduces each constraint in sets iteration The minimization. which alternat- the ing vary for bi-convex,The bi-convex constraint introduces sets which vary for each iteration in the alternat- ing minimization.,32327
Their using is then dimension PCA. reduced,Their dimension is then reduced using PCA.,32328
time. and follows feature the specific over channel of series time a represents Each it,Each channel of the time series represents a specific feature and follows it over time.,32329
obtained main series of method time is proposed classify the the effectively. to The focus,The main focus of the proposed method is to classify the obtained time series effectively.,32330
"the learn this, the temporal let idea machine Towards is to features.","Towards this, the idea is to let the machine learn temporal features.",32331
"extended filter used. Kalman detection, been has harmonic For","For harmonic detection, extended Kalman filter has been used.",32332
the zone. uses non-detection decreasing proposed for strategy a The method,The proposed method uses a strategy for decreasing the non-detection zone.,32333
"compare limit algorithm, algorithm, central Box-Muller polarization decision three also and algorithms ofGRNgeneration: algorithm. We","We also compare three algorithms ofGRNgeneration: Box-Muller algorithm, polarization decision algorithm, and central limit algorithm.",32334
"or efficiency, environmental limitations However, solutions constraints. most practicability, have their regarding","However, most solutions have their limitations regarding efficiency, practicability, or environmental constraints.",32335
different do easy who identify. are activities to Persons,Persons who do different activities are easy to identify.,32336
"is are their approach. two results, Despite with known the problems there it impressive that","Despite their impressive results, it is known that there are two problems with the approach.",32337
"flow fails of small estimation fast. the situations in First, move where optical multi-resolution objects","First, the multi-resolution estimation of optical flow fails in situations where small objects move fast.",32338
"when warping or dis-occlusion Second, artifacts happens. creates occlusion","Second, warping creates artifacts when occlusion or dis-occlusion happens.",32339
"manageable break-up learning allows large It to problems subtasks. agents more smaller, into","It allows learning agents to break-up large problems into smaller, more manageable subtasks.",32340
"set A however, a reusable suitable question, major how of open is to skills. identify","A major open question, however, is how to identify a suitable set of reusable skills.",32341
"The desired overall the description captures subtasks skills. capture abstract while problem structure,","The abstract description captures the overall problem structure, while subtasks capture desired skills.",32342
optimize We of we demonstrate levels both that jointly can over learning.,We demonstrate that we can jointly optimize over both levels of learning.,32343
now. used has for for a while localization regression Heatmap quite landmark been,Heatmap regression has been used for landmark localization for quite a while now.,32344
segmentation. also the The image u-net lesion but provides classifies,The u-net classifies the image but also provides lesion segmentation.,32345
must between exploit models performance. predictive Multi-output maximise outputs to regression dependencies,Multi-output regression models must exploit dependencies between outputs to maximise predictive performance.,32346
"and discussed. their are tools, of functions advantages these and main features disadvantages The","The main functions and features of these tools, their advantages and disadvantages are discussed.",32347
comparative analysis ontologies for A means systematic is engineering of presented.,A systematic comparative analysis of means for engineering ontologies is presented.,32348
"Typically, on vehicles. mobile are mounted sensors","Typically, mobile sensors are mounted on vehicles.",32349
the object. distance target Each the from sensor measures to it continuously,Each sensor continuously measures the distance from it to the target object.,32350
circuit validated simulation The is through comparison with the the electromagnetic proposed model results.,The proposed circuit model is validated through the comparison with the electromagnetic simulation results.,32351
The a thirdorder filter prototype. design procedure by developing explained is,The design procedure is explained by developing a thirdorder filter prototype.,32352
"applications. one One become the learning most technique, these deep U-Net, of popular for has","One deep learning technique, U-Net, has become one of the most popular for these applications.",32353
"Residual the as of models utilize well proposed as power Network, RCNN. The U-Net,","The proposed models utilize the power of U-Net, Residual Network, as well as RCNN.",32354
segmentation are of tasks. proposed advantages architectures There for several these,There are several advantages of these proposed architectures for segmentation tasks.,32355
"a deep residual architecture. First, training helps when unit","First, a residual unit helps when training deep architecture.",32356
"residual accumulation feature convolutional better representation for recurrent feature Second, tasks. with segmentation ensures layers","Second, feature accumulation with recurrent residual convolutional layers ensures better feature representation for segmentation tasks.",32357
object for an paper feaTure This framework Aggregating Multi-Level Amulet) Agile proposes salient (Agile detection.,This paper proposes an Agile Aggregating Multi-Level feaTure framework (Agile Amulet) for salient object detection.,32358
maps Agile multi-level to previous builds convolutional features. saliency The works on Amulet using predict,The Agile Amulet builds on previous works to predict saliency maps using multi-level convolutional features.,32359
"convolutional novel propose aggregate to addition, method we features. multi-level In deep a","In addition, we propose a novel method to aggregate multi-level deep convolutional features.",32360
verified MATLAB expressions simulations. Derived through are,Derived expressions are verified through MATLAB simulations.,32361
"of structure maintaining consumption power significantly reduces the Therefore, system. while performance proposed","Therefore, proposed structure significantly reduces power consumption while maintaining performance of the system.",32362
way to such novice in doctors. developed help be It the also a can,It can also be developed in such a way to help the novice doctors.,32363
morphing the We kernel show spatial results. the that also improves,We also show that the spatial morphing kernel improves the results.,32364
velocity of features explorative monocular Another an is for contribution vehicle estimation. study,Another contribution is an explorative study of features for monocular vehicle velocity estimation.,32365
no directly derive assumption. requires from loss this axiom function distributional that We a,We derive a loss function directly from this axiom that requires no distributional assumption.,32366
"In to manifolds. a method this intersected paper, multiple cluster we propose","In this paper, we propose a method to cluster multiple intersected manifolds.",32367
resulting of as Isomap. a algorithm implemented is used algorithm in variation Dijkstras simple The,The resulting algorithm is implemented as a simple variation of Dijkstras algorithm used in Isomap.,32368
traffic of of estimation future demand. is the enablers of the key accurate ATFCM One,One of the key enablers of ATFCM is the accurate estimation of future traffic demand.,32369
"plans, available (schedules, etc.) information flight The","The available information (schedules, flight plans, etc.)",32370
"Hence AS and power consumption complexity, to is of latency recommended. more due AC,","Hence due to more complexity, power consumption and latency of AC, AS is recommended.",32371
effectiveness each the An of study method. ablation proposed confirms,An ablation study confirms the effectiveness of each proposed method.,32372
connected series thyristors. Many are realized voltage high applications with,Many high voltage applications are realized with series connected thyristors.,32373
and imbalance is network. This static voltage balancing by using mitigated dynamic,This voltage imbalance is mitigated by using static and dynamic balancing network.,32374
derives on paper models its two the balancing for based dynamic network The cycle. charge-discharge,The paper derives two models for the dynamic balancing network based on its charge-discharge cycle.,32375
"Therefore, a this to be capacity in there waste of practice. seems","Therefore, there seems to be a waste of capacity in this practice.",32376
a problem. not Producing identity results photo-realistic well multi-view synthesis still is preserving and defined,Producing photo-realistic and identity preserving multi-view results is still a not well defined synthesis problem.,32377
"process Particularly, effort. semantic human annotation considerable the segmentation, or object in requires detection","Particularly, in object detection or semantic segmentation, the annotation process requires considerable human effort.",32378
approaches limited to predicting matrices. Previous covariance have mostly been diagonal,Previous approaches have been mostly limited to predicting diagonal covariance matrices.,32379
covariances the of structure illustrate preserving for predicted image denoising. these also use We,We also illustrate the use of these predicted covariances for structure preserving image denoising.,32380
Both are resolution the these to related camera. tasks of limitations,Both these tasks are related to resolution limitations of the camera.,32381
artifacts simply demosaicing. demosaiced by super-resolution a exacerbates Performing images the on introduced,Performing super-resolution on a demosaiced images simply exacerbates the artifacts introduced by demosaicing.,32382
"Osteoarthritis world, disease in in Indonesia. including the found is a","Osteoarthritis is a disease found in the world, including in Indonesia.",32383
"their pooling explosion. this a layers serious family drawback However, of is of dimensionality","However, a serious drawback of this family of pooling layers is their dimensionality explosion.",32384
Approximate properties have explored with compact towards pooling methods this resolving weakness. been,Approximate pooling methods with compact properties have been explored towards resolving this weakness.,32385
accepted become paradigm toxicity. part has suggestion the soil a This of on,This suggestion has become a part of the accepted paradigm on soil toxicity.,32386
"to control monitor stringent Cd fertilizer have programs launched. Consequently, been","Consequently, stringent fertilizer control programs to monitor Cd have been launched.",32387
Cd to diseases fertilizers common. are to chronic toxicity and Attempts link,Attempts to link Cd toxicity and fertilizers to chronic diseases are common.,32388
data scenarios Alternative that explain the examined. are,Alternative scenarios that explain the data are examined.,32389
These point clarify difficulties the rectify found to of and in view. standard help,These help to clarify and rectify difficulties found in the standard point of view.,32390
"a that can time, i.e. RANSAC method evaluate hypothesis We a in present constant","We present a method that can evaluate a RANSAC hypothesis in constant time, i.e.",32391
the data. the independent of size of,independent of the size of the data.,32392
"Then, model been studied. cascaded and regarding impedances loop has voltage the converter close gain","Then, the cascaded converter model regarding close loop impedances and voltage gain has been studied.",32393
the plant. performance of is controller proposed converter to the The stabilize for this,The controller for this converter is proposed to stabilize the performance of the plant.,32394
The tested converter. of has typical proposed effectiveness the buck controller on been a,The effectiveness of the proposed controller has been tested on a typical buck converter.,32395
"has features. model our Furthermore, unique several","Furthermore, our model has several unique features.",32396
"maps. non-rigid predicted we algorithm tracking a present saliency the based on object Finally,","Finally, we present a non-rigid object tracking algorithm based on the predicted saliency maps.",32397
endoscopy introduced and methods advances systems capsule have Recent in capabilities. new,Recent advances in capsule endoscopy systems have introduced new methods and capabilities.,32398
capsule is unit. important of One parts of image the a compression the system endoscopy,One of the important parts of a capsule endoscopy system is the image compression unit.,32399
the hence rate process. and images the compression of increases Better frame improves diagnosis,Better compression of images increases the frame rate and hence improves the diagnosis process.,32400
with this paper is In ratio high high compression precision compression algorithm proposed. a,In this paper a high precision compression algorithm with high compression ratio is proposed.,32401
In compress the to algorithm use more efficiently. similarity this data between we frames the,In this algorithm we use the similarity between frames to compress the data more efficiently.,32402
"are However, the CDRL of theoretical not algorithms well understood. yet properties","However, the theoretical properties of CDRL algorithms are not yet well understood.",32403
captions to generate images. text Prior has work used,Prior work has used text captions to generate images.,32404
learning the by deep refine to results generated models. segmentation propose a method We such,We propose a method to refine the segmentation results generated by such deep learning models.,32405
networks powerful provide neural for image learning capability classification. Deep a feature convolutional,Deep convolutional neural networks provide a powerful feature learning capability for image classification.,32406
collaborative enables and to other. weight classic deep representations This each weigh,This collaborative weight enables deep and classic representations to weigh each other.,32407
to our going published code repository. is in public be The,The code is going to be published in our public repository.,32408
accurate for the a person prediction attention will enough. particular not make This,This will make the attention prediction for a particular person not accurate enough.,32409
an effective endoscopy capsule of diagnosis of is (WCE) gastrointestinal means Wireless disorders.,Wireless capsule endoscopy (WCE) is an effective means of diagnosis of gastrointestinal disorders.,32410
selected (MLP) a channels and Suitable classified by perceptron are color multi-layer structure.,Suitable color channels are selected and classified by a multi-layer perceptron (MLP) structure.,32411
is implementation require The such quantized does structure not the multiplications. that MLP,The MLP structure is quantized such that the implementation does not require multiplications.,32412
The dataset. tested method bleeding image by on WCE simulation is proposed,The proposed method is tested by simulation on WCE bleeding image dataset.,32413
proposed The designed WCE that structure constrains exist systems. hardware resource in is considering,The proposed structure is designed considering hardware resource constrains that exist in WCE systems.,32414
to This as epistemic argumentation. graphs of approach generalization probabilistic epistemic a paper the introduces,This paper introduces epistemic graphs as a generalization of the epistemic approach to probabilistic argumentation.,32415
statistical approach models We a propose estimation likelihood. to simulator-based intractable with for novel parameter,We propose a novel approach to parameter estimation for simulator-based statistical models with intractable likelihood.,32416
"large for protein Classical of flexibility remain especially proteins. expensive, simulations computationally","Classical simulations of protein flexibility remain computationally expensive, especially for large proteins.",32417
set of the free/admissible arguments). conflict of each,each of the conflict free/admissible set of arguments).,32418
examples exist Many in cooperation biology. of,Many examples of cooperation exist in biology.,32419
multilabel This task classification. work of image addresses the,This work addresses the task of multilabel image classification.,32420
to In models representation image CNN (i.e. that learn contrast an conventional,In contrast to conventional CNN models that learn an image representation (i.e.,32421
"model embedding developed the learns a vector), mapping image the (i.e.","the image embedding vector), the developed model learns a mapping (i.e.",32422
"present impacts control of results Hence, local intervention. global functional these a","Hence, these results present global functional impacts of a local control intervention.",32423
reactive The loads into studied. amplification of is effects,The effects of reactive loads into amplification is studied.,32424
upper gain implies a this arcs asymmetry along of loop. This or lower,This implies a gain asymmetry along upper or lower arcs of this loop.,32425
propose segmentation a of loss We cluttered cells. instance for multiclass new function weighted,We propose a new multiclass weighted loss function for instance segmentation of cluttered cells.,32426
results of microscope our T-cells. on We annotated images validate manually,We validate our results on manually annotated microscope images of T-cells.,32427
"we reward for the Therefore, learning constraints. design methodology function with a reinforcement innovate problems","Therefore, we innovate a reward function design methodology for the reinforcement learning problems with constraints.",32428
the demonstrate The on datasets. experimental evaluations of our effectiveness large-scale real framework,The experimental evaluations demonstrate the effectiveness of our framework on large-scale real datasets.,32429
Wright-Fisher the The diffusion. probabilities sample for the approximation an from characterize distribution stationary,The sample probabilities characterize an approximation for the stationary distribution from the Wright-Fisher diffusion.,32430
not is when small. rates solution known interest has solution The because not are the,The solution has interest because the solution is not known when rates are not small.,32431
slowly show the dataset size the increases. of Results as increases rate error the that,Results show that the error rate increases slowly as the size of the dataset increases.,32432
ConvNet to hand regions. a fast We likely first use detect,We use a fast ConvNet to first detect likely hand regions.,32433
a for mask hand. This pixel-level step each generates,This step generates a pixel-level mask for each hand.,32434
focused objectives These typically and separately addressed has work most image were data. previous on,These objectives were typically addressed separately and most previous work has focused on image data.,32435
"in ecosystems, and agricultural plants outbreaks food crops threaten disease Infectious trade.","Infectious disease outbreaks in plants threaten ecosystems, agricultural crops and food trade.",32436
computer-aided and important an diagnosis in plays Automatic role segmentation treatment. liver,Automatic liver segmentation plays an important role in computer-aided diagnosis and treatment.,32437
of boundary growing. incorporate target to organs the region also restrict the We,We also incorporate the boundary of target organs to restrict the region growing.,32438
These real-world of contributions modelling demonstrated inferential are and range a and signals. on synthetic,These modelling and inferential contributions are demonstrated on a range of synthetic and real-world signals.,32439
watermarking important image of parameters. distortion two reversible under and the are Capacity,Capacity and distortion of the image under reversible watermarking are two important parameters.,32440
result would compensation distortion low This method rate. in distortion,This distortion compensation method would result in low distortion rate.,32441
"of of However, the the visual image parts affect the quality perception as well. those","However, the quality of those parts affect the visual perception of the image as well.",32442
angiographic background foreground images compress methods of techniques. different using Existing and,Existing methods compress foreground and background of angiographic images using different techniques.,32443
been the for have solving networks successful very problem Two-stream detection. of action,Two-stream networks have been very successful for solving the problem of action detection.,32444
end-to-end action real-time a present two-stream for network We detection. trainable,We present a real-time end-to-end trainable two-stream network for action detection.,32445
"parameters we Additionally, observe Kinetics. an initializing when using with improvement pre-trained","Additionally, we observe an improvement when initializing with parameters pre-trained using Kinetics.",32446
the points. of repulsion: provide (DPPs) Determinantal sets modeling processes diverse enable of point they,Determinantal point processes (DPPs) enable the modeling of repulsion: they provide diverse sets of points.,32447
shakes. essential videos for to captured hand-held technique stabilization Video most is due high-frequency,Video stabilization technique is essential for most hand-held captured videos due to high-frequency shakes.,32448
"the convolutional we video problem this (ConvNet). paper, solve neural a using stabilization network In","In this paper, we solve the video stabilization problem using a convolutional neural network (ConvNet).",32449
an ability. hearing There with between people deaf people undeniable normal barrier and communication is,There is an undeniable communication barrier between deaf people and people with normal hearing ability.,32450
to DeepASL non-intrusively uses ASL signs. infrared capture the mechanism light its as sensing,DeepASL uses infrared light as its sensing mechanism to non-intrusively capture the ASL signs.,32451
is information screening helpful diagnosis. vessel disease in retinal and Retinal,Retinal vessel information is helpful in retinal disease screening and diagnosis.,32452
classification and tools (CNNs) powerful segmentation networks Convolutional images. medical are of neural for,Convolutional neural networks (CNNs) are powerful tools for classification and segmentation of medical images.,32453
by a is levels reduce competition operating and costs. high driven to of need This,This is driven by high levels of competition and a need to reduce operating costs.,32454
We for adversarial semi-supervised using network. a propose method an segmentation semantic,We propose a method for semi-supervised semantic segmentation using an adversarial network.,32455
major cause of women. Cancer a is death Breast worldwide among,Breast Cancer is a major cause of death worldwide among women.,32456
of the classes. is entire determined image majority the class over voting using The nuclear,The class of the entire image is determined using majority voting over the nuclear classes.,32457
techniques. several Duality play comes these in into steps at,Duality comes into play at several steps in these techniques.,32458
"this problem. paper, we propose techniques this addressing for two In","In this paper, we propose two techniques for addressing this problem.",32459
accounted for further negative to transfer. avoid Data are also knowledge outliers,Data outliers are also accounted for to further avoid negative knowledge transfer.,32460
have same. addressing Several paradigms processing been explored image learning the for and machine,Several image processing and machine learning paradigms have been explored for addressing the same.,32461
objective Minimizing with a non-differentiable problem. a in deblurring optimization blind RGTV results image non-convex,Minimizing a blind image deblurring objective with RGTV results in a non-convex non-differentiable optimization problem.,32462
"strategy discovered i.e. studies Adversarial vulnerable have Examples, that is recent However, modelling such to","However, recent studies have discovered that such modelling strategy is vulnerable to Adversarial Examples, i.e.",32463
"trained the of independently employ Particularly, posterior layer for models we features as detection. decision","Particularly, we employ the decision layer of independently trained models as features for posterior detection.",32464
technique localization. propose We improved weakly-supervised object for an,We propose an improved technique for weakly-supervised object localization.,32465
enables computer key pertaining This dataset to satellite research vision imagery progress four frontiers.,This satellite imagery dataset enables research progress pertaining to four key computer vision frontiers.,32466
to We to build models machine a propose such train too.,We propose to train a machine to build such models too.,32467
organs. for abdominal of and for growing therapy CT is analysis pre-diagnosis need The scan,The need for CT scan analysis is growing for pre-diagnosis and therapy of abdominal organs.,32468
field the organ's on of conditional segmented mask The is means border. enhanced by random,The segmented mask is enhanced by means of conditional random field on the organ's border.,32469
much attention complexities researchers from have of fatigue The disciplines. drawn various across,The complexities of fatigue have drawn much attention from researchers across various disciplines.,32470
"eventually and Long-term and chronic affect physical lead health. fatigue psychological individuals could syndromes, to","Long-term fatigue could lead to chronic syndromes, and eventually affect individuals physical and psychological health.",32471
of also equipment but not evaluating time. require Traditional methodologies only enormous consume fatigue sophisticated,Traditional methodologies of evaluating fatigue not only require sophisticated equipment but also consume enormous time.,32472
is Our goal on to based predict selfie. fatigue rate a,Our goal is to predict fatigue rate based on a selfie.,32473
identify first faces algorithms. all automatic using the detect demographics We their and,We first detect all the faces and identify their demographics using automatic algorithms.,32474
"distribution ethnic investigate Next, groups. fatigue and gender, age, we different by the over weekday","Next, we investigate the fatigue distribution by weekday over different age, gender, and ethnic groups.",32475
"glimpse Instead, attention a in visual sequences frame. each module learns to predict","Instead, a visual attention module learns to predict glimpse sequences in each frame.",32476
"prediction. jointly performing subsequent glimpses, receive motion These and workers activity tracking","These workers receive glimpses, jointly performing subsequent motion tracking and activity prediction.",32477
"assigned all workers, to glimpse with different importance. point is each existing albeit","each glimpse point is assigned to all existing workers, albeit with different importance.",32478
"(re-ID). we paper, this learning propose a feature framework re-identification novel video for In person","In this paper, we propose a novel feature learning framework for video person re-identification (re-ID).",32479
"of specular detection cost reflections introduced. a function for Then, is","Then, a cost function for detection of specular reflections is introduced.",32480
power This favorable using than economically or consuming processing techniques. more way is,This way is more economically favorable than consuming power or using processing techniques.,32481
"a oversegmented For a is into of it superpixels. homogeneous image, first set testing","For a testing image, it is first oversegmented into a set of homogeneous superpixels.",32482
vegetation scenarios. also the in demonstrate We of the segmentation real-world effectiveness for approach,We also demonstrate the effectiveness of the approach for vegetation segmentation in real-world scenarios.,32483
Langevin Algorithms sampling from on are discretizing distributions. based popular diffusion for high-dimensional tools,Algorithms based on discretizing Langevin diffusion are popular tools for sampling from high-dimensional distributions.,32484
step in tumors. Accurate segmentation of a the evaluating characteristics breast of crucial is lesions,Accurate segmentation of breast lesions is a crucial step in evaluating the characteristics of tumors.,32485
algorithms outperformed Watershed results the that marker-controlled transformation investigated. The all indicate other,The results indicate that the marker-controlled Watershed transformation outperformed all other algorithms investigated.,32486
"optimization to demonstrate learning. from separated that than end-to-end higher accuracy we significantly leads Additionally,","Additionally, we demonstrate that optimization from end-to-end leads to significantly higher accuracy than separated learning.",32487
Point al. The Feature (Drost Pair et,The Point Pair Feature (Drost et al.,32488
"of have several the algorithm years, During the last proposed. variations been","During the last years, several variations of the algorithm have been proposed.",32489
"introduced al. solution these Among Hinterstoisser extensions, the by et","Among these extensions, the solution introduced by Hinterstoisser et al.",32490
and semantic is of branch. branch composed an a SA-Siam appearance,SA-Siam is composed of a semantic branch and an appearance branch.,32491
Siamese branch network. Each is similarity-learning a,Each branch is a similarity-learning Siamese network.,32492
"we attention propose mechanism branch. for a the channel semantic In addition,","In addition, we propose a channel attention mechanism for the semantic branch.",32493
to the the target computed activations position. according Channel-wise channel weights around are,Channel-wise weights are computed according to the channel activations around the target position.,32494
"unknown. In and random parameters reality, are some","In reality, some parameters are random and unknown.",32495
the programming formulate customer demand problem use stochastic the optimization uncertainty. We integer incorporating to,We use stochastic integer programming to formulate the optimization problem incorporating the customer demand uncertainty.,32496
"task. Generative for Face Aging Deep Models,","Deep Generative Models, for Face Aging task.",32497
"approach for (GANs). this work, Adversarial present we Networks a Generative novel In training","In this work, we present a novel approach for training Generative Adversarial Networks (GANs).",32498
in of leading mortality one is cancer causes women. of Breast the,Breast cancer is one of the leading causes of mortality in women.,32499
"constrained we the features, invariant a as formulate the nonlinear least-squares. problem Using","Using the invariant features, we formulate the problem as a constrained nonlinear least-squares.",32500
initialization. problem non-convexity to of sensitivity The leads the to the its,The non-convexity of the problem leads to its sensitivity to the initialization.,32501
scans. shows cardiac the clinical technique potential as This the protocol of a for free-breathing,This shows the potential of the technique as a clinical protocol for free-breathing cardiac scans.,32502
can paper follow. model a statistics a to presents asked that This mathematician teach,This paper presents a model that a mathematician asked to teach statistics can follow.,32503
presence of better Antifragile in measurably the hazards. grow systems,Antifragile systems grow measurably better in the presence of hazards.,32504
robot study An is of case autonomous AI planning-based presented. wheeled an,An AI planning-based case study of an autonomous wheeled robot is presented.,32505
"we investigate why paper, this the case. this is In","In this paper, we investigate why this is the case.",32506
"recognition real-world particularly tolerant Automatic understood, less to partial scenarios. remains expression occlusion in","Automatic expression recognition tolerant to partial occlusion remains less understood, particularly in real-world scenarios.",32507
light field a and proposes novel the information method simultaneously. paper capture to This spectral,This paper proposes a novel method to capture the spectral and light field information simultaneously.,32508
the method. the proposed of verify effectiveness The experiments,The experiments verify the effectiveness of the proposed method.,32509
problem with the of given binary generation We consider properties. image,We consider the problem of binary image generation with given properties.,32510
"solver the PDE However, search embedding a into computationally procedure expensive. is","However, embedding a PDE solver into the search procedure is computationally expensive.",32511
binarized a to network approximate use a neural We solver. PDE,We use a binarized neural network to approximate a PDE solver.,32512
logical This allows entire a formula. us as to problem the encode,This allows us to encode the entire problem as a logical formula.,32513
"directly, second processing. segments but post mostly The one text needs complex regions","The second one segments text regions directly, but mostly needs complex post processing.",32514
considered Both step in the online offline scenarios are second DTER. of and,Both offline and online scenarios are considered in the second step of DTER.,32515
well. developed as is suboptimal strategy online The transmission,The online suboptimal transmission strategy is developed as well.,32516
billion in quality Five access surgical to people world lack care. the,Five billion people in the world lack access to quality surgical care.,32517
"skill many and avoidable varies Surgeon complications surgical and dramatically, patients suffer harm.","Surgeon skill varies dramatically, and many surgical patients suffer complications and avoidable harm.",32518
independence constructions Most assume current full of resolutions. the across,Most of the current constructions assume full independence across resolutions.,32519
resolutions. which construction among assumes instead contribution new independence Our across is GPs a conditional,Our contribution is a new construction which instead assumes conditional independence among GPs across resolutions.,32520
"little particular, exhibits to of signs no In overfitting. it","In particular, it exhibits little to no signs of overfitting.",32521
correct We intervals. split-panel also confidence a construction of method to consider and jackknife bias,We also consider a split-panel jackknife method to correct bias and construction of confidence intervals.,32522
and in Monte illustrate samples. application empirical finite our An procedure Carlo simulations,An empirical application and Monte Carlo simulations illustrate our procedure in finite samples.,32523
energy grid demand. meeting to BSs additional are for supply the connected also conventional,BSs are also connected to the conventional grid supply for meeting additional energy demand.,32524
Traditional population repertoire to immune healthy chronic from methods distinguish response genetics fail evolution.,Traditional population genetics methods fail to distinguish chronic immune response from healthy repertoire evolution.,32525
batches and samples similar Conflicting distributed groups. grouped are relations into are compact different over,Conflicting relations are distributed over different batches and similar samples are grouped into compact groups.,32526
visual sequence similarities categorization a as Learning tasks. of then framed is,Learning visual similarities is then framed as a sequence of categorization tasks.,32527
competitive proposed The analysis approach has on unsupervised and posture shown classification. detailed object performance,The proposed unsupervised approach has shown competitive performance on detailed posture analysis and object classification.,32528
training been advantages data for have section using level in outlined The I. character of,The advantages of using character level data for training have been outlined in section I.,32529
"to Thereafter, modality network. our pair we this feed offline-online","Thereafter, we feed this offline-online modality pair to our network.",32530
"constrain using demonstrations. Instead, to exploration we propose","Instead, we propose to constrain exploration using demonstrations.",32531
the Workflows exploration to discover agent's ability prune bad accelerate out and directions rewards.,Workflows prune out bad exploration directions and accelerate the agent's ability to discover rewards.,32532
by training way is One occupants. can this be achieved,One way this can be achieved is by training occupants.,32533
shift problems as prediction in both these pose of We design. a under,We pose both of these problems as prediction under a shift in design.,32534
"classifiers one-class Therefore, efficiently such can problems. model","Therefore, one-class classifiers can efficiently model such problems.",32535
effectiveness experiments method. The demonstrate the proposed of,The experiments demonstrate the effectiveness of proposed method.,32536
accurately done efficiently. across fusing by object This proposals is and sensors,This is done by fusing object proposals across sensors accurately and efficiently.,32537
"directly are interest in of the tracked world. (targets) objects the Second, real","Second, the objects of interest (targets) are tracked directly in the real world.",32538
skills. already learned many ONE has Suppose,Suppose ONE has already learned many skills.,32539
harvesting communications. problem radio a transmission The is in (RF) critical scheduling frequency energy,The transmission scheduling is a critical problem in radio frequency (RF) energy harvesting communications.,32540
"established matching ensure some measures. different To combine handcrafted robustness, drive the similarity we and","To drive the matching and ensure some robustness, we combine different established handcrafted similarity measures.",32541
"stereogrammetry sensing, optical data . Radar Synthetic fusion, images, Keywords: remote (SAR), Aperture","Keywords: Synthetic Aperture Radar (SAR), optical images, remote sensing, data fusion, stereogrammetry .",32542
"(SNR) major in SHI. the signal-to-noise problem However, low ratio is a","However, the low signal-to-noise ratio (SNR) is a major problem in SHI.",32543
"minimum second for Eigenspace-based harmonic (EIBMV) In this paper, USI. variance has been beamformer employed","In this paper, Eigenspace-based minimum variance (EIBMV) beamformer has been employed for second harmonic USI.",32544
(THI) Imaging Inversion is by (PI) Harmonic technique. Tissue achieved The Pulse,The Tissue Harmonic Imaging (THI) is achieved by Pulse Inversion (PI) technique.,32545
"provide However, accurate image. low they a","However, they provide a low accurate image.",32546
in extinction of (SIS) disease the susceptible-infectious-susceptible to the spreading. expected model time We investigate,We investigate the expected time to extinction in the susceptible-infectious-susceptible (SIS) model of disease spreading.,32547
enables descriptive stochastic with be relations simulations. approach impossible would discover to This that us,This approach enables us to discover descriptive relations that would be impossible with stochastic simulations.,32548
important the of factor ranking most this vertices. infectious determining The structural is the degrees,The most important structural factor determining this ranking is the degrees of the infectious vertices.,32549
impressive The resulting in datasets. results ALPR achieved approach two,The resulting ALPR approach achieved impressive results in two datasets.,32550
networks Characterizing functional humans perception. during magnetophosphene brain in,Characterizing functional brain networks in humans during magnetophosphene perception.,32551
consistently activate right occipito-temporal Magnetophosphene to pathway. inferior perception appears the,Magnetophosphene perception appears to consistently activate the right inferior occipito-temporal pathway.,32552
very the magnetophosphene study characterizing in perception first results provides neuroimaging humans. This,This study provides the very first neuroimaging results characterizing magnetophosphene perception in humans.,32553
source in is a approach bioelectromagnetics. connectivity use of field promising of dense-EEG The the,The use of dense-EEG source connectivity is a promising approach in the field of bioelectromagnetics.,32554
"labeled we framework, fully our knowledge first existing domain-invariant In the categories. objectness from learn","In our framework, we first learn domain-invariant objectness knowledge from the existing fully labeled categories.",32555
"general, it although in not solution, cases. in does may The exist exceptional exist,","The solution, in general, does not exist, although it may exist in exceptional cases.",32556
"to approach of to is, This models. unlikely wider therefore, a economic set applicable be","This approach is, therefore, unlikely to be applicable to a wider set of economic models.",32557
alternately a clique network neural architecture new Here (CliqueNet). updated with propose convolutional we,Here we propose a new convolutional neural network architecture with alternately updated clique (CliqueNet).,32558
and are constructed loop layers The are a as updated alternately.,The layers are constructed as a loop and are updated alternately.,32559
a feature effectively that strategy the multi-scale avoids progressive parameters. adopt We of growth,We adopt a multi-scale feature strategy that effectively avoids the progressive growth of parameters.,32560
"standard dataset, by metrics. ActivityNet improved demonstrates the Captions measured as large-scale results On JEDDi-Net","On the large-scale ActivityNet Captions dataset, JEDDi-Net demonstrates improved results as measured by standard metrics.",32561
dataset. dense We the results also TACoS-MultiLevel present on first captioning the,We also present the first dense captioning results on the TACoS-MultiLevel dataset.,32562
This in paper a acquisition activity multi-disciplinary review knowledge systems. methods offers human of,This paper offers a multi-disciplinary review of knowledge acquisition methods in human activity systems.,32563
"consider the we both placement and interference-limited under caching noise-limited strategy particular, In environments.","In particular, we consider the caching placement strategy under both noise-limited and interference-limited environments.",32564
"the demonstrate of albeit scheme over results superiority proposed trade-offs. certain the Numerical others, caching","Numerical results demonstrate the superiority of the proposed caching scheme over others, albeit certain trade-offs.",32565
"trajectory method we models. behavior segmentation semantic propose paper, In a based on learned this","In this paper, we propose a trajectory semantic segmentation method based on learned behavior models.",32566
"the video behavior models some proposed we method, from In learn sequences.","In the proposed method, we learn some behavior models from video sequences.",32567
"the effectiveness with show we the Ramer-Douglas-Peucker method. Comparing algorithm, of proposed the","Comparing with the Ramer-Douglas-Peucker algorithm, we show the effectiveness of the proposed method.",32568
"generalization transform In (GFT). we present Fourier novel paper, of the a graph this","In this paper, we present a novel generalization of the graph Fourier transform (GFT).",32569
"Further, a the model recursive can into architecture. proposed stacked be multi-layer","Further, the proposed recursive model can be stacked into a multi-layer architecture.",32570
"synthesized attenuation data (CT) X-ray therapy Especially, can computed tomography map radiation provide for planning.","Especially, synthesized computed tomography (CT) data can provide X-ray attenuation map for radiation therapy planning.",32571
alternatively fashion. other each and in Generators segmentors prompt an end-to-end training,Generators and segmentors prompt each other alternatively in an end-to-end training fashion.,32572
"expression, of Tracy-Widom distribution it the Since brings processing. has no closed-form difficulty great in","Since the distribution of Tracy-Widom has no closed-form expression, it brings great difficulty in processing.",32573
"Gaussian the theory, eigenvalue According random the maximum shifted with matrix is consistent distribution. to","According to the random matrix theory, the shifted maximum eigenvalue is consistent with Gaussian distribution.",32574
environments. verified signal results in The simulations algorithm are different the theoretical by and,The algorithm and theoretical results are verified by the simulations in different signal environments.,32575
and the potential There benefits system. are SP of many applications,There are many potential benefits and applications of the SP system.,32576
"needed in Thus, extra formats. process is an to data prepare specific","Thus, an extra process is needed to prepare data in specific formats.",32577
easy-trained of because an CSRNet convolutional pure model structure. is its,CSRNet is an easy-trained model because of its pure convolutional structure.,32578
"is inference computationally processing to frames video due on However, expensive CNN dense individually.","However, CNN inference on video is computationally expensive due to processing dense frames individually.",32579
running them. to algorithms machine before hyperparameters Modern supervised that have involve be set learning,Modern supervised machine learning algorithms involve hyperparameters that have to be set before running them.,32580
this goal two-fold. The paper is of,The goal of this paper is two-fold.,32581
We their to of tunability assess our apply parameters. measures the,We apply our measures to assess the tunability of their parameters.,32582
Markov natural decision model problems. observable are capture a processes to such (POMDPs) formal Partially,Partially observable Markov decision processes (POMDPs) are a natural formal model to capture such problems.,32583
"strategy into Via translate a the cloning, we the for POMDP. data behavior","Via behavior cloning, we translate the data into a strategy for the POMDP.",32584
"identify is and images. the seek, Image to the localize feature matching similarities across","Image feature matching is to seek, localize and identify the similarities across the images.",32585
can between of indicate content. The images the different similarities features local their matched,The matched local features between different images can indicate the similarities of their content.,32586
stages can sufficient resilience this of view proposed Both changes. algorithm point to new provide,Both stages of this new proposed algorithm can provide sufficient resilience to view point changes.,32587
state-of-the-art. demonstrate experiments our other that perform outperforms to approach We,We perform experiments to demonstrate that our approach outperforms other state-of-the-art.,32588
make in will our facilitate wound database available to We research publicly assessment.,We will make our database publicly available to facilitate research in wound assessment.,32589
"aesthetics. is rising image interest there in a perceiving Recently,","Recently, there is a rising interest in perceiving image aesthetics.",32590
deal aesthetics with as works image or regression problem. classification a existing The,The existing works deal with image aesthetics as a classification or regression problem.,32591
approach We to multi-dimensional a modelling. domain present,We present a multi-dimensional approach to domain modelling.,32592
"respect of operator, matrix the functions shift the polynomial filtering. to graph perform With shift","With respect to the graph shift operator, polynomial functions of the shift matrix perform filtering.",32593
on graph shift eigendecomposition. Design filters depends of matrix the,Design of graph filters depends on the shift matrix eigendecomposition.,32594
"Consequently, this information graphs is as difficult a often present challenge obtain. random to","Consequently, random graphs present a challenge as this information is often difficult to obtain.",32595
to approximations filters. consensus the application Numerical support simulations and derived,Numerical simulations support the derived approximations and application to consensus filters.,32596
"between typically map theoretical However, the competition measured traits. knowledge of complete assume predictions and","However, theoretical predictions typically assume complete knowledge of the map between competition and measured traits.",32597
in These for limit inferring interactions the of these patterns competitive plausible application assumptions nature.,These assumptions limit the plausible application of these patterns for inferring competitive interactions in nature.,32598
"visible However, measured unless not close for proxies be strategies. it traits niche may are","However, it may not be visible unless measured traits are close proxies for niche strategies.",32599
"in coding this paper, non-sparse a difficulty. studied remain","non-sparse coding studied in this paper, remain a difficulty.",32600
"communications constraints considered theory. However, in not neural commonly face populations","However, neural populations face constraints not commonly considered in communications theory.",32601
research problem. intake an Measuring is accurate dietary open,Measuring accurate dietary intake is an open research problem.,32602
"for introduce concept each of an food the We distribution"" ""energy image.","We introduce the concept of an ""energy distribution"" for each food image.",32603
image the energy. mapping the of food is to learn Our to goal food the,Our goal is to learn the mapping of the food image to the food energy.,32604
based energy distribution. then food on estimate energy the can We,We can then estimate food energy based on the energy distribution.,32605
"composed deconvolutional Specifically, it and two convolutional sub-network i.e., of is sub-network. parts,","Specifically, it is composed of two parts, i.e., convolutional sub-network and deconvolutional sub-network.",32606
visual results demonstrate experimental the approach. effectiveness and of the our of analysis quantitative Both,Both visual and quantitative analysis of the experimental results demonstrate the effectiveness of our approach.,32607
"an optical First, each image generate flow we video frame. for","First, we generate an optical flow image for each video frame.",32608
periodicity. the assumption shows Existing results and under work static of good stationary,Existing work shows good results under the assumption of static and stationary periodicity.,32609
"to handle dynamics. wavelet adopt non-stationary better non-static video and the transform we Instead,","Instead, we adopt the wavelet transform to better handle non-static and non-stationary video dynamics.",32610
complicated This task uncertain eigenvalues. iteration matrix by with networks is random,This task is complicated by random networks with uncertain iteration matrix eigenvalues.,32611
"This upon networks. results by analysis these extending large-scale, builds to directed constant, work random","This work builds upon these results by extending analysis to large-scale, constant, random directed networks.",32612
"discussed. of are Additionally, limitations proposed the method","Additionally, limitations of the proposed method are discussed.",32613
also is future framework plan A described. improvements of,A future plan of framework improvements is also described.,32614
depict weighting illumination-aware condition of present to a scene. accurately mechanism illumination A is novel,A novel illumination-aware weighting mechanism is present to accurately depict illumination condition of a scene.,32615
Latent appear various hidden applications. with variable binary models units in,Latent variable models with hidden binary units appear in various applications.,32616
"noise, in in challenging particular is the such models, of a problem. presence Learning computational","Learning such models, in particular in the presence of noise, is a challenging computational problem.",32617
search fast a This the implies memory of contents.,This implies a fast search of the memory contents.,32618
"respectively and As coil). hermeneutics consequence, a emerge two circle (called","As a consequence, two hermeneutics emerge (called respectively circle and coil).",32619
acts one of a features. space in pre-assigned The first,The first one acts in a pre-assigned space of features.,32620
the scaling of deep consider shape high-resolution. problem generative We models to,We consider the problem of scaling deep generative shape models to high-resolution.,32621
This generate easily individual method than more allows network. to our edges sharp an,This allows our method to generate sharp edges more easily than an individual network.,32622
state-of-the-art We network across demonstrate compression datasets and array rates an of architectures.,We demonstrate state-of-the-art compression rates across an array of datasets and network architectures.,32623
"missing a for still piece. is However, transfer style stereoscopic images","However, style transfer for stereoscopic images is still a missing piece.",32624
at having the illustrates This encoder. difference in factor importance of the performance the decimation,This difference in performance illustrates the importance of having the decimation factor at the encoder.,32625
BSs necessary the utilize BSs is when it quite sufficient. to So rationally are resources,So it is quite necessary to utilize the BSs resources rationally when BSs are sufficient.,32626
We advantage. transform canyons an to the strong weakness (urban relief) and terrain,We transform the weakness (urban canyons and strong terrain relief) to an advantage.,32627
terrain using is the a of map relief. navigation a It vision-based,It is a vision-based navigation using a map of the terrain relief.,32628
technology demonstrate this in Xiaoshan. investigate region the and effectiveness We Chinese of,We investigate and demonstrate the effectiveness of this technology in Chinese region Xiaoshan.,32629
corresponds these expected vision-based navigation the of for the conditions. The to accuracy system,The accuracy of the vision-based navigation system corresponds to the expected for these conditions.,32630
"approach of this the exactly. condition guarantees is By construction, mass conservation discretization satisfied that","By construction, this discretization approach guarantees that the conservation of mass condition is satisfied exactly.",32631
"is inevitably in However, noise. applications the by contaminated practical CSI","However, in practical applications the CSI is inevitably contaminated by noise.",32632
of is cause worldwide. leading cancer cancer among second the women Breast death,Breast cancer is the second leading cause of cancer death among women worldwide.,32633
"most Nevertheless, detected malignances of it also one if the early. is treatable","Nevertheless, it is also one of the most treatable malignances if detected early.",32634
Screening been with for digital (DM) used. mammography has breast widely cancer,Screening for breast cancer with digital mammography (DM) has been widely used.,32635
women it limited with demonstrates However dense breasts. for sensitivity,However it demonstrates limited sensitivity for women with dense breasts.,32636
DM. accuracy better diagnostic shown than CEDM has,CEDM has shown better diagnostic accuracy than DM.,32637
"centers. across widely not While medical CEDM available is promising, yet","While promising, CEDM is not yet widely available across medical centers.",32638
of important inputs The ability is classifier a many unknown recognize for to systems. classification-based,The ability of a classifier to recognize unknown inputs is important for many classification-based systems.,32639
"the discuss detection, problem simultaneous and We of novelty classification i.e.","We discuss the problem of simultaneous classification and novelty detection, i.e.",32640
framework. Generative We a (GAN) Networks on based propose method the Adversarial,We propose a method based on the Generative Adversarial Networks (GAN) framework.,32641
corroborated our has been method of designed strength experiments. by several The carefully,The strength of our method has been corroborated by several carefully designed experiments.,32642
"is the improved efficiency significantly In running the accuracy filtering sacrificing at without time. particular,","In particular, the filtering accuracy is significantly improved without sacrificing the efficiency at running time.",32643
for analysis. of neural used recently a been variety Convolutional (CNNs) histology networks image have,Convolutional neural networks (CNNs) have been recently used for a variety of histology image analysis.,32644
cancer proposed classification. the state-of-the-art breast The framework method for outperformed,The proposed framework outperformed the state-of-the-art method for breast cancer classification.,32645
further of bottleneck The samples shortage a of performance. improvement training become has for,The shortage of training samples has become a bottleneck for further improvement of performance.,32646
"from extracted is feature and target pixel Firstly, a neighbors. the its spectral-spatial","Firstly, the spectral-spatial feature is extracted from a target pixel and its neighbors.",32647
"is fed as CNN. standard the an image into matrix two-dimensional considered Finally,","Finally, the two-dimensional matrix considered as an image is fed into standard CNN.",32648
it why we HSI-CNN. This is call,This is why we call it HSI-CNN.,32649
is with the framework. classification improved HSI-CNN Experiments hyperspectral show efficiently performance of image that,Experiments show that the performance of hyperspectral image classification is improved efficiently with HSI-CNN framework.,32650
the The image. of depends the the segmentation consistency of over the on contrast reliability,The reliability of the segmentation depends on the consistency of the contrast over the image.,32651
was and proposed two evaluated The DRIVE. approach datasets on public STARE,The proposed approach was evaluated on two public datasets STARE and DRIVE.,32652
promising choices qualitatively their We aspect. motivating demonstrating present methodological mosaics and our,We present mosaics motivating qualitatively our methodological choices and demonstrating their promising aspect.,32653
"pattern probing, is which in metric, consists for i.e. double-sided Aspl{\""u}nd's useful a matching,","Aspl{\""u}nd's metric, which is useful for pattern matching, consists in a double-sided probing, i.e.",32654
function a the jointly. the are probed sub-graph and of over-graph,the over-graph and the sub-graph of a function are probed jointly.,32655
grey-scale Processing defined images the has using framework. It for Logarithmic Image previously (LIP) been,It has previously been defined for grey-scale images using the Logarithmic Image Processing (LIP) framework.,32656
significant Numerous experiments interesting conducted numbers presented. with recommendations of and are are intensive results,Numerous numbers of intensive experiments are conducted and interesting results with significant recommendations are presented.,32657
methods in known successful tasks. and classification to machine are many pattern be Graph-based learning,Graph-based methods are known to be successful in many machine learning and pattern classification tasks.,32658
characterize the primitives. relationships edges and these between,and edges characterize the relationships between these primitives.,32659
graph variations should be intra-class embedding being highly to resilient process discriminant. This while,This embedding process should be resilient to intra-class graph variations while being highly discriminant.,32660
summarize main motivations the We and the concepts about semiparametric intuitive models.,We summarize the main motivations and the intuitive concepts about semiparametric models.,32661
online one advertising. the optimization critical problems is in of Bidding most,Bidding optimization is one of the most critical problems in online advertising.,32662
"consider paper, sponsored RTB SS-RTB. auction, we in In this problem named search the","In this paper, we consider the RTB problem in sponsored search auction, named SS-RTB.",32663
previous for applied. DA be Most cannot methods,Most previous methods for DA cannot be applied.,32664
(RL) environment. We dynamic the learning handling for a propose solution reinforcement complex,We propose a reinforcement learning (RL) solution for handling the complex dynamic environment.,32665
also extend the to We problem. handle method the multi-agent,We also extend the method to handle the multi-agent problem.,32666
Alibaba. auction in SS-RTB of search e-commerce deployed the the system platform We,We deployed the SS-RTB system in the e-commerce search auction platform of Alibaba.,32667
of the our offline A/B method. effectiveness evaluation test of demonstrate online and experiments Empirical,Empirical experiments of offline evaluation and online A/B test demonstrate the effectiveness of our method.,32668
of decision tumors analysis clinical for critical making. brain Quantitative is,Quantitative analysis of brain tumors is critical for clinical decision making.,32669
the the solution known. closed is computation form No quantiles for of,No closed form solution for the computation of the quantiles is known.,32670
and across Then efficiency domain. we accuracy restore algorithms provide which the alternate whole,Then we provide alternate algorithms which restore accuracy and efficiency across the whole domain.,32671
based (DNN) expressive network and capacity. Deep models architecture high learning neural have power,Deep neural network (DNN) architecture based models have high expressive power and learning capacity.,32672
powerful be a methodology. for Games learning tool can about statistical,Games can be a powerful tool for learning about statistical methodology.,32673
Question Answering core is detection a Base component Knowledge (KBQA). for Relation,Relation detection is a core component for Knowledge Base Question Answering (KBQA).,32674
perspectives two matching multiple is The texts thoroughly. each view input inside compare through to,The matching inside each view is through multiple perspectives to compare two input texts thoroughly.,32675
designed All these model. end-to-end components an trainable are neural in network,All these components are designed in an end-to-end trainable neural network model.,32676
on WebQSP state-of-the-art yield results. and SimpleQuestions Experiments,Experiments on SimpleQuestions and WebQSP yield state-of-the-art results.,32677
achieves the success of emergence learning. exceptional Face thanks to deep recognition,Face recognition achieves exceptional success thanks to the emergence of deep learning.,32678
is feature to Visual approaches segment in objects cost-effective clustering one videos. of the,Visual feature clustering is one of the cost-effective approaches to segment objects in videos.,32679
in strong baselines low extremely resolution also outperform handling face We matching.,We also outperform strong baselines in handling extremely low resolution face matching.,32680
task dense images. (NRSfM) addresses paper multiple non-rigid the of structure-from-motion using This,This paper addresses the task of dense non-rigid structure-from-motion (NRSfM) using multiple images.,32681
"expensive and hurdled by methods scalability, this often noisy State-of-the-art to problem measurements. are computations,","State-of-the-art methods to this problem are often hurdled by scalability, expensive computations, and noisy measurements.",32682
for the of naturally a allows frames. complex compact This representation deformation non-rigid over,This naturally allows for a compact representation of the complex non-rigid deformation over frames.,32683
benchmark datasets. real and experimental We provide several results synthetic on,We provide experimental results on several synthetic and real benchmark datasets.,32684
parts. OCSVM Several space input several the divide into models,Several OCSVM models divide the input space into several parts.,32685
also datasets experiments the the method. verify different of Various proposed efficiency on,Various experiments on different datasets also verify the efficiency of the proposed method.,32686
on and entropy Tsallis The of based approach. thermodynamics Renyi calculation,The calculation of Renyi and Tsallis entropy based on thermodynamics approach.,32687
is Filter (GF) complexity. Guided for The linear its well-known,The Guided Filter (GF) is well-known for its linear complexity.,32688
existing on algorithms inverse of inefficient current are To hardwares. the best matrix knowledge our,To the best of our knowledge existing matrix inverse algorithms are inefficient on current hardwares.,32689
multichannel such of computation multi-label as system system. applications guidance limits in intensive shortcoming This,This shortcoming limits applications of multichannel guidance in computation intensive system such as multi-label system.,32690
We image new a need filtering. filter guided GF-like multichannel that can perform fast,We need a new GF-like filter that can perform fast multichannel image guided filtering.,32691
to are agent not training. often known during will The tasks an solve that need,The tasks that an agent will need to solve often are not known during training.,32692
"varied to problem, are to this There which are applied settings. approaches different","There are varied approaches to this problem, which are applied to different settings.",32693
a excludes outliers. determine real-time The range to help that QTFs,The QTFs help to determine a real-time range that excludes outliers.,32694
CMTF responsible is range is into This for mitigating fed noise. which impulsive the,This range is fed into the CMTF which is responsible for mitigating impulsive noise.,32695
scale were exhibit feature. and conducted of invariance the rotational to Experiments,Experiments were conducted to exhibit rotational and scale invariance of the feature.,32696
tackle networks adversarial this We (GANs). using problem generative,We tackle this problem using generative adversarial networks (GANs).,32697
"GANs of been samples. sets find between utilized to correspondences Recently, have try to","Recently, GANs have been utilized to try to find correspondences between sets of samples.",32698
"for these are of manifolds. explicitly alignment proper designed not However, GANs","However, these GANs are not explicitly designed for proper alignment of manifolds.",32699
applications measurement biology We in single-cell two together. demonstrate integrating types MAGAN different of in,We demonstrate applications of MAGAN in single-cell biology in integrating two different measurement types together.,32700
the standard a also techniques conditions. (EKF) highly provide under with comparison We noisy,We also provide a comparison with the standard techniques (EKF) under highly noisy conditions.,32701
transfer constrains the style proposed method around The area. only designated,The proposed method constrains style transfer only around the designated area.,32702
We of proposed demonstrate method. the characteristics,We demonstrate the characteristics of proposed method.,32703
"we input Finally, with the logo show various results generation of images.","Finally, we show the results of logo generation with various input images.",32704
"of power and estimate we signal Specifically, the from bispectrum spectrum, mean, the observations. the","Specifically, we estimate the mean, power spectrum, and bispectrum of the signal from the observations.",32705
"In the method to is spectral robust noise. addition,","In addition, the spectral method is robust to noise.",32706
"an Hence, approximate not the approach localization initialization. for does our humans require of","Hence, our approach does not require an approximate localization of the humans for initialization.",32707
convolutional are trained share stages All jointly. the three feature layers and,All three stages share the convolutional feature layers and are trained jointly.,32708
great Deep complex have shown neural networks achievements solving problems. in,Deep neural networks have shown great achievements in solving complex problems.,32709
"are that fundamental world limit problems real their there applications. However,","However, there are fundamental problems that limit their real world applications.",32710
to convolutional transformations networks. estimate calculate neural uncertainty random the We deep of simple prediction,We calculate simple random transformations to estimate the prediction uncertainty of deep convolutional neural networks.,32711
an uncertainty. method deep We also propose consider network to neural adaptive thresholding the,We also propose an adaptive thresholding method to consider the deep neural network uncertainty.,32712
form as structural reduced framework recently special cases. machine The nests and proposed learning approaches,The framework nests recently proposed structural and reduced form machine learning approaches as special cases.,32713
original I countries find was effect no in where for study. effects the several found,I find effects for several countries where no effect was found in the original study.,32714
model adaptive variable filter paper new This that allows scale orientation. introduces a and,This paper introduces a new adaptive filter model that allows variable scale and orientation.,32715
using can orientation of learned filters propagation. and parameters scale The back be,The scale and orientation parameters of filters can be learned using back propagation.,32716
base relatively filters. for model The a size (grid) large proposed uses,The proposed model uses a relatively large base size (grid) for filters.,32717
"differentiable grid, for a acts the an envelope function In as filters. the","In the grid, a differentiable function acts as an envelope for the filters.",32718
"are the envelope only updated in weights the during Therefore, training.","Therefore, only the weights in the envelope are updated during training.",32719
"FD communications of highlighted. are potential open Finally, challenges multi-cell research","Finally, potential open research challenges of multi-cell FD communications are highlighted.",32720
"obtained the in probability In particular, secrecy strategies outage three of closed-form. are all","In particular, the secrecy outage probability of all three strategies are obtained in closed-form.",32721
"PCA. conventional cannot hidden However, components within data one discover using multiway","However, one cannot discover hidden components within multiway data using conventional PCA.",32722
"robust learning low-rank for tensor feature extraction, subspaces linear recovery. and","learning linear subspaces for feature extraction, and robust low-rank tensor recovery.",32723
free on this images. We from limitation by training unannotated ourselves,We free ourselves from this limitation by training on unannotated images.,32724
derived accuracy verified the simulations expressions. of MATLAB,MATLAB simulations verified accuracy of the derived expressions.,32725
"this the of problem degraded by images noise. In address Poisson paper, we denoising","In this paper, we address the problem of denoising images degraded by Poisson noise.",32726
"huge satellites leading provide Modern data. large amounts resolution high to of with coverage signals,","Modern satellites provide large coverage with high resolution signals, leading to huge amounts of data.",32727
"lack exploit still countries. expertise in developing observatories to particularly datasets, volcano satellite Moreover,","Moreover, volcano observatories still lack expertise to exploit satellite datasets, particularly in developing countries.",32728
unusual employed the within Convolutional data. networks are detect radar patterns to neural (CNN),Convolutional neural networks (CNN) are employed to detect unusual patterns within the radar data.,32729
a is challenging diversity. dataset with This good,This is a challenging dataset with good diversity.,32730
"all Dataset, code website. models publicly be the and will available source on trained","Dataset, source code and trained models will all be publicly available on the website.",32731
"neuronal for paper mechanisms firstly prediction. the describes memory search, and The","The paper firstly describes the mechanisms for neuronal search, memory and prediction.",32732
"be basis, The consistent. mathematical allowing also for rule construction a language to the has","The language also has a mathematical basis, allowing for the rule construction to be consistent.",32733
"both can hierarchies memory and tree as static structures. be built process Now, dynamic","Now, both static memory and dynamic process hierarchies can be built as tree structures.",32734
is The framework two structured phases. in,The framework is structured in two phases.,32735
were images performed sunlight. direct Experiments of unmeasured folded real papers under on,Experiments on real images of folded papers were performed under unmeasured direct sunlight.,32736
on propositional a previous work retraction where defined morphological erosion. as extends This was logics,This extends previous work on propositional logics where retraction was defined as a morphological erosion.,32737
The using learning. human detection a presents deep images still improve to in technique paper,The paper presents a technique to improve human detection in still images using deep learning.,32738
"Our saliency ViS-HuD, method, the image. novel from map computes visual","Our novel method, ViS-HuD, computes visual saliency map from the image.",32739
and has to information drift Leveraging topic. history hot correct a become relocalize,Leveraging history information to relocalize and correct drift has become a hot topic.,32740
graph Current optimization. map the map previous global be together pose can by and merged,Current map and previous map can be merged together by the global pose graph optimization.,32741
large-scale We environment. our outdoor evaluate the of in ability system map the also merging,We also evaluate the map merging ability of our system in the large-scale outdoor environment.,32742
"our integrated reuse is The source code, map into code public VINS-Mono. of","The source code of map reuse is integrated into our public code, VINS-Mono.",32743
handle (NR) no-reference are Current quality this of to unable distortions. models video diversity,Current no-reference (NR) video quality models are unable to handle this diversity of distortions.,32744
"data tested predictors As video result, video a on quality NR real-world often perform poorly.","As a result, NR video quality predictors tested on real-world video data often perform poorly.",32745
via crowdsourcing. of a video collected large We scores number subjective quality,We collected a large number of subjective video quality scores via crowdsourcing.,32746
pairs We manual hundreds supervision. videos from any obtain without using of face millions of,We obtain millions of face pairs from hundreds of videos without using any manual supervision.,32747
most in of the small. scenarios faces are very detected surveillance where,in surveillance scenarios where most of the faces detected are very small.,32748
"its computational showed method In however, performance. poor first the implementation,","In its first implementation, however, the method showed poor computational performance.",32749
and freely code data Our on are github. available,Our code and data are available freely on github.,32750
"learning introduce curriculum cross-paced partial we (CPPCL) hard), the to easy framework.","easy to hard), we introduce the cross-paced partial curriculum learning (CPPCL) framework.",32751
four extensively approach evaluated (i.e. datasets on available is Our publicly,Our approach is extensively evaluated on four publicly available datasets (i.e.,32752
exploration We the of learning. meta in problem reinforcement consider,We consider the problem of exploration in meta reinforcement learning.,32753
"bridges, conditions dynamic a need on we Under management. real-time","Under dynamic conditions on bridges, we need a real-time management.",32754
"fuzzy generation aid are the rules Then, these of rule algorithms. by generalized","Then, these rules are generalized by the aid of fuzzy rule generation algorithms.",32755
Markov each pixels between Field. modelled class The by correlation is belonging to Random spatial,The spatial correlation between pixels belonging to each class is modelled by Markov Random Field.,32756
framework alternatively. the estimate Bayesian proposed abundance A vectors is to corresponding and classes,A Bayesian framework is proposed to estimate the classes and corresponding abundance vectors alternatively.,32757
An MCMC vector the samples. estimate used is generated to algorithm abundance based on,An MCMC algorithm is used to estimate the abundance vector based on generated samples.,32758
on proposed the simulated of The result approach. of data prominence implementation shows,The result of implementation on simulated data shows the prominence of proposed approach.,32759
concept of The was cognitive introduced only. originally for radar active radar,The concept of cognitive radar was introduced originally for active radar only.,32760
In how radar. applied also passive paper this this we to be paradigm describe can,In this paper we describe how this paradigm can be applied also to passive radar.,32761
information-theoretic novel propose based supervised mutual hashing information. an We on optimizing quantity: a method,We propose a novel supervised hashing method based on optimizing an information-theoretic quantity: mutual information.,32762
to range techniques neural Existing networks entirely box black solutions. from hand-designed,Existing techniques range from black box neural networks to entirely hand-designed solutions.,32763
the the while repetitive designer This labour. in manual loop keeps reducing,This keeps the designer in the loop while reducing repetitive manual labour.,32764
"our We framework, supporting present and challenges. findings open","We present our framework, supporting findings and open challenges.",32765
The procedure automated steps. primary consists three of,The automated procedure consists of three primary steps.,32766
"are resampling for image blocks. First, features calculated","First, resampling features are calculated for image blocks.",32767
resampled. of be identified to We these some expect falsely as blocks,We expect some of these blocks to be falsely identified as resampled.,32768
"the detection very business classic researched. Within is anomaly of intelligence, not frequently domain","Within the domain of business intelligence, classic anomaly detection is not very frequently researched.",32769
of number These large reaches enough oscillations the a colony once cells. emerge,These oscillations emerge once the colony reaches a large enough number of cells.,32770
setup. Modulation is radio sensing a cognitive challenging recognition spectrum a in performing task while,Modulation recognition is a challenging task while performing spectrum sensing in a cognitive radio setup.,32771
for problem important robot-assisted instruments the robotic Semantic surgery. segmentation an is of,Semantic segmentation of robotic instruments is an important problem for the robot-assisted surgery.,32772
segmentation instrument this pixel-wise to is needed address challenge. Accurate,Accurate pixel-wise instrument segmentation is needed to address this challenge.,32773
"samples (BPTP) Technology were Bogor, Agricultural Indonesia. Institute from Assessment The obtained","The samples were obtained from Agricultural Technology Assessment Institute (BPTP) Bogor, Indonesia.",32774
Primers four to The samples. were amplify all used the of eggplant able,The Primers were used able to amplify all of the four eggplant samples.,32775
"the of showed of and cluster three clusters B, analysis (A, C). The existence result","The result of cluster analysis showed the existence of three clusters (A, B, and C).",32776
Depth computer in robotic useful tasks. and proved very vision various cues have been,Depth cues have been proved very useful in various computer vision and robotic tasks.,32777
a the from problem image. monocular still addresses of single paper depth This estimation,This paper addresses the problem of monocular depth estimation from a single still image.,32778
the is the highest considered iris with probability. as The unique biometric trait,The iris is considered as the biometric trait with the highest unique probability.,32779
has vision widely in learning been a promising different applications. as approach computer Deep recognized,Deep learning has been widely recognized as a promising approach in different computer vision applications.,32780
State-of-the-art been showed dataset. has on the result DETRAC performance vehicle,State-of-the-art performance result has been showed on the DETRAC vehicle dataset.,32781
"images, demonstrate species it, using recognition the learning. With bark of we feasibility through deep","With it, we demonstrate the feasibility of species recognition through bark images, using deep learning.",32782
eavesdropping the by best the eavesdropper. which selected relay maximizes is The rate,The best relay which maximizes the eavesdropping rate is selected by the eavesdropper.,32783
option for of is appealing the vehicles. autonomous safety validating an Simulation,Simulation is an appealing option for validating the safety of autonomous vehicles.,32784
Generative shown (GAIL) to Imitation Learning models. recently driver Adversarial representative been has learn human,Generative Adversarial Imitation Learning (GAIL) has recently been shown to learn representative human driver models.,32785
extend address approach grounded parameter-sharing GAIL in a to We shortcomings these through learning. curriculum,We extend GAIL to address these shortcomings through a parameter-sharing approach grounded in curriculum learning.,32786
is known widely regression(GWR) that model. same as varying-coefficient It is essentially geographically weighted,It is widely known that geographically weighted regression(GWR) is essentially same as varying-coefficient model.,32787
"of GWR estimation word, In other is distance-kernel-based.","In other word, estimation of GWR is distance-kernel-based.",32788
"design and obtained. scale selection also relationship optimal parameters between is Furthermore, of bandwith a","Furthermore, a relationship between optimal bandwith selection and design of scale parameters is also obtained.",32789
is show that this not universal. We notion,We show that this notion is not universal.,32790
not and pixel disparity introduce We formulation a new which does stereo models. use,We introduce a new stereo formulation which does not use pixel and disparity models.,32791
are in Many problems each treated a vision pixel label. assigning as,Many problems in vision are treated as assigning each pixel a label.,32792
are for methods solving of the one Graph cuts energy cient minimization. effi-,Graph cuts are one of the effi- cient methods for solving energy minimization.,32793
"truncated etc. terms, Thus, linear quadratic smoothness or","Thus, truncated linear or quadratic smoothness terms, etc.",32794
"used, where approximate is necessary. minimization energy are","are used, where approximate energy minimization is necessary.",32795
"is the reasoning naturally this visibility energy For function. included formulation, in","For this formulation, visibility reasoning is naturally included in the energy function.",32796
images cloud for difficult. This segmentation such renders,This renders cloud segmentation for such images difficult.,32797
"compensated the artificial supplementary effects First, the lighting. is by ill of algorithmically illumination emulating","First, the ill illumination is algorithmically compensated by emulating the effects of artificial supplementary lighting.",32798
caused trained a DCNN by only Then data using missing detail recovers synthetic quantization. the,Then a DCNN trained using only synthetic data recovers the missing detail caused by quantization.,32799
"always the of channel. characteristics This is on depending task not the trivial,","This task is not always trivial, depending on the characteristics of the channel.",32800
to heuristic proposed problem. the A solve procedure is,A heuristic procedure is proposed to solve the problem.,32801
players pricing A for strategy the proposed. supply chain is data the in win-win,A win-win pricing strategy for the players in the data supply chain is proposed.,32802
in object by large-scale deep mainly detection advances Recent with learning benchmarks. driven detection are,Recent advances in object detection are mainly driven by deep learning with large-scale detection benchmarks.,32803
The described follows. as are contributions main,The main contributions are described as follows.,32804
LSTD The is results detector demonstrate scenarios. that for low-shot deep a preferable,The results demonstrate that LSTD is a preferable deep detector for low-shot scenarios.,32805
circuit MATLAB The design. was used the the Simulink tool analysis of for,The MATLAB Simulink tool was used for the analysis of the circuit design.,32806
"advanced to an comparable analysis. Simulink, accurate tool gives results real-time","Simulink, an advanced tool gives accurate results comparable to real-time analysis.",32807
propagates importance. that neural in networks is great way The information of,The way that information propagates in neural networks is of great importance.,32808
"overhead. simple with extra These to improvements subtle are implement, computational","These improvements are simple to implement, with subtle extra computational overhead.",32809
state-of-the-art also Cityscapes. It on is and MVD,It is also state-of-the-art on MVD and Cityscapes.,32810
a PickNet in picking propose frame plug-and-play captioning. informative to video We perform,We propose a plug-and-play PickNet to perform informative frame picking in video captioning.,32811
procedure until the end goes of sequence. on the video This,This procedure goes on until the end of the video sequence.,32812
the down-converted deviations. baseband reference Sine for signals analysis digitally to are phase of,Sine reference signals are digitally down-converted to baseband for the analysis of phase deviations.,32813
depth monocular and regress Previous estimation expected the a methods single view take results. directly,Previous monocular depth estimation methods take a single view and directly regress the expected results.,32814
model generalizes other to depth well monocular benchmarks. estimation also The,The model also generalizes well to other monocular depth estimation benchmarks.,32815
segmentation task nuclei paper in histopathological of addresses images. high-resolution This the,This paper addresses the task of nuclei segmentation in high-resolution histopathological images.,32816
segmentation of for show data effectiveness methods also augmentation the task. nuclei We,We also show the effectiveness of data augmentation methods for nuclei segmentation task.,32817
state-of-the- outperforms Our experiments our method methods. art showed prior,Our experiments showed our method outperforms prior state-of-the- art methods.,32818
precisely image segment it . time in whole-slide possible acceptable makes This to the,This makes it possible to precisely segment the whole-slide image in acceptable time .,32819
Breast worldwide. in the cancer cancer most common is women,Breast cancer is the most common cancer in women worldwide.,32820
is common technology The most mammography. screening,The most common screening technology is mammography.,32821
convolutional neural State-of-the-art their performance on deep the networks compared abnormalities. of are classifying,State-of-the-art deep convolutional neural networks are compared on their performance of classifying the abnormalities.,32822
"with Compared of efficient forms image other labels supervision, to obtain. quite are weak","Compared with other forms of weak supervision, image labels are quite efficient to obtain.",32823
"on work, supervised label image the segmentation we semantic weakly with annotations. In focus our","In our work, we focus on the weakly supervised semantic segmentation with image label annotations.",32824
quality been generated Recent this progress largely the for task of has dependent on pseudo-annotations.,Recent progress for this task has been largely dependent on the quality of generated pseudo-annotations.,32825
"the In paper, of we classification facial images in deep this with gender learning. tackle","In this paper, we tackle the classification of gender in facial images with deep learning.",32826
CNNs The gender with method Adience LFW the annotations. our is and of fine-tuned on,The CNNs of our method is fine-tuned on the Adience and LFW with gender annotations.,32827
tasks A other object-related of host (e.g.,A host of other object-related tasks (e.g.,32828
protein-expression optimal generates of ordering analysis The rank combinatorial coherence. an,The combinatorial analysis generates an optimal rank ordering of protein-expression coherence.,32829
also experiment. all from these This observations scattered means TMA be deduced can one,This means all these scattered observations can also be deduced from one TMA experiment.,32830
dependency TMA data. presented The gives information tool a extract algorithm to from pathologists molecular,The presented algorithm gives molecular pathologists a tool to extract dependency information from TMA data.,32831
reconstruct networks. to algorithms dependency foster certainly the refinement This will of,This will certainly foster the refinement of algorithms to reconstruct dependency networks.,32832
autonomous applications driving. robotics segmentation especially benefits related Semantic,Semantic segmentation benefits robotics related applications especially autonomous driving.,32833
It is meta-architectures multiple for methodology. decoding segmentation that also comprised of the define,It is also comprised of multiple meta-architectures for segmentation that define the decoding methodology.,32834
"and These include SkipNet, Dilation Frontend. UNet,","These include SkipNet, UNet, and Dilation Frontend.",32835
results for scenes. urban are presented the on Experimental dataset Cityscapes,Experimental results are presented on the Cityscapes dataset for urban scenes.,32836
paper algorithm. a framework multi-agent proposes well intent-aware learning an as This as planning,This paper proposes an intent-aware multi-agent planning framework as well as a learning algorithm.,32837
"maximize agent goal this plans to space expected in Under the framework, an the utility.","Under this framework, an agent plans in the goal space to maximize the expected utility.",32838
into The the takes other belief of planning consideration. agents' intents process,The planning process takes the belief of other agents' intents into consideration.,32839
It can easily and prediction ii) integrate algorithms. intent planning existing low-level,ii) It can easily integrate existing intent prediction and low-level planning algorithms.,32840
It space. the action does suffer not in from sparse iii) feedbacks,iii) It does not suffer from sparse feedbacks in the action space.,32841
an surface important is urban monitoring. Impervious for development indicator,Impervious surface is an important indicator for urban development monitoring.,32842
and Observation resolutions. sensing spatial systems temporal provide data Earth Modern different at,Modern Earth Observation systems provide sensing data at different temporal and spatial resolutions.,32843
"First, as LDA the with I explore using convolutional of activations features. layers","First, I explore using LDA with the activations of convolutional layers as features.",32844
"and distinctly clusters actual yelp align these Furthermore, photo well labels. the with","Furthermore, these clusters align well and distinctly with the actual yelp photo labels.",32845
"matrix. statistic on traditional of distance most based and spatial the methods Additionally, weighted are","Additionally, most of the traditional methods are based on distance statistic and spatial weighted matrix.",32846
"weight bounded not Specifically, this or with is matrix. distance method statistic","Specifically, this method is not bounded with distance statistic or weight matrix.",32847
compare and also We see and RNNPlanner their disadvantages. DUP advantages to,We also compare DUP and RNNPlanner to see their advantages and disadvantages.,32848
for particles detection investigate urine methods factors recognition. different CNN-based We involving object further these,We further investigate different factors involving these CNN-based object detection methods for urine particles recognition.,32849
for theory We dimensionality reduction (NLDR). nonlinear develop,We develop theory for nonlinear dimensionality reduction (NLDR).,32850
is There using NLDR deriving basis new algorithms. for existing limited theory for,There is limited basis for using existing NLDR theory for deriving new algorithms.,32851
us ones. new methods allows derive understand both This to existing and,This allows us to both understand existing methods and derive new ones.,32852
image X-ray (e.g. of The archives availability increased,The increased availability of X-ray image archives (e.g.,32853
"(i.e. ResNet multiple In a we experiment, also investigate concluding depths","In a concluding experiment, we also investigate multiple ResNet depths (i.e.",32854
deep are unified a The learning via jointly processes two scheme. end-to-end optimized above,The above two processes are jointly optimized via a unified end-to-end deep learning scheme.,32855
future(MMPPF) past The CTTC formulated mathematically multi-optional the present structures. using is many-sorted,The CTTC is formulated mathematically using the multi-optional many-sorted past present future(MMPPF) structures.,32856
processes discussed CTTC. decision-making This article the are how described in,This article discussed how decision-making processes are described in the CTTC.,32857
"operations from In are make feature HENet. removed use to maps, pooling of order greater","In order to make greater use of feature maps, pooling operations are removed from HENet.",32858
radio astronomy. (RFI) frequency Transient radio is detrimental interference to,Transient radio frequency interference (RFI) is detrimental to radio astronomy.,32859
mitigating One RFI stations independent monitoring telescope to approach at is deploy radio arrays.,One mitigating approach is to deploy independent RFI monitoring stations at radio telescope arrays.,32860
enhanced systems. controllability smart technologies of distribution and of Proliferation has observability grid,Proliferation of smart grid technologies has enhanced observability and controllability of distribution systems.,32861
"systems transmission more be can system, If of resources coordinated both the efficiently. with used","If coordinated with the transmission system, resources of both systems can be used more efficiently.",32862
distribution and systems a coordinated This in proposes a model transmission manner. paper operate to,This paper proposes a model to operate transmission and distribution systems in a coordinated manner.,32863
using (SLR) model solved Surrogate Lagrangian approach. Relaxation proposed a is The,The proposed model is solved using a Surrogate Lagrangian Relaxation (SLR) approach.,32864
existing performance (e.g. this compared methods is approach computational The of against,The computational performance of this approach is compared against existing methods (e.g.,32865
"fully the this images. address mutual not approach However, does and dependencies high-resolution of low-","However, this approach does not fully address the mutual dependencies of low- and high-resolution images.",32866
CPU used for to ORB- global Idled pose. SLAM providing is robust perform,Idled CPU is used to perform ORB- SLAM for providing robust global pose.,32867
CPU modules from and are GPU. to integrate taken Strategies,Strategies are taken to integrate modules from CPU and GPU.,32868
vivid observation deforms while model It transforms current and fuses incrementally and keeping texture. new,It transforms and deforms current model and incrementally fuses new observation while keeping vivid texture.,32869
"of this, discretely- closed form Besides distributed derived IEP case. novel is a in expression","Besides this, a novel closed form of IEP expression is derived in discretely- distributed case.",32870
the finally effectiveness verify of Simulation architecture. results proposed the,Simulation results finally verify the effectiveness of the proposed architecture.,32871
for domains. critical Ontologies information sources semantic are of application many,Ontologies are critical sources of semantic information for many application domains.,32872
"to understandability its the is First, coverage. and restructured increase ontology","First, the ontology is restructured to increase its understandability and coverage.",32873
Convolutional neural tools for networks classification. and image are segmentation powerful,Convolutional neural networks are powerful tools for image segmentation and classification.,32874
can automatically and with regions. segmented accurately Various heart dynamical be morphological parameters quantified cardiac,Various morphological and dynamical cardiac parameters can be quantified accurately with automatically segmented heart regions.,32875
"was, large requires of generalization shown that sufficient trajectories. however, relatively number It learning","It was, however, shown that sufficient generalization requires relatively large number of learning trajectories.",32876
The to countermeasure discusses problem. paper overcome possible this,The paper discusses possible countermeasure to overcome this problem.,32877
analytical The accuracy of confirm simulation the derived expressions. results,The simulation results confirm the accuracy of derived analytical expressions.,32878
Decision Process on approach (MDP). Markov Our based is,Our approach is based on Markov Decision Process (MDP).,32879
spot This conventional generalizes the techniques. metering adaptive matrix or metering module,This adaptive metering module generalizes the conventional spot or matrix metering techniques.,32880
data sets. is network on validated real proposed multispectral The,The proposed network is validated on real multispectral data sets.,32881
proposed visual performance of demonstrates mode. the analysis and in experimental competitive Both quantitative results,Both visual and quantitative analysis of experimental results demonstrates competitive performance in the proposed mode.,32882
our the approach. of results are Numerical to demonstrate conducted effectiveness,Numerical results are conducted to demonstrate the effectiveness of our approach.,32883
of and monitor for the important to sake analyze safety. events It is city crowd,It is important to monitor and analyze crowd events for the sake of city safety.,32884
DigCrowd effectiveness method the Experimental results our demonstrate . of,Experimental results demonstrate the effectiveness of our DigCrowd method .,32885
fused re-identification representations the correspondence are task. perform to These,These correspondence representations are fused to perform the re-identification task.,32886
"end-to-end Further, scheme. the framework is via deep unified a optimized proposed learning","Further, the proposed framework is optimized via a unified end-to-end deep learning scheme.",32887
effective sensor arrays Sparse cost alternative offer uniform arrays. to a,Sparse sensor arrays offer a cost effective alternative to uniform arrays.,32888
This planar with paper proposes sparse configuration array inter-element novel few a spacings. unit,This paper proposes a novel sparse planar array configuration with few unit inter-element spacings.,32889
"is built, the binary of encoders. end-to-end are An two as architecture which three-network treated","An end-to-end three-network architecture is built, two of which are treated as the binary encoders.",32890
The remarkably related works. model outperforms proposed,The proposed model remarkably outperforms related works.,32891
"needed misclassify few) to them. these no perturbations are for Thus, (or frames extra","Thus, no (or few) extra perturbations are needed for these frames to misclassify them.",32892
"the procedure, different Performance scenarios. optimization are and presented for with results, without","Performance results, with and without the optimization procedure, are presented for different scenarios.",32893
the most common is (DPN) diabetic one peripheral neuropathy of Diabetic complications.,Diabetic peripheral neuropathy (DPN) is one of the most common diabetic complications.,32894
and unclear. damage nerve large underlying remain following dysfunction Mechanisms sensory metabolic loss,Mechanisms underlying nerve damage and sensory loss following metabolic dysfunction remain large unclear.,32895
lead significant a distribution drop. a will to mismatch Such performance,Such a distribution mismatch will lead to a significant performance drop.,32896
"In work, this the improve robustness we object to of aim detection. cross-domain","In this work, we aim to improve the cross-domain robustness of object detection.",32897
the Species by impacted same invasive taxonomic are in a species. group virtual similarly,Species in the same taxonomic group are similarly impacted by a virtual invasive species.,32898
"a occur. can within group given interesting differences taxonomic However,","However, interesting differences within a given taxonomic group can occur.",32899
implications these for ecologists discussed. and land are The results managers of,The implications of these results for ecologists and land managers are discussed.,32900
kernel We landscape. density to estimate approach the apply persistence,We apply kernel density approach to estimate the persistence landscape.,32901
"several our model. algorithms In we step, to next provide solve the","In the next step, we provide several algorithms to solve our model.",32902
chirp symbols orthogonal information UW-A for We consider M-ary over proposed waveforms are carried transmissions.,We consider information symbols are carried over proposed M-ary orthogonal chirp waveforms for UW-A transmissions.,32903
"develop an Moreover, optimal proposed a waveforms. we non-coherent coherent based chirp receivers on and","Moreover, we develop a coherent and an optimal non-coherent receivers based on proposed chirp waveforms.",32904
is experiments. evaluated Performance chirp water in waveforms M-ary tank orthogonal of,Performance of M-ary orthogonal chirp waveforms is evaluated in water tank experiments.,32905
an expression data-based tumor Microarray issue. classification and is active challenging gene,Microarray gene expression data-based tumor classification is an active and challenging issue.,32906
"e.g. images, that form We learn generator of the invertable conditions an can trans-","We learn an invertable generator that can trans- form the conditions of images, e.g.",32907
"etc. summer from night, to winter day to","from day to night, summer to winter etc.",32908
is semantic information achieved like associating their with attributes. by usually This categories,This is usually achieved by associating categories with their semantic information like attributes.,32909
"this fully not exploited. offered yet the believe paradigm that potential we However, is by","However, we believe that the potential offered by this paradigm is not yet fully exploited.",32910
generate in task always Learning challenging computer vision. to natural a has been scenes,Learning to generate natural scenes has always been a challenging task in computer vision.,32911
single architecture and discriminator X-Fork has single generator. a a,X-Fork architecture has a single discriminator and a single generator.,32912
generator The both and in image the semantic hallucinates target its view. segmentation the,The generator hallucinates both the image and its semantic segmentation in the target view.,32913
generate cGAN the the sharper feedback from helps The first second cGAN images.,The feedback from the second cGAN helps the first cGAN generate sharper images.,32914
The problem power thus optimization solved is convex as classic a allocation problem.,The power allocation problem is thus solved as a classic convex optimization problem.,32915
pre-trained the discuss generic for that also uses SLADS-Net We a training. of development images,We also discuss the development of a pre-trained SLADS-Net that uses generic images for training.,32916
to problems. propose Bayesian for extensions We dynamic solving optimization practical,We propose practical extensions to Bayesian optimization for solving dynamic problems.,32917
"our problems. synthetic real-world Lastly, on evaluate technique we and","Lastly, we evaluate our technique on synthetic and real-world problems.",32918
"the this that is is said of data It ""dimension"" reduced.","It is said that the ""dimension"" of this data is reduced.",32919
"the Despite BTD GPR-based literature. large-scale in rare comparisons algorithms of this, are","Despite this, large-scale comparisons of GPR-based BTD algorithms are rare in the literature.",32920
five of effort algorithms. GPR-based experience the The BTD involved with with development substantial institutions,The effort involved five institutions with substantial experience with the development of GPR-based BTD algorithms.,32921
"noise the label supervision. to tackle more we leveraging Instead, category-level propose accessible by","Instead, we propose to tackle the label noise by leveraging more accessible category-level supervision.",32922
"its Furthermore, than may make real-valued counterparts. the nature better it generalize binary","Furthermore, the binary nature may make it generalize better than its real-valued counterparts.",32923
"datasets, image benchmark outperforms retrieval also existing methods. On hashing ABC","On benchmark image retrieval datasets, ABC also outperforms existing hashing methods.",32924
"atlas, to further facilitate There a of development i.e. a is need the digital","There is a further need to facilitate the development of a digital atlas, i.e.",32925
"slices Currently, of registration manual of landmarks. requires identification","Currently, registration of slices requires manual identification of landmarks.",32926
registration basis atlas These as will serve the results for and building.,These results will serve as the basis for registration and atlas building.,32927
significant robotics and related mobile is semantic Real-time applications. of segmentation importance for,Real-time semantic segmentation is of significant importance for mobile and robotics related applications.,32928
term which We we efficient segmentation propose ShuffleSeg. network as computationally a,We propose a computationally efficient segmentation network which we term as ShuffleSeg.,32929
on tradeoff accuracy speed Interesting discussed. the insights and is,Interesting insights on the speed and accuracy tradeoff is discussed.,32930
neural large-margin for in deep We Gaussian propose networks a loss tasks. (L-GM) Mixture classification,We propose a large-margin Gaussian Mixture (L-GM) loss for deep neural networks in classification tasks.,32931
"this In kernel-based broadly filtering we (e.g. work, connect","In this work, we broadly connect kernel-based filtering (e.g.",32932
the central surgery. is Tracking a image-guided pose instruments in problem of,Tracking the pose of instruments is a central problem in image-guided surgery.,32933
"coherence used scenarios, increasingly modality. is microscopic optical (OCT) an as For imaging tomography","For microscopic scenarios, optical coherence tomography (OCT) is increasingly used as an imaging modality.",32934
data a We OCT deep from with pose estimation address learning-based tracking volume framework. new,We address pose estimation from OCT volume data with a new deep learning-based tracking framework.,32935
deep label's We show learning our our that resolution. errors at ground-truth reaches approach,We show that our deep learning approach reaches errors at our ground-truth label's resolution.,32936
a pose fiducial estimation. marker by used for,used by a fiducial marker for pose estimation.,32937
"we versus comparison control of Finally, the of a points. present configuration number optimized","Finally, we present a comparison of optimized configuration versus the number of control points.",32938
in in environment. These the a process dynamic are designing parameters important,These parameters are important in designing the process in a dynamic environment.,32939
Modeling coalition the forming of link on of the activation/inactivation the process. new effect d),d) Modeling of the effect of the new link activation/inactivation on the coalition forming process.,32940
of e) the Modeling the link on process. coalition-forming mobility effect,e) Modeling the effect of link mobility on the coalition-forming process.,32941
people fashion clothing common other in share aesthetics each The with the dressing. reflects that,The clothing fashion reflects the common aesthetics that people share with each other in dressing.,32942
effectiveness demonstrate of body the results proposed segmentation The classification and methods. the fashion,The results demonstrate the effectiveness of the proposed body segmentation and fashion classification methods.,32943
network and them. sequence-to-sequence A to language encode processing this query symbol structures learns,A sequence-to-sequence network processing this language learns to encode symbol structures and query them.,32944
We significance then study the functional the SCCs. of,We then study the functional significance of the SCCs.,32945
egocentric have object action works A concentrated vision on and in recognition. of large number,A large number of works in egocentric vision have concentrated on action and object recognition.,32946
"Detection segmentation has less videos, however, first-person and been explored. of hands in","Detection and segmentation of hands in first-person videos, however, has less been explored.",32947
"an problem. in-depth Here, hand look we at the segmentation take","Here, we take an in-depth look at the hand segmentation problem.",32948
collected laboratory settings. segmentation in are hand the datasets Existing,Existing hand segmentation datasets are collected in the laboratory settings.,32949
can refine segmentations. generated fields help We whether hand explore conditional random further,We further explore whether conditional random fields can help refine generated hand segmentations.,32950
"we stochastic layers. introduce different variance layers, In this a of paper, kind","In this paper, we introduce variance layers, a different kind of stochastic layers.",32951
"learning and complete novel, deep We framework multi-person tracking. localisation present a for","We present a novel, complete deep learning framework for multi-person localisation and tracking.",32952
data Graph (GSP) on that described graphs. processing kernels irregular live signals studies signal by,Graph signal processing (GSP) studies signals that live on irregular data kernels described by graphs.,32953
"Chebyshev matrix using the ideal a polynomial. approximate filter we Finally,","Finally, we approximate the ideal filter using a Chebyshev matrix polynomial.",32954
iteratively a design objective. greedy algorithm simplified the We to minimize,We design a greedy algorithm to iteratively minimize the simplified objective.,32955
comprehensive paper this to survey topic. a background This research provides,This survey paper provides a comprehensive background to this research topic.,32956
"and practical analysis. no not This is there tutorial, code, little a is data computer","This is not a practical tutorial, there is little computer code, and no data analysis.",32957
"we study Furthermore, the square the error nonuniform of sampling. mean","Furthermore, we study the mean square error of the nonuniform sampling.",32958
"we theoretical to do of results. simulations verify our some correctness Finally, the","Finally, we do some simulations to verify the correctness of our theoretical results.",32959
automated driving participants. precise requires of traffic models Highly,Highly automated driving requires precise models of traffic participants.,32960
models of learning art on state are Many based machine techniques. the currently,Many state of the art models are currently based on machine learning techniques.,32961
"others, is the required one data Among challenge. of labeled major amount","Among others, the required amount of labeled data is one major challenge.",32962
this addressing problem autonomous learning is An proposed. process,An autonomous learning process addressing this problem is proposed.,32963
"Recent e.g. either assumptions, make oversimplifying approaches","Recent approaches either make oversimplifying assumptions, e.g.",32964
"method encourages a preservation. Therefore, that identity propose we data-driven","Therefore, we propose a data-driven method that encourages identity preservation.",32965
of our experimental demonstrates method. superiority The evaluation the,The experimental evaluation demonstrates the superiority of our method.,32966
cloud building SO-Net models (SOM). Map point the by The of a spatial distribution Self-Organizing,The SO-Net models the spatial distribution of point cloud by building a Self-Organizing Map (SOM).,32967
Our code is at available the project website.,Our code is available at the project website.,32968
cooperation user in of performance makes improvement of systems. CDMA Use the much,Use of user cooperation makes much improvement in the performance of CDMA systems.,32969
"the for of Moreover, the novel calls stream asynchronous algorithms. event nature","Moreover, the asynchronous nature of the event stream calls for novel algorithms.",32970
"This object asynchronous with a cameras. paper approach presents tracking to new, efficient","This paper presents a new, efficient approach to object tracking with asynchronous cameras.",32971
"process and are Stochastic biological, used models ecological now industrial analyse commonly complex systems. to","Stochastic process models are now commonly used to analyse complex biological, ecological and industrial systems.",32972
parallel is ideal multi-core It from for computers. embarrassingly and benefits running being on,It benefits from being embarrassingly parallel and is ideal for running on multi-core computers.,32973
"more the the is plausible. As result, visibly inpainting","As the result, the inpainting is visibly more plausible.",32974
running framework. RANSAC solvers a These methods minimal on rely robust in often,These robust methods often rely on running minimal solvers in a RANSAC framework.,32975
"bases based been polynomial on have Gr\""obner a ideal. monomial basis the for traditionally These","These monomial bases have traditionally been based on a Gr\""obner basis for the polynomial ideal.",32976
bases enumerate in Here describe way. an we can how efficient all such we,Here we describe how we can enumerate all such bases in an efficient way.,32977
a problems. a evaluate number scheme on basis that We we novel of present sampling,We present a novel basis sampling scheme that we evaluate on a number of problems.,32978
"model settings, formulated. parametric In ODE cannot these be","In these settings, parametric ODE model cannot be formulated.",32979
"approach of an paper, proposed more images. In efficient to collect texture this is databases","In this paper, an approach is proposed to collect more efficient databases of texture images.",32980
The approach proposed included stages. is two,The proposed approach is included two stages.,32981
"model the In theory competitive problem, order game the nature of to used. is","In order to model the competitive nature of the problem, game theory is used.",32982
"in context the cooperative multicarrier communication efficiently optimized is investigated and networks. SWIPT Moreover, of","Moreover, SWIPT is investigated and efficiently optimized in the context of multicarrier cooperative communication networks.",32983
both existing dataset the on Our methods based tasks. learning KITTI outperforms method driving in,Our method outperforms existing learning based methods on the KITTI driving dataset in both tasks.,32984
"Visual limited on available (VAMs). Range are (HDR) However, High Dynamic Attention Models studies","However, limited studies are available on High Dynamic Range (HDR) Visual Attention Models (VAMs).",32985
method proposed existing superiority evaluations fusion the of other the Performance fusion demonstrate against methods.,Performance evaluations demonstrate the superiority of the proposed fusion method against other existing fusion methods.,32986
are the timestamps events The occurrence exact unknown. of,The exact occurrence timestamps of the events are unknown.,32987
Our outperforms and real synthetic on classical sets. methods both algorithm count three panel,Our algorithm outperforms classical methods on both synthetic and three real panel count sets.,32988
to improve aspect a generation. way propose caption of this We,We propose a way to improve this aspect of caption generation.,32989
over methods performance detection the proposed the evaluate of saliency the existing We dataset.,We evaluate the performance of the existing saliency detection methods over the proposed dataset.,32990
the This transparent of object addresses problem matting. paper,This paper addresses the problem of transparent object matting.,32991
"In paper, object first we this transparent a formulate as matting problem. estimation refractive flow","In this paper, we first formulate transparent object matting as a refractive flow estimation problem.",32992
"propose learning flow. then for learning framework, TOM-Net, a called refractive We deep the","We then propose a deep learning framework, called TOM-Net, for learning the refractive flow.",32993
characteristics. learning image at dictionary multiple coherent capture seeks Deep to dictionaries different scales complementary,Deep dictionary learning seeks multiple dictionaries at different image scales to capture complementary coherent characteristics.,32994
"in rate adversarial proposed presence The fooling of shows remarkably furthermore, lower perturbation. algorithm","The proposed algorithm furthermore, shows remarkably lower fooling rate in presence of adversarial perturbation.",32995
the Information decoder. Side one components of the of key is Wyner-Ziv,Side Information is one of the key components of the Wyner-Ziv decoder.,32996
generation better side result coder. in of functionality will Better information Wyner-Ziv,Better side information generation will result in better functionality of Wyner-Ziv coder.,32997
based We criterion. as quality system) image quality used metrics our HVS visual have (human,We have used HVS (human visual system) based image quality metrics as our quality criterion.,32998
a that popular KAC two provides on Experiments improvement datasets. show Net significant,Experiments show that KAC Net provides a significant improvement on two popular datasets.,32999
organizations algorithms are mission-critical Yet settings. AI in deploying many,Yet organizations are deploying AI algorithms in many mission-critical settings.,33000
the main Terrains electronic are game. of part an,Terrains are the main part of an electronic game.,33001
"techniques effort terrains. used development, procedural game reduce To synthetic to on generate human are","To reduce human effort on game development, procedural techniques are used to generate synthetic terrains.",33002
trivial task. terrain a not is However a rendering,However rendering a terrain is not a trivial task.,33003
must for gaming. Their optimal rendering techniques be,Their rendering techniques must be optimal for gaming.,33004
"scale for planetary must and which conversion. terrains, Specially account precision","Specially planetary terrains, which must account for precision and scale conversion.",33005
fit to planetary models are Multi-resolution best terrains.,Multi-resolution models are best fit to planetary terrains.,33006
decrease can change of without quality. An his any noticing point in observer view visual,An observer can change his point of view without noticing any decrease in visual quality.,33007
describe to processes. biological are complex models kinetic Stochastic often used,Stochastic kinetic models are often used to describe complex biological processes.,33008
"a measurements only at However, few relatively are practice, time-points. in taken","However, in practice, measurements are taken only at a relatively few time-points.",33009
problematic. This inference the makes more task even,This makes the inference task even more problematic.,33010
The correlating scores subjective evaluated human measures objective their ratings. assessment by are quality with,The quality assessment measures are evaluated by correlating their objective scores with human subjective ratings.,33011
RBQI approaches. proposed the results outperforms correlation the existing all The that show,The correlation results show that the proposed RBQI outperforms all the existing approaches.,33012
"small data. sample performs SRC not very However, for well","However, SRC performs not very well for small sample data.",33013
evaluation performance. test present a results We of and number analyze overall subjects for,We present evaluation results for a number of test subjects and analyze overall performance.,33014
our methods. effectiveness are employed showcase simulations to of Numerical the,Numerical simulations are employed to showcase the effectiveness of our methods.,33015
Results revealed higher clearly prefer that frame rates. subjects,Results revealed that subjects clearly prefer higher frame rates.,33016
"In this DVS Recognition the for Human we (HAR). Activity using paper, explore feasibility of","In this paper, we explore the feasibility of using DVS for Human Activity Recognition (HAR).",33017
pre-clinical analyzed vitro Selected be hits a further trial. through and examined can in,Selected hits can be further analyzed in vitro and examined through a pre-clinical trial.,33018
algorithm correctness The theoretical their intensities by simulated with moments. verifying the shown is,The algorithm correctness is shown by verifying the simulated intensities with their theoretical moments.,33019
data. linear mixed are (SGLMMs) models analyzing Spatial for spatial popular non-Gaussian generalized,Spatial generalized linear mixed models (SGLMMs) are popular for analyzing non-Gaussian spatial data.,33020
desirable choices present sensible functions possess We properties. parametric link which of,We present sensible choices of parametric link functions which possess desirable properties.,33021
Laplace estimator. GIS the choosing multiple densities importance for We use approximation the in,We use Laplace approximation for choosing the multiple importance densities in the GIS estimator.,33022
simulated using data methodology examples. is The and real proposed illustrated,The proposed methodology is illustrated using simulated and real data examples.,33023
Motion video is estimation the of all in the procedures one encoders. important,Motion estimation is one of the important procedures in the all video encoders.,33024
producers and thus non-producers constitute and between interactions can public good. a Pyoverdine mediates,Pyoverdine thus mediates interactions between producers and non-producers and can constitute a public good.,33025
"increase at of the saturating need transient. Because be concentrations, declines the benefit pyoverdine only","Because the benefit of pyoverdine declines at saturating concentrations, the increase need only be transient.",33026
"to generalized a weighting more combines MIS different sampling GMIS extending framework. functions, and techniques","GMIS combines different sampling techniques and weighting functions, extending MIS to a more generalized framework.",33027
"powerful, very but are GANs perfect. not","GANs are very powerful, but not perfect.",33028
train the often to results artifacts. are hard suffer and from still They,They are hard to train and the results still often suffer from artifacts.,33029
scene method segmentation. a RNNs proposes for paper Multimodal new RGB-D This semantic called,This paper proposes a new method called Multimodal RNNs for RGB-D scene semantic segmentation.,33030
baselines proposed to efficiently We designed carefully model examine our structure. various,We carefully designed various baselines to efficiently examine our proposed model structure.,33031
trans-disease therapeutic opportunities. CD-targeting antibodies monoclonal (mAB) rich repositioning afford,CD-targeting therapeutic monoclonal antibodies (mAB) afford rich trans-disease repositioning opportunities.,33032
logistic expression CD-associated then We profiles analyses. enrichment regression regularized using and compiled pathway,We then compiled CD-associated expression profiles using regularized logistic regression and pathway enrichment analyses.,33033
co-exist. challenging instances The segmentation can of problem object video when extremely become multiple,The problem of video object segmentation can become extremely challenging when multiple instances co-exist.,33034
of also are actively these working technologies. towards the MPEG/ITU standardization,MPEG/ITU are also actively working towards the standardization of these technologies.,33035
"the capture, taken videos. procedure the calibrate, explain and We to post-process","We explain the procedure taken to capture, calibrate, and post-process the videos.",33036
images generative Our face style-aggregated original transforms adversarial method by images to a module.,Our method transforms original face images to style-aggregated images by a generative adversarial module.,33037
paper breast explores problem the classification This images. of of microscopy tissue,This paper explores the problem of breast tissue classification of microscopy images.,33038
"we techniques deep learning suitable a training Given to the address utilize classification problem. dataset,","Given a suitable training dataset, we utilize deep learning techniques to address the classification problem.",33039
short We used connections side-output transfer information semantic layers. between to,We used short connections to transfer semantic information between side-output layers.,33040
"experiments. better our than cross-training behaves methods model other in Moreover,","Moreover, our model behaves better than other methods in cross-training experiments.",33041
vessel public performance BTS-DSN task segmentation Conclusions: three achieves on datasets. competitive in,Conclusions: BTS-DSN achieves competitive performance in vessel segmentation task on three public datasets.,33042
set a data different from We Kaggle. EyePACS used,We used a different EyePACS data set from Kaggle.,33043
Hyper-parameter in original the not settings described were study.,Hyper-parameter settings were not described in the original study.,33044
these were later some But of published.,But some of these were later published.,33045
study. not original able the to We were replicate,We were not able to replicate the original study.,33046
evaluated decision is given error tracking rate. channel error method for this In,In this method channel tracking error is evaluated for given decision error rate.,33047
"Then, decision rate error. approximated error channel given the tracking is for","Then, the decision error rate is approximated for given channel tracking error.",33048
"a can the $T$. obtain we of As result, shape","As a result, we can obtain the shape of $T$.",33049
a derived this work. is model in Such,Such a model is derived in this work.,33050
model The only shot attributed is considering our to deviation noise.,The deviation is attributed to our model considering only shot noise.,33051
"Thereby, over one approach evaluating with acquisitions different metastases. the of hundred","Thereby, evaluating the approach with over one hundred different acquisitions of metastases.",33052
assessed algorithm before or two acquisitions the never the evaluation. the physicians The had,The two physicians or the algorithm had never assessed the acquisitions before the evaluation.,33053
"a experts in to ninety segmentation results. satisfied of percent algorithmic the up As result,","As a result, the experts satisfied in up to ninety percent of algorithmic segmentation results.",33054
"a we adapt cells pacemaker recent model et In of study, al. (Severi this","In this study, we adapt a recent model of pacemaker cells (Severi et al.",33055
"the electrical predicts of Further, reduced irregular parameter model regimes phase dynamics.","Further, the reduced phase model predicts parameter regimes of irregular electrical dynamics.",33056
"observed. is temperature dependence activation a strong Instead,","Instead, a strong activation temperature dependence is observed.",33057
and chaotic nonlinear of Previous signal. nature researches revealed the EEG,Previous researches revealed the chaotic and nonlinear nature of EEG signal.,33058
Generative (GANs) plausible images. low-dimensional latent transform visually vectors into networks adversarial,Generative adversarial networks (GANs) transform low-dimensional latent vectors into visually plausible images.,33059
"noisy a Next, demonstrate vectors we images latent exhibit that consistent from recovered bias.","Next, we demonstrate that latent vectors recovered from noisy images exhibit a consistent bias.",33060
to years. Stereoscopic the introduced video consumer in have past few been technologies market the,Stereoscopic video technologies have been introduced to the consumer market in the past few years.,33061
"in possible. subjective evaluation is some not consuming, cases time and However, expensive,","However, subjective evaluation is time consuming, expensive, and in some cases not possible.",33062
ultimate subjective The through tests. video way assess is to quality,The ultimate way to assess video quality is through subjective tests.,33063
"is in and time cases possible. not expensive, subjective even evaluation However, most consuming,","However, subjective evaluation is time consuming, expensive, and in most cases not even possible.",33064
accurate often algorithms Standard and supervised fusion precise training labels. require,Standard supervised fusion algorithms often require accurate and precise training labels.,33065
"difficult many However, applications. in labels be to obtain may sensing remote accurate","However, accurate labels may be difficult to obtain in many remote sensing applications.",33066
effective proposed performance. classification show and The algorithms regression,The proposed algorithms show effective classification and regression performance.,33067
"for are proposed. two this novel in Thus features anti-spoofing robust paper, and","Thus in this paper, two novel and robust features for anti-spoofing are proposed.",33068
is on Multi-modal features. based implemented robust two face spoofing these detection,Multi-modal face spoofing detection is implemented based on these two robust features.,33069
Experiments and dataset comprehensive used a by constructed are a on ourselves. dataset conducted widely,Experiments are conducted on a widely used dataset and a comprehensive dataset constructed by ourselves.,33070
The techniques. are fusion compared conventional with simulated proposed and algorithms,The proposed algorithms are simulated and compared with conventional fusion techniques.,33071
is that the techniques cost. It result proposed shown lower in,It is shown that the proposed techniques result in lower cost.,33072
evaluated dataset both quality subjectively of each and distorted this in The objectively. video is,The quality of each distorted video in this dataset is evaluated both subjectively and objectively.,33073
"annotation is Manual error-prone. tedious, slow, and","Manual annotation is tedious, slow, and error-prone.",33074
"Specifically, to views. same pose the we train all in system the predict","Specifically, we train the system to predict the same pose in all views.",33075
to a necessary is sufficient not Such poses. consistency accurate but constraint predict,Such a consistency constraint is necessary but not sufficient to predict accurate poses.,33076
paper been have spectrum studied. this in band different three In wide scenarios sensing,In this paper three different scenarios in wide band spectrum sensing have been studied.,33077
of showed Simulation results scheme. proposed performance the superior the,Simulation results showed the superior performance of the proposed scheme.,33078
"is recognition for designed. neural First, network Long-Short-Term-Memory speech a acoustic","First, a Long-Short-Term-Memory neural network for acoustic speech recognition is designed.",33079
"used lip-reading Second, are model Neural features. Convolutional Networks to","Second, Convolutional Neural Networks are used to model lip-reading features.",33080
"acoustic-speech Finally, features simultaneously. visual lip-reading networks are to and process combined visual acoustic and","Finally, acoustic-speech and visual lip-reading networks are combined to process acoustic and visual features simultaneously.",33081
environments. noisy mechanism ensures of model in performance An attention the,An attention mechanism ensures performance of the model in noisy environments.,33082
and prediction of examples taken Numerical are the the analysis. confirm results,Numerical examples are taken and the results confirm the prediction of analysis.,33083
"a Specifically, network for neural we decoding. nonlinear, approach recurrent iterative propose","Specifically, we propose a recurrent neural network approach for nonlinear, iterative decoding.",33084
detectors. these three of achieves to state-of-the-art comparable networks results the face Each,Each of these three networks achieves comparable results to the state-of-the-art face detectors.,33085
"However, changing are in practice. demands the","However, the demands are changing in practice.",33086
cargos. The transporting is rail objective maximize function to income the of net express the,The objective function is to maximize the net income of transporting the rail express cargos.,33087
"centerline the object illustrate to stitching and panorama extraction. usefulness detection, with applications We","We illustrate the usefulness with applications to object detection, panorama stitching and centerline extraction.",33088
"methods changes extract novel mutual In needed. are in suited to correlations particular, dynamic","In particular, novel methods suited to extract dynamic changes in mutual correlations are needed.",33089
crowd framework present We for a synthesis. video-driven,We present a framework for video-driven crowd synthesis.,33090
from crowd video motion are vectors paths. processed compute to input extracted Motion global,Motion vectors extracted from input crowd video are processed to compute global motion paths.,33091
the paths the These video. observed input in encode dominant motions,These paths encode the dominant motions observed in the input video.,33092
crowd on different We videos proposed collected approach demonstrate under the settings.,We demonstrate the proposed approach on crowd videos collected under different settings.,33093
has Blind both important military signal in applications civilian and communications. identification,Blind signal identification has important applications in both civilian and military communications.,33094
"a a Then, metric employed decision of making. decision distance for special tree is","Then, a decision tree of a special distance metric is employed for decision making.",33095
of central is to issue upsampling. limited handle One the joint FCNs capacity,One central issue of FCNs is the limited capacity to handle joint upsampling.,33096
tasks. proposed processing five image evaluated The on is network,The proposed network is evaluated on five image processing tasks.,33097
also helps show DGF multiple to the vision computer We tasks. of that performance improve,We also show that DGF helps to improve the performance of multiple computer vision tasks.,33098
"this presented method, is (SAM), Agnostic causal discovery Modeling in new A paper. Structural","A new causal discovery method, Structural Agnostic Modeling (SAM), is presented in this paper.",33099
"of image-caption function In a pairs. mapping training is learned general, set the from","In general, the mapping function is learned from a training set of image-caption pairs.",33100
"be might However, available. for large language, some paired not image-caption corpus scale","However, for some language, large scale image-caption paired corpus might not be available.",33101
captioning to unpaired this image We an approach by pivoting. problem language present,We present an approach to this unpaired image captioning problem by language pivoting.,33102
Quantitative of comparisons several the baseline against method. demonstrate effectiveness approaches our,Quantitative comparisons against several baseline approaches demonstrate the effectiveness of our method.,33103
"the noise. However, does this distinguish technique signal between and the not","However, this technique does not distinguish between the signal and the noise.",33104
GNU using software USRP tested units. model Radio The implemented and proposed was and,The proposed model was implemented and tested using GNU Radio software and USRP units.,33105
effective reasoning do state-of-the-art approaches the provide understanding not mechanism Current an for process.,Current state-of-the-art approaches do not provide an effective mechanism for understanding the reasoning process.,33106
networks compared. of the neural The and results deep generative model traditional are,The results of the generative model and traditional deep neural networks are compared.,33107
soft-biometric How we can classification human visual system? the the help facial of with improve,How can we improve the facial soft-biometric classification with help of the human visual system?,33108
in information the linear Such implies redundant. is CNNs that relationship,Such linear relationship implies that the information in CNNs is redundant.,33109
to clustering feature subspace eliminates maps. filters redundancy by applying the convolutional in method Our,Our method eliminates the redundancy in convolutional filters by applying subspace clustering to feature maps.,33110
theory intelligence. artificial a general is AI Fractal for,Fractal AI is a theory for general artificial intelligence.,33111
in computer prediction vision. well is Saliency problem a studied,Saliency prediction is a well studied problem in computer vision.,33112
"fine-tuning our we to Furthermore, different models datasets. explore","Furthermore, we explore fine-tuning our models to different datasets.",33113
in recent have Deep demonstrated high for performances neural years. prediction convolutional fixation networks,Deep convolutional neural networks have demonstrated high performances for fixation prediction in recent years.,33114
negligible. is spectral channels shown short be power The for also to body-to-body variation time,The short time power spectral variation for body-to-body channels is also shown to be negligible.,33115
paper the is in that function. residual The identity,The residual in that paper is the identity function.,33116
"lower layers, the propose create to include residuals residual. to We suitably normalized, all from","We propose to include residuals from all lower layers, suitably normalized, to create the residual.",33117
"contribute of layer. a all previous to equally the way, layers output This","This way, all previous layers contribute equally to the output of a layer.",33118
"optical by are or typically Images solid-state relying scanners, on technologies. modern captured","Images are captured by modern scanners, typically relying on solid-state or optical technologies.",33119
Each to works post-processor one. as the network previous,Each network works as post-processor to the previous one.,33120
These in at degradation negligible a quality. segmentation savings come,These savings come at a negligible degradation in segmentation quality.,33121
robust naturally to motion in individual or occlusion frames. approach blur This renders the,This naturally renders the approach robust to occlusion or motion blur in individual frames.,33122
"technical by part trips the A of followed gallery article. of the picture comes next,","A picture gallery of trips comes next, followed by the technical part of the article.",33123
"(IAF With in Faster mind, this RCNN). R-CNN propose Illumination-aware an we","With this in mind, we propose an Illumination-aware Faster R-CNN (IAF RCNN).",33124
"to is an Network image. input introduced of give Specifically, the illumination Illumination-aware an measure","Specifically, an Illumination-aware Network is introduced to give an illumination measure of the input image.",33125
the performance. approaches non-zeros Conventional to leading tend learning to overestimate poor machine,Conventional machine learning approaches tend to overestimate the non-zeros leading to poor performance.,33126
process powerful Gaussian The the useful more particularly ZiGPs networks are making for interpretable.,The ZiGPs are particularly useful for making the powerful Gaussian process networks more interpretable.,33127
We networks modelling through variable-order is latent sparse where signals. achieved introduce GP mixing sparse,We introduce sparse GP networks where variable-order latent modelling is achieved through sparse mixing signals.,33128
"fusion facial parts as are However, extraction tasks. as accurate their challenging well","However, accurate facial parts extraction as well as their fusion are challenging tasks.",33129
approaches. segmentation Modern image have algorithms triggered deep various learning,Modern deep learning algorithms have triggered various image segmentation approaches.,33130
pixel deal most segmentation. them of with based However,However most of them deal with pixel based segmentation.,33131
"certain However, provide contextual degree superpixels information computation while of cost. a reducing","However, superpixels provide a certain degree of contextual information while reducing computation cost.",33132
"max-voting we weighted-average. have Furthermore, ensemble of like enlisted number a and approaches","Furthermore, we have enlisted a number of ensemble approaches like max-voting and weighted-average.",33133
analyze We classes. Dempster-Shafer theory confusion of uncertainty among used to various have the also,We have also used the Dempster-Shafer theory of uncertainty to analyze confusion among various classes.,33134
"account. the and takes it priori the coherence minimizes Unlike region spatiotemporal into overlapping works,","Unlike the priori works, it minimizes the overlapping region and takes spatiotemporal coherence into account.",33135
the flicker proposed to solve registration subpixel The method is image problem.,The subpixel image registration method is proposed to solve the flicker problem.,33136
Reality (MR) user yields of types technology Mixed new that is powerful a interactive experience.,Mixed Reality (MR) is a powerful interactive technology that yields new types of user experience.,33137
expensive. training of neural The is deep nets,The training of deep neural nets is expensive.,33138
training corrector neural of predictor- present for a method We the nets. deep,We present a predictor- corrector method for the training of deep neural nets.,33139
SGD modifications special required to this backpropagation with is methodology. No by,No special modifications to SGD with backpropagation is required by this methodology.,33140
reconstruction. image of medical Deep networks have demonstrated for neural field promising potential the,Deep neural networks have demonstrated promising potential for the field of medical image reconstruction.,33141
Previous require (e.g. QSM data approaches multiple of orientation,Previous approaches of QSM require multiple orientation data (e.g.,33142
regularization or or Orientation (e.g. of terms Sampling Susceptibility Multiple through Calculation COSMOS),Calculation of Susceptibility through Multiple Orientation Sampling or COSMOS) or regularization terms (e.g.,33143
"long or artifacts. multiple require suffer scans Unfortunately, either from orientation they","Unfortunately, they either require long multiple orientation scans or suffer from artifacts.",33144
using modified network trained U-net and COSMOS structure maps. The a QSM gold-standard has is,The network has a modified U-net structure and is trained using gold-standard COSMOS QSM maps.,33145
"patients. preliminary two application, network a As tested for was the","As a preliminary application, the network was tested for two patients.",33146
"describe paper and submitted agents Within framework, for the rules, we competition. this results this","Within this paper we describe the framework, rules, submitted agents and results for this competition.",33147
of There different with associated are BCI classification problems tasks.,There are problems associated with classification of different BCI tasks.,33148
indices In this BCI. use of paper the propose the we chaotic of,In this paper we propose the use of chaotic indices of the BCI.,33149
quantify Ut is misleading. Ul be of The and way to using inappropriate conventional can,The conventional way of using Ut to quantify Ul is inappropriate and can be misleading.,33150
"from model different attributes. learns In sets image this unpaired addition, with","In addition, this model learns from unpaired image sets with different attributes.",33151
results fast high-resolution speed. in It produces high-quality and,It produces high-resolution and high-quality results in fast speed.,33152
"biomedical apply quantization In accurate techniques FCNs segmentation. this paper, image we for to","In this paper, we apply quantization techniques to FCNs for accurate biomedical image segmentation.",33153
"via large-scale particle Gaussian can Second, variations effectively it a handle filter.","Second, it can effectively handle large-scale variations via a Gaussian particle filter.",33154
the using synthesis performed views maps. available and is depth View,View synthesis is performed using the available views and depth maps.,33155
modelled constraints can how show We as be a instance. the Max-SAT,We show how the constraints can be modelled as a Max-SAT instance.,33156
evolve states. such We toward solutions of numerical steady give that examples weak,We give numerical examples of solutions that evolve toward such weak steady states.,33157
greatly and learning. approach the of robustness depth This increases accuracy the,This approach greatly increases the accuracy and robustness of the depth learning.,33158
dataset results. our on Evaluation state-of-the-art outperforms method the shows that KITTI,Evaluation on KITTI dataset shows that our method outperforms the state-of-the-art results.,33159
call We strategy this heatmap regulation (HR).,We call this strategy heatmap regulation (HR).,33160
"attention this In angle steering driving the we to prediction. apply autonomous paper, for mechanism","In this paper, we apply the attention mechanism to autonomous driving for steering angle prediction.",33161
tool Probabilistic applications. graphical key in machine learning are a models,Probabilistic graphical models are a key tool in machine learning applications.,33162
popular variational of family and methods successful approaches. a Iterative are,Iterative variational methods are a popular and successful family of approaches.,33163
"paper, function In consider sequential partition we the instead over summation variables. via computing this","In this paper, we instead consider computing the partition function via sequential summation over variables.",33164
to learning in study This best which assess models lung diagnosis. aims perform deep cancer,This study aims to assess which deep learning models perform best in lung cancer diagnosis.,33165
Several biopsies. to trained non-cancer cancer and were models learning discriminate deep,Several deep learning models were trained to discriminate cancer and non-cancer biopsies.,33166
have metabolic tool become Genome-scale for metabolic models examining principles. fundamental a,Genome-scale metabolic models have become a fundamental tool for examining metabolic principles.,33167
"reactions. of enzyme regulation take into and translation, we regulation, account also Besides activities transcriptional","Besides transcriptional regulation, we also take into account regulation of translation, enzyme activities and reactions.",33168
centrality new reveals of three-domain functional representation known structural relevance. the This high with components,This new three-domain representation reveals the structural centrality of components with known high functional relevance.,33169
nucleotide single type studied The polymorphism of is the most (SNP) variation. genetic widely,The single nucleotide polymorphism (SNP) is the most widely studied type of genetic variation.,33170
genome-phenotype Haplotype in unravelling the information association. is essential,Haplotype information is essential in unravelling the genome-phenotype association.,33171
Minimum metric for The Correction reads. of (MEC) reconstruction Error haplotypes from often is used,The Minimum Error Correction (MEC) metric is often used for reconstruction of haplotypes from reads.,33172
"However, problems with have been the MEC metric reported.","However, problems with the MEC metric have been reported.",33173
is coverage The of explored data. different performance levels of rates and MEC for error,The performance of MEC is explored for different coverage levels and error rates of data.,33174
"based an LCANet, neural lipreading propose paper, network system. we end-to-end deep In this","In this paper, we propose LCANet, an end-to-end deep neural network based lipreading system.",33175
The short-term both information. effectively encoder spatio-temporal and captures long-term,The encoder effectively captures both short-term and long-term spatio-temporal information.,33176
"a More to generate cascaded decoder importantly, attention-CTC LCANet incorporates output texts.","More importantly, LCANet incorporates a cascaded attention-CTC decoder to generate output texts.",33177
models. in a We algorithm message-passing variational inference variational propose such for,We propose a variational message-passing algorithm for variational inference in such models.,33178
Our show our that current methods accuracy. and visual method outperforms in experiments both quality,Our experiments show that our method outperforms current methods in both visual quality and accuracy.,33179
health a active structural We transceiver describe highly-integrated (SHM). monitoring for CMOS,We describe a highly-integrated CMOS transceiver for active structural health monitoring (SHM).,33180
"sparse By advantages. structural two-fold this can representation applying appearance we take into model,","By applying this representation into structural sparse appearance model, we can take two-fold advantages.",33181
"the target time of over First, is correlation considered. patches","First, the correlation of target patches over time is considered.",33182
tracking that on trackers. show the Evaluations challenging benchmark outperforms our state-of-the-art recent method,Evaluations on recent challenging benchmark show that our tracking method outperforms the state-of-the-art trackers.,33183
transfiguration images. in problem paper This object the studies wild,This paper studies the object transfiguration problem in wild images.,33184
"should transfiguration, before No matter maps remain or attention after object constant.","No matter before or after object transfiguration, attention maps should remain constant.",33185
"hardware of challenging. more in CNNs has deployment become the Consequently,","Consequently, the deployment of CNNs in hardware has become more challenging.",33186
"to setting. we present baseline extended Furthermore, approach a detection from recognition","Furthermore, we present a baseline approach extended from recognition to detection setting.",33187
This paper approach object instance a presents segmentation. weakly-supervised to,This paper presents a weakly-supervised approach to object instance segmentation.,33188
re-identification context the of our person propose (re-ID). We in method,We propose our method in the context of person re-identification (re-ID).,33189
stability GANs. guaranteeing essential for in constraint Lipschitz is training,Lipschitz constraint is essential in guaranteeing training stability for GANs.,33190
We conditions Lipschitz meeting GANs satisfy with constraint. theoretically prove the that BC the discriminators,We prove theoretically that GANs with discriminators meeting the BC conditions satisfy the Lipschitz constraint.,33191
dyadic models network data. for a interaction propose of novel We temporal class,We propose a novel class of network models for temporal dyadic interaction data.,33192
"or augmented reality, personal automation. industrial robotics","augmented reality, personal robotics or industrial automation.",33193
to supports E-RPN well generalize The during training.,The E-RPN supports to generalize well during training.,33194
Convolutional are Topic-related Neuronal Networks. a using then of categories topic-specific estimated by set,Topic-related categories are then estimated by using a set of topic-specific Convolutional Neuronal Networks.,33195
comprehensive We object provide salient models. detection of (SOD) evaluation a,We provide a comprehensive evaluation of salient object detection (SOD) models.,33196
"when however, models, being perform The satisfactory far applied from daily to real-world scenes. still","The models, however, still perform far from being satisfactory when applied to real-world daily scenes.",33197
"update the and a dataset benchmark. previous we new propose Then, saliency quality high","Then, we propose a new high quality dataset and update the previous saliency benchmark.",33198
"attribute-based dataset. assessment performance on we Finally, report our","Finally, we report attribute-based performance assessment on our dataset.",33199
gradient based outperform They speed. both the in traditional methods running descent and accuracy,They outperform the traditional gradient descent based methods in both accuracy and running speed.,33200
"multiple based methods. regressors computing learning especially resources, lots for training of deep Moreover, requires","Moreover, training multiple regressors requires lots of computing resources, especially for deep learning based methods.",33201
"Regression this develop to Self-Iterative efficiency. paper, a (SIR) model we In improve framework the","In this paper, we develop a Self-Iterative Regression (SIR) framework to improve the model efficiency.",33202
"Learning framework of paper, within In problem model the this Multiple we (MIL). this Instance","In this paper, we model this problem within the framework of Multiple Instance Learning (MIL).",33203
"addition, event-stream the correlative tracking of In has high based this filter advantage speed.","In addition, this correlative filter based event-stream tracking has the advantage of high speed.",33204
past Autonomous driving attention tremendous in attracted has few the years. especially,Autonomous driving has attracted tremendous attention especially in the past few years.,33205
"Compared with from real public scenes, existing datasets e.g.","Compared with existing public datasets from real scenes, e.g.",33206
used observation are environment. given types motion characterize Both in to a agents',Both observation types are used to characterize agents' motion in a given environment.,33207
trajectory for Obtained multi-perspective predicting using data and results feasibility the analyzing of information. exemplify,Obtained results exemplify the feasibility of using multi-perspective data for predicting and analyzing trajectory information.,33208
commonly limitations. Synthetic they but datasets have used are,Synthetic datasets are commonly used but they have limitations.,33209
components basic expressions combinations (AU). are Facial of Action called Units,Facial expressions are combinations of basic components called Action Units (AU).,33210
expression for developing AUs analysis. general key facial is Recognizing,Recognizing AUs is key for developing general facial expression analysis.,33211
is present like here these. to aim two statistical My to approaches questions,My aim here is to present two statistical approaches to questions like these.,33212
These more models the Gaussian than accurately the are to model. traditional error realistic represent,These models are more realistic to accurately represent the error than the traditional Gaussian model.,33213
"orientation objects images, remote sensing of is absolute the arbitrary. In","In remote sensing images, the absolute orientation of objects is arbitrary.",33214
observations. empirical two driven method by is Our,Our method is driven by two empirical observations.,33215
from their complementary to is algorithms exploit disparity different still advantages Refining maps challenging. raw,Refining raw disparity maps from different algorithms to exploit their complementary advantages is still challenging.,33216
a disparity network refiner inputs. raw into to refine better,into a refiner network to better refine raw disparity inputs.,33217
A discriminator fields at and different receptive scales. disparities network classifies,A discriminator network classifies disparities at different receptive fields and scales.,33218
Both versions proposed. of semi-supervised are the algorithm fully and supervised,Both fully supervised and semi-supervised versions of the algorithm are proposed.,33219
of can fuse depth from different depths kinds sources. be generalized to The algorithm,The algorithm can be generalized to fuse depths from different kinds of depth sources.,33220
"different fusion stereo-ToF stereo-stereo fusion opportunities: fusion. stereo-monocular Experiments explored and fusion,","Experiments explored different fusion opportunities: stereo-monocular fusion, stereo-ToF fusion and stereo-stereo fusion.",33221
second hypotheses. and Results our confirm third,Results confirm our second and third hypotheses.,33222
scene a surveillance for road of Appropriate traffic. detection in is of modeling essential anomalies,Appropriate modeling of a surveillance scene is essential for detection of anomalies in road traffic.,33223
"it intuitive distance-based In makes model to the the learning estimate scene addition, parameters.","In addition, the distance-based scene learning makes it intuitive to estimate the model parameters.",33224
methods. have with state-of-the-art the compared results various We,We have compared the results with various state-of-the-art methods.,33225
We overcome buildings. by the exploiting distinct of geometric issue properties to this propose,We propose to overcome this issue by exploiting the distinct geometric properties of buildings.,33226
"which encompass discuss mission-critical IoT, vehicle the communications autonomous and applications, services. industrial We control","We discuss the applications, which encompass industrial IoT, autonomous vehicle control and mission-critical communications services.",33227
VOC results bells and Experiments PASCAL show without new and state-of-the-art any whistles. COCO on,Experiments show new state-of-the-art results on PASCAL VOC and COCO without any bells and whistles.,33228
"malicious In are algorithms paper, we against behavior attack). distributed this (jamming that study robust","In this paper, we study distributed algorithms that are robust against malicious behavior (jamming attack).",33229
jammers and We launching the coordinated consider uncoordinated of both attacks. cases,We consider both the cases of jammers launching coordinated and uncoordinated attacks.,33230
as We and bandit in algorithms. each develop scenario a setup problem multi-player the,We setup the problem in each scenario as a multi-player bandit and develop algorithms.,33231
data RGB-D with depth progress availability of segmentation enables in The CNNs. semantic,The availability of depth data enables progress in RGB-D semantic segmentation with CNNs.,33232
high suffer methods memory from cost. and computation These,These methods suffer from high computation and memory cost.,33233
"both integrated easily CNNs. into can Without operators introducing additional be existing any parameters,","Without introducing any additional parameters, both operators can be easily integrated into existing CNNs.",33234
into of space RGB compressed the This reduces representation cost. and decoding computational bypasses,This bypasses decoding of the compressed representation into RGB space and reduces computational cost.,33235
information gradient even prohibitive field) spatial is a provided. computationally with random,a spatial random field) is computationally prohibitive even with gradient information provided.,33236
of evolution life transition was a major history multicellularity on of in the earth. The,The evolution of multicellularity was a major transition in the history of life on earth.,33237
Conditions multicellularity theoretically favored have studied is and experimentally. which been under,Conditions under which multicellularity is favored have been studied theoretically and experimentally.,33238
of station measurements. are mobile localize (LOS) sight methods Well-known using (MS) employed to line,Well-known methods are employed to localize mobile station (MS) using line of sight (LOS) measurements.,33239
other performance of conventional conducted Simulations to algorithms. the comparison of SRNI investigate in are,Simulations are conducted to investigate the performance of SRNI in comparison of other conventional algorithms.,33240
"the present by learning method deep Co-Training Deep based Co-Training, framework. a inspired We","We present Deep Co-Training, a deep learning based method inspired by the Co-Training framework.",33241
semantic Extensive experiments datasets (i.e. on four segmentation public,Extensive experiments on four public semantic segmentation datasets (i.e.,33242
been known Euclidean and studied. case widely is The well has,The Euclidean case is well known and has been widely studied.,33243
theory Our ranking by statistical motivated is problems.,Our theory is motivated by statistical ranking problems.,33244
"supervisory lack difficulty. unsupervised the setting, signals exacerbate of In this","In the unsupervised setting, lack of supervisory signals exacerbate this difficulty.",33245
The anatomies difficulties. adds presence with brain enhanced other further diseases or intensities of,The presence of other brain anatomies or diseases with enhanced intensities adds further difficulties.,33246
orthopedic CT commonly in procedures. used is,CT is commonly used in orthopedic procedures.,33247
"structures. bone However, contrast poor has for MRI","However, MRI has poor contrast for bone structures.",33248
a painting beautiful childhood. of Drawing a is many people dream since,Drawing a beautiful painting is a dream of many people since childhood.,33249
data. to is Sketch generate model (SIE) The training the Extraction applied Image,The Sketch Image Extraction (SIE) model is applied to generate the training data.,33250
pencil images sketch It extraction. and smoothing includes reality,It includes smoothing reality images and pencil sketch extraction.,33251
then are images appealing optimization by generated iterations. artistic The,The appealing artistic images are then generated by optimization iterations.,33252
the dataset The and Dataset. Cats Buildings Kaggle Experiments operated Oxford on are,Experiments are operated on the Kaggle Cats dataset and The Oxford Buildings Dataset.,33253
"results synthesis are to and Our artistic, proved robust. be beautiful","Our synthesis results are proved to be artistic, beautiful and robust.",33254
significant performance improvement with The respect results the show of the experimental state to art.,The experimental results show significant performance improvement with respect to the state of the art.,33255
multiple TGM layers convolutional our models for present We detection. video with activity fully,We present our fully convolutional video models with multiple TGM layers for activity detection.,33256
"subjects simple expresses abstract, view. is art an that exaggerated and in form Caricature","Caricature is an art form that expresses subjects in abstract, simple and exaggerated view.",33257
algorithm to Fully trained Convolutional instances. simple a classify is First Network (FCN) object,First algorithm is a simple Fully Convolutional Network (FCN) trained to classify object instances.,33258
masks. objects trained an ConvAE of output dataset generated on is to localize artificially segmentation,ConvAE is trained to localize objects on an artificially generated dataset of output segmentation masks.,33259
end-to-end hybrid solved effectively We using problems networks. both demonstrate can be that CNN-LSTM,We demonstrate that both problems can be solved effectively using hybrid end-to-end CNN-LSTM networks.,33260
available. made be code The will source,The source code will be made available.,33261
for to and autonomous vehicle. is objects an The crucial pedestrians detect ability moving other,The ability to detect pedestrians and other moving objects is crucial for an autonomous vehicle.,33262
be overhead. minimum must This in system real-time done with,This must be done in real-time with minimum system overhead.,33263
a on image superpixel presents decomposition. This segmentation approach region-growing work based,This work presents a region-growing image segmentation approach based on superpixel decomposition.,33264
"Secondly, guide merging process. region we to strategy merging a efficiently the propose global","Secondly, we propose a global merging strategy to efficiently guide the region merging process.",33265
allows a regions boundary adherence. This to with consistent strong final segmentation into reach,This allows to reach a final segmentation into consistent regions with strong boundary adherence.,33266
promising results proposed approach. obtained demonstrate the of potential The the,The obtained results demonstrate the promising potential of the proposed approach.,33267
"In framework, learning propose for we SeqFace, discriminative paper, this face called a features.","In this paper, we propose a framework, called SeqFace, for learning discriminative face features.",33268
"critical circulators, communication as and wireless such and isolators radar components, are applications. Non-reciprocal to","Non-reciprocal components, such as isolators and circulators, are critical to wireless communication and radar applications.",33269
"In based approach paper, on switched a present to lines. non-reciprocity transmission general this we","In this paper, we present a general approach to non-reciprocity based on switched transmission lines.",33270
analysis as results of wireless communication. be obtained This for generalization a can seen,This analysis can be seen as a generalization of results obtained for wireless communication.,33271
"(RMT) random bound. results direct of a theory matrix Further, application lower from yields trivial","Further, direct application of results from random matrix theory (RMT) yields a trivial lower bound.",33272
other This in unique linear systems. for been observed not behavior decoders has,This unique behavior has not been observed for linear decoders in other systems.,33273
"procedure. saliency per-pixel masks However, inefficient tedious a annotating and is","However, annotating per-pixel saliency masks is a tedious and inefficient procedure.",33274
high is method segmentation and robust of An accurate automated interest.,An accurate and robust automated segmentation method is of high interest.,33275
"approach starts the a CNN values, to The low-level from successful similar style methods. in","The approach starts from low-level values, in a similar style to the successful CNN methods.",33276
building by that assumption is similar is similar The appearance. motion characterized,The building assumption is that similar motion is characterized by similar appearance.,33277
aquifers. the rain mountain the the the propagation of of towards water top from,the propagation of rain water from the top of the mountain towards the aquifers.,33278
data illustrate on hydrological the datasets. to method potential Tests real of analyze this,Tests on real data illustrate the potential of this method to analyze hydrological datasets.,33279
This the increasing environmental includes or occlusions varying effects. weather occurrence of and,This includes increasing the occurrence of occlusions or varying environmental and weather effects.,33280
"few addressed modeling the in have domain. sensor variation However,","However, few have addressed modeling variation in the sensor domain.",33281
"of designing damping technique controls. greatly may proposed facilitate capable factors participation the Therefore, estimating","Therefore, the proposed technique capable of estimating participation factors may greatly facilitate designing damping controls.",33282
of recognition processing. image Detection the a and one Text aspect important is of,Text Detection and recognition is a one of the important aspect of image processing.,33283
compares this task. handle and methods paper analyzes the to This,This paper analyzes and compares the methods to handle this task.,33284
problems. enumerates fundamental the It that and consideration summarizes need factors these problems addressing when,It summarizes the fundamental problems and enumerates factors that need consideration when addressing these problems.,33285
efficiently operations be ware direct can These on including platforms different implemented hard- implementation.,These operations can be implemented efficiently on different platforms including direct hard- ware implementation.,33286
We and standard on applied variants its LBPNet benchmarks.,We applied LBPNet and its variants on standard benchmarks.,33287
"a estimating range of of marginalized the is over sigma, it noise Instead scales. noise","Instead of estimating the noise sigma, it is marginalized over a range of noise scales.",33288
"marginalization the for approach. criterion a built Also, is proposed termination on RANSAC new","Also, a new termination criterion for RANSAC is built on the proposed marginalization approach.",33289
in for component organization is popular brain tool investigating analysis Independent neuroscience research. (ICA) a,Independent component analysis (ICA) is a popular tool for investigating brain organization in neuroscience research.,33290
new group a multi-level HINT probabilistic ICA ICA. implements model for,HINT implements a new multi-level probabilistic ICA model for group ICA.,33291
covariate for brain provides It framework ICA effects investigating a on networks. modeling principled statistically,It provides a statistically principled ICA modeling framework for investigating covariate effects on brain networks.,33292
enhanced abilities. of range have to The attention descriptors due local gained discriminative wide their,The local descriptors have gained wide range of attention due to their enhanced discriminative abilities.,33293
order the factor in is relationship direction. particular directional Local multi-radius a,Local directional order is the multi-radius relationship factor in a particular direction.,33294
orders. the center of is to It neighboring the into required range value transform,It is required to transform the center value into the range of neighboring orders.,33295
"the over descriptor. image of histogram to LDOP computed is whole Finally, construct the","Finally, the histogram of LDOP is computed over whole image to construct the descriptor.",33296
the experimental superiority descriptor. and confirm results of robustness The LDOP the,The experimental results confirm the superiority and robustness of the LDOP descriptor.,33297
"vehicle localization. strategies work, evaluate multiple fusion this for research we In pose-graph and","In this work, we research and evaluate multiple pose-graph fusion strategies for vehicle localization.",33298
"a i.e. on focus We localization single system, absolute fusing","We focus on fusing a single absolute localization system, i.e.",33299
pooling proposed heuristic most removes choices outperforms its counterparts. method The and RoI,The proposed method removes most heuristic choices and outperforms its RoI pooling counterparts.,33300
fully learnable object towards moves It detection. further,It moves further towards fully learnable object detection.,33301
identification human. strongly of information this All contributes to,All this information contributes strongly to identification of human.,33302
discussed on existing techniques their based are performances. The,The existing techniques are discussed based on their performances.,33303
for further directions also This paper future provides research.,This paper also provides future directions for further research.,33304
over recently synthesis verification that We cross-spectrum improves our rates demonstrate approach approaches. published,We demonstrate that our approach improves cross-spectrum verification rates over recently published synthesis approaches.,33305
"(VQA) question has topical. In years, answering visual become recent","In recent years, visual question answering (VQA) has become topical.",33306
(iVQA). paper of this the propose problem In inverse VQA we,In this paper we propose the inverse problem of VQA (iVQA).,33307
window This a about VQA images. into what new completely `believe' provides models,This provides a completely new window into what VQA models `believe' about images.,33308
going made Suggestions weaknesses address forward. these then to on are how,Suggestions are then made on how to address these weaknesses going forward.,33309
the Test were noted. and runs application of outputs made were,Test runs were made and outputs of the application were noted.,33310
"considered biases nuisance. are Indeed, typically idiosyncratic choice as","Indeed, idiosyncratic choice biases are typically considered as nuisance.",33311
a and discrimination we perceptual in Here a choice motor idiosyncratic task task. biases quantify,Here we quantify idiosyncratic choice biases in a perceptual discrimination task and a motor task.,33312
biases in report cases. both substantial and significant We,We report substantial and significant biases in both cases.,33313
sensor a from cheap low-resolution Registering point clouds is task. a challenging accurately,Registering accurately point clouds from a cheap low-resolution sensor is a challenging task.,33314
"implement mixture architecture existing Additionally, alignment these Gaussian changes. dynamic be efficiently cannot the","Additionally, the existing Gaussian mixture alignment architecture cannot be efficiently implement these dynamic changes.",33315
"investigated have problem. paper, this the unconstrained extensively recognition ear we In","In this paper, we have extensively investigated the unconstrained ear recognition problem.",33316
"the further, convolutional different network improve have deep To we models. combined neural performance","To improve the performance further, we have combined different deep convolutional neural network models.",33317
"have problem the in Finally, of the field. recognition bias dataset we addressed ear","Finally, we have addressed the problem of dataset bias in the ear recognition field.",33318
It influence image on the observed has an been also that has quality results.,It has also been observed that image quality has an influence on the results.,33319
Lane components maps. markings important the of such are one of,Lane markings are one of the important components of such maps.,33320
convey the of drivers. roads rules to markings Lane,Lane markings convey the rules of roads to drivers.,33321
"shape the In property bias paper, of examine CNNs. this we","In this paper, we examine the shape bias property of CNNs.",33322
feature and good depend learning learning algorithms of representations. success heavily The deep generalisation on,The success and generalisation of deep learning algorithms heavily depend on learning good feature representations.,33323
"such factorisation date, To not been has attempted.","To date, such factorisation has not been attempted.",33324
framework. a novel active We propose feature-driven,We propose a novel feature-driven active framework.,33325
GP the and by variograms using are a discrete estimated search method. grid of Kernels,Kernels of the GP are estimated by using variograms and a discrete grid search method.,33326
Deep rises significant tasks. learning performance the yields in object recognition in,Deep learning yields significant rises in performance in the object recognition tasks.,33327
plays essential role life. change Global an in our climate daily,Global climate change plays an essential role in our daily life.,33328
"for deep In retrieval. novel image unsupervised method hashing we paper, propose large-scale this a","In this paper, we propose a novel unsupervised deep hashing method for large-scale image retrieval.",33329
a extraction images (IR) challenging from task. infrared remains Feature,Feature extraction from infrared (IR) images remains a challenging task.,33330
Learning based therefore raw can work that imagery/patches methods have on assumed significance.,Learning based methods that can work on raw imagery/patches have therefore assumed significance.,33331
part computer extraction a plays vision significant in Feature tasks.,Feature extraction plays a significant part in computer vision tasks.,33332
abstract that more neural extracts We the from layers. stacked networks deep leverage features,We leverage the deep neural networks that extracts more abstract features from stacked layers.,33333
form interpolation module intermediate the predictions blends final result. The to the,The interpolation module blends the intermediate predictions to form the final result.,33334
evaluation in heterogeneous policy been emphasized effects literature. the Unobserved treatment have,Unobserved heterogeneous treatment effects have been emphasized in the policy evaluation literature.,33335
and is simple to The implement. Kolmogorov-Smirnov-type modified test proposed consistent,The proposed modified Kolmogorov-Smirnov-type test is consistent and simple to implement.,33336
Monte simulations performs that finite Carlo in show samples. well our test,Monte Carlo simulations show that our test performs well in finite samples.,33337
"factors analysing time. their Here, disentangle these we influence over by","Here, we disentangle these factors by analysing their influence over time.",33338
"DNN-based recent the a high-level we use of model. To measure bottom-up influence saliency features,","To measure the influence of high-level bottom-up features, we use a recent DNN-based saliency model.",33339
"by processes. dataset, top-down predominantly search locations fixation In determined the are","In the search dataset, fixation locations are determined predominantly by top-down processes.",33340
Reward shaping reinforcement accelerate reward (RL) receiving allows additional learning to learning signals. agents by,Reward shaping allows reinforcement learning (RL) agents to accelerate learning by receiving additional reward signals.,33341
"manually, for tasks. However, these complex difficult can RL signals especially design to be","However, these signals can be difficult to design manually, especially for complex RL tasks.",33342
almost computer descriptors vision. progress every problem on fueled Interest point have in,Interest point descriptors have fueled progress on almost every problem in computer vision.,33343
is and/or adversarial State-of-art recursive video prediction based networks learning. generative on network LSTM,State-of-art video prediction is based on LSTM recursive networks and/or generative adversarial network learning.,33344
related works. recent framework Experiments the proposed show that outperforms,Experiments show that the proposed framework outperforms recent related works.,33345
power multiple cells. schemes schemes in paper This various distribution co-channel for discusses,This paper discusses various schemes for power distribution schemes in multiple co-channel cells.,33346
"The computation. similarity suggests frames, of nearby to reuse opportunity video spatial however,","The spatial similarity of nearby video frames, however, suggests opportunity to reuse computation.",33347
"approach this on to part In accelerating inference present a video. new, paper, we two","In this paper, we present a new, two part approach to accelerating inference on video.",33348
"many great potential tasks. vision-related Recently, recognition learning proved its algorithms in deep have","Recently, deep learning algorithms have proved its great potential in many vision-related recognition tasks.",33349
for Experiments challenges. various datasets action recognition video human conducted two are public on containing,Experiments are conducted on two public video datasets for human action recognition containing various challenges.,33350
results method The comparing our approaches. that performance the existing with achieves show state-of-the-art,The results show that our method achieves the state-of-the-art performance comparing with existing approaches.,33351
as is An application MT-HVDC offshore wind selected including system two farms test case.,An offshore MT-HVDC system including two wind farms is selected as application test case.,33352
"first, of is theory. a network the dynamic using generalized At proposed, model hypergraph","At first, a generalized dynamic model of the network is proposed, using hypergraph theory.",33353
"associated we easily with show keypoints be the Additionally, semantic shapes. that predicted can","Additionally, we show that semantic keypoints can be easily associated with the predicted shapes.",33354
"for In tutorial provided. a addition, SurvBoost detailed hands-on is","In addition, a detailed hands-on tutorial for SurvBoost is provided.",33355
system that present automatic moment capture in cameras. We mobile an on real-time runs,We present an automatic moment capture system that runs in real-time on mobile cameras.,33356
of the state schemes. simulation non-robust show performance art The compared to results gains the,The simulation results show the performance gains compared to non-robust state of the art schemes.,33357
critical expression a facial in interactions. human-machine recognition role Reliable plays,Reliable facial expression recognition plays a critical role in human-machine interactions.,33358
two the We datasets evaluate model on proposed threat scenarios. various public under,We evaluate the proposed model on two public datasets under various threat scenarios.,33359
"recurrent achieved audio we using landmarks on features. By based network, neural a mouth","By using a recurrent neural network, we achieved mouth landmarks based on audio features.",33360
"However, solutions and regions. to undesired the repaired network artifacts introduce noise still current","However, current network solutions still introduce undesired artifacts and noise to the repaired regions.",33361
content. paper their strives segmentation video actions of pixel-level for This and in actors,This paper strives for pixel-level segmentation of actors and their actions in video content.,33362
"Many vision on and superpixels. object such recognition increasingly computer build as segmentation, applications,","Many computer vision applications, such as object recognition and segmentation, increasingly build on superpixels.",33363
"that systematically algorithms deal have so with images. superpixel there few noisy far been However,","However, there have been so far few superpixel algorithms that systematically deal with noisy images.",33364
derived the presence is of for Recursion random delays. importance weights equation under,Recursion equation for importance weights is derived under the presence of random delays.,33365
"Further, the work proposed convergence this of ensure conditions the particle the which explores filter.","Further, this work explores the conditions which ensure the convergence of the proposed particle filter.",33366
"on external only the intra-operatively is displays. data acquired indirectly visualized Moreover,","Moreover, the intra-operatively acquired data is only visualized indirectly on external displays.",33367
under evaluated for different performance the framework to The the datasets is visualize factors. dependency,The framework is evaluated under different datasets to visualize the performance for the dependency factors.,33368
This single estimation. problem the depth paper of considers image,This paper considers the problem of single image depth estimation.,33369
errors training. other about loss measuring inference is used functions The for in,The other is about loss functions for measuring inference errors used in training.,33370
"Tree ensemble specifically for modified purpose. were methods, this forests random","Tree ensemble methods, specifically random forests were modified for this purpose.",33371
"particular, we is noise consider white temporally uncorrelated. spatially and/or that In","In particular, we consider noise that is temporally white and/or spatially uncorrelated.",33372
that the noise additional information performance. enhanced process of about Detectors use have this make,Detectors that make use of this additional information about the noise process have enhanced performance.,33373
"exist. does LMPIT white noise, an For not","For white noise, an LMPIT does not exist.",33374
object-part The semantic-level tackles both modeling. using and PersonLab model proposed part-based associations reasoning,The proposed PersonLab model tackles both semantic-level reasoning and object-part associations using part-based modeling.,33375
most review architectures. will prominent multilayer proposals This the using convolutional document,This document will review the most prominent proposals using multilayer convolutional architectures.,33376
"are design friendly existing all not networks quantization. Unfortunately, to","Unfortunately, not all existing networks design are friendly to quantization.",33377
iterative residuals. reweighted the Here we on data analyse algorithms MaxCon as,Here we analyse MaxCon as iterative reweighted algorithms on the data residuals.,33378
This an makes it randomized alternative for global optimizers. attractive and methods,This makes it an attractive alternative for randomized methods and global optimizers.,33379
"to Further, address framework single-view formulated efficient multi-view an ambiguity. is","Further, an efficient multi-view framework is formulated to address single-view ambiguity.",33380
the consistently this single-view framework of We network. performance improves that the show,We show that this framework consistently improves the performance of the single-view network.,33381
over current competitive achieves or methods. superior approach state-of-the-art the performance Our,Our approach achieves competitive or superior performance over the current state-of-the-art methods.,33382
"a in In novel unconstrained introduce localization paper, problem this videos. we audio-visual event of","In this paper, we introduce a novel problem of audio-visual event localization in unconstrained videos.",33383
"easy training. solution is compact, requires compute and to no Our","Our solution is compact, easy to compute and requires no training.",33384
"camera the statistical In all known, outperforms methods. the is MSGP where setting","In the setting where the camera is known, MSGP outperforms all statistical methods.",33385
using have object achievements exciting witnessed many deep detection years techniques. for Recent learning,Recent years have witnessed many exciting achievements for object detection using deep learning techniques.,33386
but Filters are visual Correlation effects. efficient from unwanted in tracking Discriminative suffer (DCF) boundary,Discriminative Correlation Filters (DCF) are efficient in visual tracking but suffer from unwanted boundary effects.,33387
"Besides, of be can method solved it via alternating (ADMM). the direction efficiently multipliers","Besides, it can be efficiently solved via the alternating direction method of multipliers (ADMM).",33388
to effects and regressions Linear fixed group used are widely with estimate treatment effects. period,Linear regressions with period and group fixed effects are widely used to estimate treatment effects.,33389
another this that propose We issue. solves estimator,We propose another estimator that solves this issue.,33390
"different linear estimator. two significantly the from the revisit, In applications it we is regression","In the two applications we revisit, it is significantly different from the linear regression estimator.",33391
in B-scan hyperbola (GPR) radar reflections) penetrating ground images.,hyperbola reflections) in B-scan ground penetrating radar (GPR) images.,33392
a requires precise Tracking an Multi-People in open-world effort setting special in detection.,Multi-People Tracking in an open-world setting requires a special effort in precise detection.,33393
image machine to generate captioning aim image captions contents. of is The to by describe,The aim of image captioning is to generate captions by machine to describe image contents.,33394
"images captions remains discriminative many efforts, Despite non-trivial. for generating","Despite many efforts, generating discriminative captions for images remains non-trivial.",33395
clipping operators. on a generic rep- relies polygon-based polygon propose resentation We which,We propose a polygon-based generic rep- resentation which relies on polygon clipping operators.,33396
reported. on results localization pedestrian a Experimental problem are,Experimental results on a pedestrian localization problem are reported.,33397
may genetic Behavioural arise differences or absence environmental variation. of the in,Behavioural differences may arise in the absence of genetic or environmental variation.,33398
"this so development, behavioural dynamics among-individual Chaotic variation. may and influence","Chaotic dynamics may influence behavioural development, and so this among-individual variation.",33399
this methods We and experimental idea. to test discuss designs,We discuss methods and experimental designs to test this idea.,33400
"Ultimately, development much nonlinear behavioural may explain natural of variation. chaotic and","Ultimately, nonlinear and chaotic behavioural development may explain much of natural variation.",33401
We approach. these using texture propose classification to abnormalities detect a,We propose to detect these abnormalities using a texture classification approach.,33402
networks. a two is cascade method convolutional of Our neural,Our method is a cascade of two convolutional neural networks.,33403
the tissues. of detects network The presence abnormal first,The first network detects the presence of abnormal tissues.,33404
a with approach. neural method We single and random network compare our forest,We compare our method with random forest and a single neural network approach.,33405
"have performance demonstrated the in correlation filters tracking. excellent Recently, visual","Recently, correlation filters have demonstrated the excellent performance in visual tracking.",33406
"the is Region(IR). object training region However, the region,including than Interference base the sample larger","However, the base training sample region is larger than the object region,including the Interference Region(IR).",33407
"(RFCT) Tracking problem. address this paper, to Correlation we propose novel Region-filtering the this In","In this paper, we propose the novel Region-filtering Correlation Tracking (RFCT) to address this problem.",33408
spatial into training standard filter CF formulation. the immediately a samples We by introducing map,We immediately filter training samples by introducing a spatial map into the standard CF formulation.,33409
"CS-MRI. To for end, network a error correction deep (DECN) we propose this","To this end, we propose a deep error correction network (DECN) for CS-MRI.",33410
serve guiding as Existing can for template module reconstruction. the algorithms CS-MRI the,Existing CS-MRI algorithms can serve as the template module for guiding the reconstruction.,33411
approach evaluated proposed benchmark datasets. was on several The,The proposed approach was evaluated on several benchmark datasets.,33412
morphing this approach new Coherent Iterative Point We (ICPD). Drift call,We call this new morphing approach Iterative Coherent Point Drift (ICPD).,33413
to shown The framework give proposed is state-of-the-art performance.,The proposed framework is shown to give state-of-the-art performance.,33414
two could and approaches broadly incremental global. Structure-from-Motion be classes: into divided,Structure-from-Motion approaches could be broadly divided into two classes: incremental and global.,33415
"geometry epipolar into is communities. the separate partitioned First, graph","First, the epipolar geometry graph is partitioned into separate communities.",33416
"for community each the reconstruction Then, solved in parallel. problem is","Then, the reconstruction problem is solved for each community in parallel.",33417
image synthesizing models surrounding automatically regions success generative shown missing context. have in Deep using,Deep generative models have shown success in automatically synthesizing missing image regions using surrounding context.,33418
"with what users directly decide cannot approaches. to content synthesize However, such","However, users cannot directly decide what content to synthesize with such approaches.",33419
X-ray complex percutaneous alternatives enables procedures. guidance to image,X-ray image guidance enables percutaneous alternatives to complex procedures.,33420
common in tasks One of semantic the is medical imaging segmentation. most,One of the most common tasks in medical imaging is semantic segmentation.,33421
Several to networks convolutional images. of neural have variations deep successfully applied been medical,Several variations of deep convolutional neural networks have been successfully applied to medical images.,33422
"both chemistry In relationship and mysterious. underlying biology the between cases, emergent remains","In both cases, the relationship between underlying chemistry and emergent biology remains mysterious.",33423
of regulators. the The SWI/SNF Background: family critical remodelers transcriptional chromatin are,Background: The chromatin remodelers of the SWI/SNF family are critical transcriptional regulators.,33424
neural (CNNs) demonstrated have (Re-ID). in convolutional networks dominant re-identification person performance Deep,Deep convolutional neural networks (CNNs) have demonstrated dominant performance in person re-identification (Re-ID).,33425
introduce surfaces detecting RGB-D We a SLAM designed novel for patch descriptor in reconstruction. coplanar,We introduce a novel RGB-D patch descriptor designed for detecting coplanar surfaces in SLAM reconstruction.,33426
"speeds, video distinct find of we types in input architectures. network variations two","variations in input video speeds, we find two distinct types of network architectures.",33427
a by We the do model. given so inputs resampling to adaptively,We do so by adaptively resampling the inputs to a given model.,33428
require sizes special fast enrollment database can to searches. optimizations in achieve Large order,Large enrollment sizes can require special optimizations in order to achieve fast database searches.,33429
"two sharing common features. Besides, the subnetworks by these cascaded are","Besides, these two subnetworks are cascaded by sharing the common features.",33430
for a as VeSC-CoL replace quality measure uses model cardinality to space cross-validation. version,VeSC-CoL uses version space cardinality as a measure for model quality to replace cross-validation.,33431
target when that allowed. Experiments accurately concept the resource learn show computational can is VeSC-CoL,Experiments show that VeSC-CoL can accurately learn the target concept when computational resource is allowed.,33432
"we on this context-aware framework kernel In a paper, novel learning. design deep based introduce","In this paper, we introduce a novel context-aware kernel design framework based on deep learning.",33433
weights context network geometric method Our spatial the (DN). discriminatively a learns deep as of,Our method discriminatively learns spatial geometric context as the weights of a deep network (DN).,33434
have and theorists on this build continued to investigators approach. Countless,Countless investigators and theorists have continued to build on this approach.,33435
the nature workings systems. by the determined interactions And their of inner their is of,And the nature of their interactions is determined by the inner workings of their systems.,33436
assessments on microscopic diagnostic cell of samples to Pathologists the characteristic basis analyze make features.,Pathologists analyze microscopic samples to make diagnostic assessments on the basis of characteristic cell features.,33437
image detection automated paper for algorithm This based leukemia acute presents an systems.,This paper presents an algorithm for automated image based acute leukemia detection systems.,33438
"methods presented. In clustering particular, deep been embeddings have employ new that","In particular, new clustering methods that employ deep embeddings have been presented.",33439
semi-supervised constitute a to with labeled Our pair-based information model pairs the augmenting framework. allows,Our pair-based model allows augmenting the information with labeled pairs to constitute a semi-supervised framework.,33440
object two mainly There detectors. of are state-of-the-art types,There are mainly two types of state-of-the-art object detectors.,33441
"highest Such but the are slower. accuracy models typically reach rates,","Such models reach the highest accuracy rates, but are typically slower.",33442
"lower object faster much than detectors. are accuracy rates, but models two-stage reach Such","Such models reach lower accuracy rates, but are much faster than two-stage object detectors.",33443
prioritization algorithm and executed approach MVS of two before steps. operates is Our consists the,Our prioritization approach operates before the MVS algorithm is executed and consists of two steps.,33444
"we the (i.e. view rank second step, the clusters In resulting","In the second step, we rank the resulting view clusters (i.e.",33445
cultural preservation family and houses. heritage reconstruction of single,cultural heritage preservation and reconstruction of single family houses.,33446
absence The event-based large datasets. of real-world,The absence of large real-world event-based datasets.,33447
paper we problems. In both address this,In this paper we address both problems.,33448
"new novel a we a learning together feature event-based with representation architecture. First, machine introduce","First, we introduce a novel event-based feature representation together with a new machine learning architecture.",33449
"release for dataset object Second, first real-world event-based the classification. we large","Second, we release the first large real-world event-based dataset for object classification.",33450
"and expensive a is mask detailed However, segmentation obtaining time-consuming.","However, obtaining a detailed segmentation mask is expensive and time-consuming.",33451
"a this eigendecomposition ability to requires the However, performing network differentiate operation. within","However, performing eigendecomposition within a network requires the ability to differentiate this operation.",33452
"in numerical process theoretically Unfortunately, instability optimization in this while introduces practice. the doable,","Unfortunately, while theoretically doable, this introduces numerical instability in the optimization process in practice.",33453
"from Vision escape estimation, pose human trend. tasks, did as this not such","Vision tasks, such as human pose estimation, did not escape from this trend.",33454
convolutional a neural networks top linear regression layer. with,convolutional neural networks with a linear regression top layer.,33455
of comprehensive analysis deep is techniques. the This first regression,This is the first comprehensive analysis of deep regression techniques.,33456
"which, to Our a according hypothesis the (e.g. network reinforce general, results in general-purpose","Our results reinforce the hypothesis according to which, in general, a general-purpose network (e.g.",33457
manipulation or image applications BSD-GAN. We by demonstrate improved further and enabled several generation,We further demonstrate several image generation and manipulation applications enabled or improved by BSD-GAN.,33458
the is often. sequences in viewpoint a changes acquired because This drones concern by,This is a concern in sequences acquired by drones because the viewpoint changes often.,33459
to critical insult neuro-degenerative first is the in Studying multiple understanding implicated disorders. this phenomenon,Studying this phenomenon is critical to understanding the first insult implicated in multiple neuro-degenerative disorders.,33460
voltage on results axon that high sinusoidally hillock Numerical fluctuated only. water simulation indicate potential,Numerical simulation results indicate that axon hillock water potential fluctuated sinusoidally on high voltage only.,33461
"caused and trend resistance, Fluctuations the and in calcium storage amplitude respectively. were flow by","Fluctuations in the amplitude and trend were caused by calcium flow and storage resistance, respectively.",33462
in the The potential. change water in proportional water axon hillock-stored to rate change was,The change in axon hillock-stored water was proportional to the change rate in water potential.,33463
"fashion, procedure in end-to-end The be an overall resulting can in fine-tuned better performance.","The overall procedure can be fine-tuned in an end-to-end fashion, resulting in better performance.",33464
visual spatio-temporal data. patterns of Videos express structured highly,Videos express highly structured spatio-temporal patterns of visual data.,33465
"Based a generative future prediction. propose and generation we intuition, framework this on for video","Based on this intuition, we propose a generative framework for video generation and future prediction.",33466
a occurs This over mode populations. microbial wide of range for parameters evolutionary appropriate,This mode occurs over a wide range of evolutionary parameters appropriate for microbial populations.,33467
superlinear that sexual cost. the eliminates generate mode evolution distinct can of a Recombination,Recombination can generate a distinct mode of sexual evolution that eliminates the superlinear cost.,33468
"mmWave very UAVs. available communications using is limited for literature %However, there","%However, there is very limited literature available for mmWave communications using UAVs.",33469
multipath into the components. and non-persistent components~(MPCs) received persistent classify We,We classify the received multipath components~(MPCs) into persistent and non-persistent components.,33470
"deep methods learning applied recent super-resolution been have single-image years, tasks. In successfully to","In recent years, deep learning methods have been successfully applied to single-image super-resolution tasks.",33471
"upon In an architecture that design we implements a network. cascading a mechanism residual detail,","In detail, we design an architecture that implements a cascading mechanism upon a residual network.",33472
further of proposed models variant efficiency. present cascading the improve also residual to network We,We also present variant models of the proposed cascading residual network to further improve efficiency.,33473
code publicly are complete available. for system The the source,The source code for the complete system are publicly available.,33474
published algorithms the In pace. new community an at are deep-learning incredible,In the deep-learning community new algorithms are published at an incredible pace.,33475
"the the At design are and upper used time, Qt to same application. OpenGL","At the same time, Qt and OpenGL are used to design the upper application.",33476
"the functional provided. of Finally, software verification a control is","Finally, a functional verification of the control software is provided.",33477
introduce OxUvA and evaluating We tracking the single-object algorithms. benchmark for dataset,We introduce the OxUvA dataset and benchmark for evaluating single-object tracking algorithms.,33478
created deep FRN networks. features by hierarchical exploits multi-scale convolutional neural the,FRN exploits the multi-scale hierarchical features created by deep convolutional neural networks.,33479
method and weighting effectively. use enables proposed FRN features alternatively channel to The make of,The proposed channel weighting method enables FRN to make use of features alternatively and effectively.,33480
state-of-the-art that has method Experiments SCUT-HEAD. show our on achieved performance,Experiments show that our method has achieved state-of-the-art performance on SCUT-HEAD.,33481
phase contains steps. key training three The,The training phase contains three key steps.,33482
revisit of the in the vehicles. depth autonomous We problem context visual estimation of,We revisit the problem of visual depth estimation in the context of autonomous vehicles.,33483
"result, our stereo. focus a estimation by on As we depth efforts","As a result, we focus our efforts on depth estimation by stereo.",33484
systems. essential of Accurate is tracking part camera an localization,Accurate camera localization is an essential part of tracking systems.,33485
"are However, by greatly affected results localization illumination.","However, localization results are greatly affected by illumination.",33486
"and this is consuming. very tedious time However,","However, this is very tedious and time consuming.",33487
"the To gap, this Transform of introduce we (REST)."" impact ""REal-to-Synthetic reduce","To reduce the impact of this gap, we introduce ""REal-to-Synthetic Transform (REST).""",33488
is an their counterpart. real synthetic REST converts to network features that autoencoder-like,REST is an autoencoder-like network that converts real features to their synthetic counterpart.,33489
converted the can accumulated features against localization. then database camera for matched be robust The,The converted features can then be matched against the accumulated database for robust camera localization.,33490
of of is an complete The depth to the our goal image. RGB-D channel work,The goal of our work is to complete the depth channel of an RGB-D image.,33491
"bright, depth cameras fail transparent, for sense to depth Commodity-grade distant often surfaces. and shiny,","Commodity-grade depth cameras often fail to sense depth for shiny, bright, transparent, and distant surfaces.",33492
"data-driven a to this introduce approach In problem. paper, we this","In this paper, we introduce a data-driven approach to this problem.",33493
approach. effectiveness our demonstrate to are Numerical experiments conducted of the,Numerical experiments are conducted to demonstrate the effectiveness of our approach.,33494
"four alignments each. samples Our algorithm taxa involving small, first gap-free","Our algorithm first samples small, gap-free alignments involving four taxa each.",33495
approach of quality. with calculated that high Experimental show are our trees results,Experimental results show that trees calculated with our approach are of high quality.,33496
each proposed Previous methods the tasks independently. have works of three for solving tailored,Previous works have proposed tailored methods for solving each of the three tasks independently.,33497
publicly code All paper be in will available. this the,All the code in this paper will be publicly available.,33498
"call We propose we novel $\textit{cluttered a Omniglot}$. which dataset,","We propose a novel dataset, which we call $\textit{cluttered Omniglot}$.",33499
two pathways. proposed based The method contains CNN convolutional,The proposed CNN based method contains two convolutional pathways.,33500
banks multiple pathway first filter parallel The multiple MR to modalities. catering consists of convolutional,The first pathway consists of multiple parallel convolutional filter banks catering to multiple MR modalities.,33501
comparison publicly The against lesion is four MS segmentation methods. available performed,The comparison is performed against four publicly available MS lesion segmentation methods.,33502
"are economical. computationally Compared convolutions these ""fully-connected"" with more standard convolutions,","Compared with standard ""fully-connected"" convolutions, these convolutions are more computationally economical.",33503
"inter-group degradation. severe which exchange, induces the However, performance ""sparsely-connected"" information block convolutions","However, ""sparsely-connected"" convolutions block the inter-group information exchange, which induces severe performance degradation.",33504
"networks MENet family a we propose based neural compact modules. ME Finally, called on of","Finally, we propose a family of compact neural networks called MENet based on ME modules.",33505
comparisons date. analyses histological and lacking Direct are to between quantitative imaging,Direct quantitative comparisons between imaging and histological analyses are lacking to date.,33506
transmission in the channels of strength finding. account this the can relative Differences for,Differences in the relative strength of the transmission channels can account for this finding.,33507
"remarkable neural demonstrated have convolutional deep networks image single (CNNs) been on progress Recently, super-resolution.","Recently, deep convolutional neural networks (CNNs) have been demonstrated remarkable progress on single image super-resolution.",33508
This adaptation. paper domain multi-task learning-based a presents method unsupervised novel for,This paper presents a novel multi-task learning-based method for unsupervised domain adaptation.,33509
computer of neural for vision mobile (CNN) characteristics systems. convolutional performance We study networks,We study performance characteristics of convolutional neural networks (CNN) for mobile computer vision systems.,33510
such proven to CNNs to have systems. powerful be implement and a efficient approach,CNNs have proven to be a powerful and efficient approach to implement such systems.,33511
The real real conducted processing using measurements and are workloads platforms.,The measurements are conducted using real workloads and real processing platforms.,33512
"on On TensorFlow and TensorRT. side, concentrate platform especially we the","On the platform side, we concentrate especially on TensorFlow and TensorRT.",33513
significant there but show We the behavior that exists trade-offs latency--throughput is very complex.,We show that there exists significant latency--throughput trade-offs but the behavior is very complex.,33514
affect and behavior. We that factors yield the discuss several this performance demonstrate complex and,We demonstrate and discuss several factors that affect the performance and yield this complex behavior.,33515
to truth. a help ground proxy generate,to help generate a proxy ground truth.,33516
many of are data composed types in online fashion items Outfits different of (e.g.,Outfits in online fashion data are composed of items of many different types (e.g.,33517
"one some top, another. shoes) bottom, with that share stylistic relationship","top, bottom, shoes) that share some stylistic relationship with one another.",33518
attributes. two soft patch-based This and signature components: the has existing facial the features,This signature has two components: the existing patch-based features and the soft facial attributes.,33519
also is extended attributes. The to that facial different matcher so assigned different weights are,The matcher is also extended so that different weights are assigned to different facial attributes.,33520
Differently we consideration. take them into both,Differently we take them both into consideration.,33521
compared with approach approaches. state-of-the-art improves significantly performance previous the Our,Our approach significantly improves the performance compared with previous state-of-the-art approaches.,33522
We networks. low-rank to convolutional iterative applying propose neural approximation deep approach of compress,We propose iterative approach of applying low-rank approximation to compress deep convolutional neural networks.,33523
ones. method Our to of proposed the superiority non-repetitive show results compared the,Our results show the superiority of the proposed method compared to non-repetitive ones.,33524
demonstrate We less loss. higher providing compression accuracy ratio,We demonstrate higher compression ratio providing less accuracy loss.,33525
"has a filter paper, proposed. on In sparse-grid nonlinear based quadrature new this method been","In this paper, a new nonlinear filter based on sparse-grid quadrature method has been proposed.",33526
filter is Gauss-Hermite (ASGHF). The named as proposed adaptive sparse-grid filter,The proposed filter is named as adaptive sparse-grid Gauss-Hermite filter (ASGHF).,33527
with the two is nonlinear filter filtering problems. illustrated The proposed of performance,The performance of the proposed filter is illustrated with two nonlinear filtering problems.,33528
a near efficiency online the This operate with full budget. to estimator enables computational predefined,This enables the online estimator to operate near full efficiency with a predefined computational budget.,33529
field. the an is Saliency in topic multimedia detection active,Saliency detection is an active topic in the multimedia field.,33530
"detection. supplies powerful saliency a Recently, for cue information depth","Recently, depth information supplies a powerful cue for saliency detection.",33531
The proposed performance and complex maintains in algorithm robustness a situations. the good shows,The proposed algorithm shows a good performance and maintains the robustness in complex situations.,33532
saliency that framework proposed approaches. the to results other Experiments' show superior is existing,Experiments' results show that the proposed framework is superior to other existing saliency approaches.,33533
DeepWarp and is under shapes robust different model tessellations.,DeepWarp is robust under different model shapes and tessellations.,33534
"methods or flaws. results unsatisfying existing easily be However, detectable slow with can generate","However, existing methods can be slow or generate unsatisfying results with easily detectable flaws.",33535
new also to a strategy reduce We annealing propose adversarial artifacts. called loss the,We also propose a new strategy called adversarial loss annealing to reduce the artifacts.,33536
their describe for several losses inpainting effectiveness. and show We further designed specifically,We further describe several losses specifically designed for inpainting and show their effectiveness.,33537
Metropolis-Hastings distribution to any asymptotically from The sample allows algorithm probability one $\pi$.,The Metropolis-Hastings algorithm allows one to sample asymptotically from any probability distribution $\pi$.,33538
than standard PM have algorithms algorithms. properties can better These,These algorithms can have better properties than standard PM algorithms.,33539
"independent variability, averaging computations, to amenable of parallel simple A consists reduce approach to estimators.","A simple approach to reduce variability, amenable to parallel computations, consists of averaging independent estimators.",33540
implementation a methodology for propose which allows correct We original this idea. of an,We propose an original methodology which allows for a correct implementation of this idea.,33541
parallel establish properties those PM algorithms We and above. available theoretical discussed for which standard,We establish theoretical properties which parallel those available for standard PM algorithms and discussed above.,33542
various demonstrate problems. inference approach the on the interest We of,We demonstrate the interest of the approach on various inference problems.,33543
in results portion data. insufficient significant training such Ignoring,Ignoring such significant portion results in insufficient training data.,33544
deep with conventionally-trained networks. biased in leads classifiers to data Training under-represented,Training with under-represented data leads to biased classifiers in conventionally-trained deep networks.,33545
distribution the to be regular closer distribution. to encourages This the under-represented,This encourages the under-represented distribution to be closer to the regular distribution.,33546
datasets consuming. masks with labeling Manually extremely is time object,Manually labeling datasets with object masks is extremely time consuming.,33547
CNN alternating and step. fusion a temporal step a algorithm This proceeds between feed-forward by,This algorithm proceeds by alternating between a temporal fusion step and a feed-forward CNN step.,33548
"each provide learning combine stereo ray deep label. semantic we and matching Specifically, to a","Specifically, we combine deep learning and stereo matching to provide each ray a semantic label.",33549
weighting for then We blending tailored the design schemes rays.,We then design tailored weighting schemes for blending the rays.,33550
"foreground on residues Although background. focusing simple, our when can remove the effectively solution","Although simple, our solution can effectively remove foreground residues when focusing on the background.",33551
"focal At depths. same smooth the in varying transitions maintains SST time,","At the same time, SST maintains smooth transitions in varying focal depths.",33552
the unit. data from been to site wirelessly has control transmitted the The,The data has been transmitted wirelessly from the site to the control unit.,33553
We present attributes. to modeling a visual approach new,We present a new approach to modeling visual attributes.,33554
"operators. model attributes propose Instead, we as to","Instead, we propose to model attributes as operators.",33555
We on demonstrate challenging improvements approach validate and over the significant state-of-the-art. datasets our two,We validate our approach on two challenging datasets and demonstrate significant improvements over the state-of-the-art.,33556
calibration. then camera-projector for correspondence is used Such initial,Such correspondence is then used for initial camera-projector calibration.,33557
We and minimal entailment. both consider entailment rational,We consider both rational entailment and minimal entailment.,33558
of visual concise abstractions a tasks. Semantic keypoints for variety provide understanding,Semantic keypoints provide concise abstractions for a variety of visual understanding tasks.,33559
"generalized be we show representation to Finally, novel categories. can keypoint that our category-agnostic","Finally, we show that our category-agnostic keypoint representation can be generalized to novel categories.",33560
used great Neural Network(CNN) been success. widely with Convolutional has for image recognition,Convolutional Neural Network(CNN) has been widely used for image recognition with great success.,33561
"based image the paradigm. recognition there However, number of limitations a CNN are of current","However, there are a number of limitations of the current CNN based image recognition paradigm.",33562
"lacks dealing with computational images the with sizes. different Second, for scalability it","Second, it lacks the computational scalability for dealing with images with different sizes.",33563
"other any training Unlike does involve proposed approaches, not extraction. or most the technique feature","Unlike most other approaches, the proposed technique does not involve any training or feature extraction.",33564
a are sets in image The subspaces as represented high space. dimensional gallery,The gallery image sets are represented as subspaces in a high dimensional space.,33565
the then test on gallery projected set Images subspaces. the of are,Images of the test set are then projected on the gallery subspaces.,33566
machine data. Most the of and methods learning are exploit training known biases capture to,Most machine learning methods are known to capture and exploit biases of the training data.,33567
"beneficial are some While others for are biases harmful. learning,","While some biases are beneficial for learning, others are harmful.",33568
of In we work generation words gender-specific caption this investigate (e.g.,In this work we investigate generation of gender-specific caption words (e.g.,33569
"person's man, or the context. woman) the on image appearance based","man, woman) based on the person's appearance or the image context.",33570
migraine. common Epilepsy after is second the brain most disorder,Epilepsy is the second most common brain disorder after migraine.,33571
epileptic quality of seizures can patients' considerably Automatic detection the of improve life.,Automatic detection of epileptic seizures can considerably improve the patients' quality of life.,33572
Electroencephalogram systems Current challenges detection (EEG)-based encounter in situations. seizure real-life many,Current Electroencephalogram (EEG)-based seizure detection systems encounter many challenges in real-life situations.,33573
and vary recording patients non-stationary EEGs patterns signals are seizure across The sessions. and,The EEGs are non-stationary signals and seizure patterns vary across patients and recording sessions.,33574
"function Third, for Softmax representations fed classification. into and are training these","Third, these representations are fed into Softmax function for training and classification.",33575
"our conditions. robust be Furthermore, to shown approach is in and real-life noisy","Furthermore, our approach is shown to be robust in noisy and real-life conditions.",33576
have face manipulations editing a been The approach. state-of-the-art generated with,The manipulations have been generated with a state-of-the-art face editing approach.,33577
order exceeds magnitude. video all an of existing least It datasets manipulation at by,It exceeds all existing video manipulation datasets by at least an order of magnitude.,33578
away from cup grasped looking a after gaze at moving a bottle.,looking at a cup after moving gaze away from a grasped bottle.,33579
the Understanding mechanism has a endeavor subjective in underlying cognitive-neuroscience. central neural become representation,Understanding the neural mechanism underlying subjective representation has become a central endeavor in cognitive-neuroscience.,33580
"implications. its Here, hypothesis and envisage this discuss we","Here, we discuss this hypothesis and envisage its implications.",33581
"become a deep-networks-based leading has for (deep approach hashing) large-scale Recently, hashing image retrieval.","Recently, deep-networks-based hashing (deep hashing) has become a leading approach for large-scale image retrieval.",33582
"with and downlink harvested shared uplink energy users. both Hence, is","Hence, harvested energy is shared with both downlink and uplink users.",33583
blocks. multiple stacked of composed is MLFN,MLFN is composed of multiple stacked blocks.,33584
task a images vision. processing of is Noisy computer fundamental,Noisy images processing is a fundamental task of computer vision.,33585
multi-scale approach of image. Their utilizes a partitioning the binary,Their approach utilizes a multi-scale binary partitioning of the image.,33586
a Our GPU. on of milliseconds a runtime image noisy is,Our runtime of a noisy image is milliseconds on a GPU.,33587
problems prove classical it classification the on of image denoising. objects and We noisy,We prove it on the problems of noisy objects classification and classical image denoising.,33588
loss. show out can denoising multi-scale edge preservation carried novel a by We that be,We show that multi-scale denoising can be carried out by a novel edge preservation loss.,33589
"standard input. loss functions we over partial contrast, regularizers integrate directly In such into the","In contrast, we integrate such standard regularizers directly into the loss functions over partial input.",33590
paper and proposes regularization MRF/CRF compares terms. experimentally This different losses integrating,This paper proposes and experimentally compares different losses integrating MRF/CRF regularization terms.,33591
explicit regularization We losses or with earlier juxtapose methods steps using layers. our proposal-generation regularized,We juxtapose our regularized losses with earlier proposal-generation methods using explicit regularization steps or layers.,33592
semantic Our quality. approach full-supervision accuracy with achieves in near state-of-the-art segmentation,Our approach achieves state-of-the-art accuracy in semantic segmentation with near full-supervision quality.,33593
"(AE). system, In by constrain our an the Autoencoder generator we","In our system, we constrain the generator by an Autoencoder (AE).",33594
"Importantly, propose the distance generator. constraints to we improve novel two","Importantly, we propose two novel distance constraints to improve the generator.",33595
constraint from generator the use this collapse. to explicitly mode We prevent,We use this constraint to explicitly prevent the generator from mode collapse.,33596
modalities. populations We acquired stored images by dermatoscopic from collected different and different,We collected dermatoscopic images from different populations acquired and stored by different modalities.,33597
benchmark dataset be with for experts. machine learning comparisons can This human used for and,This benchmark dataset can be used for machine learning and for comparisons with human experts.,33598
network develop multi-scale estimation. robust human pose neural structure-aware We a for,We develop a robust multi-scale structure-aware neural network for human pose estimation.,33599
learning training hard advantages enhances samples. on keypoint The these masking occlusion focus to,The keypoint masking training enhances these advantages to focus learning on hard occlusion samples.,33600
state-of-the-art methods. in MPII achieves method the the the challenge leaderboard Our leading among position,Our method achieves the leading position in the MPII challenge leaderboard among the state-of-the-art methods.,33601
progress patch-based very Scene-agnostic challenging in remains inpainting despite methods. visual,Scene-agnostic visual inpainting remains very challenging despite progress in patch-based methods.,33602
et with optimization-based of refinement Yang the Combined al.,Combined with the optimization-based refinement of Yang et al.,33603
attribute achieved great on success. Recent face transfer studies have,Recent studies on face attribute transfer have achieved great success.,33604
to an able transfer of input with are image. face lot attributes A models,A lot of models are able to transfer face attributes with an input image.,33605
stochastic their identical serial defined Standard independence times. and by models distributions are processing of,Standard serial models are defined by stochastic independence and identical distributions of their processing times.,33606
intercompletion times Processing the are time models in serial statistics. identical the to,Processing times in the serial models are identical to the intercompletion time statistics.,33607
"and times. models identical Similarly, parallel standard stochastically processing independent assume","Similarly, standard parallel models assume stochastically independent and identical processing times.",33608
completion the Their equivalent are known to as statistic total processing times times.,Their processing times are equivalent to the statistic known as total completion times.,33609
"neither standard distributions. the typically models intercompletion times independence nor Comparably, predict of identical parallel","Comparably, standard parallel models typically predict neither independence of the intercompletion times nor identical distributions.",33610
"Nevertheless, that dramatically channels of can defeat hazard functions the tendency. increasing","Nevertheless, dramatically increasing hazard functions of the channels can defeat that tendency.",33611
"explicit between Therefore, these classes. model for testing set findings the our stage","Therefore, our findings set the stage for explicit model testing between these classes.",33612
start Block-Q. we right First giving the,First we start giving the right Block-Q.,33613
matrix and for decomposition D. spectral (Quotient Difference) factorization. algorithm polynomial,D. (Quotient Difference) algorithm for spectral decomposition and matrix polynomial factorization.,33614
on video image We both benchmark ReID datasets. experiments and based conduct extensive,We conduct extensive experiments on both image and video based ReID benchmark datasets.,33615
state-of-the-art Experimental our the to methods. of results demonstrate significant the approach advantages compared,Experimental results demonstrate the significant advantages of our approach compared to the state-of-the-art methods.,33616
vision important based outdoor systems. removal robustness is improving Rain the for of,Rain removal is important for improving the robustness of outdoor vision based systems.,33617
CNNs setting. the readily point PCNN to adapting allows cloud their image generalizes and architectures,PCNN generalizes image CNNs and allows readily adapting their architectures to the point cloud setting.,33618
"rear driving. human and side-view For safe mirrors for drivers, vital is having","For human drivers, having rear and side-view mirrors is vital for safe driving.",33619
more complete car. the a of what deliver They view is around happening,They deliver a more complete view of what is happening around the car.,33620
Human navigation. map drivers heavily mental exploit for also their,Human drivers also heavily exploit their mental map for navigation.,33621
information of lack renders intractable. task quite self-driving This the,This lack of information renders the self-driving task quite intractable.,33622
and drivers. steering speed) angle human by,steering angle and speed) by human drivers.,33623
visually We directly intelligent agent. task modeling introduce a the of,We introduce the task of directly modeling a visually intelligent agent.,33624
on related intelligence. typically focuses visual vision Computer subtasks to solving various,Computer vision typically focuses on solving various subtasks related to visual intelligence.,33625
agent. directly model the takes and the of Our predicts information input actions as visual,Our model takes visual information as input and directly predicts the actions of the agent.,33626
"case This effect resulting dramatic small is especially in baselines, in important of failures.","This effect is especially important in case of small baselines, resulting in dramatic failures.",33627
network work on acoustic This delay the first non-reciprocal low-loss switched lines. demonstrates based,This work demonstrates the first non-reciprocal network based on switched low-loss acoustic delay lines.,33628
design and time-domain process with the The employs results simulation circuit match full well measurements.,The design process employs time-domain full circuit simulation and the results match well with measurements.,33629
this of contributions The three-fold. are work,The contributions of this work are three-fold.,33630
"targeted large-scale develop visual indoor environments. for we First, method a new localization","First, we develop a new large-scale visual localization method targeted for indoor environments.",33631
that of sets change gestures as over vectors be These time. can represented feature,These gestures can be represented as sets of feature vectors that change over time.,33632
be expensive. and might dense multiple point often requires Accurate computationally frames of computation tracks,Accurate computation of dense point tracks often requires multiple frames and might be computationally expensive.,33633
"including reconstruction learns texture-dependent and HDM-Net shading different deformations, surface cues contours.","HDM-Net learns different reconstruction cues including texture-dependent surface deformations, shading and contours.",33634
data. of step Dimension reduction in high-dimensional often the analysis an is important,Dimension reduction is often an important step in the analysis of high-dimensional data.,33635
to data. popular find is low-dimensional best approximation high-dimensional a of the PCA technique,PCA is a popular technique to find the best low-dimensional approximation of high-dimensional data.,33636
"data. atypical very PCA However, classical sensitive to is","However, classical PCA is very sensitive to atypical data.",33637
"expensive. become However, for data computationally high-dimensional algorithms his","However, for high-dimensional data his algorithms become computationally expensive.",33638
of new The ingredients are algorithms twofold. the main,The main ingredients of the new algorithms are twofold.,33639
"allows We architecture, novel task-specific feature-level multi-task which attention. of learning a propose learning","We propose a novel multi-task learning architecture, which allows learning of task-specific feature-level attention.",33640
are construction. analyzed multi-stage AP and These the on studies gathered of concept based,These studies are analyzed and gathered based on the concept of multi-stage AP construction.,33641
"the design overcome To and to our efficiency keys are scalability. the limitations,","To overcome the limitations, the keys to our design are efficiency and scalability.",33642
competitive methods. performance state-of-the-art Extensive or show our to experimental that methods superior achieve results,Extensive experimental results show that our methods achieve competitive or superior performance to state-of-the-art methods.,33643
"presented. vision the standard tools deep-learning for are designing based methodology used systems and Then,","Then, the standard methodology and tools used for designing deep-learning based vision systems are presented.",33644
of robot be vision to This survey systems. for intended the is developers guide a,This survey is intended to be a guide for the developers of robot vision systems.,33645
"the fields computationally propose cheaper an vector from on based pre-processing way alternative, We trajectories.","We propose an alternative, computationally cheaper way based on pre-processing vector fields from the trajectories.",33646
"the camera labour is consuming. trap and expensive, intensive, images of analysis However, time","However, the analysis of camera trap images is expensive, labour intensive, and time consuming.",33647
"trained datasets, large success. recognition labeled object When shown methods have on","When trained on large labeled datasets, object recognition methods have shown success.",33648
each affinity optimal size We select semantic use for the category. adversarial learning to field,We use adversarial learning to select the optimal affinity field size for each semantic category.,33649
results obtain the final reconstruction. to the fuse We then,We then fuse the results to obtain the final reconstruction.,33650
that be for non-invasive used cardiac activity can monitoring. (SCG) method a is Seismocardiography,Seismocardiography (SCG) is a non-invasive method that can be used for cardiac activity monitoring.,33651
against ECG measurements. the tested was standard proposed performance The of the algorithm,The performance of the proposed algorithm was tested against the standard ECG measurements.,33652
"are web-technology Browser-compatible, therefore based required. solutions","Browser-compatible, web-technology based solutions are therefore required.",33653
"scope, prototype. Within provide we first this with VIOLA a","Within this scope, with VIOLA we provide a first prototype.",33654
these and fully model. Distributed be Turing's by properties lack cannot captured representations standard,Distributed representations lack these properties and cannot be fully captured by Turing's standard model.,33655
"and extensionally and are entropic. the are expressed extended functions abstractions machine, computations In","In the extended machine, functions and abstractions are expressed extensionally and computations are entropic.",33656
walk AffinityNet. by then propagation with The the affinities semantic is predicted random realized by,The semantic propagation is then realized by random walk with the affinities predicted by AffinityNet.,33657
sons. training works of subset Existing large deep learn unique using a per- representations,Existing works learn deep representations using a large training subset of unique per- sons.,33658
"unseen However, persons critical algorithm. re-ID a for is good identifying","However, identifying unseen persons is critical for a good re-ID algorithm.",33659
"learning robust step Finally, criminative matching a enables on subspace. a distance dis- metric","Finally, a metric learning step enables robust distance matching on a dis- criminative subspace.",33660
a proposes signals. for algorithm seismocardiographic This adaptive extraction (SCG) paper novel feature,This paper proposes a novel adaptive feature extraction algorithm for seismocardiographic (SCG) signals.,33661
"example, versa. vice and when For are bins the signal shorter is steeper, the variation","For example, when the signal variation is steeper, the bins are shorter and vice versa.",33662
"bins was non-adaptive algorithm, The especially the outperformed as number proposed algorithm of the reduced.","The proposed algorithm outperformed the non-adaptive algorithm, especially as the number of bins was reduced.",33663
"Mask recently, like methods More jointly. R-CNN deep them perform learning","More recently, deep learning methods like Mask R-CNN perform them jointly.",33664
Depthwise convolutions performance significant in mult-adds. the to owing benefits parameters and provide both reduction,Depthwise convolutions provide significant performance benefits owing to the reduction in both parameters and mult-adds.,33665
five popular in training our deep learning frameworks. We method have implemented,We have implemented our training method in five popular deep learning frameworks.,33666
this implementations methods. paper we these In FPGA present of,In this paper we present FPGA implementations of these methods.,33667
access a proposes This a for communication (NOMA) multiple paper cooperative system. non-orthogonal new protocol,This paper proposes a new communication protocol for a cooperative non-orthogonal multiple access (NOMA) system.,33668
been segmentation standard. proved segmentation proposed by manual Many as gold techniques and have,Many segmentation techniques have been proposed and proved by manual segmentation as gold standard.,33669
selected demonstrate results effectiveness methods. Our of the experimental,Our experimental results demonstrate the effectiveness of selected methods.,33670
within a scheme pooling is trainable end-to-end deep framework. Our,Our pooling scheme is end-to-end trainable within a deep framework.,33671
"to After at decreased the compared spTMS, medium contraction. HFD and both SampEn mild","After spTMS, both HFD and SampEn decreased at medium compared to the mild contraction.",33672
"SampEn HFD strong at medium compared significantly not contraction. increased, while did change to","SampEn increased, while HFD did not change significantly at strong compared to medium contraction.",33673
decreased all contraction parameters significantly spTMS levels. both at,spTMS significantly decreased both parameters at all contraction levels.,33674
of result. confirmations other obtained scientific to Consistent a independently each credibility lend,Consistent confirmations obtained independently of each other lend credibility to a scientific result.,33675
"developed. been To (AL) methods have Learning many tackle Active this problem,","To tackle this problem, many Active Learning (AL) methods have been developed.",33676
Neuronal in reside bodies cortex. mostly cerebral the cell,Neuronal cell bodies mostly reside in the cerebral cortex.,33677
"popular process. spherical costly geometrical inflations, simplifications, but a on rely approaches as Current such","Current approaches rely on geometrical simplifications, such as spherical inflations, a popular but costly process.",33678
are Surface between brain indeed bases geometries. incompatible,Surface bases are indeed incompatible between brain geometries.,33679
bases. surface direct approach This compatible enables data of novel a across learning surface,This novel approach enables a direct learning of surface data across compatible surface bases.,33680
filters over surface spectral neighborhoods. It of representations intrinsic exploits,It exploits spectral filters over intrinsic representations of surface neighborhoods.,33681
application to the parcellation. benefits illustrate brain this with of We approach an,We illustrate the benefits of this approach with an application to brain parcellation.,33682
people video non-overlapping matches person Video-based of cameras. across re-identification clips,Video-based person re-identification matches video clips of people across non-overlapping cameras.,33683
"practice, are which partially features. occluded, the people corrupt extracted often In can","In practice, people are often partially occluded, which can corrupt the extracted features.",33684
shown experiments. are outdoor and several Results indoor for,Results are shown for several indoor and outdoor experiments.,33685
"per auto-encoder one phase, is pre-training category. the In expert trained","In the pre-training phase, one expert auto-encoder is trained per category.",33686
applied to vision widely been have methods generation many proposal tasks. computer Object,Object proposal generation methods have been widely applied to many computer vision tasks.,33687
"the TOPG method result, significantly As a recall proposed of the increased. is","As a result, the recall of the proposed TOPG method is significantly increased.",33688
"the strategy a face for such detection. computational increases burden However,","However, such a strategy increases the computational burden for face detection.",33689
each represents of of are not image Images sets a web relationships. objects: simply interconnected,Images are not simply sets of objects: each image represents a web of interconnected relationships.,33690
noise. method shown been an adversarial for has effective be to reducing JPEG also,JPEG has been also shown to be an effective method for reducing adversarial noise.,33691
"novel accurate and paper video memory-based a proposes predictive. is online that This efficient, representation","This paper proposes a novel memory-based online video representation that is efficient, accurate and predictive.",33692
(MTMCT) video cameras. several people Tracking Multi-Target many from Multi-Camera tracks through taken,Multi-Target Multi-Camera Tracking (MTMCT) tracks many people through video taken from several cameras.,33693
neural learn both features for and network. with good We convolutional a Re-ID MTMCT,We learn good features for both MTMCT and Re-ID with a convolutional neural network.,33694
such due The challenging models large space. extremely is to discrete state of analysis their,The analysis of such models is extremely challenging due to their large discrete state space.,33695
both discuss We efficient performance. methods for their analysis approaches compare modeling and,We discuss efficient analysis methods for both modeling approaches and compare their performance.,33696
"on positive the SR task. Intuitively, gives object a impact detection","Intuitively, SR gives a positive impact on the object detection task.",33697
detector. differentiable for to preprocessing any train This procedure allows end-to-end us training SR,This end-to-end training procedure allows us to train SR preprocessing for any differentiable detector.,33698
state-of-the-art have vision provided in Networks computer Convolutional problems. several Neural results,Convolutional Neural Networks have provided state-of-the-art results in several computer vision problems.,33699
they poor deployed yield So generalized performance ZSL after the being settings. in,So they yield poor performance after being deployed in the generalized ZSL settings.,33700
"pancreas. performance on convolutional segmentation neural have (CNNs) Recently, promising accurate of demonstrated networks","Recently, convolutional neural networks (CNNs) have demonstrated promising performance on accurate segmentation of pancreas.",33701
bifurcation transition. can this A Andronov-Hopf explain supercritical,A supercritical Andronov-Hopf bifurcation can explain this transition.,33702
prerequisite behavior. events Anticipating intelligent towards an important is future,Anticipating future events is an important prerequisite towards intelligent behavior.,33703
Video goal. been towards proxy as this has forecasting task studied a,Video forecasting has been studied as a proxy task towards this goal.,33704
a in we of Can image detect domains variety objects without common instance-level annotations?,Can we detect common objects in a variety of image domains without instance-level annotations?,33705
issue representations. existing work addresses this employing voxel-based Most by,Most existing work addresses this issue by employing voxel-based representations.,33706
"been to solve have an In this representations point problem, attempt cloud proposed.","In an attempt to solve this problem, point cloud representations have been proposed.",33707
noise a maps modeling saliency with way. deal in enables Explicit noisy to probabilistic us,Explicit noise modeling enables us to deal with noisy saliency maps in a probabilistic way.,33708
networks tasks computer vision (CNN). convolutional (CV) neural increasingly on Continuous rely,Continuous computer vision (CV) tasks increasingly rely on convolutional neural networks (CNN).,33709
Our in between is pixel motion. changes frames data observation key that visual consecutive represents,Our key observation is that changes in pixel data between consecutive frames represents visual motion.,33710
to mobile new architecture of System-on-a-Chip algorithm. the efficiency a the We co-design maximize (SoC),We co-design a mobile System-on-a-Chip (SoC) architecture to maximize the efficiency of the new algorithm.,33711
"semi-supervision. straightforward grants way a Additionally, of it using","Additionally, it grants a straightforward way of using semi-supervision.",33712
"parts human improve extracted. usually, local representation body are features To learning, from the","To improve the representation learning, usually, local features from human body parts are extracted.",33713
verbal require automatic interaction manual text does the and is Learning not fully annotations.,Learning the verbal interaction is fully automatic and does not require manual text annotations.,33714
Video understanding important Question structure. video temporal is task Answering (QA) an in,Video Question Answering (QA) is an important task in understanding video temporal structure.,33715
"a propose observations, video for network Based these comemory motion-appearance we QA.","Based these observations, we propose a motion-appearance comemory network for video QA.",33716
"deep in of techniques artificial learning, data-driven. are State-of-the-art particular mostly intelligence,","State-of-the-art techniques of artificial intelligence, in particular deep learning, are mostly data-driven.",33717
"difficult and a labeling manually dataset However, scale expensive. and both collecting large is","However, collecting and manually labeling a large scale dataset is both difficult and expensive.",33718
paper synthesized learning data This effectively. an approach for from presents,This paper presents an approach for learning from synthesized data effectively.,33719
of motivation -- The seeing increasing the data. difficult probability is straightforward training,The motivation is straightforward -- increasing the probability of seeing difficult training data.,33720
the efficiency in exploring its data space. demonstrates infinite This,This demonstrates its efficiency in exploring the infinite data space.,33721
important inputs. is gating derived dehazed final the the features image of The by yielded,The final dehazed image is yielded by gating the important features of the derived inputs.,33722
enhance to the system Rear-end driving great collision has role warning safety. a,Rear-end collision warning system has a great role to enhance the driving safety.,33723
pedestrians re-id images from match attempt camera disjoint techniques of views. the to The non-overlapping,The re-id techniques attempt to match the images of pedestrians from disjoint non-overlapping camera views.,33724
viewpoints. caused by re-id major changing serious of A variations is challenge the intra-class,A major challenge of re-id is the serious intra-class variations caused by changing viewpoints.,33725
manipulated is and part our technique. video generation with reassembled separately proposed Each,Each part is manipulated separately and reassembled with our proposed video generation technique.,33726
to way. faces attributes not need any the of annotate in proposed does Our framework,Our proposed framework does not need to annotate the attributes of faces in any way.,33727
experiments the framework. Our demonstrate proposed of efficacy the,Our experiments demonstrate the efficacy of the proposed framework.,33728
handle Such analysis. CNN memory shape performance the to allows high-resolution superior,Such superior memory performance allows the CNN to handle high-resolution shape analysis.,33729
Cityscapes extensively sets. the data approach and analyze We KITTI our evaluate on and,We extensively evaluate and analyze our approach on the KITTI and Cityscapes data sets.,33730
this work for a hard mining. unsupervised we novel example present framework In training,In this work we present a novel unsupervised framework for hard training example mining.,33731
between examples Both revealed of and disagreements Euclidean types are manifold by similarities.,Both types of examples are revealed by disagreements between Euclidean and manifold similarities.,33732
any discriminative with be examples training in can discovered The loss. used,The discovered examples can be used in training with any discriminative loss.,33733
protocols Three of new are difficulty introduced. varying,Three new protocols of varying difficulty are introduced.,33734
"a methods, using pre-processing stage. protocols dataset those different comparison allow between The fair including","The protocols allow fair comparison between different methods, including those using a dataset pre-processing stage.",33735
the state-of-the-art An benchmark. performed new comparison the methods of extensive on is,An extensive comparison of the state-of-the-art methods is performed on the new benchmark.,33736
"CNN are local-feature-based types modern from of methods. evaluated, to Different methods ranging based","Different types of methods are evaluated, ranging from local-feature-based to modern CNN based methods.",33737
two worlds. the results The by taking the best achieved are of best,The best results are achieved by taking the best of the two worlds.,33738
"far Most importantly, solved. from image retrieval appears being","Most importantly, image retrieval appears far from being solved.",33739
efficient often of development deep algorithms This are data-hungry. which learning supervised the jeopardizes,This jeopardizes the efficient development of supervised deep learning algorithms which are often data-hungry.,33740
The collection supports both scenes user-configured from scenes. and data framework auto-driving,The framework supports data collection from both auto-driving scenes and user-configured scenes.,33741
recognition increasingly applications. used in Iris is large-scale,Iris recognition is increasingly used in large-scale applications.,33742
"recognition iris for result, attack fundamental detection a As takes presentation on importance.","As a result, presentation attack detection for iris recognition takes on fundamental importance.",33743
literature This on covers the diverse survey topic. research this,This survey covers the diverse research literature on this topic.,33744
"blend two effectively a remains frames still However, challenging warped problem. the to how","However, how to effectively blend the two warped frames still remains a challenging problem.",33745
is fully convolutional end. trained to network end is Our and neural,Our neural network is fully convolutional and is trained end to end.,33746
framework We for novel iterative present a reasoning. visual,We present a novel framework for iterative visual reasoning.,33747
the missing that shows framework to reasoning. for resilient is Analysis also regions,Analysis also shows that the framework is resilient to missing regions for reasoning.,33748
the from to anti-spoofing breach. Face is recognition face security a prevent crucial step systems,Face anti-spoofing is the crucial step to prevent face recognition systems from a security breach.,33749
classification approaches a as binary Previous learning face formulate problem. deep anti-spoofing,Previous deep learning approaches formulate face anti-spoofing as a binary classification problem.,33750
them struggle adequate grasp of poorly. and cues Many spoofing generalize to,Many of them struggle to grasp adequate spoofing cues and generalize poorly.,33751
we live the estimated to fuse vs. Then rPPG and faces. spoof distinguish depth,Then we fuse the estimated depth and rPPG to distinguish live vs. spoof faces.,33752
training clean label annotations crucial for datasets (CNNs). Large-scale Networks Neural possessing are Convolutional,Large-scale datasets possessing clean label annotations are crucial for training Convolutional Neural Networks (CNNs).,33753
Our discriminative features learns in fashion. detects an labels approach noisy iterative deep and,Our approach detects noisy labels and learns deep discriminative features in an iterative fashion.,33754
"likely are most prominent differences, to or noticeable described first. differences, These be","These most noticeable differences, or prominent differences, are likely to be described first.",33755
"addition, In differences, mentioned not many present, may all. although at be","In addition, many differences, although present, may not be mentioned at all.",33756
"rater datasets. are of bias medical features, and measurements inaccurate missing real-life However, ailments typical","However, missing features, rater bias and inaccurate measurements are typical ailments of real-life medical datasets.",33757
The code and database publicly source are available.,The source code and database are publicly available.,33758
"curation is time well-labeled of cost, manual samples sufficient Within scale-prohibitive. imaging, medical and","Within medical imaging, manual curation of sufficient well-labeled samples is cost, time and scale-prohibitive.",33759
is methods. literature of There rich a variant,There is a rich literature of variant methods.,33760
"are mostly However, procedures qualitative, evaluation studies. user involving","However, evaluation procedures are qualitative, mostly involving user studies.",33761
novel a We evaluation quantitative describe procedure.,We describe a novel quantitative evaluation procedure.,33762
number a recent EC of We plots transfer style methods. construct comparing,We construct EC plots comparing a number of recent style transfer methods.,33763
a style number intriguing methods. plots transfer of recent of reveal properties EC These,These EC plots reveal a number of intriguing properties of recent style transfer methods.,33764
"used effect methods. style strong has The a outcome, all on the for","The style used has a strong effect on the outcome, for all methods.",33765
"improve and results. does worse style large necessarily effectiveness, weights produce can not Using","Using large style weights does not necessarily improve effectiveness, and can produce worse results.",33766
"evidence to evaluations. our human comparing evaluations Finally, show EC we","Finally, we show evidence comparing our EC evaluations to human evaluations.",33767
distributional or with ubiquitous constraints is Committee diversity problem. a selection,Committee selection with diversity or distributional constraints is a ubiquitous problem.,33768
"and their However, be optimize also is hard stable. often formulation to known to not","However, their formulation is also known to be hard to optimize and often not stable.",33769
its augmenting this discriminator improve By with a approach we accuracy.,By augmenting this approach with a discriminator we improve its accuracy.,33770
"incorporation the analytical These differences of impact of baselines, uncertainty. estimate and age recognition our","These differences impact our recognition of baselines, and the analytical incorporation of age estimate uncertainty.",33771
be key Two for success. turn to ingredients out,Two ingredients turn out to be key for success.,33772
"by FloorNet, tackles neural deep This this paper problem novel proposing architecture. a","This paper tackles this problem by proposing FloorNet, a novel deep neural architecture.",33773
of across features branches exploit the the exchanges architectures. intermediate the to all FloorNet best,FloorNet exchanges intermediate features across the branches to exploit the best of all the architectures.,33774
"In thrombus accurate post-operative required. this is of quantification context, volume","In this context, accurate quantification of post-operative thrombus volume is required.",33775
intermediate discarding feature. performs downsampling any PGP without,PGP performs downsampling without discarding any intermediate feature.,33776
data techniques. augmentation and complementary It works as is augmentation used data to commonly,It works as data augmentation and is complementary to commonly used data augmentation techniques.,33777
Experimental image results on the classification method. popular effectiveness proposed benchmarks based of demonstrate the,Experimental results based on popular image classification benchmarks demonstrate the effectiveness of the proposed method.,33778
recognition. by action motivated such We human the of task representations investigate,We investigate such representations motivated by the task of human action recognition.,33779
"we re-identification. (GCT) this correspondence graph approach person propose paper, In transfer a for","In this paper, we propose a graph correspondence transfer (GCT) approach for person re-identification.",33780
by The distances matching from derived is aggregating different score references.,The matching score is derived by aggregating distances from different references.,33781
subset optimal as that identified this function. set the maximizes An submodular is,An optimal subset is identified as the set that maximizes this submodular function.,33782
and set parameter. It is regularization properly to therefore their step necessary size,It is therefore necessary to properly set their step size and regularization parameter.,33783
algorithms. results illustrate the proposed effectiveness of Simulation the,Simulation results illustrate the effectiveness of the proposed algorithms.,33784
requires conditions. (ReID) Person different of Re-Identification images under comparing two captured person,Person Re-Identification (ReID) requires comparing two images of person captured under different conditions.,33785
"extract we layers Specifically, use two first, of input the to convolutional features several images.","Specifically, first, we use several convolutional layers to extract the features of two input images.",33786
"inputs. we compute maps similarity for Network Then, propose Similarity to Convolution the score the","Then, we propose Convolution Similarity Network to compute the similarity score maps for the inputs.",33787
determine transformer attention. (STNs) networks spatial spatial use We to,We use spatial transformer networks (STNs) to determine spatial attention.,33788
depth-wise efficient the propose to We apply convolution compute to similarity.,We propose to apply efficient depth-wise convolution to compute the similarity.,33789
"improved further use we Furthermore, the an loss ranking to improve performance.","Furthermore, we use an improved ranking loss to further improve the performance.",33790
editing for is video important compression. segmentation video an Instance and technique level object,Instance level video object segmentation is an important technique for video editing and compression.,33791
consideration Logic in Programming Practice Under of and Theory (TPLP).,Under consideration in Theory and Practice of Logic Programming (TPLP).,33792
two-stream We detection. for a tampering network propose face,We propose a two-stream network for face tampering detection.,33793
the newly network evaluate dataset. We proposed our on collected two-stream,We evaluate the proposed two-stream network on our newly collected dataset.,33794
of Experimental our results method. demonstrate the effectiveness,Experimental results demonstrate the effectiveness of our method.,33795
discover spaces. methods complex optimization efficiently explore sought if these modern to We,We sought to discover if modern optimization methods efficiently explore these complex spaces.,33796
with physics-based simulation a model human They a were musculoskeletal environment. provided and,They were provided with a human musculoskeletal model and a physics-based simulation environment.,33797
across persons aims views/scenes. (ReID) at different matching Person re-identification,Person re-identification (ReID) aims at matching persons across different views/scenes.,33798
"the manner. network More importantly, optimized and end-to-end an simultaneously ABC triplet are in","More importantly, the ABC and triplet network are simultaneously optimized in an end-to-end manner.",33799
by number problem on based image this of methods attention. focused A solving using have,A number of methods have focused on solving this problem by using image based attention.,33800
Humans so when this solving also do problem.,Humans also do so when solving this problem.,33801
limited drawback. this to accuracy is due The,The accuracy is limited due to this drawback.,33802
"exemplar to based an we In problem this using paper, by solve this method. propose","In this paper, we propose to solve this problem by using an exemplar based method.",33803
one obtain obtain more region. We differential exemplars opposing attention to or and a supporting,We obtain one or more supporting and opposing exemplars to obtain a differential attention region.,33804
based methods. to attention image closer This human attention than is differential other attention,This differential attention is closer to human attention than other image based attention methods.,33805
improved accuracy obtaining also answering questions. when in helps It,It also helps in obtaining improved accuracy when answering questions.,33806
datasets. on is The method challenging evaluated benchmark,The method is evaluated on challenging benchmark datasets.,33807
the is challenging remote reduction problem of sensing. a (CAR) artifacts Compression field in,Compression artifacts reduction (CAR) is a challenging problem in the field of remote sensing.,33808
have hand-crafted learning methods recent demonstrated methods. superior deep based over the performance previous Most,Most recent deep learning based methods have demonstrated superior performance over the previous hand-crafted methods.,33809
"important is interaction, challenging Textual robotics human-computer knowledge for task grounding and but mining. an","Textual grounding is an important but challenging task for human-computer interaction, robotics and knowledge mining.",33810
 seen  Recent progress semantic have years in segmentation. remarkable,  Recent years have seen remarkable progress in semantic segmentation.,33811
"Yet, remains it a apply task to video-based segmentation to applications. challenging techniques","Yet, it remains a challenging task to apply segmentation techniques to video-based applications.",33812
"autonomous video present of the challenge framework. the a segmentation design driving, significant to","autonomous driving, present a significant challenge to the design of the video segmentation framework.",33813
low latency work to Both together ensure maintaining quality. segmentation components high while,Both components work together to ensure low latency while maintaining high segmentation quality.,33814
"the However, levels features from of semantic usually different layers inconsistent. are","However, the semantic levels of features from different layers are usually inconsistent.",33815
"On proposed the state-of-the-art performance. datasets, achieves method pedestrian detection these","On these pedestrian datasets, the proposed method achieves state-of-the-art detection performance.",33816
an interactions to between be It for can image-question pair. hierarchy multi-step stacked form a,It can be stacked to form a hierarchy for multi-step interactions between an image-question pair.,33817
without used can be directly therefore any learning. TVNet extra,TVNet can therefore be used directly without any extra learning.,33818
"by further can training. Finally, the be fine-tuned TVNet parameters of end-to-end the","Finally, the parameters of the TVNet can be further fine-tuned by end-to-end training.",33819
and beyond optical exact flow. to This enables richer learn TVNet patterns task-specific,This enables TVNet to learn richer and task-specific patterns beyond exact optical flow.,33820
recognition the experiments verify effectiveness the Extensive proposed of benchmarks action two approach. on,Extensive experiments on two action recognition benchmarks verify the effectiveness of the proposed approach.,33821
"an check step heavily post-processing left-right isolated consistency is conventional the However, hand-crafted. and","However, the conventional left-right consistency check is an isolated post-processing step and heavily hand-crafted.",33822
"are by way, In maps disparity recurrent proposed model. improved the this progressively generated the","In this way, the generated disparity maps are progressively improved by the proposed recurrent model.",33823
"Thanks has cross-modal of progress success significant the made retrieval deep recently. to learning,","Thanks to the success of deep learning, cross-modal retrieval has made significant progress recently.",33824
network DVSNet network. consists of convolutional two a and flow networks: neural a segmentation,DVSNet consists of two convolutional neural networks: a segmentation network and a flow network.,33825
"The but segmentations, accurate highly former is slower. generates deeper and semantic","The former generates highly accurate semantic segmentations, but is deeper and slower.",33826
higher Frame the network. confidence flow score expected regions traverse a with,Frame regions with a higher expected confidence score traverse the flow network.,33827
through regions confidence segmentation Frame lower a to network. the score with have expected pass,Frame regions with a lower expected confidence score have to pass through the segmentation network.,33828
new method based is for analysis. networks complex A color-texture on proposed,A new method based on complex networks is proposed for color-texture analysis.,33829
adaptive threshold is selection for and also approach proposed. An automatic,An automatic and adaptive approach for threshold selection is also proposed.,33830
"methods literature with pre-trained including architectures. various are networks convolutional Results among neural compared, deep","Results among various literature methods are compared, including deep convolutional neural networks with pre-trained architectures.",33831
"each task. recent for detailed algorithms performed using state-of-the-art Then, is study a quantitative most","Then, a detailed quantitative study is performed using most recent state-of-the-art algorithms for each task.",33832
approach whether viable. proceed We this to is investigate,We proceed to investigate whether this approach is viable.,33833
MNIST is Experiments this the the for manifold that show conducted case. indeed,Experiments conducted for the MNIST manifold show that this indeed is the case.,33834
"supervision can guidance. quality improve of to Additionally, be employed discriminative further the","Additionally, discriminative supervision can be employed to further improve the quality of guidance.",33835
human-annotated algorithm ground-truth with comparing visual counts. We our evaluate by,We evaluate our algorithm by comparing with ground-truth human-annotated visual counts.,33836
"DeepMVS, for multi-view present neural stereo We deep convolutional (ConvNet) network reconstruction. a","We present DeepMVS, a deep convolutional neural network (ConvNet) for multi-view stereo reconstruction.",33837
metric conventionally errors. pixel-wise the of adopted,the conventionally adopted metric of pixel-wise errors.,33838
model ismodeled planar model. a by a Signature horizontal image with neuronal and verticalprincipal secondarymodels,Signature image ismodeled by a planar neuronal model with horizontal secondarymodels and a verticalprincipal model.,33839
The been tested proposed on methodhas two databases.,The proposed methodhas been tested on two databases.,33840
automatically classification We visual technique of classifier. for detecting a a pre-trained the develop errors,We develop a technique for automatically detecting the classification errors of a pre-trained visual classifier.,33841
"Weakly unexplored. labels, image-level segmentation masks, instead instance with pixel-level supervised of remains expensive","Weakly supervised instance segmentation with image-level labels, instead of expensive pixel-level masks, remains unexplored.",33842
"of we network distance Second, present multi-branch to the features. intra-class minimize decision a","Second, we present a multi-branch decision network to minimize the intra-class distance of features.",33843
These improvement. lightweight prominent accuracy but detection techniques complexity zero bring extra time,These lightweight techniques bring zero extra time complexity but prominent detection accuracy improvement.,33844
proposed genetic on algorithms. based approach is The,The proposed approach is based on genetic algorithms.,33845
achieved approach. the attest results the of proposed The robustness,The results achieved attest the robustness of the proposed approach.,33846
designed applications planar DC-blocker for suitable mode is fabricated. and signaling differential A,A planar DC-blocker suitable for differential mode signaling applications is designed and fabricated.,33847
"tolerances. much to fabrication sensitive the structure So, less is","So, the structure is much less sensitive to fabrication tolerances.",33848
"the tolerate higher is to DC-blocker Moreover, differences. likely DC-voltage","Moreover, the DC-blocker is likely to tolerate higher DC-voltage differences.",33849
for all in events video video. Dense generate text aims an to untrimmed descriptions captioning,Dense video captioning aims to generate text descriptions for all events in an untrimmed video.,33850
and involves describing events. detecting both This,This involves both detecting and describing events.,33851
"sub-problems. captioning a these proposal two an event model, for and","an event proposal and a captioning model, for these two sub-problems.",33852
The alternation. either in or are trained models separately,The models are either trained separately or in alternation.,33853
"model dense propose transformer end-to-end for we this problem, an video address To captioning.","To address this problem, we propose an end-to-end transformer model for dense video captioning.",33854
appropriate the video encodes The representations. encoder into,The encoder encodes the video into appropriate representations.,33855
video the encoding to decoder event with proposal anchors decodes The from different form proposals.,The proposal decoder decodes from the encoding with different anchors to form video event proposals.,33856
palette-based for color visual framework We present for a composition applications.,We present a palette-based framework for color composition for visual applications.,33857
"Color composition a in of visualization. art, is and critical visual aspect applications design,","Color composition is a critical aspect of visual applications in art, design, and visualization.",33858
geometry images of is in on based the approach Our RGBXY-space.,Our approach is based on the geometry of images in RGBXY-space.,33859
layer a We tool. demonstrate decomposition real-time,We demonstrate a real-time layer decomposition tool.,33860
"system. case, is that challenge In one evaluating the arises latter learning the","In the latter case, one challenge that arises is evaluating the learning system.",33861
of method to applying the step artificial systems. this take novel learning We,We take the novel step of applying this method to artificial learning systems.,33862
"an ill-posed, set is which is and equation many The thus solutions. have framework","The framework is an equation set which is ill-posed, and thus have many solutions.",33863
the and We strategy cross-layer prove of information cross-task by comparative the study. fusion utility,We prove the utility of the cross-layer and cross-task information fusion strategy by comparative study.,33864
cross-modal provide matching. both testing of for scenarios and These testing training dynamic and static,These provide training and testing scenarios for both static and dynamic testing of cross-modal matching.,33865
CNN matches performance The examples human (e.g. easy on,The CNN matches human performance on easy examples (e.g.,33866
more examples but different (e.g. challenging human gender faces) performance exceeds across on,different gender across faces) but exceeds human performance on more challenging examples (e.g.,33867
"nationality). the faces age with same gender, and","faces with the same gender, age and nationality).",33868
scale data. vital Learning large for representation and compact challenging multimedia is,Learning compact representation is vital and challenging for large scale multimedia data.,33869
proposed key following components. The DCVH has the,The proposed DCVH has the following key components.,33870
"image addition, competitive DCVH can annotation/tagging. provide In for performance","In addition, DCVH can provide competitive performance for image annotation/tagging.",33871
"In detection paper, a this student for new dataset introduce and we localization. engagement","In this paper, we introduce a new dataset for student engagement detection and localization.",33872
subject so performed to generalized users. that can is new independent it be The analysis,The subject independent analysis is performed so that it can be generalized to new users.,33873
modeled engagement problem of prediction as weakly learning is The problem. a supervised,The problem of engagement prediction is modeled as a weakly supervised learning problem.,33874
comparison. problem the discuss of image This color content article,This article discuss the problem of color image content comparison.,33875
considers This locations. color uses histograms the method and color,This method uses the color histograms and considers color locations.,33876
Testing performed. and algorithms analyzing of modified and based are,Testing and analyzing of based and modified algorithms are performed.,33877
method paper indices presents for a correction estimation. noise Sobol' This simple,This paper presents a simple noise correction method for Sobol' indices estimation.,33878
is features analytical One the expression is there that of actual appealing noise correction derived.,One of the appealing features is that there is actual analytical noise correction expression derived.,33879
controlling parameters remotely. the platform A software the allows optimizing and loop interface PC-based,A PC-based software interface allows controlling the platform and optimizing the loop parameters remotely.,33880
of allow of interest are Several and quantities rapidly. included smoothly to tools measurement,Several tools are included to allow measurement of quantities of interest smoothly and rapidly.,33881
existing can loss deep The frameworks. be learning easily using implemented,The loss can be easily implemented using existing deep learning frameworks.,33882
Bone decades. images been has that task CT worked on for a is from segmentation,Bone segmentation from CT images is a task that has been worked on for decades.,33883
scans on CT patients approach myeloma. of We the suffering whole-body multiple from evaluated,We evaluated the approach on whole-body CT scans of patients suffering from multiple myeloma.,33884
"social immense far-reaching surrounding However, are its also environmental impacts to areas. and and the","However, the social and environmental impacts are also immense and far-reaching to its surrounding areas.",33885
Mapping land cover critical is the impacts. /land for use tracking changed (LUCC),Mapping land use /land cover changed (LUCC) is critical for tracking the impacts.,33886
for in better topographic was area. accuracy The also information added mountain,The topographic information was also added for better accuracy in mountain area.,33887
actions in has increased Analyzing human gained attention videos recently.,Analyzing human actions in videos has gained increased attention recently.,33888
main of The method working consists stages. six,The working method consists of six main stages.,33889
appearance Cell cell motion correlation have cycle progression. strong and a with and disease,Cell motion and appearance have a strong correlation with cell cycle and disease progression.,33890
"ConvNet is method a representations, developed based for recognition. the Upon proposed action","Upon the proposed representations, a ConvNet based method is developed for action recognition.",33891
for correspondences Most to frame accurate dense video in-between synthesize interpolation frame. require approaches an,Most approaches for video frame interpolation require accurate dense correspondences to synthesize an in-between frame.,33892
"in scenarios with well challenging do they not perform e.g. Therefore,","Therefore, they do not perform well in challenging scenarios with e.g.",33893
"amount applicable motion. a limited of for only are However, they","However, they are only applicable for a limited amount of motion.",33894
to This offers chapter channel-based Bayesian to an probability the accessible approach introduction theory.,This chapter offers an accessible introduction to the channel-based approach to Bayesian probability theory.,33895
"it at scale However, intrinsic training to images. data ground difficult for is truth collect","However, it is difficult to collect ground truth training data at scale for intrinsic images.",33896
time be views. to can at then test trained model applied single Our,Our trained model can then be applied at test time to single views.,33897
illustrating football are player student assessment results. presented to Two and applications our also,Two applications to student and football player assessment are also presented illustrating our results.,33898
"many In extract texture color jointly. have features and to authors tried related researches,","In many related researches, authors have tried to extract color and texture features jointly.",33899
"efficient combined are respect, image In analysis considered methods this as descriptors.","In this respect, combined methods are considered as efficient image analysis descriptors.",33900
"image this methods. efficient most review we analysis paper, state-of-the-art and texture In","In this paper, we review most efficient and state-of-the-art image texture analysis methods.",33901
"some survived. classification Also, are approaches texture","Also, some texture classification approaches are survived.",33902
computer a in prediction depth problem is Single-view vision. fundamental,Single-view depth prediction is a fundamental problem in computer vision.,33903
"its objects. unreconstructable derived and MVS challenges, own noise with comes from Data including","Data derived from MVS comes with its own challenges, including noise and unreconstructable objects.",33904
the relations. associated (FRE)are Fuzzy fuzzy with binary composition equations of relation,Fuzzy relation equations (FRE)are associated with the composition of binary fuzzy relations.,33905
"embedding learning an Deep neural aims to metric network. as modeled function, deep learn","Deep metric learning aims to learn an embedding function, modeled as deep neural network.",33906
"to been results. yield deep has metric Recently, state-of-the-art applied ensemble learning to","Recently, ensemble has been applied to deep metric learning to yield state-of-the-art results.",33907
"in diverse the As aspect feature of one should important ensemble, their embeddings. be learners","As one important aspect of ensemble, the learners should be diverse in their feature embeddings.",33908
"divergence learners. among the encourages diversity propose which We also a loss,","We also propose a divergence loss, which encourages diversity among the learners.",33909
brain studies. extraction Brain most imaging step is fundamental for a,Brain extraction is a fundamental step for most brain imaging studies.,33910
"prior Many inverse regularize solution require ill-posed, are the problems space. a to hence and","Many inverse problems are ill-posed, and hence require a prior to regularize the solution space.",33911
"this In (AQuaSI) Sparse the Adaptive propose work, prior. we Image Quantile","In this work, we propose the Adaptive Quantile Sparse Image (AQuaSI) prior.",33912
"watershed-based Using around these individual boundaries with algorithm. boundaries, closed-contour detected a estimated are leaves","Using these detected boundaries, closed-contour boundaries around individual leaves are estimated with a watershed-based algorithm.",33913
The for dense leaves. instance segmenter an is result,The result is an instance segmenter for dense leaves.,33914
for results dense Promising are segmentation foliage in leaves obtained.,Promising segmentation results for leaves in dense foliage are obtained.,33915
"However, significantly cross-wavelength matching presents worse performance.","However, cross-wavelength matching presents significantly worse performance.",33916
important play many in biological roles MicroRNAs processes.,MicroRNAs play important roles in many biological processes.,33917
signalling different pathways. to clustered pro-tumorigenic are miRNAs target components of frequently co-expressed Genomically,Genomically clustered miRNAs are frequently co-expressed to target different components of pro-tumorigenic signalling pathways.,33918
many concept Manifold learning central theory has methods. the been of,Manifold theory has been the central concept of many learning methods.,33919
"learning In this paper we encourage to novel method present manifold-aware of representations. ManifoldNet, a","In this paper we present ManifoldNet, a novel method to encourage learning of manifold-aware representations.",33920
manifold a Our into input the of segments set fragments. approach,Our approach segments the input manifold into a set of fragments.,33921
"to Due the segmentation unsupervised tends its be nature, noisy. to","Due to its unsupervised nature, the segmentation tends to be noisy.",33922
this We by segmentation ensemble manifold (EMS). mitigate introducing,We mitigate this by introducing ensemble manifold segmentation (EMS).,33923
with task-specific can with or be only ManifoldNet pseudo the labels trained labels. together,ManifoldNet can be trained with only the pseudo labels or together with task-specific labels.,33924
evaluate semi-supervised learning. and tasks: on different network two We imitation (distillation) ManifoldNet,We evaluate ManifoldNet on two different tasks: network imitation (distillation) and semi-supervised learning.,33925
systems The inside-out tracking these issues. of overcome aims to introduction,The introduction of inside-out tracking systems aims to overcome these issues.,33926
"like (CNN) AlexNet, etc. neural networks VGGNet, convolutional GoogleNet, The","The convolutional neural networks (CNN) like AlexNet, GoogleNet, VGGNet, etc.",33927
descriptor very been discriminative as problems. the for proven have many feature vision computer,have been proven as the very discriminative feature descriptor for many computer vision problems.,33928
introduce to layer in non-linearity. The discards information the some Unit order Rectified Linear (ReLU),The Rectified Linear Unit (ReLU) layer discards some information in order to introduce the non-linearity.,33929
years. have registration results recent methods set Probabilistic for in point demonstrated competitive,Probabilistic methods for point set registration have demonstrated competitive results in recent years.,33930
point estimate model a distribution probability clouds. of These techniques the,These techniques estimate a probability distribution model of the point clouds.,33931
primarily the is sensor fundamental sets. point by across in caused location problem changes This,This fundamental problem is primarily caused by changes in the sensor location across point sets.,33932
probabilistic foundations revisit of the paradigm. the registration We,We revisit the foundations of the probabilistic registration paradigm.,33933
encountered density handles applications. severe Lidar in variations density-adaptive commonly registration Our successfully terrestrial,Our density-adaptive registration successfully handles severe density variations commonly encountered in terrestrial Lidar applications.,33934
datasets. extensive on real-world Lidar perform several challenging experiments We,We perform extensive experiments on several challenging real-world Lidar datasets.,33935
"mmWave fundamental presented. the this first article, characteristics In of are","In this article, the fundamental characteristics of mmWave are first presented.",33936
"Then, two discussed. channel modeling main methods are","Then, two main channel modeling methods are discussed.",33937
in important one a The is components system. the of most antenna WCE,The antenna is one of the most important components in a WCE system.,33938
"are ready shared. of and videos billions be viewed Nowadays, online to","Nowadays, billions of videos are online ready to be viewed and shared.",33939
"segments different each Furthermore, within video, attract of may significantly numbers different views.","Furthermore, within each video, different segments may attract significantly different numbers of views.",33940
"video problem, fine-grained a yet important attractiveness prediction. This to challenging namely phenomenon leads","This phenomenon leads to a challenging yet important problem, namely fine-grained video attractiveness prediction.",33941
"present relationships. different engagements that different attractiveness demonstrate and video First, we","First, we demonstrate that video attractiveness and different engagements present different relationships.",33942
"attractiveness FVAD us provides study problem. Second, prediction an to opportunity video fine-grained the","Second, FVAD provides us an opportunity to study the fine-grained video attractiveness prediction problem.",33943
on ideas so We detection. in recent by several unsupervised building landmark do,We do so by building on several recent ideas in unsupervised landmark detection.,33944
show the semantic for object We representation excellent matching. is pre-trained also that,We also show that the pre-trained representation is excellent for semantic object matching.,33945
antenna linearization predistortion large-scale investigates transmitters. (DPD) hybrid beamforming of This article digital,This article investigates digital predistortion (DPD) linearization of hybrid beamforming large-scale antenna transmitters.,33946
effect is new in critical dispersal. population predicted A,A new critical effect is predicted in population dispersal.,33947
"be of has cost, measured the should in the changes which coefficient. a Mobility reproduction","Mobility has a cost, which should be measured in the changes of the reproduction coefficient.",33948
Lambert solved a It be form can $W$-function. using closed in,It can be solved in a closed form using Lambert $W$-function.,33949
"worse solution is there no for For conditions, mobility.","For worse conditions, there is no solution for mobility.",33950
ill-posed. required is registration alignment problems It the image is by known that,It is known that the required alignment by image registration problems is ill-posed.,33951
this to CNN-architecture solve We a problem. propose specific,We propose a specific CNN-architecture to solve this problem.,33952
"experiments, that various we approach results. the proposed excellent yields In show","In various experiments, we show that the proposed approach yields excellent results.",33953
Our for and aforementioned the frameworks multiple extends studies work tasks.,Our work studies and extends multiple frameworks for the aforementioned tasks.,33954
the database. evaluated Faces methods Cartoon the (IIIT-CFW) Wild in on are the All,All the methods are evaluated on the Cartoon Faces in the Wild (IIIT-CFW) database.,33955
in Semantic in and challenging vision-based are images computer tasks aerial segmentation geolocalization vision.,Semantic segmentation and vision-based geolocalization in aerial images are challenging tasks in computer vision.,33956
"is tractable GPU. to Furthermore, limited be embedded its on an size","Furthermore, its size is limited to be tractable on an embedded GPU.",33957
"segmentation loss. the is loss by learning cosine trained Furthermore, and with FPAN box multi-task","Furthermore, the FPAN is trained by multi-task learning with box segmentation loss and cosine loss.",33958
demonstrate superior benefit design. our of sketch-specific Such the performances effectively,Such superior performances effectively demonstrate the benefit of our sketch-specific design.,33959
paper patch-based This matcher proposes recognition. a for hierarchical face multi-label,This paper proposes a hierarchical multi-label matcher for patch-based face recognition.,33960
"patches. divided image signature multi-level generation, is face into iteratively a In","In signature generation, a face image is iteratively divided into multi-level patches.",33961
The training of consists steps. matcher three,The matcher training consists of three steps.,33962
"learn of to are the classifiers local First, matching built patch. each local","First, local classifiers are built to learn the local matching of each patch.",33963
"the Last, of final matchings matching. are as different levels the global combined","Last, the global matchings of different levels are combined as the final matching.",33964
"However, `out-of-the-box' such to information. most CNN are prior still models blind","However, most `out-of-the-box' CNN models are still blind to such prior information.",33965
masks partial approaches full (e.g. from weakly-supervised Common generate input,Common weakly-supervised approaches generate full masks from partial input (e.g.,33966
scribbles interactive segmentation methods preprocessing. standard or as seeds) using,scribbles or seeds) using standard interactive segmentation methods as preprocessing.,33967
"(e.g. such in errors functions poorer in loss masks result training But, since standard","But, errors in such masks result in poorer training since standard loss functions (e.g.",33968
potentially seeds pixels. cross-entropy) do from other not distinguish mislabeled,cross-entropy) do not distinguish seeds from potentially mislabeled other pixels.,33969
analysis examples. has of revealed networks their Recent to vulnerability adversarial deep structured neural carefully,Recent analysis of deep neural networks has revealed their vulnerability to carefully structured adversarial examples.,33970
"main a our two of point composed pathways. network From is view, of structural","From a structural point of view, our network is composed of two main pathways.",33971
reliable ones. It the ones inconsistent promotes removes and the,It promotes the reliable ones and removes the inconsistent ones.,33972
"has tracking. full Thus, the in system our main approaches two benefit of","Thus, our system has the full benefit of two main approaches in tracking.",33973
state-of-the-art by results outperform significant margin. a methods These,These results outperform state-of-the-art methods by a significant margin.,33974
"can removing improve prediction, assumption While challenging. fitting this such is models","While removing this assumption can improve prediction, fitting such models is challenging.",33975
"paper, a effective aggregation but (SBA) we method. propose semantic-based this simple In","In this paper, we propose a simple but effective semantic-based aggregation (SBA) method.",33976
discriminative of deep semantic as The filters convolutional SBA utilizes proposed the detectors. layers,The proposed SBA utilizes the discriminative filters of deep convolutional layers as semantic detectors.,33977
and Our various excellent tasks. unsupervised SBA performance easy is to generalize on achieves,Our unsupervised SBA is easy to generalize and achieves excellent performance on various tasks.,33978
"propose to quality regression In work, network address the a this issue. hallucination-guided we","In this work, we propose a hallucination-guided quality regression network to address the issue.",33979
large all Our previous method outperforms state-of-the-art significantly methods the by margins.,Our method significantly outperforms all the previous state-of-the-art methods by large margins.,33980
We approximation approach that propose an SP-HO-CRFs new these resolves to problems.,We propose an new approximation approach to SP-HO-CRFs that resolves these problems.,33981
"introduces ratio on maximum combining the unfairness. user hand, The other","The maximum ratio combining on the other hand, introduces user unfairness.",33982
shocks. results global towards pronounced key macroeconomic to reactions climate of point quantities Our,Our results point towards pronounced global reactions of key macroeconomic quantities to climate shocks.,33983
and is essential for vertebrae automated of identification Robust spine localisation analysis.,Robust localisation and identification of vertebrae is essential for automated spine analysis.,33984
these severely of that This incapture study distortions concludes trackers. performance affect the,This study concludes that incapture distortions severely affect the performance of these trackers.,33985
unseen image on is pairs. task the Evaluation on of performed registration separate test,Evaluation is performed on the task of registration on separate unseen test image pairs.,33986
"(QTA). this propose Question Attention we work, In Type-guided","In this work, we propose Question Type-guided Attention (QTA).",33987
regularization deep We custom networks. training functions use for of construct supervised in neural,We construct custom regularization functions for use in supervised training of deep neural networks.,33988
The first labels an autoencoder. phase with models,The first phase models labels with an autoencoder.,33989
"After training, auxiliary branch. we discard this","After training, we discard this auxiliary branch.",33990
also convolutional are Gains architecture. network choices of different consistent over,Gains are also consistent over different choices of convolutional network architecture.,33991
"in practice. must solve schemes issue, be this To relaying implemented","To solve this issue, relaying schemes must be implemented in practice.",33992
"For two these node scenarios, of optimal relay the is the obtained. location","For these two scenarios, the optimal location of the relay node is obtained.",33993
can complicated a range effects of subtle connectivity. dependent These create and competing time,These competing effects can create a complicated and subtle range of time dependent connectivity.,33994
a pharmaceutical in neuromodulator is with compared (muscarine) employed. one A is which control situation,A control situation is compared with one in which a pharmaceutical neuromodulator (muscarine) is employed.,33995
actions spatiotemporal This the of addresses problem videos. in localization of paper,This paper addresses the problem of spatiotemporal localization of actions in videos.,33996
increasing demand. in diagnosis is Histological Digitized,Digitized Histological diagnosis is in increasing demand.,33997
"diagnosis process. imposing are color to factors various to due variations obstacles However, the","However, color variations due to various factors are imposing obstacles to the diagnosis process.",33998
is proposed a of solutions. The with variations problem many color problem stain well-defined,The problem of stain color variations is a well-defined problem with many proposed solutions.,33999
Most template reference on solutions are highly these dependent of a slide.,Most of these solutions are highly dependent on a reference template slide.,34000
publicly be available. code will made The,The code will be made publicly available.,34001
present of for age We buildings automated first from photographs. unconstrained method the a estimation,We present a first method for the automated age estimation of buildings from unconstrained photographs.,34002
We in that be substantially improved can this performance show the way.,We show that the performance can be substantially improved in this way.,34003
is based final face correlation the made matrix. The attribute prediction on,The final face attribute prediction is made based on the correlation matrix.,34004
"outside and occurring Eurasia, Africa). (in region North the America,","occurring outside the region (in Eurasia, North America, and Africa).",34005
first sp. The subgenus all comprising SA Colias,The first subgenus comprising all SA Colias sp.,34006
"str. name which the s. Colias (Colias) for Worlds), from the and New Old","from the Old and New Worlds), for which the name Colias (Colias) s. str.",34007
monotonic design continuous convergence $D$-optimal are provided. criterion for alternative on An the spaces proof,An alternative proof the monotonic convergence for $D$-optimal criterion on continuous design spaces are provided.,34008
proposed converges the to the We $D$-optimal that algorithm further design. show,We further show that the proposed algorithm converges to the $D$-optimal design.,34009
numerical to the algorithms. demonstrated used usefulness of and performance the examples proposed Different are,Different numerical examples are used to demonstrated the usefulness and performance of the proposed algorithms.,34010
proposed shoeprints in significantly to crime performance matching impressions. test improves laboratory metric scene Our,Our proposed metric significantly improves performance in matching crime scene shoeprints to laboratory test impressions.,34011
depth high-quality systems sensors into SLAM depth or dense maps. lower-power,lower-power depth sensors or SLAM systems into high-quality dense depth maps.,34012
"encoder motion trained, images. for blurred employed Once training a is guided it for of","Once trained, it is employed for guided training of a motion encoder for blurred images.",34013
AdaLEA gradually proposed a allows progresses. training to model earlier The an anticipation learn as,The proposed AdaLEA allows a model to gradually learn an earlier anticipation as training progresses.,34014
"a construct for we DataBase Additionally, Incident anticipation. Near-miss","Additionally, we construct a Near-miss Incident DataBase for anticipation.",34015
"data. Therefore, storing the are and transmitting space required EEG bandwidth large for more and","Therefore, large space and more bandwidth are required for storing and transmitting the EEG data.",34016
to for The EEG system of data develop this is an compression. objective paper efficient,The objective of this paper is to develop an efficient system for EEG data compression.,34017
"In firstly are the preprocessing system, recorded unit. data EEG preprocessed in the this","In this system, the recorded EEG data are firstly preprocessed in the preprocessing unit.",34018
data segmentation done EEG Standardization and are this unit. of in,Standardization and segmentation of EEG data are done in this unit.,34019
"compression the the data EEG are resulting unite. to passed Then,","Then, the resulting EEG data are passed to the compression unite.",34020
followed by lossy algorithm. composes lossless a unit compression a compression The of algorithm compression,The compression unit composes of a lossy compression algorithm followed by a lossless compression algorithm.,34021
into high The randomness EEG transforms the lossy compression algorithm data data redundancy. with,The lossy compression algorithm transforms the randomness EEG data into data with high redundancy.,34022
"Run and (RLE) Arithmetic compression Encoding algorithm. Length lossless a Furthermore, Encoding are as proposed","Furthermore, Arithmetic Encoding and Run Length Encoding (RLE) are proposed as a lossless compression algorithm.",34023
"photometric for stereo media. we paper, In this a propose method participating","In this paper, we propose a photometric stereo method for participating media.",34024
forward previous method proposed studies differs with The to respect scatter. from shape-dependent modeling,The proposed method differs from previous studies with respect to modeling shape-dependent forward scatter.,34025
"Human fluctuating is visual perception complex, a dynamic process. and","Human visual perception is a complex, dynamic and fluctuating process.",34026
to have signals temporal We the with various incoming timescales. manipulate structures characteristic,We manipulate the incoming signals to have temporal structures with various characteristic timescales.,34027
"We and timescales. distinct operate that perception, find biases opposing over shape two","We find that two opposing biases shape perception, and operate over distinct timescales.",34028
"recency e.g. short Positive times, appears over","Positive recency appears over short times, e.g.",34029
"over of an probability trends increased times. longer changed Adaptation, reflects response, entailing","Adaptation, entailing an increased probability of changed response, reflects trends over longer times.",34030
in reproduces the experimental simple all regimes. stimulus A data model mathematical,A simple mathematical model reproduces the experimental data in all stimulus regimes.,34031
of significantly evaluation has hindered field. FSS This the the development unreliable,This unreliable evaluation has significantly hindered the development of the FSS field.,34032
critical (LV) disease cardiac volumes a diagnosis. Left is for procedure ventricular estimation,Left ventricular (LV) volumes estimation is a critical procedure for cardiac disease diagnosis.,34033
paper objective task. The is prediction volumes direct address to LV of this,The objective of this paper is to address direct LV volumes prediction task.,34034
The aspects. are of this following main the paper contributions,The main contributions of this paper are the following aspects.,34035
"cardiac preprocessing on method (CMR). propose we resonance magnetic First, data a new","First, we propose a new data preprocessing method on cardiac magnetic resonance (CMR).",34036
"estimation. Second, a for we new volumes end-to-end LV networks propose structure","Second, we propose a new networks structure for end-to-end LV volumes estimation.",34037
"playing multiple during fetal is placenta roles a complex organ, The development.","The placenta is a complex organ, playing multiple roles during fetal development.",34038
Very the placental physiology. association and little between is known abnormalities fetal morphological about,Very little is known about the association between placental morphological abnormalities and fetal physiology.,34039
"challenging for paper, introduce dataset, MLB-YouTube, a fine-grained designed detection. this activity we In new","In this paper, we introduce a challenging new dataset, MLB-YouTube, designed for fine-grained activity detection.",34040
We valuable that learning recognition. temporal is activity structure fine-grained find for,We find that learning temporal structure is valuable for fine-grained activity recognition.,34041
Quantitative experiments dataset method. effectiveness demonstrate on the conducted and of proposed qualitative the,Quantitative and qualitative experiments conducted on the dataset demonstrate the effectiveness of proposed method.,34042
"and explanations both In non-classical effects order we them. specific, classical and for consider discuss","In specific, we consider order effects and discuss both classical and non-classical explanations for them.",34043
iterations efficient. We closed-form enjoys the each computationally ADMM show makes solution which very sub-problem,We show each sub-problem enjoys closed-form solution which makes the ADMM iterations computationally very efficient.,34044
"bootstrap construct proposed bands. to procedure confidence is a multiplier uniform addition, In","In addition, a multiplier bootstrap procedure is proposed to construct uniform confidence bands.",34045
"the methodology of housing. demand to for applied heterogeneity analyze elasticity Finally, is income in","Finally, the methodology is applied to analyze heterogeneity in income elasticity of demand for housing.",34046
"its adaptation into metrics incorporates problem, solve To this algorithm. direct NetAdapt","To solve this problem, NetAdapt incorporates direct metrics into its adaptation algorithm.",34047
semi-supervised image Current are problem mainstream biomedical segmentation address the to works graph-based. mostly,Current mainstream works to address the semi-supervised biomedical image segmentation problem are mostly graph-based.,34048
Gaussian previous noise image (AWGN). focus Most on of methods denoising additive white,Most of previous image denoising methods focus on additive white Gaussian noise (AWGN).,34049
"computer vision real-world advancing with of the image the problem noisy techiniques. However,the denoising","However,the real-world noisy image denoising problem with the advancing of the computer vision techiniques.",34050
by These are under different settings. camera images different cameras captured,These images are captured by different cameras under different camera settings.,34051
as datasets. new We as dataset on denoising different methods well previous our the evaluate,We evaluate the different denoising methods on our new dataset as well as previous datasets.,34052
"dataset on deep a learning essential for is Usually, AAA applying segmentation. training large","Usually, a large training dataset is essential for applying deep learning on AAA segmentation.",34053
"Then, AAA. U-net to trained was the segment","Then, U-net was trained to segment the AAA.",34054
recognition have approaches attributes. importance the ignored Existing of mostly facial sketch-photo,Existing sketch-photo recognition approaches have mostly ignored the importance of facial attributes.,34055
performed and on databases. (E-PRIP) experiments are composite Semi-forensic) semi-forensic Extensive (IIIT-D,Extensive experiments are performed on composite (E-PRIP) and semi-forensic (IIIT-D Semi-forensic) databases.,34056
the significantly state-of-the-art. outperforms proposed method The,The proposed method significantly outperforms the state-of-the-art.,34057
of highly spaces of conceptual a influential geometric way framework knowledge. representing The provides,The highly influential framework of conceptual spaces provides a geometric way of representing knowledge.,34058
upon acceptance. code Source paper be will released,Source code will be released upon paper acceptance.,34059
tend distances dimensional vector constant. high to Many a,Many high dimensional vector distances tend to a constant.,34060
distributions apparently chaotically means actually separable. intrinsically hollow The center are overlapping,The hollow center means apparently chaotically overlapping distributions are actually intrinsically separable.,34061
"multiple landmarks, for time-consuming. tracking, extremely becomes especially Manual body","Manual tracking, especially for multiple body landmarks, becomes extremely time-consuming.",34062
"outlined a the manually we Finally, segmented regions region. to compared","Finally, we compared the segmented regions to a manually outlined region.",34063
"been In years, there in jointly action. perception has modeling a renewed interest recent and","In recent years, there has been a renewed interest in jointly modeling perception and action.",34064
"used words, it the object actions In for). can other what be represents","In other words, it represents what actions can the object be used for).",34065
a in to scene. also two-step a new propose affordances We predict approach,We also propose a two-step approach to predict affordances in a new scene.,34066
possible us distribution at of to the test The sample VAE allows poses time.,The VAE allows us to sample the distribution of possible poses at test time.,34067
rely Previous text manually approaches detection windows. defined sliding on scene usually for,Previous approaches for scene text detection usually rely on manually defined sliding windows.,34068
many active applications. video research Joint and of understanding an language with area is,Joint understanding of video and language is an active research area with many applications.,34069
relies in on embeddings. text-video typically work Prior domain this learning,Prior work in this domain typically relies on learning text-video embeddings.,34070
"learning this at embeddings To text-video aim data from heterogeneous sources. issue, we address","To address this issue, we aim at learning text-video embeddings from heterogeneous data sources.",34071
as input of modalities generalization We descriptors. such the to other show MEE face also,We also show the generalization of MEE to other input modalities such as face descriptors.,34072
natural under enables us selection simulate conditions. algorithm to The mutations,The algorithm enables us to simulate mutations under natural selection conditions.,34073
the obtained content image segmentation proposed screen algorithm. using for results Promising are,Promising results are obtained for screen content image segmentation using the proposed algorithm.,34074
range at are computational of effective a Deep neural networks tasks. highly,Deep neural networks are highly effective at a range of computational tasks.,34075
"networks results in accuracy most for However, applied to when drops naive binarization tasks.","However, naive binarization results in accuracy drops when applied to networks for most tasks.",34076
implementations across also develop approach for We different proposed architectures. efficient our,We also develop efficient implementations for our proposed approach across different architectures.,34077
it find defined each pathways. function flux-balanced The can such is of element that,The function of each element is defined such that it can find flux-balanced pathways.,34078
of regions sets. a set their closed and boundaries) causal,a closed set of regions and their boundaries) causal sets.,34079
of better. changes design made little We make bunch to a it,We made a bunch of little design changes to make it better.,34080
swell. network also We trained new this pretty that's,We also trained this new network that's pretty swell.,34081
but a time It's little accurate. bigger last than more,It's a little bigger than last time but more accurate.,34082
achieved with on par better person re-identification performance is The other techniques. or RNN-based,The performance achieved is better or on par with other RNN-based person re-identification techniques.,34083
for multimedia and Rain vision in many effect annoying is images tasks. typically computer,Rain effect in images typically is annoying for many multimedia and computer vision tasks.,34084
"Furthermore, removal. tasks to also design dust other applicable our decomposition layer is like","Furthermore, our design is also applicable to other layer decomposition tasks like dust removal.",34085
convolutional Experimental show that results key in learning results supervised advantages. using two,Experimental results show that using supervised convolutional learning results in two key advantages.,34086
a novel multi-human model deep propose parsing. for Network further We (NAN) Adversarial Nested,We further propose a novel deep Nested Adversarial Network (NAN) model for multi-human parsing.,34087
"for POMDP work cases, some failing well for solvers catastrophically others. while MDP based","MDP based POMDP solvers work well for some cases, while catastrophically failing for others.",34088
solving problems lead gaining However efficient information solutions. can for to POMDP,However for solving POMDP problems gaining information can lead to efficient solutions.,34089
tiger then to such demonstrate We sub-optimality. use the well-known problem,We then use the well-known tiger problem to demonstrate such sub-optimality.,34090
"based We MDP budgeted multi-resolution, POMDP cannot be gathering addressed show that solvers. using information","We show that multi-resolution, budgeted information gathering cannot be addressed using MDP based POMDP solvers.",34091
"handle simultaneously objectives these existing methods two well. cannot However,","However, existing methods cannot simultaneously handle these two objectives well.",34092
"a convolutional mapped first is a vector through latent to face encoder. the Specifically,","Specifically, the face is first mapped to a latent vector through a convolutional encoder.",34093
latent aging features face The and rejuvenation. age preserves facial personalized and vector controls the,The latent vector preserves personalized face features and the age controls facial aging and rejuvenation.,34094
"Besides, representation to identity of utilized face. feature preserve generated is the a personalized high-level","Besides, a high-level feature representation is utilized to preserve personalized identity of the generated face.",34095
method datasets appealing demonstrate on two the proposed Experiments the over performance state-of-the-art. benchmark of,Experiments on two benchmark datasets demonstrate appealing performance of the proposed method over the state-of-the-art.,34096
and also in academia is Predictive maturing companies. in gained traction process monitoring recently has,Predictive process monitoring has recently gained traction in academia and is maturing also in companies.,34097
architecture We multi-view learning deep present map from for an end-to-end depth images. inference,We present an end-to-end deep learning architecture for depth map inference from multi-view images.,34098
MVSNet proposed the indoor DTU large-scale is demonstrated on The dataset.,The proposed MVSNet is demonstrated on the large-scale indoor DTU dataset.,34099
motion the artifacts. to risk This scanning leads increasing clinically of prolonged times,This leads to clinically prolonged scanning times increasing the risk of motion artifacts.,34100
"is advanced processing and an of techniques need there ASL. in immense Thus, imaging","Thus, there is an immense need of advanced imaging and processing techniques in ASL.",34101
"can dramatically. objects Empirically, show the performance improved for we pose-varying be that","Empirically, we show that the performance for pose-varying objects can be improved dramatically.",34102
"For dataset manually called task, Palette-and-Text we this curated introduce (PAT). our","For this task, we introduce our manually curated dataset called Palette-and-Text (PAT).",34103
semantics captures produce color and the the input The of former text palettes. relevant,The former captures the semantics of the text input and produce relevant color palettes.,34104
the colorizes latter image color grayscale The palette. generated a using,The latter colorizes a grayscale image using the generated color palette.,34105
of dimension distinct RankCGAN associates latent of space. some each interest subjective to a measure,RankCGAN associates each subjective measure of interest to a distinct dimension of some latent space.,34106
improving health Personalization the powerful in of is very interventions. effectiveness,Personalization is very powerful in improving the effectiveness of health interventions.,34107
"passes. e-Health window optimal in crucial the can be Also, before intervention timing","Also, in e-Health intervention timing can be crucial before the optimal window passes.",34108
e-Health our intervention benchmark. compare with an We methods,We compare our methods with an e-Health intervention benchmark.,34109
setting. online learning learning for that We outperforms our batch demonstrate,We demonstrate that batch learning outperforms online learning for our setting.,34110
by codelets executed a running tiny Machine. partitioning is Turing space The simulated on,The space partitioning is executed by tiny codelets running on a simulated Turing Machine.,34111
is The quantized ambiguity coherent also for transmission. phase (PA) cooperative,The phase ambiguity (PA) is also quantized for coherent cooperative transmission.,34112
"for approach, is the evaluation useless actuators. This of however,","This approach, however, is useless for the evaluation of actuators.",34113
architecture the publicly for is The code proposed available.,The code for the proposed architecture is publicly available.,34114
for a restoration investigate by novel learning. approach reinforcement We image,We investigate a novel approach for image restoration by reinforcement learning.,34115
data filtered the on data. labeled of depend Additionally quality variation strongly of the input,Additionally the quality of labeled data depend strongly on the variation of filtered input data.,34116
model to are functions choice lenses. the Polynomial nonlinearity a usual of,Polynomial functions are a usual choice to model the nonlinearity of lenses.,34117
identification This investigates protection. real-world paper issue better of fulfill the species to,This paper investigates the issue of real-world identification to fulfill better species protection.,34118
identification and hot as species plant focus is it on issue. a classic We,We focus on plant species identification as it is a classic and hot issue.,34119
and tradition simple. In the scanned are identification is samples the background plant species specimen,In tradition plant species identification the samples are scanned specimen and the background is simple.,34120
"challenging. recognition more real-world species is However,","However, real-world species recognition is more challenging.",34121
in We of attention image first recognition. visual terms general crop the before,We first crop the image in terms of visual attention before general recognition.,34122
"a data augmentation as it method. we Besides, apply","Besides, we apply it as a data augmentation method.",34123
(AC). the cropping call augmentation data novel approach We attention,We call the novel data augmentation approach attention cropping (AC).,34124
from trained of neural amount to convolutional are a Deep species large data. predict networks,Deep convolutional neural networks are trained to predict species from a large amount of data.,34125
achieves different state-of-the-art of approach that types Experiments our results on datasets. demonstrate first,Experiments first demonstrate that our approach achieves state-of-the-art results on different types of datasets.,34126
"AC. of Besides, performance method the evaluate data we also augmentation","Besides, we also evaluate the performance of data augmentation method AC.",34127
performance. that provides Results superior AC show,Results show that AC provides superior performance.,34128
"precision with AC, of AC with results without improvement. the achieve methods substantial the Compared","Compared with the precision of methods without AC, the results with AC achieve substantial improvement.",34129
and quantitative idea. our to Extensive validate qualitative experiments are performed,Extensive qualitative and quantitative experiments are performed to validate our idea.,34130
model data IJB-A. We on sets: the challenging LFW two evaluated and,We evaluated the model on two challenging data sets: LFW and IJB-A.,34131
"propose model first reconstruction. In paper, deep learning the this for multi-contrast we MRI","In this paper, we propose the first deep learning model for multi-contrast MRI reconstruction.",34132
"feature number sharing We significantly which of reduces achieve through information units, sharing the parameters.","We achieve information sharing through feature sharing units, which significantly reduces the number of parameters.",34133
degenerated the used quality can to The optimizer improve be diverse of repeatedly observations. learned,The learned optimizer can be repeatedly used to improve the quality of diverse degenerated observations.,34134
generalization. The proposed possesses and interpretability method strong high,The proposed method possesses strong interpretability and high generalization.,34135
under an to the this problem Solving consideration. us proximity specific graphs measure gives optimized,Solving this problem gives us an optimized proximity measure specific to the graphs under consideration.,34136
The that indicate variable results constant models. resolution significantly models resolution outperform,The results indicate that variable resolution models significantly outperform constant resolution models.,34137
based both outperforms LSTM state-of-the-art model non-LSTM algorithms. The and proposed,The proposed model outperforms both LSTM and non-LSTM based state-of-the-art algorithms.,34138
"However, we solution a simple effective such deal ambiguities. yet to with provide","However, we provide a simple yet effective solution to deal with such ambiguities.",34139
segmentation. in brought image Convolutional have advances neural networks significant (CNNs),Convolutional neural networks (CNNs) have brought significant advances in image segmentation.,34140
paper approach for a This deep learning and unsupervised to proposes clustering representation segmentation. unified,This paper proposes a unified approach to unsupervised deep representation learning and clustering for segmentation.,34141
method Our consists phases. of two proposed,Our proposed method consists of two phases.,34142
abilities Our medical the experiments deep of segmentation. potential unsupervised image show for representation learning,Our experiments show the potential abilities of unsupervised deep representation learning for medical image segmentation.,34143
two The classification deep networks by aspects. trackers of limited using performance existing is,The performance of existing trackers using deep classification networks is limited by two aspects.,34144
"class negative extreme there samples. between and exists Second, positive imbalance","Second, there exists extreme class imbalance between positive and negative samples.",34145
address adversarial presents two these problems the to VITAL via This paper learning. algorithm,This paper presents the VITAL algorithm to address these two problems via adversarial learning.,34146
favorably performs tracker that benchmark the demonstrate experiments against on state-of-the-art approaches. Extensive datasets proposed,Extensive experiments on benchmark datasets demonstrate that the proposed tracker performs favorably against state-of-the-art approaches.,34147
"via Inspired image work by the recent stylization, address problem on neural we texture transfer.","Inspired by recent work on image stylization, we address the problem via neural texture transfer.",34148
lower for errors productivity for relative assessment configurations. Relative were multi-stock biomass parameters and model,Relative errors for productivity and relative biomass parameters were lower for multi-stock assessment model configurations.,34149
"Networks the features based adaptable specific goal, Convolutional Towards such learn and can on data.","Towards such goal, Convolutional Networks can learn specific and adaptable features based on the data.",34150
"limitation, is size the overcome image such processed patches. To using fixed","To overcome such limitation, the image is processed using fixed size patches.",34151
could and best size. suffer Both patch strategies not from drawbacks the lead to,Both strategies suffer from drawbacks and could not lead to the best patch size.,34152
"networks exploited multi-context this combining layers. or several alleviate problem, works To by information","To alleviate this problem, several works exploited multi-context information by combining networks or layers.",34153
more increases a train. model resulting parameters This process difficult number the of to in,This process increases the number of parameters resulting in a more difficult model to train.,34154
"is of misalignment principal and emphasize, between economic analysis. As we a core agent focus","As we emphasize, misalignment between principal and agent is a core focus of economic analysis.",34155
"VO, metrics, semi-dense real-time. feature-less Umeyama performance & Keywords method method, Method, feature-based","Keywords VO, performance metrics, Umeyama Method, feature-based method, feature-less method & semi-dense real-time.",34156
if early. curable a is skin aggressive cancer detected Melanoma,Melanoma is a curable aggressive skin cancer if detected early.,34157
"Typically, initial diagnosis if screening biopsy with necessary. examination involves and the subsequent histopathological","Typically, the diagnosis involves initial screening with subsequent biopsy and histopathological examination if necessary.",34158
"generally clinical requires and significant confirmation, good which requires effort. histological A annotation","A good annotation generally requires clinical and histological confirmation, which requires significant effort.",34159
"Moreover, images. method generating like is dermoscopy also the real-world proposed capable of","Moreover, the proposed method is also capable of generating real-world like dermoscopy images.",34160
motion-blurred image. present single a a to a method extract from We video sequence,We present a method to extract a video sequence from a single motion-blurred image.,34161
"temporal Firstly, the the frames. averaging of destroys ordering","Firstly, averaging destroys the temporal ordering of the frames.",34162
is to introduce the order. Our to loss temporal invariant main functions contribution,Our main contribution is to introduce loss functions invariant to the temporal order.,34163
measure and f equation Q. mutation function depends This on fitness on a,This equation depends on a fitness function f and on mutation measure Q.,34164
"More cycle al. the Poole Darwin-Eigen by proposed et exactly,","More exactly, the Darwin-Eigen cycle proposed by Poole et al.",34165
"length of the takes place if the M_{c}(f,Q). exceeds genomes","takes place if the length of the genomes exceeds M_{c}(f,Q).",34166
increasingly ecological important components indirect as Trait-mediated acknowledged are systems. in dynamics effects the of,Trait-mediated indirect effects are increasingly acknowledged as important components in the dynamics of ecological systems.,34167
in single paper objects images. of we In the localization problem this of unsupervised address,In this paper we address the problem of unsupervised localization of objects in single images.,34168
"we any do treat image on neighboring image each and individually rely not Furthermore, similarity.","Furthermore, we treat each image individually and do not rely on any neighboring image similarity.",34169
to deep-learning generation We employ region this maps of tackle and problem. proposals based saliency,We employ deep-learning based generation of saliency maps and region proposals to tackle this problem.,34170
image encoder/decoder determined First an in the regions salient are architecture. using,First salient regions in the image are determined using an encoder/decoder architecture.,34171
regions the further These refined overlap based similarity ratios. on are and,These regions are further refined based on the overlap and similarity ratios.,34172
allows data training algorithm. using the back-propagation learn model from This us parameters to,This allows us to learn the model parameters from training data using back-propagation algorithm.,34173
is capable images Our projection natural creating of X-ray appearance. with approach,Our approach is capable of creating X-ray projection images with natural appearance.,34174
"clear baseline extensions our to show Additionally, method. improvement compared the","Additionally, our extensions show clear improvement compared to the baseline method.",34175
"dataset. extensive subjective we image omnidirectional conducted experiments with evaluation, For","For evaluation, we conducted extensive subjective experiments with omnidirectional image dataset.",34176
for of method presents novel a This unsupervised pathology images. segmentation paper,This paper presents a novel method for unsupervised segmentation of pathology images.,34177
factor Staging lung is cancer of major of a prognosis.,Staging of lung cancer is a major factor of prognosis.,34178
of images cancer methods specimen. We on lung evaluated our pathology,We evaluated our methods on pathology images of lung cancer specimen.,34179
problem fundamental Optical vision. character a recognition computer is in (OCR),Optical character recognition (OCR) is a fundamental problem in computer vision.,34180
printed work topology characters. the This revealed to also recognize robustness machine RNN-LSTM learning of,This machine learning work also revealed the robustness of RNN-LSTM topology to recognize printed characters.,34181
"coherence intravascular order In used. tomography optical can image to (plaque), be atherosclerotic (IVOCT) lesions","In order to image atherosclerotic lesions (plaque), intravascular optical coherence tomography (IVOCT) can be used.",34182
"profiles single to examined detect plaque. IVOCT attenuation images example, of For in A-Scans are","For example, attenuation profiles in single A-Scans of IVOCT images are examined to detect plaque.",34183
"In induce sickness. is that motion the detect we exceptional likely VR paper, this to","In this paper, we detect the exceptional motion that is likely to induce VR sickness.",34184
Carlo usefulness Monte simulations an demonstrate empirical proposed procedure. of and the illustration the,Monte Carlo simulations and an empirical illustration demonstrate the usefulness of the proposed procedure.,34185
with The entire trainable supervision. end-to-end is only weak network,The entire network is end-to-end trainable with only weak supervision.,34186
of encodes the PSF rotation position the the of depth The of source. amount point,The amount of rotation of the PSF encodes the depth position of the point source.,34187
often training. paired requires of number large image-sentence annotations a for It,It often requires a large number of paired image-sentence annotations for training.,34188
applications. in It thus real-world limits and is labor usage expensive its,It is labor expensive and thus limits its usage in real-world applications.,34189
novel The unseen represents object. placeholder an,The placeholder represents an unseen novel object.,34190
"descriptions. novel from decoupled the Thus, be model sequence can the object","Thus, the sequence model can be decoupled from the novel object descriptions.",34191
to the will SM-P from object words the generate retrieve query The a memory.,The SM-P will generate a query to retrieve the words from the object memory.,34192
form measure changes. from areas that local Overlapping dense capacitors conductive layers area separate,Overlapping conductive areas from separate layers form local capacitors that measure dense area changes.,34193
"and directions are issues AGMEN challenging discussed, research the highlighted. potential Furthermore, are in","Furthermore, the challenging issues in AGMEN are discussed, and potential research directions are highlighted.",34194
"Neural recognition. efficient of Convolutional tools (CNNs) widespread, Networks Deep are visual","Deep Convolutional Neural Networks (CNNs) are widespread, efficient tools of visual recognition.",34195
"hosting online years, and In both video exponentially recent growing. service retail are","In recent years, both online retail and video hosting service are exponentially growing.",34196
"is network, deep this proposed neural AsymNet, to A novel problem. called explore","A novel deep neural network, called AsymNet, is proposed to explore this problem.",34197
"efficiency the when training. Moreover, method an to proposed achieve approximate is training","Moreover, an approximate training method is proposed to achieve the efficiency when training.",34198
essential that require Knowledge to systems structured are to (KGs) data. access becoming Graphs information,Knowledge Graphs (KGs) are becoming essential to information systems that require access to structured data.,34199
"identifying relevant, nodes We measure and propose specificity entity-specific, most of edges. accurate as an","We propose specificity as an accurate measure of identifying most relevant, entity-specific, nodes and edges.",34200
scalable bidirectional a method We walks compute develop based specificity. to random on,We develop a scalable method based on bidirectional random walks to compute specificity.,34201
(PPIs) as usually networks. are modelled Motivation: Protein-protein interactions,Motivation: Protein-protein interactions (PPIs) are usually modelled as networks.,34202
be tend similar revealed that functions similarly They wired. to involved in proteins,They revealed that proteins involved in similar functions tend to be similarly wired.,34203
"complexes. this, address of graphlets generalization To to simplicial we define a simplets,","To address this, we define simplets, a generalization of graphlets to simplicial complexes.",34204
either samples based on groups their Clustering as closeness or identifies homogeneity. analysis mutual,Clustering analysis identifies samples as groups based on either their mutual closeness or homogeneity.,34205
are sparse assumed The separated regions. clusters by be relatively to,The clusters are assumed to be separated by relatively sparse regions.,34206
their according sequentially to The are dynamic boundary samples eroded densities.,The samples are eroded sequentially according to their dynamic boundary densities.,34207
"between more this boundaries manner, different By clusters more and apparent. become","By this manner, boundaries between different clusters become more and more apparent.",34208
Mode problem in critical generative training is collapse a networks. adversarial,Mode collapse is a critical problem in training generative adversarial networks.,34209
"often sacrificing result image the the quality. However, is of their achievement","However, their achievement is often the result of sacrificing the image quality.",34210
"easily is we our that algorithm particular, extendable existing In to various demonstrate GANs.","In particular, we demonstrate that our algorithm is easily extendable to various existing GANs.",34211
is Discriminant improve Analysis used ranking to approach initial Context based lists. post-ranking The Information,The Discriminant Context Information Analysis based post-ranking approach is used to improve initial ranking lists.,34212
We of video. in problem retrieval activity address text-based the,We address the problem of text-based activity retrieval in video.,34213
also A task. auxiliary adding is multi-task an re-generation as by employed query loss,A multi-task loss is also employed by adding query re-generation as an auxiliary task.,34214
Our two significantly benchmarks: and Charades-STA challenging work approach ActivityNet prior outperforms on Captions.,Our approach significantly outperforms prior work on two challenging benchmarks: Charades-STA and ActivityNet Captions.,34215
error (BDPIM-OSD) OSD two can between the barriers. limit propagation with BDPIM adjacent,BDPIM with OSD (BDPIM-OSD) can limit the error propagation between two adjacent barriers.,34216
analysis. conducted Simulations our are corroborate to,Simulations are conducted to corroborate our analysis.,34217
settings Optimal uncoded by in systems also and simulations. investigated coded parameter are,Optimal parameter settings are also investigated in uncoded and coded systems by simulations.,34218
semantic concept abilities space. similar Computers by ascertain can a utilizing to start,Computers can start to ascertain similar abilities by utilizing a semantic concept space.,34219
in are the explored. strategies Two space semantic,Two strategies in the semantic space are explored.,34220
Fast and image landmark benefit many detection analysis can medical anatomical methods. accurate,Fast and accurate anatomical landmark detection can benefit many medical image analysis methods.,34221
"to medical landmarks method a propose automatically we detect in anatomical images. Here,","Here, we propose a method to automatically detect anatomical landmarks in medical images.",34222
"that contain to is landmark. used patches classification the identify Simultaneously,","Simultaneously, classification is used to identify patches that contain the landmark.",34223
simulations approach. of effectiveness the show Numerical our,Numerical simulations show the effectiveness of our approach.,34224
Decoding disease. latent the apparent without an produces representation resultant image,Decoding the resultant latent representation produces an image without apparent disease.,34225
"a the Context adopt Branch, local scale In adaptive mechanism for we competition context selection.","In the Context Branch, we adopt a local competition mechanism for adaptive context scale selection.",34226
two challenging evaluations Comprehensive (i.e. detection on datasets pedestrian,Comprehensive evaluations on two challenging pedestrian detection datasets (i.e.,34227
of PCN. the well Caltech demonstrated proposed effectiveness INRIA) the and,Caltech and INRIA) well demonstrated the effectiveness of the proposed PCN.,34228
Previous ZSL algorithms tested several on attributes. are datasets benchmark annotated with,Previous ZSL algorithms are tested on several benchmark datasets annotated with attributes.,34229
"defective terms and attribute these of distribution the datasets However, are image in diversity.","However, these datasets are defective in terms of the image distribution and attribute diversity.",34230
"To problems, propose Large-scale Attribute these a Dataset overcome we (LAD).","To overcome these problems, we propose a Large-scale Attribute Dataset (LAD).",34231
analyze learning zero-shot dataset our supervised tasks. learning We both and conducting by,We analyze our dataset by conducting both supervised learning and zero-shot learning tasks.,34232
new Seven this on dataset. state-of-the-art ZSL are tested algorithms,Seven state-of-the-art ZSL algorithms are tested on this new dataset.,34233
of The challenge zero-shot the implementing experimental dataset. reveal on results our learning,The experimental results reveal the challenge of implementing zero-shot learning on our dataset.,34234
a in problem untrimmed challenging has content been actions video analysis. long videos Locating in,Locating actions in long untrimmed videos has been a challenging problem in video content analysis.,34235
"location refined for Network further to Localization regression. sent the Finally, proposals a are fine-grained","Finally, the refined proposals are sent to a Localization Network for further fine-grained location regression.",34236
"remarkable is with high precise performance thresholds. localization gain Specifically, under IoU the","Specifically, the performance gain is remarkable under precise localization with high IoU thresholds.",34237
Motion yet boundary challenging problem. a is detection crucial,Motion boundary detection is a crucial yet challenging problem.,34238
The optical with generic MoBoNet works any and flows. is proposed,The proposed MoBoNet is generic and works with any optical flows.,34239
in impacts directly the detector processing the The of efficacy succeeding SAR-ATR the chain. stages,The efficacy of the detector directly impacts the succeeding stages in the SAR-ATR processing chain.,34240
detector. reported the are implementing the for There methods numerous in literature,There are numerous methods reported in the literature for implementing the detector.,34241
"the various First, a detection is methods for taxonomy proposed.","First, a taxonomy for the various detection methods is proposed.",34242
"overviewed. strategies implementation assumptions different for are the underlying Second,","Second, the underlying assumptions for different implementation strategies are overviewed.",34243
"a examples between selections tabular Third, comparison careful representative introduced. is of","Third, a tabular comparison between careful selections of representative examples is introduced.",34244
explicit We formulation propose as structural discover an autoencoding representations. landmarks to,We propose an autoencoding formulation to discover landmarks as explicit structural representations.,34245
Github . code on is available Our,Our code is available on Github .,34246
embeddings visual-semantic present principled adapting first ZSD. for by a We approach,We present a principled approach by first adapting visual-semantic embeddings for ZSD.,34247
"imperfections performance. In the radar with antenna degrade the antennas, multiple practical system the","In the practical radar with multiple antennas, the antenna imperfections degrade the system performance.",34248
"all the for theoretically are derived. Additionally, unknown parameters distributions the prior","Additionally, the prior distributions for all the unknown parameters are theoretically derived.",34249
performance. superior CS reconstruction been obtained have proposed and models Several,Several CS reconstruction models have been proposed and obtained superior performance.,34250
"sampling we the utilize to sub-network, operator. layer mimic convolutional the sampling a In","In the sampling sub-network, we utilize a convolutional layer to mimic the sampling operator.",34251
have modelled widely interactions Molecular as been networks.,Molecular interactions have widely been modelled as networks.,34252
with are in linked The patterns molecules biological local networks around wiring their functions. molecular,The local wiring patterns around molecules in molecular networks are linked with their biological functions.,34253
apply proteins. We hypergraphlets predict uncharacterised successfully functions biological of our to,We apply our hypergraphlets to successfully predict biological functions of uncharacterised proteins.,34254
statistics Knowledge second-order (e.g. channels of of,Knowledge of second-order statistics of channels (e.g.,34255
paper measurements. estimating instead considers problem This downlink the the from channel covariance of uplink,This paper considers instead the problem of estimating the downlink channel covariance from uplink measurements.,34256
"of their assumptions. This and the reveals techniques, limitations some questions of current underlying","This reveals the limitations of current techniques, and questions some of their underlying assumptions.",34257
convolutional large network The benchmarks. superior recent deep performance neural achieved (CNN) very proposed on,The proposed very deep convolutional neural network (CNN) achieved superior performance on recent large benchmarks.,34258
performance synthesis. great image generative in models have demonstrated Deep,Deep generative models have demonstrated great performance in image synthesis.,34259
Experiments the generation enables image transfer. model and conditional that show,Experiments show that the model enables conditional image generation and transfer.,34260
"stochastic representation, preserving sampled its latent be shape. while due to appearance Moreover, can","Moreover, appearance can be sampled due to its stochastic latent representation, while preserving shape.",34261
arteries. real A coronary similarities strong synthesized qualitative showed and between analysis,A qualitative analysis showed strong similarities between real and synthesized coronary arteries.,34262
geometries. Wasserstein to can generative be blood show synthesize networks Results that adversarial used vessel,Results show that Wasserstein generative adversarial networks can be used to synthesize blood vessel geometries.,34263
as pricing such well path options. and processes dependent filtering We consider of,We consider filtering such processes and well as pricing of path dependent options.,34264
computational recent dehazing the imaging years. topic Image become important an has in,Image dehazing has become an important computational imaging topic in the recent years.,34265
and The generation. degradation cost intermittent energy renewable incorporates storage mod-el,The mod-el incorporates energy storage degradation cost and intermittent renewable generation.,34266
"security, and people logistics. has systems, counting intelligent Vision-based widespread of automatic in applications transportation","Vision-based automatic counting of people has widespread applications in intelligent transportation systems, security, and logistics.",34267
"this currently there no large-scale problem. dataset for approaches benchmarking public is However, on","However, there is currently no large-scale public dataset for benchmarking approaches on this problem.",34268
human The for heads. image resulting analyzed potential depth is identifying,The resulting depth image is analyzed for identifying potential human heads.,34269
refined again counting. trajectories ascertain are to reliable The,The trajectories are again refined to ascertain reliable counting.,34270
are the head counted trajectories scene. accumulating the leaving by People eventually,People are eventually counted by accumulating the head trajectories leaving the scene.,34271
"and identification, To trajectory also two different effective enable head features. propose compound we","To enable effective head and trajectory identification, we also propose two different compound features.",34272
"update on based the over estimates to time responses. students Second, the","Second, to update the estimates over time based on the students responses.",34273
"wild, e-learning the results. system a in MAPLE in promising showed implemented live When","When implemented in a live e-learning system in the wild, MAPLE showed promising results.",34274
"Human animation fashion and important an for video is shape editing, estimation industry. task","Human shape estimation is an important task for video editing, animation and fashion industry.",34275
results demonstrated by in Each our performance as of them experiments. improvement,Each of them results in performance improvement as demonstrated by our experiments.,34276
"enables also achieving segmentation. Besides body-part performance, volumetric our state-of-the-art method","Besides achieving state-of-the-art performance, our method also enables volumetric body-part segmentation.",34277
"this a for we in SoccerNet, introduce spotting In paper, benchmark videos. action soccer","In this paper, we introduce SoccerNet, a benchmark for action spotting in soccer videos.",34278
"is such, dataset easily scalable. the As","As such, the dataset is easily scalable.",34279
We extensive datasets. evaluate models to our conduct on real-world experiments,We conduct extensive experiments to evaluate our models on real-world datasets.,34280
approach in group detecting effective our empirical demonstrate and that results anomalies. is The robust,The empirical results demonstrate that our approach is effective and robust in detecting group anomalies.,34281
video be such remains (though tested). performance sequences) domains to in its,video sequences) (though its performance in such domains remains to be tested).,34282
and using subjective is with assessed objective algorithm criteria results. very The good,The algorithm is assessed using subjective and objective criteria with very good results.,34283
"between distortion. reduce the angular the in representations to Third, order spectral difference minimized are","Third, the angular difference between representations are minimized in order to reduce the spectral distortion.",34284
"unsupervised or We refer Sparse proposed the as to Dirichlet-Net, uSDN. architecture","We refer to the proposed architecture as unsupervised Sparse Dirichlet-Net, or uSDN.",34285
demonstrate uSDN performance of state-of-the-art. Extensive compared experimental to the superior results as the,Extensive experimental results demonstrate the superior performance of uSDN as compared to the state-of-the-art.,34286
quality with verify the the trajectory Experiments quality. obtained corresponds that requested well,Experiments verify that the obtained trajectory quality corresponds well with the requested quality.,34287
exploit The information vary. and way to the can it,The information and the way to exploit it can vary.,34288
temporal data sets. We on network test these empirical,We test these on empirical temporal network data sets.,34289
learn to images. input Learning-based approaches color enhancement map typically from to retouched images,Learning-based color enhancement approaches typically learn to map from input images to retouched images.,34290
the we train enhancement optimal of our Then global actions. learn agent to sequence the,Then we train our agent to learn the optimal global enhancement sequence of the actions.,34291
lenses the distortions. from Images assumption violate and camera captured fisheye by pinhole suffer,Images captured by fisheye lenses violate the pinhole camera assumption and suffer from distortions.,34292
vision crucial preprocessing Rectification is many a computer images for of therefore fisheye step applications.,Rectification of fisheye images is therefore a crucial preprocessing step for many computer vision applications.,34293
will publicly be and available. Our dataset code synthesized made,Our code and synthesized dataset will be made publicly available.,34294
full-wave also verified (MoM). The the model using is against Method Moments simplified of simulation,The simplified model is also verified against full-wave simulation using the Method of Moments (MoM).,34295
"are In contrast, image to tags easier simple gather.","In contrast, simple image tags are easier to gather.",34296
great planar which performances estimation. detection markers are markers pose Circular for offer and,Circular markers are planar markers which offer great performances for detection and pose estimation.,34297
"to considered be standard"". Usually, is manual ""gold annotation the","Usually, manual annotation is considered to be the ""gold standard"".",34298
"larger results (i.e., we Dice current coefficients) the showed SS that state-of-the-art Our methods. outperformed","Our results showed that we outperformed (i.e., larger Dice coefficients) the current state-of-the-art SS methods.",34299
This domain known problem. as is adaptation the,This is known as the domain adaptation problem.,34300
"framework novel domain-adaptive this dictionary learning a visual cross-domain In recognition. work, propose for we","In this work, we propose a novel domain-adaptive dictionary learning framework for cross-domain visual recognition.",34301
of intermediate set domains. method Our generates a,Our method generates a set of intermediate domains.,34302
dictionary steps. dictionaries sparse These alternating by updating and between domain-adaptive coding learned are,These dictionaries are learned by alternating between domain-adaptive sparse coding and dictionary updating steps.,34303
outperforms approach methods. three experiments state-of-the-art datasets most that our demonstrates on public Extensive,Extensive experiments on three public datasets demonstrates that our approach outperforms most state-of-the-art methods.,34304
of method. Extensive the validate experiments effectiveness our,Extensive experiments validate the effectiveness of our method.,34305
against predicting candidate protection human This tool by circulating vaccine strains. selection assists,This tool assists candidate vaccine selection by predicting human protection against circulating strains.,34306
signals. to one combines Each up utilized multiple snapshot signal with interfering,Each snapshot combines one utilized signal with up to multiple interfering signals.,34307
space. infinite Most can a be represented measure a in as Gaussian processes dimensional Gaussian,Most Gaussian processes can be represented as a Gaussian measure in a infinite dimensional space.,34308
"algorithm we for statistical proposed our three efficiency of different demonstrate models. the Finally,","Finally, we demonstrate the efficiency of our proposed algorithm for three different statistical models.",34309
vectors the two a coordinates foreground. indicate by the space Only few expanded in base,Only a few coordinates in the space expanded by the two base vectors indicate foreground.,34310
"to a Furthermore, FCN masked-convolution operation based the on calculation. accelerate attention result we adopt","Furthermore, we adopt a masked-convolution operation based on the attention result to accelerate FCN calculation.",34311
videos. spatial temporal features problem disentangling and approach by We this in,We approach this problem by disentangling spatial and temporal features in videos.,34312
representations. for video MSnet can be used show prediction using disentangled how We,We show how MSnet can be used for video prediction using disentangled representations.,34313
method disentangle experiments assess of also features. We effectiveness to the carry our to out,We also carry out experiments to assess the effectiveness of our method to disentangle features.,34314
trivial relationship expectation. intertwining problem for is not data The characteristics between and user,The problem is not trivial for intertwining relationship between data characteristics and user expectation.,34315
"for objects accurate critical in detection salient is of interaction. Thus, human-computer scenes complex","Thus, accurate detection of salient objects in complex scenes is critical for human-computer interaction.",34316
first entropy atmospheric for estimating Gaussian used is that the is This image time visibility.,This is the first time that Gaussian image entropy is used for estimating atmospheric visibility.,34317
"were foggy hazy and videos collected constantly the Jiangsu, The from China. in expressway Tongqi","The videos were collected from the constantly foggy and hazy Tongqi expressway in Jiangsu, China.",34318
assessments. many are cornerstone methods stock the of population Length-based and studies,Length-based methods are the cornerstone of many population studies and stock assessments.,34319
"possible Bias in without corrections. Z/K a complex behavior, showed","Bias in Z/K showed a complex behavior, without possible corrections.",34320
"be Estimation tentatively may external very of verifications. necessary with conditions, Linf specific possible under","Estimation of Linf may be tentatively possible under very specific conditions, with necessary external verifications.",34321
"after, (before, relations  temporal etc.) Extracting  overlapping,","  Extracting temporal relations (before, after, overlapping, etc.)",34322
described a key understanding is language. in events natural aspect of,is a key aspect of understanding events described in natural language.,34323
this existing resource. can extraction show be systems via that improved temporal We,We show that existing temporal extraction systems can be improved via this resource.,34324
system The are available. publicly resource proposed and both,The proposed system and resource are both publicly available.,34325
facilitated space. bas-relief has modeling in normal image bas-relief images from successfully Creating,Creating bas-relief from normal images has successfully facilitated bas-relief modeling in image space.,34326
"method bas-relief produces easy is and Our models. to implement, good-quality","Our method is easy to implement, and produces good-quality bas-relief models.",34327
one due in is information contained image. single Single-image challenging haze-removal to limited,Single-image haze-removal is challenging due to limited information contained in one single image.,34328
largely this handcrafted Previous solutions rely for deficiency. priors on to compensate,Previous solutions largely rely on handcrafted priors to compensate for this deficiency.,34329
towards haze image propose ap- this novel paper single proach we semantic In a removal.,In this paper we propose a novel semantic ap- proach towards single image haze removal.,34330
"based features. we priors Unlike semantic existing on methods, extracted color infer","Unlike existing methods, we infer color priors based on extracted semantic features.",34331
image. regions color sky in and saturated illumination,saturated illumination color and sky regions in image.,34332
formulation. by mathematical its simplify developing We an alternative calculation,We simplify its calculation by developing an alternative mathematical formulation.,34333
and the scope new applicability. extends formula The the computational significantly therefore of efficiency improves,The new formula significantly improves the computational efficiency and therefore extends the scope of applicability.,34334
"the the formula, Based develop on ordered Shapley method. value simplified further we","Based on the simplified formula, we further develop the ordered Shapley value method.",34335
using illustrated are online proposed a campaign approaches advertising dataset. real-world The,The proposed approaches are illustrated using a real-world online advertising campaign dataset.,34336
"real-time low-cost object FPGA-based we Moreover, hardware present tracking a architecture.","Moreover, we present a low-cost FPGA-based real-time object tracking hardware architecture.",34337
"high-pass has Gaussian applied. interaction response wavelet before is First, the the be to filtered","First, the wavelet response has to be high-pass filtered before the Gaussian interaction is applied.",34338
"interaction. also subband Gaussian per Then, required after are distinct the weights","Then, distinct weights per subband are also required after the Gaussian interaction.",34339
limited information-processing Here with and perception as consider channels we capacity. information serial action two,Here we consider perception and action as two serial information channels with limited information-processing capacity.,34340
We simplified a illustrate lifting with task. robot in method simulator our cup NAO a,We illustrate our method in a NAO robot simulator with a simplified cup lifting task.,34341
require real time. streams objects those from in that video detecting,those that require detecting objects from video streams in real time.,34342
while maintaining competitive performance. reducing the cost computing,reducing the computing cost while maintaining competitive performance.,34343
"focus previous efforts architectures. seek To on balance, good the optimizing a usually model","To seek a good balance, previous efforts usually focus on optimizing the model architectures.",34344
"and can various balance this one performance explore framework, to On cost. strategies","On this framework, one can explore various strategies to balance performance and cost.",34345
genetic optimization using The algorithms. is performed,The optimization is performed using genetic algorithms.,34346
how the process. objects IterGANs show on Experiments scenes rotated and help with generation,Experiments on rotated objects and scenes show how IterGANs help with the generation process.,34347
"identification this handwriting. from effective online new a In paper, writer present biometric system we","In this paper, we present a new effective biometric writer identification system from online handwriting.",34348
"as used classifier. Neural is Deep Network Finally,","Finally, Deep Neural Network is used as classifier.",34349
"a To of our series method, experiments. perform we evaluate","To evaluate our method, we perform a series of experiments.",34350
use data accuracy images the We two-dimensional our method. to of synthetic first demonstrate and,We first use two-dimensional images and synthetic data to demonstrate the accuracy of our method.,34351
technique assess proposes to quality of the photographs. This paper a aesthetic,This paper proposes a technique to assess the aesthetic quality of photographs.,34352
"from the the system a downloaded proposed large images validate internet. dataset, We on containing","We validate the proposed system on a large dataset, containing images downloaded from the internet.",34353
"periodicities are two sound double present not, making fundamental If periodicity, around the the arbitrary.","If not, two periodicities are present around double the fundamental periodicity, making the sound arbitrary.",34354
"on is real proposed of world approach the scenes. Further, feasibility illustrated the several","Further, the feasibility of the proposed approach is illustrated on several real world scenes.",34355
convolutional networks training. way compare neural We quantitatively conventional new of training this with,We quantitatively compare this new way of training convolutional neural networks with conventional training.,34356
image. an from normals train CNN to this surface and a SVBRDF regress We,We train a CNN to regress an SVBRDF and surface normals from this image.,34357
high end-to-end a quality trained results is for variety The of and produces framework materials.,The framework is trained end-to-end and produces high quality results for a variety of materials.,34358
approach matrix matricized to methods State-of-the-art the MSI. and problem via low-rank approximations HSI the,State-of-the-art methods approach the problem via low-rank matrix approximations to the matricized HSI and MSI.,34359
"challenges These remain. of effective methods are extent, number a some but to","These methods are effective to some extent, but a number of challenges remain.",34360
"we the propose In from this to perspective. work, problem a super-resolution tackle tensor","In this work, we propose to tackle the super-resolution problem from a tensor perspective.",34361
identifiability realistic the approach mild the and of under conditions. proposed The SRI guarantees,The proposed approach guarantees the identifiability of the SRI under mild and realistic conditions.,34362
the approach. the effectiveness included to numerical are show experiments proposed of Semi-real,Semi-real numerical experiments are included to show the effectiveness of the proposed approach.,34363
paper. three main contributions are of There this,There are three main contributions of this paper.,34364
the benefits datasets benchmark our Experiments model. on show two of,Experiments on two benchmark datasets show the benefits of our model.,34365
"due data limitations. However obtaining challenging in both to many channels is reality,","However in reality, obtaining both data channels is challenging due to many limitations.",34366
speech automatically can training does for data require WCSEPD extract not labeled any segments. and,WCSEPD does not require any labeled data for training and can automatically extract speech segments.,34367
provide to gains in significant is anticipated scenarios. (mmWave) communication Millimeter-wave throughout urban,Millimeter-wave (mmWave) communication is anticipated to provide significant throughout gains in urban scenarios.,34368
new We under this show conjunctive that is query answering of decidable. class TGDs,We show that conjunctive query answering under this new class of TGDs is decidable.,34369
"on we retrieval paper, for triplet-based embedding task. binary this image In networks deep focus","In this paper, we focus on triplet-based deep binary embedding networks for image retrieval task.",34370
The for be been has the loss to effective triplet problem. most shown ranking,The triplet loss has been shown to be most effective for the ranking problem.,34371
"important is retrieval do task. consider relations, Such strategies which not the order for","Such strategies do not consider the order relations, which is important for retrieval task.",34372
"at address the levels. we structure-function both Importantly, group and individual-based the relationship","Importantly, we address the structure-function relationship both at the group and individual-based levels.",34373
clusters subspace of Sparse $n$ lie low-dimensional a that subspaces. near clustering points union (SSC),Sparse subspace clustering (SSC) clusters $n$ points that lie near a union of low-dimensional subspaces.,34374
"and German. five for Dutch, English, Spanish, PNNs languages: study We French,","We study PNNs for five languages: English, Spanish, French, Dutch, and German.",34375
biological are functions. into their Proteins structures fold complex MOTIVATION: that crucial for,MOTIVATION: Proteins fold into complex structures that are crucial for their biological functions.,34376
of databases the on protein We of transfer structures. also show learning applying possibility known,We also show the possibility of applying transfer learning on databases of known protein structures.,34377
of brain proposed the image an input identifies MR automatically CNN volume. contrast The,The proposed CNN automatically identifies the MR contrast of an input brain image volume.,34378
"in and training framework of forward is a decision-making paper. Therefore, this the put learning","Therefore, a framework of the decision-making training and learning is put forward in this paper.",34379
"microscopic within Then the framework. basic is car-following, trained this behavior,","Then the basic microscopic behavior, car-following, is trained within this framework.",34380
that Automated for learning. is effective vehicle's our It indicates decision-making framework,It indicates that our framework is effective for Automated vehicle's decision-making learning.,34381
entropy functions computation SpatEntropy entropy measures. and for of R package The spatial contains the,The R package SpatEntropy contains functions for the computation of entropy and spatial entropy measures.,34382
measures a of feature entropy SpatEntropy. extension The spatial unique is to,The extension to spatial entropy measures is a unique feature of SpatEntropy.,34383
The able package and point with data. both work is to areal,The package is able to work with both areal and point data.,34384
Virtual reality the and significant years. in past have growth technologies few augmented seen,Virtual and augmented reality technologies have seen significant growth in the past few years.,34385
our mobile models constraints time CPU inference. operate under the of strict real All,All our models operate under the strict constraints of real time mobile CPU inference.,34386
capsules This deep. networks implies can't very be supervised that,This implies that supervised capsules networks can't be very deep.,34387
"soft-attention solve To incorporating mechanisms. we self-gated propose this,","To solve this, we propose incorporating self-gated soft-attention mechanisms.",34388
training paired motion is different characters expensive. from sequences Collecting,Collecting paired motion training sequences from different characters is expensive.,34389
kernel derived. is its Various by satisfied transformation properties,Various properties satisfied by its transformation kernel is derived.,34390
leads paired users the low to for spectral This efficiency NOMA. in,This leads to low spectral efficiency for the paired users in NOMA.,34391
outperforms is MIMO-NOMA. that shown It NOMA-GSSK,It is shown that NOMA-GSSK outperforms MIMO-NOMA.,34392
co-occurrence propose end-to-end an framework. we learning this paper convolutional In feature,In this paper we propose an end-to-end convolutional co-occurrence feature learning framework.,34393
information Firstly each joint of point-level is independently. encoded,Firstly point-level information of each joint is encoded independently.,34394
in into they Then assembled spatial representation semantic temporal domains. and are both,Then they are assembled into semantic representation in both spatial and temporal domains.,34395
also We practical and considerations. the most questions biological highlight statistical of,We also highlight the biological and statistical questions of most practical considerations.,34396
applied key frames. Light image detector weight object is on sparse,Light weight image object detector is applied on sparse key frames.,34397
"A very frames. establishing network, across small for is designed correspondence Light Flow,","A very small network, Light Flow, is designed for establishing correspondence across frames.",34398
designed A key GRU flow-guided to is effectively aggregate frames. features module on,A flow-guided GRU module is designed to effectively aggregate features on key frames.,34399
"performed. is propagation frames, sparse feature For non-key","For non-key frames, sparse feature propagation is performed.",34400
whole end-to-end. trained can network The be,The whole network can be trained end-to-end.,34401
"pathological and backgrounds, from types, lighting Challenges states. arise varied conditions, skin","Challenges arise from varied lighting conditions, skin types, backgrounds, and pathological states.",34402
Residual evaluated: variants U-Net. are of Dense U-Net Two U-Net and,Two variants of U-Net are evaluated: U-Net and Dense Residual U-Net.,34403
have remarkable networks vision computer progresses various neural made Deep tasks. on,Deep neural networks have made remarkable progresses on various computer vision tasks.,34404
"our further boost module Moreover, is to attention network's an introduced performance.","Moreover, an attention module is introduced to further boost our network's performance.",34405
generate paper novel systems. to handwritten recognition dataset synthetic approach a for This presents word,This paper presents a novel approach to generate synthetic dataset for handwritten word recognition systems.,34406
"lack to hard owing Hence, becomes train recognition proper dataset. to systems it of","Hence, it becomes hard to train recognition systems owing to lack of proper dataset.",34407
isolated We to words. experimented recognition of accuracy characters the synthetic data using and improve,We experimented using synthetic data to improve the recognition accuracy of isolated characters and words.,34408
have results obtained We the experiment. encouraging from,We have obtained encouraging results from the experiment.,34409
"Latin experiment Finally, verifies utility text the with the of approach. the","Finally, the experiment with Latin text verifies the utility of the approach.",34410
state-of-the-art perspectives. A summary made of from works the two these is,A summary of the state-of-the-art works is made from these two perspectives.,34411
simple methods. effective and baseline This work provides,This work provides simple and effective baseline methods.,34412
inspiring field. for helpful are new evaluating They for and the ideas,They are helpful for inspiring and evaluating new ideas for the field.,34413
State-of-the-art are benchmarks. results challenging on achieved,State-of-the-art results are achieved on challenging benchmarks.,34414
"tasks image and classification of differences detection. there the several More importantly, object are between","More importantly, there are several differences between the tasks of image classification and object detection.",34415
will for be the reproduction. released code The,The code will be released for the reproduction.,34416
"caused a blur clear by pixel degradation. recovers from deconvolution, image superposition image","image deconvolution, recovers a clear image from pixel superposition caused by blur degradation.",34417
deep addressing this convolutional succeed neural Few (CNN) task. in networks,Few deep convolutional neural networks (CNN) succeed in addressing this task.,34418
"propose we iterative deconvolution novel this analysis, (IRD) residual a Based algorithm. on","Based on this analysis, we propose a novel iterative residual deconvolution (IRD) algorithm.",34419
"involves the expected When perfectly. objective the reward this cumulative only, works simplification","When the objective involves the expected cumulative reward only, this simplification works perfectly.",34420
"this to value. risk-sensitive, when However, an simplification the is incorrect objective leads","However, when the objective is risk-sensitive, this simplification leads to an incorrect value.",34421
proposed problem. Secondly in SATs two situations illustrate inventory control that an the we require,Secondly we illustrate that two situations require the proposed SATs in an inventory control problem.,34422
with speed accuracy handle compromise to or large methods Most RIP the variations. existing,Most existing methods compromise with speed or accuracy to handle the large RIP variations.,34423
face to a real-time detector. designs Such rotation-invariant lead,Such designs lead to a real-time rotation-invariant face detector.,34424
"simulations. but we not our least, by validate numerical Last analysis","Last but not least, we validate our analysis by numerical simulations.",34425
"performance handcrafted in improve still cannot Learning-based but approaches tasks, features general. certain replace in","Learning-based approaches improve performance in certain tasks, but still cannot replace handcrafted features in general.",34426
"metric, a Average neural ranking-based directly networks. optimize deep Specifically, we performance Precision, using retrieval","Specifically, we directly optimize a ranking-based retrieval performance metric, Average Precision, using deep neural networks.",34427
of provided the to proposed validate are Empirical the also studies effectiveness method.,Empirical studies are also provided to validate the effectiveness of the proposed method.,34428
"for Thresholding drains subtle noise Curvelet image out of Coefficients, in subspace. component denoising, signal","Thresholding of Curvelet Coefficients, for image denoising, drains out subtle signal component in noise subspace.",34429
denoised near edges effect artifacts in image. granular ringing produces and the This,This produces ringing artifacts near edges and granular effect in the denoised image.,34430
"to the information. apply keep Bilateral finest scale, we Filter (BF) In edge","In the finest scale, we apply Bilateral Filter (BF) to keep edge information.",34431
a novel synthesis paper image. single This the problem view of tackles from,This paper tackles the problem of novel view synthesis from a single image.,34432
scheme power is also an prove the We that achieve sufficient to On-Off asymptotically capacity.,We also prove that an On-Off power scheme is sufficient to asymptotically achieve the capacity.,34433
"CNN-based address Siamese To propose we architecture. encoder-decoder this a task,","To address this task, we propose a CNN-based Siamese encoder-decoder architecture.",34434
"communities. Yet, in mostly both been separate discussed have distributions","Yet, both distributions have been discussed in mostly separate communities.",34435
"using assembly both patterns community spatially we approach. simultaneously explicit, Here, evolutionary study a","Here, we study both patterns simultaneously using a spatially explicit, evolutionary community assembly approach.",34436
"Colorful Voxel Here, problem. end-to-end tackle an Network this framework, (CVN), propose we to trainable","Here, we propose an end-to-end trainable framework, Colorful Voxel Network (CVN), to tackle this problem.",34437
seen (ZSL) objects via recognize attributes. unseen to using aims disjoint objects learning Zero-shot sharing,Zero-shot learning (ZSL) aims to recognize unseen objects using disjoint seen objects via sharing attributes.,34438
"To aggregation to rank is framework proposed solve the assumption. this circumvent novel a problem,","To solve this problem, a novel rank aggregation framework is proposed to circumvent the assumption.",34439
Polynomial of large number problem applications. recurrent a is regression with a,Polynomial regression is a recurrent problem with a large number of applications.,34440
appears analysis. motion In computer often vision it in,In computer vision it often appears in motion analysis.,34441
"Moreover, problem harder have structure. even outliers the strong when is","Moreover, the problem is even harder when outliers have strong structure.",34442
with robust value standard demonstrate by comparing our of findings methods. the regression We,We demonstrate the value of our findings by comparing with standard robust regression methods.,34443
results. show ontologies promising with real-world Experiments,Experiments with real-world ontologies show promising results.,34444
the update rate. CANDID adaptively initializes pixel-level threshold and distance The,The CANDID adaptively initializes the pixel-level distance threshold and update rate.,34445
the at parameters These computing a change location. updated dynamics by are,These parameters are updated by computing the change dynamics at a location.,34446
"background update is maintained a Further, by formulating the model deterministic policy.","Further, the background model is maintained by formulating a deterministic update policy.",34447
state-of-the-art approaches. of measures the qualitative The existing quantitative and method the outperform proposed,The qualitative and quantitative measures of the proposed method outperform the existing state-of-the-art approaches.,34448
automated techniques This calls to refereeing. order assist detection offside accurate for in,This calls for automated offside detection techniques in order to assist accurate refereeing.,34449
"similarity. Hough KLT Keywords: color connected components, tracking, transform,","Keywords: Hough transform, connected components, KLT tracking, color similarity.",34450
is spatio-temporal for of crucial most classification. action Acquiring step the action states an,Acquiring spatio-temporal states of an action is the most crucial step for action classification.,34451
signal phase method. have stationary already We by NLFM designed,We have already designed NLFM signal by stationary phase method.,34452
Underwater The is gaining as Communications popularity. Acoustic use systems of sensors vector for receivers,The use of vector sensors as receivers for Underwater Acoustic Communications systems is gaining popularity.,34453
was channel novel recent capacity of evaluating underwater framework work for In a proposed.,In a recent work novel framework for evaluating capacity of underwater channel was proposed.,34454
Bellhop tool. channel The statistics are by the analysing parameters modeled with generated simulation,The channel parameters are modeled by analysing the statistics generated with Bellhop simulation tool.,34455
"are base First, decomposed images detail parts content. into the source and","First, the source images are decomposed into base parts and detail content.",34456
by Then weighted-averaging. the are base parts fused,Then the base parts are fused by weighted-averaging.,34457
"to network deep learning use features. we For detail extract content, a the multi-layer","For the detail content, we use a deep learning network to extract multi-layer features.",34458
in generalized recurrency is show a rigidity. fact We that,We show that recurrency is in fact a generalized rigidity.,34459
simulated data sequences of both effectiveness the and on demonstrate Experiments the real method.,Experiments on both simulated sequences and real data demonstrate the effectiveness of the method.,34460
baseline present neural convolutional using We networks. a system,We present a baseline system using convolutional neural networks.,34461
encodings proposed real-world are evaluated a using experimentally All dataset.,All proposed encodings are experimentally evaluated using a real-world dataset.,34462
in manuscript under is consideration for TPLP. acceptance This,This manuscript is under consideration for acceptance in TPLP.,34463
Approximation geometric problems. many a of data often is scattered task in engineering,Approximation of scattered geometric data is often a task in many engineering problems.,34464
"The of two points. is is a on RBF distance as it based non-separable, approximation","The RBF approximation is non-separable, as it is based on a distance of two points.",34465
to (LSE). of a solution Equations leads of System overdetermined It Linear,It leads to a solution of overdetermined Linear System of Equations (LSE).,34466
new method this presented. In a derived is and paper RBF approximation,In this paper a new RBF approximation method is derived and presented.,34467
The for general. cases d in dimensional applicable presented approach is,The presented approach is applicable for d dimensional cases in general.,34468
architecture a then We present including two parallel-classifiers localization. network simple object for,We then present a simple network architecture including two parallel-classifiers for object localization.,34469
a the of our experiments. variety ACoL of in approach demonstrate superiority We,We demonstrate the superiority of our ACoL approach in a variety of experiments.,34470
"training for evaluation. summarize and we model and databases commonly Then, both used the compare","Then, we summarize and compare the commonly used databases for both model training and evaluation.",34471
"several promising highlighted. technical the are challenges directions and Finally,","Finally, the technical challenges and several promising directions are highlighted.",34472
designed and video laboriously hand Traditional is compression optimized. hand,Traditional video compression is laboriously hand designed and hand optimized.,34473
alternative an This codec. end-to-end in deep paper an learning presents,This paper presents an alternative in an end-to-end deep learning codec.,34474
Video interpolation. simple image builds repeated compression codec one is Our idea: on,Our codec builds on one simple idea: Video compression is repeated image interpolation.,34475
generation. and It benefits from image deep advances recent interpolation in thus,It thus benefits from recent advances in deep image interpolation and generation.,34476
are face threat attacks Spoofing a recognition systems. to modern,Spoofing attacks are a threat to modern face recognition systems.,34477
authentication of This makes design highly possible a the active-light secure framework.,This makes possible the design of a highly secure active-light authentication framework.,34478
We part-aligned learns representation re-identification. that novel person a network propose a for,We propose a novel network that learns a part-aligned representation for person re-identification.,34479
representation significantly. misalignment problem This part-aligned the reduces part,This part-aligned representation reduces the part misalignment problem significantly.,34480
object synthetic detection neural system for deep present training using networks images. We for a,We present a system for training deep neural networks for object detection using synthetic images.,34481
dataset. the approach box is KITTI The of bounding cars on detection on evaluated,The approach is evaluated on bounding box detection of cars on the KITTI dataset.,34482
Recent achieved state-of-the-art object convolutional neural based accuracy. detectors deep network have,Recent deep convolutional neural network based object detectors have achieved state-of-the-art accuracy.,34483
popular are techniques computation requirements networks reducing neural footprint. for and Low-precision the memory,Low-precision neural networks are popular techniques for reducing the computation requirements and memory footprint.,34484
We fast beamforming independent scheme a simple with likelihood maximum detection. symbol-wise propose channel (ML),We propose a simple channel independent beamforming scheme with fast symbol-wise maximum likelihood (ML) detection.,34485
the visually that images The achieve with can result proposed identical show method accuracy. comparable,The result show that the proposed method can achieve visually identical images with comparable accuracy.,34486
computational a plenoptic of framework presents map paper estimating accurately This disparity the for images.,This paper presents a computational framework for accurately estimating the disparity map of plenoptic images.,34487
variational intrinsic provides the The principle is proposed and on sub-pixel precision. framework based,The proposed framework is based on the variational principle and provides intrinsic sub-pixel precision.,34488
the warping A problem. large embedded is our tackling framework strategy for displacement in,A warping strategy is embedded in our framework for tackling the large displacement problem.,34489
"of a In detecting radiography. method in novel paper, this we propose chest pneumothorax","In this paper, we propose a novel method of detecting pneumothorax in chest radiography.",34490
be is medical Our method to images. when effective CNN applying to size expected large,Our method is expected to be effective when applying CNN to large size medical images.,34491
promising solution low-resolution A (ADCs). is employ to analog-to-digital converters,A promising solution is to employ low-resolution analog-to-digital converters (ADCs).,34492
"it performance by and Most degradation ADCs error. quantifies importantly, caused low-resolution the estimation channel","Most importantly, it quantifies the performance degradation caused by low-resolution ADCs and channel estimation error.",34493
like a topic applications many detection serves for Face as fundamental research face recognition.,Face detection serves as a fundamental research topic for many applications like face recognition.,34494
networks. convolutional made the progress been has recent Impressive development neural with of especially,Impressive progress has been made especially with the recent development of convolutional neural networks.,34495
to across of state efficiently Another objects temporally is the represent frames. challenge,Another challenge is to efficiently represent the state of objects temporally across frames.,34496
is multi-object (GLMB) tracker filter proposed. multi-Bernoulli on A labeled based generalised large-scale the,A large-scale multi-object tracker based on the generalised labeled multi-Bernoulli (GLMB) filter is proposed.,34497
"a model For applied using against pretraining uncontrollable a training, approach stable we obstacles. robust","For stable training, we applied a pretraining approach using a model robust against uncontrollable obstacles.",34498
"worthwhile framework breast automatic to identification cancer classification Consequently, an develop. is for","Consequently, an automatic classification framework for breast cancer identification is worthwhile to develop.",34499
deep development technique. of witnessed learning the years Recent,Recent years witnessed the development of deep learning technique.,34500
medical learning diagnosis use Increasing applications to improve start number accuracy. deep to of,Increasing number of medical applications start to use deep learning to improve diagnosis accuracy.,34501
"current al. visual (Vaquero itSIMPLE et However, modeling, knowledge engineering with tools like","However, current knowledge engineering tools with visual modeling, like itSIMPLE (Vaquero et al.",34502
it The understanding is problem challenging the and needs of as sentence. video both,The problem is challenging as it needs the understanding of both video and sentence.,34503
trained manner. end-to-end is in an ABLR jointly,ABLR is jointly trained in an end-to-end manner.,34504
treatments to movement estimate across effects. multiple leverage Researchers causal increasingly,Researchers increasingly leverage movement across multiple treatments to estimate causal effects.,34505
to Future al. the will theory apply et the drafts Finkelstein,Future drafts will apply the theory to the Finkelstein et al.,34506
distribution rely Their graphical normalized observed the incomes. analysis on of of finding the,Their finding rely on the graphical analysis of the observed distribution of normalized incomes.,34507
from test the at posterior Monte are samples generated distribution Carlo time. dropout using efficiently,Monte Carlo samples from the posterior distribution are efficiently generated using dropout at test time.,34508
"analysis manual results that group uncertainty sizes with to annotations. effect of Furthermore, closer in","Furthermore, group analysis with uncertainty results in effect sizes closer to that of manual annotations.",34509
challenging experiments are four performed Extensive on datasets.,Extensive experiments are performed on four challenging datasets.,34510
frames. between change of the We strength obtain similarity calculate content-motion of to consecutive the,We calculate the similarity of content-motion to obtain the strength of change between consecutive frames.,34511
concisely. of categories underlying introduced The each research the areas are,The research areas underlying each of the categories are introduced concisely.,34512
structural another. one deformation from form a drastically Extreme into morph can structure,Extreme deformation can drastically morph a structure from one structural form into another.,34513
algorithms impact made in medical field have a imaging. tremendous of the Machine (ML) learning,Machine learning (ML) algorithms have made a tremendous impact in the field of medical imaging.,34514
"and connections We scenarios, discuss for these opportunities learning future also research. between","We also discuss connections between these learning scenarios, and opportunities for future research.",34515
"perturbations(i.e. modified such is, by images That adding","That is, images modified by adding such perturbations(i.e.",34516
"with models; we proposed CNN addition, and In AlexNet compared our two well-known VGGNet. classifier","In addition, we compared our proposed classifier with two well-known CNN models; AlexNet and VGGNet.",34517
for classifier. were arrhythmia from of ECG the the database MIT-BIH the evaluation used recordings,ECG recordings from the MIT-BIH arrhythmia database were used for the evaluation of the classifier.,34518
describe in the We detail statistical the process generation of and analysis data.,We describe in detail the generation process and statistical analysis of the data.,34519
studies model co-occurrence learn events latent of Existing EHR embeddings. to capturing clinical their by,Existing studies model EHR by capturing co-occurrence of clinical events to learn their latent embeddings.,34520
Our handling modeling values data. and of heterogeneity missing straightforward of for allows approach,Our modeling approach allows for straightforward handling of missing values and heterogeneity of data.,34521
higher contributing diagnosis. and HeteroMed level exploits to disease semantically metapaths important to capture relations,HeteroMed exploits metapaths to capture higher level and semantically important relations contributing to disease diagnosis.,34522
possibilities introduces for safety-critical of applications. localization new precise Integration,Integration of precise localization introduces new possibilities for safety-critical applications.,34523
are restart strategies search. for two tabu The local fundamental and,The tabu and restart are two fundamental strategies for local search.,34524
human Tasks long been the computer have hands community. related part of to vision,Tasks related to human hands have long been part of the computer vision community.,34525
everyday multimedia and We are creating everywhere. contents,We are creating multimedia contents everyday and everywhere.,34526
"approach images. For rewarding to synthesize (GANs) Adversarial example, is Networks a the Generative","For example, the Generative Adversarial Networks (GANs) is a rewarding approach to synthesize images.",34527
"to videos. on GANs not it is Nevertheless, capitalizing when generate trivial","Nevertheless, it is not trivial when capitalizing on GANs to generate videos.",34528
us This semantic in coherence temporal generate designing videos. to and explore GANs motivates to,This motivates us to explore semantic and temporal coherence in designing GANs to generate videos.,34529
"paper, provide approach develop problem. solution In this an this effective we to to an","In this paper, we develop an approach to provide an effective solution to this problem.",34530
"for lane-level are to reach able nearly truth we Having ground accuracy. training, locations","Having ground truth locations for training, we are able to reach nearly lane-level accuracy.",34531
problem common is (e.g. computer Counting vision in a,Counting is a common problem in computer vision (e.g.,34532
a or the street pedestrians crowd). on in traffic,traffic on the street or pedestrians in a crowd).,34533
"However, at different the (e.g. in occur objects scales image can","However, objects in the image can occur at different scales (e.g.",34534
in proposal. of turn locations the event temporal This each adjusts,This in turn adjusts the temporal locations of each event proposal.,34535
method. The of trend of minimum the error decrement guarantees proposed convergence,The trend decrement of minimum error guarantees convergence of the proposed method.,34536
We under accurate assumptions. then derive distributional expressions more common,We then derive more accurate expressions under common distributional assumptions.,34537
the Single-Image recent Super-Resolution. of progress Neural have been in Convolutional Networks rapid backbone,Convolutional Neural Networks have been the backbone of recent rapid progress in Single-Image Super-Resolution.,34538
approach. of our the effectiveness benchmark Extensive demonstrate evaluations,Extensive benchmark evaluations demonstrate the effectiveness of our approach.,34539
the tags short Then their extract to of only videos take features. we,Then we take short videos of only the tags to extract their features.,34540
diagnostic been support Expert studied. extensively systems have,Expert diagnostic support systems have been extensively studied.,34541
cerevisiae characterized best eukaryotic Saccharomyces is yeast one The of models. the,The yeast Saccharomyces cerevisiae is one of the best characterized eukaryotic models.,34542
"trafficking S. of we Here, present different cerevisiae. pathways yeast the","Here, we present the different trafficking pathways of yeast S. cerevisiae.",34543
quickly extensively high-quality Planning preferences to plans. generate has with employed been,Planning with preferences has been employed extensively to quickly generate high-quality plans.,34544
more for the design It detection to researchers produce in robust inspires face wild.,It inspires researchers to produce more robust design for face detection in the wild.,34545
in tasks. have great based image found deraining Deep learning methods success,Deep learning based methods have found great success in image deraining tasks.,34546
into blocks reconstruction. merged the different of final outputs The are,The outputs of different blocks are merged into the final reconstruction.,34547
conditions. detachable rainy is different meet ResGuideNet to,ResGuideNet is detachable to meet different rainy conditions.,34548
vision other can low- benefit Experiments and that ResGuideNet high-level validate tasks.,Experiments validate that ResGuideNet can benefit other low- and high-level vision tasks.,34549
We a single consider period. single product and,We consider a single product and single period.,34550
a we duopoly country. in the competition foreign consider,we consider a duopoly competition in the foreign country.,34551
the random exchange model We a variable. rate as,We model the exchange rate as a random variable.,34552
We and impact optimal competition exchange prices. investigate rate of on capacities and the optimal,We investigate the impact of competition and exchange rate on optimal capacities and optimal prices.,34553
"resources time, the traditional As space, etc.) (frequency,","As the traditional resources (frequency, time, space, etc.)",34554
is helical wavefront kind momentum phase. angular (OAM) The with of orbital a,The orbital angular momentum (OAM) is a kind of wavefront with helical phase.,34555
We data readouts the (TE). two using that in echo-time differ acquire EPI,We acquire the data using two EPI readouts that differ in echo-time (TE).,34556
using data crops pixel-precise are The of augmentation. and single instances annotations for allow articial,The annotations are pixel-precise and allow using crops of single instances for articial data augmentation.,34557
as are details is mathematical concerned. However far as surprise no there,However there is no surprise as far as mathematical details are concerned.,34558
and that Are evolution? laws quantitative constrain there govern,Are there quantitative laws that govern and constrain evolution?,34559
"based of classifying species, and typically involved morphology. study evolution comparing on the Historically,","Historically, the study of evolution involved classifying and comparing species, typically based on morphology.",34560
evolutionary predictions of can dynamics. of be diversity to models This quantitative compared,This diversity can be compared to predictions of quantitative models of evolutionary dynamics.,34561
"environments. and evolution In physical variable biological shaped these is by all often systems,","In all these systems, evolution is shaped by often variable biological and physical environments.",34562
an train image simultaneously more can in two on unsupervised network than datasets manner. Our,Our network can simultaneously train on more than two image datasets in an unsupervised manner.,34563
radiology. a Positron (PET) of modern tomography cornerstone emission is,Positron emission tomography (PET) is a cornerstone of modern radiology.,34564
synthesized Perceptual for quality task. a is challenging textures assessment,Perceptual quality assessment for synthesized textures is a challenging task.,34565
"a knowledge, domain experts. require queries no domain priori and no human hence These","These queries require no a priori domain knowledge, and hence no human domain experts.",34566
naturally. also to proposed techniques be knowledge The ingested extended allow,The proposed techniques also allow ingested knowledge to be extended naturally.,34567
"are of techniques. early, and paper Our description investigations provides this the a","Our investigations are early, and this paper provides a description of the techniques.",34568
Assessment future next is of our step for work. efficacy their,Assessment of their efficacy is our next step for future work.,34569
as to and artifacts leads color blurriness. often such discrepancy This,This often leads to artifacts such as color discrepancy and blurriness.,34570
"but such Post-processing are expensive is may used artifacts, usually and reduce to fail.","Post-processing is usually used to reduce such artifacts, but are expensive and may fail.",34571
Our methods outperforms for model masks. other irregular,Our model outperforms other methods for irregular masks.,34572
comparisons with qualitative other show and We our to approach. validate quantitative methods,We show qualitative and quantitative comparisons with other methods to validate our approach.,34573
harmonic in this and voltage research. sub-harmonic Third studied differential schemes are,Third harmonic differential voltage and sub-harmonic schemes are studied in this research.,34574
"leading color appearance and artifacts. inconsistent noticeable to is neglected, colors Their often","Their color appearance is often neglected, leading to inconsistent colors and noticeable artifacts.",34575
system HDR space. in We extending state-of-the-art to by colors SLAM rectify this a accumulate,We rectify this by extending a state-of-the-art SLAM system to accumulate colors in HDR space.,34576
controller. a Our is main map-aware contribution time exposure,Our main contribution is a map-aware exposure time controller.,34577
square given. models least using to is regression explanation Lord's ordinary An paradox,An explanation to Lord's paradox using ordinary least square regression models is given.,34578
features applications. learning machine Discriminative are critical for,Discriminative features are critical for machine learning applications.,34579
"labeled datasets annotations with extremely expert pixel-level collecting Nevertheless, an process. expensive especially is","Nevertheless, collecting expert labeled datasets especially with pixel-level annotations is an extremely expensive process.",34580
intelligent is to perform pivotal for agents intuition complex tasks. Physical,Physical intuition is pivotal for intelligent agents to perform complex tasks.,34581
"modularization, for ontology and versioning, forgetting, example, exchange. It ontology knowledge underpins,","It underpins, for example, ontology versioning, ontology modularization, forgetting, and knowledge exchange.",34582
the replacement ontology. means the depends What intended safe of on application,What safe replacement means depends on the intended application of the ontology.,34583
"others. psychology, fields, has studied signal among processing Personality been widely neuropsychology, in analysis and","Personality analysis has been widely studied in psychology, neuropsychology, and signal processing fields, among others.",34584
"the research also visual an years, attractive in past area it From computing. became few","From the past few years, it also became an attractive research area in visual computing.",34585
are the and of Future identified research venues in field discussed.,Future venues of research in the field are identified and discussed.,34586
"payoff models. additive some monotone equilibria randomly By may contrast, refute empirical disturbed","By contrast, some empirical equilibria may refute monotone additive randomly disturbed payoff models.",34587
performed detection for is angiodysplasia localization using and endoscopy. Gold-standard wireless capsule,Gold-standard for angiodysplasia detection and localization is performed using wireless capsule endoscopy.,34588
"mask. of predicted we connected each analyze component Then,","Then, we analyze connected component of each predicted mask.",34589
aim discriminant neural We compact network. features to beginning our at convolutional of deep learn,We aim to learn discriminant compact features at beginning of our deep convolutional neural network.,34590
collected The based on models from are accelerometer measurements. built-in data,The models are based on data collected from built-in accelerometer measurements.,34591
We human synthesis. pose computational novel address of problem the,We address the computational problem of novel human pose synthesis.,34592
"a we of desired of sequence coherent videos actions. poses, produce also Given","Given a sequence of desired poses, we also produce coherent videos of actions.",34593
deal been designed objects. specifically has The small network to with,The network has been specifically designed to deal with small objects.,34594
It at proposals uses on region multiple an architecture based generated scales.,It uses an architecture based on region proposals generated at multiple scales.,34595
detection hand is compared against It recent algorithms.,It is compared against recent hand detection algorithms.,34596
central Multiple is system (CNS). of a sclerosis (MS) demyelinating nervous disease the,Multiple sclerosis (MS) is a demyelinating disease of the central nervous system (CNS).,34597
"is due tracer. expensive to the and radioactive invasive imaging injection PET a However, of","However, PET imaging is expensive and invasive due to the injection of a radioactive tracer.",34598
railway massive has a system.  America ,  America has a massive railway system.,34599
how The works discussed of details in this implementation report. algorithm this are,The implementation details of how this algorithm works are discussed in this report.,34600
frames. highly contain redundant between Videos information,Videos contain highly redundant information between frames.,34601
"practice, close users. In assumption physically to for fails even usually however, hold, oversimplified this","In practice, however, this oversimplified assumption usually fails to hold, even for physically close users.",34602
baselines. Simulation performance existing gains over state-of-the-art the substantial reveal results,Simulation results reveal substantial performance gains over the existing state-of-the-art baselines.,34603
"are preceded priming stimuli performed prime response In stimuli. target tasks, responses toward by speeded","In response priming tasks, speeded responses are performed toward target stimuli preceded by prime stimuli.",34604
for The saliency individual classification map also visualization. of the using sensitivity was determined wavelengths,The sensitivity of individual wavelengths for classification was also determined using the saliency map visualization.,34605
"discrepancies data progress. the impede various from depth However, acquired sensors with still real noticeably","However, discrepancies with the real data acquired from various depth sensors still noticeably impede progress.",34606
leveraging gap. approach only propose a models novel bridge CAD the thus realism to We,We thus propose a novel approach leveraging only CAD models to bridge the realism gap.,34607
(RVFL) model. is randomized network well-grounded The an functional-link random theoretically vector learning and established,The random vector functional-link (RVFL) network is an established and theoretically well-grounded randomized learning model.,34608
derived algorithm expectation-maximization of An for estimation is the maximum randomized MM. likelihood our (EM),An expectation-maximization (EM) algorithm is derived for the maximum likelihood estimation of our randomized MM.,34609
"the are information. preserve parts weighted-average strategy more contour to fused lowrank by Then,","Then, the lowrank parts are fused by weighted-average strategy to preserve more contour information.",34610
the surface. the on the One friction condition major of impact coefficient is road,One major impact on the friction coefficient is the condition of the road surface.,34611
of Nine sample gestures learned classes from gestures data. were,Nine classes of gestures were learned from gestures sample data.,34612
"data, KDEs When PDFs. bona positive applied not do to fide these typical yield","When applied to positive data, these typical KDEs do not yield bona fide PDFs.",34613
We transformation call log-KDEs. obtained this KDEs via the,We call the KDEs obtained via this transformation log-KDEs.,34614
"the demonstrate logKDE. R via package, methodology We our log-KDEs","We demonstrate the log-KDEs methodology via our R package, logKDE.",34615
demonstrate to the case Real studies provided are data approach. log-KDE,Real data case studies are provided to demonstrate the log-KDE approach.,34616
intensive in Many methods noise low-light enhancement ignore images. original,Many low-light enhancement methods ignore intensive noise in original images.,34617
"As they enhance as result, noise simultaneously well. the a often","As a result, they often simultaneously enhance the noise as well.",34618
"procedures extra the by denoising adopted most ruin details. Furthermore, methods","Furthermore, extra denoising procedures adopted by most methods ruin the details.",34619
practicability of method. the our extensive Results demonstrate of and effectiveness experiments,Results of extensive experiments demonstrate the effectiveness and practicability of our method.,34620
we or of model explicit $f(\cdot)$. have Therefore no form,Therefore we have no explicit form or model of $f(\cdot)$.,34621
the of are Nor interested workings in black internal necessarily box. the we,Nor are we necessarily interested in the internal workings of the black box.,34622
parameters. performed the extrinsic with then The again refined is fusion,The fusion is then performed again with the refined extrinsic parameters.,34623
patterns. and applicability significant results driving the approach finding preliminary of for show Analysis,Analysis and preliminary results show the applicability of approach for finding significant driving patterns.,34624
We latent style then discover the to documents in these images. over the train LDA,We then train LDA over these documents to discover the latent style in the images.,34625
resulting to indicators The our topics are of visual be excellent shown across style platform.,The resulting topics are shown to be excellent indicators of visual style across our platform.,34626
"is region just of modeled Instead, one of entire the set displacement parameters. by motion","Instead, displacement of the entire region is modeled by just one set of motion parameters.",34627
"the to Therefore, the model regions. are decoder only for texture the parameters reconstructing transmitted","Therefore, only the model parameters are transmitted to the decoder for reconstructing the texture regions.",34628
the in Non-texture conventionally. are regions frame coded,Non-texture regions in the frame are coded conventionally.,34629
learning be examples. State-of-the-art algorithms fooled by carefully machine can crafted adversarial,State-of-the-art machine learning algorithms can be fooled by carefully crafted adversarial examples.,34630
"problem present AI adversarial As examples in concrete such, safety. a","As such, adversarial examples present a concrete problem in AI safety.",34631
related to image. tasks determining Many if text string an in exists a particular are,Many tasks are related to determining if a particular text string exists in an image.,34632
"called Based a Guided developed Attention. on model this framework, we","Based on this framework, we developed a model called Guided Attention.",34633
real-world text perspective task new for provides scene This new studying problems. related a,This new real-world task provides a new perspective for studying scene text related problems.,34634
performs show against that state-of-the-art our trackers. favorably Experimental results other tracker,Experimental results show that our tracker performs favorably against other state-of-the-art trackers.,34635
"classification with results. fine-grained for model than SomethingSomething, better Our impressive performs baselines existing","Our model performs better than existing classification baselines for SomethingSomething, with impressive fine-grained results.",34636
it new on yields task. strong a the baseline And captioning Something-Something,And it yields a strong baseline on the new Something-Something captioning task.,34637
state-of-the-art estimation omnidirectional flow approach optical images. we adapt for paper a In this to,In this paper we adapt a state-of-the-art approach for optical flow estimation to omnidirectional images.,34638
the truth motion ground results fields we evaluation of For synthetically. create,For evaluation of the results we create ground truth motion fields synthetically.,34639
contains truth with The static background. ground cubes,The ground truth contains cubes with static background.,34640
"baseline our provides both a performance. For strong datasets, model","For both datasets, our model provides a strong baseline performance.",34641
Bayesian critical are of quantifying for systems. methods behaviors the,Bayesian methods are critical for quantifying the behaviors of systems.,34642
"intensive. Bayesian are computationally are often While they very useful, methods","While Bayesian methods are very useful, they are often computationally intensive.",34643
necessitates more the efficient of This development algorithms.,This necessitates the development of more efficient algorithms.,34644
critical game and is task. design video levels complex a of The,The design of video game levels is a complex and critical task.,34645
fun challenge while Levels avoiding at need costs. all frustration to elicit and,Levels need to elicit fun and challenge while avoiding frustration at all costs.,34646
unknown of length. handwritten of segmentation-free paper numeral strategies strings the presents This recognition for,This paper presents segmentation-free strategies for the recognition of handwritten numeral strings of unknown length.,34647
"by the patches several source technique. sliding into window images Firstly, divided are","Firstly, the source images are divided into several patches by sliding window technique.",34648
"classified to patches of according Oriented Then, the the Gradient Histogram (HOG) are features.","Then, the patches are classified according to the Histogram of Oriented Gradient (HOG) features.",34649
each value sub-dictionaries by learned of the are algorithm. decomposition K-singular class And (K-SVD),And the sub-dictionaries of each class are learned by K-singular value decomposition (K-SVD) algorithm.,34650
"combining constructed sub-dictionaries. global by is dictionary Secondly, these a","Secondly, a global dictionary is constructed by combining these sub-dictionaries.",34651
The interactive primary of modeling our digital application for fabrication. is method,The primary application of our method is interactive modeling for digital fabrication.,34652
enables high-resolution Our technique a user interactively models. process to,Our technique enables a user to interactively process high-resolution models.,34653
fuse to And these designed layers(fusion two strategies) fusion features. are,And two fusion layers(fusion strategies) are designed to fuse these features.,34654
"is fused the image Finally, decoder. by reconstructed","Finally, the fused image is reconstructed by decoder.",34655
utilized different two is at TL stages; channel exploitation. generation channel and,TL is utilized at two different stages; channel generation and channel exploitation.,34656
"pretrained by knowledge the is Finally, a gained employing exploited CNN TL. by","Finally, the knowledge gained by a pretrained CNN is exploited by employing TL.",34657
"in However, reality are these simultaneously. seldom met requirements","However, in reality these requirements are seldom met simultaneously.",34658
performance method demonstrates all datasets. on leading Our,Our method demonstrates leading performance on all datasets.,34659
"barrier computational implementing its to The is most high technique, however, complexity. significant this","The most significant barrier to implementing this technique, however, is its high computational complexity.",34660
assumed variables main evolutionary transmission be linked through to virulence trade-offs. and broadly Pathogen are,Pathogen transmission and virulence are main evolutionary variables broadly assumed to be linked through trade-offs.,34661
"errors. and limitations future advantages, work computing Finally, with the we discuss of","Finally, we discuss the advantages, limitations and future work of computing with errors.",34662
in also Competitive widely section. performances used are on benchmarks summarized this,Competitive performances on widely used benchmarks are also summarized in this section.,34663
application scenarios. to related then additional We extend survey our issues and,We then extend our survey to additional related issues and application scenarios.,34664
"focus density evaluation. mainly we forecasts, also point Although forecast we analyze on","Although we focus mainly on point forecasts, we also analyze density forecast evaluation.",34665
methods. proposed the compared achieved state-of-the-art competitive results the method performance that show Experimental to,Experimental results show that the proposed method achieved competitive performance compared to the state-of-the-art methods.,34666
remote gaze is tracking challenging using problem. Unconstrained cameras off-the-shelf a,Unconstrained remote gaze tracking using off-the-shelf cameras is a challenging problem.,34667
"Recently, proposed. gaze have for networks appearance-based neural using convolutional (CNN) algorithms estimation been promising","Recently, promising algorithms for appearance-based gaze estimation using convolutional neural networks (CNN) have been proposed.",34668
be that recorded doubles neurons seven can years.  of  simultaneously every The number,  The number of neurons that can be simultaneously recorded doubles every seven years.,34669
observable number performance required particles exponentially certain the the a with of scales dimensions. for,the number of particles required for a certain performance scales exponentially with the observable dimensions.,34670
malaria studied like by and inspection Many micro-organisms manual expert are counting. hand parasites still,Many micro-organisms like malaria parasites are still studied by expert manual inspection and hand counting.,34671
problem task. classification is Video a learning vital representation for,Video representation learning is a vital problem for classification task.,34672
"to methods video suffer limitations classification. from regard extended when existing two this in However,","However, existing methods in this regard suffer from two limitations when extended to video classification.",34673
"and high scenarios. real-world computational hinders memory in cost their Second, application","Second, high computational and memory cost hinders their application in real-world scenarios.",34674
via between exploit existing semantic methods seen-to-unseen correlation Most projection spaces. and learning a visual,Most existing methods exploit seen-to-unseen correlation via learning a projection between visual and semantic spaces.,34675
methods to often multiagent learning are games. in Policy applied gradient continuous reinforcement,Policy gradient methods are often applied to reinforcement learning in continuous multiagent games.,34676
of The role in remains coding sensory synchronous cognition and controversial. firing,The role of synchronous firing in sensory coding and cognition remains controversial.,34677
both coherent to attempt a of We within understand lines framework. evidence theoretical,We attempt to understand both lines of evidence within a coherent theoretical framework.,34678
electrical data signal problem a parameters is using in of The recurrent engineering. estimation quantized,The estimation of signal parameters using quantized data is a recurrent problem in electrical engineering.,34679
"based is Gauss-Markov theorem. quantile-based presented, the which estimator on a In is paper, this","In this paper, a quantile-based estimator is presented, which is based on the Gauss-Markov theorem.",34680
problems lacking adaptation Domain is widely used in labels. learning,Domain adaptation is widely used in learning problems lacking labels.,34681
"train. latter former has However, ability hard is whereas poor generalization to very the the","However, the former has poor generalization ability whereas the latter is very hard to train.",34682
condition is factor estate for building a The real an valuation. important of,The condition of a building is an important factor for real estate valuation.,34683
vision. computer topic in is a tracking hot Object,Object tracking is a hot topic in computer vision.,34684
"fused image inverse is the fused with by Finally, transforms multi-scale coefficients. reconstructed","Finally, the fused image is reconstructed by inverse multi-scale transforms with fused coefficients.",34685
"synchronization However, usually base this demands between stations. accurate","However, this usually demands accurate synchronization between base stations.",34686
method called transform radiation-invariant is The (RIFT). feature proposed,The proposed method is called radiation-invariant feature transform (RIFT).,34687
"RIFT a proposes map originally (MIM) for Second, description. maximum feature index","Second, RIFT originally proposes a maximum index map (MIM) for feature description.",34688
more Experimental show SIFT SAR-SIFT. and results much superior RIFT to is that,Experimental results show that RIFT is much more superior to SIFT and SAR-SIFT.,34689
made datasets The and code multi-modal image of source . are remote public RIFT sensing,The source code of RIFT and multi-modal remote sensing image datasets are made public .,34690
"present effective a called simple automatic for We method but SegFinNet. latent fingerprint segmentation,","We present a simple but effective method for automatic latent fingerprint segmentation, called SegFinNet.",34691
three databases Experimental latent results (i.e. different on,Experimental results on three different latent databases (i.e.,34692
"Additionally, analysis in-depth and we these results provide of cases the failure methods. an of","Additionally, we provide an in-depth analysis of the results and failure cases of these methods.",34693
The baseline dataset well in will due available as as made time. results publicly be,The dataset as well as baseline results will be made publicly available in due time.,34694
dataset baseline well . The UFDD available are www.ufdd.info/ at: as results as,The UFDD dataset as well as baseline results are available at: www.ufdd.info/ .,34695
surveillance in cameras extremely across person illumination challenging. variations problem Drastic re-identification the make,Drastic variations in illumination across surveillance cameras make the person re-identification problem extremely challenging.,34696
"data for image and inefficient for volumes, are small-scale solvers large inadequate Existing registration.","Existing solvers are inadequate for large data volumes, and inefficient for small-scale image registration.",34697
"presented. accurate solver efficient is work, this In matrix-based method an and","In this work, an efficient and accurate matrix-based solver method is presented.",34698
"For data source the internet. set, augmenting more data the from we","For augmenting the data set, we source more data from the internet.",34699
the effectiveness the the The on data. approach DFW show experiments of,The experiments show the effectiveness of the approach on the DFW data.,34700
This stage. training image-level detection supervision with supervised paper addresses only at weakly object,This paper addresses weakly supervised object detection with only image-level supervision at training stage.,34701
"plan components. we the hardware sensory a provide Furthermore, low-cost system based on for construction","Furthermore, we provide a construction plan for the sensory system based on low-cost hardware components.",34702
"sensors cameras, represent perception a state-of-the-art from LiDAR Aside systems. component of central","Aside from cameras, LiDAR sensors represent a central component of state-of-the-art perception systems.",34703
is CNN LiDAR significantly LiLaNet data. state-of-the-art to outperform architectures current shown for,LiLaNet is shown to significantly outperform current state-of-the-art CNN architectures for LiDAR data.,34704
on Different and external integrated have approaches sensors force based been proposed.,Different approaches based on external and integrated force sensors have been proposed.,34705
"are sensor hampered by friction, size, sterilizability. and These","These are hampered by friction, sensor size, and sterilizability.",34706
architecture with increased We efficiency. for employ deep bottlenecks residual a,We employ a deep residual architecture with bottlenecks for increased efficiency.,34707
compare the We using difference projections. volumes approach and methods two-dimensional to Siamese,We compare the Siamese approach to methods using difference volumes and two-dimensional projections.,34708
"on the between tissue in dataset, methods shows good our different generalization Based subjects.","Based on the tissue dataset, our methods shows good generalization in between different subjects.",34709
coherence estimation using \textit{Conclusions} We tomography. method force optical a novel image-based propose,\textit{Conclusions} We propose a novel image-based force estimation method using optical coherence tomography.,34710
force We substantially capturing the that improves of subsurface estimation. structures illustrate deformation,We illustrate that capturing the deformation of subsurface structures substantially improves force estimation.,34711
well as content with as visual provide mass a Movies stories. us attracting of,Movies provide us with a mass of visual content as well as attracting stories.,34712
"sentences subtitles. Particularly, training from words and movie firstly the we extract","Particularly, we firstly extract words and sentences from the training movie subtitles.",34713
We extensive the on MovieQA dataset. conduct experiments,We conduct extensive experiments on the MovieQA dataset.,34714
"only large LMN visual inputs, With performance content representation obtains with as a frame-level improvement.","With only visual content as inputs, LMN with frame-level representation obtains a large performance improvement.",34715
the re-identification aims Person same at recognizing images different cameras. (Re-ID) from person across taken,Person re-identification (Re-ID) aims at recognizing the same person from images taken across different cameras.,34716
"introduced an these estimator Then, is overcomes that limitations.","Then, an estimator is introduced that overcomes these limitations.",34717
important an Fingertip interaction. plays in detection role computer human,Fingertip detection plays an important role in human computer interaction.,34718
binocular transform images. works into Previous depth images,Previous works transform binocular images into depth images.,34719
for and low TSBnet right shares convolutional images. level features layers of left first,TSBnet first shares convolutional layers for low level features of right and left images.,34720
networks Then extracts high separately. convolutional level it two-stream in features,Then it extracts high level features in two-stream convolutional networks separately.,34721
"that we finite natural traces argue and when reasoning Also, action are about change. more","Also, we argue that finite traces are more natural when reasoning about action and change.",34722
"two generic outline a implementations, one extension and of we clingo. an Finally,","Finally, we outline two implementations, a generic one and an extension of clingo.",34723
In-depth on are investigations to insights newly layer. provide introduced conducted the,In-depth investigations are conducted to provide insights on the newly introduced layer.,34724
generalizing to We new with architecture examples. a learning propose for network-based capsule few data,We propose a capsule network-based architecture for generalizing learning to new data with few examples.,34725
train the multiMNIST dataset the digit. We on lacking one networks,We train the networks on the multiMNIST dataset lacking one digit.,34726
paper solve model those of to issues. aims investigation anomaly-based at further This,This paper aims at further investigation of anomaly-based model to solve those issues.,34727
"enhance (MS) pan-sharpened PAN First, to Spectral data together with discrimination. Multi incorporated is ship","First, pan-sharpened Multi Spectral (MS) data is incorporated together with PAN to enhance ship discrimination.",34728
samples. the numerical scales number measured cost of linearly with Its,Its numerical cost scales linearly with the number of measured samples.,34729
"observer actor perspectives. between and the This two, the learning enables link","This enables learning the link between the two, actor and observer perspectives.",34730
the computer recognition Face investigated most in vision. extensively of is one problems (FR),Face recognition (FR) is one of the most extensively investigated problems in computer vision.,34731
photos by high-resolution professional taken of faces photo-journalists. celebrity,high-resolution photos of celebrity faces taken by professional photo-journalists.,34732
"the in images surveillance remains and challenging low-resolution under-studied. largely unconstrained more However, FR","However, the more challenging FR in unconstrained and low-resolution surveillance images remains largely under-studied.",34733
"FR benchmark. presents it As consequence, a extremely an challenging","As a consequence, it presents an extremely challenging FR benchmark.",34734
"by thousands of physicians of is time consuming. annotation However, costly and manual histopathological images","However, manual annotation of thousands of histopathological images by physicians is costly and time consuming.",34735
is consideration under for This acceptance paper TPLP. in,This paper is under consideration for acceptance in TPLP.,34736
"ConvNet specify summarizer framework, discriminator. new of and a consisting a We","We specify a new ConvNet framework, consisting of a summarizer and discriminator.",34737
investigated for fundamental topic a many has decades. as research compression Image been,Image compression has been investigated as a fundamental research topic for many decades.,34738
rapid The years. recent hippocampal interneuron has development in achieved technology characterization,The hippocampal interneuron characterization technology has achieved rapid development in recent years.,34739
"However, features characterization capturing interneuron has segmentation accurate and remained true for neuronal elusive.","However, capturing true neuronal features for accurate interneuron characterization and segmentation has remained elusive.",34740
"have the purpose, For domain this intra-logistics. chosen of robotic we","For this purpose, we have chosen the domain of robotic intra-logistics.",34741
"strategies, This etc. reasoning action, planning, includes about resources, change, multi-agent","This includes multi-agent planning, reasoning about action, change, resources, strategies, etc.",34742
"In allows to and regards as scalability. study solutions users return, effectiveness asprilo alternative","In return, asprilo allows users to study alternative solutions as regards effectiveness and scalability.",34743
teaching for attractive logic beyond and it benchmarking well programming. This makes,This makes it attractive for benchmarking and teaching well beyond logic programming.,34744
challenging  Convolutional various great vision tasks. have (ConvNets) achieved  successes networks in,  Convolutional networks (ConvNets) have achieved great successes in various challenging vision tasks.,34745
"encountering shift. domain the performance degrade would when ConvNets the of However,","However, the performance of ConvNets would degrade when encountering the domain shift.",34746
"fully based is for dilated model a on prediction. our pixel-wise Specifically, network convolutional","Specifically, our model is based on a dilated fully convolutional network for pixel-wise prediction.",34747
"image (AIR) In is an proposed. registration paper, this adversarial framework","In this paper, an adversarial image registration (AIR) framework is proposed.",34748
(VAEs) generation Variational auto-encoders problem. image an solution attractive to provide,Variational auto-encoders (VAEs) provide an attractive solution to image generation problem.,34749
"shown convolutional this networks neural performance on task. outstanding Recently, have (CNNs)","Recently, convolutional neural networks (CNNs) have shown outstanding performance on this task.",34750
issue. crucial this mitigate sensors Detection on attacks to is of,Detection of attacks on sensors is crucial to mitigate this issue.,34751
"avoided. doing, so is sensing matrix problem by the large-scale In the induced","In so doing, the problem induced by the large-scale sensing matrix is avoided.",34752
"both and no improved. computational interpolations quality can the image be the speed required, With","With no interpolations required, both the computational speed and the image quality can be improved.",34753
acceptance This for in is paper under consideration TPLP.,This paper is under consideration for acceptance in TPLP.,34754
techniques reasonably technology. reliable recognition made deep learning Contemporary a image have,Contemporary deep learning techniques have made image recognition a reasonably reliable technology.,34755
from are developed (LCs). lymphoma neoplasms (LL) Lymphoid lymphoid leukemia and cells,Lymphoid leukemia (LL) and lymphoma are neoplasms developed from lymphoid cells (LCs).,34756
accumulate in are mild can thus and cells. MECCs PDMs mostly,PDMs and MECCs are mostly mild thus can accumulate in cells.,34757
contribute of to cell Some the PDMs/MECCs transformation.,Some of the PDMs/MECCs contribute to cell transformation.,34758
drive can transformation. cell affects more one or alone GECC A genes and,A GECC affects one or more genes and can alone drive cell transformation.,34759
and participates more transformation. genes affects in An one cell IECC or,An IECC affects one or more genes and participates in cell transformation.,34760
"be more rapidly CC. transformed LC a can Hence, a by","Hence, a LC can be more rapidly transformed by a CC.",34761
is by pathway of Slow accumulation driven PDMs/MECCs.,Slow pathway is driven by accumulation of PDMs/MECCs.,34762
"GECC by step"". in is a driven pathway ""one Rapid","Rapid pathway is driven by a GECC in ""one step"".",34763
PDMs/MECCs/IECC(s). Accelerated of is by accumulation pathway driven,Accelerated pathway is driven by accumulation of PDMs/MECCs/IECC(s).,34764
occur at different a transformations LC via pathways different ages. Cell of,Cell transformations of a LC via different pathways occur at different ages.,34765
transformation A occurs slow adults. mainly pathway via in,A transformation via slow pathway occurs mainly in adults.,34766
"source, serve important This an overlays can why meta-data of as those is extraction e.g.","This is why extraction of those overlays can serve as an important meta-data source, e.g.",34767
"on solution extraction, detection and The text several recognition. proposed keyframe processing relies text steps:","The proposed solution relies on several processing steps: keyframe extraction, text detection and text recognition.",34768
"The component i.e. main our system, of","The main component of our system, i.e.",34769
"CRAM task. the two tackles performance, vision ensure To computer","To ensure the performance, CRAM tackles two computer vision task.",34770
"presented methods terms algorithms, are of in motivations, experiment and results. These","These methods are presented in terms of motivations, algorithms, and experiment results.",34771
light-field is datasets. Temporal of achieved a variety coherence on,Temporal coherence is achieved on a variety of light-field datasets.,34772
and of systems principal production storage are and Automated facilities. retrieval modern components warehouse,Automated storage and retrieval systems are principal components of modern production and warehouse facilities.,34773
is new introduced unit temporal dilated to (DTR) relation capturing. representation temporal enhance A,A new dilated temporal relation (DTR) unit is introduced to enhance temporal representation capturing.,34774
the of public the experiments approach. on show Comprehensive datasets three proposed effectiveness,Comprehensive experiments on three public datasets show the effectiveness of the proposed approach.,34775
"graph-based propose local fast the address a To problem, we algorithm. denoising cloud point","To address the point cloud denoising problem, we propose a fast graph-based local algorithm.",34776
alternately are partites The until of nodes optimized two convergence.,The two partites of nodes are alternately optimized until convergence.,34777
transformations and perspectives. illumination from The different limited different to is levels affine domain problem,The problem domain is limited to different illumination levels and affine transformations from different perspectives.,34778
competitive single can be tasks a CNN-architecture across show We with used that results.,We show that a single CNN-architecture can be used across tasks with competitive results.,34779
local three is framework The based on steps. learning,The local learning framework is based on three steps.,34780
"Vector Support the trained (SVM) Second, training on is a one-versus-all selected samples. classifier Machines","Second, a one-versus-all Support Vector Machines (SVM) classifier is trained on the selected training samples.",34781
"on EOPG results problem, to several present the algorithm first the We solve including it.","We present several results on the EOPG problem, including the first algorithm to solve it.",34782
fundamental is statistical of the Probabilistic analysis to data. complex modeling,Probabilistic modeling is fundamental to the statistical analysis of complex data.,34783
attention Understanding multimedia is visual human important for applications.,Understanding human visual attention is important for multimedia applications.,34784
from studies models. saliency and attempted data learn build computational have Many eye-tracking prediction to,Many studies have attempted to learn from eye-tracking data and build computational saliency prediction models.,34785
"generative deep end-to-end for paper, saliency novel a In pages. this model Web propose we","In this paper, we propose a novel end-to-end deep generative saliency model for Web pages.",34786
size the of data and linearly training w.r.t.,the size of training data and linearly w.r.t.,34787
improvements over two on previous the We filling. exemplar-based hole propose work,We propose two improvements over the previous work on exemplar-based hole filling.,34788
new Tracking Echocardiography function evaluation. for relatively a cardiac is (STE) method Speckle,Speckle Tracking Echocardiography (STE) is a relatively new method for cardiac function evaluation.,34789
And by labels. above detection is the the model face trained,And the face detection model is trained by the above labels.,34790
"are second IoU anchors not However, and between threshold threshold first used. with","However, anchors with IoU between first threshold and second threshold are not used.",34791
"Precise models. a training Score(PBS), detection strategy, train to We Box novel propose object","We propose a novel training strategy, Precise Box Score(PBS), to train object detection models.",34792
"training Our making strategy better more from proposed information of extracts existing utilization datasets, datasets.","Our proposed training strategy extracts more information from datasets, making better utilization of existing datasets.",34793
"problem. respiratory to Consequently, extraction of features linear simplifies a","Consequently, extraction of respiratory features simplifies to a linear problem.",34794
and the Idling through pollute environment waste vehicles emission. exhaust energy,Idling vehicles waste energy and pollute the environment through exhaust emission.,34795
collected the We image first detection. car for IR idling dataset sequence,We collected the first IR image sequence dataset for idling car detection.,34796
"detected cars Then, we track parked. are over time the to the cars that identify","Then, we track the detected cars over time to identify the cars that are parked.",34797
results IR dataset. experimental We present promising sequence on our image,We present promising experimental results on our IR image sequence dataset.,34798
"visual. domains fashion primarily as However, such are certain","However, certain domains such as fashion are primarily visual.",34799
"neural two this problems of benchmark study networks. challenging we identification using paper, the In","In this paper, we study the identification of two challenging benchmark problems using neural networks.",34800
hypothesize mechanism. segregates telosome We a that the conservative by,We hypothesize that the telosome segregates by a conservative mechanism.,34801
chromatin of data suggest an active These segregation yeast. the in presence for mechanism,These data suggest the presence of an active mechanism for chromatin segregation in yeast.,34802
method refinement called Predictors. efficient We usinga Linear present an learning-based approach subpixel,We present an efficient subpixel refinement method usinga learning-based approach called Linear Predictors.,34803
paper. this shown Two are key in ideas,Two key ideas are shown in this paper.,34804
experiments. method efficiency extensive show We and our of the accuracy through,We show the efficiency and accuracy of our method through extensive experiments.,34805
of the provides version a optimParallel of package optim(). The parallel R methods optimization gradient-based,The R package optimParallel provides a parallel version of the gradient-based optimization methods of optim().,34806
can significantly reduce optimParallel() Using optimization times.,Using optimParallel() can significantly reduce optimization times.,34807
and all is fundamental behaviour. to biological activity physical Energetic processes drives causally,Energetic activity is fundamental to all physical processes and causally drives biological behaviour.,34808
"present Finally, simulations our supporting we results findings.","Finally, we present simulations results supporting our findings.",34809
been efficient interpretable as controllers for Fuzzy industry and decades controllers. used system have in,Fuzzy controllers have been used in industry for decades as interpretable and efficient system controllers.,34810
modeling Adversarial for a We human motion (GANs). Networks generative model using present Generative novel,We present a novel generative model for human motion modeling using Generative Adversarial Networks (GANs).,34811
is of model Our capable completion. and generation motion,Our model is capable of motion generation and completion.,34812
class the call the shared class.} We \doublequote{known,We call the shared class the \doublequote{known class.},34813
allows samples. unknown extracting target samples separate This approach that known features from target,This approach allows extracting features that separate unknown target samples from known target samples.,34814
"each task, separation for the dataset enables utilize to right This i.e. us","This separation enables us to utilize each dataset for the right task, i.e.",34815
"data requires algorithms labeled for large, applications Training sets. learning speech training machine","Training machine learning algorithms for speech applications requires large, labeled training data sets.",34816
of We evaluate using both efficacy the criteria. and our approach objective subjective,We evaluate the efficacy of our approach using both objective and subjective criteria.,34817
image a memory-efficient deformable We present computational approach fast and novel to registration.,We present a novel computational approach to fast and memory-efficient deformable image registration.,34818
"evaluated real-world the datasets. we on medical approach Furthermore,","Furthermore, we evaluated the approach on real-world medical datasets.",34819
learning. is a Autonomous lifestyle greatly driving practical future driven deep by becoming,Autonomous driving is becoming a future practical lifestyle greatly driven by deep learning.,34820
"detection effective sign plays for critical a Specifically, deep learning an it. traffic by role","Specifically, an effective traffic sign detection by deep learning plays a critical role for it.",34821
"sensor  remote provide can or In  information. reinforcing each complementary sensing,","  In remote sensing, each sensor can provide complementary or reinforcing information.",34822
valuable to overall boost from It is sensors outputs performance. multiple to fuse,It is valuable to fuse outputs from multiple sensors to boost overall performance.,34823
pixel require methods the data. fusion each often Previous for accurate training in labels supervised,Previous supervised fusion methods often require accurate labels for each pixel in the training data.,34824
"difficult infeasible obtain. to in or are pixel-level labels However, remote many applications, sensing","However, in many remote sensing applications, pixel-level labels are difficult or infeasible to obtain.",34825
"modalities. outputs In addition, multiple often different have resolution from sensors or","In addition, outputs from multiple sensors often have different resolution or modalities.",34826
"multi-resolution difficult to fuse is such multi-modal, directly It data. often","It is often difficult to directly fuse such multi-modal, multi-resolution data.",34827
an the because is differ and challenging photo task domains This sketch extremely significantly.,This is an extremely challenging task because the photo and sketch domains differ significantly.,34828
fingerprint a localization alterations. detection paper for proposes of This method and,This paper proposes a method for detection and localization of fingerprint alterations.,34829
the can of fingerprint research. images alleviate successfully GAN availability for limited A altered trained,A successfully trained GAN can alleviate the limited availability of altered fingerprint images for research.,34830
synthetically dataset The open-sourced. fingerprint generated altered be will,The synthetically generated altered fingerprint dataset will be open-sourced.,34831
"of sets detection, object naturally entities. that are as outputs expressed have","object detection, have outputs that are naturally expressed as sets of entities.",34832
for (HRS) satellite presented experiment high-resolution eye an images. is study tracking The,The presented study is an eye tracking experiment for high-resolution satellite (HRS) images.,34833
values Line kernel density estimation. is based using on fitting comparison color,Line fitting is based on color values comparison using kernel density estimation.,34834
match. needed reduction of to It a the depth computations find best significant provides,It provides a significant reduction of computations needed to find the best depth match.,34835
analysis several our tasks on of We temporal expression. evaluate to facial related method all,We evaluate our method on several tasks all related to temporal analysis of facial expression.,34836
based improved We support-agnostic grouping implement collaborative intensity-invariance sparse reconstruction. for,We implement intensity-invariance based improved grouping for collaborative support-agnostic sparse reconstruction.,34837
"correlation patches stack similar-structured firstly, we collaboration via For intensity-invariant measure.","For collaboration firstly, we stack similar-structured patches via intensity-invariant correlation measure.",34838
filtering. collaborate patches noise for yield grouped to sparse desirable estimates The,The grouped patches collaborate to yield desirable sparse estimates for noise filtering.,34839
method also for color extend We images. proposed denoising our of,We also extend our proposed method for denoising of color images.,34840
"tomography. this Kalman dimension dynamic X-ray consider we prior-based reduction for In filter paper, undersampled","In this paper, we consider prior-based dimension reduction Kalman filter for undersampled dynamic X-ray tomography.",34841
"method, parameterized With basis. are a low-dimensional X-ray by the this reconstructions","With this method, the X-ray reconstructions are parameterized by a low-dimensional basis.",34842
propose and identity embedding joint and of sensitive We face voice. investigate an,We propose and investigate an identity sensitive joint embedding of face and voice.,34843
face to cross-modal voice. and face voice from an to retrieval embedding enables Such from,Such an embedding enables cross-modal retrieval from voice to face and from face to voice.,34844
social recognition their is to with environment. robots for a Action meaningfully engage critical task,Action recognition is a critical task for social robots to meaningfully engage with their environment.,34845
"by primitive the joints. trajectory are the corresponding of described actions Then, vectors","Then, the primitive actions are described by the trajectory vectors of corresponding joints.",34846
by simulations. discussed this are The of analyzed properties at estimator and first,The properties of this estimator are discussed and analyzed at first by simulations.,34847
of location work invariance. the the approaches existing on of Most assumption,Most of the existing approaches work on the assumption of location invariance.,34848
approaches perform These object-centric to matching. require images,These approaches require object-centric images to perform matching.,34849
"transformation. into complete matching is segmentation and sub-objectives: divided three localization, objective The object","The complete objective is divided into three sub-objectives: object localization, segmentation and matching transformation.",34850
localization between establishes Object images. initial correspondence step an the,Object localization step establishes an initial correspondence between the images.,34851
modified of detector version shot multi-box purpose. for single A used this is,A modified version of single shot multi-box detector is used for this purpose.,34852
make images to detected object-centric. cropped is The region the,The detected region is cropped to make the images object-centric.,34853
estimating are pivotal when physiologic Uncertainty measurements parameters.,Uncertainty measurements are pivotal when estimating physiologic parameters.,34854
standard by the in PSM is here the The uncertainty data percent represented drift.,The standard uncertainty in the PSM data is here represented by the percent drift.,34855
"operation it best which However, results. clear leads to is not","However, it is not clear which operation leads to best results.",34856
connected of a backbone FC-DRN composed residual networks. densely has,FC-DRN has a densely connected backbone composed of residual networks.,34857
proven has Depth for useful information been recognition. face,Depth information has been proven useful for face recognition.,34858
of for State-of-the-art visual rely supervised a models tasks wide on perception range pretraining.,State-of-the-art visual perception models for a wide range of tasks rely on supervised pretraining.,34859
these the pretraining task is facto de classification models. for ImageNet,ImageNet classification is the de facto pretraining task for these models.,34860
"and by standards old is nearly ImageNet is years ""small"". ten Yet, now modern","Yet, ImageNet is now nearly ten years old and is by modern standards ""small"".",34861
collect reasons are and to difficult obvious: such are datasets The annotate.,The reasons are obvious: such datasets are difficult to collect and annotate.,34862
demonstrate to Our results. experiments leads excellent that prediction hashtag large-scale training for,Our experiments demonstrate that training for large-scale hashtag prediction leads to excellent results.,34863
practice are optimization potentials also discussed. for in further The,The potentials for further optimization in practice are also discussed.,34864
standard and on tested model our datasets by face We annotated two attributes. identities,We tested our model on two standard datasets annotated by identities and face attributes.,34865
(GANs) have translation Image-to-image been with and investigated Generative tasks widely Adversarial dual Networks learning.,Image-to-image translation tasks have been widely investigated with Generative Adversarial Networks (GANs) and dual learning.,34866
on based GANs problem We and tackle dual unpaired with data this learning.,We tackle this problem with unpaired data based on GANs and dual learning.,34867
results The proposed our the demonstrate effectiveness method. of,The results demonstrate the effectiveness of our proposed method.,34868
is analytics tracking of visual cornerstone many the systems. Object,Object tracking is the cornerstone of many visual analytics systems.,34869
tracking is compared on dataset. proposed approach The with trackers several well-known OTB the recent,The proposed approach is compared with several well-known recent trackers on the OTB tracking dataset.,34870
predict novel present from framework to We human that anatomy surface. a learns body,We present a novel framework that learns to predict human anatomy from body surface.,34871
"Furthermore, predicted to is response stimulation calculated varied stochastically. strength","Furthermore, predicted response to varied stimulation strength is calculated stochastically.",34872
a static function AI regard researchers memory. without as Most,Most researchers regard AI as a static function without memory.,34873
"ask we am I? can ourselves: we memory, have ""Where When","When we have memory, we can ask ourselves: ""Where am I?",34874
more to their Our illustrate that prone higher findings despite accuracy. ResNets are overfitting,Our findings illustrate that ResNets are more prone to overfitting despite their higher accuracy.,34875
considered one be just complementary a as should later This resource.,This later one should be considered just as a complementary resource.,34876
Human years. progress witnessed in detection has impressive recent,Human detection has witnessed impressive progress in recent years.,34877
"is However, in the environments crowded solved. human far from detecting issue highly occlusion of","However, the occlusion issue of detecting human in highly crowded environments is far from solved.",34878
"scenarios current in To still detection benchmarks. worse, human matters are make crowd under-represented","To make matters worse, crowd scenarios are still under-represented in current human detection benchmarks.",34879
"and is contains dataset The diversity. rich-annotated high large, CrowdHuman","The CrowdHuman dataset is large, rich-annotated and contains high diversity.",34880
frameworks state-of-the-art is of on CrowdHuman presented. detection Baseline performance,Baseline performance of state-of-the-art detection frameworks on CrowdHuman is presented.,34881
DeepDIVA the uses is Python implemented PyTorch. framework and deep in learning,DeepDIVA is implemented in Python and uses the deep learning framework PyTorch.,34882
"completely source, as through It accessible open is DIVAServices. Service and Web","It is completely open source, and accessible as Web Service through DIVAServices.",34883
physics Training provided is data simulator. by a,Training data is provided by a physics simulator.,34884
in partial The network fast to used applications views. real-time be from is enough,The network is fast enough to be used in real-time applications from partial views.,34885
properties the and of viability Experimental the architecture. proposed show the results generalisation,Experimental results show the viability and the generalisation properties of the proposed architecture.,34886
motion is in classic problem Human a and graphics. modeling vision computer,Human motion modeling is a classic problem in computer vision and graphics.,34887
is the at website. available project Our code,Our code is available at the project website.,34888
images many MRI Machine in learning results produce state-of-the-art segmentation. models,Machine learning models produce state-of-the-art results in many MRI images segmentation.,34889
consuming work. very This costs is time labeling experts process and,This labeling process is very time consuming and costs experts work.,34890
reduce high is this to finding a Therefore on way cost demand.,Therefore finding a way to reduce this cost is on high demand.,34891
model the datum again. result train another is used Inference to,Inference result is another datum used to train the model again.,34892
ventricle We on and our very results promising. segmentation method human left applied are,We applied our method on human left ventricle segmentation and results are very promising.,34893
to This method used also labels. be can generate,This method can also be used to generate labels.,34894
A is coordinate maximum convenient computationally algorithm likelihood developed for descent-type estimation.,A computationally convenient coordinate descent-type algorithm is developed for maximum likelihood estimation.,34895
"to main to objective, a and this however, task. domain model transfer is The target","The main objective, however, is to transfer this model to a target domain and task.",34896
model is sub-class A representing queuing proposed first management scheme with service the multi-class introduced.,A queuing model representing the proposed multi-class management scheme with sub-class service is first introduced.,34897
are city a given system of zone the The conditions then stability in derived.,The stability conditions of the system in a given city zone are then derived.,34898
frontier pushing recognition have over years. of past the been CNNs Deep visual,Deep CNNs have been pushing the frontier of visual recognition over past years.,34899
We learn loss representations. more spatial propose diversity structured activation to face a,We propose a spatial activation diversity loss to learn more structured face representations.,34900
"training and reduces the testing computational addition, In complexity. our greatly method","In addition, our method greatly reduces the training and testing computational complexity.",34901
and cues both exhibit in rich appearance gender Human behavior.,Human exhibit rich gender cues in both appearance and behavior.,34902
probabilistic BC+. We action language present of a extension,We present a probabilistic extension of action language BC+.,34903
attribute coupled classification. three to tightly sub-networks are facial The effective perform,The three sub-networks are tightly coupled to perform effective facial attribute classification.,34904
"fed subnetworks layer classifiers the oc-SVM are the representations voxel-level. middle at Finally, of to","Finally, the middle layer representations of the subnetworks are fed to oc-SVM classifiers at voxel-level.",34905
should be give described. which estimators for We the recommendations estimands used,We give recommendations which estimators should be used for the estimands described.,34906
"describe data solutions. AE sketch with possible of and Finally, problems meta-analyses we","Finally, we describe problems with meta-analyses of AE data and sketch possible solutions.",34907
aim to proposed large text conceptual methods representations Several space collections. learn from recently,Several recently proposed methods aim to learn conceptual space representations from large text collections.,34908
released. is source The of code MC-GAN,The source code of MC-GAN is released.,34909
selects model a The dynamics. immune-tumour of concept rich,The model selects a rich concept of immune-tumour dynamics.,34910
focus is autonomous of The driving accuracy. research to improve driving primary,The primary focus of autonomous driving research is to improve driving accuracy.,34911
"been progress algorithms While great at fail times. made, state-of-the-art has still","While great progress has been made, state-of-the-art algorithms still fail at times.",34912
is that as foresee cars ahead early therefore problems as possible. It important automated,It therefore is important that automated cars foresee problems ahead as early as possible.,34913
take if of driver paramount the to is importance will be This over. asked also,This is also of paramount importance if the driver will be asked to take over.,34914
failures that randomly. occur do We conjecture not,We conjecture that failures do not occur randomly.,34915
"learn to a presents occurrence predict these the of to i.e. This work failures, method","This work presents a method to learn to predict the occurrence of these failures, i.e.",34916
over developed and datasets. trained real driving driving model A camera-based is,A camera-based driving model is developed and trained over real driving datasets.,34917
predicted. show be and can score failure learned indeed Experimental that the results,Experimental results show that the failure score can indeed be learned and predicted.,34918
"First, for encoding cell novel view a laser into bird's projected projection. information is eye","First, laser information is projected into a novel cell encoding for bird's eye view projection.",34919
on the KITTI framework proposed state-of-the-art achieves that dataset Experiments among comparable results methods. show,Experiments on KITTI dataset show that the proposed framework achieves state-of-the-art results among comparable methods.,34920
"are frequency known these an particular, only functions uncertainty class. in response to lie In","In particular, these frequency response functions are only known to lie in an uncertainty class.",34921
"Finally, demonstrate of are effectiveness to method. several proposed the presented the numerical results","Finally, several numerical results are presented to demonstrate the effectiveness of the proposed method.",34922
to by goal means is group The items identify of all testing.,The goal is to identify all items by means of group testing.,34923
is the (hereafter problem GGTP). This testing generalized group,This is the generalized group testing problem (hereafter GGTP).,34924
also short provide survey a We GGTP. of,We also provide a short survey of GGTP.,34925
"printing has on proposed little time. Furthermore, effect the technique","Furthermore, the proposed technique has little effect on printing time.",34926
FDM printer on show a Experiments dual-nozzle results. promising,Experiments on a dual-nozzle FDM printer show promising results.,34927
calibrate the work tone. is to Future perceived required,Future work is required to calibrate the perceived tone.,34928
hypotheses. improves strategy of updating proposed further the quality Then the model,Then the proposed updating strategy further improves the quality of model hypotheses.,34929
"domain applied adaptation generative with Therefore, networks. generation study image to we adversarial","Therefore, we study domain adaptation applied to image generation with generative adversarial networks.",34930
for are determining also forgery important These localization. improvements,These improvements are also important for determining forgery localization.,34931
"we investigate this of to our To first approach. the are best the knowledge,","To the best of our knowledge, we are the first to investigate this approach.",34932
"mixtures can in most regarded practice, be as However, multiple datasets of domains.","However, in practice, most datasets can be regarded as mixtures of multiple domains.",34933
our of results Experimental evaluation protocol. the show proposed validity,Experimental results show the validity of our proposed evaluation protocol.,34934
Boolean can using The networks. gene be in cells well-approximated dynamics living of expression complex,The complex dynamics of gene expression in living cells can be well-approximated using Boolean networks.,34935
"interest of results subset a in only points true identified However, initially correspondences (inliers). the","However, only a subset of the initially identified interest points results in true correspondences (inliers).",34936
"pre-processing. the requires learned to detectors, amount In of data previous least method comparison our","In comparison to previous learned detectors, our method requires the least amount of data pre-processing.",34937
performance state-of-the-art allows that these competitive tasks. on We PAG demonstrate or,We demonstrate that PAG allows competitive or state-of-the-art performance on these tasks.,34938
investigated currently a disease. Electrical neurological for stimulation is brain therapy as being,Electrical brain stimulation is currently being investigated as a therapy for neurological disease.,34939
a that often separates representation rotation with Deep learning pose and parameterise methods a translation.,Deep learning methods often parameterise a pose with a representation that separates rotation and translation.,34940
"pose a Riemannian paper, problem. general formulation propose we this of estimation In the","In this paper, we propose a general Riemannian formulation of the pose estimation problem.",34941
"experience fashion. in allow the acquire agent First, self-supervised a we to","First, we allow the agent to acquire experience in a self-supervised fashion.",34942
schemes. to experimentally is state-of-the-art favorably contemporary shown The proposed scheme with compare,The proposed scheme is experimentally shown to compare favorably with contemporary state-of-the-art schemes.,34943
"significant attracted attention attribute community. vision (FAC) Recently, in the computer facial has classification","Recently, facial attribute classification (FAC) has attracted significant attention in the computer vision community.",34944
with been of challenging along has Great the made progress datasets. availability FAC,Great progress has been made along with the availability of challenging FAC datasets.,34945
flow. natural hypothesis these optic reflect that We the statistics specificities investigate the input of,We investigate the hypothesis that these specificities reflect the input statistics of natural optic flow.,34946
graphics were self-motion motion generated simulating computer in underwater Realistic using scene. sequences an,Realistic motion sequences were generated using computer graphics simulating self-motion in an underwater scene.,34947
component of important is initialisation tracking. object an Model,Model initialisation is an important component of object tracking.,34948
"scarce data annotations. hand, the is on often short and imaging other On medical","On the other hand, medical imaging data is often scarce and short on annotations.",34949
unified second is one recognition. The and micro both a approach macro expression for,The second one is a unified approach for both macro and micro expression recognition.,34950
"propagation. multipath require Novel real-time applications due channel industrial complex wideband, wireless characterization to","Novel industrial wireless applications require wideband, real-time channel characterization due to complex multipath propagation.",34951
"highly channels radio Additionally, measurements. flexible inhomogeneous demand","Additionally, inhomogeneous radio channels demand highly flexible measurements.",34952
"a we utilizing a software-defined paper, sounding architecture. this correlative approach propose channel In","In this paper, we propose a correlative channel sounding approach utilizing a software-defined architecture.",34953
"fast real-time, The with active immune wide-band enables measurements approach to time interference. variance","The approach enables real-time, wide-band measurements with fast time variance immune to active interference.",34954
validated The with a is example. demanding industrial desired performance application,The desired performance is validated with a demanding industrial application example.,34955
leverages is in many areas complementarity the different that of approaches It between AI. present,It leverages the complementarity between different approaches that is present in many areas of AI.,34956
are activities an integral day our part of life. day to Sports,Sports activities are an integral part of our day to day life.,34957
Label plays an in embedding role important applications. real-world many,Label embedding plays an important role in many real-world applications.,34958
"To enhance embeddings, contexts adopted. the captured multiple label relatedness can the by be","To enhance the label relatedness captured by the embeddings, multiple contexts can be adopted.",34959
derived effective solve alternative sparse factorization is problem. An to further algorithm the matrix optimization,An effective alternative optimization algorithm is further derived to solve the sparse matrix factorization problem.,34960
event. Intelligence development General of Artificial to The a major (AGI) promises be,The development of Artificial General Intelligence (AGI) promises to be a major event.,34961
identified. problems A number been safety significant for of have AGI,A significant number of safety problems for AGI have been identified.,34962
"list survey on recent and these, research We them. solving","We list these, and survey recent research on solving them.",34963
"current Finally, policy review public the on we AGI.","Finally, we review the current public policy on AGI.",34964
to where well model Our training areas far generalizes from was away available. data,Our model generalizes well to areas far away from where training data was available.,34965
Recent the neural (CNN). object by networks driven improvements success in detection are convolutional of,Recent improvements in object detection are driven by the success of convolutional neural networks (CNN).,34966
able learn rich hand-crafted features to They outperforming features. are,They are able to learn rich features outperforming hand-crafted features.,34967
and classification proposals a SSD CNN. using creation performs object single,SSD performs object proposals creation and classification using a single CNN.,34968
Traffic evaluation an DriveU the extensive on present Dataset (DTLD). We Light,We present an extensive evaluation on the DriveU Traffic Light Dataset (DTLD).,34969
"both, and high rates. false low We reach positive accuracy","We reach both, high accuracy and low false positive rates.",34970
size modern labeled to learning easy data approaches deep limited renders of The ordinal overfit.,The limited size of labeled ordinal data renders modern deep learning approaches easy to overfit.,34971
"and evaluated matcher image datasets. recognition has proposed with recognition character been The face recognition,","The proposed matcher has been evaluated with image recognition, character recognition and face recognition datasets.",34972
"been decision have unaddressed. Somewhat foundations processes surprisingly, for the such sequential","Somewhat surprisingly, the foundations for such sequential decision processes have been unaddressed.",34973
"Q-learning RL From sets show with an we sound. sampled action that perspective, is","From an RL perspective, we show that Q-learning with sampled action sets is sound.",34974
is temporally a The of result clusters interleaved set separated of points. by sub-sequences unclustered,The result is a set of temporally separated clusters interleaved by sub-sequences of unclustered points.,34975
ground points of trajectories movement. labeled consists The of animal set a truth of simulating,The ground truth consists of a set of trajectories of labeled points simulating animal movement.,34976
safety Advanced significantly. Driver improve (ADAS) Systems Assistance driving,Advanced Driver Assistance Systems (ADAS) improve driving safety significantly.,34977
They a when drivers alert appears. traffic unsafe dangerous conditions from maneuver,They alert drivers from unsafe traffic conditions when a dangerous maneuver appears.,34978
Traditional data-driven maneuvers mostly alone. on based models are methods driving predict to,Traditional methods to predict driving maneuvers are mostly based on data-driven models alone.,34979
channel based the expectation-maximization (EM) algorithm. We estimation propose on,We propose the channel estimation based on expectation-maximization (EM) algorithm.,34980
"we algorithms joint decoding. and (BCJR) for adopt Furthermore, Bahl-Cocke-Jelinek-Raviv multi-user iterative Viterbi","Furthermore, we adopt Viterbi and Bahl-Cocke-Jelinek-Raviv (BCJR) algorithms for joint iterative multi-user decoding.",34981
"and Numerical performance illustrate experimental of proposed results joint and channel detection estimation, the decoding.","Numerical and experimental results illustrate the performance of proposed channel estimation, joint detection and decoding.",34982
extractor (e.g. of system person parts: feature consists an reID A three typical video-based image-level,A typical video-based person reID system consists of three parts: an image-level feature extractor (e.g.,34983
"a modeling CNN), aggregate and temporal loss temporal function. features to a method","CNN), a temporal modeling method to aggregate temporal features and a loss function.",34984
"desmoplasia, most an fibrous-like stroma, encompassing the PDAC tumor of mass. includes anisotropic","PDAC includes an anisotropic fibrous-like stroma, desmoplasia, encompassing most of the tumor mass.",34985
"CDMs. produced cancer-associated is by and CAFs, fibroblasts, extracellular cell-derived Desmoplasia their matrices,","Desmoplasia is produced by cancer-associated fibroblasts, CAFs, and their cell-derived extracellular matrices, CDMs.",34986
to tumor-suppressive render a architecture CDMs of manipulated microenvironment. the suggested be can Results that,Results suggested that the architecture of CDMs can be manipulated to render a tumor-suppressive microenvironment.,34987
We of also unsupervised an adaptation. applying domain to approach RegCGAN introduce,We also introduce an approach of applying RegCGAN to unsupervised domain adaptation.,34988
available. We code make will publicly our,We will make our code publicly available.,34989
Many modified been of are authentic. these not and images have,Many of these images have been modified and are not authentic.,34990
"that the of sparse the is weighted we atoms. dictionary Further, a prove combination KL-center","Further, we prove that the weighted KL-center is a sparse combination of the dictionary atoms.",34991
dictionary-based problems Vision. involving variety present of We Computer classification reconstruction and experiments in a,We present experiments involving a variety of dictionary-based reconstruction and classification problems in Computer Vision.,34992
"However, studying received their and properties little has attention. labels very","However, studying labels and their properties has received very little attention.",34993
machine Current several algorithms. principles to learning paradigms and challenges of labeling impose,Current principles and paradigms of labeling impose several challenges to machine learning algorithms.,34994
"ambiguous, and Labels often incomplete, redundant. are","Labels are often incomplete, ambiguous, and redundant.",34995
We a gain models. using refined show wide significant labels of across range,We show significant gain using refined labels across a wide range of models.,34996
We methods: quantization. acceleration and types mimic two utilize of,We utilize two types of acceleration methods: mimic and quantization.,34997
of by transfering student network from improves the a network. knowledge a performance Mimic teacher,Mimic improves the performance of a student network by transfering knowledge from a teacher network.,34998
a of Quantization without full-precision quantized a converts degradation network to large performance. one,Quantization converts a full-precision network to a quantized one without large degradation of performance.,34999
"this Mimic. feature Using propose of we the quantization, Quantization","Using this feature of the quantization, we propose Quantization Mimic.",35000
"quantizes the quantized mimic a then It large network, small network. first","It first quantizes the large network, then mimic a quantized small network.",35001
is Safety pools. a aspect all critical in swimming,Safety is a critical aspect in all swimming pools.,35002
paper drowning near technique equations describes early (NEPTUNE). novel prediction a using This,This paper describes a near drowning early prediction technique using novel equations (NEPTUNE).,35003
to These by used near-drowning. be equations identify would variables the,These variables would be used by the equations to identify near-drowning.,35004
structures. waveform with OFDM enriched are multi-numerology flexible parameters,OFDM waveform parameters are enriched with flexible multi-numerology structures.,35005
a are structured manner. systems presented Research for multi-numerology in opportunities,Research opportunities for multi-numerology systems are presented in a structured manner.,35006
a segmentation high images challenging in Semantic sensing resolution fundamental is task. remote and,Semantic segmentation in high resolution remote sensing images is a fundamental and challenging task.,35007
has a stream It backward forward a stream. and,It has a forward stream and a backward stream.,35008
pre-trained fine-tuned encoders. reconstruction with and shared are subnetworks The two,The two subnetworks are pre-trained and fine-tuned with shared reconstruction encoders.,35009
The into are segmentation. outputs final the merged,The outputs are merged into the final segmentation.,35010
It the is inbreeding seems why reason avoided.,It seems the reason why inbreeding is avoided.,35011
"On suggests interesting other hand, the phenomenon this an possibility.","On the other hand, this phenomenon suggests an interesting possibility.",35012
in think more way. reasonable would be this to It,It would be more reasonable to think in this way.,35013
the mathematically this We drift. genetic studied on based of concept,We studied this mathematically based on the concept of genetic drift.,35014
of useful for mutation. as trigger suggested be a fixation that This inbreeding could,This suggested that inbreeding could be useful as a trigger for fixation of mutation.,35015
"app"" ""killer is Besides, real still missing. a","Besides, a real ""killer app"" is still missing.",35016
therefore levels support. scalable providing new of while to be it We user expect more,We therefore expect it to be more scalable while providing new levels of user support.,35017
prototype and is early this demo. implemented presented An been in has,An early prototype has been implemented and is presented in this demo.,35018
"in images. networks Recently, generative have generating (GANs) realistic adversarial shown promising performance","Recently, generative adversarial networks (GANs) have shown promising performance in generating realistic images.",35019
"focus voxels. In multi-organ network on labels a learning that we deep this paper, segmentation","In this paper, we focus on learning a deep multi-organ segmentation network that labels voxels.",35020
The BMC cells algorithm. will separated be the by watershed connected with marker-controlled,The connected cells with BMC will be separated by the marker-controlled watershed algorithm.,35021
classification the segmentation. features extracted will be The the for after,The features will be extracted for the classification after the segmentation.,35022
"The with proposed expert an manual state-of-the-art compared algorithm measurement, is relevant algorithms. and","The proposed algorithm is compared with an expert manual measurement, and state-of-the-art relevant algorithms.",35023
"and major As appearance. thus extinction, show, this our results on has effect a","As our results show, this has a major effect on extinction, and thus appearance.",35024
introduce also for prior We haze-level estimate effective valid a transmission. of,We also introduce a valid haze-level prior for effective estimate of transmission.,35025
"a introduce counting model, (GU-Net). U-Net novel Gated We named","We introduce a novel counting model, named Gated U-Net (GU-Net).",35026
"U-Net Specifically, concept propose to short-cut learnable we of enrich the with connections. architecture the","Specifically, we propose to enrich the U-Net architecture with the concept of learnable short-cut connections.",35027
"we Instead to these learn simply setting connections, connections of data. from short-cut propose","Instead of simply setting short-cut connections, we propose to learn these connections from data.",35028
"GU-Nets achieve outperform state-of-the-art architecture, base the and U-Net performance. consistently","GU-Nets consistently outperform the base U-Net architecture, and achieve state-of-the-art performance.",35029
"we Furthermore, (termed method propose the method on two-stage basic TSCORRCA). based CORRCA a","Furthermore, we propose a two-stage method based on the basic CORRCA method (termed TSCORRCA).",35030
SSVEP-based that CORRCA-based have great systems. for demonstrates study implementing potential BCI This high-performance methods,This study demonstrates that CORRCA-based methods have great potential for implementing high-performance SSVEP-based BCI systems.,35031
this Using suppress FBMC interference. we significantly intrinsic system the can,Using this system we can significantly suppress the FBMC intrinsic interference.,35032
we different In based different structures system our propose multiplexing DP-FBMC on three techniques.,In our DP-FBMC system we propose three different structures based on different multiplexing techniques.,35033
DP-FBMC simulations. We with systems compare through other,We compare DP-FBMC with other systems through simulations.,35034
an component are Programming. in Effective strategies general-purpose search Constraint important,Effective general-purpose search strategies are an important component in Constraint Programming.,35035
"a idea, namely, We using to guide search. new correlations introduce variables between","We introduce a new idea, namely, using correlations between variables to guide search.",35036
correlations measured domain propagation. constraint Variable are maintained and by changes during using,Variable correlations are measured and maintained by using domain changes during constraint propagation.,35037
"matrix, based correlation on crbs-max. heuristics We and variable propose the two crbs-sum","We propose two variable heuristics based on the correlation matrix, crbs-sum and crbs-max.",35038
"has and makes for more macro-expression, which duration it machine. shorter than human Micro-expression challenging","Micro-expression has shorter duration than macro-expression, which makes it more challenging for human and machine.",35039
future some We by micro-expressions research. conclude proposing directions advancing promising to,We conclude by proposing some promising future directions to advancing micro-expressions research.,35040
"segmentation this paper, steps. decompose In two into we","In this paper, we decompose segmentation into two steps.",35041
primitive label on map a The second step based involves protocol-specific computing segmentation. the,The second step involves computing a protocol-specific label map based on the primitive segmentation.,35042
for objects a the interactions the leverage also We grounding. as textual among guide,We also leverage the interactions among objects as a textual guide for the grounding.,35043
"in (MDLP) descriptor, Pattern feature Multichannel A novel Distributed color Local is manuscript. this proposed","A novel color feature descriptor, Multichannel Distributed Local Pattern (MDLP) is proposed in this manuscript.",35044
robust the information the computed aids MDLP arrangement. of texture extraction multi-distance by in The,The multi-distance information computed by the MDLP aids in robust extraction of the texture arrangement.,35045
"MDLP color an channel extracted for each features image. of Further, are","Further, MDLP features are extracted for each color channel of an image.",35046
results. results These with the experimental contrast,These results contrast with the experimental results.,35047
"memory attempt for the video our Second, leverage model is networks summarization first tasks. to","Second, our model is the first attempt to leverage memory networks for video summarization tasks.",35048
"experiments. perform of For evaluation, sets three we","For evaluation, we perform three sets of experiments.",35049
"VIST performance dataset. image-based in we Finally, storytelling with model's experiment domain, another our","Finally, we experiment our model's performance in another domain, with image-based storytelling VIST dataset.",35050
that on tasks. our We model state-of-the-art achieves verify performance the all,We verify that our model achieves state-of-the-art performance on all the tasks.,35051
"made great Skeleton-based problems many remain unsolved. has still but action recognition recently, progress","Skeleton-based action recognition has made great progress recently, but many problems still remain unsolved.",35052
"to model. training, clip-based loss we During incremental propose the a optimize","During training, we propose a clip-based incremental loss to optimize the model.",35053
that approach results illustrate methods. results The our much state-of-the-art achieves than better comparison,The comparison results illustrate that our approach achieves much better results than state-of-the-art methods.,35054
image consuming to consist Conventional time approaches methods. registration iterative of,Conventional approaches to image registration consist of time consuming iterative methods.,35055
method end-to-end registering images. We an multimodal for propose DL,We propose an end-to-end DL method for registering multimodal images.,35056
their and MR retinal registration. cardiac accuracy Experiments demonstrate for multimodal image,Experiments demonstrate their accuracy for multimodal retinal and cardiac MR image registration.,35057
"downlink overhead. Furthermore, for training alleviating tensor algorithm compressed the decomposition developed is a","Furthermore, a compressed tensor decomposition algorithm is developed for alleviating the downlink training overhead.",35058
the proposed showcase to are effectiveness methods. of presented Numerical the simulations,Numerical simulations are presented to showcase the effectiveness of the proposed methods.,35059
understanding in are Spatio-temporal human actions videos. crucial in contexts,Spatio-temporal contexts are crucial in understanding human actions in videos.,35060
to with proposed ontologies this way deal to Fuzzy uncertainty. offer are a,Fuzzy ontologies are proposed to offer a way to deal with this uncertainty.,35061
the fuzzy state the ontologies. describes of in art developing paper This,This paper describes the state of the art in developing fuzzy ontologies.,35062
"part is vital embedded. of valuable information is of a and Foot human, lots","Foot is a vital part of human, and lots of valuable information is embedded.",35063
contains which information is it human describes one this walking features. Plantar of pressure and,Plantar pressure is one of which contains this information and it describes human walking features.,35064
according Plantar converted pressure images standards. to can be some into simple,Plantar pressure can be converted into images according to some simple standards.,35065
"full plantar medical paper, usage. for these In we images pressure of take this advantage","In this paper, we take full advantage of these plantar pressure images for medical usage.",35066
named a neural propose network based novel We PSGAN. deep method,We propose a novel deep neural network based method named PSGAN.,35067
"Recently, accuracy. was joint largely improved in estimation by human predicting pose locations","Recently, human pose estimation by predicting joint locations was largely improved in accuracy.",35068
consists and (AFC). components: Pose-guided Composition (PPA) Part Attention-aware Feature Attention AACN two of main,AACN consists of two main components: Pose-guided Part Attention (PPA) and Attention-aware Feature Composition (AFC).,35069
applied and mask undesirable pedestrian out maps. feature learned to features PPA background in is,PPA is learned and applied to mask out undesirable background features in pedestrian feature maps.,35070
radar this resolve and position are autofocus. known to correct errors ambiguity Approaches generally as,Approaches to resolve this ambiguity and correct position errors are generally known as radar autofocus.,35071
for tracking given evaluation of systems. is object the unified A metric,A unified metric is given for the evaluation of object tracking systems.,35072
"of several augmentation envisage pathology. image of We unsupervised and the including applications classification approach,","We envisage several applications of the approach, including image augmentation and unsupervised classification of pathology.",35073
induced traditionally formulated zones as regions. were constraints multiple disjoint The prohibited (PZs) by,The constraints induced by prohibited zones (PZs) were traditionally formulated as multiple disjoint regions.,35074
a for paper OPF This problem the formulation PZs. new with proposes,This paper proposes a new formulation for the OPF problem with PZs.,35075
algorithm formulation significantly the proposed expedites implementation. The,The proposed formulation significantly expedites the algorithm implementation.,35076
"discriminability. limitation its works recent feature on exposed have However,","However, recent works have exposed its limitation on feature discriminability.",35077
viewpoint the of This casts on new weakness paper a softmax loss.,This paper casts a new viewpoint on the weakness of softmax loss.,35078
a hence between discrimination different We softmax the introduce function soft-margin classes. explicitly to encourage,We hence introduce a soft-margin softmax function to explicitly encourage the discrimination between different classes.,35079
"the loss the of other weak. hand, softmax is learned On classifier","On the other hand, the learned classifier of softmax loss is weak.",35080
"the diversity, Independence Criterion we achieve Hilbert-Schmidt (HSIC). adopt To the","To achieve the diversity, we adopt the Hilbert-Schmidt Independence Criterion (HSIC).",35081
learning of uncertainty deep quantification. with power Probabilistic deep combine expressive learning models the,Probabilistic deep learning models combine the expressive power of deep learning with uncertainty quantification.,35082
to distribution predict distributions von object a Our angle. pose model uses Mises over,Our model uses von Mises distributions to predict a distribution over object pose angle.,35083
Our model training time. likelihood-based at efficient for allows and test inference,Our model allows for likelihood-based training and efficient inference at test time.,35084
a the regions detected human detector. label state-of-the-art We with ts,We label the detected regions ts with a state-of-the-art human detector.,35085
one is once. Each at of the detected as regions marked least human,Each one of the detected regions is at least marked as human once.,35086
tasks Face as detection analysis reenactment recognition. such and facial face is essential facial to,Face detection is essential to facial analysis tasks such as facial reenactment and face recognition.,35087
"cascade called paper, devise anchor-based framework cascade. this we anchor an In efficient","In this paper, we devise an efficient anchor-based cascade framework called anchor cascade.",35088
dehazing an network. training neural end-to-end in,in training an end-to-end dehazing neural network.,35089
an whole-slide of in segmentation step images Tumor is important computer-assisted towards slides diagnosis. histology,Tumor segmentation in whole-slide images of histology slides is an important step towards computer-assisted diagnosis.,35090
CNN algorithms. outperforms the competing results demonstrate and combination Experimental PHPs that of features,Experimental results demonstrate that the combination of PHPs and CNN features outperforms competing algorithms.,35091
Overall persistent the highlights utility image homology of for framework histopathology proposed the analysis.,Overall the proposed framework highlights the utility of persistent homology for histopathology image analysis.,35092
are for high algorithms applications. many importance of Optical Flow,Optical Flow algorithms are of high importance for many applications.,35093
"as addition, In propose we improved optimization post-processing. variational","In addition, we propose improved variational optimization as post-processing.",35094
factors improves Intelligent control controllers. two these in,Intelligent control improves these two factors in controllers.,35095
an architecture in Altera The Stratix V FPGA. is implemented,The architecture is implemented in an Altera Stratix V FPGA.,35096
"deformation between embeddings Second, LSTM into frames. fed the an the to learn are network","Second, the embeddings are fed into an LSTM network to learn the deformation between frames.",35097
"Best operator, includes Evaporation. a we novel method Our call which anti-stagnation Value","Our method includes a novel anti-stagnation operator, which we call Best Value Evaporation.",35098
"research directions could methods. to Finally, future identify where lead we supertree improved","Finally, we identify directions where future research could lead to improved supertree methods.",35099
"we problem properties generator. rethink make this paper, of In the a proposal good what","In this paper, we rethink the problem of what properties make a good proposal generator.",35100
an social human-computer interaction. non-verbal processing and behavior Gaze in important is cue signal,Gaze behavior is an important non-verbal cue in social signal processing and human-computer interaction.,35101
in computations. The is executing extremely data-extensive challenge their biggest DNNs,The biggest challenge in executing DNNs is their extremely data-extensive computations.,35102
"data the deployment. directly networks quantizing %However, accuracy deep is in by limited system","%However, the system accuracy is limited by quantizing data directly in deep networks deployment.",35103
Previous inter-layer mainly works signals neglected. discretize on mainly weights while are focus,Previous works mainly focus on weights discretize while inter-layer signals are mainly neglected.,35104
example. the DNNs system implement memristor-based a proposed on the SNC We deployment as,We implement the proposed DNNs on the memristor-based SNC system as a deployment example.,35105
of finite previously a neurons. for uniform We size effects globally developed theory coupled generalize,We generalize a previously developed theory of finite size effects for uniform globally coupled neurons.,35106
discriminative an are and around object? views where But informative such,But where are such informative and discriminative views around an object?,35107
"and venues. this work, travel are as destination location In taken business","In this work, travel destination and business location are taken as venues.",35108
by is applications. very a Discovering important for a photo venue context-aware,Discovering a venue by a photo is very important for context-aware applications.,35109
data. heterogeneous discovery goal venue multimodal is Our fine-grained from social,Our goal is fine-grained venue discovery from heterogeneous social multimodal data.,35110
on proposed the of confirm method. results feasibility the this Experimental dataset,Experimental results on this dataset confirm the feasibility of the proposed method.,35111
estimates likelihood marginal unbiased. is An additional benefit are that,An additional benefit is that marginal likelihood estimates are unbiased.,35112
AU. The of determined the Gaussian is size the of the and by amplitude intensity,The amplitude and size of the Gaussian is determined by the intensity of the AU.,35113
has bandstop A in proposed varactor-based been article. tunable filter this,A varactor-based tunable bandstop filter has been proposed in this article.,35114
have networks classification. neural used Convolutional for widely image (CNNs) been,Convolutional neural networks (CNNs) have been widely used for image classification.,35115
"propose novel we a improve convolutional the To learning prototype (CPL). learning robustness, framework called","To improve the robustness, we propose a novel learning framework called convolutional prototype learning (CPL).",35116
"train classification we to criteria of CPL, framework the design the Under multiple network.","Under the framework of CPL, we design multiple classification criteria to train the network.",35117
Positive geometry matrices covariance The of of (SPD) that space Definite matrices. the Symmetric is,The space geometry of the covariance matrices is that of Symmetric Positive Definite (SPD) matrices.,35118
systems Automated rely vehicles. proper passing on classification toll the of,Automated toll systems rely on proper classification of the passing vehicles.,35119
vehicle. whole obtain about To the information,To obtain information about the whole vehicle.,35120
Existing interactive as promising approaches venue. a propose search image,Existing approaches propose interactive image search as a promising venue.,35121
three and datasets baselines three experimental We extensive on outperform settings.,We outperform three baselines on three datasets and extensive experimental settings.,35122
"this interpretation. with comes of approach complications However,","However, this approach comes with complications of interpretation.",35123
Optimal parameter Instrumental Linear The approximation the (OLIVA). in called Variables the Approximation is linear,The parameter in the linear approximation is called the Optimal Linear Instrumental Variables Approximation (OLIVA).,35124
instrument is in unknown the be identified. The estimand not may IV and,The instrument in the IV estimand is unknown and may not be identified.,35125
discuss least We also to weighted criteria. extensions squares,We also discuss extensions to weighted least squares criteria.,35126
excellent suggest performance an simulations finite sample the Carlo for proposed inferences. Monte,Monte Carlo simulations suggest an excellent finite sample performance for the proposed inferences.,35127
"set a is At training-time, filters. explicit the boundary with provided of network separate","At training-time, the network is provided with a separate set of explicit boundary filters.",35128
tracking in a compared plays role performance trackers. Correlation major filter existing improved to,Correlation filter plays a major role in improved tracking performance compared to existing trackers.,35129
The the tracker of the to location predict correlation the response target. uses adaptive,The tracker uses the adaptive correlation response to predict the location of the target.,35130
frame were proposed accuracy high correlation recently varieties of and trackers with rates. Many,Many varieties of correlation trackers were proposed recently with high accuracy and frame rates.,35131
approach validated sequences. Benchmark Tracking Object is The using,The approach is validated using Object Tracking Benchmark sequences.,35132
The optimize frame-wise usually likelihood the employ maximal existing a models. to loss methods,The existing methods usually employ a frame-wise maximal likelihood loss to optimize the models.,35133
can that performance. recognition text substantially boost show scene Experimental EP results the,Experimental results show that the EP can substantially boost scene text recognition performance.,35134
constraint specific search requires good problems of often heuristics. or Effective choosing solving,Effective solving of constraint problems often requires choosing good or specific search heuristics.,35135
is to heuristics give search robust which automatic The goal performance. obtain,The goal is to obtain automatic search heuristics which give robust performance.,35136
our adaptive than technique experiments is original heuristics. more show that the robust Preliminary search,Preliminary experiments show that our adaptive technique is more robust than the original search heuristics.,35137
outperform the also heuristics. can original It,It can also outperform the original heuristics.,35138
"Using reduction, analysis. significant we PWM-based dimension identify subspaces for","Using dimension reduction, we identify significant PWM-based subspaces for analysis.",35139
ML-based sequence PWM-based These subspaces tool. an form analysis,These PWM-based subspaces form an ML-based sequence analysis tool.,35140
improved the TFs. on performance mammalian It also,It also improved the performance on mammalian TFs.,35141
easily to tools PWM-based future well information. as is more as The include extendable ensemble,The ensemble is easily extendable to include more tools as well as future PWM-based information.,35142
personalized full Reaching the on precision of genome of interpretation. the medicine potential depends quality,Reaching the full potential of precision medicine depends on the quality of personalized genome interpretation.,35143
the compared better the original that achieve results results show proposed with Evaluation FlowNet. network,Evaluation results show that the proposed network achieve better results compared with the original FlowNet.,35144
This block-matching optical flow called proposes (ABMOF). algorithm event-driven adaptive OF an paper,This paper proposes an event-driven OF algorithm called adaptive block-matching optical flow (ABMOF).,35145
ABMOF accumulated events. time of uses slices DVS,ABMOF uses time slices of accumulated DVS events.,35146
results. The OF based time slices adaptively input events are on rotated the and,The time slices are adaptively rotated based on the input events and OF results.,35147
conventional standards as Lucas-Kanade Results such achieves accuracy (LK). comparable show that to ABMOF,Results show that ABMOF achieves comparable accuracy to conventional standards such as Lucas-Kanade (LK).,35148
implemented. our adapted using also is An slices LK method,An LK method using our adapted slices is also implemented.,35149
"strategies influence and this sampling conformational how explore space how might predictions? do Namely,","Namely, how do sampling strategies explore conformational space and how might this influence predictions?",35150
"we kinetics. compare in describing thermodynamics MSM Additionally, and estimators","Additionally, we compare MSM estimators in describing thermodynamics and kinetics.",35151
"dilated equipped we convolutional blocks a different generic classification of Then, with rates. network design","Then, we design a generic classification network equipped with convolutional blocks of different dilated rates.",35152
"obtains apparent performance approach Despite state-of-the-arts. over proposed the our superior simplicity,","Despite the apparent simplicity, our proposed approach obtains superior performance over state-of-the-arts.",35153
of different facial landmarks. distortions into expressions categories facial requires capturing regional Classifying,Classifying facial expressions into different categories requires capturing regional distortions of facial landmarks.,35154
best are of results results aware these are Both we the of.,Both of these results are the best results we are aware of.,35155
considered. for adaptive is worst-case problem robust The beamforming signal general-rank model,The worst-case robust adaptive beamforming problem for general-rank signal model is considered.,35156
algorithm to Herein inner second-order it. solve an is (SOCP) proposed cone approximate program,Herein an inner second-order cone program (SOCP) approximate algorithm is proposed to solve it.,35157
"relaxation does result, computationally As a technique. algorithm heavy SDP use not our","As a result, our algorithm does not use computationally heavy SDP relaxation technique.",35158
significance fundus doctors The retinal vessels of segmentation to diseases. for diagnose of the is,The segmentation of retinal vessels is of significance for doctors to diagnose the fundus diseases.,35159
the mainly generator. We structure the network improve of,We mainly improve the network structure of the generator.,35160
multiple in has enjoyed years. Reinforcement learning recent successes,Reinforcement learning has enjoyed multiple successes in recent years.,35161
in is topic Parsimony signal active representation a research. of,Parsimony in signal representation is a topic of active research.,35162
a conditional GAN model HiGAN \emph{high-level} of \emph{low-level} consists conditional The a GAN. and,The HiGAN model consists of a \emph{low-level} conditional GAN and a \emph{high-level} conditional GAN.,35163
(i.e. datasets experiments on recognition Comprehensive video two challenging,Comprehensive experiments on two challenging video recognition datasets (i.e.,35164
"better? by Inspired we advancement, asked do ourselves, such can we","Inspired by such advancement, we asked ourselves, can we do better?",35165
Networks (DCNet) propose We (DCNet++). and Networks Diverse Capsule Capsule Dense,We propose Dense Capsule Networks (DCNet) and Diverse Capsule Networks (DCNet++).,35166
in different maps This primary forming in the incorporating layers helps capsules. learned by feature,This helps in incorporating feature maps learned by different layers in forming the primary capsules.,35167
"network, which deeper adds essentially leads to of feature discriminative DCNet, maps. convolution a learning","DCNet, essentially adds a deeper convolution network, which leads to learning of discriminative feature maps.",35168
classification on the architectures. of efficacy using proposed Experiments image task datasets demonstrate the benchmark,Experiments on image classification task using benchmark datasets demonstrate the efficacy of the proposed architectures.,35169
wind the forecasts. of This around on energy confidence construction aims bands work,This work aims on the construction of confidence bands around wind energy forecasts.,35170
"'open (e.g. and differing distinguish ones door', 'open cupboard'),","'open door', 'open cupboard'), and distinguish differing ones (e.g.",35171
vs verb-only 'open labels. 'open door' using bottle'),'open door' vs 'open bottle') using verb-only labels.,35172
legitimate ambiguities class (Fig. Current between approaches recognition and for verbs neglect action semantic overlaps,Current approaches for action recognition neglect legitimate semantic ambiguities and class overlaps between verbs (Fig.,35173
"two-stream learn multi-output fusion using regression, CNN. these as We a representations","We learn these representations as multi-output regression, using a two-stream fusion CNN.",35174
"Finally, real evaluate collected framework using data. we developed effectiveness our the of vehicle the","Finally, we evaluate the effectiveness of our developed framework using the collected real vehicle data.",35175
wavelet identification neural nonlinear system. network online of performs the (WNN) The the,The wavelet neural network (WNN) performs the online identification of the nonlinear system.,35176
is prove to used the theory the of The MPC. stability Lyapunov,The Lyapunov theory is used to prove the stability of the MPC.,35177
vehicle. unmanned We control methodology identification online of an the apply the and to autonomous,We apply the methodology to the online identification and control of an unmanned autonomous vehicle.,35178
"time videos is for this longer However, consuming. very","However, for longer videos this is very time consuming.",35179
produce accurate architectures deep datasets. Modern on highly learning segmentation challenging many semantic results,Modern deep learning architectures produce highly accurate results on many challenging semantic segmentation datasets.,35180
"neonates Limited-channel of analytic by lack research in is open, tools. Goal: accessible hindered EEG","Goal: Limited-channel EEG research in neonates is hindered by lack of open, accessible analytic tools.",35181
three experienced by validated was The reviewers. aEEG algorithm,The aEEG algorithm was validated by three experienced reviewers.,35182
"reviewers background classification. Using methodology, a pattern assigned standard","Using standard methodology, reviewers assigned a background pattern classification.",35183
using Pearson's coefficient. Results compared were correlation,Results were compared using Pearson's correlation coefficient.,35184
the work the reduces manual facilitates and identity processing. of It recognition automatic,It reduces the manual work of identity recognition and facilitates the automatic processing.,35185
the from the data. layers features learns intermediate CNN The automatically hierarchical at,The CNN learns the hierarchical features at intermediate layers automatically from the data.,35186
does paper for performance anti-spoofing. This CNNs face of evaluation a,This paper does a performance evaluation of CNNs for face anti-spoofing.,35187
and The CNN are used this ResNet Inception study. architectures in,The Inception and ResNet CNN architectures are used in this study.,35188
Spoofing results are Database. Face MSU Mobile computed over The benchmark,The results are computed over benchmark MSU Mobile Face Spoofing Database.,35189
The in for face results using these settings. different obtained favorable anti-spoofing CNN architectures are,The favorable results are obtained using these CNN architectures for face anti-spoofing in different settings.,35190
a machine model. replacement to We this apply estimator,We apply this estimator to a machine replacement model.,35191
to challenging Hierarchical is approach these learning a able that is principled reinforcement tackle tasks.,Hierarchical reinforcement learning is a principled approach that is able to tackle these challenging tasks.,35192
hierarchical learning POMDP. We propose reinforcement approach a in hierarchical for learning deep,We propose a hierarchical deep reinforcement learning approach for learning in hierarchical POMDP.,35193
The RL is hierarchical to apply MDP learning. proposed to algorithm deep POMDP and both,The deep hierarchical RL algorithm is proposed to apply to both MDP and POMDP learning.,35194
evaluate We hierarchical challenging the algorithm proposed various POMDP. on,We evaluate the proposed algorithm on various challenging hierarchical POMDP.,35195
tasks scene visual are parsing understanding. estimation two Depth in particularly and scene important,Depth estimation and scene parsing are two particularly important tasks in visual scene understanding.,35196
challenging are datasets conducted two (i.e. on experiments Extensive,Extensive experiments are conducted on two challenging datasets (i.e.,35197
novel paper method presents for correspondence. cross-domain sparse a This,This paper presents a novel method for sparse cross-domain correspondence.,35198
input unorganized data. perspective five network of point representations the of cloud The consists,The network input consists of five perspective representations of the unorganized point cloud data.,35199
point. box the The an values and objectness each outputs bounding network for offset map,The network outputs an objectness map and the bounding box offset values for each point.,35200
KITTI the evaluation the dataset quantitative evaluations help We with server. carried of out,We carried out quantitative evaluations with the help of the KITTI dataset evaluation server.,35201
extract and spatial both for pose network features frames. we From the features image,From the spatial network we extract image features and pose features for both frames.,35202
serve model that (TFF). Temporal temporal Fields These for as Flow input our predicts features,These features serve as input for our temporal model that predicts Temporal Flow Fields (TFF).,35203
formulate representation similarity novel to joints. of This allows detected a measure,This novel representation allows to formulate a similarity measure of detected joints.,35204
"as driving, crucial estimation are autonomous pose technologies. self-localization/camera applications scene For and parsing such","For applications such as autonomous driving, self-localization/camera pose estimation and scene parsing are crucial technologies.",35205
"propose problems In paper, tackle two this these to framework a simultaneously. unified we","In this paper, we propose a unified framework to tackle these two problems simultaneously.",35206
semantically-labeled. Both point are and the images the clouds,Both the point clouds and the images are semantically-labeled.,35207
sensors. Each has highly from video frame motion ground truth pose accurate,Each video frame has ground truth pose from highly accurate motion sensors.,35208
"studies of proposed various effectiveness ablation performed, the the Finally, are system. demonstrate which","Finally, various ablation studies are performed, which demonstrate the effectiveness of the proposed system.",35209
learning alter behavioral is perception. evidence and physiological that category can This,This is behavioral and physiological evidence that category learning can alter perception.,35210
net this a We effect. neural model sketch for,We sketch a neural net model for this effect.,35211
of training be Deep to large learning effective. data requires amounts,Deep learning requires large amounts of training data to be effective.,35212
"iteratively on training, currently predicted the based we errors of clicks add During the segmentation.","During training, we iteratively add clicks based on the errors of the currently predicted segmentation.",35213
data these and an domain-specific expensive laborious to is Labeling names task. color learn,Labeling data to learn these domain-specific color names is an expensive and laborious task.,35214
"article learn aim labeled color from weakly data. names Therefore, to in we this","Therefore, in this article we aim to learn color names from weakly labeled data.",35215
"we For branch naming network. an color attention purpose, the this add to","For this purpose, we add an attention branch to the color naming network.",35216
attention pixel-wise used The predictions modulate the to of branch the color is network. naming,The attention branch is used to modulate the pixel-wise color naming predictions of the network.,35217
"the illustrate experiments, relevant branch we the that attention regions. identifies In correctly","In experiments, we illustrate that the attention branch correctly identifies the relevant regions.",35218
apply of model to self-consistency splices. detecting We task localizing this image and the,We apply this self-consistency model to the task of detecting and localizing image splices.,35219
"optimization experiences Thereby, techniques. important parameter report detailed initialization we concerning and","Thereby, we report important experiences concerning detailed parameter initialization and optimization techniques.",35220
very they flexible constraints Inequality exact because prior not assume knowledge. are do,Inequality constraints are very flexible because they do not assume exact prior knowledge.,35221
evaluation conditions. non-studio This us allows present qualitative to in and extension quantitative,This extension allows us to present quantitative and qualitative evaluation in non-studio conditions.,35222
idea issues address challenging simple This style transfer can methods. existing the effectively in,This simple idea can effectively address the challenging issues in existing style transfer methods.,35223
"we approach. loss, also progressive a on proposed present optimization the feature-domain Based","Based on the proposed loss, we also present a progressive feature-domain optimization approach.",35224
deadline increases The as bets very rapidly comes betting near. for the of amount,The amount of bets increases very rapidly as the deadline for betting comes near.,35225
bets gives the bettor which value Each on largest of the expectation benefit. a horse,Each bettor bets on a horse which gives the largest expectation value of the benefit.,35226
"alternative be Using simulations, biased. we approximate-Bayesian (ABC) computation one to on find method based","Using simulations, we find one alternative method based on approximate-Bayesian computation (ABC) to be biased.",35227
"we for to argument, methods them. acquiring improve need and Yet, of apply models","Yet, to improve and apply models of argument, we need methods for acquiring them.",35228
"out fallen of post-processing more In works has favour. however, recent CRF","In more recent works however, CRF post-processing has fallen out of favour.",35229
can CRFs All backpropagation. optimized using of parameters convolutional easily the be,All parameters of the convolutional CRFs can easily be optimized using backpropagation.,35230
implementation further research CRF our publicly make facilitating we available. To,To facilitating further CRF research we make our implementation publicly available.,35231
"learning new \emph{e.g. very with a are Humans little of capable concept fine-grained supervision,","Humans are capable of learning a new fine-grained concept with very little supervision, \emph{e.g.",35232
"is collecting challenging. datasets annotated such However,","However, collecting such annotated datasets is challenging.",35233
complete existing entries. new We databases database by propose to generating,We propose to complete existing databases by generating new database entries.,35234
(GAN) the generative training algorithm process. networks adopts training adversarial Our,Our training algorithm adopts the generative adversarial networks (GAN) training process.,35235
facial Robust localization partially are remains challenging  landmark faces  occluded. task when a,  Robust facial landmark localization remains a challenging task when faces are partially occluded.,35236
on of We dataset RICPR COFW. the evaluate challenging,We evaluate RICPR on the challenging dataset of COFW.,35237
weights the then fine We propagation. role tune using player back,We then fine tune the player role weights using back propagation.,35238
and The the interpretability hierarchical of ensures representation. integrity the the group architecture,The hierarchical architecture ensures the interpretability and the integrity of the group representation.,35239
RMS error. values show of results an improvement of terms Our in prediction,Our results show an improvement in terms of RMS values of prediction error.,35240
"counting paper, In model this to a introduce (PMC). we novel solve algorithm projected","In this paper, we introduce a novel algorithm to solve projected model counting (PMC).",35241
primal the small of algorithm treewidth input of instance. exploits Our graph the,Our algorithm exploits small treewidth of the primal graph of the input instance.,35242
"deep recently they been incorporated As architectures. such, into have","As such, they have recently been incorporated into deep architectures.",35243
the kernel boost method MvCCDA. performance develop We a of to further MvCCDA of,We develop a kernel method of MvCCDA to further boost the performance of MvCCDA.,35244
"kernel optimization analysis for MvCCDA extension, and completeness. presented Beyond are of complexity also","Beyond kernel extension, optimization and complexity analysis of MvCCDA are also presented for completeness.",35245
years. recent analysis become under has the Medical image topic spotlight a in,Medical image analysis has become a topic under the spotlight in recent years.,35246
medical in of machine a concerning There the is research image usage learning. significant progress,There is a significant progress in medical image research concerning the usage of machine learning.,35247
"and answers questions awaiting still problems and solutions, numerous there However, are respectively.","However, there are still numerous questions and problems awaiting answers and solutions, respectively.",35248
physical subject. extremely is an optical of relevant of layer the networks Supervision,Supervision of the physical layer of optical networks is an extremely relevant subject.,35249
"in are complexity, Characteristics simulated as time, processing specificity, such and environments. analysed sensitivity,","Characteristics such as sensitivity, specificity, processing time, and complexity, are analysed in simulated environments.",35250
for an There explain is their to demand algorithms outcomes. increasing,There is an increasing demand for algorithms to explain their outcomes.,35251
"is So produced no far, ranking explains rankings method there the algorithm. a that by","So far, there is no method that explains the rankings produced by a ranking algorithm.",35252
in generations. network problem deep a focus neural We on training of the,We focus on the problem of training a deep neural network in generations.,35253
of approach. also Model the and extraction ensemble feature our transfer verify effectiveness,Model ensemble and transfer feature extraction also verify the effectiveness of our approach.,35254
Carlo algorithm a on Monte inference powerful manifolds. Bayesian Geodesic non-Euclidean is (gMC) for,Geodesic Monte Carlo (gMC) is a powerful algorithm for Bayesian inference on non-Euclidean manifolds.,35255
object are applications. one Human class interesting numerous faces with,Human faces are one interesting object class with numerous applications.,35256
"problem images deblurring In by this exploiting we the address of structures. face facial paper,","In this paper, we address the problem of deblurring face images by exploiting facial structures.",35257
affect can the Inter-pathologist test diagnostic accuracy. results in variability,Inter-pathologist variability in the test results can affect diagnostic accuracy.,35258
in have Multi-label manifested learning machine problems various applications. themselves learning,Multi-label learning problems have manifested themselves in various machine learning applications.,35259
"challenges. framework space Recently, transformation these targeting proposed a label been (LST) has","Recently, a label space transformation (LST) framework has been proposed targeting these challenges.",35260
"(DLST) paper, transformation this In distribution-based we model. propose space label a","In this paper, we propose a distribution-based label space transformation (DLST) model.",35261
"the Consequently, using performance. classifier better codes trained dense latent multi-label yields","Consequently, multi-label classifier trained using the dense latent codes yields better performance.",35262
The the information of to enables fill out leverage DLST correlations. additional about distribution label,The leverage of distribution enables DLST to fill out additional information about the label correlations.,35263
employed to ML-KNN the label Then is transformed vector the code. original latent recover from,Then ML-KNN is employed to recover the original label vector from the transformed latent code.,35264
"human the paper, associations faces we In voices. and study this between","In this paper, we study the associations between human faces and voices.",35265
outcome which increases poor leads rates. mortality treatment This the to,This leads to poor treatment outcome which increases the mortality rates.,35266
Tehran tested Center. Service Emergency system as in The been has prototype designed a,The designed system has been tested as a prototype in Tehran Emergency Service Center.,35267
to in changes This leads contact patterns.,This leads to changes in contact patterns.,35268
population demonstrate negative the at level. models potential to consequences the use network We epidemic,We use epidemic network models to demonstrate the potential negative consequences at the population level.,35269
through take models. structure of account the several into network population social the We,We take into account the social structure of the population through several network models.,35270
"infectious evolves, the epidemic their themselves distance susceptible individuals contacts. from As may","As the epidemic evolves, susceptible individuals may distance themselves from their infectious contacts.",35271
replace social ties. Some lost seeking individuals by connections their new,Some individuals replace their lost social connections by seeking new ties.,35272
"real-world in same can negative effect Moreover, arise networks. the","Moreover, the same negative effect can arise in real-world networks.",35273
importance findings the measures analysis careful models. preventive in epidemic of of highlight These,These findings highlight the importance of careful analysis of preventive measures in epidemic models.,35274
tasks. related among generalization sharing knowledge improves learning by Multi-task performance,Multi-task learning improves generalization performance by sharing knowledge among related tasks.,35275
segmentation detection proposed using architecture evaluated We a combination two and on our of datasets.,We evaluated our proposed architecture on a combination of detection and segmentation using two datasets.,35276
how demonstrate wild with limited Experiments birds CNN from representations general our datasets. learns,Experiments with wild birds demonstrate how our CNN learns general representations from limited datasets.,35277
finding constraint to a is problems. solutions QGS satisfaction systematic approach of,QGS is a systematic approach to finding solutions of constraint satisfaction problems.,35278
provide QGS better shows with Comparison that significantly classification. neural network trained networks classifiers other,Comparison with other neural network classifiers shows that QGS trained networks provide significantly better classification.,35279
generated. material Gene genetic is duplication which through is a new mechanism major,Gene duplication is a major mechanism through which new genetic material is generated.,35280
tasks is an important computer different and vision for Texture cue applications.,Texture is an important cue for different computer vision tasks and applications.,35281
efficient yet Pattern the (LBP) of one considered is descriptors. Local Binary best texture,Local Binary Pattern (LBP) is considered one of the best yet efficient texture descriptors.,35282
"noise. sensitivity notable limitations, However, mostly some to has the LBP","However, LBP has some notable limitations, mostly the sensitivity to noise.",35283
distributions models. are in building a block Gaussian used generative key commonly as many,Gaussian distributions are commonly used as a key building block in many generative models.,35284
"However, well their networks. explored deep applicability been has not in","However, their applicability has not been well explored in deep networks.",35285
multivariate can model the functions be data. dependence copula to structure used of Empirical,Empirical copula functions can be used to model the dependence structure of multivariate data.,35286
algorithm numerically benefits the computational theoretically and The and error approximation of is assessed.,The computational benefits and approximation error of the algorithm is theoretically and numerically assessed.,35287
"we for network (DR-ResNet) based counting. deeply-recursive crowd propose a Consequently, blocks on ResNet","Consequently, we propose a deeply-recursive network (DR-ResNet) based on ResNet blocks for crowd counting.",35288
"Beijing data video-monitoring from Besides, we dataset a bus the new generate station. of","Besides, we generate a new dataset from the video-monitoring data of Beijing bus station.",35289
MEnet several public desirable that Tests performance. show benchmarks has on achieved,Tests on several public benchmarks show that MEnet has achieved desirable performance.,35290
data tera-voxel current brain interest. neuroanatomy much of light-microscopic is Whole sets using,Whole brain neuroanatomy using tera-voxel light-microscopic data sets is of much current interest.,35291
tasks. hence They for facial important analysis various are,They are hence important for various facial analysis tasks.,35292
utilize the shape and to appearance facial ways in information. the They differ,They differ in the ways to utilize the facial appearance and shape information.,35293
models explicitly global build holistic represent and appearance to shape The the methods information. facial,The holistic methods explicitly build models to represent the global facial appearance and shape information.,35294
the models. appearance local model build global shape explicitly CLMs The but the leverage,The CLMs explicitly leverage the global shape model but build the local appearance models.,35295
facial shape capture appearance implicitly information. regression-based methods and The,The regression-based methods implicitly capture facial shape and appearance information.,35296
"theories as discuss we differences. category, each as algorithms well their their within For underlying","For algorithms within each category, we discuss their underlying theories as well as their differences.",35297
"strengths out and their weaknesses. on evaluations, the point respective we Based","Based on the evaluations, we point out their respective strengths and weaknesses.",35298
section There separate the deep review to learning-based is also a latest algorithms.,There is also a separate section to review the latest deep learning-based algorithms.,35299
listing of software. existing also and includes The survey a benchmark the databases,The survey also includes a listing of the benchmark databases and existing software.,35300
This cameras. is performance with calibrated three recorded,This performance is recorded with three calibrated cameras.,35301
observed a accuracy. It classification is that at energy efficiency sacrifice is with slight achieved,It is observed that energy efficiency is achieved with a slight sacrifice at classification accuracy.,35302
"other all energy energy networks, novel ensemble models. efficient outperforms our efficient model Among","Among all energy efficient networks, our novel ensemble model outperforms other energy efficient models.",35303
pathway cortex via sensory early modulates nuclei. The connections sensory cerebral feed-back to processing,The cerebral cortex modulates early sensory processing via feed-back connections to sensory pathway nuclei.,35304
of poorly human are for modulation understood. functions top-down The behavior this,The functions of this top-down modulation for human behavior are poorly understood.,35305
frameworks on recognition cerebral Speech commonly explained is cortex areas. in focusing,Speech recognition is commonly explained in frameworks focusing on cerebral cortex areas.,35306
"To improve used. EP performance, the be can algorithm the","To improve the performance, the EP algorithm can be used.",35307
"center on we based channel priors. dark and Then, a channel center-dark generate saliency map","Then, we generate a center-dark channel map based on center saliency and dark channel priors.",35308
of the in of is The dimensionality realm curse association twofold. rules,The curse of dimensionality in the realm of association rules is twofold.,35309
network balance The incorporates transmitter isolation. on-chip enhance architecture an (TX)-receiver also the (RX) to,The architecture also incorporates an on-chip balance network to enhance the transmitter (TX)-receiver (RX) isolation.,35310
neural-nets a We for novel propose to method stage. the merge inference convolutional,We propose a novel method to merge convolutional neural-nets for the inference stage.,35311
the performance further The model. re-trained to of weights fine-tune shared merged the are,The shared weights are further re-trained to fine-tune the performance of the merged model.,35312
the satisfactory of validate a method. effectiveness and demonstrate Experimental results performance the,Experimental results demonstrate a satisfactory performance and validate the effectiveness of the method.,35313
"parameters. atmospheric of Moreover, the approach on does model estimation scattering the not proposed rely","Moreover, the proposed approach does not rely on estimation of the atmospheric scattering model parameters.",35314
"experiments on datasets. O-HAZE and NYU-Depth, We I-HAZE, conduct","We conduct experiments on NYU-Depth, I-HAZE, and O-HAZE datasets.",35315
approach both CycleGAN Extensive method the experiments proposed quantitatively improves that and demonstrate qualitatively.,Extensive experiments demonstrate that the proposed approach improves CycleGAN method both quantitatively and qualitatively.,35316
far molecular However being from the understood. is mechanism underlying,However the underlying molecular mechanism is far from being understood.,35317
Convolutional networks Learning solving applications of Machine for in neural problems. are ubiquitous variety a,Convolutional neural networks are ubiquitous in Machine Learning applications for solving a variety of problems.,35318
for collection stiffening. chest results treatments These strategies investigation future motivate data and wall into,These results motivate future data collection strategies and investigation into treatments for chest wall stiffening.,35319
areas growth Microbial processes many division fundamental life to and science. are relevant of,Microbial growth and division are fundamental processes relevant to many areas of life science.,35320
"extending cycles. operate cells, mechanisms over single division These possibly within several","These mechanisms operate within single cells, possibly extending over several division cycles.",35321
"distinct However, measurements cells. all on relied experimental from pooled date many have to studies","However, all experimental studies to date have relied on measurements pooled from many distinct cells.",35322
"comparison the and accuracy, processing trade-off This memory, between time. assessed","This comparison assessed the trade-off between accuracy, memory, and processing time.",35323
simulations yields the that through mathematically We optimal solution. proposed approach show and an,We show mathematically and through simulations that the proposed approach yields an optimal solution.,35324
its use the of additional maps The encoder network generated image this latent to representation.,The use of the additional encoder network maps this generated image to its latent representation.,35325
to person's match camera aims a across Person streams. identity multiple re-identification,Person re-identification aims to match a person's identity across multiple camera streams.,35326
applied challenging been Deep person re-identification successfully neural the to have networks task.,Deep neural networks have been successfully applied to the challenging person re-identification task.,35327
agent the outperforms find RL heuristic the solutions solutions. that We learns near-optimal that,We find that the RL agent learns near-optimal solutions that outperforms the heuristic solutions.,35328
the recognition and method ensures considerable in than This previous time accuracy research. discrimination,This method ensures considerable discrimination in accuracy and recognition time than the previous research.,35329
"results, competitive techniques. unsupervised with can produce adaption domain ReID We that scene-independent show leading-edge","We show that scene-independent ReID can produce leading-edge results, competitive with unsupervised domain adaption techniques.",35330
"and we person for comparing introduce a ReID. Finally, dataset new across-camera within-camera","Finally, we introduce a new dataset for comparing within-camera and across-camera person ReID.",35331
"small due to time-consuming tend to datasets annotation. be However,","However, datasets tend to be small due to time-consuming annotation.",35332
and of This study aims structural non-structural at bridge images scenes. components recognizing from urban,This study aims at recognizing bridge structural and non-structural components from images of urban scenes.,35333
a filter response impulse function. A with filter Gaussian of is Gaussian,A Gaussian filter is a filter with impulse response of Gaussian function.,35334
"(e.g. for Also, data could be helpful they transmission","Also, they could be helpful for data transmission (e.g.",35335
"filters by Gaussian be methods. practice, could In approximately designed several the","In practice, the Gaussian filters could be approximately designed by several methods.",35336
Some problem of are marriage the aspects of stable discussed.,Some aspects of the problem of stable marriage are discussed.,35337
continue cases. discuss We partial intermediate to transferable,We continue to discuss intermediate partial transferable cases.,35338
image task Image is processing. classification an machine field and the important in of learning,Image classification is an important task in the field of machine learning and image processing.,35339
The K quantum Nearest-Neighbor on based scheme algorithm. is,The scheme is based on quantum K Nearest-Neighbor algorithm.,35340
"extracted images classical on computers. of are Firstly, vectors feature the","Firstly, the feature vectors of images are extracted on classical computers.",35341
"algorithm is for searching process used similarity. minimum the up search Next, quantum to speed","Next, the quantum minimum search algorithm is used to speed up searching process for similarity.",35342
"is Finally, classified measurement. image the quantum by","Finally, the image is classified by quantum measurement.",35343
"the the measurement only ensure step scheme. the to validity Moreover, once of is executed","Moreover, the measurement step is executed only once to ensure the validity of the scheme.",35344
"classification Hence, a while performance good quantum improving greatly our the has scheme efficiency.","Hence, our quantum scheme has a good classification performance while greatly improving the efficiency.",35345
patients. multiple cord sclerosis in atrophy affected lesions and/or (MS) frequently spinal The is by,The spinal cord is frequently affected by atrophy and/or lesions in multiple sclerosis (MS) patients.,35346
in is available Spinal Cord The proposed readily framework and the open-source Toolbox.,The proposed framework is open-source and readily available in the Spinal Cord Toolbox.,35347
optimal can with the This performance jammer. scheme interactions enough after the achieve,This scheme can achieve the optimal performance after enough interactions with the jammer.,35348
images (RLFC) in the a Our field redundancies method constructing tree light exploits by structure.,Our method (RLFC) exploits redundancies in the light field images by constructing a tree structure.,35349
understanding of how is evolution we of conceive Our strongly its by causes. shaped fundamental,Our understanding of evolution is shaped strongly by how we conceive of its fundamental causes.,35350
on that dynamics dynamics of the long-term Today clear is evolutionary depend it mutational introduction.,Today it is clear that long-term evolutionary dynamics depend on the dynamics of mutational introduction.,35351
"been among imagined, previously observed Mutation-biased diverse taxa. possibility adaptation, has a not","Mutation-biased adaptation, a possibility not previously imagined, has been observed among diverse taxa.",35352
sustained Directional under trends a are bias. possible,Directional trends are possible under a sustained bias.,35353
may biases. in origin effect Biases mutational an analogous have that to developmental are,Biases that are developmental in origin may have an effect analogous to mutational biases.,35354
"our a classes. maneuver trajectories on predictive distribution multi-modal over Additionally, future model outputs based","Additionally, our model outputs a multi-modal predictive distribution over future trajectories based on maneuver classes.",35355
scenarios. analysis the of a also present traffic We for qualitative model's various predicted distributions,We also present a qualitative analysis of the model's predicted distributions for various traffic scenarios.,35356
"existing remain philosophical, lacks and studies and many clear on RSI results. However, systems formulation","However, many existing studies on RSI systems remain philosophical, and lacks clear formulation and results.",35357
the landmark ended they target up covering.,the target landmark they ended up covering.,35358
"training intention mechanisms, promoting believe e.g. more We reading that protocols active","We believe that training protocols promoting more active intention reading mechanisms, e.g.",35359
"possible result, As populations. revealed number we real in effects, new subtle of a a","As a result, we revealed a number of new subtle effects, possible in real populations.",35360
"direction effects Moreover, propagation. the both in get of stronger","Moreover, both effects get stronger in the direction of propagation.",35361
the The almost open mean-field ion on channels of solely existing state. focuses theory,The existing mean-field theory of ion channels focuses almost solely on the open state.,35362
"test work, to we In hypothesis. aim this this","In this work, we aim to test this hypothesis.",35363
"study. switching that of prohibits noise Furthermore, system the show in we","Furthermore, we show that noise prohibits switching in the system of study.",35364
"extract to three features different at different uses columns, MCNN scales. each filter sizes, with","MCNN uses three columns, each with different filter sizes, to extract features at different scales.",35365
show popular very experiments datasets results. on Extensive three compelling,Extensive experiments on three popular datasets show very compelling results.,35366
"was it segmentation judge the However, humans necessary for result. to CNN","However, it was necessary for humans to judge the CNN segmentation result.",35367
"and OF thus computationally is is flow However, task-oriented or pre-computed. costly, typically other","However, OF or other task-oriented flow is computationally costly, and is thus typically pre-computed.",35368
used window widely years. approaches Sliding been recognition have for tasks recent in object,Sliding window approaches have been widely used for object recognition tasks in recent years.,35369
presented of sliding the conventional first. A window theoretical is approach investigation,A theoretical investigation of the conventional sliding window approach is presented first.,35370
rough date of mentioned estimations publications computational to Other the cost. only,Other publications to date only mentioned rough estimations of the computational cost.,35371
"the approach window detail. in disparity is sliding Subsequently, presented proposed","Subsequently, the proposed disparity sliding window approach is presented in detail.",35372
"two with comparison Furthermore, made. state-of-the-art a is methods","Furthermore, a comparison with two state-of-the-art methods is made.",35373
from bones data This body investigates action skeleton based for recognition. paper skeleton,This paper investigates body bones from skeleton data for skeleton based action recognition.,35374
designed network A skeleton edge graph is based neural recognition. for action then convolutional,A graph edge convolutional neural network is then designed for skeleton based action recognition.,35375
"more hybrid networks. performance can improvements by achieved be the Additionally,","Additionally, more performance improvements can be achieved by the hybrid networks.",35376
The suggest experimental results method for results ZSD. yields promising that our,The experimental results suggest that our method yields promising results for ZSD.,35377
strategy standard the to is One training synthetically apply set. enlarge augmentation data to,One standard strategy is to apply data augmentation to synthetically enlarge the training set.,35378
and tested in component of consistency. accuracy of terms the is performance The bridge extraction,The performance of the bridge component extraction is tested in terms of accuracy and consistency.,35379
"initialization proposed. background In detection a this is robust superpixel based approach motion on paper,","In this paper, a robust background initialization approach based on superpixel motion detection is proposed.",35380
characteristics objects. of effectively eliminate adopted Both to spatial frames and temporal are foreground,Both spatial and temporal characteristics of frames are adopted to effectively eliminate foreground objects.,35381
for with is subsequence selected estimation. A illumination condition background first stable,A subsequence with stable illumination condition is first selected for background estimation.,35382
visual autonomous What representation a good is agents? for,What is a good visual representation for autonomous agents?,35383
the plans. people predictable bias that This prefer adds model,This model adds the bias that people prefer predictable plans.,35384
bias simple This with plan prediction. calculated is,This bias is calculated with simple plan prediction.,35385
human Our with higher inference. a had model correlation,Our model had a higher correlation with human inference.,35386
"This potential example application, mobile devices. in for their limits","This limits their potential application, for example in mobile devices.",35387
"of deraining. (LPNet) paper, lightweight we for pyramid this networks propose single image a In","In this paper, we propose a lightweight pyramid of networks (LPNet) for single image deraining.",35388
discuss of LPNet also tasks. We value and for the low- other vision potential high-level,We also discuss the potential value of LPNet for other low- and high-level vision tasks.,35389
results and coal-mining the data obtained disasters with The coherent previous radon literature. on are,The results obtained on the coal-mining disasters and radon data are coherent with previous literature.,35390
"depth Learning-based, to generalizes estimation unseen often datasets. poorly single-view","Learning-based, single-view depth estimation often generalizes poorly to unseen datasets.",35391
"learning-based, multi-view few depth There estimation methods. exists","There exists few learning-based, multi-view depth estimation methods.",35392
It depth estimation. to two-view is single- generalizable or dense,It is generalizable to single- or two-view dense depth estimation.,35393
"the keeping loss we further utilize improve while same, To performance, the distillation complexity function.","To further improve the performance, while keeping the complexity same, we utilize distillation loss function.",35394
of model data We the unlabeled with using labels train network. our soft the teacher,We train our model with unlabeled data using the soft labels of the teacher network.,35395
"play. context-dependent sparsity, game include strategy reversals abrupt These of time-sensitive and reward","These include reward sparsity, abrupt context-dependent reversals of strategy and time-sensitive game play.",35396
rates This a improvement introduces (i.e. paper method linking technological for,This paper introduces a method for linking technological improvement rates (i.e.,35397
Moore's (i.e. adoption and Law) technology curves,Moore's Law) and technology adoption curves (i.e.,35398
the first 'knee'? corresponding to is timing as for the hypothesized cross-over The,The timing for the cross-over is hypothesized as corresponding to the first 'knee'?,35399
for the analysis uncertainty organizations uncertainty. to future technological highlight The reduce around opportunities cases,The uncertainty analysis around the cases highlight opportunities for organizations to reduce future technological uncertainty.,35400
regresses rotations Previous either joint joint work positions. or,Previous work regresses either joint rotations or joint positions.,35401
requires skeleton bone onto latter avoid invalid stretching constraints and to configurations. The re-projection,The latter requires re-projection onto skeleton constraints to avoid bone stretching and invalid configurations.,35402
"On short-term state-of-the-art predictions, the quantitatively. QuaterNet improves","On short-term predictions, QuaterNet improves the state-of-the-art quantitatively.",35403
The consists steps. matcher three proposed of,The proposed matcher consists of three steps.,35404
popular Cricket games one in of this the most era. modern is undoubtedly,Cricket is undoubtedly one of the most popular games in this modern era.,35405
"establish graph positive Specifically, the patch-wise of to via correspondences we matching. training propose pairs","Specifically, we propose to establish the patch-wise correspondences of positive training pairs via graph matching.",35406
"of descriptive these Here, the symbolic languages. we about plasticity ask","Here, we ask about the plasticity of these symbolic descriptive languages.",35407
presented raw results focusing. data Preliminary for ERS are,Preliminary results are presented for ERS raw data focusing.,35408
also detecting method is capable incompletion. The of,The method is also capable of detecting incompletion.,35409
estimating in good object methods pose. causes fail registration This to a,This causes registration methods to fail in estimating a good object pose.,35410
"Furthermore, novel other desired, are statistical advanced if methods strategies computational are needed.","Furthermore, if other advanced statistical methods are desired, novel computational strategies are needed.",35411
captures is of practical compilers. that language an important KC most the Decision-DNNF,Decision-DNNF is an important KC language that captures most of the practical compilers.,35412
of of celebrities of from The movies. scenes extensive consists set violent dataset an,The dataset consists of an extensive set of violent scenes of celebrities from movies.,35413
pairs voting decision-making. well-studied STV are (RP) rules for and group two ranked,STV and ranked pairs (RP) are two well-studied voting rules for group decision-making.,35414
"surprisingly However, literature be broken. about the is should how ties vague","However, the literature is surprisingly vague about how ties should be broken.",35415
"novel under also respectively. ILP propose We STV RP, and formulations for PUT-winners","We also propose novel ILP formulations for PUT-winners under STV and RP, respectively.",35416
displacement technique novel field. decorrelation This which study addresses proposes estimating elastography a in the,This study proposes a novel elastography technique which addresses the decorrelation in estimating displacement field.,35417
accuracy recently benefited adding of from distillation residual The knowledge layers.,The accuracy of knowledge distillation recently benefited from adding residual layers.,35418
"allotted with counterparts vehicles subchannels. the communicate Accordingly, directly their utilizing","Accordingly, vehicles communicate directly with their counterparts utilizing the allotted subchannels.",35419
first This sensor-based work that to features. is data knowledge activity our visualizes abstract the,This is the first work to our knowledge that visualizes abstract sensor-based activity data features.,35420
in tasks. forefront Deep at stands the many computer vision learning,Deep learning stands at the forefront in many computer vision tasks.,35421
solution. this address the task We following proposing by,We address this task by proposing the following solution.,35422
fine-tuning trained The is then annotations such procedure. with by an model noisy improved online,The model trained with such noisy annotations is then improved by an online fine-tuning procedure.,35423
will be released acceptance. Code and the paper upon data,Code and data will be released upon the paper acceptance.,35424
"Afterwards, implement maps with Block. low-level the connections skip via Fusion feature we","Afterwards, we implement the skip connections with low-level feature maps via Fusion Block.",35425
"is often supervision available not in full practice. However,","However, full supervision is often not available in practice.",35426
radar requires transmit-receive a function number tasks. to the execute of Each,Each function requires the radar to execute a number of transmit-receive tasks.,35427
"task is NP-hard problem. In an general, scheduling","In general, task scheduling is an NP-hard problem.",35428
"methods low performance. but hand, the other provide have complexity relatively heuristic poor On","On the other hand, heuristic methods have low complexity but provide relatively poor performance.",35429
assumption present the in images. structural natural regularity neglects,assumption neglects the structural regularity present in natural images.,35430
chronic is irreversible Glaucoma vision a eye loss. disease to that leads,Glaucoma is a chronic eye disease that leads to irreversible vision loss.,35431
"accuracy, rely However, segmentation ignore methods on heavily visual the measurement-based these and features. various","However, these measurement-based methods rely heavily on the segmentation accuracy, and ignore various visual features.",35432
"the as result. fused different the are probabilities screening final Finally, output of streams","Finally, the output probabilities of different streams are fused as the final screening result.",35433
"by into dividing image the image the Thus, evaluated overlapping regions. is locally","Thus, the image is evaluated locally by dividing the image into overlapping regions.",35434
Each set fragmented local is region on Gaussians patches. small of a further into,Each region is further fragmented into a set of local Gaussians on small patches.,35435
"a multi-level Gaussians A Gaussian these global creating encodes, each for region structure. local","A global Gaussian encodes, these local Gaussians for each region creating a multi-level structure.",35436
"the Also, earlier we learning metric have on this of descriptor. analyzed methods efficiency","Also, we have analyzed the efficiency of earlier metric learning methods on this descriptor.",35437
"a image is available. Conventionally, reconstructed object good fidelity classification with after is performed object","Conventionally, object classification is performed after a reconstructed object image with good fidelity is available.",35438
the label. of concatenation The model to class three is predict CNN the outputs used,The concatenation of the three CNN model outputs is used to predict the class label.,35439
experiments datasets show proposed model. advantage the the The of over benchmark,The experiments over benchmark datasets show the advantage of the proposed model.,35440
neural features Image extracted a using convolutional pre-trained were network.,Image features were extracted using a pre-trained convolutional neural network.,35441
"ego-motion the are quantized, are features. histograms as windowed Eye and the used and","Eye and ego-motion are quantized, and the windowed histograms are used as the features.",35442
compared combination better activity obtains for features. individual classification as to of accuracy The features,The combination of features obtains better accuracy for activity classification as compared to individual features.,35443
convolutional detecting neural use networks fence for We pixels.,We use convolutional neural networks for detecting fence pixels.,35444
"or images relationship one-to-many between can many-to-many. domain one-to-one, The two be","The relationship between two domain images can be one-to-one, one-to-many or many-to-many.",35445
"performed tasks, generation unpaired translation edges-to-objects and including Experiments image on are image translation. facial","Experiments are performed on unpaired image generation tasks, including edges-to-objects translation and facial image translation.",35446
"feedback computational either suffer from from reranking require or current However, users approaches burdensome costs.","However, current reranking approaches either require feedback from users or suffer from burdensome computational costs.",35447
effective results proposals The our experimental and that are demonstrate efficient.,The experimental results demonstrate that our proposals are effective and efficient.,35448
voltage-dependent are and Kinetic conductances. parameters closing the that the describe those of opening,Kinetic parameters are those that describe the opening and closing of the voltage-dependent conductances.,35449
supported theoretical empirical by are arguments evidence. extensive Our,Our theoretical arguments are supported by extensive empirical evidence.,35450
"semantics core idea is called this and The of to SemStyle, model, style. separate","The core idea of this model, called SemStyle, is to separate semantics and style.",35451
recognition object manipulation. for task is Understanding planning robotic and states important object as as,Understanding object states is as important as object recognition for robotic task planning and manipulation.,35452
"Resnet the based solution problem, model state a to deep proposed. a As is identification","As a solution to the state identification problem, a Resnet based deep model is proposed.",35453
on classes. is with model eleven trained The of and the Imagenet dataset weights initialized,The model is initialized with Imagenet weights and trained on the dataset of eleven classes.,35454
generative (e.g. refers an target object Bionic which creativity of design to in a approach,Bionic design refers to an approach of generative creativity in which a target object (e.g.,35455
of a features objects to designed contain is (e.g. biological source floor lamp),a floor lamp) is designed to contain features of biological source objects (e.g.,35456
"generative unsupervised realising propose a We approach novel design. deep to DesignGAN, bionic","We propose DesignGAN, a novel unsupervised deep generative approach to realising bionic design.",35457
Networks have images. Generative photo-realistic in promise shown generating great,Generative Networks have shown great promise in generating photo-realistic images.,35458
"still this, the surrounding Despite is theory area. an them active research","Despite this, the theory surrounding them is still an active research area.",35459
single MNIST We dataset on the labels. moving our with model train class,We train our model on the single moving MNIST dataset with class labels.,35460
is unsupervised on Adversarial algorithm feature based learning visual hybrid It Generative an prediction. context,It is an unsupervised visual feature learning hybrid Generative Adversarial algorithm based on context prediction.,35461
approaches overview till an This of presents now. that developed been have SLAM paper,This paper presents an overview of SLAM approaches that have been developed till now.,35462
"detailed. algorithm, and its further Kinect variants, Fusion in discussed are approaches developed","Kinect Fusion algorithm, its variants, and further developed approaches are discussed in detailed.",35463
as agent sensory whole The a space task its may posit (i.e. space,The agent may posit as a task space its whole sensory space (i.e.,35464
to one individual. is common of describe an the Gender most attributes used,Gender is one of the most common attributes used to describe an individual.,35465
"AutoGen termed the for We Autoencoder model, as robust Representative a propose Class same.","We propose a robust Class Representative Autoencoder model, termed as AutoGen for the same.",35466
academic communities. driving from learning for end-to-end has drawn Imitation autonomous attention,Imitation learning for end-to-end autonomous driving has drawn attention from academic communities.,35467
"applied we issue, order address to technique. In this SMOTE","In order to address this issue, we applied SMOTE technique.",35468
Different and form cell into variable types compartments patterns. migrate tissue distribution tend various to,Different cell types tend to migrate into various tissue compartments and form variable distribution patterns.,35469
tissue between correlative Drug of in and requires compartments. analysis various biomarkers development the,Drug development requires correlative analysis of various biomarkers in and between the tissue compartments.,35470
"by are slides pathologists. annotated manually expert To tissue that, enable","To enable that, tissue slides are manually annotated by expert pathologists.",35471
"task. a tissue error-prone annotation is slides labor intensive, of Manual and tedious","Manual annotation of tissue slides is a labor intensive, tedious and error-prone task.",35472
to for data humans gathering impossible difficult from the UAV robot area or is reach.,UAV is data gathering robot from the difficult or impossible area for humans to reach.,35473
problem which an dynamics proposed contains by RNN. restoration modeled is a control The,The proposed control problem contains a restoration dynamics which is modeled by an RNN.,35474
proposed restorer the model unfolding dynamically (DURR). the recurrent We call,We call the proposed model the dynamically unfolding recurrent restorer (DURR).,35475
is Then often F-measure. saliency of detected evaluated in maps the quality of terms,Then the quality of detected saliency maps is often evaluated in terms of F-measure.,35476
"FLoss activations. produce the continuously network the force to Consequently, can polarized","Consequently, the FLoss can continuously force the network to produce polarized activations.",35477
structured time (STS) dependency. modeling the interweaved of The requires problem series spatiotemporal classification,The structured time series (STS) classification problem requires the modeling of interweaved spatiotemporal dependency.,35478
spatial temporal model STS previous dependencies most the methods classification independently. and,most previous STS classification methods model the spatial and temporal dependencies independently.,35479
motivates merits. network a neural to deep This us with design such,This motivates us to design a deep neural network with such merits.,35480
"and datasets. for and SVHN, test PCN CIFAR We with train image classification ImageNet","We train and test PCN for image classification with SVHN, CIFAR and ImageNet datasets.",35481
reveal or bottom-up saliency visual attention. Errors prediction of also top-down,Errors of top-down prediction also reveal visual saliency or bottom-up attention.,35482
"and falls and However, efficiency. challenging of remains it accuracy short","However, it remains challenging and falls short of accuracy and efficiency.",35483
the evaluation conventional We and CNN Model-based methods. Gaussian traditional compared Mixture conduct with the,We conduct the evaluation compared with the conventional CNN and traditional Gaussian Mixture Model-based methods.,35484
"Moreover, using this paper to further, the accuracy improve classification mixup of explores the method.","Moreover, to improve the classification accuracy further, this paper explores the using of mixup method.",35485
{Radio easy gained for deployment. technology popularity cheap and Frequency Identification has,{Radio Frequency Identification technology has gained popularity for cheap and easy deployment.,35486
relies extended on Kalman decoupled scenario adjustment filter parameter method. The,The parameter adjustment scenario relies on decoupled extended Kalman filter method.,35487
outperforms real when experiments that regularization on conventional data. trained DNNs OLE-GRSVNet show with Numerical,Numerical experiments show that OLE-GRSVNet outperforms DNNs with conventional regularization when trained on real data.,35488
"Our method with a navigable portraits space. latent smooth, of photorealistic fictive generates identities","Our method generates photorealistic portraits of fictive identities with a smooth, navigable latent space.",35489
core people in a categories Understanding cognitive represent is how problem science.,Understanding how people represent categories is a core problem in cognitive science.,35490
method's a and quantitative qualitative results proof-of-concept feasibility. We for as the provide,We provide qualitative and quantitative results as a proof-of-concept for the method's feasibility.,35491
from frame face trained to tracker An frame. follows online the,An online trained tracker follows the face from frame to frame.,35492
schemes constraints from allocation prevent to Both emerging. conflicts incorporate,Both schemes incorporate constraints to prevent allocation conflicts from emerging.,35493
Plain field convolutional expense networks at enlarge of receptive (CNNs) the the cost. computational generally,Plain convolutional networks (CNNs) generally enlarge the receptive field at the expense of computational cost.,35494
"filtering dilated Recently, to this issue. been adopted has address","Recently, dilated filtering has been adopted to address this issue.",35495
"decrease further used layer channels another of is maps. to the feature convolutional Furthermore,","Furthermore, another convolutional layer is further used to decrease the channels of feature maps.",35496
dependent the fingerprint immensely of quality are recognition Fingerprint techniques images. on,Fingerprint recognition techniques are immensely dependent on quality of the fingerprint images.,35497
"is proposed. fingerprint In this enhancement a algorithm paper, adaptive quality","In this paper, a quality adaptive fingerprint enhancement algorithm is proposed.",35498
quality images enhancement process. fingerprint of two-stage a through each undergo quality Fingerprint class,Fingerprint images of each quality class undergo through a two-stage fingerprint quality enhancement process.,35499
is derived truncated distribution. under Gaussian An achievable the rate input,An achievable rate is derived under the truncated Gaussian input distribution.,35500
"the exploited. arrays Meanwhile, is symmetrical of LED geometry","Meanwhile, the symmetrical geometry of LED arrays is exploited.",35501
"the genetic algorithm. association transmitter-user solved Finally, formulated and a by is problem","Finally, the transmitter-user association problem is formulated and solved by a genetic algorithm.",35502
our methods against on three existing evaluate LRMR-based We datasets. framework benchmark,We evaluate our framework against existing LRMR-based methods on three benchmark datasets.,35503
"""grows"" using scaffold. This as grammar model incrementally examples positive a a","This model incrementally ""grows"" a grammar using positive examples as a scaffold.",35504
regexes ability from with examples. is competitive We to this that human show approach learn,We show that this approach is competitive with human ability to learn regexes from examples.,35505
"reinforcement guided word-level by via rewards. and is updated HRGR-Agent learning, sentence-level","HRGR-Agent is updated via reinforcement learning, guided by sentence-level and word-level rewards.",35506
in great learning has successes spawned many Progress engineering deep applications. in,Progress in deep learning has spawned great successes in many engineering applications.,35507
tracking-by-clustering in formulation. a descriptor used The discriminative adapted is,The adapted discriminative descriptor is used in a tracking-by-clustering formulation.,35508
"Location heatmaps, are regression landmark. to Gaussian by made predictions heatmap one per","Location predictions are made by regression to Gaussian heatmaps, one heatmap per landmark.",35509
of problem content. the research aesthetics assessment generic has explored with images Recent widely of,Recent research has widely explored the problem of aesthetics assessment of images with generic content.,35510
aesthetic for images a paper introduces quality of method faces. This assessment with,This paper introduces a method for aesthetic quality assessment of images with faces.,35511
"binary, methods outperforms Experimental both approach for our i.e. results existing show that","Experimental results show that our approach outperforms existing methods for both binary, i.e.",35512
"prediction in different four continuous the aesthetic score on low/high, and state-of-the-art. databases","low/high, and continuous aesthetic score prediction on four different databases in the state-of-the-art.",35513
quite has many of real-time algorithms. heuristics research search The produced area,The research area of real-time heuristics search has produced quite many algorithms.,35514
times Overall segmentation about one minute. were,Overall segmentation times were about one minute.,35515
results or planning postoperative visualization advantages. offers and several treatment surgical of,surgical treatment planning or visualization of postoperative results and offers several advantages.,35516
"view, In their to from environment. parasites exposure the novel DPL increases increased also","In their view, increased DPL also increases exposure to novel parasites from the environment.",35517
"the briefly this contextualize we main criticism. In object reply, of","In this reply, we briefly contextualize the main object of criticism.",35518
competing We two the discuss the of then scenarios. putative parsimony,We then discuss the putative parsimony of the two competing scenarios.,35519
synthetically This by has using previously addressed data. problem been generated,This problem has previously been addressed by using synthetically generated data.,35520
"networks However, to data. generalize on real trained to synthetic often fail data","However, networks trained on synthetic data often fail to generalize to real data.",35521
methods deep Channel up is of speed family to inference. model's an important pruning,Channel pruning is an important family of methods to speed up deep model's inference.,35522
regard model independent Previous pruning channel two fine-tuning as filter and pruning algorithms steps.,Previous filter pruning algorithms regard channel pruning and model fine-tuning as two independent steps.,35523
"after to be zero training. the removed corresponding values filters Hence, all safely can index","Hence, all the filters corresponding to zero index values can be removed safely after training.",35524
"we model in prevent drop By an can accuracy. erasing several excessive gradually filters, weak","By gradually erasing several weak filters, we can prevent an excessive drop in model accuracy.",35525
"(including achieves significantly better performance. from training previous with Compared scratch), algorithms AutoPruner state-of-the-art pruning","Compared with previous state-of-the-art pruning algorithms (including training from scratch), AutoPruner achieves significantly better performance.",35526
"exist solved. be two further However, problems still to there","However, there still exist two problems to be further solved.",35527
"i.e., PSRN and proposed We Sub-JHMDB on PennAction. challenging the evaluate two benchmarks,","We evaluate the proposed PSRN on two challenging benchmarks, i.e., Sub-JHMDB and PennAction.",35528
traffic and detection object involving emphasizes This research mainly detection essentially on thus classification.,This research mainly emphasizes on traffic detection thus essentially involving object detection and classification.,35529
introduce optical flow orientation of also use the We to information. features accommodate,We also introduce the use of optical flow features to accommodate orientation information.,35530
performance results rate. results Experimental in steady this a rise in that demonstrate the,Experimental results demonstrate that this results in a steady rise in the performance rate.,35531
show hashing the proposed large-scale Experiments on datasets that state-of-the-art algorithms. method superior to is,Experiments on large-scale datasets show that the proposed method is superior to state-of-the-art hashing algorithms.,35532
"deep methods tasks. to In recent image learning years, have applied successfully been classification","In recent years, deep learning methods have been successfully applied to image classification tasks.",35533
deep Many such exist neural easily cats today dogs. from that networks can differentiate,Many such deep neural networks exist today that can easily differentiate cats from dogs.,35534
the process. at paper involved and need discusses first The ranking for the channel,The paper at first discusses the need for channel ranking and the involved process.,35535
boundary. each continuous contrasts Other opposite induced images included along,Other images included opposite contrasts along each induced continuous boundary.,35536
algorithm. and the and quality of the show results model effectiveness Experimental proposed,Experimental results show the effectiveness and quality of the proposed model and algorithm.,35537
impaired visually represents cost-effective paper a system towards This scene perception individual. aimed,This paper represents a cost-effective scene perception system aimed towards visually impaired individual.,35538
provided in outdoor scenes. traffic results Experimental this paper use,Experimental results provided in this paper use outdoor traffic scenes.,35539
of This reduces number overlap negatives. false the,This overlap reduces the number of false negatives.,35540
purpose For soft this non-maximum smoothing. fine-tuned suppression with we Gaussian the,For this purpose we fine-tuned the soft non-maximum suppression with Gaussian smoothing.,35541
"general, events New the internally paced of method: such timing (e.g. In","New method: In general, the timing of such internally paced events (e.g.",35542
heartbeats cycles) respiratory may a variability. large display or,heartbeats or respiratory cycles) may display a large variability.,35543
CR and detect validate these Results: to some events. practical tools CARE-rCortex provides,Results: CARE-rCortex provides some practical tools to detect and validate these CR events.,35544
phase define or to time-locked cycle. heart Users a of respiratory can baselines,Users can define baselines time-locked to a phase of respiratory or heart cycle.,35545
real with datasets. CARE-rCortex the We the cardio-respiratory use of of two illustrate analysis,We illustrate the use of CARE-rCortex with the analysis of two real cardio-respiratory datasets.,35546
of Different be baseline epochs. in can used EEG the analysis TF normalizations,Different baseline normalizations can be used in the TF analysis of EEG epochs.,35547
processing local features. and on applications computer rely image vision Many,Many computer vision and image processing applications rely on local features.,35548
well-known descriptors. It blur motion of that traditional performance decreases feature is detectors and the,It is well-known that motion blur decreases the performance of traditional feature detectors and descriptors.,35549
"shutter algorithms, handle method distortion. can deblurring blur and rolling spatially-variant Unlike most the","Unlike most deblurring algorithms, the method can handle spatially-variant blur and rolling shutter distortion.",35550
"of Furthermore, it state-of-the-art in contrary running capable to real-time algorithms. is","Furthermore, it is capable of running in real-time contrary to state-of-the-art algorithms.",35551
"Besides, information. discriminative regions different contain distinct","Besides, different regions contain distinct discriminative information.",35552
regions to should important prediction. The more the contribute,The important regions should contribute more to the prediction.,35553
"deep a into Finally, a network the map pooling score. regresses quality","Finally, a deep pooling network regresses the quality map into a score.",35554
BIQA methods. many that predictions Experiments our outperform state-of-the-art have demonstrated,Experiments have demonstrated that our predictions outperform many state-of-the-art BIQA methods.,35555
"by data still is eye-tracking consuming and analyzing hand mobile limited. However, time","However, analyzing mobile eye-tracking data by hand is time consuming and still limited.",35556
at levels. and hardware requires optimization both software Success,Success requires optimization at both software and hardware levels.,35557
"non-uniform representing distributed parameters However, data and (e.g.","However, representing non-uniform data and distributed parameters (e.g.",35558
"the manner. Recently, was posit proposed, in non-uniform system a number represents numbers which","Recently, the posit number system was proposed, which represents numbers in a non-uniform manner.",35559
the very of connectome part a signatures. codes the small show We that,We show that a very small part of the connectome codes the signatures.,35560
correlated are characteristics. their in wildly and vary,are correlated and vary wildly in their characteristics.,35561
proposed methods show by approach Experimentally the baseline surpasses the provided we the organizers. that,Experimentally we show that the proposed approach surpasses the baseline methods provided by the organizers.,35562
each error and accuracy the of pixel presented absolute formulation. average average We for report,We report average accuracy and average absolute pixel error for each of the presented formulation.,35563
segmentation corresponding The function analyzed. is and error,The segmentation function and corresponding error is analyzed.,35564
for novel We propose part-based object in arbitrary tracking a sequences. video an challenging method,We propose a novel part-based method for tracking an arbitrary object in challenging video sequences.,35565
"disentangled improve disentanglement Better in analysis extraction, the and feature synthesis. helps turn, further common","Better common feature extraction, in turn, helps further improve the disentanglement analysis and disentangled synthesis.",35566
methods. an re-localization camera regression coordinate has essential current become of Scene part,Scene coordinate regression has become an essential part of current camera re-localization methods.,35567
and local The objects' information. flow semantic encode stochastic graph correlation,The stochastic flow graph encode objects' local correlation and semantic information.,35568
are training. on using surgical phase based recognition supervised State-of-the-art algorithms laparoscopic fully videos,State-of-the-art surgical phase recognition algorithms using laparoscopic videos are based on fully supervised training.,35569
networks surgical It also performance. recognition that boosts observed phase CNN-LSTM of end-to-end training is,It is also observed that end-to-end training of CNN-LSTM networks boosts surgical phase recognition performance.,35570
We noise. block impulsive its oriented pulse and describe modulation resilience multidimensional against position,We describe block oriented multidimensional pulse position modulation and its resilience against impulsive noise.,35571
modulation and of of encoder decoder part algorithm. The the the implements the BBC,The modulation implements the encoder and part of the decoder of the BBC algorithm.,35572
"classifier is as For network extractor. a purpose, data-driven, neural feature used that visually-aware a","For that purpose, a neural network classifier is used as a data-driven, visually-aware feature extractor.",35573
similarity-based as for a latter ranking input algorithm. The recommendations serves then using,The latter then serves as input for similarity-based recommendations using a ranking algorithm.,35574
available tested on Fashion is Our dataset. publicly the approach,Our approach is tested on the publicly available Fashion dataset.,35575
product from larger learning using databases presented. are strategies Initialization transfer,Initialization strategies using transfer learning from larger product databases are presented.,35576
shortest using in paper shaders. for present several improvements OpenGL computing path maps this We,We present in this paper several improvements for computing shortest path maps using OpenGL shaders.,35577
"side-information ""cannot-link"" partial to in applications, In clustering is ""must-link"" or available. some indicate","In some applications, partial side-information to indicate ""must-link"" or ""cannot-link"" in clustering is available.",35578
clustering This subspace to with side-information. task of the leads,This leads to the task of subspace clustering with side-information.,35579
OurDifNet trained complex end-to-end without loss is functions.,OurDifNet is trained end-to-end without complex loss functions.,35580
show mathematically solution. yields optimal approach We an proposed that and simulations through the,We show mathematically and through simulations that the proposed approach yields an optimal solution.,35581
algorithm is Motion background detection subtraction. based mostly on,Motion detection algorithm is mostly based on background subtraction.,35582
We compared algorithms. its processing with other speed also,We also compared its processing speed with other algorithms.,35583
smartphone videos cover The are usage. selected carefully to typical,The videos are carefully selected to cover typical smartphone usage.,35584
face through cannot tracking approaches. results solved existing suggest The mobile that be,The results suggest that mobile face tracking cannot be solved through existing approaches.,35585
The one x-rays allows objects. of penetration to large image power,The penetration power of x-rays allows one to image large objects.,35586
"synchrotron micron-level For specimens sources. resolution with example, using centimeter-sized can imaged be","For example, centimeter-sized specimens can be imaged with micron-level resolution using synchrotron sources.",35587
This current procedure with the iterates comparing seam until negligibly seams. evaluation-reestimation changes previous the,This evaluation-reestimation procedure iterates until the current seam changes negligibly comparing with the previous seams.,35588
a sampling disk. We present uniformly new methods by solid angle for the subtended,We present new methods for uniformly sampling the solid angle subtended by a disk.,35589
the onto the the of projection disk unit sphere).,the projection of the disk onto the unit sphere).,35590
allow low-variance sampling from direct for of sources. These disk-shaped illumination mappings stratified light,These mappings allow for low-variance stratified sampling of direct illumination from disk-shaped light sources.,35591
"novel training paper, we method second-order neural for develop In feed-forward this nets. a","In this paper, we develop a novel second-order method for training feed-forward neural nets.",35592
the good the algorithm proposed show of performance benchmark datasets. We on,We show the good performance of the proposed algorithm on benchmark datasets.,35593
The issue model of selection vital importance. is research in applied of,The issue of model selection in applied research is of vital importance.,35594
"paper, show is that indeed possible. Within this this we","Within this paper, we show that this is indeed possible.",35595
image in search. achieved remarkable field neural of networks success the Deep have,Deep neural networks have achieved remarkable success in the field of image search.,35596
"However, the and images tested trained algorithms natural captured cameras. ordinary with state-of-the-art for are","However, the state-of-the-art algorithms are trained and tested for natural images captured with ordinary cameras.",35597
adversarial results in suggest from terms of solved far being that is MNIST robustness. These,These results suggest that MNIST is far from being solved in terms of adversarial robustness.,35598
presented are arguments in of second. favour Some the,Some arguments are presented in favour of the second.,35599
the by correlation interference of is The degree determined different trials. between,The degree of interference is determined by the correlation between different trials.,35600
which show FORCE of this specific avoid phenomenon. We ingredients,We show which specific ingredients of FORCE avoid this phenomenon.,35601
"in difference magnitude that of slower results is orders LMS. convergence for this Overall,","Overall, this difference results in convergence that is orders of magnitude slower for LMS.",35602
Learning of task. accumulating form overall to across implies trials the concept information the multiple,Learning implies accumulating information across multiple trials to form the overall concept of the task.,35603
"lesion approach classification problem. Finally, skin confirm we of the our on effectiveness","Finally, we confirm the effectiveness of our approach on skin lesion classification problem.",35604
serious profiled. want threat This for not individuals a be poses do to who privacy,This poses a serious privacy threat for individuals who do not want to be profiled.,35605
the revolutionize Mathematics role medical to has substantial a science. played,Mathematics has played a substantial role to revolutionize the medical science.,35606
various Intelligent on in to systems efficient diagnosing diseases. based have proved be mathematical theories,Intelligent systems based on mathematical theories have proved to be efficient in diagnosing various diseases.,35607
analytical cascading solve deploy subproblems. to (ATC) method We optimization target distributedly,We deploy analytical target cascading (ATC) method to distributedly solve optimization subproblems.,35608
into while an preserving algorithm. We this reflection symmetry through iterative the image partition superpixels,We partition the image into superpixels while preserving this reflection symmetry through an iterative algorithm.,35609
"two the and these we this disentanglement. cross-domain objectives concept of paper, bridge introduce In","In this paper, we bridge these two objectives and introduce the concept of cross-domain disentanglement.",35610
parts. the three aim into We to representation internal separate,We aim to separate the internal representation into three parts.,35611
for contains part shared The both information domains.,The shared part contains information for both domains.,35612
propose adversarial this augmentation address data limitation. to We,We propose adversarial data augmentation to address this limitation.,35613
also We for design strategy training. effective reward/penalty joint a,We also design a reward/penalty strategy for effective joint training.,35614
during shown information Scientific that processing representations influence studies have experience. stimuli conscious non-conscious and,Scientific studies have shown that non-conscious stimuli and representations influence information processing during conscious experience.,35615
"asymptotic an the also expression ASC presented. Additionally, analytical for is","Additionally, an asymptotic analytical expression for the ASC is also presented.",35616
performance investigated. and impact fading the multipath shadowing is secrecy on of The,The impact of shadowing and multipath fading on the secrecy performance is investigated.,35617
with to simulations analytical the compared results are validate Monte-Carlo analysis. The,The analytical results are compared with Monte-Carlo simulations to validate the analysis.,35618
"learning Recently, technology CADx. image been successfully applied has including to deep a medical analysis","Recently, a deep learning technology has been successfully applied to medical image analysis including CADx.",35619
"explaining based Existing however, deep have the CADx diagnostic decision. in learning approaches, limitation a","Existing deep learning based CADx approaches, however, have a limitation in explaining the diagnostic decision.",35620
"reasonable explanation. clinical real be In could with practice, decisions clinical made","In real clinical practice, clinical decisions could be made with reasonable explanation.",35621
world current learning deep limited So are in approaches in deployment. real CADx,So current deep learning approaches in CADx are limited in real world deployment.,35622
"with framework. this the In paper, we CADx (ICADx) proposed interpretability CADx interpretable investigate in","In this paper, we investigate interpretability in CADx with the proposed interpretable CADx (ICADx) framework.",35623
the The proposed mammogram was on database. of method validated public effectiveness,The effectiveness of the proposed method was validated on public mammogram database.,35624
a feature in point Structure Motion(SFM). matching from Image step is key,Image feature point matching is a key step in Structure from Motion(SFM).,35625
"we a paper, accelerated this image method Hashing. matching GPU Cascade improved In proposed with","In this paper, we proposed a GPU accelerated image matching method with improved Cascade Hashing.",35626
"method parallelize Cascade on Next, the Hashing GPU. we","Next, we parallelize the Cascade Hashing method on GPU.",35627
"nearest skip proposed achieving neighbor uses almost with Another similar a architecture connection interpolation, results.","Another proposed architecture uses a skip connection with nearest neighbor interpolation, achieving almost similar results.",35628
program trade classification formulated is accuracy rule An for simplicity. to integer optimally,An integer program is formulated to optimally trade classification accuracy for rule simplicity.,35629
"approximate algorithm using randomization. handle CG To we an propose datasets, large","To handle large datasets, we propose an approximate CG algorithm using randomization.",35630
Facial recognition a image. conventionally computed is from single attribute,Facial attribute recognition is conventionally computed from a single image.,35631
"may face subject practice, multiple have images. In each","In practice, each subject may have multiple face images.",35632
"issue. address approaches inconsistency to the develop Then, two we","Then, we develop two approaches to address the inconsistency issue.",35633
experiments on large two attributes. databases annotations public facial The are with of conducted,The experiments are conducted on two large public databases with annotations of facial attributes.,35634
and were deleted LSTM. the high-noise The by waveband found intervals,The high-noise waveband intervals were found and deleted by the LSTM.,35635
waveband determined waveband-based optimum-data CNN's were by intervals The detection.,The optimum-data waveband intervals were determined by CNN's waveband-based detection.,35636
wavebands. modeling higher than The full-waveband the other of was modeling significantly or accuracy any,The accuracy was significantly higher than the full-waveband modeling or modeling of any other wavebands.,35637
constraints. four key vision through are Our these geometric fundamental that is problems coupled insight,Our key insight is that these four fundamental vision problems are coupled through geometric constraints.,35638
"neurons approach relevant model saliency that penalizes are prediction, for This i.e. high most","This approach penalizes high saliency neurons that are most relevant for model prediction, i.e.",35639
the This clustering studies paper subspace problem.,This paper studies the subspace clustering problem.,35640
block existing special The of property methods many our into case. diagonal falls,The block diagonal property of many existing methods falls into our special case.,35641
of real BDR. demonstrate Experiments on datasets effectiveness the,Experiments on real datasets demonstrate the effectiveness of BDR.,35642
rendering algorithm in participating introduce work for we this media. transient novel In a,In this work we introduce a novel algorithm for transient rendering in participating media.,35643
extend We the domain. estimates the steady-state include temporal radiance to beam,We extend the beam steady-state radiance estimates to include the temporal domain.,35644
"comprehensive by platform. research is RGB-T evaluation limited a lacking However,","However, RGB-T research is limited by lacking a comprehensive evaluation platform.",35645
"a with the represented image In with patches object is nodes. tracked particular, as graph","In particular, the tracked object is represented with a graph with image patches as nodes.",35646
"degrades indirect illumination interference significantly resulting depth. from the (MPI) However, captured multipath","However, multipath interference (MPI) resulting from indirect illumination significantly degrades the captured depth.",35647
during attention lots decades. Electricity of drawn issue theft last detection has,Electricity theft detection issue has drawn lots of attention during last decades.,35648
"few are can amyloid be identified, in water. insoluble relatively From these, since structures amyloids","From these, relatively few amyloid structures can be identified, since amyloids are insoluble in water.",35649
"the PDB. are Therefore, mostly solid state in NMR-recorded deposited structures amyloid","Therefore, mostly solid state NMR-recorded amyloid structures are deposited in the PDB.",35650
"al. of to I. (Stankovic, et the Our relate method results","Our results relate to the method of (Stankovic, I. et al.",35651
the The in of outperforms terms accuracy. state-of-the-art segmentation model proposed methods,The proposed model outperforms the state-of-the-art methods in terms of segmentation accuracy.,35652
"geometry case, We tackle the unknown. addressed where parameters is rarely and camera scene","We tackle the rarely addressed case, where camera parameters and scene geometry is unknown.",35653
inference new how We learning can and language. existing for adapted be techniques show the,We show how existing inference and learning techniques can be adapted for the new language.,35654
"solutions such equipment special cameras as or require breathalyzers. Existing infrared electrocardiogram,","Existing solutions require special equipment such as electrocardiogram, infrared cameras or breathalyzers.",35655
the and performed validate approach experiments baselines. are to Extensive,Extensive experiments are performed to validate the approach and baselines.,35656
"the we errors multi-object In propose handle temporal to during tracking. this methods paper,","In this paper, we propose the methods to handle temporal errors during multi-object tracking.",35657
near appear the occurs or error noisy object. detections Temporal occluded objects are when,Temporal error occurs when objects are occluded or noisy detections appear near the object.,35658
"drift tracking ID-switching those errors and fail various or In occur. may situations, like","In those situations, tracking may fail and various errors like drift or ID-switching occur.",35659
shape It hard to temporal motion is using errors by and information. overcome only,It is hard to overcome temporal errors only by using motion and shape information.,35660
detections to provide useful scene effectively according We noisy also to condition. remove technique,We also provide useful technique to remove noisy detections effectively according to scene condition.,35661
"highly identity especially consistency, is Tracking performance, by methods. our improved attaching","Tracking performance, especially identity consistency, is highly improved by attaching our methods.",35662
important and many in classification detection an is applications. task Image distortion,Image distortion classification and detection is an important task in many applications.,35663
"these deep convolutional tackle to use network. neural we fully a To propose issues,","To tackle these issues, we propose to use a deep fully convolutional neural network.",35664
of methods. lossy widely is used JPEG compression the one,JPEG is one of the widely used lossy compression methods.,35665
"and suffer blocking bit-rates. usually including especially blurring, images artifacts compression at low JPEG-compressed from","JPEG-compressed images usually suffer from compression artifacts including blocking and blurring, especially at low bit-rates.",35666
primary to with increase inference training Our overhead. goal is minimal mappingaccuracy,Our primary goal is to increase inference mappingaccuracy with minimal training overhead.,35667
"we a namely conditioned conditionalimage the proposed a spatially suggest generation algorithm, Furthermore,using GAN. algorithm,","Furthermore,using the proposed algorithm, we suggest a conditionalimage generation algorithm, namely a spatially conditioned GAN.",35668
"image-retrieval showed the applied and effectiveness to Furthermore, its in tasks DTLC-GAN we learning. representation","Furthermore, we applied the DTLC-GAN to image-retrieval tasks and showed its effectiveness in representation learning.",35669
important most learning. factors machine is in of the one Data,Data is one of the most important factors in machine learning.,35670
"analysis adversarial semi-supervised Recently, performances exhibited image networks generative in excellent scenarios.","Recently, generative adversarial networks exhibited excellent performances in semi-supervised image analysis scenarios.",35671
"previous heterogeneous. as studies whole, structure this the most a While PMC treated is functionally","While most previous studies treated the PMC as a whole, this structure is functionally heterogeneous.",35672
PMC whether The are with particular investigated subdivisions study the DOCs. specifically of present associated,The present study investigated whether particular subdivisions of the PMC are specifically associated with DOCs.,35673
The fluctuations structural within subdivision. validated by PMC impaired finding this were functional substrates,The structural substrates were validated by finding impaired functional fluctuations within this PMC subdivision.,35674
a primarily make approach algorithm to overall contextually more the aware. propose learning We multitask,We propose a multitask learning approach primarily to make the overall algorithm more contextually aware.,35675
Observation of replacement Sensor also SSN's acts Stimulus (SSO) a SOSA core. as,SOSA also acts as a replacement of SSN's Stimulus Sensor Observation (SSO) core.,35676
youth effects on study This making development. investigates dose-response the of music,This study investigates the dose-response effects of making music on youth development.,35677
high-dimensional balancing (ii) to assess covariate in How settings?,(ii) How to assess covariate balancing in high-dimensional settings?,35678
Biological systems environment. under process information noisy,Biological systems process information under noisy environment.,35679
adaptation coli of its model because investigation of suitable E. is for Sensory simplicity.,Sensory adaptation model of E. coli is suitable for investigation because of its simplicity.,35680
"approach quantitatively, been stochastic the adaptation has To thermodynamic processing understand attempted.","To understand the adaptation processing quantitatively, stochastic thermodynamic approach has been attempted.",35681
under piloted of The conventional operate are aircrafts. conditions from UAVs those different which,The conditions under which UAVs operate are different from those of conventional piloted aircrafts.,35682
air-to-ground necessitates UAVs. development channel for propagation (AG) This of new models,This necessitates development of new air-to-ground (AG) propagation channel models for UAVs.,35683
measured domain analysis carried of out. Both frequency data are the time and,Both time and frequency domain analysis of the measured data are carried out.,35684
is relationships key objects. to between visual the of One the issues describe,One of the key issues is to describe the visual relationships between objects.,35685
"world interactions When problem. real capturing diverse dealing these very data, with difficult is a","When dealing with real world data, capturing these very diverse interactions is a difficult problem.",35686
can common in It alleviated sense network. by a incorporating be,It can be alleviated by incorporating common sense in a network.,35687
"decision. reasons predictor the that we a a explanations %Therefore, need certain reveals why takes","%Therefore, we need explanations that reveals the reasons why a predictor takes a certain decision.",35688
"faithful and method provide able We agnostic LORE, propose interpretable to an explanations.","We propose LORE, an agnostic method able to provide interpretable and faithful explanations.",35689
are these nonlinear models both two which parameters/inputs has Most variables in consequences. of and,Most of these models are nonlinear both in variables and parameters/inputs which has two consequences.,35690
"as parameters in are bifurcations models varied. behaviour many undergo Second,","Second, many models undergo bifurcations in behaviour as parameters are varied.",35691
metric' classifier. learning a introduce to novel `certainty We active a guide probabilistic multi-class for,We introduce a novel `certainty metric' to guide active learning for a multi-class probabilistic classifier.,35692
"outstanding ""key selected persons"", are The neighbor namely measurement. by saliency persons, based the K-nearest","The outstanding persons, namely ""key persons"", are selected by the K-nearest neighbor based saliency measurement.",35693
dispatch problem. We focus economic on the,We focus on the economic dispatch problem.,35694
multiple into horizon The decomposed considered smaller scheduling is sub-horizons.,The considered scheduling horizon is decomposed into multiple smaller sub-horizons.,35695
The ramp the generating rates units. of mathematically sub-horizons are using between modeled interdependencies,The interdependencies between the sub-horizons are mathematically modeled using ramp rates of generating units.,35696
the rate convergence This enhances significantly. technique,This technique enhances the convergence rate significantly.,35697
is by to This model the conditions. these approximate satisfy distorting done,This is done by distorting the approximate model to satisfy these conditions.,35698
"data, non-linearity. importance neglected the the we Using investigate of empirical US","Using US data, we investigate the empirical importance of the neglected non-linearity.",35699
"each the only two. and In one contrast, concern vehicle segmentation of detection semantic","In contrast, vehicle detection and semantic segmentation each only concern one of the two.",35700
propose tackle this to boundary-aware a We with multi-task semantic problem network. learning,We propose to tackle this problem with a semantic boundary-aware multi-task learning network.,35701
reconstructed can material fiber The model. reinforced using a structure be composite of a tracking-by-detection,The structure of a fiber reinforced composite material can be reconstructed using a tracking-by-detection model.,35702
"accuracy, the heavily performance. on detection rely recall Tracking-by-detection algorithms especially","Tracking-by-detection algorithms rely heavily on detection accuracy, especially the recall performance.",35703
significant tracking advanced over state-of-the-art with algorithms fiber show together performance. Experiments improvements detection,Experiments show significant improvements over state-of-the-art fiber detection algorithms together with advanced tracking performance.,35704
powerful are feedforward for tool functional (DFNNs) a neural approximation. Deep networks,Deep feedforward neural networks (DFNNs) are a powerful tool for functional approximation.,35705
learning our also in method. We perform to how deep variable describe selection,We also describe how to perform variable selection in our deep learning method.,35706
experiments our that SISR method validate methods. outperforms state-of-the-art Extensive the existing,Extensive experiments validate that our SISR method outperforms the existing state-of-the-art methods.,35707
model perspective assume focal a distortion. unknown We and radial unknown length with camera,We assume a perspective camera model with unknown focal length and unknown radial distortion.,35708
one modelled is distortion The using the parameter. radial model division with,The radial distortion is modelled using the division model with one parameter.,35709
This a results solver. numerically in fast stable very and,This results in a very fast and numerically stable solver.,35710
achieve with dynamic routing fewer generalization capsules model more helps parameters. many The capacity,The dynamic routing helps capsules achieve more generalization capacity with many fewer model parameters.,35711
with methods multiple parallel benchmarks. The achieve other leading a models in proposed performance,The proposed models achieve a parallel performance with other leading methods in multiple benchmarks.,35712
neural to stimuli. Artificial behavior in activation internal also have response networks,Artificial neural networks also have internal activation behavior in response to stimuli.,35713
"identifies Further, both tasks. high with performance on networks HMS","Further, HMS identifies networks with high performance on both tasks.",35714
transmit to The explanations knowledge humans. main of to objective is,The main objective of explanations is to transmit knowledge to humans.,35715
construct for proposes This machine explanations made models. informative predictions learning to work from,This work proposes to construct informative explanations for predictions made from machine learning models.,35716
human ranked These to understanding. that explanations informative improves filtered produce and concepts are,These concepts are filtered and ranked to produce informative explanations that improves human understanding.,35717
to are developed experiments the illustrate Numerical provided theory.,Numerical experiments are provided to illustrate the developed theory.,35718
"manifold. Definite Specifically, Symmetric a it is (SPD) Positive","Specifically, it is a Symmetric Positive Definite (SPD) manifold.",35719
to input being SPD layer we the rectifying the However singular. matrices exploit from prevent,However we exploit the rectifying layer to prevent the input SPD matrices from being singular.,35720
with objective innovative a introduce function. the also We non-linear an of network learning proposed,We also introduce a non-linear learning of the proposed network with an innovative objective function.,35721
"version of We developed a also named deep have RieMNet, as DRieMNet.","We have also developed a deep version of RieMNet, named as DRieMNet.",35722
the results our the method experimental superiority Extensive of over state-of-the-art. show,Extensive experimental results show the superiority of our method over the state-of-the-art.,35723
next is to of KT to question. response The the objective a predict student's,The objective of KT is to predict a student's response to the next question.,35724
propose address regularization to a these novel then problems. We,We then propose a novel regularization to address these problems.,35725
method to us proposed DLKT. The allows achieve trustworthy,The proposed method allows us to achieve trustworthy DLKT.,35726
Such in cortex. the observed mechanisms activity can patterns models spatiotemporal of search underlying facilitate,Such models can facilitate the search of mechanisms underlying observed spatiotemporal activity patterns in cortex.,35727
"spiking (LFPs) the With from local forward activity. models, can be simulated field potentials computed","With forward models, local field potentials (LFPs) can be computed from the simulated spiking activity.",35728
"probabilities distance-dependent the densities, introduces and connection and neuron upscaling delays. The preserves","The upscaling preserves the neuron densities, and introduces distance-dependent connection probabilities and delays.",35729
disentangled a crucial task. yet representations interpretable is challenging Learning,Learning interpretable disentangled representations is a crucial yet challenging task.,35730
a of takes the form Our dual autoencoder model structure.,Our model takes the form of a dual autoencoder structure.,35731
"dual be semi-supervised effective. setting, This out very to for swap mechanism, turns tailored","This dual swap mechanism, tailored for semi-supervised setting, turns out to be very effective.",35732
biometric introduced. security these privacy and address been concerns has cancelable To,To address these security and privacy concerns cancelable biometric has been introduced.,35733
"nearest ridge-based each sector and neighbor are computed. in The minutiae identified, are features","The nearest neighbor minutiae in each sector are identified, and ridge-based features are computed.",35734
"In stimulating experiments, consolidation be presynaptic neurons. synaptic by repeatedly induced can","In experiments, synaptic consolidation can be induced by repeatedly stimulating presynaptic neurons.",35735
- tool. method gold LCModel the is standard commercial quantification a The fitting current,The current gold standard quantification method is the LCModel - a commercial fitting tool.,35736
involve simulated Experiments testing brain and in-vivo on and spectra. training human,Experiments involve training and testing on simulated and in-vivo human brain spectra.,35737
improvement standard consistent classification image are reported datasets. experiments over demonstrate which a Preliminary,Preliminary experiments are reported which demonstrate a consistent improvement over standard image classification datasets.,35738
tasks is extremely Applying in reinforcement physical-world challenging. learning,Applying reinforcement learning in physical-world tasks is extremely challenging.,35739
disclose that results recovers The faithfully the Taobao of environment. important Virtual real properties,The results disclose that Virtual Taobao faithfully recovers important properties of the real environment.,35740
is boundary between relationship estimation What landmarks localisation? the and,What is the relationship between boundary estimation and landmarks localisation?,35741
other Our easily can method information datasets. from also integrate,Our method can also easily integrate information from other datasets.,35742
our selection and demonstrate various We in technique segmentation of keypoint effectiveness the applications.,We demonstrate the effectiveness of our technique in various segmentation and keypoint selection applications.,35743
an information enrich acquisitions Multi-contrast for available of MRI magnitude the anatomy diagnosis. of,Multi-contrast MRI acquisitions of an anatomy enrich the magnitude of information available for diagnosis.,35744
"factor. excessive associated be additional contrasts with may scan times Yet, limiting a","Yet, excessive scan times associated with additional contrasts may be a limiting factor.",35745
"diminished at higher factors acceleration decreases sampling particularly density performance towards reconstruction, In high-spatial-frequencies. with","In reconstruction, performance decreases towards higher acceleration factors with diminished sampling density particularly at high-spatial-frequencies.",35746
The improve multi-contrast of MRI quality proposed method and can the efficiency exams. scan help,The proposed method can help improve the quality and scan efficiency of multi-contrast MRI exams.,35747
correlated one-self to protect Physiological preparations biological danger. are with from changes these,Physiological changes are correlated with these biological preparations to protect one-self from danger.,35748
extend core We idea key the with contributions. two,We extend the core idea with two key contributions.,35749
"the show qualitative and results both segmentations of we inpaintings. Moreover,","Moreover, we show qualitative results of both the segmentations and inpaintings.",35750
summarization. This the addresses paper problem video of,This paper addresses the problem of video summarization.,35751
"In summarization paper, this a problem. labeling sequence we formulate as video","In this paper, we formulate video summarization as a sequence labeling problem.",35752
effectiveness Extensive benchmark the two of our on and models. analysis datasets demonstrate experiments,Extensive experiments and analysis on two benchmark datasets demonstrate the effectiveness of our models.,35753
"combining results. the our global prediction local features, we By the final and obtain","By combining the global and the local features, we obtain our final prediction results.",35754
"we After convolution layers, hashing and pooling. deploy for block-wise binary the histogram","After the convolution layers, we deploy binary hashing and block-wise histogram for pooling.",35755
is identification use for biometrics. known The as physical and behavioral human of characteristics,The use of physical and behavioral characteristics for human identification is known as biometrics.,35756
"is traits biometrics fingerprint Among used. most many the widely the available, the","Among the many biometrics traits available, the fingerprint is the most widely used.",35757
"also have due In addition, recognition focused on multibiometrics researches to many advantages. its recent","In addition, recent researches have also focused on multibiometrics recognition due to its many advantages.",35758
properly to to A understand kitchen the needs environment cooking cooking robot any activities. continue,A kitchen robot properly needs to understand the cooking environment to continue any cooking activities.,35759
state not so But like researched detection detection. as object's has well far object been,But object's state detection has not been researched well so far as like object detection.,35760
"of The model optimizers. terms layers, is robustly and analyzed different in","The model is analyzed robustly in terms of different layers, and optimizers.",35761
a results show qualitative progression state-of-the-art. clear model delivers our quantitative over and Both,Both quantitative and qualitative results show our model delivers a clear progression over state-of-the-art.,35762
of preventable irreversible the world. in blindness cause but the leading Glaucoma is,Glaucoma is the leading cause of irreversible but preventable blindness in the world.,35763
the in that observed recently glaucoma been ocular hemodynamics. has It changes induces,It has been recently observed that glaucoma induces changes in the ocular hemodynamics.,35764
characterizing approach first this propose hemodynamics. those paper for In changes using we computational a,In this paper we propose a first approach for characterizing those changes using computational hemodynamics.,35765
is from on patient-specific simulation images. The extracted fundus arterial trees performed,The simulation is performed on patient-specific arterial trees extracted from fundus images.,35766
computation. weight involved score MCW The the technique for matcher achieves in each the optimal,The MCW technique achieves the optimal weight for each matcher involved in the score computation.,35767
"DS to applied to final Further, decision. the induced the theory scores is output","Further, DS theory is applied to the induced scores to output the final decision.",35768
each A of study the performance shrinkage the evaluate priors. to conducted of is simulation,A simulation study is conducted to evaluate the performance of each of the shrinkage priors.,35769
illustration empirical data. pan-European regional we use an For growth economic,For an empirical illustration we use pan-European regional economic growth data.,35770
"yet to e.g. well-understood knowledge, not ontological However, what is extent it","However, it is not yet well-understood to what extent ontological knowledge, e.g.",35771
"be (existential) in given principled set a of can a as way. embedded rules,","given as a set of (existential) rules, can be embedded in a principled way.",35772
"convex relations model a are in modelled which regions. Second, we as study","Second, we study a model in which relations are modelled as convex regions.",35773
previously method improvements methods. the considerably ensures than proposed This accuracy discrimination,This method ensures considerably discrimination accuracy improvements than the previously proposed methods.,35774
regions. find Recent object instances while categorizing candidate detectors,Recent object detectors find instances while categorizing candidate regions.,35775
find in We meta-measures. improvements almost large all the,We find large improvements in almost all the meta-measures.,35776
Experimental evaluations method current to the the state-of-the-art. outperforms in that comparison indicate,Experimental evaluations indicate that the method outperforms in comparison to the current state-of-the-art.,35777
"shown favorable trackers (SVM) machines based performance. have vector Recently, part-based support and","Recently, part-based and support vector machines (SVM) based trackers have shown favorable performance.",35778
"training limit updating applications. the their and process online real-time time-consuming Nonetheless,","Nonetheless, the time-consuming online training and updating process limit their real-time applications.",35779
"In work, recursion to extend GLMB propagate posterior. multi-object analytic we this the the","In this work, we extend the analytic GLMB recursion to propagate the multi-object posterior.",35780
on mathematical models mass-action based focus deterministic we Here on kinetics.,Here we focus on mathematical models based on deterministic mass-action kinetics.,35781
positive equations These models of systems coupled nonlinear orthant. the on differential are,These models are systems of coupled nonlinear differential equations on the positive orthant.,35782
"be Recent accurate. surprisingly fast, can low-bitwidth that compact has neural work networks shown","Recent work has shown that fast, compact low-bitwidth neural networks can be surprisingly accurate.",35783
to experiments We verify effectiveness extensive of the method. the perform proposed,We perform extensive experiments to verify the effectiveness of the proposed method.,35784
SBU and RGB+D dataset. Kinect proposed Interaction NTU evaluate We on method our,We evaluate our proposed method on NTU RGB+D and SBU Kinect Interaction dataset.,35785
"method interactions. actions of improves addition, the accuracy our in significantly the two-person In","In addition, our method significantly improves the accuracy of the actions in two-person interactions.",35786
by signals measured and EEG Rating Scale scores. (HDRS) were Depression The responses Hamilton ketamine,The ketamine responses were measured by EEG signals and Hamilton Depression Rating Scale (HDRS) scores.,35787
EEG patterns may advance. ketamine The for baseline prefrontal recognising account effects at in,The prefrontal EEG patterns at baseline may account for recognising ketamine effects in advance.,35788
irregular their because trees mostly twisted are of branches. While other shape in,While other trees are irregular in shape mostly because of their twisted branches.,35789
The other self-shadows than leaves. palms ordinary have large,The palms have large self-shadows other than ordinary leaves.,35790
in shown palms in of the satellite Then image. field the any it number counts,Then it counts the number of palms in any field shown in the satellite image.,35791
Dataset). (UCSD Dataset on two Avenue approach our benchmarks We test and popular,We test our approach on two popular benchmarks (UCSD Dataset and Avenue Dataset).,35792
of potential networks zero-shot great task neural convolutional shown have for learning. the Graph recently,Graph convolutional neural networks have recently shown great potential for the task of zero-shot learning.,35793
connections node's descendants. on its are These and based relationship a ancestors to added,These connections are added based on a node's relationship to its ancestors and descendants.,35794
that constraints in specified research in operations the model. be require problems Many,Many problems in operations research require that constraints be specified in the model.,35795
is the Determining laborsome a right task. hard and constraints,Determining the right constraints is a hard and laborsome task.,35796
propose process machine learning an principles. to approach We automate this artificial intelligence and using,We propose an approach to automate this process using artificial intelligence and machine learning principles.,35797
"how adjustment to the However, degree never been determine has of discussed.","However, how to determine the degree of adjustment has never been discussed.",35798
thus testing structure. for market unique market organization a the The environment provides,The market thus provides a unique testing environment for the market organization structure.,35799
method random warping bounds data functional for for a constructing develop with variability. tolerance We,We develop a method for constructing tolerance bounds for functional data with random warping variability.,35800
functional Component to tolerance factor. bounds utilize Principal second a tolerance construct Analysis The functional,The second functional tolerance bounds utilize functional Principal Component Analysis to construct a tolerance factor.,35801
control by and is This applications: work disease main motivated process two monitoring.,This work is motivated by two main applications: process control and disease monitoring.,35802
"these identify desirable context, signals. to is this in In abnormalities it","In this context, it is desirable to identify abnormalities in these signals.",35803
U-Net. segmentation promising encoder-decoders as the approaches Recent using results such show convolutional,Recent approaches show promising segmentation results using convolutional encoder-decoders such as the U-Net.,35804
has the years. recognition over past Low-resolution few increasing (LRFR) attention received face,Low-resolution face recognition (LRFR) has received increasing attention over the past few years.,35805
demands for is video biggest technologies the of One LRFR surveillance.,One of the biggest demands for LRFR technologies is video surveillance.,35806
these are generally small in in Faces size. images,Faces in these images are generally small in size.,35807
"given. definition general First, is problem a","First, a general problem definition is given.",35808
"catogory. on this topic by of the works is systematically Later, analysis presented","Later, systematically analysis of the works on this topic is presented by catogory.",35809
"In methods, settings. to datasets we on experiment focus also and describing addition the","In addition to describing the methods, we also focus on datasets and experiment settings.",35810
"we and effort. summarized general a the Finally, future priorities speculate for the limitations","Finally, we summarized the general limitations and speculate a priorities for the future effort.",35811
RGB-D reconstruction a We novel method camera. moving a using non-rigid present,We present a novel non-rigid reconstruction method using a moving RGB-D camera.,35812
rigid use the background. of only approaches scene and completely non-rigid Current the ignore part,Current approaches use only non-rigid part of the scene and completely ignore the rigid background.,35813
large tracking sufficient lack information frame-to-frame for parts often geometric Non-rigid photometric and motion.,Non-rigid parts often lack sufficient geometric and photometric information for tracking large frame-to-frame motion.,35814
foreground pose camera approach from rigid background Our uses estimated the for tracking.,Our approach uses camera pose estimated from the rigid background for foreground tracking.,35815
robust occurs. enables frame-to-frame motion where large situations This in foreground tracking,This enables robust foreground tracking in situations where large frame-to-frame motion occurs.,35816
"is network neural termed The i.e. CocoNet,","The neural network is termed CocoNet, i.e.",35817
one a single image encodes only. network,one network encodes a single image only.,35818
general and a technique present artifact compression. removal We that performs both image,We present a general technique that performs both artifact removal and image compression.,35819
"removal, image try compression and a artifact its artifacts. remove JPEG we For to input","For artifact removal, we input a JPEG image and try to remove its compression artifacts.",35820
reuses and legacy technique compression routines. Our decompression JPEG's,Our technique reuses JPEG's legacy compression and decompression routines.,35821
paper of localization strives in actions videos. human for This spatio-temporal,This paper strives for spatio-temporal localization of human actions in videos.,35822
"cumbersome and we propose to As annotating video boxes error-prone, bypass expensive, in box-supervision. is","As annotating boxes in video is expensive, cumbersome and error-prone, we propose to bypass box-supervision.",35823
"action on Instead, we based point-supervision. localization introduce","Instead, we introduce action localization based on point-supervision.",35824
"unsupervised a We provide spatio-temporal which start proposals, candidate of videos. regions in set from","We start from unsupervised spatio-temporal proposals, which provide a set of candidate regions in videos.",35825
"randomly For MCNN. each of an layer activation selected is for a function applications, both","For both applications, an activation function is randomly selected for each layer of a MCNN.",35826
surveillance. is tumor consistent for detection reliable A fully necessary and meningioma tool for automated,A fully automated tool for meningioma detection is necessary for reliable and consistent tumor surveillance.,35827
studies concerning detection. been automated There have lesion various,There have been various studies concerning automated lesion detection.,35828
machine in networks learning performance neural various impressive demonstrated Deep have tasks.,Deep neural networks have demonstrated impressive performance in various machine learning tasks.,35829
"to in However, changes notoriously are distribution. data they sensitive","However, they are notoriously sensitive to changes in data distribution.",35830
"drastic can a slight distribution lead performance change even in the reduction. Often, to","Often, even a slight change in the distribution can lead to drastic performance reduction.",35831
that to standard important certain shows adapt fine-tuning in the Our work fails model cases.,Our work shows that standard fine-tuning fails to adapt the model in certain important cases.,35832
"learning propose dual-memory for this lifelong self-organizing scenarios. paper, architecture a In we","In this paper, we propose a dual-memory self-organizing architecture for lifelong learning scenarios.",35833
so as minimize brain The the that costs pervades certain to neuroscience. theoretical idea functions,The idea that the brain functions so as to minimize certain costs pervades theoretical neuroscience.,35834
"consider partial the often derivatives researchers vector practice, of In the as gradient.","In practice, researchers often consider the vector of partial derivatives as the gradient.",35835
contribute human Rice food one significantly supplies. that to staple the been has foods of,Rice has been one of the staple foods that contribute significantly to human food supplies.,35836
"been have exported worldwide. and rice cultivated, Numerous varieties imported,","Numerous rice varieties have been cultivated, imported, and exported worldwide.",35837
varieties mixed could rice Different during and production rice be trading.,Different rice varieties could be mixed during rice production and trading.,35838
data determined the CNN. spatio-spectral from then a using varieties acquired rice are The deep,The rice varieties are then determined from the acquired spatio-spectral data using a deep CNN.,35839
significant have detection a Pulmonary results later on the impact diagnosis. nodules,Pulmonary nodules detection results have a significant impact on the later diagnosis.,35840
"with traditional Furthermore, experimental the U-Net. the we compare results","Furthermore, we compare the experimental results with the traditional U-Net.",35841
implementation considerable a of discrete Therefore is satisfactory interest.,Therefore a satisfactory discrete implementation is of considerable interest.,35842
continuous that concrete both entities and have manifestations. discrete abstract Operators can are,Operators are abstract entities that can have both continuous and discrete concrete manifestations.,35843
"games, of the We for propose generation tutorial problem i.e.","We propose the problem of tutorial generation for games, i.e.",35844
"as play to tutorials problem. teach players an can generate which games, to AI","to generate tutorials which can teach players to play games, as an AI problem.",35845
model the observed is A Gaussian mixture for fitted data.,A Gaussian mixture model is fitted for the observed data.,35846
"update further i.e. to We resort bounds function value estimation, approximations to","We further resort to function approximations to update value bounds estimation, i.e.",35847
are learned actor gradient-based critic and parameters approaches. via Both,Both actor and critic parameters are learned via gradient-based approaches.,35848
are required Visual to today adaptive recognition exhibit algorithms abilities.,Visual recognition algorithms are required today to exhibit adaptive abilities.,35849
Image-to-image learning. deep translation has due advances recently in to significant received attention,Image-to-image translation has recently received significant attention due to advances in deep learning.,35850
essential to neural any the functioning system. is of Control,Control is essential to the functioning of any neural system.,35851
Evaluations SCFace datasets. on are conducted portions UCCSface of challenging and the,Evaluations are conducted on challenging portions of the SCFace and UCCSface datasets.,35852
carrier FBS tier for is sensing its and equipped transmissions. to The support NOMA,The FBS tier is equipped to support NOMA and carrier sensing for its transmissions.,35853
the The goals of two-fold. paper are,The goals of the paper are two-fold.,35854
"errors lexicon-free requires explicit in confidence filtering However, calculation. scenarios, an","However, in lexicon-free scenarios, filtering errors requires an explicit confidence calculation.",35855
of applications in signal reconstruction matrix The communications processing. arises in low-rank problem various and,The problem of low-rank matrix reconstruction arises in various applications in communications and signal processing.,35856
"this we work, problem. In technique optimisation a a based novel on propose","In this work, we propose a technique based on a novel optimisation problem.",35857
saliency a in page prediction challenge vision. problem is transformation Web and computer image,Web page saliency prediction is a challenge problem in image transformation and computer vision.,35858
part integral Video in become games have an of times. people's most lives recent,Video games have become an integral part of most people's lives in recent times.,35859
This being abundance of an led data online. to games shared to related video,This led to an abundance of data related to video games being shared online.,35860
recommendations. help powerful with providing that users by tools them are systems Recommendation meaningful,Recommendation systems are powerful tools that help users by providing them with meaningful recommendations.,35861
the on propose predict (unimodal). based G-Score the trailer to first models We alone,We first propose models to predict the G-Score based on the trailer alone (unimodal).,35862
"is greedy of hereby fast heuristic. usually construction The involving solutions a problem-specific, feasible","The construction of feasible solutions is hereby problem-specific, usually involving a fast greedy heuristic.",35863
GAPF (CRB). Carlo bound the the estimator simulations matches that Cramer-Rao Monte show performance,Monte Carlo simulations show that the GAPF estimator performance matches the Cramer-Rao bound (CRB).,35864
to an is also REM. used estimator The create,The estimator is also used to create an REM.,35865
during common biopsy Needle interventions minimally or as brachytherapy. such invasive insertion is,Needle insertion is common during minimally invasive interventions such as biopsy or brachytherapy.,35866
"Furthermore, scenario. ex-vivo our an needle prostate we in approach insertion validate","Furthermore, we validate our approach in an ex-vivo prostate needle insertion scenario.",35867
"detection. may contain some However, edge points of to imperfect the due outliers scattering","However, some of the scattering points may contain outliers due to imperfect edge detection.",35868
"applied. is optimization a LPNN to of problem, calculate this real-time Then, solution","Then, to calculate a real-time solution of this optimization problem, LPNN is applied.",35869
"problem, ambiguous captioning is for image. suitable an with Image many captions an","Image captioning is an ambiguous problem, with many suitable captions for an image.",35870
"facto for beam To captions. de the method is ambiguity, search multiple sampling address","To address ambiguity, beam search is the de facto method for sampling multiple captions.",35871
"However, captions. generic to known produce is search beam computationally expensive and","However, beam search is computationally expensive and known to produce generic captions.",35872
"VAE diverse, accurate. and GAN less are Though","Though diverse, GAN and VAE are less accurate.",35873
"drive as summary since generation. should use our part-of-speech We summaries, caption","We use part-of-speech as summaries, since our summary should drive caption generation.",35874
"a introduce probabilistic inference models. compilation, graphical for discrete algorithm approximate collapsed We novel","We introduce collapsed compilation, a novel approximate inference algorithm for discrete probabilistic graphical models.",35875
compilation shows that well evaluation on benchmarks. standard collapsed performs experimental Our,Our experimental evaluation shows that collapsed compilation performs well on standard benchmarks.,35876
the to model scheme the decompose the We develop mode-decomposition study OAM-modes. and channel,We study the channel model and develop the mode-decomposition scheme to decompose the OAM-modes.,35877
"the angles we model of impact the scenario. under non-coaxial channel on included Then, discuss","Then, we discuss the impact of included angles on the channel model under non-coaxial scenario.",35878
"the studies pairing performance of schemes device carried three been have MTC comparison out. Furthermore,","Furthermore, performance comparison studies of the three MTC device pairing schemes have been carried out.",35879
gradients target (e.g. the of method uses Our concept any,Our method uses the gradients of any target concept (e.g.,35880
convolutional flows the that into of a target class) layer. score,the score of target class) that flows into a convolutional layer.,35881
"""Inception"" efficiency. fuses classification and module features The as network adopts for representation multi-scale an","The classification network fuses multi-scale features as representation and adopts an ""Inception"" module for efficiency.",35882
evaluate experiments. proposed its extensive components TSR system with We and the,We evaluate the proposed TSR system and its components with extensive experiments.,35883
represents of countries. habits of attributes its food and different peoples the Diversity culinary from,Diversity of food and its attributes represents the culinary habits of peoples from different countries.,35884
proteins bacterial We model study biosynthesis of generic a in of stochastic operons.,We study a stochastic model of biosynthesis of proteins in generic bacterial operons.,35885
a as The process effect process commonly known a with leads to combined diffusion jumps.,The combined effect leads to a process commonly known as a diffusion process with jumps.,35886
for is of technique generation mobile millimeter a wave next communication. the The promising,The millimeter wave is a promising technique for the next generation of mobile communication.,35887
"channel the information the key is the accurate design. for precoding state However,","However, the accurate channel state information is the key for the precoding design.",35888
"stage, of first subspace in channel column matrix. we the the Specifically, the estimate","Specifically, in the first stage, we estimate the column subspace of the channel matrix.",35889
simultaneously the and Both network. to spectral the proposed are assigned information spatial,Both the spatial and spectral information are simultaneously assigned to the proposed network.,35890
focuses a distributed radar target system. on multiple-input-multiple-output (MIMO) widely paper localization This in,This paper focuses on target localization in a widely distributed multiple-input-multiple-output (MIMO) radar system.,35891
accurate We of the position estimated target by can errors. obtain measurement the minimizing an,We can obtain an accurate estimated position of the target by minimizing the measurement errors.,35892
"constraints. LPNN is it and function twice-differentiable objective However, that requires well known","However, it is well known that LPNN requires twice-differentiable objective function and constraints.",35893
"the second locally in While is method, algorithm competitive utilized.","While in the second method, locally competitive algorithm is utilized.",35894
prediction pairs. Existing with verification kinship kinship focuses on primarily image in research,Existing research in kinship verification primarily focuses on kinship prediction with image pairs.,35895
sparsity class-specific matrix. introduces This in new autoencoder weight the formulation,This new autoencoder formulation introduces class-specific sparsity in the weight matrix.,35896
Lip-reading. movement the recognition using is speech Visual of called lip,Visual recognition of speech using the lip movement is called Lip-reading.,35897
sentence can level end end also model same The for classification. be extended to,The same model can also be extended for end to end sentence level classification.,35898
"neural building convolutional this efficient paper, networks. and interested in are In we lightweight","In this paper, we are interested in building lightweight and efficient convolutional neural networks.",35899
infer is lineage from data the to alternative phenotype estimation. latent-variable An approach the by,An alternative approach is to infer the phenotype from the lineage data by latent-variable estimation.,35900
"statistical work, how this we estimations. distorts In the survivor clarify bias","In this work, we clarify how the survivor bias distorts statistical estimations.",35901
TNM the determinant and prognosis treatment. Predicting is stage of cancer breast major,Predicting TNM stage is the major determinant of breast cancer prognosis and treatment.,35902
detecting Pathologic commonly histological in (pN-stage) metastasis by slides. N-stage pathologists performed is,Pathologic N-stage (pN-stage) is commonly performed by pathologists detecting metastasis in histological slides.,35903
by is classifier. combining metastasis node CNN lymph patch-level pN-stage based and detector slide-level predicted,pN-stage is predicted by combining patch-level CNN based metastasis detector and slide-level lymph node classifier.,35904
denoising LF parallax propose on anisotropic a framework based We novel (APA). analysis,We propose a novel LF denoising framework based on anisotropic parallax analysis (APA).,35905
to reinforcement the video learn Humans play algorithms. significantly (RL) state-of-the-art than learning faster games,Humans learn to play video games significantly faster than the state-of-the-art reinforcement learning (RL) algorithms.,35906
"efficiency. we Inspired model-based leveraging sample investigate in by for two issues this, RL","Inspired by this, we investigate two issues in leveraging model-based RL for sample efficiency.",35907
suffers from output This ineffiency can discontinuities at result computational and boundaries. approach in,This approach suffers from computational ineffiency and can result in discontinuities at output boundaries.,35908
"substantially mitigates not does but avoid This limitations. stitching altogether, its","This does not avoid stitching altogether, but substantially mitigates its limitations.",35909
(overall approach competition. labeling our in entry performance) to the contributed This wining building INRIA,This approach contributed to our wining entry (overall performance) in the INRIA building labeling competition.,35910
are exploring visual of current on on Most focusing features. recognition researches emotion facial,Most of current researches on visual emotion recognition are focusing on exploring facial features.,35911
detecting our We test method the pelvic frontal task on hip of fractures from x-rays.,We test our method on the task of detecting hip fractures from frontal pelvic x-rays.,35912
of compression a BPG artifact removal method propose images. We compressed learned for post-processing by,We propose a method for learned compression artifact removal by post-processing of BPG compressed images.,35913
networks trained three different We of sizes.,We trained three networks of different sizes.,35914
We encoded images input different using with QP values. BPG,We encoded input images using BPG with different QP values.,35915
combination is an the integer posed best as problem. programming The of selection,The selection of the best combination is posed as an integer programming problem.,35916
imaging in diffraction-photons. problem This paper addresses the of presence the of,This paper addresses the problem of imaging in the presence of diffraction-photons.,35917
"experiment conducted demonstrate of a Additionally, designed proof-of-concept the to the scheme. is viability","Additionally, a proof-of-concept experiment is conducted to demonstrate the viability of the designed scheme.",35918
has dermoscopic analysis image rapid growth performance. in Automated diagnostic witnessed,Automated dermoscopic image analysis has witnessed rapid growth in diagnostic performance.,35919
"evidence faces decisions. part, in provided adoption support is because resistance, to no Yet","Yet adoption faces resistance, in part, because no evidence is provided to support decisions.",35920
"work, an is this for classification approach In evidence-based presented.","In this work, an approach for evidence-based classification is presented.",35921
"optimal parameters size, etc.) rate, (dilution choosing volume operational Moreover,","Moreover, choosing optimal operational parameters (dilution rate, volume size, etc.)",35922
with different learning domains. are generalize well concerned models \emph{unseen} to that We,We are concerned with learning models that generalize well to different \emph{unseen} domains.,35923
novel (GANs). unsupervised Networks approach Generative a introduces Adversarial using co-localization paper for This object,This paper introduces a novel approach for unsupervised object co-localization using Generative Adversarial Networks (GANs).,35924
consider problem the of summarization. video We,We consider the problem of video summarization.,35925
"such it very However, labeled to and difficult training expensive is examples. create","However, it is very expensive and difficult to create such labeled training examples.",35926
is this collect. type argue much that to We easier of data,We argue that this type of data is much easier to collect.,35927
integration. brain and successfully such Our segregation functional important ERGMs reproduced as pooled functional properties,Our pooled ERGMs successfully reproduced important brain properties such as functional segregation and functional integration.,35928
improve and robustness a with Our larger sample promising are size. in would results,Our results are promising and would improve in robustness with a larger sample size.,35929
"with confounding or variables, hidden (models with insufficient or etc.) latent Causally structures","Causally insufficient structures (models with latent or hidden variables, or with confounding etc.)",35930
developed the data. belief been from Algorithms network to acquire structure therefore have,Algorithms have been therefore developed to acquire the belief network structure from data.,35931
"already Regrettably, erroneous them are some (e.g. be to of known","Regrettably, some of them are known already to be erroneous (e.g.",35932
this described the for in paper. case,for the case described in this paper.,35933
this proposed. remedy is of A failure,A remedy of this failure is proposed.,35934
validate method through We computer simulations. the new,We validate the new method through computer simulations.,35935
"the considerably healthy Additionally, alters statistics network find we controls. of aripiprazole that","Additionally, we find that aripiprazole considerably alters the network statistics of healthy controls.",35936
"establish aripiprazole their test performance. of adversely we also cognitive a that Using ability, affects","Using a test of cognitive ability, we establish that aripiprazole also adversely affects their performance.",35937
evidence that network brain This measurable result differences. architecture macroscopic in to provides changes behavioural,This provides evidence that changes to macroscopic brain network architecture result in measurable behavioural differences.,35938
is This different assessed medications time in have been the first way. this,This is the first time different medications have been assessed in this way.,35939
averaging multiple the acquiring techniques to improve rely and Current signal acquisitions SNR. DT-CMR on,Current DT-CMR techniques rely on acquiring and averaging multiple signal acquisitions to improve the SNR.,35940
"It interaction, and has visually subtitling. in the video human-robot applications impaired helping","It has applications in human-robot interaction, helping the visually impaired and video subtitling.",35941
nor visual neither diversity of linguistic datasets complexity adequate Existing structures. contain,Existing datasets neither contain adequate visual diversity nor complexity of linguistic structures.,35942
"introduced. briefly representation which eight six Omega of and networks, languages are classes includes neural","Omega includes eight representation languages and six classes of neural networks, which are briefly introduced.",35943
ontologies systems. still ontology challenges serious state-of-the-art alignment Large pose to,Large ontologies still pose serious challenges to state-of-the-art ontology alignment systems.,35944
Evaluation evaluation have of datasets Alignment a conducted comprehensive We the Ontology using the Initiative.,We have conducted a comprehensive evaluation using the datasets of the Ontology Alignment Evaluation Initiative.,35945
"propose tracking, communications. For new we model a mmWave vehicular beam practical for","For beam tracking, we propose a new practical model for mmWave vehicular communications.",35946
"accurate more show produces that optic we traditional than algorithms. SOFAS results flow Finally,","Finally, we show that SOFAS produces more accurate results than traditional optic flow algorithms.",35947
"is the a efficient simple lower EE obtained. Furthermore, bound yet of","Furthermore, a simple yet efficient lower bound of the EE is obtained.",35948
"consists usually weeds in modern agriculture, In the over field. spraying all control herbicides agricultural","In modern agriculture, usually weeds control consists in spraying herbicides all over the agricultural field.",35949
This herbicide farmers and of environmental practice and pollution. waste cost significant for involves,This practice involves significant waste and cost of herbicide for farmers and environmental pollution.,35950
Learning impressive problem. different complex Recently shown approach in Deep classification has results,Recently Deep Learning approach has shown impressive results in different complex classification problem.,35951
method three main consists phases. in proposed The,The proposed method consists in three main phases.,35952
automatically identify interline to the using detect lines them and crop we First weeds. the,First we automatically detect the crop lines and using them to identify the interline weeds.,35953
"used In second interline phase, to constitute dataset. training are weeds the the","In the second phase, interline weeds are used to constitute the training dataset.",35954
comparable supervised obtained traditional results training are the labeling. data The to,The results obtained are comparable to the traditional supervised training data labeling.,35955
when produce divergence characterizes error. view keypoint Epipolar zero distributions reprojection two,Epipolar divergence characterizes when two view keypoint distributions produce zero reprojection error.,35956
undergone recognition has few adaptation visual progress the in for Domain past great years.,Domain adaptation for visual recognition has undergone great progress in the past few years.,35957
a nonmonotonic Programming based well-known on Answer is approach Set programs. solving problem (ASP) logic,Answer Set Programming (ASP) is a well-known problem solving approach based on nonmonotonic logic programs.,35958
"this answer inconsistent consider sets. without HEX-programs, In programs work and we ASP- i.e.,","In this work we consider inconsistent ASP- and HEX-programs, i.e., programs without answer sets.",35959
"two theoretical motivated by we This applications, concrete work which is also present.","This theoretical work is motivated by two concrete applications, which we also present.",35960
"for a algorithm a single is Typically, used. ground-truth single","Typically, for a single algorithm a single ground-truth is used.",35961
illuminant performance review evaluation algorithms. estimation of the range then a of We,We then review the performance evaluation of a range of illuminant estimation algorithms.,35962
images. paper  method two-stage proposes the This  hyperspectral novel a of classification for,  This paper proposes a novel two-stage method for the classification of hyperspectral images.,35963
"information. spatial propose to incorporate approach two-stage In a paper, this we the","In this paper, we propose a two-stage approach to incorporate the spatial information.",35964
probability will class for be map The resulting noisy. each,The resulting probability map for each class will be noisy.,35965
information utilizes effectively both data spatial hyperspectral method and sets. spectral of the proposed Our,Our proposed method effectively utilizes both spectral and spatial information of the hyperspectral data sets.,35966
"labelled in a Depth ""Relative depths. (RDIS) densely We dataset introduce new Stereo"" with relative","We introduce a new ""Relative Depth in Stereo"" (RDIS) dataset densely labelled with relative depths.",35967
dataset. RDIS pretrain ResNet on We model first our a,We first pretrain a ResNet model on our RDIS dataset.,35968
Then we ground-truth on depths. the datasets RGB-D finetune with model metric,Then we finetune the model on RGB-D datasets with metric ground-truth depths.,35969
"During we estimation depth classification task. our formulate a as finetuning,","During our finetuning, we formulate depth estimation as a classification task.",35970
"example-mining concept exploring Instead we of of hard constraint. strategy, large margin introduce the","Instead of exploring hard example-mining strategy, we introduce the concept of large margin constraint.",35971
two-center on a database organ for of task the We MR experiments perform segmentation.,We perform experiments on a two-center MR database for the task of organ segmentation.,35972
functioning fundamental Weber's describes of perceptual systems. the law,Weber's law describes the fundamental functioning of perceptual systems.,35973
pruning. and characteristic important is to an measure ultimately control with It therefore to,It is therefore an important characteristic to measure and ultimately to control with pruning.,35974
and point individual measured using represented clouds. by The of LiDAR trees was geometry,The geometry of individual trees was measured using LiDAR and represented by point clouds.,35975
was distribution of discrete sky energy A synthesised public the hemispherical model weather records. using,A discrete energy distribution model of the hemispherical sky was synthesised using public weather records.,35976
"is target domain unsupervised, any domain labels. using Our adaptation without procedure","Our domain adaptation procedure is unsupervised, without using any target domain labels.",35977
method public validated different chest on segmentation. datasets left/right for our X-ray We two lung,We validated our method on two different chest X-ray public datasets for left/right lung segmentation.,35978
semantic difficult computer been always segmentation task vision. has a Indoor in,Indoor semantic segmentation has always been a difficult task in computer vision.,35979
computationally for presents This paper methodology density estimation. efficient new kernel,This paper presents new methodology for computationally efficient kernel density estimation.,35980
compute density same exactly. estimates to be can The derivative used methodology,The same methodology can be used to compute density derivative estimates exactly.,35981
is an the size. Given linear computational the in complexity ordered sample sample,Given an ordered sample the computational complexity is linear in the sample size.,35982
methods in fast the approximation proposed with density results existing extremely Combining methodology estimation.,Combining the proposed methodology with existing approximation methods results in extremely fast density estimation.,35983
the compared Extensive documents with this approach state-of-the-art. experimentation existing efficiency the and effectiveness of,Extensive experimentation documents the effectiveness and efficiency of this approach compared with the existing state-of-the-art.,35984
"problem, use propose to translation image-to-image models. this we address To","To address this problem, we propose to use image-to-image translation models.",35985
image both unpaired explore the purpose. We translation of paired this for models usage and,We explore both the usage of paired and unpaired image translation models for this purpose.,35986
data Combining improvement. to leads further both sources,Combining both data sources leads to further improvement.,35987
selected is nearest The reference classifier used with the neighbour set.,The nearest neighbour classifier is used with the selected reference set.,35988
a model. as categorisation can a interpreted be set Such data-driven,Such a set can be interpreted as a data-driven categorisation model.,35989
cross-referencing. the two models to juxtapose We from the fields enable,We juxtapose the models from the two fields to enable cross-referencing.,35990
models (VQA) questions open-ended question language answering natural about respond images. to Visual,Visual question answering (VQA) models respond to open-ended natural language questions about images.,35991
buildings. of derived integrating GBI by is saliency the geometric measured The resulting,The resulting GBI is measured by integrating the derived geometric saliency of buildings.,35992
"degrade face eyeglasses accuracy. significantly recognition of In recognition, the could the application","In the application of face recognition, eyeglasses could significantly degrade the recognition accuracy.",35993
"we a In paper, propose this to synthesis address manner. virtual problem in this","In this paper, we propose to address this problem in a virtual synthesis manner.",35994
networks actual Convolutional are neural and (ConvNets) image standard classification. recognizement the for,Convolutional neural networks (ConvNets) are the actual standard for image recognizement and classification.,35995
"SqueezeNet, and AlexNet, models study on ConvNets Inception. platform: PyTorch ResNet uses the The available","The study uses ConvNets models available on the PyTorch platform: AlexNet, SqueezeNet, ResNet and Inception.",35996
goal. of advances approaching the Some recent that we demonstrate are,Some of the recent advances demonstrate we are approaching that goal.,35997
release research. also our promote We further to code,We also release our code to promote further research.,35998
rewards. find require agents random in Traditional actions to RL to perform methods exploration,Traditional exploration methods in RL require agents to perform random actions to find rewards.,35999
relevance Video most of the one prediction streaming service. tasks online important for is,Video relevance prediction is one of the most important tasks for online streaming service.,36000
"video approach i.e. to analyzing promising content One solve itself, it is","One promising approach to solve it is analyzing video content itself, i.e.",36001
"predicting by relevance and subtitle video metadata. audio, frame, video","predicting video relevance by video frame, audio, subtitle and metadata.",36002
experiments achieve accuracy with methods. traditional compared Our classification outstanding,Our experiments achieve outstanding classification accuracy compared with traditional methods.,36003
from measurements. is a undersampled method Compressive recover image to the sensing original,Compressive sensing is a method to recover the original image from undersampled measurements.,36004
methods to data. also They truth yielded ground with the performances comparable,They also yielded comparable performances to the methods with ground truth data.,36005
in show are modifications neglected some of required We often which examples implementations.,We show some examples of required modifications which are often neglected in implementations.,36006
"(CP) technique propose In Cube follows. contrast, simple as effective Padding a and we","In contrast, we propose a simple and effective Cube Padding (CP) technique as follows.",36007
"methods in experiments, outperforms In speed our method and baseline both quality.","In experiments, our method outperforms baseline methods in both speed and quality.",36008
"be the should cost cost saved lower CFA. using Moreover, by the associated than","Moreover, the associated cost should be lower than the cost saved by using CFA.",36009
"a Inspired sensing weakness, low scheme. strength we edge very and cost design HA's by","Inspired by HA's strength and weakness, we design a very low cost edge sensing scheme.",36010
"between directional of it Briefly, variations. functional a difference logistic guides by demosaicking the","Briefly, it guides demosaicking by a logistic functional of the difference between directional variations.",36011
theoretical experiments Carlo Monte the in setting dynamic to conducted finding. data support are panel,Monte Carlo experiments are conducted in dynamic panel data setting to support the theoretical finding.,36012
result that higher. suggests the be The may to returns schooling,The result suggests that the returns to schooling may be higher.,36013
surgical to outcomes costly patients Adverse and hospitals. are,Adverse surgical outcomes are costly to patients and hospitals.,36014
efficiency performance of and suggest a framework. results generation our high morphology The,The results suggest a high efficiency and performance of our morphology generation framework.,36015
Each standard backpropagation. learns through with a along specialization of its kernel location weights,Each kernel learns a location of specialization along with its weights through standard backpropagation.,36016
ROIs. differentiable optimize A supervision attention to is no the used requiring additional mechanism,A differentiable attention mechanism requiring no additional supervision is used to optimize the ROIs.,36017
synthetic domains. architecture both We natural tasks proposed multiple on and evaluate across our,We evaluate our proposed architecture on both synthetic and natural tasks across multiple domains.,36018
et. approach by extend (Xie (BNs) the learning We for networks Bayesian proposed decomposition,We extend the decomposition approach for learning Bayesian networks (BNs) proposed by (Xie et.,36019
"CGs), graphs special as which regression to multivariate (MVR chain BNs case. a learning include","to learning multivariate regression chain graphs (MVR CGs), which include BNs as a special case.",36020
"running except PC-like the algorithm the decomposition-based in outperforms fact, time. algorithm usually In","In fact, the decomposition-based algorithm usually outperforms the PC-like algorithm except in running time.",36021
when sparse. better The is is of underlying the much algorithms graph both performance,The performance of both algorithms is much better when the underlying graph is sparse.,36022
believe field. the be in We researchers instructive this for will directions potential these,We believe these potential directions will be instructive for the researchers in this field.,36023
aims transfer arbitrary to content Style styles to visual images. transfer,Style transfer aims to transfer arbitrary visual styles to content images.,36024
strategies appealing compare measure these We visually images. up to and compare they produce how,We compare these strategies and compare how they measure up to produce visually appealing images.,36025
adopt recurrence or based mainly methods Existing networks. convolution,Existing methods mainly adopt recurrence or convolution based networks.,36026
a supervised neuroanatomical We propose for network neural shapes. deep learning on,We propose a deep neural network for supervised learning on neuroanatomical shapes.,36027
unary and DCCRF consists The terms. pairwise of,The DCCRF consists of unary and pairwise terms.,36028
objects' unary based across displacements on tracked estimate appearance time visual The information. terms,The unary terms estimate tracked objects' displacements across time based on visual appearance information.,36029
visual use learning makes spectral from to range. a transfer It wider to possible,It makes possible to use transfer learning from visual to a wider spectral range.,36030
This real-time drone a paper areas. introduces individuals in violent to system public surveillance identify,This paper introduces a real-time drone surveillance system to identify violent individuals in public areas.,36031
Network images. from uses detect to first The aerial system Pyramid Feature humans the,The system first uses the Feature Pyramid Network to detect humans from aerial images.,36032
identification and pose is violent the with compared estimation The individuals state-of-the-art techniques. performance,The pose estimation and violent individuals identification performance is compared with the state-of-the-art techniques.,36033
lesions cystic classification for on certain achieved have Previous CT. in segmented accuracy computer-aided developments,Previous computer-aided developments have achieved certain accuracy for classification on segmented cystic lesions in CT.,36034
to data. We novel iterative to a image propose graph adapt d-dimensional a a method,We propose a novel iterative method to adapt a a graph to d-dimensional image data.,36035
the image The features. graph drives of the method nodes towards,The method drives the nodes of the graph towards image features.,36036
method fully is parallelisable. efficient proposed and The computationally,The proposed method is computationally efficient and fully parallelisable.,36037
behavior finite In sample the Monte-Carlo simulations analyze we of estimator. our,In Monte-Carlo simulations we analyze the finite sample behavior of our estimator.,36038
"health Connected security and is communication technologies. devices, medical technology devices that a associates","Connected health is a technology that associates medical devices, security devices and communication technologies.",36039
and be enables patients to home. It monitored remotely treated their from,It enables patients to be monitored and treated remotely from their home.,36040
detection. investigation division) A (cell step the fundamental mitotic is cell of,A fundamental step of the investigation is mitotic cell (cell division) detection.,36041
recognition. action complementary information key for temporal features are two human Spatial and and,Spatial and temporal features are two key and complementary information for human action recognition.,36042
feature recalibration. operations the and implement to are Squeeze Excitation The used,The Squeeze and Excitation operations are used to implement the feature recalibration.,36043
is The of proposed using the minimization energy optimization. performed stochastic,The minimization of the proposed energy is performed using stochastic optimization.,36044
of dataset order created the degraded to method. images realistically consisting is in evaluate A,A dataset consisting of realistically degraded images is created in order to evaluate the method.,36045
source information flow in of the is estimation. optical context Temporal coherence valuable of a,Temporal coherence is a valuable source of information in the context of optical flow estimation.,36046
"However, suitable a a motion leverage non-trivial finding task. model information is this to","However, finding a suitable motion model to leverage this information is a non-trivial task.",36047
usefulness demonstrate Experiments our approach. new the of,Experiments demonstrate the usefulness of our new approach.,36048
the search Similarity of of strategy speed final searching and accelerates accuracy the results. boosts,Similarity search strategy accelerates the speed of searching and boosts the accuracy of final results.,36049
"years, (DL) biomedical powerful In learning segmentation. recent have image become deep methods tools for","In recent years, deep learning (DL) methods have become powerful tools for biomedical image segmentation.",36050
inserted devices. supporting Catheters commonly are life,Catheters are commonly inserted life supporting devices.,36051
"results. deep learning, resurgence With begining are of the supervised training showing approaches to promising","With the resurgence of deep learning, supervised training approaches are begining to showing promising results.",36052
algorithms that suggests the Experimental RED evidence state-of-the-art. are,Experimental evidence suggests that the RED algorithms are state-of-the-art.,36053
"that the explicit does regularization We not algorithms. claim, explain RED however,","We claim, however, that explicit regularization does not explain the RED algorithms.",36054
"al. Furthermore, et interpret the algorithms Romano from we RED","Furthermore, we interpret the RED algorithms from Romano et al.",36055
algorithms new and acceleration guarantees. and convergence propose with,and propose new algorithms with acceleration and convergence guarantees.,36056
in analysis and variance models many regressors. tests with of restrictions of linear Applications include,Applications include analysis of variance and tests of linear restrictions in models with many regressors.,36057
RGB vs sketch near-infrared. vs or RGB,RGB vs sketch or RGB vs near-infrared.,36058
these of optimal. none We is scenarios two that observed,We observed that none of these two scenarios is optimal.,36059
monocular Automatically challenging determining from image three-dimensional RGB human is pose a data problem.,Automatically determining three-dimensional human pose from monocular RGB image data is a challenging problem.,36060
"maintaining coherent heatmaps whilst differentiability. visually produces model, MargiPose, Our resulting","Our resulting model, MargiPose, produces visually coherent heatmaps whilst maintaining differentiability.",36061
segmentation topic processing computer vision. a and in Image is image research fundamental,Image segmentation is a fundamental research topic in image processing and computer vision.,36062
"the of algorithms large decades, applications. for number developed various In a segmentation last researchers","In the last decades, researchers developed a large number of segmentation algorithms for various applications.",36063
undesirable results. to lead will segmentation This some flaw,This flaw will lead to some undesirable segmentation results.,36064
method can Our measure similarity the adaptively update estimating by parameters. function some,Our method can adaptively update the similarity measure function by estimating some parameters.,36065
computational cost greater of at have the networks power increased representational generally Larger complexity.,Larger networks generally have greater representational power at the cost of increased computational complexity.,36066
as all-in-focus can generated potentially images The reference be images. used,The generated images can potentially be used as reference all-in-focus images.,36067
the achieves dataset. competitive and state-of-the-art on on the results performance FVTA MovieQA MemexQA dataset,FVTA achieves state-of-the-art performance on the MemexQA dataset and competitive results on the MovieQA dataset.,36068
for temperatures calculation method A presented. engine is,A calculation method for engine temperatures is presented.,36069
behaviors thermal measuring of can be various simulated positions successfully. Different,Different thermal behaviors of various measuring positions can be simulated successfully.,36070
Previous jointly successful interactions. and model social physical been work to has not,Previous work has not been successful to jointly model physical and social interactions.,36071
"integration foundation the and Concepts learning, human understanding, of and knowledge deep transfer. are","Concepts are the foundation of human deep learning, understanding, and knowledge integration and transfer.",36072
stable behavior and is general is maximum learning How reached? once,How stable and general is behavior once maximum learning is reached?,36073
of Harris al. experiment the Using Pavlovian data et by and this another,Using the data of this and another Pavlovian experiment by Harris et al.,36074
the research personalized of empowering for is areas education. one key tracing Knowledge,Knowledge tracing is one of the key research areas for empowering personalized education.,36075
"problems our have in major extensive two model. experimentation, through However, the we DKT noticed","However, through our extensive experimentation, we have noticed two major problems in the DKT model.",36076
the The that the fails model reconstruct first is problem to input. observed,The first problem is that the model fails to reconstruct the observed input.,36077
"time-steps predicted consistent. KCs not the for across is performance Second,","Second, the predicted performance for KCs across time-steps is not consistent.",36078
time. and expected transit gradually to over performance is because is undesirable This student's unreasonable,This is undesirable and unreasonable because student's performance is expected to transit gradually over time.,36079
Two variables. sets and are provided: data invalid Combining examples instrumental,Two examples are provided: Combining data sets and invalid instrumental variables.,36080
"adaptive indirect An variations to is fuzzy parameter resistant proposed. controller, then","An indirect adaptive fuzzy controller, resistant to parameter variations is then proposed.",36081
structure of The first in specified the the controller is general stage. adaptive,The general structure of the adaptive controller is specified in the first stage.,36082
"of two second In fuzzy with its are stage, regulated aid the systems. parameters the","In the second stage, its parameters are regulated with the aid of two fuzzy systems.",36083
"approaches Finally, conventional compared. and results the of proposed are the the","Finally, the results of the proposed and the conventional approaches are compared.",36084
"configurations, For observed no or we contrast inducers' color the other induction.","For the other inducers' configurations, we observed color contrast or no induction.",36085
representation solution to tracking. a is visual Sparse viable,Sparse representation is a viable solution to visual tracking.,36086
"With system a limited power, task. becomes develop how a computing challenging robust to","With limited computing power, how to develop a robust system becomes a challenging task.",36087
"are small inflammation. and related bioactive species mediators critical, human and physiology Eicosanoids of","Eicosanoids and related species are critical, small bioactive mediators of human physiology and inflammation.",36088
"of diagnostic a making. essential for decision measure Furthermore, uncertainty is","Furthermore, a measure of uncertainty is essential for diagnostic decision making.",36089
"energy makes Mitochondrial currency phosphorylation oxidative (mOxPhos) ATP, of life. the","Mitochondrial oxidative phosphorylation (mOxPhos) makes ATP, the energy currency of life.",36090
"is so, channel new combination create crucial. merged In layers, to FMs are convolutional FMs,","In convolutional layers, FMs are merged to create new FMs, so, channel combination is crucial.",36091
"class. all not relevance FMs given However, the have a same for","However, not all FMs have the same relevance for a given class.",36092
"we brain evaluate segmentation, data. available using Finally, the in tumor proposed publicly methods","Finally, we evaluate the proposed methods in brain tumor segmentation, using publicly available data.",36093
Pavement that capacity. one road important is the most affects of the reasons cracks,Pavement cracks is one of the most important reasons that affects the road capacity.,36094
"important new Therefore, significance. has a detection designing method","Therefore, designing a new detection method has important significance.",36095
"Hausdorff evaluate is used distance At to detection precision. last,","At last, Hausdorff distance is used to evaluate detection precision.",36096
"However, of urban phenotypic remains on traits selection directional in empirical scarce. evidence areas","However, empirical evidence of directional selection on phenotypic traits in urban areas remains scarce.",36097
between survival intensities. difference different of colouration was local in adults No found,No difference in local survival was found between adults of different colouration intensities.,36098
critical rate in been deep issue network Learning training. scheduler has the neural a,Learning rate scheduler has been a critical issue in the deep neural network training.,36099
"(HTD). new method, proposes a This named paper hyperbolic-tangent decay scheduling","This paper proposes a new scheduling method, named hyperbolic-tangent decay (HTD).",36100
"modulation. secure for work, linear this investigate precoding In we spatial","In this work, we investigate linear precoding for secure spatial modulation.",36101
to optimum. but method much local The usually GD-based complexity converges a lower has,The GD-based method has much lower complexity but usually converges to a local optimum.,36102
secrecy even design Our achieves than SCA-based design. rate the existing higher a GD-based,Our SCA-based design even achieves a higher secrecy rate than the existing GD-based design.,36103
"the a book"" humans recognize How do action ""opening ?","How do humans recognize the action ""opening a book"" ?",36104
this Graph We representation Convolutional perform graph Networks. on reasoning via,We perform reasoning on this graph representation via Graph Convolutional Networks.,36105
Something-Something state-of-the-art results Charades datasets. on achieve We both and,We achieve state-of-the-art results on both Charades and Something-Something datasets.,36106
International become the biomedical standard challenges validation analysis image for methods. have of,International challenges have become the standard for validation of biomedical image analysis methods.,36107
"Humans knowledge novel generalize, applying situations prior readily stimuli. to and","Humans readily generalize, applying prior knowledge to novel situations and stimuli.",36108
for have a unified representation to acting planning. and,to have a unified representation for acting and planning.,36109
"past or work remained informal However, has subject on implementation-focused. the","However, past work on the subject has remained informal or implementation-focused.",36110
its the properties of planning. HTN We HTN relation to and acting study also,We also study the properties of HTN acting and its relation to HTN planning.,36111
"hand, approaches, scalable and algorithms. on Model-based other the models require","Model-based approaches, on the other hand, require models and scalable algorithms.",36112
Convolution execution of time Neural Convolutional overall (CNNs). Networks the operations dominate,Convolution operations dominate the overall execution time of Convolutional Neural Networks (CNNs).,36113
by The and operations and are Easy Convolution Random convolution Pooling conventional pooling (ECP). replaced,The conventional convolution and pooling operations are replaced by Easy Convolution and Random Pooling (ECP).,36114
"computations a the needed. As quarter of result, are a convolution conventional only","As a result, only a quarter of the conventional convolution computations are needed.",36115
a detection crucial It early of is where disease is importance.,It is a disease where early detection is of crucial importance.,36116
Colonoscopy precancerous early is used of for pathology. commonly detection cancer and,Colonoscopy is commonly used for early detection of cancer and precancerous pathology.,36117
are along encoders specific the encoder trained learning with Each of rate the decoder.,Each of the encoders are trained with encoder specific learning rate along the decoder.,36118
"technique system. makes implement easy our localized In a measurement millimeter-precision a to situation, it","In a localized situation, our technique makes it easy to implement a millimeter-precision measurement system.",36119
image (SISR). achieved have success remarkable in networks single super-resolution neural Deep,Deep neural networks have achieved remarkable success in single image super-resolution (SISR).,36120
"training joint importance problem. into Specifically, and network a optimization formulate learning the the we","Specifically, we formulate the network training and the importance learning into a joint optimization problem.",36121
"treatment identifies Under average heterogeneity, effect an instrument-specific the effect local treatment (LATE). instrument","Under treatment effect heterogeneity, an instrument identifies the instrument-specific local average treatment effect (LATE).",36122
a Graph-based machine role key in representations play learning.,Graph-based representations play a key role in machine learning.,36123
to process convolution systems Deep extensively data. input operations use learning,Deep learning systems extensively use convolution operations to process input data.,36124
techniques have for to convolutions developed conditions. restricted approximations Previous,Previous techniques have developed approximations to convolutions for restricted conditions.,36125
"applicability be their Unfortunately, general and is used limited clouds. cannot point for","Unfortunately, their applicability is limited and cannot be used for general point clouds.",36126
segmentation. a for propose novel vessel We system deep-learning-based,We propose a novel deep-learning-based system for vessel segmentation.,36127
"as They reduced flexibility. benefits such crew costs, promise increased and safety increased","They promise benefits such as reduced crew costs, increased safety and increased flexibility.",36128
"resulting We framework call the G\""odel design machine.","We call the resulting framework design G\""odel machine.",36129
solve novel problems machine by Such design could a a concepts. of generating variety,Such a machine could solve a variety of design problems by generating novel concepts.,36130
"it the are way change itself. addition, could generated In these by modifying concepts","In addition, it could change the way these concepts are generated by modifying itself.",36131
"new perceptual opinion. human is with PieAPP, The our metric, error estimated by well-correlated","The perceptual error estimated by our new metric, PieAPP, is well-correlated with human opinion.",36132
vital importance analysis retinopathy. Automatic tasks of is blood retinal of images diagnosis of in,Automatic analysis of retinal blood images is of vital importance in diagnosis tasks of retinopathy.,36133
a step is vessels Segmenting retinal fundamental images. analysing accurately in,Segmenting vessels accurately is a fundamental step in analysing retinal images.,36134
to U-net We connection detect utilize vessels. with residual,We utilize U-net with residual connection to detect vessels.,36135
"large is deep For required. a our training dataset network,","For training our deep network, a large dataset is required.",36136
highlights learning-based A proposed thorough classic approach. robustness the techniques comparison of the with and,A thorough comparison with classic and learning-based techniques highlights the robustness of the proposed approach.,36137
"to time. off it solution is necessary inference, against accuracy In often trade approximate","In approximate inference, it is often necessary to trade off accuracy against solution time.",36138
settings algorithm to find trade-off The optimal algorithm. run-time pre-processing for performs extended the,The trade-off algorithm performs pre-processing to find optimal run-time settings for the extended algorithm.,36139
task. a cluttered and crowded a every person's in environment action is challenging Recognizing,Recognizing every person's action in a crowded and cluttered environment is a challenging task.,36140
"this solve novel In methods we paper, propose this problem. to","In this paper, we propose novel methods to solve this problem.",36141
"accurate. that fast, proposed show the is reliable results experimental The and method","The experimental results show that the proposed method is fast, reliable and accurate.",36142
word task. is descriptions Generating from images challenging a,Generating images from word descriptions is a challenging task.,36143
of generate adversarial shown are be able realistic to Generative real-life objects. networks(GANs) images to,Generative adversarial networks(GANs) are shown to be able to generate realistic images of real-life objects.,36144
demonstrate approaches. results that other the produces our better proposed We model state-of-art surpassing,We demonstrate that our proposed model produces the better results surpassing other state-of-art approaches.,36145
a in parsing Face image basic is task analysis. face,Face parsing is a basic task in face image analysis.,36146
eyes such as amounts labeling pixel facial each to It and appropriate nose. with parts,It amounts to labeling each pixel with appropriate facial parts such as eyes and nose.,36147
scales. different (CNNs) input consists taking neural of networks convolutional It multiple in,It consists of multiple convolutional neural networks (CNNs) taking input in different scales.,36148
two-stage is parsing A for stages pipeline proposed use and iCNN. face both,A two-stage pipeline is proposed for face parsing and both stages use iCNN.,36149
have obtained dataset the state-of-the-art On results methods. better a benchmark we than,On a benchmark dataset we have obtained better results than the state-of-the-art methods.,36150
"data imaging step nonlinear effective toward Finally, harmonisation. normalisation intensity that is validate we a","Finally, we validate that nonlinear intensity normalisation is a step toward effective imaging data harmonisation.",36151
and validated real method synthetic using is images. vascular The,The method is validated using synthetic and real vascular images.,36152
"and a tree), report recall. the we precision accuracy, and extraction","a tree), and we report the extraction accuracy, precision and recall.",36153
method representation a such representation. a using requires and reasoning formal a PSAW,PSAW requires a formal representation and a reasoning method using such a representation.,36154
"MEBN developed. algorithms efficient been for reasoning Also, have","Also, efficient reasoning algorithms for MEBN have been developed.",36155
detection. object accurately This of formulation to while the hemorrhages us localize complexity bypassing allows,This formulation allows us to accurately localize hemorrhages while bypassing the complexity of object detection.,36156
also may medical segmentation apply These tasks. to other findings,These findings may also apply to other medical segmentation tasks.,36157
learning train. to data of for huge those is the key requirements processes One,One of the key requirements for those learning processes is huge data to train.,36158
pancreas variations. methods such failed with segmentation deal Previous to,Previous pancreas segmentation methods failed to deal with such variations.,36159
novel that automated fully segmentation. and We method segmentation propose localization pancreas a contains,We propose a fully automated pancreas segmentation method that contains novel localization and segmentation.,36160
include surrounding both global pancreas the organ and information. The features,The global features include both the pancreas and surrounding organ information.,36161
"is After localization, pancreas segmentation probabilistic patient-specific atlas-based a performed.","After localization, a patient-specific probabilistic atlas-based pancreas segmentation is performed.",36162
dataset IvS million We face conduct LBL more with a on identities. than two,We conduct LBL on a IvS face dataset with more than two million identities.,36163
and behaviors information different Petri interesting very Nets for studying tool of systems. simulating is,Petri Nets is very interesting tool for studying and simulating different behaviors of information systems.,36164
these Image techniques solve retrieval to have problems. potential great,Image retrieval techniques have great potential to solve these problems.,36165
of It images. firearm and can be retrieval for used fine-grained recognition both,It can be used for both fine-grained recognition and retrieval of firearm images.,36166
by deal margin the We propose to double a unbalanced loss with loss. contrastive employing,We propose to deal with the unbalanced loss by employing a double margin contrastive loss.,36167
among the of methods. one standards most used compression lossy commonly is JPEG image,JPEG is one of the most commonly used standards among lossy image compression methods.,36168
new removal. that the artifact state-of-the-art for DMCNN of show task JPEG a sets Experiments,Experiments show that DMCNN sets a new state-of-the-art for the task of JPEG artifact removal.,36169
"to deformations tracking during However, occurred observed large were colonoscope errors insertions. colon due","However, large tracking errors were observed due to colon deformations occurred during colonoscope insertions.",36170
estimation colon a deformation propose method. We,We propose a colon deformation estimation method.,36171
deformed used colon a is from estimate a This shape colonoscope method shape. to,This method is used to estimate a deformed colon shape from a colonoscope shape.,36172
use deformed shape. to colon the We forests a estimate regression algorithm,We use the regression forests algorithm to estimate a deformed colon shape.,36173
"As a the utilized preliminary colon estimate method a phantom. of study, to we deformations","As a preliminary study, we utilized the method to estimate deformations of a colon phantom.",36174
"shapes. method estimated experiments, In deformed colon the proposed correctly phantom our","In our experiments, the proposed method correctly estimated deformed colon phantom shapes.",36175
"examples object variability. well for With standard large class, train each inter-class of datasets","With large examples of each object class, standard datasets train well for inter-class variability.",36176
available ratio altitude (SINR). All DSCs improve signal-to-interference-plus-noise adjust their the to,All DSCs adjust their altitude to improve the available signal-to-interference-plus-noise ratio (SINR).,36177
optimal The distribution field controls mean the illustrate corresponding DSCs. power of and simulations,The simulations illustrate the optimal power controls and corresponding mean field distribution of DSCs.,36178
"a automating practice. improving very surgical step towards evaluation skills important Thus, is surgical","Thus, automating surgical skills evaluation is a very important step towards improving surgical practice.",36179
"recognizing great challenging. techniques, face recognition under remains Despite unconstrained persons success settings the of","Despite the great success of face recognition techniques, recognizing persons under unconstrained settings remains challenging.",36180
"and substantial occlusions unfavorable difficulties. cause lighting, like views, profile Issues can","Issues like profile views, unfavorable lighting, and occlusions can cause substantial difficulties.",36181
"have e.g. this exploiting problem works Previous by tackle to context, the attempted","Previous works have attempted to tackle this problem by exploiting the context, e.g.",36182
"combines skill-sets, differentiable. from forgetting, previous suffer does model and effectively Our not is fully","Our model effectively combines previous skill-sets, does not suffer from forgetting, and is fully differentiable.",36183
models approach presents learning This vehicles. novel autonomous for a for self-awareness paper,This paper presents a novel approach for learning self-awareness models for autonomous vehicles.,36184
"as modalities. used case, perception the localization are presented In visual and","In the presented case, visual perception and localization are used as modalities.",36185
"the a is make, task model, year Fine-grained vehicle vehicle. and of classifying of classification","Fine-grained vehicle classification is the task of classifying make, model, and year of a vehicle.",36186
case. of method proposed the show for Simulation optimality results a two-user the,Simulation results show the optimality of the proposed method for a two-user case.,36187
propose Bayesian a fields recursive RSS and time-varying method to crowdsourcing. by estimate We,We propose to estimate time-varying RSS fields by a recursive Bayesian method and crowdsourcing.,36188
parameters. present of also the estimated Cram\'er-Rao We Bayesian the (BCRB) bound,We also present the Bayesian Cram\'er-Rao bound (BCRB) of the estimated parameters.,36189
"results experimental the synthetic we our on method of data Finally, illustrate with performance sets.","Finally, we illustrate the performance of our method with experimental results on synthetic data sets.",36190
"With there is the security. of interchange growing increase a of data, in necessity","With the increase in interchange of data, there is a growing necessity of security.",36191
"forms Copy technique widespread possible, Forgery CMF. Move one Among many is the of tampering","Among the many forms of tampering possible, one widespread technique is Copy Move Forgery CMF.",36192
effectiveness demonstrate and quantitative The results obtained of qualitative the algorithm. this,The qualitative and quantitative results obtained demonstrate the effectiveness of this algorithm.,36193
Navigating find target and is operator-dependent. anatomy through to required view tedious the plane,Navigating through target anatomy to find the required view plane is tedious and operator-dependent.,36194
proposed that on Experiments videos. show state-of-the-art benchmark trackers tracker our outperforms various,Experiments show that our proposed tracker outperforms state-of-the-art trackers on various benchmark videos.,36195
different real The environments conditions. in can proposed of method perform,The proposed method can perform in real environments of different conditions.,36196
"approximation the the scheme Here, Euler-Maruyama standard is technique.","Here, the Euler-Maruyama scheme is the standard approximation technique.",36197
"is computationally However, the MCMC method expensive.","However, the MCMC method is computationally expensive.",36198
"remain and benefit it, the but accelerate higher-order unclear. specific implementation Using may approximations","Using higher-order approximations may accelerate it, but the specific implementation and benefit remain unclear.",36199
various deployed has widely been recognition applications. in Fingerprint-based,Fingerprint-based recognition has been widely deployed in various applications.,36200
Derived simulations. expressions with MATLAB verified are,Derived expressions are verified with MATLAB simulations.,36201
in the avoid A artifacts this checkerboard is paper. proposed condition to,A condition to avoid the checkerboard artifacts is proposed in this paper.,36202
robustness This to images. estimation correlation severely design inaccurate against boosts degraded due new,This new design boosts robustness against inaccurate correlation estimation due to severely degraded images.,36203
the whole existing This contrast practice deploys that in to is image.,This is in contrast to existing practice that deploys the whole image.,36204
tasks both and conducted. Extensive are on denoising image experiments super-resolution,Extensive experiments on both image denoising and super-resolution tasks are conducted.,36205
"we decay Second, weight and dropout. data contrast augmentation with","Second, we contrast data augmentation with weight decay and dropout.",36206
"In gains does provides and re-tuning. augmentation generalization large require systematically not data contrast, hyperparameter","In contrast, data augmentation systematically provides large generalization gains and does not require hyperparameter re-tuning.",36207
when operate first they operand the on itself matrix.,itself when they operate on the first operand matrix.,36208
the detection pixel-wise and segmentation. class-wise the It stem semantic learns jointly,It jointly learns the class-wise stem detection and the pixel-wise semantic segmentation.,36209
known analysis. in problem is alignment document a Historical widely manuscript,Historical manuscript alignment is a widely known problem in document analysis.,36210
manuscript is done mostly Finding editions manually. between differences the,Finding the differences between manuscript editions is mostly done manually.,36211
"capital per Our in employment, series time sector. worker, includes this value-added and data","Our time series data includes value-added per worker, employment, and capital in this sector.",36212
"applications, specific which usually costs, unknown. are on And depend the error","And the error costs, which depend on specific applications, are usually unknown.",36213
experiments classifier outperforms based demonstrate NR classifiers. representation (NRC) based the proposed that previous Our,Our experiments demonstrate that the proposed NR based classifier (NRC) outperforms previous representation based classifiers.,36214
"classification as features also inputs, on various deep achieves performance With state-of-the-art it tasks. visual","With deep features as inputs, it also achieves state-of-the-art performance on various visual classification tasks.",36215
of information purpose this The unclear. remains,The purpose of this information remains unclear.,36216
"performance. we (noise) Furthermore, how neuronal mechanism's affects crosstalk analyze the","Furthermore, we analyze how neuronal crosstalk (noise) affects the mechanism's performance.",36217
direction imperfect The performance. finding the array degrades,The imperfect array degrades the direction finding performance.,36218
through ratio Signal-to-noise computer are multi-numerology simulations. for results (SIR) obtained systems,Signal-to-noise ratio (SIR) results for multi-numerology systems are obtained through computer simulations.,36219
networks: proposed successive discriminator. consists The model generator and two of,The proposed model consists of two successive networks: generator and discriminator.,36220
"segmentation in less on a Moreover, than GPU. performed is image recent an second","Moreover, an image segmentation is performed in less than a second on recent GPU.",36221
processing. satellite in-demand imagery in is Semantic segmentation,Semantic segmentation is in-demand in satellite imagery processing.,36222
"has branches architecture input two the and functions. loss networks, conventional two proposed Unlike","Unlike conventional networks, the proposed architecture has two input branches and two loss functions.",36223
can show for that We this image single the classification. improve performance,We show that this can improve the performance for single image classification.,36224
how we neural But with the model can input net train size? different,But how can we train the neural net model with different input size?,36225
of input accept different audio. text size can like and,can accept different size of input like text and audio.,36226
In a new Attention propose network this Network(AIN). Incorporate called structure paper we,In this paper we propose a new network structure called Attention Incorporate Network(AIN).,36227
"and operation multi-rate these of (e.g. minimal Leveraging perform amount pipelining computation models a clocks,","Leveraging operation pipelining and multi-rate clocks, these models perform a minimal amount of computation (e.g.",36228
each an per for as four as to frame convolutional output. timestep few produce layers),as few as four convolutional layers) for each frame per timestep to produce an output.,36229
"of estimating liberalization. trade is, cost this goal of paper Another the","Another goal of this paper is, estimating the cost of trade liberalization.",36230
"nuisance VAT used tariff, reforming policy a scenarios. For is in both","For reforming nuisance tariff, a VAT policy is used in both scenarios.",36231
"this used. In TRIST method is study,","In this study, TRIST method is used.",36232
of mirroring shortcut Typical hourglass between networks layers. contain instances connections,Typical instances of hourglass networks contain shortcut connections between mirroring layers.,36233
applicable The hourglass approach proposed is any architecture. to,The proposed approach is applicable to any hourglass architecture.,36234
representation two ways. We the of utility this in demonstrate,We demonstrate the utility of this representation in two ways.,36235
"clustering performing discover by space. automatically First, we embedding the in learned can objects","First, we can automatically discover objects by performing clustering in the learned embedding space.",36236
cluster of and one contains various resulting examples scales. instance seen from Each viewpoints,Each resulting cluster contains examples of one instance seen from various viewpoints and scales.,36237
used of estimation Filter(PF) is target Particle and state. Non-gaussian Non-linear extensively for,Particle Filter(PF) is used extensively for estimation of target Non-linear and Non-gaussian state.,36238
"of However, sample suffers degeneracy problem to performance inherent due and impoverishment. its","However, its performance suffers due to inherent problem of sample degeneracy and impoverishment.",36239
convergence proposed framework. reliability with achieve PF of outlier transductive Proposed mechanism faster detection tracking,Proposed outlier detection mechanism with transductive reliability achieve faster convergence of proposed PF tracking framework.,36240
These results great . demonstrated application has for method that potential proposed have our clinical,These results have demonstrated that our proposed method has great potential for clinical application .,36241
the possibility This receiver. any implementing of method opens existing in the,This opens the possibility of implementing the method in any existing receiver.,36242
IF robust more hybrid receiver errors. to systematic is an with A,A receiver with an IF hybrid is more robust to systematic errors.,36243
method its operational under telescope. any is use normal The robust at allowing conditions,The method is robust allowing its use under normal operational conditions at any telescope.,36244
demonstrate also is full We analog receiver a errors. against robust more that systematic,We also demonstrate that a full analog receiver is more robust against systematic errors.,36245
rodent recordings. from recent findings This consistent finding with is,This finding is consistent with recent findings from rodent recordings.,36246
Faster/Mask via R-CNN. this on achieve We a framework based,We achieve this via a framework based on Faster/Mask R-CNN.,36247
are GP on of effective frameworks range a simulated highly and benchmarks. Current real,Current GP frameworks are highly effective on a range of real and simulated benchmarks.,36248
observa tions of diverse growing generating availability nature. and a of huge is Technology,Technology is generating a huge and growing availability of observa tions of diverse nature.,36249
learning as is central data data This big discipline. scientific a placing,This big data is placing data learning as a central scientific discipline.,36250
"of statistical It batches preprocessing, data. includes collection, visualization enormous of and, storage, essentially, analysis","It includes collection, storage, preprocessing, visualization and, essentially, statistical analysis of enormous batches of data.",36251
a visual answering novel We model propose probabilistic question for QA). (Visual,We propose a novel probabilistic model for visual question answering (Visual QA).,36252
of role learning plays the systems. performance for a critical availability Data deep,Data availability plays a critical role for the performance of deep learning systems.,36253
state-of-art. to method effectiveness the of the demonstrate compared Qualitative our results,Qualitative results demonstrate the effectiveness of our method compared to the state-of-art.,36254
of agrochemicals important towards sustainable the use component Reducing is an agriculture.,Reducing the use of agrochemicals is an important component towards sustainable agriculture.,36255
This and description processes. of focusses numerical developmental on simulation mathematical chapter the,This chapter focusses on the mathematical description and numerical simulation of developmental processes.,36256
us underlying a with provides tissue to non-invasive micro-structure. method analyze Tractography,Tractography provides us with a non-invasive method to analyze underlying tissue micro-structure.,36257
Gold We performance our and network with Standard parameters. benchmark its several evaluate across,We evaluate our network with Gold Standard and benchmark its performance across several parameters.,36258
"with time offset can every control. a Besides, desirable be implemented closed-loop channel of a","Besides, a desirable time offset of every channel can be implemented with a closed-loop control.",36259
increasing interests. verification have speaker received systems End-to-end,End-to-end speaker verification systems have received increasing interests.,36260
this contributions two work. We in make,We make two contributions in this work.,36261
improvements of of Experiments conducted dataset are on the and demonstrate $Fisher$ A-softmax. part using,Experiments are conducted on part of $Fisher$ dataset and demonstrate the improvements of using A-softmax.,36262
crucial is occlusion computer boundary in a Object vision. detection and research fundamental problem,Object occlusion boundary detection is a fundamental and crucial research problem in computer vision.,36263
(Visual investigate visual QA). adaptation cross-dataset problem question for of We the answering,We investigate the problem of cross-dataset adaptation for visual question answering (Visual QA).,36264
algorithm. adaptation We this proposing by novel difficulty domain overcome a,We overcome this difficulty by proposing a novel domain adaptation algorithm.,36265
qualitatively can and analyze when effective. We adaptation both mostly the be quantitatively,We both quantitatively and qualitatively analyze when the adaptation can be mostly effective.,36266
data and is solely proposed trained end-to-end The architecture relies for inference. raw perfusion on,The proposed architecture is trained end-to-end and solely relies on raw perfusion data for inference.,36267
the of demonstrate on experiments hyperspectral proposed large dataset Extensive effectiveness method. a the,Extensive experiments on a large hyperspectral dataset demonstrate the effectiveness of the proposed method.,36268
and promising applicability Previous real-world works problems. to control its for planning results showed,Previous works showed promising results for its applicability to real-world planning and control problems.,36269
"we propose prove solving For algorithms corresponding their and novel both constructs, correctness.","For both constructs, we propose corresponding novel solving algorithms and prove their correctness.",36270
challenging is because complex task of in a industrial distorted It setting. data various,It is a challenging task because of various distorted data in complex industrial setting.,36271
high performance DBCF demonstrate Extensive experiments exceeds state-of-the-art the methods. that produces and,Extensive experiments demonstrate that DBCF produces high performance and exceeds the state-of-the-art methods.,36272
from objects. plausible results The can predictor that generate lighting indicate even diffuse estimations the,The results indicate that the predictor can generate plausible lighting estimations even from diffuse objects.,36273
in can problem this interference alleviate systems. the intrinsic Using system FBMC we,Using this system we can alleviate the intrinsic interference problem in FBMC systems.,36274
(BER) results dispersive In our in ratio DP-FBMC performance channels. simulation bit error better has,In our simulation results DP-FBMC has better bit error ratio (BER) performance in dispersive channels.,36275
two of alleles. class natural models selection general define between a We representing,We define a general class of models representing natural selection between two alleles.,36276
"but and population spatial size The are arbitrary, fixed. structure","The population size and spatial structure are arbitrary, but fixed.",36277
"be can asexual diploid, sexual. otherwise; haploid, be Genetics can or or reproduction","Genetics can be haploid, diploid, or otherwise; reproduction can be asexual or sexual.",36278
"deaths, (e.g.~births, in events Biological on fashion depend the mating, state. population current arbitrary dispersal)","Biological events (e.g.~births, deaths, mating, dispersal) depend in arbitrary fashion on the current population state.",36279
on the Our is idea formalism sites. based of genetic,Our formalism is based on the idea of genetic sites.,36280
at and genetic single Each site resides a locus houses particular allele. a,Each genetic site resides at a particular locus and houses a single allele.,36281
of formalize the selection. method We weak then,We then formalize the method of weak selection.,36282
two develop filtering linear algorithms model systems. new is to Gaussian This exploited conditionally for,This model is exploited to develop two new filtering algorithms for conditionally linear Gaussian systems.,36283
here extraction. and is One automatic problems road precise the of,One of the problems here is automatic and precise road extraction.,36284
incoming traffic priority instant for The higher perpetually scheduling. URLLC scheduler proposed grants a,The proposed scheduler perpetually grants incoming URLLC traffic a higher priority for instant scheduling.,36285
presented and level system simulations extensive analysis to support conclusions. are Analytical paper,Analytical analysis and extensive system level simulations are presented to support paper conclusions.,36286
"paper, with on data classification manifolds. the tasks In this we SPD training limited tackle","In this paper, we tackle the classification tasks with limited training data on SPD manifolds.",36287
facial We two and evidence are finding present distinct faces look-alikes tasks. that recognizing,We present evidence that finding facial look-alikes and recognizing faces are two distinct tasks.,36288
in a under cafeteria conditions. normal Data were collected eating,Data were collected in a cafeteria under normal eating conditions.,36289
choices not or were scripted restricted. beverage Food and,Food and beverage choices were not scripted or restricted.,36290
Unit in the Facial detection the topic The (FAU) of explores paper wild. Action,The paper explores the topic of Facial Action Unit (FAU) detection in the wild.,36291
The on datasets. are and EmotionNet experiments DISFA the performed,The experiments are performed on the EmotionNet and DISFA datasets.,36292
experiments The usefulness facial for expression information AU show detection. the of,The experiments show the usefulness of facial expression information for AU detection.,36293
databases. achieves two proposed the state-of-art on results The network,The proposed network achieves state-of-art results on the two databases.,36294
results database cross of of the network. the Analysis for shows protocol effectiveness the,Analysis of the results for cross database protocol shows the effectiveness of the network.,36295
type of originating a is from the ventricles. Premature contraction(PVC) beat premature ectopic ventricular,Premature ventricular contraction(PVC) is a type of premature ectopic beat originating from the ventricles.,36296
monitoring valuable echocardiography imaging (TEE) diagnostic modality. is a and Transoesophageal,Transoesophageal echocardiography (TEE) is a valuable diagnostic and monitoring imaging modality.,36297
"scoring and different were criteria-based a Two used. general strategies, percentage impression, overall an","Two different scoring strategies, a criteria-based percentage and an overall general impression, were used.",36298
a method structure. common searching of for used is molecular Similarity identification,Similarity searching is a common method used for identification of molecular structure.,36299
the Extensive proposed datasets on demonstrate experiments real-world of effectiveness method. the various,Extensive experiments on various real-world datasets demonstrate the effectiveness of the proposed method.,36300
"data meet a serious with model movie vision would difficulties. However, from learning","However, learning a vision model from movie data would meet with serious difficulties.",36301
"from movies. In we explore learning vision alternative this an to approach models paper,","In this paper, we explore an alternative approach to learning vision models from movies.",36302
"temporal module. of a we module framework and analysis consider visual comprised Specifically, a a","Specifically, we consider a framework comprised of a visual module and a temporal analysis module.",36303
proven data sensing promising The is compressed have methods that aggregation existing data for aggregation.,The existing data aggregation methods have proven that compressed sensing is promising for data aggregation.,36304
"efficiency fidelity However, on-node they require or compromise computations. either and energy recovery on complex","However, they compromise either on energy efficiency or recovery fidelity and require complex on-node computations.",36305
"sailing we can find through the prediction. berthing direction ship Additionally, and of","Additionally, we can find the berthing and sailing direction of ship through prediction.",36306
"number addition, describe of propose In viewers' in behavior VR. to attentional a we metrics","In addition, we propose a number of metrics to describe viewers' attentional behavior in VR.",36307
brain a Lesion Resonance remains task. Magnetic detection in (MRI) challenging Images,Lesion detection in brain Magnetic Resonance Images (MRI) remains a challenging task.,36308
on annotated learning State-of-the-art large mostly datasets. making use of are based supervised approaches,State-of-the-art approaches are mostly based on supervised learning making use of large annotated datasets.,36309
provides a of learning. importance This to study MI the mechanism,This provides a mechanism to study the importance of MI learning.,36310
to are a synthesized classifier. The samples used train then,The synthesized samples are then used to train a classifier.,36311
Upon acceptance available. will made code be,Upon acceptance code will be made available.,36312
influenced that performance substrate way has antenna The the the of antenna. the,The way that the antenna substrate has influenced the performance of the antenna.,36313
of key component A fusion is image registration. MR-TRUS,A key component of MR-TRUS fusion is image registration.,36314
"is a metric smooth order optimization. approach in to for used the Further, multi-pass","Further, a multi-pass approach is used in order to smooth the metric for optimization.",36315
large registration indicate that results range. the overall a capture has framework The,The results indicate that the overall registration framework has a large capture range.,36316
"detect trained localize Finally, the and target types. to we a tissue model","Finally, we trained a model to detect and localize the target tissue types.",36317
"than approximately convnets, the original descent. Like faster existing solves much ours gradient the problem","Like existing convnets, ours approximately solves the original problem much faster than the gradient descent.",36318
descent our can gradient network. and on-the-fly in These reported be modifications affect directly the,These modifications affect directly the gradient descent and can be reported on-the-fly in our network.,36319
stems These modifications unrolling gradient from as the descent. are possible architecture the proposed,These modifications are possible as the proposed architecture stems from unrolling the gradient descent.,36320
task hierarchy and approach reinforcement We new decomposition introduce formation hierarchical learning. in a to,We introduce a new approach to hierarchy formation and task decomposition in hierarchical reinforcement learning.,36321
of HAMs. The environment in changes leads internal action the this to hierarchy,The internal action in this environment leads to changes the hierarchy of HAMs.,36322
method. the Preliminary the experiments of prospects demonstrated,Preliminary experiments demonstrated the prospects of the method.,36323
anatomical of frameworks. as in Cytoarchitectonic brain references human multimodal parcellations the atlas serve,Cytoarchitectonic parcellations of the human brain serve as anatomical references in multimodal atlas frameworks.,36324
"a results segmentations. these networks better Compared from fine-tuning to significantly in initialization, random","Compared to a random initialization, fine-tuning from these networks results in significantly better segmentations.",36325
work be Our into divided two contributions. can,Our work can be divided into two contributions.,36326
intricate machine. scene an a classification problem for is Acoustic,Acoustic scene classification is an intricate problem for a machine.,36327
"convincing (CNN) Convolutional emerging research, of deep Networks achieve Neural field an As results.","As an emerging field of research, deep Convolutional Neural Networks (CNN) achieve convincing results.",36328
with loss. hearing treat neural which Cochlear to used implants patients are are prosthetics (CIs),Cochlear implants (CIs) are neural prosthetics which are used to treat patients with hearing loss.,36329
"programmed. CIs need to surgery, After be","After surgery, CIs need to be programmed.",36330
and apply suppression positive false shape-based We strategy a in a constraint. incorporate training,We incorporate a false positive suppression strategy in training and apply a shape-based constraint.,36331
and Mixture functions. OWA They both: generalize,They generalize both: OWA and Mixture functions.,36332
"H, simple A a and provided in generalized example. special mixture applied operator, toy is","A special generalized mixture operator, H, is provided and applied in a simple toy example.",36333
"depth Furthermore, this per-pixel novel depth. to a recover introduces dense parameterization work","Furthermore, this work introduces a novel depth parameterization to recover dense per-pixel depth.",36334
is also maps basis training. generator depth via end-to-end The learned,The basis depth maps generator is also learned via end-to-end training.,36335
whole system combines (i.e. nicely domain The knowledge,The whole system nicely combines domain knowledge (i.e.,36336
multi-view deep geometry learning (i.e. constraints) hard-coded and,hard-coded multi-view geometry constraints) and deep learning (i.e.,36337
learning depth address maps problem. challenging dense to basis feature SfM and learning) the,feature learning and basis depth maps learning) to address the challenging dense SfM problem.,36338
real of success the proposed prove method. Experiments data large scale on the,Experiments on large scale real data prove the success of the proposed method.,36339
method second video compared first magnification Eulerian based is to algorithms. motion order and Our,Our method is compared to first and second order motion based Eulerian video magnification algorithms.,36340
amounts distinct training of a training has when advantage also with data. Our approach limited,Our approach also has a distinct advantage when training with limited amounts of training data.,36341
achieved computer Neural field. success Convolutional vision Deep (CNN) Networks significant in has,Deep Convolutional Neural Networks (CNN) has achieved significant success in computer vision field.,36342
another understood machine detect has Can concept? if a to we train machine a,Can we train a machine to detect if another machine has understood a concept?,36343
"this concept. of by principle, is tests on In the that subject conducting possible","In principle, this is possible by conducting tests on the subject of that concept.",36344
be we questions. to this avoiding However want by direct done procedure,However we want this procedure to be done by avoiding direct questions.,36345
mechanical material. bulk was implemented of characterizing the A hyperelastic for properties fiber-reinforced model (Holzapfel-Gasser-Ogden),A fiber-reinforced hyperelastic model (Holzapfel-Gasser-Ogden) was implemented for characterizing the mechanical properties of bulk material.,36346
tearing failure Results the this fibrous cap of processes. tissue study demonstrate,Results of this study demonstrate the fibrous cap tissue tearing failure processes.,36347
a train of location We cGAN given generate overhead views imagery. a to ground-level,We train a cGAN to generate ground-level views of a location given overhead imagery.,36348
"outcome rCBV) MTT, lesion to Tmax, rCBF, TTP, perform prediction.","Tmax, TTP, MTT, rCBF, rCBV) to perform lesion outcome prediction.",36349
counterparts the of lacking address at price this Online a limitation teacher. highcapacity,Online counterparts address this limitation at the price of lacking a highcapacity teacher.,36350
"present strategy an this work, (ONE) for In On-the-fly distillation. one-stage Ensemble online we Native","In this work, we present an On-the-fly Native Ensemble (ONE) strategy for one-stage online distillation.",36351
techniques different have measuring few last in been the appearance Many material years. for proposed,Many different techniques for measuring material appearance have been proposed in the last few years.,36352
"which appearance modeling. public These produced have large been have data-driven accurate, datasets, used for","These have produced large public datasets, which have been used for accurate, data-driven appearance modeling.",36353
appearance. excellent predictors We that the functionals of our of attributes perceived show are,We show that our functionals are excellent predictors of the perceived attributes of appearance.,36354
"functionals derive classic Moreover, used representations. to applicable analytic to our BRDF methodology be can","Moreover, our methodology can be used to derive functionals applicable to classic analytic BRDF representations.",36355
"of must capable on-line including single/few trials. It be learning, with","It must be capable of on-line learning, including with single/few trials.",36356
"over be safe catastrophic forgetting. from must Memories/knowledge durations, permanent lifelong","Memories/knowledge must be permanent over lifelong durations, safe from catastrophic forgetting.",36357
"retrieval time. over may i.e., semantically gradually confabulation, errors, accumulate plausible Some","Some confabulation, i.e., semantically plausible retrieval errors, may gradually accumulate over time.",36358
all computational shown have these has artificial to system been No properties.,No artificial computational system has been shown to have all these properties.,36359
We classification by networks. localization produced initial from start,We start from initial localization produced by classification networks.,36360
of and object more localization masks object. accurate more provide contain regions These,These object masks provide more accurate localization and contain more regions of object.,36361
fine produce processes progressively networks. segmentation These masks to conducted and optimize are object iteratively,These processes are conducted iteratively to progressively produce fine object masks and optimize segmentation networks.,36362
validation These samples issues protocols processes the include generation the used. and,These issues include the samples generation processes and the validation protocols used.,36363
self-supervised method a adversarial introduce on learning We based training. novel,We introduce a novel self-supervised learning method based on adversarial training.,36364
network by inpainting more helps generate images dropped the realistic feature entries. The repair,The repair network helps generate more realistic images by inpainting the dropped feature entries.,36365
separating counting track the major of is automatic overlapped One using tracks. photomicrographs of difficulties,One of the major difficulties of automatic track counting using photomicrographs is separating overlapped tracks.,36366
Image recent style broad attention transfer drawn has years. in,Image style transfer has drawn broad attention in recent years.,36367
to of framework merit new such its is and styles contents. generalizability key A,A key merit of such framework is its generalizability to new styles and contents.,36368
"The decoder. encoder, content of consists overall mixer and framework style encoder,","The overall framework consists of style encoder, content encoder, mixer and decoder.",36369
results effectiveness proposed have the methods. Extensive robustness and demonstrated experimental the of,Extensive experimental results have demonstrated the effectiveness and robustness of the proposed methods.,36370
Villegas hierarchical video et The method by prediction al.,The hierarchical video prediction method by Villegas et al.,36371
target radar (GPR) detection a classification is task. and Ground challenging penetrating,Ground penetrating radar (GPR) target detection and classification is a challenging task.,36372
efficient are obtain strategies Model-based to for learning. control sample critical,Model-based strategies for control are critical to obtain sample efficient learning.,36373
explored elegant been in strategy the mostly has tabular setting. planning This,This elegant planning strategy has been mostly explored in the tabular setting.,36374
model Replay afforded by first the Experience highlight a over flexibility We (ER).,We first highlight the flexibility afforded by a model over Experience Replay (ER).,36375
also on non-local-based recognition new for We accuracy. propose models further improvement the,We also propose new non-local-based models for further improvement on the recognition accuracy.,36376
test us earn will be they that called the say scores Let IQ. at the,Let us say that the scores they earn at the test will be called IQ.,36377
"ear we both appearance-based and features for study, utilized the geometric In have representation. features","In the study, we have utilized both geometric features and appearance-based features for ear representation.",36378
representation and classification. networks The methods convolutional employ neural deep for appearance-based,The appearance-based methods employ deep convolutional neural networks for representation and classification.,36379
"This have a we adaptation. performed domain way,","This way, we have performed a domain adaptation.",36380
over introduce regular for grids. lossy We data novel a compression multidimensional algorithm,We introduce a novel lossy compression algorithm for multidimensional data over regular grids.,36381
of The on images. vision-based temporal works proposed sequence a system,The proposed vision-based system works on a temporal sequence of images.,36382
"all Based detected. estimated lane, on other events the are","Based on the estimated lane, all other events are detected.",36383
"to speed. extensive their received and Recently, have filter-based due simplicity attention superior correlation trackers","Recently, correlation filter-based trackers have received extensive attention due to their simplicity and superior speed.",36384
"To tracking forecast is novel a more tracking to be advocated failure. indicator specific, monitoring","To be more specific, a novel tracking monitoring indicator is advocated to forecast tracking failure.",36385
segmentation semantic neural networks been applied have to problems. Convolutional successfully,Convolutional neural networks have been successfully applied to semantic segmentation problems.,36386
training the This substantially process. adversarial stabilizes,This substantially stabilizes the adversarial training process.,36387
fine-grained released This will recognition. research the image facilitate dataset to acceptance be upon of,This dataset will be released upon acceptance to facilitate the research of fine-grained image recognition.,36388
The for in all transformations piece-wise trained vertices iteration. agent linear each directly predicts,The trained agent directly predicts piece-wise linear transformations for all vertices in each iteration.,36389
tissue Numerous characterize arterial used to delamination. the models have been,Numerous models have been used to characterize the arterial tissue delamination.,36390
"capture arterial failure well. can CZMs the tissue triangular Thus, exponential-linear response and","Thus, triangular and exponential-linear CZMs can capture the arterial tissue failure response well.",36391
image significant and speed results for a improves up in segmentation. performance training This learning,This results in a significant training speed up and improves learning performance for image segmentation.,36392
processing. of in critical (VSR) problems the one become most has Video super-resolution video,Video super-resolution (VSR) has become one of the most critical problems in video processing.,36393
"propose for Adversarial formulation a VSR. In Network(GAN)-based work, we Generative this","In this work, we propose a Generative Adversarial Network(GAN)-based formulation for VSR.",36394
describe and introduction to neuroimaging We collaborative analysis. a reproducible project-based,We describe a project-based introduction to reproducible and collaborative neuroimaging analysis.,36395
"To we our problems, based substantial project. a group open-ended around these address course","To address these problems, we based our course around a substantial open-ended group project.",36396
in working successful on real in reproducibly scientific questions. projects broadly were engaging students The,The projects were broadly successful in engaging students in working reproducibly on real scientific questions.,36397
"has counterparts. are been Recently, than proven more CovDs their discriminative it that infinite-dimensional low-dimensional","Recently, it has been proven that infinite-dimensional CovDs are more discriminative than their low-dimensional counterparts.",36398
"and infinite-dimensional is However, the high. CovDs the computational implicit is form load of","However, the form of infinite-dimensional CovDs is implicit and the computational load is high.",36399
"infinite-dimensional approximate Finally, these we via CovDs approximations.","Finally, we approximate infinite-dimensional CovDs via these approximations.",36400
"task our of set image classification. Empirically, the to apply framework we","Empirically, we apply our framework to the task of image set classification.",36401
for outperforms existing abnormality detection. Cardiac-MOS all work also binary,Cardiac-MOS also outperforms all existing work for binary abnormality detection.,36402
"future in potential scoring motion demonstrates clinical As first automatic the Cardiac-MOS great solution, application.","As the first automatic motion scoring solution, Cardiac-MOS demonstrates great potential in future clinical application.",36403
Both research. available are dataset the for the and code,Both the code and the dataset are available for research.,36404
two step follows process. a method Our,Our method follows a two step process.,36405
"a is First, trained. B domain autoencoder variational for","First, a variational autoencoder for domain B is trained.",36406
"using model Hashing our Deep Supervised we Embeddings (WDHT). name Accordingly, Tag Weakly","Accordingly, we name our model Weakly Supervised Deep Hashing using Tag Embeddings (WDHT).",36407
"mentioned. some for is work ideal that In future addition, have areas","In addition, some areas that is ideal for future work have mentioned.",36408
"the the Strength is node Received unknown trilateration. location estimated In Signal using (RSS),","In the Received Signal Strength (RSS), the unknown node location is estimated using trilateration.",36409
attention which Specific in become coherent). is (or spatially multipath scenarios correlated for signals given,Specific attention is given for multipath scenarios in which signals become spatially correlated (or coherent).,36410
approach and useful proposed Our offers practically into neural is interpretability. insights,Our proposed approach is practically useful and offers insights into neural interpretability.,36411
assumption algorithms. The of typical rigidity SLAM is scene in,The assumption of scene rigidity is typical in SLAM algorithms.,36412
"scenarios monocular, dynamic DynaSLAM is for RGB-D in stereo robust configurations. and","DynaSLAM is robust in dynamic scenarios for monocular, stereo and RGB-D configurations.",36413
"monocular, our stereo in datasets. evaluate public We and RGB-D system","We evaluate our system in public monocular, stereo and RGB-D datasets.",36414
visual DynaSLAM dynamic standard accuracy baselines outperforms the in scenarios. SLAM highly of,DynaSLAM outperforms the accuracy of standard visual SLAM baselines in highly dynamic scenarios.,36415
small-scale achieved results of in excellent datasets. landmark problem recognition has The,The problem of landmark recognition has achieved excellent results in small-scale datasets.,36416
"of the design effectiveness the we assess waveform approach proposed Finally, by simulations. numerical","Finally, we assess the effectiveness of the proposed waveform design approach by numerical simulations.",36417
to to add network. set samples identified neural the Bayesian using are Informative a training,Informative samples to add to the training set are identified using a Bayesian neural network.,36418
piecewise function Mumford-Shah approximates The smooth a functional function. by a,The Mumford-Shah functional approximates a function by a piecewise smooth function.,36419
"an discretize Ambrosio-Tortorelli this Calculus via Exterior Discrete In we approximation formulation. paper, a","In this paper, we discretize an Ambrosio-Tortorelli approximation via a Discrete Exterior Calculus formulation.",36420
of for is localization the automatic videos. endoscopic essential tool task Surgical analysis an,Surgical tool localization is an essential task for the automatic analysis of endoscopic videos.,36421
"annotated the data weak with circumvent supervision. In we to this lack work, of propose","In this work, we propose to circumvent the lack of annotated data with weak supervision.",36422
proposed great spinal disease method clinical in diagnoses. potential has The,The proposed method has great potential in clinical spinal disease diagnoses.,36423
"becomes image an When translation is data not accessible, problem. training ill-posed paired","When paired training data is not accessible, image translation becomes an ill-posed problem.",36424
to a single separating approach reflection an from image. present We,We present an approach to separating reflection from a single image.,36425
separation. layer novel We loss pixel-level exclusion also that enforces a propose,We also propose a novel exclusion loss that enforces pixel-level layer separation.,36426
"foveated images, pixel image produce spatially density. with we an To rasterize varying to","To produce foveated images, we rasterize to an image with spatially varying pixel density.",36427
set approach. of our perceptual of user A the demonstrates studies efficacy,A set of perceptual user studies demonstrates the efficacy of our approach.,36428
"the These processing task not classification jointly hand. are stages, at optimizing however,","These processing stages, however, are not jointly optimizing the classification task at hand.",36429
of information is label into Contextual means recurrent by ReuseNet. incorporated FuseNet a version called,Contextual label information is incorporated into FuseNet by means of a recurrent version called ReuseNet.,36430
"regularization fields. and resampling and such map conditional as random pansharpening through interpolation,","pansharpening and resampling through interpolation, and map regularization such as conditional random fields.",36431
FuseNet approaches baseline quantitative surpass qualitative in results. ReuseNet and the and both,FuseNet and ReuseNet surpass the baseline approaches in both quantitative and qualitative results.,36432
"Finally, games present suite. Atari preliminary the we results in of a subset","Finally, we present preliminary results in a subset of the Atari games suite.",36433
enables new image This acquisition results. analysis of applications optimising possible towards best further,This further enables new applications of optimising image acquisition towards best possible analysis results.,36434
in surveillance. It and motion capture applications has broad,It has broad applications in motion capture and surveillance.,36435
"are detecting to due and not complexity dynamic of the However, straightforward. tracking structure, human","However, due to the complexity of human dynamic structure, detecting and tracking are not straightforward.",36436
"detection Consequently, developed. different techniques have with and tracking and been different performance applications","Consequently, different detection and tracking techniques with different applications and performance have been developed.",36437
"technique a To indispensable. is fusion this problem, avoid using multi-sensor","To avoid this problem, using a multi-sensor fusion technique is indispensable.",36438
"survey detection, are In multi-sensor this presented. tracking methods of a paper, fusion comparative and","In this paper, a comparative survey of detection, tracking and multi-sensor fusion methods are presented.",36439
as can tractography. as a direct TOMs prior input even act for or,TOMs can act as a prior or even as direct input for tractography.,36440
the learn fully-convolutional We use encoder-decoder mapping. neural an network required architecture to,We use an encoder-decoder fully-convolutional neural network architecture to learn the required mapping.,36441
available. data code are openly Our and our,Our code and our data are openly available.,36442
grid outages Simulations the due reproduce behavior the of of dynamic post-contingency generators. to,Simulations reproduce the post-contingency dynamic behavior of the grid due to outages of generators.,36443
the paper. images in self-ception Some included are,Some self-ception images are included in the paper.,36444
numerous records from scene. field light A real-world rays light a,A light field records numerous light rays from a real-world scene.,36445
"dense devices a by light existing However, capturing time-consuming is a process. field","However, capturing a dense light field by existing devices is a time-consuming process.",36446
significant correlation is constraint spatial dimensions. The and related between a depth most angular,The most significant constraint is a depth related correlation between angular and spatial dimensions.,36447
method constraint employing the deep out a network. by working error-sensitive avoids neural Our,Our method avoids working out the error-sensitive constraint by employing a deep neural network.,36448
"careful fine-grained datasets. part categorization many fact, annotations In are in available","In fact, careful part annotations are available in many fine-grained categorization datasets.",36449
is image-level of obtained aggregating region-level prediction by products element-wise The these probabilities.,The image-level prediction is obtained by aggregating element-wise products of these region-level probabilities.,36450
"proposed. To detection problem, and address this efficient YOLO were as SSD frameworks object such","To address this problem, efficient object detection frameworks such as YOLO and SSD were proposed.",36451
"improved, speed comes Although accuracy. greatly the it with price lower of a is","Although the speed is greatly improved, it comes with a price of lower accuracy.",36452
"Fire In SSD. paper, an named propose efficient SSD this we","In this paper, we propose an efficient SSD named Fire SSD.",36453
input transformations. framework and takes sets images labeled The preprocessing predefined as of,The framework takes as input sets of labeled images and predefined preprocessing transformations.,36454
and the the individual optimal preprocessing images. learns jointly for It transformations classifier,It jointly learns the classifier and the optimal preprocessing transformations for individual images.,36455
computationally cheaper said version) distribution. and of an approximate,an approximate and computationally cheaper version) of said distribution.,36456
"exact MCMC sampling still accelerates the DA-MCMC distribution. applications, complex in targeting while","DA-MCMC accelerates MCMC sampling in complex applications, while still targeting the exact distribution.",36457
"a algorithm. computationally design faster, We approximate, DA-MCMC albeit","We design a computationally faster, albeit approximate, DA-MCMC algorithm.",36458
"highly dependent. problem However, acceleration the is","However, the acceleration is highly problem dependent.",36459
antigen onto update predicted for The maps competitive receptors. scheme dynamics optimal commonly considered,The predicted optimal update scheme maps onto commonly considered competitive dynamics for antigen receptors.,36460
labeled deep Existing large models amount of data. require network a,Existing deep network models require a large amount of labeled data.,36461
to framework TD-GAN be can easily adapted general other The is tasks. learning and,The TD-GAN framework is general and can be easily adapted to other learning tasks.,36462
performances. filter currently terms of (CF) based are trackers top Correlation ranked their in,Correlation filter (CF) based trackers are currently ranked top in terms of their performances.,36463
applications. a high intensity value in diagnosis pain has assessment Automatic disease,Automatic pain intensity assessment has a high value in disease diagnosis applications.,36464
"simulation promising filtering evaluated algorithm scenario, used results. a highly with is The over commonly","The algorithm is evaluated over a commonly used filtering simulation scenario, with highly promising results.",36465
"breast cancerous poses coherency (BC) cells high to inherently significant challenges Also, cancer multi-classification. of","Also, inherently high coherency of cancerous cells poses significant challenges to breast cancer (BC) multi-classification.",36466
investigates structural bridge of recognition video paper using the data. This components automated,This paper investigates the automated recognition of structural bridge components using video data.,36467
is machine A train learning created video to algorithms. the simulated new dataset,A new simulated video dataset is created to train the machine learning algorithms.,36468
"video collected field. Results presented simulated as in the as for well video are data,","Results are presented for simulated video data, as well as video collected in the field.",36469
the indispensable understanding in Scene within scene. an a component parsing is semantics,Scene parsing is an indispensable component in understanding the semantics within a scene.,36470
"on parsing. networks on based scene fully have neural Recently, new convolutional records methods achieved","Recently, methods based on fully convolutional neural networks have achieved new records on scene parsing.",36471
task. efficiency counterpart FFT of CCIM The in that of remarkably the exceeds our,The efficiency of CCIM exceeds that of the FFT counterpart remarkably in our task.,36472
is time applied the architecture attributes This is such first for problem. fine-grained the classification,This is the first time such architecture is applied for the fine-grained attributes classification problem.,36473
a Convolutional with method competitive baseline. compared Neural Network We the proposed deep,We compared the proposed method with a competitive deep Convolutional Neural Network baseline.,36474
"exist practice. disadvantages there in actually two with However, GRM","However, there actually exist two disadvantages with GRM in practice.",36475
reconstruction by the measurements randomly performances. CS cannot obtained is Another sufficient GRM that provide,Another is that the CS measurements randomly obtained by GRM cannot provide sufficient reconstruction performances.,36476
"solve STP. this an algorithm propose paper, In we the to","In this paper, we propose an algorithm to solve the STP.",36477
for New proposed. score a vertex intensification heuristics system constructive are and purpose,New constructive heuristics and a vertex score system for intensification purpose are proposed.,36478
benchmarks which shows on algorithm is set The encouraging of a tested results.,The algorithm is tested on a set of benchmarks which shows encouraging results.,36479
"achieve PSV evaluate result. our We method an and dataset, advanced on","We evaluate our method on PSV dataset, and achieve an advanced result.",36480
step therefore automating workflow. the in of key cells a inspection Automated segmentation visual is,Automated segmentation of cells is therefore a key step in automating the visual inspection workflow.,36481
"have Their as anisotropy, such analyzed extensively. been spectra, discrepancy, properties or progressiveness","Their properties such as discrepancy, spectra, anisotropy, or progressiveness have been analyzed extensively.",36482
correlations extractor can The via CNN spatial benefit component. considering also CRF from the feature,The CNN feature extractor can also benefit from considering spatial correlations via the CRF component.,36483
"Answering. propose for Question Visual paper, graph-based this we In approach a novel","In this paper, we propose a novel graph-based approach for Visual Question Answering.",36484
by experience visuo-tactile simple body as theirs parts just stimulation. can Humans fake synchronous,Humans can experience fake body parts as theirs just by simple visuo-tactile synchronous stimulation.,36485
"example, procedurally, can or from For manually. abstracted they created measurements, generated be raw","For example, they can be abstracted from raw measurements, generated procedurally, or created manually.",36486
consistent distributions The neighborhoods. to are various GANs and buildings synchronized style produce over,The various GANs are synchronized to produce consistent style distributions over buildings and neighborhoods.,36487
the variability user We of the with output. provide the control over direct,We provide the user with direct control over the variability of the output.,36488
on We large-scale examples. demonstrate several our system,We demonstrate our system on several large-scale examples.,36489
image important patient-specific an anatomy. plays in Semantic role modeling segmentation,Semantic image segmentation plays an important role in modeling patient-specific anatomy.,36490
and is medical both time-consuming image Manual annotation expensive.,Manual medical image annotation is both time-consuming and expensive.,36491
time vessels to hours minutes. Kid-Net reduces segmentation from of matter kidney,Kid-Net reduces kidney vessels segmentation time from matter of hours to minutes.,36492
and classification utilized and Feature up-sampling higher down-sampling to achieve accuracies. localization are,Feature down-sampling and up-sampling are utilized to achieve higher classification and localization accuracies.,36493
competence. on challenging evaluation dataset show a results Quantitative qualitative and testing Kid-Net,Quantitative and qualitative evaluation results on a challenging testing dataset show Kid-Net competence.,36494
This to cases. degradation the in leads some performance fusion,This leads to the fusion performance degradation in some cases.,36495
"features source Firstly, deep from network the extract residual is images. to used (ResNet)","Firstly, the residual network (ResNet) is used to extract deep features from source images.",36496
maps. the and obtain is Then weight ZCA to utilized normalize features deep initial,Then ZCA is utilized to normalize the deep features and obtain initial weight maps.,36497
"using Finally, a reconstructed strategy. is image the weighted-averaging fused","Finally, the fused image is reconstructed using a weighted-averaging strategy.",36498
integrates provide information Multi-task network neural that a jointly distinct from architectures mechanism sources.,Multi-task neural network architectures provide a mechanism that jointly integrates information from distinct sources.,36499
and network that synCTs allows sampling multiple This their representation. share segmentations of,This allows sampling of multiple segmentations and synCTs that share their network representation.,36500
for networks multi-focus fusion. Convolutional used been image neural have recently,Convolutional neural networks have recently been used for multi-focus image fusion.,36501
vector popular recognition spaces a strategy pattern Embedding very data methods. into of is,Embedding data into vector spaces is a very popular strategy of pattern recognition methods.,36502
"ambiguous. distances metrics When embeddings are between quantized, performance become","When distances between embeddings are quantized, performance metrics become ambiguous.",36503
systems. a it measurable state-of-the-art that have demonstrate in can empirical data in effect We,We demonstrate that it can have a measurable effect in empirical data in state-of-the-art systems.,36504
"this model, of Point-of-care at roles. systems, two important the play heart","Point-of-care systems, at the heart of this model, play two important roles.",36505
"will the be of assessing progression such used they Second, therapies. for","Second, they will be used for assessing the progression of such therapies.",36506
deconvolution local detection. present alternate We to as an approach maxima,We present deconvolution as an alternate approach to local maxima detection.,36507
a are filter truth artifical points generate convolved The ground with to labels. mapping,The ground truth points are convolved with a mapping filter to generate artifical labels.,36508
CNN cell generate is detection. of then deconvolved as the to Output trained points,Output of the trained CNN is then deconvolved to generate points as cell detection.,36509
"directions for these rewards. we leveraging further discuss Finally, intrinsic between the future interplay two","Finally, we discuss future directions for further leveraging the interplay between these two intrinsic rewards.",36510
"are methods surgical for achieve To workflow analysis crucial. this,","To achieve this, methods for surgical workflow analysis are crucial.",36511
"video-based the performance best convolutional for Currently, networks tasks. workflow analysis neural provide","Currently, convolutional neural networks provide the best performance for video-based workflow analysis tasks.",36512
"data annotated For of such networks, training necessary. large amounts are","For training such networks, large amounts of annotated data are necessary.",36513
"always a feasible. and time-consuming, not costly, of amount is collecting sufficient often However, data","However, collecting a sufficient amount of data is often costly, time-consuming, and not always feasible.",36514
one system simulating conditioned consists particle path. particle step of a on Each Gibbs,Each Gibbs step consists of simulating a particle system conditioned on one particle path.,36515
Monte to method create the system. conditional a Sequential (cSMC) relies It particle on Carlo,It relies on a conditional Sequential Monte Carlo (cSMC) method to create the particle system.,36516
can a This get used uniform result. ergodicity to be,This can be used to get a uniform ergodicity result.,36517
"continuum. modelling cell either typically macro-scale, are to two There discrete micro-scale, or approaches migration:","There are typically two approaches to modelling cell migration: either micro-scale, discrete or macro-scale, continuum.",36518
mathematical have Deterministic advantage to generally analysis. they more models are the amenable that,Deterministic models have the advantage that they are generally more amenable to mathematical analysis.,36519
"phenomenological predictive their models power. often specified which Deterministic on are basis, reduces a","Deterministic models are often specified on a phenomenological basis, which reduces their predictive power.",36520
continuum ABMs over deterministic Stochastic advantages have their counterparts.,Stochastic ABMs have advantages over their deterministic continuum counterparts.,36521
actions. This the procedures allows consisting multiple of of derivation,This allows the derivation of procedures consisting of multiple actions.,36522
Tomography (PET) Positron as functional from perfusion and metabolism. such Images Emission deliver data,Images from Positron Emission Tomography (PET) deliver functional data such as perfusion and metabolism.,36523
"Imaging structures. hand, Resonance (MRI) provide from Magnetic the information other On describing anatomical images","On the other hand, images from Magnetic Resonance Imaging (MRI) provide information describing anatomical structures.",36524
Fusing is from two the helpful in complementary the information oncology. modalities,Fusing the complementary information from the two modalities is helpful in oncology.,36525
hierarchical A two-stage approach proposed. is registration,A two-stage hierarchical registration approach is proposed.,36526
"global a is applied. registration affine First,","First, a global affine registration is applied.",36527
optimization. as function Mutual used metric the for information is,Mutual information is used as metric function for the optimization.,36528
stages. in is both A used approach multi-resolution,A multi-resolution approach is used in both stages.,36529
in world. our is repetition Visual ubiquitous,Visual repetition is ubiquitous in our world.,36530
"continuities temporal field's three addition, dynamics. from the motion temporal emerge In","In addition, three temporal motion continuities emerge from the field's temporal dynamics.",36531
repetitive filters dynamics. the estimate We wavelet motion with to maps temporally convolve,We temporally convolve the motion maps with wavelet filters to estimate repetitive dynamics.,36532
rotor resulted targeting in termination. Ablation sites arrhythmia has,Ablation targeting rotor sites has resulted in arrhythmia termination.,36533
to However rotors leading are not abundance at clear. mechanisms locations the of these,However the mechanisms leading to abundance of rotors at these locations are not clear.,36534
with change from This monomorphic ventricular process coalesces tachycardia. to polymorphic a,This process coalesces with a change from polymorphic to monomorphic ventricular tachycardia.,36535
estimation. proposed approach on for available datasets Our pose is evaluated publicly several,Our proposed approach is evaluated on several publicly available datasets for pose estimation.,36536
classification model and semi-supervised on our settings. unsupervised in localization We and also evaluated,We also evaluated our model on classification and localization in unsupervised and semi-supervised settings.,36537
"retrieval because tasks techniques hashing retrieval of their gained large-scale speed. Recently, have importance in","Recently, hashing techniques have gained importance in large-scale retrieval tasks because of their retrieval speed.",36538
data cross-view existing paired. frameworks assume the Most are well of that,Most of the existing cross-view frameworks assume that data are well paired.,36539
"situation in applications. is However, fully-paired not universal real multiview the","However, the fully-paired multiview situation is not universal in real applications.",36540
detection under presence missing robustness annotations. of object the study of the We,We study the robustness of object detection under the presence of missing annotations.,36541
detailed for this We explanation provide result. a,We provide a detailed explanation for this result.,36542
a re-weights gradients instances. of function the overlap positive RoIs Sampling with of Soft as,Soft Sampling re-weights the gradients of RoIs as a function of overlap with positive instances.,36543
HexaLab by simply opening used can be www.hexalab.net.,HexaLab can be used by simply opening www.hexalab.net.,36544
tool both visualization targets scholars. users and Our,Our visualization tool targets both users and scholars.,36545
"inspection support of variety end, we this visualization and a To wide volume tools.","To this end, we support a wide variety of visualization and volume inspection tools.",36546
People sensing issues one hottest applications. counting of in is the,People counting is one of the hottest issues in sensing applications.,36547
occlusion shading network. and calculation encoded in automatically the Spatial are,Spatial occlusion and shading calculation are automatically encoded in the network.,36548
"feature used perturbation this chain assess be may robustness. to Thus,","Thus, this perturbation chain may be used to assess feature robustness.",36549
heart the lies tracking. of Model at update object,Model update lies at the heart of object tracking.,36550
"training CNNs However, requires annotated image data.","However, training CNNs requires annotated image data.",36551
emphysema of CT We scans. consider in chest the assessing extent task,We consider the task of assessing emphysema extent in chest CT scans.,36552
how our CAPTCHAs. these are system towards work provided well defeating showing Results,Results are provided showing how our system work well towards defeating these CAPTCHAs.,36553
continuously systems physical interact complex the robotic dynamical systems with The in world.,The robotic systems continuously interact with complex dynamical systems in the physical world.,36554
"classification. algorithms on discovery poorly Unfortunately, fully-automated score this state-of-the-art even the","Unfortunately, even the state-of-the-art fully-automated discovery algorithms score poorly on this classification.",36555
at aims work addressing This points. two those,This work aims at addressing those two points.,36556
the first identifies approach of GBM. two-stage presence A,A two-stage approach first identifies the presence of GBM.,36557
"slice. each ""abnormal"" in localization MR is followed GBM a This by","This is followed by a GBM localization in each ""abnormal"" MR slice.",36558
"viz. As two system, of architectures CADe part the CNN","As part of the CADe system, two CNN architectures viz.",36559
Detection (C-CNN) and are CNN Classification employed. CNN (D-CNN),Classification CNN (C-CNN) and Detection CNN (D-CNN) are employed.,36560
of surgery assessment surgical efficient for Recognition training. and is surgical crucial gesture skill,Recognition of surgical gesture is crucial for surgical skill assessment and efficient surgery training.,36561
usually current segment-level from Most edit the therefore over-segmentation of scores. low approaches suffer and,Most of the current approaches usually suffer from over-segmentation and therefore low segment-level edit scores.,36562
deep from a learning features with trained intelligent using hierarchical is agent reinforcement model. An,An intelligent agent is trained using reinforcement learning with hierarchical features from a deep model.,36563
consistency and mechanism Temporal action into design over-segmentation reduce errors. reward our integrated is to,Temporal consistency is integrated into our action design and reward mechanism to reduce over-segmentation errors.,36564
"approaches of Unfortunately, existing problem. from the same majority suffers the","Unfortunately, the majority of existing approaches suffers from the same problem.",36565
CNNs I present use In i.e. this of another paper,In this paper I present another use of CNNs i.e.,36566
and detectors on light perception. integral cars Traffic for autonomous scene sign are road,Traffic light and sign detectors on autonomous cars are integral for road scene perception.,36567
"region boxes, proposed method of or bounding notion has sliding no windows. proposals, The","The proposed method has no notion of bounding boxes, region proposals, or sliding windows.",36568
object detectors fine-tuned and We methods state-of-the-art tracking. outperform for pupil generic,We outperform state-of-the-art generic object detectors and methods fine-tuned for pupil tracking.,36569
This paper EH-powered some theoretical CPS architecture the problems. discusses and,This paper discusses the EH-powered CPS architecture and some theoretical problems.,36570
"partially EH EH-powered and fully addressed. (e.g., CPSs Different for models EH-powered) are","Different EH models for CPSs (e.g., fully EH-powered and partially EH-powered) are addressed.",36571
directions discussed research in the Future are end. briefly,Future research directions are briefly discussed in the end.,36572
"not However, representations. do most deep networks feature such enforce","However, most deep networks do not enforce such feature representations.",36573
"important it not is per In the is candidate. performance context, this what","In this context, it is not important what is the performance per candidate.",36574
Instead objective. group per the the ultimate performance is,Instead the performance per group is the ultimate objective.,36575
millions algorithm of was handle hundreds of The to designed examples. efficiently,The algorithm was designed to efficiently handle hundreds of millions of examples.,36576
"the being automatic capture is virtual environment is screen Concurrently, executed while explored.","Concurrently, automatic screen capture is executed while the virtual environment is being explored.",36577
We self-supervised using spatio-temporal between frames for signals action present recognition. a video approach,We present a self-supervised approach using spatio-temporal signals between video frames for action recognition.,36578
tangle architecture to A temporal is and spatial representation two-stream learning. leveraged,A two-stream architecture is leveraged to tangle spatial and temporal representation learning.,36579
is as verification Our and a task formulated tasks. both sequence spatio-temporal alignment,Our task is formulated as both a sequence verification and spatio-temporal alignment tasks.,36580
The action weights self-supervised validated is on the effectiveness pre-trained task. recognition,The self-supervised pre-trained weights effectiveness is validated on the action recognition task.,36581
are to and still validity boost required. investigations Further generalize performance,Further investigations to boost performance and generalize validity are still required.,36582
"MRIs of plausibility this gate we In the for paper, using diagnosing PD. automatically investi-","In this paper, we investi- gate the plausibility of using MRIs for automatically diagnosing PD.",36583
the FreeSurfer is second and library the used steps. for first The,The FreeSurfer library is used for the first and the second steps.,36584
evaluate method. set used proposed to Markers data the Progression Parkinsons is The (PPMI) Initiative,The Parkinsons Progression Markers Initiative (PPMI) data set is used to evaluate the proposed method.,36585
system be The proposed the in of PD. prove promising diagnosis assisting to,The proposed system prove to be promising in assisting the diagnosis of PD.,36586
in prerequisites appropriate provided the mathematical are The appendices.,The appropriate mathematical prerequisites are provided in the appendices.,36587
as used (SMO). Spectral Sequential features Optimization input features have for based Minimal been,Spectral based features have been used as input features for Sequential Minimal Optimization (SMO).,36588
segmentation tasks. Random image medical forests (RF) successfully applied many have to been,Random forests (RF) have been successfully applied to many medical image segmentation tasks.,36589
"pixels. contextual pixel-wise However, relationships the outputs ignores between classifier label RF of individual","However, the pixel-wise RF classifier ignores contextual relationships between label outputs of individual pixels.",36590
transformation explore using as ITN. effect representations different regression of outputs We the of,We explore the effect of using different transformation representations as regression outputs of ITN.,36591
"complete to are manipulation Also, subtle most involved there actions. object of movements","Also, there are subtle movements involved to complete most of object manipulation actions.",36592
"determine to i.e. graph the p-Lapalcian fitting is difficult it However,","However, it is difficult to determine the fitting graph p-Lapalcian i.e.",36593
parameter performance the is graph factor the critical a which p-Laplacian. for of,the parameter which is a critical factor for the performance of graph p-Laplacian.,36594
"on the conduct then, graph. learning And framework semi-supervised fused we","And then, we conduct semi-supervised learning framework on the fused graph.",36595
the data and set Extensive efficiency UC-Merced proposed demonstrate the experiments method. on of effectiveness,Extensive experiments on UC-Merced data set demonstrate the effectiveness and efficiency of the proposed method.,36596
"available new, on In text/video a to we large-scale datasets, dataset. publicly introduce testing addition","In addition to testing on publicly available datasets, we introduce a new, large-scale text/video dataset.",36597
(FPS) First-Person in the game. We the of Shooter technique test context a,We test the technique in the context of a First-Person Shooter (FPS) game.,36598
"closed is scene images, once, to allowing such geometries. recover at With seen all","With such images, all scene is seen at once, allowing to recover closed geometries.",36599
leading the cancer in several is the cancer-related years. past of Lung deaths cause,Lung cancer is the leading cause of cancer-related deaths in the past several years.,36600
automated networks (CNNs). in neural deep classification State-of-the-art approaches nodule convolutional lung use,State-of-the-art approaches in automated lung nodule classification use deep convolutional neural networks (CNNs).,36601
"number large samples a However, these training generalize well. of require to networks","However, these networks require a large number of training samples to generalize well.",36602
alternative investigates as an of use capsule (CapsNets) CNNs. networks This paper to the,This paper investigates the use of capsule networks (CapsNets) as an alternative to CNNs.,36603
significantly show training We CNNs that CapsNets when is the of outperforms small. samples number,We show that CapsNets significantly outperforms CNNs when the number of training samples is small.,36604
The real-world method validated problems. on several is proposed Radiomics,The proposed method is validated on several real-world Radiomics problems.,36605
"Another is of GANs application e.g. image-to-image translations,","Another application of GANs is image-to-image translations, e.g.",36606
and We supervised evaluate model. a pure generator models; CycleGAN modification two a also of,We also evaluate two supervised models; a modification of CycleGAN and a pure generator model.,36607
"there Therefore, the pressing arterial of is understand a to delamination. micromechanical need mechanisms","Therefore, there is a pressing need to understand the micromechanical mechanisms of arterial delamination.",36608
this (e.g. The failure of objective study investigate to mechanisms various was,The objective of this study was to investigate various failure mechanisms (e.g.,36609
of interfacial arterial for collagen responsible fibers) delamination. failure,failure of collagen fibers) responsible for arterial interfacial delamination.,36610
tests a scanning were microscope. performed tensile of the in electron on In-situ micro-tester fibers,In-situ tensile tests of fibers were performed on a micro-tester in the scanning electron microscope.,36611
"elegant provides paper complete. of explicit proof algebraic matrix, that construction an the This making","This paper provides an elegant explicit construction of that matrix, making the algebraic proof complete.",36612
constructive thus proof in included is proposed being very The for textbooks. here algebraic suitable,The constructive algebraic proof proposed here is thus very suitable for being included in textbooks.,36613
"more provides Therefore, preserving. HpLapR proposed to local potential the the structure exploiting","Therefore, the proposed HpLapR provides more potential to exploiting the local structure preserving.",36614
to We remote regression and recognition. image apply for the logistic HpLapR conduct implementations sensing,We apply HpLapR to logistic regression and conduct the implementations for remote sensing image recognition.,36615
HpLapR. experimental results the demonstrate of superiority The the proposed,The experimental results demonstrate the superiority of the proposed HpLapR.,36616
"performance. correspondences challenge registration inaccurate However, the","However, inaccurate correspondences challenge the registration performance.",36617
"estimate optimal correspondences minimize To for influence, PointNet. using their we weights","To minimize their influence, we estimate optimal weights for correspondences using PointNet.",36618
We the to minimize the criterion directly registration train error. with the network,We train the network directly with the criterion to minimize the registration error.,36619
for this method. justification We theoretical provide,We provide theoretical justification for this method.,36620
problem. is Object a detection solving to the by Tracking approach Tracking common Multiple,Tracking by detection is a common approach to solving the Multiple Object Tracking problem.,36621
"approaches. evaluation, competitive online, real-time among achieves results especially in This method","This method achieves competitive results in evaluation, especially among online, real-time approaches.",36622
and feedbacks by Eco-evolutionary can formalisms. demographic coupling be modelled evolutionary readily,Eco-evolutionary feedbacks can be readily modelled by coupling demographic and evolutionary formalisms.,36623
avenues. of integrative future approaches suggest modelling overview We provide and an these,We provide an overview of these approaches and suggest future integrative modelling avenues.,36624
our 'Spatio-Temporal name novel Channel Correlation' (STC). We block,We name our novel block 'Spatio-Temporal Channel Correlation' (STC).,36625
acceptance as so degrees We accommodate for ADFs arguments. the to arbitrary generalize further,We further generalize ADFs so as to accommodate arbitrary acceptance degrees for the arguments.,36626
problems related complexity for We results present to weighted also ADFs.,We also present complexity results for problems related to weighted ADFs.,36627
can theorems. physics unique geometrical well Such interpreted that be using filtering understood possesses properties,Such filtering possesses unique geometrical properties that can be interpreted using well understood physics theorems.,36628
social involves Successful processes. and both cognitive explanation,Successful explanation involves both cognitive and social processes.,36629
this of neurons. to on process efforts the Current automate focus mainly segmentation,Current efforts to automate this process focus mainly on the segmentation of neurons.,36630
"network and cross-validated SPHERE_RGBD trained available on a The + publicly is Inertial_calorie. dataset,","The network is trained and cross-validated on a publicly available dataset, SPHERE_RGBD + Inertial_calorie.",36631
achieve proposed was and shown to The optimization-based substantially method state-of-the-art results. outperform baselines,The proposed method was shown to substantially outperform optimization-based baselines and achieve state-of-the-art results.,36632
"accuracy networks, wider and using deeper be rapidly. can improved By network","By using deeper and wider networks, network accuracy can be improved rapidly.",36633
set of attention. problem selecting received the The images of much has,The problem of selecting the set of images has received much attention.,36634
in they identification the walk on based (i.e. People video way,People identification in video based on the way they walk (i.e.,36635
relevant vision task is using computer gait) a a in non-invasive approach.,gait) is a relevant task in computer vision using a non-invasive approach.,36636
"each at relay In this received is signal and forwarded. structure, demodulated","In this structure, received signal at each relay is demodulated and forwarded.",36637
are Bit system and (BER) Rate as Error performance Probability (P_out) criteria. investigated Outage,Bit Error Rate (BER) and Outage Probability (P_out) are investigated as system performance criteria.,36638
at atmospheric The proposed strong performance structure shows moderate and regimes. independent turbulence,The proposed structure shows independent performance at moderate and strong atmospheric turbulence regimes.,36639
the side protection DC One a multi-terminal against major faults. is of challenges in VSCHVDC,One of the major challenges in a multi-terminal VSCHVDC is protection against DC side faults.,36640
(DMR) Redundancy for is fault decision of scenario. Modular making Dual utilized technique a The,The Dual Modular Redundancy (DMR) technique is utilized for decision making of a fault scenario.,36641
dual other. result a when contradicts uses one system voting This the,This uses a dual voting system when one result contradicts the other.,36642
combined for change considered threshold breaker the operation. with directional Current is,Current threshold combined with directional change is considered for the breaker operation.,36643
proposed Mignotte has [M. a the model Mignotte. problem for non-convex,Mignotte has proposed a non-convex model for the problem [M. Mignotte.,36644
edge-histogram specification image for model energy-based problem. An the,An energy-based model for the image edge-histogram specification problem.,36645
"Furthermore, constraints on we applications. additional include our model based in different","Furthermore, we include additional constraints in our model based on different applications.",36646
method decent Numerical given results are successfully produces efficiently. that our to illustrate results,Numerical results are given to illustrate that our method successfully produces decent results efficiently.,36647
for from virus Identifying downstream viral data metagenomic critical is sequences analyses.,Identifying viral sequences from metagenomic data is critical for downstream virus analyses.,36648
"(e.g., information or tweets) eyewitnesses by posted of images disaster. a includes text Such","Such information includes text (e.g., tweets) or images posted by eyewitnesses of a disaster.",36649
use to images The of disaster damage limited. social assess is media,The use of social media images to assess disaster damage is limited.,36650
"represented feature sets vectors (\emph{instances}). multiple by instance (\emph{bags}) objects (MIL), In are learning of","In multiple instance learning (MIL), objects are represented by sets (\emph{bags}) of feature vectors (\emph{instances}).",36651
"datasets classifiers for Like classification, benchmark and available numerous for MIL. several are supervised","Like for supervised classification, several benchmark datasets and numerous classifiers are available for MIL.",36652
the what drawn be has kind for of results. implications This conclusions may from comparison,This has implications for what kind of conclusions may be drawn from the comparison results.,36653
very similar can that conceptually behave show datasets Our differently. results,Our results show that conceptually similar datasets can behave very differently.,36654
method can recover has and local Our details real-time performance. strand,Our method can recover local strand details and has real-time performance.,36655
our a synthetic of rendered large to train network. use set models We hair,We use a large set of rendered synthetic hair models to train our network.,36656
Fluid three principles: Strong (I) on is based Machine-Learning annotation aid.,Fluid annotation is based on three principles: (I) Strong Machine-Learning aid.,36657
edit model. also the are operations The by assisted,The edit operations are also assisted by the model.,36658
single image annotation Full in a (II) pass.,(II) Full image annotation in a single pass.,36659
We annotator to and choose order. in which what annotate empower to the,We empower the annotator to choose what to annotate and in which order.,36660
"machine This on already the know, what enables does i.e. concentrating not","This enables concentrating on what the machine does not already know, i.e.",36661
made. putting human it the effort on errors only,putting human effort only on the errors it made.,36662
effectively. budget using helps the This annotation,This helps using the annotation budget effectively.,36663
"extraction IVUS lumen arterial segmentation involves media. The namely, components of walls two and the","The IVUS segmentation involves the extraction of two arterial walls components namely, lumen and media.",36664
image interpretation. automatic preliminary towards is an segmentation  Semantic important step  medical,  Semantic segmentation is an important preliminary step towards automatic medical image interpretation.,36665
previous results attention focused more values and This methods. in plausible than more visualisation,This results in more focused attention values and more plausible visualisation than previous methods.,36666
apply method tasks. different three to proposed We the,We apply the proposed method to three different tasks.,36667
local forming hypotheses. to image are tubular fit Regularly spaced templates data,Regularly spaced tubular templates are fit to image data forming local hypotheses.,36668
This track enables starting to single a seed trees from method point. our,This enables our method to track trees starting from a single seed point.,36669
on coronary chest extract trees to is data and Our evaluated arteries. method CT airway,Our method is evaluated on chest CT data to extract airway trees and coronary arteries.,36670
"during greatly the ""db stage, the regions"" retrieval performance improved. this is called descriptors Using","Using this descriptors called ""db regions"" during the retrieval stage, the performance is greatly improved.",36671
"may Like field empirical science, any axiomatically. approached be AI of","Like any field of empirical science, AI may be approached axiomatically.",36672
"We human-level formulate AI system general-purpose, a in of terms postulates. for requirements","We formulate requirements for a general-purpose, human-level AI system in terms of postulates.",36673
research. learning deep relate advances theoretical in neuroscience to We also,We also relate deep learning to advances in theoretical neuroscience research.,36674
these state-of-the-art tasks. performance results demonstrate across all Our,Our results demonstrate state-of-the-art performance across all these tasks.,36675
We new clustering method propose a optimal transportation. based on,We propose a new clustering method based on optimal transportation.,36676
on focus learning vehicle works mainly and global metrics. distance Previous features Re-ID extracting on,Previous works on vehicle Re-ID mainly focus on extracting global features and learning distance metrics.,36677
"on i.e., datasets, our two methods VehicleID. VeRi vehicle Re-ID evaluate and We large-scale","We evaluate our methods on two large-scale vehicle Re-ID datasets, i.e., VeRi and VehicleID.",36678
with performance comparison results promising recent show methods works. in achieve Experimental our,Experimental results show our methods achieve promising performance in comparison with recent works.,36679
complex functions. provide convolutional neural to more mapping Deeper networks capacity approximate,Deeper convolutional neural networks provide more capacity to approximate complex mapping functions.,36680
"depth and increasing increases However, on complexity. difficulties model network imposes training","However, increasing network depth imposes difficulties on training and increases model complexity.",36681
"SSIMLayer nonlinear not it and inherently hence, does transformations. require proposed is nonlinear The subsequent","The proposed SSIMLayer is inherently nonlinear and hence, it does not require subsequent nonlinear transformations.",36682
"the robust To features, a integrate pre-trained kinship to face we face generator. model extract","To extract robust features, we integrate a pre-trained face model to the kinship face generator.",36683
"transformation stable attain more we Lastly, a results. cycle-domain adapt to","Lastly, we adapt cycle-domain transformation to attain a more stable results.",36684
Wild conducted (FIW) Families are dataset. on Experiments the in,Experiments are conducted on Families in the Wild (FIW) dataset.,36685
"are Particularly, important this three presented contributions throughout paper.","Particularly, three important contributions are presented throughout this paper.",36686
Personalized is and of the online task broadcasters. a crucial many video businesses for advertisement,Personalized advertisement is a crucial task for many of the online businesses and video broadcasters.,36687
"registration require a transfer the or training Moreover, does not to keypoint an atlas phase.","Moreover, keypoint transfer does not require the registration to an atlas or a training phase.",36688
"allows with Finally, for the the field-of-view. variable method highly segmentation of scans","Finally, the method allows for the segmentation of scans with highly variable field-of-view.",36689
epidemic allowed distinct of and us criticality processes. phases to identify This,This allowed us to identify criticality and distinct phases of epidemic processes.,36690
We classification. propose decision-tree approach also based a for,We also propose a decision-tree based approach for classification.,36691
"samples where In training recognition, all image are cannot there cover many cases classes. target","In image recognition, there are many cases where training samples cannot cover all target classes.",36692
is top Non-invasive currently in or among testing NIPT obstetric the care. prenatal topic researched,Non-invasive prenatal testing or NIPT is currently among the top researched topic in obstetric care.,36693
"testing Trisomy, Uninformative DNA, Cell-free sequencing, Keywords: Method, . result, Prenatal Next-generation","Keywords: Next-generation sequencing, Cell-free DNA, Uninformative result, Method, Trisomy, Prenatal testing .",36694
to the task. both proposal captioning generation dense generation and important Temporal caption are,Temporal proposal generation and caption generation are both important to the dense captioning task.,36695
Reduction approach friction heat efficiency. is proven and losses of the to a engine increase,Reduction of heat and friction losses is a proven approach to increase the engine efficiency.,36696
"proposed feed like strategies, control forward compared. feedback or are controllers, Different and","Different control strategies, like feed forward or feedback controllers, are proposed and compared.",36697
"detailed differences are modelling some However, between simplified there which worth are mentioning. significant and","However, there are some significant differences between simplified and detailed modelling which are worth mentioning.",36698
"design cross-layer a Wide (LPWANs). networks Area is imperative Power Therefore, realize effectively to Low","Therefore, a cross-layer design is imperative to effectively realize Low Power Wide Area networks (LPWANs).",36699
"present a including proposed. modeling paper, In cross-layer the is power assessment framework","In the present paper, a cross-layer assessment framework including power modeling is proposed.",36700
and cross-layer energy a improves significantly throughput. that efficiency We demonstrate approach overall,We demonstrate that a cross-layer approach significantly improves energy efficiency and overall throughput.,36701
"LPWAN assessment been open-source First, an framework conceived. has","First, an open-source LPWAN assessment framework has been conceived.",36702
hypotheses schemes. allows evaluating and and It testing,It allows testing and evaluating hypotheses and schemes.,36703
"Secondly, is assessed. protocol representative a LoRaWAN as case, the","Secondly, as a representative case, the LoRaWAN protocol is assessed.",36704
"point-optimal reference our test For efficient. and a is correctly density, nearly specified","For a correctly specified reference density, our test is point-optimal and nearly efficient.",36705
inpainting frames. of We for recovering video missing regions video data-driven a new present method,We present a new data-driven video inpainting method for recovering missing regions of video frames.,36706
in Our end-to-end jointly trains an manner. sub-networks both method,Our method jointly trains both sub-networks in an end-to-end manner.,36707
"their uncertainty, errors may lead rarely these which However, methods to quantify downstream in analysis.","However, these methods rarely quantify their uncertainty, which may lead to errors in downstream analysis.",36708
"Additionally, replaced iterative is parallel, one. standard the by a multi-threaded implementation","Additionally, the standard iterative implementation is replaced by a parallel, multi-threaded one.",36709
describe We details scalability strong scaling and formulation. using the a its implementation analyze,We describe the implementation details and analyze its scalability using a strong scaling formulation.,36710
as training process. sequence a stochastic We the model,We model the training sequence as a stochastic process.,36711
models evidence We argue are empirical power-law processes. that natural from suitable for,We argue from empirical evidence that power-law models are suitable for natural processes.,36712
like Random where lunch is transfer typing work. no-free does model not learning,Random typing model is like no-free lunch where transfer learning does not work.,36713
samples independently the distribution. Zeta programs zeta process from,Zeta process independently samples programs from the zeta distribution.,36714
database of a genetics model A by sub-programs of uses inspired common sub-programs.,A model of common sub-programs inspired by genetics uses a database of sub-programs.,36715
An samples Zeta distribution. process evolutionary from mutations zeta,An evolutionary zeta process samples mutations from Zeta distribution.,36716
semi-intrusive family uncertainty for of introduced. models (UP) multiscale propagation methods A is,A family of semi-intrusive uncertainty propagation (UP) methods for multiscale models is introduced.,36717
tested have case on These algorithms based dynamics. been semi-intrusive studies two on reaction-diffusion,These semi-intrusive algorithms have been tested on two case studies based on reaction-diffusion dynamics.,36718
are important human-machine learning for electromyography Machine control. surface device and classifiers interfacing using,Machine learning classifiers using surface electromyography are important for human-machine interfacing and device control.,36719
features extracted (SVMs) vector classifiers as machines manually based Conventional use such on support e.g.,Conventional classifiers such as support vector machines (SVMs) use manually extracted features based on e.g.,36720
"advantage. - networks, can automatically extract an Deep important contrast, features neural specific person by","Deep neural networks, by contrast, can automatically extract person specific features - an important advantage.",36721
have images. images Retinal among the medical resolution highest clarity and,Retinal images have the highest resolution and clarity among medical images.,36722
Experimental state-of-the-art demonstrate evaluations methods. network that outperforms proposed current the,Experimental evaluations demonstrate that the proposed network outperforms current state-of-the-art methods.,36723
important performance that Ablative improvement. analysis the SSA in is factor shows an indeed,Ablative analysis shows that the SSA is indeed an important factor in performance improvement.,36724
resulting array non-uniform is The sparse. highly linear,The resulting non-uniform linear array is highly sparse.,36725
"patterns sparse the Also, arrays. than better lobe other exhibits side SCA","Also, the SCA exhibits better side lobe patterns than other sparse arrays.",36726
advance concern face privacy recognition leakage. The people's techniques the regarding new of also arises,The advance of new face recognition techniques also arises people's concern regarding the privacy leakage.,36727
"in protection, terms privacy preservation, evaluate utility similarity. and structure the approach proposed We of","We evaluate the proposed approach in terms of privacy protection, utility preservation, and structure similarity.",36728
is accurate with of steatosis An pathology importance. tissue clinical quantification samples high,An accurate steatosis quantification with pathology tissue samples is of high clinical importance.,36729
introduces learning control. generative new deep This synthesis for and motion paper network human a,This paper introduces a new generative deep learning network for human motion synthesis and control.,36730
describe method motion We first for data. prerecorded training efficient a an from model RNNs,We first describe an efficient method for training a RNNs model from prerecorded motion data.,36731
our We models. model against baseline the by comparison generative show superiority of,We show the superiority of our generative model by comparison against baseline models.,36732
many areas is Deep learning ubiquitous areas across computer of vision.,Deep learning is ubiquitous across many areas areas of computer vision.,36733
large It for scale scale often datasets fine-tuned on before being small-to-medium problems. requires training,It often requires large scale datasets for training before being fine-tuned on small-to-medium scale problems.,36734
"datasets. in classes paper between we specifically, overlapping the this utilize More","More specifically, in this paper we utilize the overlapping classes between datasets.",36735
available state-of-the-art We three results demonstrate benchmarks. publicly on,We demonstrate state-of-the-art results on three publicly available benchmarks.,36736
run have CNNs segmentation for time Deep requirements. memory and high semantic,Deep CNNs for semantic segmentation have high memory and run time requirements.,36737
"to shuffled, grouped, proposed efficient like Various approaches make CNNs convolutions. depth-wise separable have been","Various approaches have been proposed to make CNNs efficient like grouped, shuffled, depth-wise separable convolutions.",36738
cases model AMD normal is minimal using on and mixture a trained of data. This,This model is trained on a mixture of normal and AMD cases using minimal data.,36739
with methods. par Our model's also existing is performance on the,Our model's performance is also on par with the existing methods.,36740
"C-arm for solution Augmented marker-free proposes Reality ""technician-in-the-loop"" (AR) repositioning. work This a","This work proposes a marker-free ""technician-in-the-loop"" Augmented Reality (AR) solution for C-arm repositioning.",36741
We in simulating experiments setting trauma a orthopedic surgery. conduct,We conduct experiments in a setting simulating orthopedic trauma surgery.,36742
are relays (AF) and forward Amplifyand- often decode-and-forward used. (DF),Amplifyand- forward (AF) and decode-and-forward (DF) relays are often used.,36743
as Hybrid system this to We refer approach. the,We refer to this system as the Hybrid approach.,36744
to Carlo are validate simulations analytical results. throughout the this used paper Monte,Monte Carlo simulations are used throughout this paper to validate the analytical results.,36745
parameters this investigated is system in several of paper. The impact,The impact of several system parameters is investigated in this paper.,36746
the model ARM. Re-identification as We to proposed Metalearning refer Attention-based,We refer to the proposed Attention-based Re-identification Metalearning model as ARM.,36747
"As be accurate. such, RECIST must annotations","As such, RECIST annotations must be accurate.",36748
region method RECIST proposed stages: consists normalization of two The and estimation. lesion,The proposed method consists of two stages: lesion region normalization and RECIST estimation.,36749
learned and an in fashion. SHN both end-to-end be can STN,STN and SHN can both be learned in an end-to-end fashion.,36750
core applications. various in of tasks counting the is Crowd one surveillance,Crowd counting is one of the core tasks in various surveillance applications.,36751
"present problems. we both address this unified to work, In solution a","In this work, we present a unified solution to address both problems.",36752
accurate method signatures. verifying This online paper for presents an,This paper presents an accurate method for verifying online signatures.,36753
"a one-class modeled using are and signatures user's classifier. Finally, classified","Finally, user's signatures are modeled and classified using a one-class classifier.",36754
independent datasets proposed thanks on method signature The is self-taught learning. to,The proposed method is independent on signature datasets thanks to self-taught learning.,36755
hinders to rapid research reproduce prototyping hurdles results. researchers and This from poses,This hinders researchers from rapid prototyping and poses hurdles to reproduce research results.,36756
The paper this draft AIRLab and with analyses. of first presented performance snippets outlines code,The presented draft of this paper outlines AIRLab with first code snippets and performance analyses.,36757
more a exhaustive follow soon. introduction final A as version will,A more exhaustive introduction will follow as a final version soon.,36758
"e.g. Failure learning, cases black-box deep of","Failure cases of black-box deep learning, e.g.",36759
"might have adversarial severe in healthcare. consequences examples,","adversarial examples, might have severe consequences in healthcare.",36760
studied mostly the Yet calibrated attacks. real-world images of are in such failures with context,Yet such failures are mostly studied in the context of real-world images with calibrated attacks.,36761
"studies the demystify To adversarial examples, to need be designed. rigorous","To demystify the adversarial examples, rigorous studies need to be designed.",36762
"images. medical hinders such the study Unfortunately, complexity images directly of medical the design from","Unfortunately, complexity of the medical images hinders such study design directly from the medical images.",36763
domain-independent task planning. encoding present for a We general constraint-based,We present a general constraint-based encoding for domain-independent task planning.,36764
effects optional as expressed causal Task conditions planning and by is relationships of actions. characterized,Task planning is characterized by causal relationships expressed as conditions and effects of optional actions.,36765
It time-oriented scheduling. work of in in the spirit a constraint-based follows previous view,It follows a time-oriented view in the spirit of previous work in constraint-based scheduling.,36766
need accurate segmentation. skin This for lesion and underlines automatic for approach an the,This underlines the need for an accurate and automatic approach for skin lesion segmentation.,36767
"neural tackle network To SkinNet. issue, called (CNN) we a convolutional this propose","To tackle this issue, we propose a convolutional neural network (CNN) called SkinNet.",36768
is a of proposed CNN U-Net. The version modified,The proposed CNN is a modified version of U-Net.,36769
"strategies. injection of mode normalization that method collapse to we leads because the found However,","However, we found that the injection method leads to mode collapse because of normalization strategies.",36770
"biasing propose latent the information. on central to the normalization we inject criteria, code Based","Based on the criteria, we propose central biasing normalization to inject the latent code information.",36771
"expensive dietary prone methods errors. conventional are time-consuming, and The assessment to","The conventional dietary assessment methods are time-consuming, expensive and prone to errors.",36772
with and existing are over FOND empirically planners new planners existing compared The benchmarks. resulting,The resulting FOND planners are compared empirically with existing planners over existing and new benchmarks.,36773
scenarios. industrial vehicle are Connected deployed fleets worldwide IoT in several,Connected vehicle fleets are deployed worldwide in several industrial IoT scenarios.,36774
highlighted experimental proposed results. the The through several superiority is method of,The superiority of the proposed method is highlighted through several experimental results.,36775
"current Unfortunately, in problem benchmarks task and limited for diversity. are this size","Unfortunately, current benchmarks for this problem are limited in size and task diversity.",36776
on novel based a paper proposes This image multi-exposure fusion exposure method compensation.,This paper proposes a novel multi-exposure image fusion method based on exposure compensation.,36777
one fusion multi-exposure The image combined images are of compensated existing finally by methods.,The compensated images are finally combined by one of existing multi-exposure image fusion methods.,36778
building Games playing a (CCGs). is in Deck Collectible component crucial Card,Deck building is a crucial component in playing Collectible Card Games (CCGs).,36779
multimodal brings demonstrate information notable Practical framework using our incorporating that benefits. experiments,Practical experiments demonstrate that incorporating multimodal information using our framework brings notable benefits.,36780
is a thousand picture words. A worth,A picture is worth a thousand words.,36781
information exhibit contrasts and associated with the thus same anatomy capture similarities. These underlying,These contrasts capture information associated with the same underlying anatomy and thus exhibit similarities.,36782
a contrasts. dictionaries The stage multiple correlations capture that learns first of among group,The first stage learns a group of dictionaries that capture correlations among multiple contrasts.,36783
representative band extracts the and each as takes a globally It whole most bands.,It takes each band as a whole and globally extracts the most representative bands.,36784
"diverse. However, homogeneous whose features spectral regions to are different objects, different correspond","However, different homogeneous regions correspond to different objects, whose spectral features are diverse.",36785
"main classical PCA In four to models, has contrast properties. SuperPCA","In contrast to classical PCA models, SuperPCA has four main properties.",36786
"to classification leading resistant, discriminative, performance. are and The features improved resulting compact, HSI noise","The resulting features are discriminative, compact, and noise resistant, leading to improved HSI classification performance.",36787
systems. self-reproducing studying generation tool P. theory for Kabamba a as developed,P. Kabamba developed generation theory as a tool for studying self-reproducing systems.,36788
a exhibit theory. and also illustrating fixed-point self-replication connection between examples We,We also exhibit examples illustrating a connection between self-replication and fixed-point theory.,36789
"systems is comparison's OFDM presented For of conventional paper. the also performance the throughout sake,","For comparison's sake, the performance of conventional OFDM systems is also presented throughout the paper.",36790
results. and validates This complements currently available,This complements and validates currently available results.,36791
"supervises accurate motion, manner. that a in but learning sparse structure from","structure from motion, that supervises learning in a sparse but accurate manner.",36792
for The code publicly is available. the method proposed,The code for the proposed method is publicly available.,36793
"the data However, certain a when (i.e. body position from labeled","However, when the labeled data from a certain body position (i.e.",36794
"missing, how (i.e. domain) positions from to the target is data other leverage","target domain) is missing, how to leverage the data from other positions (i.e.",36795
of this labels domain) to position? the help activity source learn,source domain) to help learn the activity labels of this position?,36796
"selected domain, transfer source accurate With between perform domains. we knowledge need to the","With the selected source domain, we need to perform accurate knowledge transfer between domains.",36797
property. learn between distance only the global ignoring the domains Existing local methods while,Existing methods only learn the global distance between domains while ignoring the local property.,36798
property the based \textit{Stratified} local domains. our of on is capture distance to STL proposed,STL is based on our proposed \textit{Stratified} distance to capture the local property of domains.,36799
demonstrate STL. superiority on the experiments recognition public datasets of activity Extensive three,Extensive experiments on three public activity recognition datasets demonstrate the superiority of STL.,36800
an object vehicles. segmentation videos unmanned understanding by important generated plays in Primary role aerial,Primary object segmentation plays an important role in understanding videos generated by unmanned aerial vehicles.,36801
remains the Depth from estimation challenging a a image problem. wild in single,Depth estimation from a single image in the wild remains a challenging problem.,36802
images training high-quality One the main data the is lack of wild. obstacle in for,One main obstacle is the lack of high-quality training data for images in the wild.,36803
al.} our \textit{et work recent As (Holden in,As in our recent work (Holden \textit{et al.},36804
problem The is exist numerous regions which all that can object be potentially salient.,The problem is that numerous object regions exist which potentially can all be salient.,36805
most return which to limited contain a an regions are object. set likely of These,These return a limited set of regions which are most likely to contain an object.,36806
saliency used estimation methods have Several object proposals.,Several saliency estimation methods have used object proposals.,36807
"this object aim we In salient to detection. improve paper,","In this paper, we aim to improve salient object detection.",36808
several We context saliency propose computed proposals. which features the from are,We propose several saliency features which are computed from the context proposals.,36809
of AI Modern by and ever-increasing autonomy. high a robotic level are systems characterized and,Modern AI and robotic systems are characterized by a high and ever-increasing level of autonomy.,36810
pattern speckle of Recovering its the difficult. an phaseless is from image object,Recovering the image of an object from its phaseless speckle pattern is difficult.,36811
multiple media is imaging. scattering the matrix Let unknown in transmission alone,Let alone the transmission matrix is unknown in multiple scattering media imaging.,36812
"of During CVSs, productions between TWs moments of direction. tend to silence, reverse","During moments of silence, between productions of CVSs, TWs tend to reverse direction.",36813
"could be Within context, microalgae-based treatment alternative systems treatment plants. used this as","Within this context, microalgae-based treatment systems could be used as alternative treatment plants.",36814
Alzheimer chose for purpose. this disease We,We chose Alzheimer disease for this purpose.,36815
disease the Alzheimer loss. is dementia memory leading of cause and,Alzheimer disease is the leading cause of dementia and memory loss.,36816
"CNN, a Kernel Convolutions Convolutional Networks Networks. Enter Neural are of and Neural hybrid","Enter CNN, Convolutional Neural Networks are a hybrid of Kernel Convolutions and Neural Networks.",36817
"By problem can minimal solved using Convolutional Neural error Networks, with rate. be the","By using Convolutional Neural Networks, the problem can be solved with minimal error rate.",36818
is by stable dynamics simulation. be One to found molecular candidate structure,One candidate structure is found to be stable by molecular dynamics simulation.,36819
clear intuitive article of provide distinct and In descriptions measures. candidate six we this,In this article we provide clear and intuitive descriptions of six distinct candidate measures.,36820
"performance. with on simulated reveals excellent evaluation An truth, data, quantitative ground available","An evaluation on simulated data, with available ground truth, reveals excellent quantitative performance.",36821
"a associated and are with reconstruction estimate. crude The typically iterative, first seeded algorithms","The associated reconstruction algorithms are typically iterative, and seeded with a crude first estimate.",36822
"Coherent Sci. Imaging method (Pavlov et Bragg deterministic al., for Diffraction A","A deterministic method for Bragg Coherent Diffraction Imaging (Pavlov et al., Sci.",36823
comparison is the function autocorrelation performed. a as The with point starting appropriate,The appropriate comparison with the autocorrelation function as a starting point is performed.,36824
"utility has functions. achieved been using Traditionally, handcrafted this","Traditionally, this has been achieved using handcrafted utility functions.",36825
future utility viewpoints. propose usefulness a of better function the that to learn predicts We,We propose to learn a better utility function that predicts the usefulness of future viewpoints.,36826
explanations to discriminative learn discuss an class features image. textual in Current,Current textual explanations learn to discuss class discriminative features in an image.,36827
automatic metrics. using explanations counterfactual quantitatively both We proposed analyze and our then generated qualitatively,We then analyze our generated counterfactual explanations both qualitatively and quantitatively using proposed automatic metrics.,36828
"we given with TSWAP, graph undirected an vertices. are colored In","In TSWAP, we are given an undirected graph with colored vertices.",36829
in A token each colored placed is vertex.,A colored token is placed in each vertex.,36830
vertices. between of adjacent A tokens be pair swapped can,A pair of tokens can be swapped between adjacent vertices.,36831
problem. the swaps number The required the minimum in of is of optimization variant,The minimum number of swaps is required in the optimization variant of the problem.,36832
that between conditions state must consists in The local difference both problems (swaps/moves) transitions satisfy.,The difference between both problems consists in local conditions that state transitions (swaps/moves) must satisfy.,36833
This design TSWAP. first algorithms attempt for constitutes to the optimal solving,This constitutes the first attempt to design optimal solving algorithms for TSWAP.,36834
take attribute-based recent of seen past Neural Networks over this The research. has Convolutional field,The recent past has seen attribute-based Convolutional Neural Networks take over this field of research.,36835
"a other obtained the similar hand, On much results can smaller CNN? with be","On the other hand, can similar results be obtained with a much smaller CNN?",36836
goal answer questions. these paper The is of to an to this give,The goal of this paper is to give an answer to these questions.,36837
"values to need magrittr purrr, stored. intermediate with no Unlike be and","Unlike with magrittr and purrr, no intermediate values need to be stored.",36838
"implementations. favorable runtimes to fc benchmarked, comparison achieve composed in functions other When using","When benchmarked, functions composed using fc achieve favorable runtimes in comparison to other implementations.",36839
many artificial Dependence tasks an is for intelligence. in concept important,Dependence is an important concept for many tasks in artificial intelligence.,36840
task by the discarding executed be efficiently more something can task. from A independent,A task can be executed more efficiently by discarding something independent from the task.,36841
notions extension. apply update in two two We and well-known also issues: belief conservative these,We also apply these two notions in two well-known issues: belief update and conservative extension.,36842
"new update a formula-formula dependence. define Firstly, on we based operator","Firstly, we define a new update operator based on formula-formula dependence.",36843
"forgetting. formula to Furthermore, conservative we extension reduce","Furthermore, we reduce conservative extension to formula forgetting.",36844
"the discriminatory exhibit learned Nevertheless, do to applied when power. enough datasets, not features complex","Nevertheless, when applied to complex datasets, the learned features do not exhibit enough discriminatory power.",36845
"understanding, to a For variations comparison also study GDM. and better simple we","For comparison and better understanding, we also study a simple variations to GDM.",36846
is difference the key them The between function. discriminant,The key difference between them is the discriminant function.,36847
to limitations. space due optimization is Combinatorial feasible not search,Combinatorial optimization is not feasible due to search space limitations.,36848
images. maps depth single for both network A predicts image depth,A single image depth network predicts depth maps for both images.,36849
energy the of enhance power the intelligent for output potential harvesters. We intervention probe to,We probe the potential for intelligent intervention to enhance the power output of energy harvesters.,36850
electric a bi-resonant a piezo We case general harvester. study: investigate and principles,We investigate general principles and a case study: a bi-resonant piezo electric harvester.,36851
intelligent pre-programmed energy-conserving interventions consider We via operations. reversible,We consider intelligent interventions via pre-programmed reversible energy-conserving operations.,36852
"Compared characterization, compactness. dynamic motion and superior optical for exhibits imaging with efficiency flow","Compared with optical flow for motion characterization, dynamic imaging exhibits superior efficiency and compactness.",36853
"can be visual Accordingly, cues involved. view-tolerant more","Accordingly, more view-tolerant visual cues can be involved.",36854
is perform model learning CNN to feature A on multi-view images. novel proposed dynamic then,A novel CNN model is then proposed to perform feature learning on multi-view dynamic images.,36855
"performance challenging experiments, state-of-the-art three achieve approach In datasets. the can on proposed","In experiments, the proposed approach can achieve state-of-the-art performance on three challenging datasets.",36856
stages: of two reconstruction. coding proposed coupled coupled image sparse approach The and denoising consists,The proposed coupled image denoising approach consists of two stages: coupled sparse coding and reconstruction.,36857
"Therefore, information. are significant devoted resources spatiotemporal brain process to","Therefore, significant brain resources are devoted to process spatiotemporal information.",36858
"biological compaction are Nevertheless, still lacking. empirical approaches the tackling soundness of time","Nevertheless, empirical approaches tackling the biological soundness of time compaction are still lacking.",36859
"Meanwhile, control with performed the previous of regardless women men, par on conditioning.","Meanwhile, women performed on par with control men, regardless of the previous conditioning.",36860
on for focused GAN-based solution face the work we synthesis. this attribute In guided,In this work we focused on GAN-based solution for the attribute guided face synthesis.,36861
"steganalysis. has deep its shown in Recently, learning power","Recently, deep learning has shown its power in steganalysis.",36862
"the filters learned. meantime, automatically high-pass In also are the","In the meantime, the high-pass filters are also automatically learned.",36863
architecture. The proposed prove of the experimental results the effectiveness,The experimental results prove the effectiveness of the proposed architecture.,36864
do-calculus. the formula We of provide Front-Door a the adjustment using the proof,We provide a proof of the the Front-Door adjustment formula using the do-calculus.,36865
basic for algorithms We revision and update ACDFs. propose,We propose basic revision and update algorithms for ACDFs.,36866
"static knowledge, call common handle We propositional which we also constraints.","We also handle static propositional common knowledge, which we call constraints.",36867
results Our show viability of experimental the our approach.,Our experimental results show the viability of our approach.,36868
"our connectedness that while maintaining of suggest Finally, increasing limits the despotism hierarchy. followers results","Finally, our results suggest that increasing the connectedness of followers limits despotism while maintaining hierarchy.",36869
"prediction. the recognition in survey In and techniques complete the paper, this we action state-of-the-art","In this paper, we survey the complete state-of-the-art techniques in the action recognition and prediction.",36870
Deep Imaging. compressed learning have Resonance shown Magnetic sensing-based performance for promising approaches,Deep learning approaches have shown promising performance for compressed sensing-based Magnetic Resonance Imaging.,36871
"Recently, have visually results. shown to more appealing achieve loss functions adversarial and perceptual been","Recently, adversarial and perceptual loss functions have been shown to achieve more visually appealing results.",36872
different We and places and excavate the scenes. memorize of semantic stable structure,We excavate and memorize the stable semantic structure of different places and scenes.,36873
place standard The state-of-the-art on encoder intrinsic proposed recognition outperforms the methods benchmark image-level Nordland.,The proposed intrinsic encoder outperforms the state-of-the-art image-level place recognition methods on standard benchmark Nordland.,36874
experimental approach TSE the The outperforms proposed state-of-the-art results demonstrate methods. that,The experimental results demonstrate that the proposed approach outperforms state-of-the-art TSE methods.,36875
is texture network. as a dynamic The directed modeled,The dynamic texture is modeled as a directed network.,36876
"For vertex each is estimated. for transformation, activity network each the","For each network transformation, the activity for each vertex is estimated.",36877
"the activity concatenating is by Then, histograms. texture descriptor constructed","Then, texture descriptor is constructed by concatenating the activity histograms.",36878
in textures tend These provide better dynamic to performance classification.,These tend to provide better performance in dynamic textures classification.,36879
exist map that classify models and to data. learn supervised Other,Other models exist that learn to map and classify supervised data.,36880
Humans recognition. cognitive pattern in many including well tasks remarkably perform,Humans perform remarkably well in many cognitive tasks including pattern recognition.,36881
"understood. this process mechanisms are the However, neuronal well underlying not","However, the neuronal mechanisms underlying this process are not well understood.",36882
"a temporal be span considered to feature this each can sense, axis. In single","In this sense, each feature can be considered to span a single temporal axis.",36883
obtained data a We motor from cognitive our to applied task. proposed inhibitory scheme experimental,We applied our proposed scheme to experimental data obtained from a motor inhibitory cognitive task.,36884
"The the our good test classification indicating exhibits performance, of adequateness approach.","The test exhibits good classification performance, indicating the adequateness of our approach.",36885
learning. on optimization prior modeling model-based either mainly or through image inference They focus discriminative,They mainly focus on modeling image prior through either model-based optimization or discriminative inference learning.,36886
of flexibility learning levels annotation. our model from data The of enables different with joint,The flexibility of our model enables joint learning from data with different levels of annotation.,36887
architecture cognition a complex process by facilitated brain. human the intricate of is Human,Human cognition is a complex process facilitated by the intricate architecture of human brain.,36888
conceptual Human formal of classic cognition violate built. on which often probabilities representation are spaces,Human cognition often violate classic probabilities on which formal representation of conceptual spaces are built.,36889
"proposed Also, achieves cognition, particular, the consciousness. model in","Also, the proposed model achieves cognition, in particular, consciousness.",36890
constructive this achieving consciousness have process demonstrated learning with of scenario. a We,We have demonstrated this process of achieving consciousness with a constructive learning scenario.,36891
representation severity. is that the interpretable The learns a subject-level model of clinically reflective disease,The model learns a clinically interpretable subject-level representation that is reflective of the disease severity.,36892
Our for model state-of-the COPD. of severity performance measures predicting art gives in clinical,Our model gives state-of-the art performance in predicting clinical measures of severity for COPD.,36893
of in video. paper human actions localization spatio-temporal This addresses,This paper addresses spatio-temporal localization of human actions in video.,36894
gain is mainly improved localization. shown to to temporal due The be,The gain is shown to be mainly due to improved temporal localization.,36895
learning deep problem. this two encoder-decoder develop architectures address We to based,We develop two encoder-decoder based deep learning architectures to address this problem.,36896
imaging destructive capacitive paper a non This coplanar non describes invasive device.,This paper describes a coplanar non invasive non destructive capacitive imaging device.,36897
"cortical by with open surfaces is described Specifically, two triangulated boundaries. ribbon the","Specifically, the cortical ribbon is described by two triangulated surfaces with open boundaries.",36898
approach for great offers potential morphometry. The cortical,The approach offers great potential for cortical morphometry.,36899
is input challenging single-view essential from images problem. yet an Generating a multi-view,Generating multi-view images from a single-view input is an essential yet challenging problem.,36900
"and has graphics, in vision, It broad applications robotics.","It has broad applications in vision, graphics, and robotics.",36901
to propose problem. this We CR-GAN address,We propose CR-GAN to address this problem.,36902
aggregate features then these approach. We a through bag-of-word,We then aggregate these features through a bag-of-word approach.,36903
then previewed. draped and parameters can edited recovered updated the further The be garment,The recovered parameters can then be further edited and the updated draped garment previewed.,36904
further of quantitative and proposed effectiveness evaluations and the robustness the qualitative work. Both demonstrate,Both quantitative and qualitative evaluations further demonstrate the effectiveness and robustness of the proposed work.,36905
"merely of not are the similar. shapes slowest the nearly that identical, eigenvectors find We","We find that the shapes of the slowest eigenvectors are nearly identical, not merely similar.",36906
provided. updated review literature also is brief A,A brief updated literature review is also provided.,36907
show high AIT for The SNR. outperforms WM that results,The results show that WM outperforms AIT for high SNR.,36908
"a estimation MMA accurate. feasible DoA In and single with is conclusion,","In conclusion, DoA estimation with a single MMA is feasible and accurate.",36909
results modular to some Our on approaches lights shed knowledge compilation.,Our results shed some lights on modular approaches to knowledge compilation.,36910
"results apply epistemic we Furthermore, planning. the our in multi-agent","Furthermore, we apply our results in the multi-agent epistemic planning.",36911
"In samples annotated via addition, a selected and low-confidence few AL. with predictions are","In addition, a few samples with low-confidence predictions are selected and annotated via AL.",36912
We to coefficients. rise that gives a assume common distribution regression the regime-specific,We assume that a common distribution gives rise to the regime-specific regression coefficients.,36913
These flexible across coefficients a shrinkage which manner. in to enable differ regimes priors assess,These shrinkage priors enable to assess which coefficients differ across regimes in a flexible manner.,36914
may thousands false cases. such in alarms of Methods yield,Methods may yield thousands of false alarms in such cases.,36915
We CT pulmonary algorithm of scans. the a present novel for registration,We present a novel algorithm for the registration of pulmonary CT scans.,36916
the Both fields driven data gradient term. are normalized by steps same,Both steps are driven by the same normalized gradient fields data term.,36917
an paper proposes for This efficient digit Devanagari recognition handwritten script. system,This paper proposes an efficient digit recognition system for handwritten Devanagari script.,36918
"provide Observation multitude sensing heterogeneous Earth of Nowadays, a remote data. systems","Nowadays, Earth Observation systems provide a multitude of heterogeneous remote sensing data.",36919
labor-intensive phenotypic information is collection Traditional and time-consuming. of,Traditional collection of phenotypic information is labor-intensive and time-consuming.,36920
is of imagery phenotyping. popular for becoming Use,Use of imagery is becoming popular for phenotyping.,36921
then from rectified the mosaic A generated is imagery. orthophoto,A rectified orthophoto mosaic is then generated from the imagery.,36922
the plant. each statistical individual to find sorghum model present location We a of,We present a statistical model to find the location of each individual sorghum plant.,36923
we recognition devices. the with way interact natural Gesture currently enables extension a of,Gesture recognition enables a natural extension of the way we currently interact with devices.,36924
"end, propose classifier a gesture a modeling probabilistic approach. hierarchical based we To this on","To this end, we propose a gesture classifier based on a hierarchical probabilistic modeling approach.",36925
"strictly increasing symmetries, symmetries sizes. compute generalizes many more for BV can block VV and","BV strictly generalizes VV symmetries, and can compute many more symmetries for increasing block sizes.",36926
in regime driven encoding. population fluctuation the ultrafast realize can Cortical neurons,Cortical neurons in the fluctuation driven regime can realize ultrafast population encoding.,36927
"well mechanisms, The underlying however, biophysical not understood. are","The underlying biophysical mechanisms, however, are not well understood.",36928
dissociation explanation provide for ultrafast sharp and of action response. the an potential We onset,We provide an explanation for the dissociation of sharp action potential onset and ultrafast response.,36929
"to datasets. up work noisily-annotated learning recognition Much visual recent to aims scale on massive,","Much recent work on visual recognition aims to scale up learning to massive, noisily-annotated datasets.",36930
include representations. the domain-specific fashion in advances development numerical of Recent computational,Recent advances in computational fashion include the development of domain-specific numerical representations.,36931
"that from purpose, firstly extracted grid are the optical-flow, cuboid virtual-inertia video. log-covariance, For feature,","For that purpose, firstly grid optical-flow, virtual-inertia feature, log-covariance, cuboid are extracted from the video.",36932
The datasets. was proposed of evaluated egocentric a on framework number,The proposed framework was evaluated on a number of egocentric datasets.,36933
that outperforms the multi-modal The showed existing features using MKL results with methods.,The results showed that using multi-modal features with MKL outperforms the existing methods.,36934
on benchmarks Extensive show of our approach. four the effectiveness challenging experiments,Extensive experiments on four challenging benchmarks show the effectiveness of our approach.,36935
limit severely variations appearance-based the performance tracking. Personal of gaze,Personal variations severely limit the performance of appearance-based gaze tracking.,36936
standard model network is difficult. neural using Adapting methods variations these to adaptation,Adapting to these variations using standard neural network model adaptation methods is difficult.,36937
these by tackle Adaptive the GaZe (SPAZE). introducing Estimator We problems SPatial,We tackle these problems by introducing the SPatial Adaptive GaZe Estimator (SPAZE).,36938
Calibrating reduces problem. SPAZE new for optimization a person a solving small to,Calibrating SPAZE for a new person reduces to solving a small optimization problem.,36939
examining gaze model-based expected We to low-dimensionality that tracking. approaches by show is this,We show that this low-dimensionality is expected by examining model-based approaches to gaze tracking.,36940
show possible. accurate that We head pose-free tracking gaze is also,We also show that accurate head pose-free gaze tracking is possible.,36941
applications. instrumental variety becoming Intelligence of is a Artificial in,Artificial Intelligence is becoming instrumental in a variety of applications.,36942
Context recognition. accurate for visual important is,Context is important for accurate visual recognition.,36943
state. graphical Faster object infer to which model a aims R-CNN) with,Faster R-CNN) with a graphical model which aims to infer object state.,36944
connectivity a A network scale-free is density if is proportional power-law function to distribution. its,A network is scale-free if its connectivity density function is proportional to a power-law distribution.,36945
challenge. tracing path scattering using participating highly media brute a is force Rendering,Rendering highly scattering participating media using brute force path tracing is a challenge.,36946
differential The solving simple to linear reduces the diffusion problem partial a equation. approximation,The diffusion approximation reduces the problem to solving a simple linear partial differential equation.,36947
"method problems. discuss prior to our Finally, how compares we standard work for various","Finally, we discuss how our method compares to prior work for various standard problems.",36948
each two screening projections of breast. for Standard mammography breast X-ray involves the cancer acquisition,Standard breast cancer screening involves the acquisition of two mammography X-ray projections for each breast.,36949
"the views tumor supports a and task Typically, of detection localization. challenging comparison of both","Typically, a comparison of both views supports the challenging task of tumor detection and localization.",36950
"Siamese for network mammography. in dual-view introduce lesion learning, We matching patch-based a deep","We introduce a deep learning, patch-based Siamese network for lesion matching in dual-view mammography.",36951
"the matching evaluated of contribution the a pipeline. detection in we Finally, full network","Finally, we evaluated the contribution of the matching network in a full detection pipeline.",36952
our approach. improved of using promise results Experimental the accuracy detection demonstrate,Experimental results demonstrate the promise of improved detection accuracy using our approach.,36953
social form daily of the our in the life. basis network,form the basis of the social network in our daily life.,36954
"these cues studies. largely additional by are the previous However, overlooked","However, these additional cues are largely overlooked by the previous studies.",36955
We neural networks. classification deep image propose new a with framework for,We propose a new framework for image classification with deep neural networks.,36956
The to intermediate the framework outputs computational network. introduces graph a of,The framework introduces intermediate outputs to the computational graph of a network.,36957
publicly benchmarks. been have up-to-date several available Experiments on conducted,Experiments have been conducted on several publicly available up-to-date benchmarks.,36958
module PointSIFT integrated into improve architecture various the can to PointNet-based be ability. representation,PointSIFT module can be integrated into various PointNet-based architecture to improve the representation ability.,36959
datasets. our Extensive outperforms framework show experiments PointSIFT-based method benchmark on standard state-of-the-art,Extensive experiments show our PointSIFT-based framework outperforms state-of-the-art method on standard benchmark datasets.,36960
code accompanied and trained model by The be this paper. will published,The code and trained model will be published accompanied by this paper.,36961
and training of biases methods exploit known Most the capture learning machine are data. to,Most machine learning methods are known to capture and exploit biases of the training data.,36962
"for harmful. While others beneficial learning, are some biases are","While some biases are beneficial for learning, others are harmful.",36963
"training present to Specifically, in captioning models tend data. image exaggerate biases","Specifically, image captioning models tend to exaggerate biases present in training data.",36964
caption of specific gender words (e.g. investigate generation We,We investigate generation of gender specific caption words (e.g.,36965
"on the context. or appearance the man, person's image woman) based","man, woman) based on the person's appearance or the image context.",36966
have Species high also abundant. which more are influence a selection on generally,Species which have a high influence on selection are also generally more abundant.,36967
behavior from one How friendly can adversarial detect and raw data?,How can one detect friendly and adversarial behavior from raw data?,36968
active for century study field of Neuroethology been an has more now. than a,Neuroethology has been an active field of study for more than a century now.,36969
"is embedding ignored. underlying space the However, feature","However, the underlying feature embedding space is ignored.",36970
A also sampling balanced introduced. is strategy,A balanced sampling strategy is also introduced.,36971
will be code of released. The trained model our and weights,The code and trained weights of our model will be released.,36972
our understand behind lagged ability to has data always ability it. Our to collect,Our ability to understand data has always lagged behind our ability to collect it.,36973
of the areas. over system of variant is new urban variety The evaluated a,The new variant of the system is evaluated over a variety of urban areas.,36974
Computer and aided this make can processing image faster accurate. more,Computer aided image processing can make this faster and more accurate.,36975
under by I begin key ideas discussion. the paper summarizing of,I begin by summarizing key ideas of the paper under discussion.,36976
has for compression many been investigated decades. Image,Image compression has been investigated for many decades.,36977
"by CAEs compact than efficiency results, features. achieve coding experimental extracting to According better JPEG","According to experimental results, CAEs achieve better coding efficiency than JPEG by extracting compact features.",36978
and ratio quality subjective high compression reconstruction. on GANs advantages large potential show,GANs show potential advantages on large compression ratio and high subjective quality reconstruction.,36979
"to among performance (RD) the best them, comparable achieves BPG. is rate-distortion Super-resolution which","Super-resolution achieves the best rate-distortion (RD) performance among them, which is comparable to BPG.",36980
the between and elements visual is of a video. a natural There correlation auditive,There is a natural correlation between the visual and auditive elements of a video.,36981
"transform constructed on the fractional In the (FRWT). FrScatNet is approach, our based Methods: wavelet","Methods: In our approach, the FrScatNet is constructed based on the fractional wavelet transform (FRWT).",36982
and scattering coefficients modulus iteratively operators. fractional FRWTs using The computed are,The fractional scattering coefficients are iteratively computed using FRWTs and modulus operators.,36983
used signal fractional constructed feature The scattering usually for vectors are classification. coefficients by,The feature vectors constructed by fractional scattering coefficients are usually used for signal classification.,36984
and we classify background into Then those target or patches malignant) groups. (benign,Then we classify those patches into target (benign or malignant) and background groups.,36985
to Conclusion: and FrScatNet accurate The is robust achieve shown results.,Conclusion: The FrScatNet is shown to achieve accurate and robust results.,36986
work. in discriminative by this obtained the and are scattering stable FrScatNet More coefficients fractional,More stable and discriminative fractional scattering coefficients are obtained by the FrScatNet in this work.,36987
are observed signals by The corrupted noise. usually,The observed signals are usually corrupted by noise.,36988
that heavy processing different the observed signal In be may tailed. applications errors it is,In different signal processing applications it is observed that the errors may be heavy tailed.,36989
It both are has estimators strongly been consistent. that the observed,It has been observed that both the estimators are strongly consistent.,36990
the estimators. asymptotic as have squares least distribution least The the squares same estimators approximate,The approximate least squares estimators have the same asymptotic distribution as the least squares estimators.,36991
proposed simulations some see how numerical work. the to perform estimators We,We perform some numerical simulations to see how the proposed estimators work.,36992
"mesh. a such volumetric requires discretization, shapes, conforming this as a For tetrahedral boundary solid","For solid shapes, this requires a volumetric discretization, such as a boundary conforming tetrahedral mesh.",36993
Experiments approach of benchmarks the on show our methods. advantages over several state-of-art,Experiments on several benchmarks show the advantages of our approach over state-of-art methods.,36994
our results proposed baselines. benchmark method datasets outperforms state-of-the-art four on demonstrate that Experimental the,Experimental results on four benchmark datasets demonstrate that our proposed method outperforms the state-of-the-art baselines.,36995
expensive. time-consuming both and Manual is annotation,Manual annotation is both time-consuming and expensive.,36996
There three the in dataset. challenges are,There are three challenges in the dataset.,36997
"sparse naturalistic a behaviors First, setting. are driving driving in predefined","First, predefined driving behaviors are sparse in a naturalistic driving setting.",36998
"of distribution is Second, behaviors driving long-tail. the","Second, the distribution of driving behaviors is long-tail.",36999
"variation is observed. a intra-class Third, huge","Third, a huge intra-class variation is observed.",37000
one vision. computer Video is challenging most of the in topics understanding,Video understanding is one of the most challenging topics in computer vision.,37001
video significantly. oriented functional object Experiments network that understanding using improves show,Experiments show that using functional object oriented network improves video understanding significantly.,37002
"the algorithm queries Further, forward feasible. hindsight to our beginning backward relational makes very","Further, our relational forward backward algorithm makes hindsight queries to the very beginning feasible.",37003
"impinging tensor-based approach, nonlinear considered Decomposition We signals are Polyadic on (CPD). named EEG Canonical","We considered EEG signals are impinging on tensor-based approach, named nonlinear Canonical Polyadic Decomposition (CPD).",37004
"task. this during EEG collected a signals In resting-state study, we","In this study, we collected EEG signals during a resting-state task.",37005
applications. reveal the temporal possibility signals findings Our recoverable of in EEG,Our findings reveal the possibility of recoverable temporal signals in EEG applications.,37006
(FCN) estimated attributes via geometry Network Such Fully model. Convolutional are a,Such geometry attributes are estimated via a Fully Convolutional Network (FCN) model.,37007
sensitive to to initialization. be is known Gauss-Newton also,Gauss-Newton is also known to be sensitive to initialization.,37008
"far situation is from ideal. Hence, the","Hence, the situation is far from ideal.",37009
"advantages tasks, two i.e. of The SISO leveraging simple framework segmentation yet is efficient, different","The SISO framework is simple yet efficient, leveraging advantages of two different segmentation tasks, i.e.",37010
the salient segmentation semantic fuse instance final eventually segmentation to and them for object result.,semantic instance segmentation and salient object segmentation to eventually fuse them for the final result.,37011
"can in results which our demonstrated occlusions effectiveness videos. handle of baseline, Experimental SISO the","Experimental results demonstrated the effectiveness of our SISO baseline, which can handle occlusions in videos.",37012
candidate problem to addressed embedding establish jointly generally This by and is correspondence. statements images,This problem is generally addressed by jointly embedding images and candidate statements to establish correspondence.,37013
loss function. by embedding based is a attention model guided max-margin end-to-end learned Our,Our attention based embedding model is learned end-to-end guided by a max-margin loss function.,37014
suggested that observations preliminary sternum These the through endochondral ossification. quail grows,These preliminary observations suggested that the quail sternum grows through endochondral ossification.,37015
sternum development. will histological during study of explore changes post-hatching The future the further quail,The future study will further explore the histological changes of quail sternum during post-hatching development.,37016
"by a Wiener estimation. obtained is reconstructed Then, image","Then, a reconstructed image is obtained by Wiener estimation.",37017
Models the network easily can be on of this analyzed kind level.,Models of this kind can be easily analyzed on the network level.,37018
with well. model By this regression approach selection we as to can extend substituting classification,By substituting regression with classification we can extend this approach to model selection as well.,37019
accurate short problem This of of semantic videos. work labelling the addresses,This work addresses the problem of accurate semantic labelling of short videos.,37020
"and evolution Another goal is competition. under speciation ""maximal"" sympatric study to","Another goal is to study evolution and sympatric speciation under ""maximal"" competition.",37021
"the the all stochastic ""toy essentially model. standard captures of features the model"" Hence,","Hence, the ""toy model"" essentially captures all the features of the standard stochastic model.",37022
separately. bottom-up incorporate or generally norm models in Saliency either top-down cues salient,Saliency models generally incorporate salient cues either in bottom-up or top-down norm separately.,37023
integration compared with state-of-the-art-models. yields the saliency several promising RGB-D results framework,RGB-D saliency integration framework yields promising results compared with the several state-of-the-art-models.,37024
domains tackle proposed to was generation the CoGAN. The first data two attempt of by,The first attempt to tackle the data generation of two domains was proposed by CoGAN.,37025
"vulnerable domain inherently of is various for solution their similarities. levels However,","However, their solution is inherently vulnerable for various levels of domain similarities.",37026
two-step model an consisting end-to-end a network. is a attention of DSaH network and hashing,DSaH is a two-step end-to-end model consisting of an attention network and a hashing network.,37027
both We performance general fine-grained conduct on extensive experiments evaluation. datasets retrieval for and,We conduct extensive experiments on both fine-grained and general retrieval datasets for performance evaluation.,37028
performance Machine methods applied in learning applications. been and have widely achieved good various real-world,Machine learning methods have achieved good performance and been widely applied in various real-world applications.,37029
be can and time-consuming. labor-intensive process This,This process can be labor-intensive and time-consuming.,37030
indicate results exhibits outstanding both Experimental qualitatively that performance and quantitatively. our model,Experimental results indicate that our model exhibits outstanding performance both qualitatively and quantitatively.,37031
We and classes thresholds that best vary demonstrate among detectors. also significantly the confidence score,We also demonstrate that the best confidence score thresholds vary significantly among classes and detectors.,37032
other code source datasets to can Our easily as adapted well. be,Our source code can easily be adapted to other datasets as well.,37033
paper image. object addresses of given the an viewpoint a estimation problem of in This,This paper addresses the problem of viewpoint estimation of an object in a given image.,37034
types New data and (ii) added on. trained are of,(ii) New types of data are added and trained on.,37035
"several methods. lanes properties unique the of challenge detection However,","However, several unique properties of lanes challenge the detection methods.",37036
"further and merging, single, performance. lines broken, splitting the hamper solid, double,","solid, broken, single, double, merging, and splitting lines further hamper the performance.",37037
outstanding on our world performances that traffic LaneNet validate delivers consistently real Experiments scenarios.,Experiments validate that our LaneNet consistently delivers outstanding performances on real world traffic scenarios.,37038
constraints are that constraint stochastic component. a problems optimization with objectives These or involve,These are constraint optimization problems that involve objectives or constraints with a stochastic component.,37039
exact classical We probabilistic (PGMs). revisit models problem graphical on of inference the,We revisit the classical problem of exact inference on probabilistic graphical models (PGMs).,37040
"time-consuming The manual issue labelling applied is the typically main process, per image.","The main issue is the time-consuming manual labelling process, typically applied per image.",37041
itself car the is annotation. driving that form of notice a We,We notice that driving the car is itself a form of annotation.,37042
"(MSIs). multispectral image MSFA is demosaicked reconstruct a images Therein, using captured mosaicked to an","Therein, a mosaicked image captured using an MSFA is demosaicked to reconstruct multispectral images (MSIs).",37043
"reconstructed obtained a estimation. Wiener using image is Subsequently,","Subsequently, a reconstructed image is obtained using Wiener estimation.",37044
environments. game in in evolutionary traditionally theory interactions homogeneous assume Models symmetric,Models in evolutionary game theory traditionally assume symmetric interactions in homogeneous environments.,37045
in NMDA to transmission the nervous (NMDA-R) excitatory central contribute receptors synaptic typically system.,NMDA receptors (NMDA-R) typically contribute to excitatory synaptic transmission in the central nervous system.,37046
mechanism has theoretically. been This yet studied not,This mechanism has not yet been studied theoretically.,37047
"generalize to the hands. Therefore, unseen new, to existing problems has method hybrid","Therefore, the existing hybrid method has problems to generalize to new, unseen hands.",37048
"work, kinematic shape to the hand learnable. we make the layer parameters In extend this","In this work, we extend the kinematic layer to make the hand shape parameters learnable.",37049
"the In this arbitrary way, hand can generalize learnt shapes. towards network","In this way, the learnt network can generalize towards arbitrary hand shapes.",37050
"and rotated, input globally to appearance. The a scaled similar images shifted, are","The input images are shifted, rotated, and globally scaled to a similar appearance.",37051
"Possibly universal, lacks explanation. BPC a mechanistic presently","Possibly universal, BPC presently lacks a mechanistic explanation.",37052
"of electroporation versions (EP) fail model"" Multiple to cell the ""standard of account for BPC.","Multiple versions of the ""standard model"" of cell electroporation (EP) fail to account for BPC.",37053
relocate the enter or outside membrane We from that molecules a hypothesize within occluding pore.,We hypothesize that occluding molecules from outside the membrane enter or relocate within a pore.,37054
"involving additional fundamental and the combination Significantly, is general, hindrance. and of a partitioning mechanism","Significantly, the additional mechanism is fundamental and general, involving a combination of partitioning and hindrance.",37055
an model. the such of requires extension for standard behavior Accounting,Accounting for such behavior requires an extension of the standard model.,37056
single on commercial societies. strong modern has clothes a image from and impacts Understanding cultural,Understanding clothes from a single image has strong commercial and cultural impacts on modern societies.,37057
detection propose MetaAnchor for flexible We frameworks. anchor and named a novel mechanism object,We propose a novel and flexible anchor mechanism named MetaAnchor for object detection frameworks.,37058
by introduced et ago were al. These years Berzuini methods over twenty,These methods were introduced over twenty years ago by Berzuini et al.,37059
"multiple for heterogeneous sensors In holds from particular, data involving characteristics. this with applications","In particular, this holds for applications involving data from multiple sensors with heterogeneous characteristics.",37060
to been models understand have theoretical epidemics. of Various basic nature proposed the,Various theoretical models have been proposed to understand the basic nature of epidemics.,37061
epidemic mobility effects Recent the focus to process. of studies on,Recent studies focus on the effects of mobility to epidemic process.,37062
"However, movement. random uncorrelated assumed typically is as of type the walk","However, uncorrelated random walk is typically assumed as the type of movement.",37063
"focus the we of of Moreover, mobility effect agents. reduced on infected the mainly","Moreover, we mainly focus on the effect of the reduced mobility of infected agents.",37064
"that depended the on of movement agents. type However, crucially","However, that crucially depended on the movement type of agents.",37065
This when medical to comes common is it case dataset.,This is common case when it comes to medical dataset.,37066
analyses. others software apply publicly to available We for in made our their,We made our software publicly available for others to apply in their analyses.,37067
diverse scenarios. proposed performance in of method is our The demonstrated,The performance of our proposed method is demonstrated in diverse scenarios.,37068
state-of-of-the-art the the object solutions. Both with are estimation compared ego-motion localization and,Both the ego-motion estimation and object localization are compared with the state-of-of-the-art solutions.,37069
while lexicographical This defining takes of actions challenge person a eating. paper the considers,This paper considers the lexicographical challenge of defining actions a person takes while eating.,37070
repeatable is gesture and discernible goal definitions to objective establish intent. on based The,The goal is to establish objective and repeatable gesture definitions based on discernible intent.,37071
subsequently simultaneously reviewed was to and recorded Video label gestures.,Video was simultaneously recorded and subsequently reviewed to label gestures.,37072
"and calculated in boundary terms ambiguity, was mistakes. of Inter-rater agreement, reliability","Inter-rater reliability was calculated in terms of agreement, boundary ambiguity, and mistakes.",37073
disparity This has the greatly estimated accuracy the of maps. improved,This has greatly improved the accuracy of the estimated disparity maps.,37074
stereo a implementation GPU. presents fast state-of-the-art This paper the on bilateral of,This paper presents the implementation of fast bilateral stereo on a state-of-the-art GPU.,37075
deployment on a based method K-means Then clustering. investigate we,Then we investigate a deployment method based on K-means clustering.,37076
iterative is An efficient to reduced further technique improve power coverage probability power. with proposed,An iterative power efficient technique is further proposed to improve coverage probability with reduced power.,37077
language is algorithm implemented C The in near with a real-time performance.,The algorithm is implemented in C language with a near real-time performance.,37078
for polarization single presented results are a Numerical component.,Numerical results are presented for a single polarization component.,37079
"applied redundant remove Finally, is non-maximum rotational to suppression detections.","Finally, rotational non-maximum suppression is applied to remove redundant detections.",37080
color space. This using presents face posteriori a recognition maximum paper on (MAP) discriminant YCbCr,This paper presents face recognition using maximum a posteriori (MAP) discriminant on YCbCr color space.,37081
"of it times grayscale method. that computation more than time requires based However, three","However, it requires three times more computation time than that of grayscale based method.",37082
traversals spaces. lattice the on conceptual are Concept lattice represented used,Concept lattice traversals are used on the lattice represented conceptual spaces.,37083
challenged might humans. to prosthetics design well assist to This prove intelligent,This might well prove to design intelligent prosthetics to assist challenged humans.,37084
"hand with discussion conceptual we bidirectional extend our of Also, our on insights spaces prosthetics.","Also, we extend our discussion with our insights on conceptual spaces of bidirectional hand prosthetics.",37085
can body. snail on inversion Gene generate chiral the mutation,Gene mutation can generate chiral inversion on the snail body.,37086
sinistral A ecological population in this context. thrive can snail,A sinistral snail population can thrive in this ecological context.,37087
Pareas show asymmetry species Distinct different bilateral degrees.,Distinct Pareas species show different bilateral asymmetry degrees.,37088
parameters. also regimes the space in is different for the Stability of solutions of analyzed,Stability of the solutions is also analyzed for different regimes in the space of parameters.,37089
"e.g. literature, There fusion are in architectures several the","There are several fusion architectures in the literature, e.g.",37090
"conditions. even of weather Therefore, approaches the performance decreases these or adverse in fails enormously","Therefore, the performance of these approaches decreases enormously or even fails in adverse weather conditions.",37091
is Convolutional of (CNN) super-resolution depth network image neural (SR). crucial importance for,Convolutional neural network (CNN) depth is of crucial importance for image super-resolution (SR).,37092
"observe to deeper that more we for networks are difficult train. image However, SR","However, we observe that deeper networks for image SR are more difficult to train.",37093
"attention very propose solve the we (RCAN). networks channel deep problems, To residual these","To solve these problems, we propose the very deep residual channel attention networks (RCAN).",37094
connections. short skip group with residual some Each contains blocks residual,Each residual group contains some residual blocks with short skip connections.,37095
improvements and achieves visual show methods. Extensive better that state-of-the-art our RCAN accuracy experiments against,Extensive experiments show that our RCAN achieves better accuracy and visual improvements against state-of-the-art methods.,37096
at criminal scene may information investigations. Animal forensic hair in examination provide a valuable,Animal hair examination at a criminal scene may provide valuable information in forensic investigations.,37097
"databases local However, are identification hair animal for rare. reference","However, local reference databases for animal hair identification are rare.",37098
diameter medulla differ of margins the according and species. the hair the The to animal,The diameter of the hair medulla and the margins differ according to the animal species.,37099
natural is of description the process of a language image. an generating Image captioning,Image captioning is the process of generating a natural language description of an image.,37100
can recognition and load systems. as issue challenging be important for Traffic categorized such an,Traffic load recognition can be categorized as an important and challenging issue for such systems.,37101
KITTI the depth predictions for deep Our state-of-the-art monocular excel on approaches benchmark.,Our deep predictions excel state-of-the-art approaches for monocular depth on the KITTI benchmark.,37102
"tracking inter-dependence minimum of and The Bayesian classification risk. is the utilized detection, then with","The inter-dependence of detection, tracking and classification is then utilized with the minimum Bayesian risk.",37103
are simulations. numerical superiority using The verified effectiveness of the and proposed approach,The effectiveness and superiority of the proposed approach are verified using numerical simulations.,37104
"the diseases of In reparation. mechanisms impair endothelial addition,chronic","In addition,chronic diseases impair the mechanisms of endothelial reparation.",37105
"conclusion,sepsis,aging dysfunction. chronic features endothelial and induce of In similar diseases","In conclusion,sepsis,aging and chronic diseases induce similar features of endothelial dysfunction.",37106
an to approach using convolutional We analysis scene networks. semantic deep present,We present an approach to semantic scene analysis using deep convolutional networks.,37107
"to method approaches, contrast surface our volumetric on In operates geometry. directly","In contrast to volumetric approaches, our method operates directly on surface geometry.",37108
"data. point Crucially, other to noisy is construction the real-world clouds and applicable unstructured","Crucially, the construction is applicable to unstructured point clouds and other noisy real-world data.",37109
algorithm The of types alternates updates. between Metropolis-Hastings two,The Metropolis-Hastings algorithm alternates between two types of updates.,37110
"temporal variations these occur spatial gradually observe and dynamics. to we However, relative that","However, we observe that these variations occur gradually relative to spatial and temporal dynamics.",37111
"deformation the Combined method our with viewpoints. coefficient arbitrary and from texture regression, neural-network-based synthesizes","Combined with neural-network-based coefficient regression, our method synthesizes the texture and deformation from arbitrary viewpoints.",37112
processes identified The for specific and are inference implementation. reasoning future,The specific inference and reasoning processes are identified for future implementation.,37113
segmentation. deep algorithm for to semantic learning use a is One instance-aware possibility,One possibility is to use a deep learning algorithm for instance-aware semantic segmentation.,37114
a data. Such high achieve large methods amount accuracies annotated of but training require,Such methods achieve high accuracies but require a large amount of annotated training data.,37115
"only the We for that each is training class assume image, known. object","We assume that for each training image, only the object class is known.",37116
background. The system object the corresponding the automatically segments from,The system automatically segments the corresponding object from the background.,37117
datasets annotated help features. discriminative visual These richly network learn the,These richly annotated datasets help the network learn discriminative visual features.,37118
introduced to The ensures corresponding is utilization training strategy discriminator. of,The corresponding training strategy is introduced to ensures utilization of discriminator.,37119
visualization We conduct to experiment fine-grained motivation. Dogs on dataset also Stanford verify our,We also conduct visualization experiment on fine-grained Stanford Dogs dataset to verify our motivation.,37120
as as programming requires Today's in proficient individuals are software who languages possible. industry many,Today's software industry requires individuals who are proficient in as many programming languages as possible.,37121
"SQL process the However, be challenging. out turns to learning of","However, the process of learning SQL turns out to be challenging.",37122
being a descriptor only transformed MCC into is a texture originally minutia descriptor.,MCC originally being a minutia only descriptor is transformed into a texture descriptor.,37123
texture converted minutiae codes minutia codes (MTCC). are The to cylinder cylinder,The minutia cylinder codes are converted to minutiae texture cylinder codes (MTCC).,37124
"setting information. Specifically, facial conditional cycle-consistency the additional into adapt we by landmarks leveraging","Specifically, we adapt cycle-consistency into the conditional setting by leveraging additional facial landmarks information.",37125
images annotations resolution pathological cancer of and lack of areas. of high (gigapixel),high resolution (gigapixel) of pathological images and lack of annotations of cancer areas.,37126
mislabel patches remove set. proposed can The the RAL in the training,The proposed RAL can remove the mislabel patches in the training set.,37127
"also the of classification DenseNet images. is proposed (ADN), for pathological atrous","atrous DenseNet (ADN), is also proposed for the classification of pathological images.",37128
"evaluated ADN proposed and RAL on two datasets, been i.e. The have pathological","The proposed RAL and ADN have been evaluated on two pathological datasets, i.e.",37129
"the performance of The proposed + experimental framework, i.e. ADN demonstrate results excellent the RAL","The experimental results demonstrate the excellent performance of the proposed ADN + RAL framework, i.e.",37130
from a is reconstruction connectivity for detection Microscopy data. (EM) task Electron critical Synaptic neural,Synaptic connectivity detection is a critical task for neural reconstruction from Electron Microscopy (EM) data.,37131
"well-justified and clean, relies on is training The conceptually and prediction expressive resulting model procedures.","The resulting model is conceptually clean, expressive and relies on well-justified training and prediction procedures.",37132
to The relationship be explained. a existence such of need remote,The existence of such a remote relationship need to be explained.,37133
"practical overlooked point, This which obvious by literature, implications. has has been the","This point, which has been overlooked by the literature, has obvious practical implications.",37134
generate Video visual by frames dynamic future learning realistic patterns. prediction aims to,Video prediction aims to generate realistic future frames by learning dynamic visual patterns.,37135
propose this an Conditional to address Appearance-Motion challenge. We GAN,We propose an Appearance-Motion Conditional GAN to address this challenge.,37136
"dynamic In a adds new categories to dimension inventory practice, with challenge. another this fine-grained","In practice, a dynamic inventory with new fine-grained categories adds another dimension to this challenge.",37137
arbitrary suitable simulating boundaries. our This makes flows for effectively turbulent with geometrical method,This effectively makes our method suitable for simulating turbulent flows with arbitrary geometrical boundaries.,37138
Learning in a complex science. number of applications large systems data features the of dynamics,Learning the dynamics of complex systems features a large number of applications in data science.,37139
graphs with unveil provide patterns that user behavioral They also interpretable and changes. the,They also provide the user with interpretable graphs that unveil behavioral patterns and changes.,37140
to The individual accommodate changes edges. is notion at local breakpoint of proposed,The notion of local breakpoint is proposed to accommodate changes at individual edges.,37141
in tracking can time. scheme that real track novel rigid present We object objects a,We present a novel object tracking scheme that can track rigid objects in real time.,37142
uses subpixel-precise objects image approach The track with to accuracy. edges high,The approach uses subpixel-precise image edges to track objects with high accuracy.,37143
"always parameter However, available. small not models may be","However, small parameter models may not be always available.",37144
"by recent satellite the captured been a moving In years, satellite platform. have videos","In recent years, the satellite videos have been captured by a moving satellite platform.",37145
"noise In the detecting on particular, modeling. we a novel algorithm local offer based","In particular, we offer a novel detecting algorithm based on the local noise modeling.",37146
exponential noise an potential by distribution. We from the targets probability vehicle differentiate patterns,We differentiate the potential vehicle targets from noise patterns by an exponential probability distribution.,37147
"in proposed background detection Thus, foreground for been enormous decades. object methods recent have subtraction","Thus, enormous background subtraction methods for foreground object detection have been proposed in recent decades.",37148
combination solution other we is strategies. show And more than that our efficient,And we show that our solution is more efficient than other combination strategies.,37149
"During this polymerases, are ribosomes, proteome all the RNA time, doubled. and","During this time, ribosomes, RNA polymerases, and the proteome are all doubled.",37150
develops inferential factor This dimensions. theory an state-varying paper of for models large,This paper develops an inferential theory for state-varying factor models of large dimensions.,37151
"are process. recurrent models, functions some loadings Unlike state general factor constant of","Unlike constant factor models, loadings are general functions of some recurrent state process.",37152
methods Our analysis. estimator principal component nonparametric combines with,Our estimator combines nonparametric methods with principal component analysis.,37153
conditions. to system This significantly some changing overcome the suddenly helps,This significantly helps the system to overcome some suddenly changing conditions.,37154
scene proposed. spotting for trainable model An text is end-to-end network neural,An end-to-end trainable neural network model for scene text spotting is proposed.,37155
both drastically GPU for training and usage computation prediction. times a reduces of Optional the,Optional usage of a GPU drastically reduces the computation times for both training and prediction.,37156
dial-a-ride problem. paper multi-atomic annealing a static to called This solve (MATA) novel proposes algorithm,This paper proposes a novel algorithm called multi-atomic annealing (MATA) to solve static dial-a-ride problem.,37157
evolved in We theory. explore has the and involved are it how problems what,We explore how it has evolved and what problems are involved in the theory.,37158
explained terms of consciousness information. integrated IIT in can is be hypothesis that a,IIT is a hypothesis that consciousness can be explained in terms of integrated information.,37159
"In this urgently a is needed. context, survey IIT for critical","In this context, a critical survey for IIT is urgently needed.",37160
"this we IIT of and concepts fundamental To first related issues. introduce end,","To this end, we first introduce fundamental concepts of IIT and related issues.",37161
"transitions Thereafter, IIT major point related we issues. has intra-model through and out discuss been","Thereafter, we discuss major transitions IIT has been through and point out related intra-model issues.",37162
"in Finally, some issues the extra-model are in last section, IIT's presented. involved theoretical, principles","Finally, in the last section, some theoretical, extra-model issues involved in IIT's principles are presented.",37163
"crisp in such thereby often are Typically, crisp semantics answers. the of resulting queries","Typically, the semantics of such queries are often crisp thereby resulting in crisp answers.",37164
in address to have past strategies been Several proposed this.,Several strategies have been proposed in past to address this.,37165
taxonomy Entities generalized. have don't they end a being up and,Entities don't have a taxonomy and they end up being generalized.,37166
the numbers of estimating crowd Crowd is people task in images. counting,Crowd counting is the task of estimating people numbers in crowd images.,37167
hard. Density the very in is regression small area person general on,Density regression on the small person area is in general very hard.,37168
distinct a fusion streams deep TSLFN two network. The followed by includes,The TSLFN includes two distinct deep streams followed by a fusion network.,37169
datasets. on comprehensive benchmark conduct We experiments four,We conduct comprehensive experiments on four benchmark datasets.,37170
path feature is in moving important an design. architecture Human,Human moving path is an important feature in architecture design.,37171
"path, elements the architects studying (e.g. By to basic the know arrange where","By studying the path, architects know where to arrange the basic elements (e.g.",37172
"simulation. a system SimArch, path moving presents multi-agent for human This paper","This paper presents SimArch, a multi-agent system for human moving path simulation.",37173
using behavior Process. model built Decision It by a Markov a involves,It involves a behavior model built by using a Markov Decision Process.,37174
characteristics human It kinds different by transition also assigning different models of probabilities.,It also models different kinds of human characteristics by assigning different transition probabilities.,37175
agents. quickly algorithm sub-optimal path of plans A modified weighted the the search A*,A modified weighted A* search algorithm quickly plans the sub-optimal path of the agents.,37176
traditional rebinning to proposed method. is a compared method The,The proposed method is compared to a traditional rebinning method.,37177
directly works on clouds. algorithm point Our,Our algorithm works directly on point clouds.,37178
to regions input An cloud vector. encoded with point is feature missing a first,An input point cloud with missing regions is first encoded to a feature vector.,37179
and obtained analysed. plots can be different recurrence From variables RQA,From recurrence plots different RQA variables can be obtained and analysed.,37180
"such BER Viterbi-decoder, techniques and inter-leaving. Time are Different used improvement as, Frequency","Different BER improvement techniques are used such as, Viterbi-decoder, Time and Frequency inter-leaving.",37181
Effect also observed. modulation of different on BER techniques performance is,Effect of different modulation techniques on BER performance is also observed.,37182
natural metric a image an is for of important images. Learning patches analyzing tool,Learning a metric of natural image patches is an important tool for analyzing images.,37183
stage. of joint behavior present design the helps the study to predict in The configuration,The present study helps to predict the behavior of joint configuration in the design stage.,37184
interest attracted powerful of because captioning image in mechanisms performance. have its Attention considerable,Attention mechanisms have attracted considerable interest in image captioning because of its powerful performance.,37185
propose accessing We mathematical areas (LGAs). a model for among government chemotherapy local,We propose a mathematical model for accessing chemotherapy among local government areas (LGAs).,37186
the features noise the main relevant Removing without image goal. the of affecting is,Removing the noise without affecting relevant features of the image is the main goal.,37187
approaches become of solve this over to Sparse decomposition problem. dictionaries used most the redundant,Sparse decomposition over redundant dictionaries become of the most used approaches to solve this problem.,37188
"parts-based NMF naturally sparse, codes favor representations.","NMF codes naturally favor sparse, parts-based representations.",37189
"sparse redundant dictionary as representation, linear a combination represented a of atoms. In signals","In sparse representation, signals represented as a linear combination of a redundant dictionary atoms.",37190
The one. gives better from denoised proposed an N-NMF image reconstruction of a,The proposed N-NMF gives a better reconstruction of an image from denoised one.,37191
"matrix non-negative dictionary sparse pursuit, denoising, representation, matching Keywords: learning, factorization. Image","Keywords: Image denoising, sparse representation, dictionary learning, matching pursuit, non-negative matrix factorization.",37192
increasingly vision developed. networks convolutional robotic are systems neural Machine applications for being (CNNs) using,Machine vision systems using convolutional neural networks (CNNs) for robotic applications are increasingly being developed.,37193
"meaning not public affects economic just welfare, from drawn such The health. statements","The meaning drawn from such statements affects economic welfare, not just public health.",37194
"quantum-inspired the on the method shrinks proposed based function. complex threshold wavelet adaptive Then, coefficients","Then, the proposed method shrinks complex wavelet coefficients based on the quantum-inspired adaptive threshold function.",37195
characterization semantics closure. of develop the bi-preference for We defining a skeptical a semantics,We develop a bi-preference semantics semantics for defining a characterization of the skeptical closure.,37196
"this adversarial propose based problem, adaptation unsupervised To on domain alleviate networks. we framework an","To alleviate this problem, we propose an unsupervised domain adaptation framework based on adversarial networks.",37197
"predictions introduce discriminator ground from truth that masks. we Hence, a distinguishes segmentation","Hence, we introduce a discriminator that distinguishes segmentation predictions from ground truth masks.",37198
several a performing challenging performers activities. such with the this dataset In we introduce paper,In this paper we introduce such a challenging dataset with several performers performing the activities.,37199
a for human in novel We recognizing videos. method such present activities,We present a novel method for recognizing human activities in such videos.,37200
"maintains activity. information for the recognizing the proposed However, the system motion required","However, the proposed system maintains the required motion information for recognizing the activity.",37201
of plays classification in key nephropathy. diagnosis glomerulus a Pathological role the,Pathological glomerulus classification plays a key role in the diagnosis of nephropathy.,37202
alternative for provides paper glomerulus an for multi-stained This cues visual classification. solution integrating,This paper provides an alternative solution for integrating multi-stained visual cues for glomerulus classification.,37203
in We these joint a stages manner. optimize two,We optimize these two stages in a joint manner.,37204
undergo the are that Objects common real deformation in non-rigid world.,Objects that undergo non-rigid deformation are common in the real world.,37205
A and is example human faces. challenging typical the,A typical and challenging example is the human faces.,37206
and both classification are tasks. These registration useful for quantitative of evaluation,These are useful for quantitative evaluation of both registration and classification tasks.,37207
Rule Set for Multi-vAlue in-hospital model patient (MRS) predicting a We propose mortality.,We propose a Multi-vAlue Rule Set (MRS) model for in-hospital predicting patient mortality.,37208
of blood for management. is glucose diabetes essential Control,Control of blood glucose is essential for diabetes management.,37209
Performance convolutional of benchmarked is network recurrent neural against algorithms. the four,Performance of the recurrent convolutional neural network is benchmarked against four algorithms.,37210
"share pooling. We predictors max all between with scales box and replace across scales, convolution","We share box predictors across all scales, and replace convolution between scales with max pooling.",37211
model these SSD. changes hurt to empirically We show not that quality vanilla compared do,We empirically show that these changes do not hurt model quality compared to vanilla SSD.,37212
background or light was applied. artificial no,no artificial background or light was applied.,37213
"and with approach It high-throughput. novel non-invasive, inexpensive objective for is phenotyping a","It is a novel approach for non-invasive, inexpensive and objective phenotyping with high-throughput.",37214
"image First, inflorescences regions depicting and were identified localized.","First, image regions depicting inflorescences were identified and localized.",37215
"using ""inflorescence""-areas from the flowers were Hough Finally, Circular individual extracted Transform.","Finally, individual flowers were extracted from the ""inflorescence""-areas using Circular Hough Transform.",37216
is the reported Australian Age-progression in on data. cancer risk then,Age-progression in the cancer risk is then reported on Australian data.,37217
of tracking and estimation useful is Head pose medical applications. in variety,Head pose estimation and tracking is useful in variety of medical applications.,37218
"increased without to important the patient is head For track markers. comfort, it","For increased patient comfort, it is important to track the head without markers.",37219
"pose using different head this regard, In the estimation address problem approaches. we two","In this regard, we address the head pose estimation problem using two different approaches.",37220
are obtained analog (ADC). to by samples a The digital (SR) modulo self-reset converter,The modulo samples are obtained by a self-reset (SR) analog to digital converter (ADC).,37221
"a the propose also we by functions, second-order-statistics objective FT-PCA. Motivated approach, geometrical","Motivated by the geometrical objective functions, we also propose a second-order-statistics approach, FT-PCA.",37222
"of upon supervised Under category built the detectors are usually weakly supervision classifiers. labels,","Under the supervision of category labels, weakly supervised detectors are usually built upon classifiers.",37223
"been few objects environments. in uncontrolled has However, research there generating real-world","However, there has been few research generating objects in uncontrolled real-world environments.",37224
"this vehicle approach in a scenes. we paper, image for In propose novel generation real-world","In this paper, we propose a novel approach for vehicle image generation in real-world scenes.",37225
"object networks neural largely detection. Region-based (R-CNN)~\cite{fast_rcnn,faster_rcnn,mask_rcnn} have dominated convolutional","Region-based convolutional neural networks (R-CNN)~\cite{fast_rcnn,faster_rcnn,mask_rcnn} have largely dominated object detection.",37226
"propose this network paper, WSOD. novel deep In we for a","In this paper, we propose a novel deep network for WSOD.",37227
The adjacent spatially the proposals cluster associated are same and the with in object. same,The proposals in the same cluster are spatially adjacent and associated with the same object.,37228
"ImageNet are for MS-COCO Experiments the benchmarks conducted on VOC, WSOD. and detection, PASCAL","Experiments are conducted on the PASCAL VOC, ImageNet detection, and MS-COCO benchmarks for WSOD.",37229
of method outperforms our previous Results art the that state significantly. the show,Results show that our method outperforms the previous state of the art significantly.,37230
outliers module false to reduce used filter alarms. post-processing A and is out,A post-processing module is used to filter out outliers and reduce false alarms.,37231
"Typical guidance or texture local includes, coordinates field. a e.g., values, normal displacement","Typical local guidance includes, e.g., normal values, texture coordinates or a displacement field.",37232
"our transfer style techniques, approach expensive Contrary not does previous involve computationally any to optimization.","Contrary to previous style transfer techniques, our approach does not involve any computationally expensive optimization.",37233
maintain of matrix scattering mode also can information encoding completely. This polarimetric,This encoding mode can also maintain polarimetric information of scattering matrix completely.,37234
"observers. by OCT intra-coronary performed manually using characterization is images expert plaque Commonly,","Commonly, plaque characterization using intra-coronary OCT images is performed manually by expert observers.",37235
"in result Creative journeys follow memes, or unique works, form. whether their that paintings final","Creative works, whether paintings or memes, follow unique journeys that result in their final form.",37236
irregular for inpainting approach This a holes. to semi-parametric image introduces paper,This paper introduces a semi-parametric approach to image inpainting for irregular holes.,37237
nonparametric of database. part The an external consists image,The nonparametric part consists of an external image database.,37238
Markov on transition a based first of We multi-view probability chain. matrices construct the tensor,We first construct a tensor based on multi-view transition probability matrices of the Markov chain.,37239
method be direction efficiently of alternating can optimized by method multipliers~(ADMM). The the proposed,The proposed method can be efficiently optimized by the alternating direction method of multipliers~(ADMM).,37240
or supervised either unsupervised video Most based are learning. on summarisation existing methods,Most existing video summarisation methods are based on either supervised or unsupervised learning.,37241
to rewards training for A DQSN. trained provide network is also the companion classification,A companion classification network is also trained to provide rewards for training the DQSN.,37242
"classification network, a result. reward on the based global the we recognisability With classification develop","With the classification network, we develop a global recognisability reward based on the classification result.",37243
on proposed achieves show state-of-the-art the approach performance. two that experiments datasets benchmark Extensive,Extensive experiments on two benchmark datasets show that the proposed approach achieves state-of-the-art performance.,37244
powerful proposed code-rates with and varying The constructs block-lengths. scheme LDPC codes,The proposed scheme constructs powerful LDPC codes with varying code-rates and block-lengths.,37245
systems. the gains obtained over compared show The large results,The results obtained show large gains over the compared systems.,37246
for (EL) the photovoltaic of useful inspection a modules. Electroluminescence is imaging modality (PV),Electroluminescence (EL) imaging is a useful modality for the inspection of photovoltaic (PV) modules.,37247
"which application are differ their dictated by their hardware in requirements, approaches respective scenarios. The","The approaches differ in their hardware requirements, which are dictated by their respective application scenarios.",37248
"compare processing variants. To and a performance, various strong investigate obtain we","To obtain a strong performance, we investigate and compare various processing variants.",37249
"make PV highly approaches cells of Both automated continuous, monitoring accurate feasible.","Both automated approaches make continuous, highly accurate monitoring of PV cells feasible.",37250
bilateral major with significant have alterations white-matter fascicles. Patients IGE in,Patients with IGE have significant bilateral alterations in major white-matter fascicles.,37251
"scale-invariant, orientation- affine and from An (ACs) for approach is proposed correspondences recovering e.g.","An approach is proposed for recovering affine correspondences (ACs) from orientation- and scale-invariant, e.g.",37252
"locally RANSAC, magnitude order optimized is faster. of an by","by locally optimized RANSAC, is an order of magnitude faster.",37253
"particular factors depend?"" the of heuristics "", the performance does (relative) which ""On",""", ""On which factors does the (relative) performance of the particular heuristics depend?""",37254
"circumstances use or heuristic?"" ""Under I which should which","or ""Under which circumstances should I use which heuristic?""",37255
"UWCNN train underwater for models each image we multiple separately type. formation Then,","Then, we separately train multiple UWCNN models for each underwater image formation type.",37256
the cross-validation and V-Net of testing. outperformed used both implementation Results: the U-Net The in,Results: The U-Net outperformed the used implementation of the V-Net in both cross-validation and testing.,37257
task We and novel subject prediction: on Variability also identity perspective BOLD (BV). introduce a,We also introduce a novel perspective on task and subject identity prediction: BOLD Variability (BV).,37258
"each based variance Conceptually, a region-specific brain BV is within region. the on measure","Conceptually, BV is a region-specific measure based on the variance within each brain region.",37259
"to compute, simple interpret, is and BV visualize.","BV is simple to compute, interpret, and visualize.",37260
extremely unfavourable Underwater from suffer conditions. images,Underwater images suffer from extremely unfavourable conditions.,37261
"veiling scattering causes called change creates so hue, in Attenuation light.","Attenuation creates change in hue, scattering causes so called veiling light.",37262
its understanding effects. lending provides a approach deeper the and of This bank channel,This approach provides a deeper understanding of the bank lending channel and its effects.,37263
A disparity between can the observed. be countries small north-south,A small north-south disparity between the countries can be observed.,37264
create to are and architectures a vision necessary AGI? What system for frameworks,What frameworks and architectures are necessary to create a vision system for AGI?,37265
existing and our demonstrate some discriminative their We discuss for generative purposes. insufficiency models and,We discuss some existing generative and discriminative models and demonstrate their insufficiency for our purposes.,37266
"some dilemmas open and Finally, we discuss architectural questions.","Finally, we discuss some architectural dilemmas and open questions.",37267
at the the blockage. likelihood BS Placing height reduces greater a of,Placing the BS at a greater height reduces the likelihood of blockage.,37268
Kingdomino. surprisingly is shows MCE for than game like The MCTS that a evaluation stronger,The evaluation shows that surprisingly MCE is stronger than MCTS for a game like Kingdomino.,37269
"novel decoupled we paper, propose In a filtering this algorithm. Kalman","In this paper, we propose a novel decoupled Kalman filtering algorithm.",37270
equivalence also The is the example. confirmed numerical via,The equivalence is also confirmed via the numerical example.,37271
fully tutorials. automatic method video This game for paper generating introduces a,This paper introduces a fully automatic method for generating video game tutorials.,37272
"of wound doctors, significant Postoperative patients. for cause a expense are hospitals, complications and","Postoperative wound complications are a significant cause of expense for hospitals, doctors, and patients.",37273
"the method an wound desired. of onset to diagnose is Hence, effective strongly complications","Hence, an effective method to diagnose the onset of wound complications is strongly desired.",37274
accessible due care to deliver means to ubiquity. wound provide a their Smartphones increasing,Smartphones provide a means to deliver accessible wound care due to their increasing ubiquity.,37275
extremely of Networks Modern tasks. Neural powerful Convolutional computer on (CNN) vision range are a,Modern Convolutional Neural Networks (CNN) are extremely powerful on a range of computer vision tasks.,37276
algorithms by proposed measurement numerical and data. by of the evaluated simulations performance is The,The performance of the proposed algorithms is evaluated by numerical simulations and by measurement data.,37277
We main result two this develop in steps.,We develop this result in two main steps.,37278
associated The that uncovers first the be fact classes to specific image-space with directions. tend,The first uncovers the fact that classes tend to be associated with specific image-space directions.,37279
This perspective on provides of existence the universal a adversarial perturbations. novel,This provides a novel perspective on the existence of universal adversarial perturbations.,37280
"contradiction apparent and our between resolves analysis accuracy vulnerability. Thus, the","Thus, our analysis resolves the apparent contradiction between accuracy and vulnerability.",37281
an is captions captioning. in for emerging image topic image an stylized Generating,Generating stylized captions for an image is an emerging topic in image captioning.,37282
topic. image weak a medical unexplored remains with relatively network annotations a segmentation Training,Training a medical image segmentation network with weak annotations remains a relatively unexplored topic.,37283
We the techniques public segmentation evaluate (NCI-ISBI) and (ACDC) prostate datasets. cardiac on,We evaluate the techniques on public cardiac (ACDC) and prostate (NCI-ISBI) segmentation datasets.,37284
and tested publicly is proposed The datasets. on two used available method widely,The proposed method is tested on two publicly available and widely used datasets.,37285
"jointly and pose MultiPoseNet problems. segmentation can person detection, keypoint handle estimation person detection,","MultiPoseNet can jointly handle person detection, keypoint detection, person segmentation and pose estimation problems.",37286
factors that highly behavior. impacted are One interference important the of is,One of the important factors that are highly impacted is interference behavior.,37287
simulations. Monte-Carlo Analytical results validated are via,Analytical results are validated via Monte-Carlo simulations.,37288
"in the of the we setting study problem Calculus. the Here, relational Situation","Here, we study the problem in the relational setting of the Situation Calculus.",37289
techniques We the actually provide synthesizing also controller. for,We also provide techniques for actually synthesizing the controller.,37290
"version, capsule to generative This variational paper (VCs). a extends the named capsules","This paper extends the capsule to a generative version, named variational capsules (VCs).",37291
model as a a probabilistic Variational model. in entities composition of capsules an image,Variational capsules model an image as a composition of entities in a probabilistic model.,37292
"multiple a encode in variational addition, capsules way. semantically-disentangling In entities","In addition, variational capsules encode multiple entities in a semantically-disentangling way.",37293
"semantic not enforce This is because, invertibility correctness. does necessarily","This is because, invertibility does not necessarily enforce semantic correctness.",37294
(RNNs) neural networks of are of sequential capable temporal information. dynamics complex modeling the Recurrent,Recurrent neural networks (RNNs) are capable of modeling the temporal dynamics of complex sequential information.,37295
block an to with We EleAttG an RNN block. an equipped EleAtt-RNN as refer,We refer to an RNN block equipped with an EleAttG as an EleAtt-RNN block.,37296
The uniqueness algorithm the existence and proposed and are of the solution convergence analysed. of,The existence and uniqueness of the solution and convergence of the proposed algorithm are analysed.,37297
work includes main The following aspects. contribution of this three the,The main contribution of this work includes the following three aspects.,37298
"the DSurVD. pedestrian by evaluated detection seven robustness the is built of algorithms existing Third,","Third, the robustness of seven existing pedestrian detection algorithms is evaluated by the built DSurVD.",37299
Experimental pedestrian detection show for improved robustness algorithms. the be results further existing can that,Experimental results show that the robustness can be further improved for existing pedestrian detection algorithms.,37300
problem applications. Document (DIQA) assessment image and an is challenging real in quality important,Document image quality assessment (DIQA) is an important and challenging problem in real applications.,37301
"dynamic, outcomes. its potential good I derive and a contribution voluntary present mechanism, public game","I present a dynamic, voluntary contribution mechanism, public good game and derive its potential outcomes.",37302
"by engaging endogenously period, players In contribution each productivity in determine investment. costly","In each period, players endogenously determine contribution productivity by engaging in costly investment.",37303
"the of payoffs all payoff lowest players receive outcome, In zero.","In the lowest payoff outcome, all players receive payoffs of zero.",37304
"there Therefore, Nash strategies. infinitely many Equilibria are","Therefore, there are infinitely many Nash Equilibria strategies.",37305
explain goal and My this discover is point. to,My goal is to discover and explain this point.,37306
I derive computer and simulation results. the mathematical analysis use to,I use mathematical analysis and computer simulation to derive the results.,37307
is wireless in networking. of use PLE estimation assist thus to great,PLE estimation is thus of great use to assist in wireless networking.,37308
"information difficult sometimes Moreover, be unreliable, spoofed, or obtain. to might this external","Moreover, this external information might sometimes be unreliable, spoofed, or difficult to obtain.",37309
"for presented. linear start, a PLE new the model To is regression","To start, a new linear regression model for the PLE is presented.",37310
"based medical image in focus this learning paper, on deep problems we three segmentation. In","In this paper, we focus on three problems in deep learning based medical image segmentation.",37311
superior results that our state-of-the-art Our network to experimental comparable or methods. suggest is,Our experimental results suggest that our network is comparable or superior to state-of-the-art methods.,37312
linear risk re-examines pricing models. premia of paper This in factor estimating the problem,This paper re-examines the problem of estimating risk premia in linear factor pricing models.,37313
based estimation new a sample-splitting variables regression. instrumental procedure propose on We,We propose a new estimation procedure based on sample-splitting instrumental variables regression.,37314
an identification. Post-mortem additional offer method can forensic recognition personal iris of,Post-mortem iris recognition can offer an additional forensic method of personal identification.,37315
"and of points from neuroscience-cognitive, mathematical view. conjointly biological, physical","conjointly from biological, neuroscience-cognitive, physical and mathematical points of view.",37316
by unify proposes It results of theories to various and means information topology.,It proposes to unify various results and theories by means of information topology.,37317
"Notably, main allow methods of of conciliation the some those theories of consciousness. topological","Notably, those topological methods allow conciliation of some of the main theories of consciousness.",37318
handle of ranges to large and differences between How datasets? b) outdoor depth the indoor,b) How to handle the large differences of depth ranges between indoor and outdoor datasets?,37319
our method. indoor on and results demonstrate both of datasets effectiveness outdoor Experimental the,Experimental results on both indoor and outdoor datasets demonstrate the effectiveness of our method.,37320
"image fast. generation on In research has developing recent methods years, been","In recent years, research on image generation methods has been developing fast.",37321
"blurry. However, generated by generally pictures are the GANs","However, the generated pictures by GANs are generally blurry.",37322
This inherent training the to the structure reflect makes of data. DCGANs unable,This makes DCGANs unable to reflect the inherent structure of the training data.,37323
"DE-GANs is and the from loss different DCGANs. Moreover, function of GANs","Moreover, the loss function of DE-GANs is different from GANs and DCGANs.",37324
of is solution to present problem. purpose this The to this a paper,The purpose of this paper is to present a solution to this problem.,37325
agreement excellent with data. prediction model The shows published an,The model prediction shows an excellent agreement with published data.,37326
We OLGA guide that suggest useful design. to be a tool may vaccine,We suggest that OLGA may be a useful tool to guide vaccine design.,37327
simode: Integral Ordinary In paper this we Equations. for Separable Matching Differential describe,In this paper we describe simode: Separable Integral Matching for Ordinary Differential Equations.,37328
were separability such Linear ease as shown features to and optimization inference.,Linear features such as separability were shown to ease optimization and inference.,37329
using of the demonstrate the ordinary differential various package We of systems functionalities equations.,We demonstrate the functionalities of the package using various systems of ordinary differential equations.,37330
"these serious indicate of benign, they can are sometimes a many health Although risk.","Although many of these are benign, they can sometimes indicate a serious health risk.",37331
small valid sufficiently our All results of are control. values the for,All our results are valid for sufficiently small values of the control.,37332
time fields. data established clustering series technique is well mining The many of in,The data mining technique of time series clustering is well established in many fields.,37333
at a always The constant rate. micro-dynamics run,The micro-dynamics always run at a constant rate.,37334
are time by dramatic statistics. described log-Poisson These transitions a,These dramatic transitions are described by a log-Poisson time statistics.,37335
long effect time The a network. ecological is the adapted collectively of,The long time effect is a collectively adapted of the ecological network.,37336
structured classification proposed  is  for task. discriminative analysis dictionary the A,  A discriminative structured analysis dictionary is proposed for the classification task.,37337
"dictionary large address Moreover, distributed structured also analysis learning datasets. presented a scale is to","Moreover, a distributed structured analysis dictionary learning is also presented to address large scale datasets.",37338
"complexity the greatly In addition, training and reduced. also are testing computational","In addition, the training and testing computational complexity are also greatly reduced.",37339
"to ""specular extends translation network to multi-view Our diffuse"" image-to-image unsupervised translation.","Our network extends unsupervised image-to-image translation to multi-view ""specular to diffuse"" translation.",37340
"several baseline a better network As results significantly our result, than yields single-view techniques.","As a result, our network yields significantly better results than several single-view baseline techniques.",37341
real data. on performance We and world also training show testing promising,We also show promising performance on real world training and testing data.,37342
of algorithm's theoretical provide its a and numerical basis. foundation review the We mathematical verification,We review the algorithm's mathematical foundation and provide a numerical verification of its theoretical basis.,37343
the complexity. very computational had equivalent problem high Solving,Solving the equivalent problem had very high computational complexity.,37344
"the high-dimensions. in of uniqueness the Also, solution problem conditions considered the for are","Also, the uniqueness conditions for solution of the problem are considered in the high-dimensions.",37345
ones. samples by existing the principle create recombining The is method new of to,The principle of the method is to create new samples by recombining existing ones.,37346
proposed show performance over augmentation. data We basic more the improves method the that,We show that the proposed method improves the performance over more basic data augmentation.,37347
azimuth. the only of direction The runs simulation on,The simulation runs only on the direction of azimuth.,37348
"in multimedia. robotics, as This forensics, various applications surveillance, has such task fields,","This task has applications in various fields, such as surveillance, forensics, robotics, multimedia.",37349
evaluation our benchmarks. approach of extensive present five through backbones three We and,We present extensive evaluation of our approach through five backbones and three benchmarks.,37350
results person the our re-identification of effectiveness demonstrate framework. Experimental,Experimental results demonstrate the effectiveness of our person re-identification framework.,37351
are built Traditional on and architectures. trainable methods detection object handcrafted shallow features,Traditional object detection methods are built on handcrafted features and shallow trainable architectures.,37352
"These etc. function, optimization behave architecture, training network in strategy and models differently","These models behave differently in network architecture, training strategy and optimization function, etc.",37353
"a object learning detection we provide paper, based review In deep frameworks. this on","In this paper, we provide a review on deep learning based object detection frameworks.",37354
methods Experimental are and provided conclusions. draw meaningful to some also compare various analyses,Experimental analyses are also provided to compare various methods and draw some meaningful conclusions.,37355
"rapidly. of digital the various increasing use Recently, is images in fields","Recently, the use of digital images in various fields is increasing rapidly.",37356
"method poor quality. visual in However, W-plane has also the performance","However, the W-plane method also has poor performance in visual quality.",37357
in is tasks. a vision step computer Geometric key matching,Geometric matching is a key step in computer vision tasks.,37358
"for two matching. different choices provide geometric in we Finally, purposes","Finally, we provide two choices for different purposes in geometric matching.",37359
"non-rigid homography When compositing the improves. further alignment transformation, thin-plate-spline with accuracy","When compositing homography with non-rigid thin-plate-spline transformation, the alignment accuracy further improves.",37360
person attention in Video attracts re-identification recent years. much,Video person re-identification attracts much attention in recent years.,37361
from image of pedestrians different to match aims sequences camera It views.,It aims to match image sequences of pedestrians from different camera views.,37362
they understood and paper that argues can as as correction be improvement. This,This paper argues that they can be understood as correction and as improvement.,37363
on channel-based approach This based a is Bayesian to probability. account novel,This account is based on a novel channel-based approach to Bayesian probability.,37364
vectors. The determines transform collection corresponding the of sensing,The corresponding transform determines the collection of sensing vectors.,37365
"not the vectors satisfy sensing K-AHS any incoherence or isometry Therefore, restricted property. must","Therefore, the K-AHS sensing vectors must not satisfy any incoherence or restricted isometry property.",37366
real by with images. are signals and simulations synthetic supported findings world The analytical,The analytical findings are supported by simulations with synthetic signals and real world images.,37367
"achieves K-AHS benchmark errors CS. than On standard images, lower reconstruction","On standard benchmark images, K-AHS achieves lower reconstruction errors than CS.",37368
Bayesian better understanding is hydrological of for important a modeling inverse processes.,Bayesian inverse modeling is important for a better understanding of hydrological processes.,37369
"surrogate of address techniques. take can modeling issue, advantage one this To","To address this issue, one can take advantage of surrogate modeling techniques.",37370
"the quantify are surrogate strategies error. to proposed Specifically, two","Specifically, two strategies are proposed to quantify the surrogate error.",37371
"In algorithm image and an pepper is paper, for salt noise. proposed denoising this","In this paper, an image denoising algorithm is proposed for salt and pepper noise.",37372
"remove noise. switching the filter salt and pepper to Finally, the a algorithm builds non-local","Finally, the algorithm builds a non-local switching filter to remove the salt and pepper noise.",37373
classification result initial Subjects a to allowed process. of change the annotation were as their,Subjects were allowed to change their initial classification as a result of the annotation process.,37374
"our that hybrid capture Finally, temporal show indeed theories automata. basic formally action we","Finally, we formally show that our temporal basic action theories indeed capture hybrid automata.",37375
measurements to vehicles accomplish assignments. suitable eNodeBs them from utilize subchannel and such request,eNodeBs request such measurements from vehicles and utilize them to accomplish suitable subchannel assignments.,37376
"proposed algorithms. greedy SPS Furthermore, the compared pseudo-random algorithm is against and","Furthermore, the proposed algorithm is compared against pseudo-random and greedy SPS algorithms.",37377
applications. health monitoring can be personal for The proposed useful method,The proposed method can be useful for personal health monitoring applications.,37378
"useful and detection correlation objects provide the between for information However, may segmentation.","However, the correlation between objects may provide useful information for detection and segmentation.",37379
the publicly is open The analysis code source used available. and for,The source code used for the analysis is open and publicly available.,37380
and learning machine in computer Intelligent research active techniques. an field currently video-surveillance is vision,Intelligent video-surveillance is currently an active research field in computer vision and machine learning techniques.,37381
provides investigators. tools surveillance for It video and forensic useful operators,It provides useful tools for surveillance operators and forensic video investigators.,37382
one these is among (PReID) Person re-identification tools.,Person re-identification (PReID) is one among these tools.,37383
"many proposed To of have the PReID. aim, increase to been this techniques performance","To this aim, many techniques have been proposed to increase the performance of PReID.",37384
"models being we for Therefore, DNN this summarized task. used state-of-the-art","Therefore, we summarized state-of-the-art DNN models being used for this task.",37385
to ranking groups: Previous methods sliding two grouping. divided actionness score and window can be,Previous methods can be divided to two groups: sliding window ranking and actionness score grouping.,37386
"learning. accurately semantic landmarks using detect SegNet, segmentation on facial We which based deep performs","We detect facial landmarks accurately using SegNet, which performs semantic segmentation based on deep learning.",37387
are point tracker. of the points KLT landmarks detected Edge using tracked,Edge points of detected landmarks are tracked using the KLT point tracker.,37388
"of its problem. In tracker spite point loss suffers the from KLT popularity, the point","In spite of its popularity, the KLT point tracker suffers from the point loss problem.",37389
calculate solve executing of periodically problem We this by facial SegNet shape to landmarks. the,We solve this problem by executing SegNet periodically to calculate the shape of facial landmarks.,37390
any pair <subject-predicate-object> of visual localized Recognizing objects relationships understanding. is pivotal among image for,Recognizing visual relationships <subject-predicate-object> among any pair of localized objects is pivotal for image understanding.,37391
"an In perspective investigate work, orthogonal feature this we interactions. on based","In this work, we investigate an orthogonal perspective based on feature interactions.",37392
"measurement results Currently, in physical record their patients data of a and notebook.","Currently, patients record their results and data of measurement in a physical notebook.",37393
"be could that those all applications, it In likely functionals misspecified. is","In applications, it is likely that those functionals could all be misspecified.",37394
estimator not but efficient. consistent Consequently or inconsistent could be their,Consequently their estimator could be inconsistent or consistent but not efficient.,37395
problems and uncertainty methods on inference (UQ). efficient for quantification effective highly rely inverse Bayesian,Bayesian inverse problems highly rely on efficient and effective inference methods for uncertainty quantification (UQ).,37396
"algorithms, are models. MCMC under on spaces, physical robust directly Infinite-dimensional function defined of refinement","Infinite-dimensional MCMC algorithms, directly defined on function spaces, are robust under refinement of physical models.",37397
"required However, are in geometric high the obtain expensive usually quantities dimensions. to","However, the required geometric quantities are usually expensive to obtain in high dimensions.",37398
"measurement of response. Specifically, is evaluate filter the designed quality an to effective function","Specifically, an effective measurement function is designed to evaluate the quality of filter response.",37399
"Therefore, been are data automatically annotations popularized has as synthetic available.","Therefore, synthetic data has been popularized as annotations are automatically available.",37400
necessary step is samples. images in these analysing across A registration,A necessary step in analysing these images is registration across samples.,37401
for developed modalities registration were lower image of resolution (e.g. Existing methods,Existing methods of registration were developed for lower resolution image modalities (e.g.,37402
"gene analyses products, However, proteins. ignore GE between i.e., interactions physical","However, GE analyses ignore physical interactions between gene products, i.e., proteins.",37403
"condition-specific to information. data BN current However, capture fail","However, current BN data fail to capture condition-specific information.",37404
is aging. human process A dynamic interest of our,A dynamic process of our interest is human aging.,37405
"e.g., validate via search. and enrichment functional literature predictions We the analyses","We validate the predictions via e.g., functional enrichment analyses and literature search.",37406
public and evaluated framework We our three in retrieval datasets. medical classification image on,We evaluated our framework in medical image retrieval and classification on three public datasets.,37407
the remove to rain it necessary images. from is So,So it is necessary to remove the rain from images.,37408
"modify task, we rain better removal also network. the To fit the","To better fit the rain removal task, we also modify the network.",37409
rain So decompose into further stages. removal we the multiple,So we further decompose the rain removal into multiple stages.,37410
and We datasets. extensive conduct synthetic real-world experiments on both,We conduct extensive experiments on both synthetic and real-world datasets.,37411
all evaluation state-of-the-art the method Our under metrics. proposed approaches outperforms,Our proposed method outperforms the state-of-the-art approaches under all evaluation metrics.,37412
use is area that from Optimization. Combinatorial One of the benefit significantly can ML,One area that can significantly benefit from the use of ML is Combinatorial Optimization.,37413
and multiple of approaches to with draw common We highlight research. related fields themes connections,We highlight common themes to multiple approaches and draw connections with related fields of research.,37414
"multiple low-resolution variable ptychographic used microscopy, Fourier patterns collect illumination to images. In are","In Fourier ptychographic microscopy, variable illumination patterns are used to collect multiple low-resolution images.",37415
volume-varied a lesions. deal We with model aggregation propose framework multi-scale to,We propose a multi-scale aggregation model framework to deal with volume-varied lesions.,37416
that show aggregation the Results learning is information. model spatial in multi-scale effective,Results show that the aggregation model is effective in learning multi-scale spatial information.,37417
traffic role (HD) High Definition an in modern play scenes. maps important,High Definition (HD) maps play an important role in modern traffic scenes.,37418
"of the cost because grows However, of the HD maps development coverage slowly limitation.","However, the development of HD maps coverage grows slowly because of the cost limitation.",37419
in is state-of-the-art crowdsourced image unordered detection an for designed It dataset. lane,It is designed for state-of-the-art lane detection in an unordered crowdsourced image dataset.,37420
"road dataset applications. for modeling a detection introduced urban lane efficient we TTLane, in And","And we introduced TTLane, a dataset for efficient lane detection in urban road modeling applications.",37421
constructed precisely data. And even be crowdsourced maps inaccurate can with the,And the maps can be constructed precisely even with inaccurate crowdsourced data.,37422
are provided Monte Carlo analysis. accuracy to our throughout of the validate simulations,Monte Carlo simulations are provided throughout to validate the accuracy of our analysis.,37423
the then correspondences find shared the modalities. These be can used representations to between,These shared representations can then be used to find the correspondences between the modalities.,37424
"as wide Given we modems bandwidth network considered, sensors. on rely power the line","Given the wide bandwidth considered, we rely on power line modems as network sensors.",37425
"Despite behind the challenges these engineering several neuroscience basic advances, important remains poorly explored.","Despite several important engineering advances, the basic neuroscience behind these challenges remains poorly explored.",37426
recognition Spectral traction for gained in biometric recently face imaging has systems.,Spectral imaging has recently gained traction for face recognition in biometric systems.,37427
based In covariates. is on merely to it possible retrieval principle these perform the,In principle it is possible to perform the retrieval based merely on these covariates.,37428
settings. their in analytical formulae our verify correctness experiments through practical to verify We,We verify our analytical formulae through experiments to verify their correctness in practical settings.,37429
boundaries method segment We accurately used to dermoscopic lesion fully images. automated from a,We used a fully automated method to accurately segment lesion boundaries from dermoscopic images.,37430
is deep trained data on publicly U-net network learning ISIC. available A from,A U-net deep learning network is trained on publicly available data from ISIC.,37431
speed proposed determine road We in and estimator scenarios. the of straight the curved accuracy,We determine the accuracy of the proposed speed estimator in straight and curved road scenarios.,37432
"single-input, components buffers first-in, The that first-out a in manage (FIFO) fashion. are single-output data","The buffers are single-input, single-output components that manage data in a first-in, first-out (FIFO) fashion.",37433
"which the bone major synthetic have weaknesses graft substitutes. these methods of However, encourage development","However, these methods have major weaknesses which encourage the development of synthetic bone graft substitutes.",37434
"methods benchmark Besides, an the accuracy methods. have to with the proposed respect improved shown","Besides, the proposed methods have shown an improved accuracy with respect to the benchmark methods.",37435
"study estimation known Applying non-linear we merits techniques, of new the this model.","Applying known non-linear estimation techniques, we study the merits of this new model.",37436
"different power with we by In generator controls. particular, mechanical robustness test a considering","In particular, we test robustness by considering a generator with different mechanical power controls.",37437
is track video step key ReID. aggregation for video-based person Feature from a a,Feature aggregation from a video track is a key step for video-based person ReID.,37438
this RNNs pooling Many tackle methods by with or existing temporal attention. average/maximum problem,Many existing methods tackle this problem by average/maximum temporal pooling or RNNs with attention.,37439
"try solve problems, jointly these to approaches i.e. Modern","Modern approaches try to jointly solve these problems, i.e.",37440
a game is of automatic problem. AI generation challenging tutorials The,The automatic generation of game tutorials is a challenging AI problem.,37441
data which different represent by source. a is by information aspects conveyed modalities Multiple,Multiple modalities represent different aspects by which information is conveyed by a data source.,37442
embedded popularity. could information such in in useful predicting Multimodal be posts their,Multimodal information embedded in such posts could be useful in predicting their popularity.,37443
superior to demonstrate compared to solvers noise The state-of-the-art. proposed robustness the,The proposed solvers demonstrate superior robustness to noise compared to the state-of-the-art.,37444
This of to dataset. collection the road-scene a caters well believable,This caters well to the collection of a believable road-scene dataset.,37445
while are competition. to promoting cooperation Standards needed enable,Standards are needed to enable cooperation while promoting competition.,37446
"specifications are minimalistic. Such extensive, and complex","Such specifications are extensive, complex and minimalistic.",37447
communications methodology wireless paper standards. for teaching describes a This,This paper describes a methodology for teaching wireless communications standards.,37448
"as be resource/subchannel weighted problems represented graphs. In bipartite allocation can general,","In general, resource/subchannel allocation problems can be represented as weighted bipartite graphs.",37449
"into in account. Thus, mentioned approach been taken constraint our has the","Thus, in our approach the mentioned constraint has been taken into account.",37450
"Fig. To problem, gist the illustrate of the in","To illustrate the gist of the problem, in Fig.",37451
increasingly tasks. a processing important Deep solve learning of techniques image number to are becoming,Deep learning techniques are becoming increasingly important to solve a number of image processing tasks.,37452
good memory It balance a maintains performance. between prediction cost and,It maintains a good balance between memory cost and prediction performance.,37453
applications. driving for compatible quite autonomous it which makes,which makes it quite compatible for autonomous driving applications.,37454
"range, the dynamic the of near-distance region. in especially, deterioration It the causes","It causes the deterioration of the dynamic range, especially, in the near-distance region.",37455
Another internal is the of in FMCW delay problem system. radar radar,Another problem of FMCW radar is internal delay in the radar system.,37456
range. It maximum the to decrease leads the of detectable,It leads to the decrease of the maximum detectable range.,37457
"down-conversion is resolves this novel these a problems paper, which technique In proposed.","In this paper, a novel down-conversion technique which resolves these problems is proposed.",37458
implement procedures proposed technique and to theory explained. are Detailed the,Detailed theory and procedures to implement the proposed technique are explained.,37459
"are Then, results. experiment with performances of it the verified","Then, performances of it are verified with the experiment results.",37460
"type expected This are utility proposing we not a hypothesis. of new way,","This way, we are not proposing a new type of expected utility hypothesis.",37461
"are its definition. we keeping classical it contrary, the On under","On the contrary, we are keeping it under its classical definition.",37462
"the forager it consecutively empty sites, lands starve. $s$ will on then If","If the forager lands consecutively on $s$ empty sites, then it will starve.",37463
"The jumping. non-monotonically of probability according the to lifetime $p_j$, varies of forager the","The lifetime of the forager varies non-monotonically according to $p_j$, the probability of jumping.",37464
RBC much and long power distance transfer? reach can How how can RBC,How long distance can RBC reach and how much power can RBC transfer?,37465
present for (STSRN) Network Synergic Residual this problem. Spatial-Temporal We novel - method a,We present a novel method - Spatial-Temporal Synergic Residual Network (STSRN) for this problem.,37466
"smooth a a residual contains temporal a residual and spatial STSRN spatial-temporal module. extractor, processor","STSRN contains a spatial residual extractor, a temporal residual processor and a spatial-temporal smooth module.",37467
"disasters impacting structural thus Natural health affect buildings, of directly public safety.","Natural disasters affect structural health of buildings, thus directly impacting public safety.",37468
experimental and are validation Theoretical derivation presented.,Theoretical derivation and experimental validation are presented.,37469
"investigate distribution. we sensor In error modeling ratio drift interstory and (IDR) addition,","In addition, we investigate modeling sensor error and interstory drift ratio (IDR) distribution.",37470
"achieved sensor accuracy. classification of discuss Moreover, the on the impact error we building","Moreover, we discuss the impact of sensor error on the achieved building classification accuracy.",37471
"than is ternary Mathematically, binary coding efficient coding. more","Mathematically, ternary coding is more efficient than binary coding.",37472
use representation implications inability optimal mathematics are The to of or logic our examined. through,The implications of our inability to use optimal representation through mathematics or logic are examined.,37473
domain-specific for three identifying classes of appropriate the problem examine applications. We,We examine the problem of identifying appropriate three classes for domain-specific applications.,37474
areas. texture approach attention-based propose an and a between areas to smooth discrimination give We,We propose an attention-based approach to give a discrimination between texture areas and smooth areas.,37475
"positions frequency out. are carried the located, compensation frequency After of high high details is","After the positions of high frequency details are located, high frequency compensation is carried out.",37476
previously with can incorporate This approach networks. SISR proposed,This approach can incorporate with previously proposed SISR networks.,37477
"frequency enhancement, high effect better achieved. By and are visual performance providing","By providing high frequency enhancement, better performance and visual effect are achieved.",37478
DenseRes SISR also We our propose composed of own blocks. network,We also propose our own SISR network composed of DenseRes blocks.,37479
vehicles emergency and Such include traffic for applications detecting response monitoring.,Such applications include detecting vehicles for emergency response and traffic monitoring.,37480
be and will The available. publicly code models,The code and models will be publicly available.,37481
frames/shots. an selecting plays key by in understanding video summarization role Video important,Video summarization plays an important role in video understanding by selecting key frames/shots.,37482
"to adversarial a we query-conditioned propose this this tackle generative three-player network In paper, challenge.","In this paper, we propose a query-conditioned three-player generative adversarial network to tackle this challenge.",37483
"dominant become in deep force has algorithm Recently, development. a learning","Recently, deep learning has become a dominant force in algorithm development.",37484
seriously all can training efficacy. Direct undermine on training samples,Direct training on all samples can seriously undermine training efficacy.,37485
choice a Monte anytime for popular sequential Carlo (MCTS) solving tree problems. search is,Monte Carlo tree search (MCTS) is a popular choice for solving sequential anytime problems.,37486
"which to be signal, can difficult define. a it depends numeric However, on feedback","However, it depends on a numeric feedback signal, which can be difficult to define.",37487
"are domains. in such biased heuristics erroneous especially and Hence inevitable,","Hence erroneous and biased heuristics are inevitable, especially in such domains.",37488
feedback indications beneficial. into be find We translating absolute may also ordinal that,We also find indications that translating absolute into ordinal feedback may be beneficial.,37489
"a method. WMI detection segmentation-free, Hence, fast, we atlas-free unsupervised, propose","Hence, we propose a segmentation-free, fast, unsupervised, atlas-free WMI detection method.",37490
a fast linear ventricles Extremal detect Stable Maximally Regions blobs using algorithm. We the as,We detect the ventricles as blobs using a fast linear Maximally Stable Extremal Regions algorithm.,37491
"are appropriate Thereafter, discriminated positives heuristics. false using","Thereafter, false positives are discriminated using appropriate heuristics.",37492
"learning we first active paper deep the stereo this ActiveStereoNet, solution In for systems. present","In this paper we present ActiveStereoNet, the first deep learning solution for active stereo systems.",37493
depth This component predictions along is discontinuities. reduce particularly and blur improves to crucial,This component is crucial to reduce blur and particularly improves predictions along depth discontinuities.,37494
"Standard matching, computational dictionary to scalability maps reconstructs which lacks due MRF using parametric inefficiency.","Standard MRF reconstructs parametric maps using dictionary matching, which lacks scalability due to computational inefficiency.",37495
"reconstruction compared approach. time addition, the to a significantly dictionary-based In is reduced","In addition, the reconstruction time is significantly reduced compared to a dictionary-based approach.",37496
"controlling measuring reaction while and stimuli Here, investigate brain participant we time. network meaningfulness modularity","Here, we investigate brain network modularity while controlling stimuli meaningfulness and measuring participant reaction time.",37497
of and reaction time ii) network And there correlation the participants? between is a modularity,And ii) is there a correlation between network modularity and the reaction time of participants?,37498
for for were than Integration occurrence and meaningful meaningless images. greater,Integration and occurrence were greater for meaningless than for meaningful images.,37499
"general deep we the neural focus this of work, on attention In in effect networks.","In this work, we focus on the effect of attention in general deep neural networks.",37500
"Our channel along an separate pathways, map spatial. and infers attention module two","Our module infers an attention map along two separate pathways, channel and spatial.",37501
be and publicly The models code will available.,The code and models will be publicly available.,37502
"less the angles. In uncontrolled results pattern it variations on addition,","In addition, it results less pattern variations on the uncontrolled angles.",37503
Training to segmentation network labeled a semantic amounts large of perform deep data. requires,Training a deep network to perform semantic segmentation requires large amounts of labeled data.,37504
"network real on relatively performs trained a synthetic data poorly on Unfortunately, images.","Unfortunately, a network trained on synthetic data performs relatively poorly on real images.",37505
multi-script for in using architecture document end-to-end this retrieval paper. An proposed handwritten signatures is,An end-to-end architecture for multi-script document retrieval using handwritten signatures is proposed in this paper.,37506
"In features second (i.e. stage, from the foreground the","In the second stage, features from the foreground (i.e.",37507
the (i.e. strokes) and background spatial information signature,signature strokes) and the background spatial information (i.e.,37508
the characterize to object with match signature combined to the were signature. query,were combined to characterize the signature object to match with the query signature.,37509
obtained. and results The promising proposed system documents was on were also tested noisy,The proposed system was also tested on noisy documents and promising results were obtained.,37510
the A comparative approaches. state-of-the-art study that outperforms shows method proposed the,A comparative study shows that the proposed method outperforms the state-of-the-art approaches.,37511
an Image where ill-posed are restoration role. the important problems plays regularization problems typical term,Image restoration problems are typical ill-posed problems where the regularization term plays an important role.,37512
"interpolations convolution. transposed also of Further, we place in multiple have added","Further, we have also added multiple interpolations in place of transposed convolution.",37513
"computationally obtained perceptually being very while The results efficient. in quality, are similar","The results obtained are perceptually similar in quality, while being computationally very efficient.",37514
propose colorization. first for exemplar-based the We learning local deep approach,We propose the first deep learning approach for exemplar-based local colorization.,37515
automatically the The picking by be reference can fully top simply suggestion. colorization performed,The colorization can be performed fully automatically by simply picking the top reference suggestion.,37516
user and Our through methods. study favorable comparisons the-state-of-the-art is to a quantitative approach validated,Our approach is validated through a user study and favorable quantitative comparisons to the-state-of-the-art methods.,37517
"be Furthermore, extended naturally colorization. approach to our video can","Furthermore, our approach can be naturally extended to video colorization.",37518
and for models freely code available use. public Our be will,Our code and models will be freely available for public use.,37519
"previous Compared is works, our based neural edge-aware. with consolidation network","Compared with previous neural network based works, our consolidation is edge-aware.",37520
relays. the Access and assumed the for sources is Multiple Division Time,Time Division Multiple Access is assumed for the sources and the relays.,37521
Sources the time in slots phase. transmit successively first for,Sources transmit successively in time slots for the first phase.,37522
limited The second slots phase for a of number time of consists retransmissions.,The second phase consists of a limited number of time slots for retransmissions.,37523
asymmetric conclusion scenarios. source for both valid same The and rate remains symmetric,The same conclusion remains valid for both symmetric and asymmetric source rate scenarios.,37524
processes (MDPs) planning Markov such problems. to models capture decision prominent are,Markov decision processes (MDPs) are prominent models to capture such planning problems.,37525
tradeoffs the discuss exploration and between We safety. sufficient of environment in ensuring progress,We discuss tradeoffs between sufficient progress in exploration of the environment and ensuring safety.,37526
scenario. selective to in classification poorly are attribute-based Existing classification zero-shot methods the work shown,Existing attribute-based zero-shot classification methods are shown to work poorly in the selective classification scenario.,37527
the under-complete human attribute the poor accounts for We defined vocabulary argue performance.,We argue the under-complete human defined attribute vocabulary accounts for the poor performance.,37528
is residual the learning and firstly the proposed jointly. by The attributes constructed classifier defined,The proposed classifier is constructed by firstly learning the defined and the residual attributes jointly.,37529
defined Then predictions subspace the are the of within attributes. conducted the,Then the predictions are conducted within the subspace of the defined attributes.,37530
"measured the confidence and defined by prediction both Finally, attributes. the residual the is","Finally, the prediction confidence is measured by both the defined and the residual attributes.",37531
"is to be time-consuming medical a task by performed experts. However,it","However,it is a time-consuming task to be performed by medical experts.",37532
"So, automatic are methods reliable required. segmentation and","So, automatic and reliable segmentation methods are required.",37533
clinical. performs can proposed fast and be The used in method,The proposed method performs fast and can be used in clinical.,37534
mechanism for enhance recognition. novel a Networks to Convolutional Neural propose fine-grained attention We,We propose a novel attention mechanism to enhance Convolutional Neural Networks for fine-grained recognition.,37535
and analysis. and Object satellite fundamental challenging is a image detection problem in aerial,Object detection is a fundamental and challenging problem in aerial and satellite image analysis.,37536
satisfy dense scenes. boxes in can These candidate sparse and tasks detection,These candidate boxes can satisfy detection tasks in sparse and dense scenes.,37537
to Experimental superiority results the state-of-the-art. of demonstrate the the framework proposed,Experimental results demonstrate the superiority of the proposed framework to the state-of-the-art.,37538
"independently. re-id classical However, per-camera lists generate approaches","However, classical re-id approaches generate per-camera lists independently.",37539
"sequential multi-camera propose a shortcoming, novel re-id we To this approach. address","To address this shortcoming, we propose a novel sequential multi-camera re-id approach.",37540
custom-designed query improve to incrementally We formulate an procedure optimization representation.,We formulate an optimization procedure custom-designed to incrementally improve query representation.,37541
"a we conduct subject-based comparative human operator Additionally, study of performance.","Additionally, we conduct a comparative subject-based study of human operator performance.",37542
implementation as of this sort. As publicly know this is far the we first available,As far as we know this is the first publicly available implementation of this sort.,37543
"max/min novel this In a for a we work, shift-register-based architecture stochastic propose function.","In this work, we propose a novel shift-register-based architecture for a stochastic max/min function.",37544
"logo, embroidery) texture, that clothes. original depict the","texture, logo, embroidery) that depict the original clothes.",37545
"is namely, proposed. address Feature) a skeletonbased these To representation, tasks, (Skeleton Pose-Motion SPMF","To address these tasks, a skeletonbased representation, namely, SPMF (Skeleton Pose-Motion Feature) is proposed.",37546
"represent effectively to complex Therefore, able they actions. are","Therefore, they are able to effectively represent complex actions.",37547
requiring results that computation. indicated less method state-of-the-art methods proposed surpasses Experimental whilst the,Experimental results indicated that the proposed method surpasses state-of-the-art methods whilst requiring less computation.,37548
theoretical and suggest models for both directions experimental inquiry. Peptide/IMT,Peptide/IMT models suggest directions for both theoretical and experimental inquiry.,37549
"proposed shortcomings with networks capsule of deal to Neural were Networks Convolutional Recently, (ConvNets).","Recently, capsule networks were proposed to deal with shortcomings of Convolutional Neural Networks (ConvNets).",37550
Power from gain angle-steering-based then beamforming CSI analyzed. tracked is using,Power gain from angle-steering-based beamforming using tracked CSI is then analyzed.,37551
in intra-cluster spreads. under provides a beam-width different design study angle-steering The angular guideline theoretical,The theoretical study provides a design guideline beam-width in angle-steering under different intra-cluster angular spreads.,37552
"improvements. many have made, been is advancements While room there for continued","While many advancements have been made, there is room for continued improvements.",37553
DIR for its Yet usefulness not was assessed.,Yet its usefulness for DIR was not assessed.,37554
"two differentiate changes, that (e.g. populations the We which geometric show","We show that geometric changes, which differentiate the two populations (e.g.",37555
"synthesized readily as well. vs. inhale are exhale),","inhale vs. exhale), are readily synthesized as well.",37556
"In radio recent years, as spectrum. cognitive introduced efficient of for use a tool network","In recent years, cognitive radio network introduced as a tool for efficient use of spectrum.",37557
with primary must users Secondary that users are unauthorized from interferences transmission. avoid ones,Secondary users are unauthorized ones that must avoid from interferences with primary users transmission.,37558
The structured for hierarchically has tractable importance of been planning representations long acknowledged.,The importance of hierarchically structured representations for tractable planning has long been acknowledged.,37559
(MRF) Resonance exception. is Fingerprinting an Magnetic not,Magnetic Resonance Fingerprinting (MRF) is not an exception.,37560
The are existing addressing neither methods nor accurate. PVE scalable,The existing methods addressing PVE are neither scalable nor accurate.,37561
"methods. that PVE, with algorithm outperforms We show, synthetic simulations on through data state-of-the-art our","We show, through simulations on synthetic data with PVE, that our algorithm outperforms state-of-the-art methods.",37562
and approach in phantom Our on datasets. EUROSPIN vivo is validated the on,Our approach is validated on the EUROSPIN phantom and on in vivo datasets.,37563
image and This a system text proposes on cross-modal leverages that retrieval paper encoding.,This paper proposes a cross-modal retrieval system that leverages on image and text encoding.,37564
modalities. can We encodings between that capture semantic show multiple relationships text,We show that text encodings can capture semantic relationships between multiple modalities.,37565
a removal inpainting. new for and introduce system content image automatic We,We introduce a new system for automatic image content removal and inpainting.,37566
and a single performed inpainting jointly in are pass. Region segmentation,Region segmentation and inpainting are performed jointly in a single pass.,37567
"more this segmentation module. are inpainting by In way, errors alleviated naturally the potential","In this way, potential segmentation errors are more naturally alleviated by the inpainting module.",37568
GAN is trained model using The a whole strategy. conditional,The whole model is trained using a conditional GAN strategy.,37569
simulation illustrate greedy formulation studies proposed of extensive efficacy and algorithm. the the the Our,Our extensive simulation studies illustrate the efficacy of the proposed formulation and the greedy algorithm.,37570
two-level is proposed. motion A optimization node,A two-level node motion optimization is proposed.,37571
frame Time-aware a problem video video a fundamental is encoding of in understanding. in sequences,Time-aware encoding of frame sequences in a video is a fundamental problem in video understanding.,37572
"aim video To this we evaluate explicitly. lacuna, fill time to","To fill this lacuna, we aim to evaluate video time explicitly.",37573
task associated on Based to the a we quantify formulate each able property.,Based on each we formulate a task able to quantify the associated property.,37574
These in properties domain. are common medical,These properties are common in medical domain.,37575
"multiple of consider classes. most However, don't relationship them between","However, most of them don't consider relationship between multiple classes.",37576
"network designed fine-grained MSML. concept classification have on architecture based that deep incorporates Moreover, we","Moreover, we have designed deep network architecture based on fine-grained classification concept that incorporates MSML.",37577
two accuracy The and stereo of are speed. factors important vision,The two important factors of stereo vision are accuracy and speed.,37578
"promote species. can On the hand, other chaos coexistence of","On the other hand, chaos can promote coexistence of species.",37579
we this higher in biodiversity. Moreover that results a confirm,Moreover we confirm that this results in a higher biodiversity.,37580
"structural out. analysis relatively classification, regions exhaustive For and carried of contours is a","For classification, a relatively exhaustive structural analysis of contours and regions is carried out.",37581
of algorithms at the expense cost. accuracy learning computation significant Deep classification achieve high,Deep learning algorithms achieve high classification accuracy at the expense of significant computation cost.,37582
heart define disease. causes in and of well-established the the cardiovascular Alterations of function geometry,Alterations in the geometry and function of the heart define well-established causes of cardiovascular disease.,37583
phenotypes. and factors limit structural the sensitivity complex in quantifying functional Both,Both factors limit the sensitivity in quantifying complex structural and functional phenotypes.,37584
"its time we paper, first for the for this In SAR tomography. demonstrate potential","In this paper, we demonstrate for the first time its potential for SAR tomography.",37585
set. richer data used is Image a to get enhancement,Image enhancement is used to get a richer data set.,37586
"and the achieves joint testing is good test, used In model results. for the","In the test, the model is used for joint testing and achieves good results.",37587
aleatoric our combine proposed compare with We and uncertainty uncertainty. model,We compare and combine our proposed aleatoric uncertainty with model uncertainty.,37588
"their image unsatisfactory. is quality However, still","However, their image quality is still unsatisfactory.",37589
"the same proposed outperforms recent quantitative the algorithms addressing and qualitative For algorithm evaluations, problem.","For qualitative and quantitative evaluations, the proposed algorithm outperforms recent algorithms addressing the same problem.",37590
"However, in be itself cue. important an location can","However, location in itself can be an important cue.",37591
for We adding location approaches information. test two,We test two approaches for adding location information.,37592
experiments perform with We location. indirect encode ways direct both to and,We perform experiments with both direct and indirect ways to encode location.,37593
array from Our digital greatly shows multi-user analysis benefits multiplexing.,Our analysis shows digital array benefits greatly from multi-user multiplexing.,37594
knowledge data detection. from transferring) sensor and low-resolution target enhancing,transferring) knowledge from sensor data and enhancing low-resolution target detection.,37595
"and approaches that performance models. better demonstrate recent achieves we traditional teacher-student Moreover, than it","Moreover, we demonstrate that it achieves better performance than traditional approaches and recent teacher-student models.",37596
encoder-decoder built structure. segmentation an mainly architectures upon are Semantic,Semantic segmentation architectures are mainly built upon an encoder-decoder structure.,37597
downsampling perform the operations models in encoder. These subsequent,These models perform subsequent downsampling operations in the encoder.,37598
"commonly ordered diagnostic cost-effective physicians. tests the by Chest imaging Background: X-rays are performed, most","Background: Chest X-rays are the most commonly performed, cost-effective diagnostic imaging tests ordered by physicians.",37599
"Treatment measurements contemplate some protocols usually assess order to recurrent such as progress, in goniometry.","Treatment protocols usually contemplate some recurrent measurements in order to assess progress, such as goniometry.",37600
to the angle. flexion are values These estimate sufficient,These values are sufficient to estimate the flexion angle.,37601
were for coordinates. these trained predicting networks six neural Convolutional then,Convolutional neural networks were then trained for predicting these six coordinates.,37602
van T. works contraction Kampen followed of Kurtz. Their of and the,Their contraction followed the works of van Kampen and of T. Kurtz.,37603
connected variables S LE. This to the noise the in subunit channel terms the,This S connected the noise terms to the channel subunit variables in the LE.,37604
"that the of numerous Subsequently, were studies these models. simplified several showed shortcomings published","Subsequently, numerous studies were published that showed the several shortcomings of these simplified models.",37605
that results it with produced good results. agreement the in They MC very showed,They showed that it produced results in very good agreement with the MC results.,37606
"regression shown classification CNN-based performance and significant improvements this for task. methods have Recently,","Recently, CNN-based regression and classification methods have shown significant performance improvements for this task.",37607
allows instances. This pixels to object individual us group into,This allows us to group object pixels into individual instances.,37608
"Photoacoustic most beamformer. used In (DAS) commonly is the imaging, algorithm Delay-and-Sum","In Photoacoustic imaging, Delay-and-Sum (DAS) algorithm is the most commonly used beamformer.",37609
"and leads of low level sidelobes. a it high to resolution However,","However, it leads to a low resolution and high level of sidelobes.",37610
sidelobes (DMAS) introduced compared to Delay-Multiply-and-Sum was DAS. to provide lower,Delay-Multiply-and-Sum (DMAS) was introduced to provide lower sidelobes compared to DAS.,37611
"artificially examples, of reducing increasing improves it and By the generalization. overfitting helps number training","By artificially increasing the number of training examples, it helps reducing overfitting and improves generalization.",37612
"actually Otherwise, strategy we the hurts. previous show that","Otherwise, we show that the previous strategy actually hurts.",37613
"to CT pixel parameters. differences have in due varying However, spacing acquisition studies","However, CT studies have varying pixel spacing due to differences in acquisition parameters.",37614
improve the are performance. used ensembles further also Multi-model to,Multi-model ensembles are also used to further improve the performance.,37615
of  images. consider problem scene search We  person the unconstrained in,  We consider the problem of person search in unconstrained scene images.,37616
"multi-scale person barrier. is matching problem under-studied the in Faster-RCNN), severe a and more search","Faster-RCNN), and the under-studied multi-scale matching problem in person search is a more severe barrier.",37617
to facilitate simultaneously increasing Automatic segmentation potential has while efficiency. morphological great measurements,Automatic segmentation has great potential to facilitate morphological measurements while simultaneously increasing efficiency.,37618
closed is for expression CSCRB the derived. A form,A closed form expression for the CSCRB is derived.,37619
"such Up been image have approaches to limited single exclusively to now, however, super-resolution.","Up to now, however, such approaches have been exclusively limited to single image super-resolution.",37620
"recurrent proposed information architecture, previous from frames to due its leverages i.e. architecture naturally The","The proposed architecture naturally leverages information from previous frames due to its recurrent architecture, i.e.",37621
provide a novel combining merging for (Eg. method We view information view dependent,We provide a novel view merging method for combining view dependent information (Eg.,37622
paper new This learning video-based for a classification. deep approach scene presents,This paper presents a new deep learning approach for video-based scene classification.,37623
reliable assist to methods doctor. required are and Automatic segmenta-tion,Automatic and reliable segmenta-tion methods are required to assist doctor.,37624
"is deep for ordinary. the inspiration, the always which rare improvement networks But of needs","But the improvement of deep networks always needs inspiration, which is rare for the ordinary.",37625
These are own have different for data characteristic distinctive their methods and sets.,These different methods have their own characteristic and are distinctive for data sets.,37626
structure. on dynamics network network Neuronal depends,Neuronal network dynamics depends on network structure.,37627
"processors for to known bias ""burn-in"" complicate the of use parallel computations. is This MCMC","This ""burn-in"" bias is known to complicate the use of parallel processors for MCMC computations.",37628
across recognition robustly face wide variety Current recognize of identities conditions. systems a imaging,Current face recognition systems robustly recognize identities across a wide variety of imaging conditions.,37629
or problem problem polycube construction deformation of The graphics. computer in is an essential,The problem of polycube construction or deformation is an essential problem in computer graphics.,37630
corresponding its a shape. a derive polycube We between mesh relationship clear and,We derive a clear relationship between a mesh and its corresponding polycube shape.,37631
"with works meshes boundaries. Our without is algorithm or and on edge-preserved, surface","Our algorithm is edge-preserved, and works on surface meshes with or without boundaries.",37632
"in Our outperforms algorithm efficiency. ones speed, previous robustness,","Our algorithm outperforms previous ones in speed, robustness, efficiency.",37633
methods. favorably other that compares We polycube demonstrat to method deformation our state-of-the-art,We demonstrat that our method compares favorably to other state-of-the-art polycube deformation methods.,37634
"longer often generalise RNNs training. the to sequences ones to However, than during struggle encountered","However, RNNs often struggle to generalise to sequences longer than the ones encountered during training.",37635
"induction. we In networks for to work, explicitly neural optimise propose this","In this work, we propose to optimise neural networks explicitly for induction.",37636
"brain brain with which generate models, compare functional models similar popular networks. our network We","We compare our brain network models with popular models, which generate similar functional brain networks.",37637
"and advertisements in non-literal interact text in complex, Images ways.","Images and text in advertisements interact in complex, non-literal ways.",37638
Extensive conducted experiments effectiveness. were to its validate,Extensive experiments were conducted to validate its effectiveness.,37639
identify classification to skin the disease. Melanoma stage is serious a,Melanoma classification is a serious stage to identify the skin disease.,37640
cancer patients. biomarker important Tumor proliferation of indicative the of breast prognosis is an,Tumor proliferation is an important biomarker indicative of the prognosis of breast cancer patients.,37641
a of in proliferation Assessment labor-intensive tumor subjective task. and highly is setting clinical,Assessment of tumor proliferation in a clinical setting is highly subjective and labor-intensive task.,37642
investigated study that tumor This proliferation was assessment from WSIs. first the,This was the first study that investigated tumor proliferation assessment from WSIs.,37643
variation and CT in large nodules in appearance suffer Lung size images.,Lung nodules suffer large variation in size and appearance in CT images.,37644
"training speed tackle propose a curriculum strategy To we general up problem, to training. this","To tackle this problem, we propose a general curriculum training strategy to speed up training.",37645
"we state-of-the-art outperforms that the In methods. network previous experiments, proposed demonstrate","In experiments, we demonstrate that the proposed network outperforms previous state-of-the-art methods.",37646
queries suffers from the significant space latter expensive The and from former overhead.,The former suffers from expensive queries and the latter from significant space overhead.,37647
hypothesis. supports the our and study of visualization The further learned context generalizability,The visualization and generalizability study of the learned context further supports our hypothesis.,37648
energy a has succinct Calabi format. and explicit,Calabi energy has a succinct and explicit format.,37649
Its under is conditions. and certain conformal flow convergent corresponding,Its corresponding flow is conformal and convergent under certain conditions.,37650
"this present RGB-D system RGB-iD namely SLAM we for a In SLAM. complete paper cameras,","In this paper we present a complete SLAM system for RGB-D cameras, namely RGB-iD SLAM.",37651
keyframes the back-end receives which front-end. is from second thread The a,The second thread is a back-end which receives keyframes from the front-end.,37652
propose RSIR (FCN). approach novel networks fully multi-label We with a therefore convolutional,We therefore propose a novel multi-label RSIR approach with fully convolutional networks (FCN).,37653
ambiguous. from challenging self-occluded regions both and images the hidden is Inferring,Inferring the hidden regions from self-occluded images is both challenging and ambiguous.,37654
propose two-pronged We these issues. a address approach to,We propose a two-pronged approach to address these issues.,37655
achieved latent by novel probablistic a diversity view-specific learning space loss. with a is This,This is achieved by learning a probablistic latent space with a novel view-specific diversity loss.,37656
qualitative of the significance performed quantitative highlight approach. the is Thorough to and analysis proposed,Thorough quantitative and qualitative analysis is performed to highlight the significance of the proposed approach.,37657
"alternating is non-convex developed minimization solve algorithm this, to this problem. Following an","Following this, an alternating minimization algorithm is developed to solve this non-convex problem.",37658
that architecture proposed the volumetric encodes a labels knowledge the proportion. represent The,The proposed architecture encodes the knowledge that the labels represent a volumetric proportion.,37659
with is designed custom A to learn loss intervals.,A custom loss is designed to learn with intervals.,37660
"achieved performance. AI have being level human few Despite target developers, several for a","Despite being a target for several AI developers, few have achieved human level performance.",37661
surge communications. motivated is displays This VR/AR by of face-to-face recent hinder that,This is motivated by recent surge of VR/AR displays that hinder face-to-face communications.,37662
was temporal two Video shot mechanisms. on based pooling representation,Video shot representation was based on two temporal pooling mechanisms.,37663
"been evaluation have Novel visualization protocols, and result strategies, accordingly. designed methods training","Novel training strategies, evaluation protocols, and result visualization methods have been designed accordingly.",37664
"However, into large does images. dependence the not dataset well on medical translate","However, the dependence on large dataset does not translate well into medical images.",37665
"course of model, the and this disappear inter-residue during time evolution. form the In contacts","In this model, inter-residue contacts form and disappear during the course of the time evolution.",37666
differ geometrical model. substantially of that this properties We homopeptides the demonstrate various in,We demonstrate that the geometrical properties of various homopeptides differ substantially in this model.,37667
two using of the introduction The adversarial single the problems network tackles discriminator. a,The introduction of the adversarial network tackles the two problems using a single discriminator.,37668
"kinds the paper of inconsistencies, this Adversarial address Macro-Micro the Net proposes two (MMAN). To","To address the two kinds of inconsistencies, this paper proposes the Macro-Micro Adversarial Net (MMAN).",37669
"dataset generalization our impressive pre-trained ability. model small a demonstrates PPSS, relatively On","On a relatively small dataset PPSS, our pre-trained model demonstrates impressive generalization ability.",37670
is Cellular associated strain. contractility assumed imply often to an compressive,Cellular contractility is often assumed to imply an associated compressive strain.,37671
This of stiffer for on substrates. mechanism a the molecules suggests mechanotransductive stretch-activation physical,This suggests a physical mechanism for the stretch-activation of mechanotransductive molecules on stiffer substrates.,37672
"on supporting adhesions, recent results focal whole-cell experimental mechanically-driven mechanotransduction.","focal adhesions, supporting recent experimental results on whole-cell mechanically-driven mechanotransduction.",37673
non-Lambertian photometric problem This for paper of the surfaces. stereo addresses,This paper addresses the problem of photometric stereo for non-Lambertian surfaces.,37674
"on generalize we data, real datasets. synthetic can well Although PS-FCN train it on","Although we train PS-FCN on synthetic data, it can generalize well on real datasets.",37675
This gap. to paper fill aims this,This paper aims to fill this gap.,37676
"nearly under conditions prove, same the with as We i.i.d. first","We first prove, under nearly the same conditions as with i.i.d.",37677
"data, convergence clustering. the weak under processes empirical multiway of","data, the weak convergence of empirical processes under multiway clustering.",37678
"consistency estimators, establish asymptotic various of that of including We variance et al. then Cameron","We then establish consistency of various asymptotic variance estimators, including that of Cameron et al.",37679
network great Deep artificial applications. has neural intelligence achieved success in many convolution,Deep convolution neural network has achieved great success in many artificial intelligence applications.,37680
large benchmark local for detectors. the We scale of present evaluation a feature,We present a large scale benchmark for the evaluation of local feature detectors.,37681
"the loss. signal supervision to classification Conventionally, as been penalize softmax trained with have CNNs","Conventionally, CNNs have been trained with softmax as supervision signal to penalize the classification loss.",37682
deep considered face and recognition maximization features of ideal are minimization Such task. for,Such minimization and maximization of deep features are considered ideal for face recognition task.,37683
shown results. to breast cancer in Deep mammograms recently approaches promising learning detection have,Deep learning approaches to breast cancer detection in mammograms have recently shown promising results.,37684
"generating we First, that high-resolution of synthetic capable mammogram show GANs are patches.","First, we show that GANs are capable of generating high-resolution synthetic mammogram patches.",37685
"performance. evaluate cancer classification breast augmented experimentally dataset the improve Next, we to using","Next, we experimentally evaluate using the augmented dataset to improve breast cancer classification performance.",37686
arbitrary This captures paper over timestamps. describes network a that multimodal correlations,This paper describes a network that captures multimodal correlations over arbitrary timestamps.,37687
two the The streams. approach fusion existing averages,The existing fusion approach averages the two streams.,37688
fusion we propose for CNN. learning a a Shannon Here network pre-trained correlation with a,Here we propose a correlation network with a Shannon fusion for learning a pre-trained CNN.,37689
complement methods. network to fusion approach existing This the found was,This approach was found to complement the existing network fusion methods.,37690
the recognition of video multimodal the The results. accuracy correlation enhanced,The multimodal correlation enhanced the accuracy of the video recognition results.,37691
constrained We propose algorithm policy objective function. a optimize to the gradient,We propose a policy gradient algorithm to optimize the constrained objective function.,37692
applications. video many important Localization Temporal untrimmed Action (TAL) in for is,Temporal Action Localization (TAL) in untrimmed video is important for many applications.,37693
"generic, and carefully consider not special approaches scenes. remain of properties However, do the largely","However, approaches remain largely generic, and do not carefully consider the special properties of scenes.",37694
are perform regions a then to used scene source recognition. features of as These,These regions are then used as a source of features to perform scene recognition.,37695
pedestrian evaluated publicly on also Grand is available performance New the The Central database. York,The performance is also evaluated on the publicly available New York Grand Central pedestrian database.,37696
that Producing but weren't. been answer over have seen images requires an could marginalizing,Producing an answer requires marginalizing over images that could have been seen but weren't.,37697
"This a as pooling}. descriptor for \emph{discriminative used the subspace is subspace dubbed video, then","This subspace is then used as a descriptor for the video, dubbed \emph{discriminative subspace pooling}.",37698
video We and datasets experiments results. on several state-of-the-art demonstrate provide,We provide experiments on several video datasets and demonstrate state-of-the-art results.,37699
crowds a dense even Counting is for demanding humans. task in people,Counting people in dense crowds is a demanding task even for humans.,37700
appearance to variability primarily large the in people. of due This is,This is primarily due to the large variability in appearance of people.,37701
of seen a only Often blobs. people as are bunch,Often people are only seen as a bunch of blobs.,37702
"clutter further pose difficulty. variations Occlusions, the background and compound","Occlusions, pose variations and background clutter further compound the difficulty.",37703
"In semantics scenario, the requires a identifying of and spatial context larger scene. this person","In this scenario, identifying a person requires larger spatial context and semantics of the scene.",37704
"to prediction the we of Hence, correct the initial CNN. feedback propose top-down","Hence, we propose top-down feedback to correct the initial prediction of the CNN.",37705
are various bottom-up the the top-down from fed of CNN layers network. to Features,Features from various layers of the bottom-up CNN are fed to the top-down network.,37706
"of crucial. step is process, dermoscopic the first in Segmentation is accuracy images thus this","Segmentation of dermoscopic images is the first step in this process, thus accuracy is crucial.",37707
tackle paper problem. to an re-identification approach This presents the,This paper presents an approach to tackle the re-identification problem.,37708
for person are machine to More datasets models train and more learning available re-identification.,More and more datasets are available to train machine learning models for person re-identification.,37709
"These positions, season, i.e. in conditions: numbers, location, cameras camera in datasets size, vary","These datasets vary in conditions: cameras numbers, camera positions, location, season, in size, i.e.",37710
"different number of identities. of images, number","number of images, number of different identities.",37711
labeling: annotated Finally there in while are are with others datasets not. attributes,Finally in labeling: there are datasets annotated with attributes while others are not.,37712
on Convolutional multitask learning. trained is (CNN) Our Neural Network a based using and model,Our model is based on a Convolutional Neural Network (CNN) and trained using multitask learning.,37713
to different are available Several in datasets. losses extract the the information used different,Several losses are used to extract the different information available in the different datasets.,37714
task classification learned Our with main is loss. a,Our main task is learned with a classification loss.,37715
the reduce To we with intra-class variation the experiment loss. center,To reduce the intra-class variation we experiment with the center loss.,37716
We works datasets. our that also outperforms two recent show re-identification system on,We also show that our system outperforms recent re-identification works on two datasets.,37717
consider streams. the video of highlight-detection problem automatic We in game,We consider the problem of automatic highlight-detection in video game streams.,37718
"present unsupervised methodology for multi-view deep novelty-based learning this we In light, a detection. highlight","In this light, we present a multi-view unsupervised deep learning methodology for novelty-based highlight detection.",37719
frequency windowed confinement. (OFDM) signals with multiplexing low pulses exhibit spectral Orthogonal division rectangularly,Orthogonal frequency division multiplexing (OFDM) signals with rectangularly windowed pulses exhibit low spectral confinement.,37720
two This strategies. method and shaping proposes that a unifies generalizes paper spectral these,This paper proposes a spectral shaping method that generalizes and unifies these two strategies.,37721
hierarchical. structures inherently key are indoor that observation Our scene is,Our key observation is that indoor scene structures are inherently hierarchical.,37722
"is or not recursive network a neural convolutional; it RvNN. our is network Hence,","Hence, our network is not convolutional; it is a recursive neural network or RvNN.",37723
"By a autoencoder a (VAE), resulting training roughly variational the distribution. follow Gaussian codes fixed-length","By training a variational autoencoder (VAE), the resulting fixed-length codes roughly follow a Gaussian distribution.",37724
"for method GRAINS, for We Generative our Autoencoders INdoor coin Scenes. Recursive","We coin our method GRAINS, for Generative Recursive Autoencoders for INdoor Scenes.",37725
localization on methods based generate produced usually maps supervised attention networks. results by classification Weakly,Weakly supervised methods usually generate localization results based on attention maps produced by classification networks.,37726
progressively are regions learn SPG confident utilized to maps masks. the high within The attention,The high confident regions within attention maps are utilized to progressively learn the SPG masks.,37727
object Extensive high-quality demonstrate is producing that ILSVRC maps. effective on localizations SPG in experiments,Extensive experiments on ILSVRC demonstrate that SPG is effective in producing high-quality object localizations maps.,37728
"However, the by of is heterogeneous often challenged datasets. labeling this","However, this is often challenged by heterogeneous labeling of the datasets.",37729
human analyzing role Facial cognitive a state. significant in expression has,Facial expression has a significant role in analyzing human cognitive state.,37730
"scale. performance influenced user-specified inlier much is the by its First,","First, its performance is much influenced by the user-specified inlier scale.",37731
"large is it data. Second, computationally inefficient for","Second, it is computationally inefficient for large data.",37732
algorithm reduction problem MaxFS the for data the of This computationally makes realistic.,This reduction of data for the MaxFS problem makes the algorithm computationally realistic.,37733
"the As object a reconstruct LSN cluttered result, and effectively can backgrounds skeletons. suppress","As a result, LSN can effectively suppress the cluttered backgrounds and reconstruct object skeletons.",37734
results proposed of state-of-the-art the validate performance LSN. the Experimental,Experimental results validate the state-of-the-art performance of the proposed LSN.,37735
images. to human learning the in consider pose still task We estimate of,We consider the task of learning to estimate human pose in still images.,37736
denoising application essential imaging. medical which Our is fluoroscopy is image in,Our application is medical image denoising which is essential in fluoroscopy imaging.,37737
"needs a physician, flavor and tailored each different to quality individual. image towards be has","physician, has a different flavor and image quality needs to be tailored towards each individual.",37738
Also user specific best users test this perform on trained data. models a for,Also models trained for a specific user perform best on this users test data.,37739
and spatial efficiency high owing is resolution. Brovey method its popular to pan-sharpening a The,The Brovey is a popular pan-sharpening method owing to its efficiency and high spatial resolution.,37740
be method mathematical optical This sensing can model sensors. explained remote of by,This method can be explained by mathematical model of optical remote sensing sensors.,37741
model of MNIST handwritten A is presented recognition digit here. simple,A simple model of MNIST handwritten digit recognition is presented here.,37742
adaptation model is face recognition. of of an previous The theory a,The model is an adaptation of a previous theory of face recognition.,37743
"We present, of achieved example, way by the algorithms. by results state-of-the-art","We present, by way of example, the results achieved by state-of-the-art algorithms.",37744
tasks this the team all in challenge. in participates Our,Our team participates in all the tasks in this challenge.,37745
"in a is attribute detection way. task, lesion In multi-label also PSPNet the adopted modified","In lesion attribute detection task, the modified PSPNet is also adopted in a multi-label way.",37746
higher higher accuracy achieves SAW-SS than the cost sensing of computational complexity. at SAE-TF,SAE-TF achieves higher sensing accuracy than SAW-SS at the cost of higher computational complexity.,37747
"i.e. multiscale in applied method This version, is a","This method is applied in a multiscale version, i.e.",37748
the variable index instead time. is scale of,the variable index is scale instead of time.,37749
model consistent natural show with We is PAN MS our images. that and more,We show that our model is more consistent with natural MS and PAN images.,37750
for sampling single diffusion schemes novel and MRI. This presents paper multi-shell,This paper presents novel single and multi-shell sampling schemes for diffusion MRI.,37751
is . data used also qualitatively Human brain evaluate to reconstruction,Human brain data is also used to qualitatively evaluate reconstruction .,37752
and algorithms systems family wireless in radar. communication The constant-modulus used widely is of in,The family of constant-modulus algorithms is widely used in wireless communication systems and in radar.,37753
"Instead, entire spectrum. it the equalizes spatial","Instead, it equalizes the entire spatial spectrum.",37754
We a of example high-SNR provide under conditions. approach the numerical the demonstrate validity to,We provide a numerical example to demonstrate the validity of the approach under high-SNR conditions.,37755
"response. model counting strongest simply the by order, the the roots with","the model order, by simply counting the roots with the strongest response.",37756
class a to justify agents Existing generating explanation learn prediction. fluently visual,Existing visual explanation generating agents learn to fluently justify a class prediction.,37757
trust such human agents This concerning users. particularly building fail with is ultimately as in,This is particularly concerning as ultimately such agents fail in building trust with human users.,37758
"prediction, Our capable AI an for is providing arguments of alternative counter explainable agent i.e.","Our explainable AI agent is capable of providing counter arguments for an alternative prediction, i.e.",37759
"counterfactuals, along explanations that classification correct decisions. with the justify","counterfactuals, along with explanations that justify the correct classification decisions.",37760
seamlessly method be into any proposed CNN The architecture. integrated general can and is,The proposed method is general and can be seamlessly integrated into any CNN architecture.,37761
proposed experiments Extensive effectiveness the datasets four on approach. of benchmark show the,Extensive experiments on four benchmark datasets show the effectiveness of the proposed approach.,37762
a and modified based on multi-task the a DeepUNet way. It perform is in segmentation,It is based on a modified DeepUNet and perform the segmentation in a multi-task way.,37763
are different task groups The and into processed pipelines. clustered channels on,The channels are clustered into groups and processed on different task pipelines.,37764
show network the fusion efficient. Experiment is that feature results,Experiment results show that the feature fusion network is efficient.,37765
between controls strength the association unconditional parameters parameters. the top-level of of auxiliary One these,One of these top-level parameters controls the unconditional strength of association between the auxiliary parameters.,37766
"contrast algorithms, assumptions. distributed this approach In Carlo requires similar few to Monte distributional","In contrast to similar distributed Monte Carlo algorithms, this approach requires few distributional assumptions.",37767
illustrated performance of of examples. algorithms with simulated is the number The a,The performance of the algorithms is illustrated with a number of simulated examples.,37768
feature paper a of vision: In we attention. visual this focus specific on,In this paper we focus on a specific feature of vision: visual attention.,37769
"field-of-view review for of we extensive cameras. this provide paper, models large an existing In","In this paper, we provide an extensive review of existing models for large field-of-view cameras.",37770
results We the also qualitative discuss performance models. provide of all and,We also provide qualitative results and discuss the performance of all models.,37771
ability implicitly the vehicles to automated lack Today's others. cooperate with,Today's automated vehicles lack the ability to cooperate implicitly with others.,37772
"policies predefined policies within for algorithm simultaneously learns over the Without and macro-actions, macro-actions.","Without predefined policies for macro-actions, the algorithm simultaneously learns policies over and within macro-actions.",37773
present and novel algorithm We for robust generating a model hypotheses consistent multiple-structure for fitting.,We present a novel algorithm for generating robust and consistent hypotheses for multiple-structure model fitting.,37774
influence parameter on experiments. our settings clearly in of the the expose results We algorithm,We clearly expose the influence of algorithm parameter settings on the results in our experiments.,37775
"strategy multi-model an comparison. we testing propose effective Besides, also by applying","Besides, we also propose an effective testing strategy by applying multi-model comparison.",37776
"CCTV are surveillance views full where not often available. scenario, person practical This reflects","This reflects practical CCTV surveillance scenario, where full person views are often not available.",37777
"samples and gels, filled emulsion two-layered sausages. filled gels, emulsion were The","The samples were emulsion filled gels, two-layered emulsion filled gels, and sausages.",37778
simplifying principle mastication a indicate and subsequent curve The may during master swallowing.,The master curve may indicate a simplifying principle during mastication and subsequent swallowing.,37779
also particular complexity We have a calculated measure.,We have also calculated a particular complexity measure.,37780
swallowing. an increase before displays This just measure,This measure displays an increase just before swallowing.,37781
"may of of different. In a transmitter addition, and time receiver rates a be","In addition, rates of time of a transmitter and a receiver may be different.",37782
"genetic might redundant. female and deplete variance However, choice preferences make","However, female preferences might deplete genetic variance and make choice redundant.",37783
evolutionary empirical the biology. of We and our results to genetics relevance conservation discuss,We discuss the relevance of our results to conservation genetics and empirical evolutionary biology.,37784
great vision computer neural have achieved successes in Convolutional (CNNs) many problems. networks,Convolutional neural networks (CNNs) have achieved great successes in many computer vision problems.,37785
generalization on as Cityscapes) (e.g. as domain another well its capacity,Cityscapes) as well as its generalization capacity on another domain (e.g.,37786
synthesized as images. in show results data unseen realistic We promising our well in as,We show promising results in our synthesized data as well as in unseen realistic images.,37787
"maximization. for paper, efficient deterministic this an algorithm optimization we In propose consensus","In this paper, we propose an efficient deterministic optimization algorithm for consensus maximization.",37788
"fact, worse the return a challenging previous solution. In techniques instances, even may off on","In fact, on challenging instances, the previous techniques may even return a worse off solution.",37789
between since show challenging different very data detection Change can data the characteristics. be multimodal,Change detection between multimodal data can be very challenging since the data show different characteristics.,37790
detect CNN between to the employed two candidate changes epochs. A Siamese then is,A Siamese CNN is then employed to detect candidate changes between the two epochs.,37791
"individual grouped the object patch-based changes. as Finally, changes candidate and verified are","Finally, the candidate patch-based changes are grouped and verified as individual object changes.",37792
"this so explored much However, not been far in direction. has","However, not much has so far been explored in this direction.",37793
"quality Meanwhile, the control. suggestions are on given photogrammetric","Meanwhile, suggestions are given on the photogrammetric quality control.",37794
"indefinite the to FAQ Due global find to maximum. guaranteed the relaxation, is not","Due to the indefinite relaxation, FAQ is not guaranteed to find the global maximum.",37795
"generalization superior demonstrating capabilities. proposed and the U-Net, learning methods The outperform","The proposed methods outperform the U-Net, demonstrating superior learning and generalization capabilities.",37796
(coCSIT) focus channel covariance exploiting Here CSIT on only. we,Here we focus on exploiting channel covariance CSIT (coCSIT) only.,37797
Spatial upsampling precision is learned achieved by edge-aware a function. employing,Spatial precision is achieved by employing a learned edge-aware upsampling function.,37798
network Our image. Siamese and left from uses to right extract a model the features,Our model uses a Siamese network to extract features from the left and right image.,37799
"producing function of capable Leveraging is high-quality this a output. as guide, input color edge-aware","Leveraging color input as a guide, this function is capable of producing high-quality edge-aware output.",37800
"Unlike manual works, matching of point clusters. existing require not we do annotation many","Unlike many existing works, we do not require manual annotation of matching point clusters.",37801
"this set-based The i.e. of face is work recognition, objective","The objective of this work is set-based face recognition, i.e.",37802
implicitly these Both of set-wise classification. for during training learnt qualities are,Both of these qualities are learnt implicitly during training for set-wise classification.,37803
"these this In paper, problem explore the relationship we domains. two between","In this paper, we explore the relationship between these two problem domains.",37804
capacity. the discussed is the system performance average of proposed of in terms hybrid The,The performance of the proposed hybrid system is discussed in terms of the average capacity.,37805
derived by analytical are Monte validated The simulations. Carlo expressions,The derived analytical expressions are validated by Monte Carlo simulations.,37806
of a is also provided. hybrid/PLC between A the hybrid comparison and parallel performance systems,A comparison between the performance of a parallel hybrid/PLC and hybrid systems is also provided.,37807
hybrid one. It found that the hybrid/PLC system the outperforms is,It is found that the hybrid/PLC system outperforms the hybrid one.,37808
"attention. a lot Recently, style has of received transfer","Recently, style transfer has received a lot of attention.",37809
at intersections for intentions the drivers Understanding is autonomous critical vehicles. component a of,Understanding the intentions of drivers at intersections is a critical component for autonomous vehicles.,37810
we give paper contribution. this two In,In this paper we give two contribution.,37811
affects it system. such condition Epilepsy is a brain the neurological the nervous and that,Epilepsy is a neurological condition such that it affects the brain and the nervous system.,37812
"great seizure patients. has of importance clinical epileptic in therapy Hence, identification","Hence, seizure identification has great importance in clinical therapy of epileptic patients.",37813
results higher than accuracies quite are published. These earlier,These accuracies are quite higher than earlier results published.,37814
models used framework. train and the scans Then were annotations developed of to segmentation these,Then these scans and annotations were used to train segmentation models of the developed framework.,37815
attacks into Adversarial find models fool can images. misclassifying perturbations that,Adversarial attacks find perturbations that can fool models into misclassifying images.,37816
"simulate lighting can natural In and shadows. find addition, perturbations natural we phenomena harmonic like","In addition, we find harmonic perturbations can simulate natural phenomena like natural lighting and shadows.",37817
the neighbor image addresses of This nearest applied large-scale retrieval. approximate domain in search work,This work addresses approximate nearest neighbor search applied in the domain of large-scale image retrieval.,37818
off-line construction the we an the propose of framework testing Within search structures. efficient group,Within the group testing framework we propose an efficient off-line construction of the search structures.,37819
embedding combined methods. method existing be can The with naturally,The method can be naturally combined with existing embedding methods.,37820
We evaluate the databases. on multiple face anti-spoofing proposed method,We evaluate the proposed method on multiple face anti-spoofing databases.,37821
results due The spoof improvements promising to modeling. show our noise,The results show promising improvements due to our spoof noise modeling.,37822
Deep many achievement in has remarkable learning made fields.,Deep learning has made remarkable achievement in many fields.,37823
"usually of the amount of parameters demands networks a neural However, data. large labeled learning","However, learning the parameters of neural networks usually demands a large amount of labeled data.",37824
called few-shot task is learning. This specific,This specific task is called few-shot learning.,37825
The the measurement of scatter. volume of for class used is simplex the,The volume of the simplex is used for the measurement of class scatter.,37826
learning. verify effectiveness miniImageNet simplex few-shot the of algorithm Omniglot our on Experiments on and,Experiments on Omniglot and miniImageNet verify the effectiveness of our simplex algorithm on few-shot learning.,37827
"variability this of examine the across objects different visual paper, i.e. ad categories, we In","In this paper, we examine the visual variability of objects across different ad categories, i.e.",37828
persuasive. causes visually advertisement an be to what,what causes an advertisement to be visually persuasive.,37829
properties are these derived. filters of The,The properties of these filters are derived.,37830
to their we grouped be particular frequency their outputs response. can show according that In,In particular we show that their outputs can be grouped according to their frequency response.,37831
understanding. an surveillance camera task is scene important for network a tracking Target in and,Target tracking in a camera network is an important task for surveillance and scene understanding.,37832
views illumination and task variation challenging due disjoint in cameras. different is to The,The task is challenging due to disjoint views and illumination variation in different cameras.,37833
"this In features. were proposed using graph-based many direction, methods appearance-based","In this direction, many graph-based methods were proposed using appearance-based features.",37834
"illumination variation However, in the FOVs. high appearance information camera fades different the with","However, the appearance information fades with high illumination variation in the different camera FOVs.",37835
will network the that policy topology. learns camera We the show,We will show that the policy learns the camera network topology.,37836
the on NLPR of the dataset. proposed the MCT We performance method demonstrate,We demonstrate the performance of the proposed method on the NLPR MCT dataset.,37837
"advantages and of discuss we Finally, the various disadvantages schemes.","Finally, we discuss advantages and disadvantages of the various schemes.",37838
mesh methods Markov the over as State-of-the-art Random model a Field facets. labeling,State-of-the-art methods model mesh labeling as a Markov Random Field over the facets.,37839
compared the our compared We Matlab-toolbox Intlab. with results have,We have compared compared our results with the Matlab-toolbox Intlab.,37840
have producing narrower superior that intervals. our is shown experiments method Numerical,Numerical experiments have shown that our method is superior producing narrower intervals.,37841
"of have principles, a been alternative supplemented DP the number postulates, ways. in These","These principles, the DP postulates, have been supplemented in a number of alternative ways.",37842
"criticised position unacceptably this has However, as recently been being strong.","However, this position has recently been criticised as being unacceptably strong.",37843
privacy-preserving data variables These are mining for latent applications. features as proposed,These latent variables are proposed as features for privacy-preserving data mining applications.,37844
"of the we problem paper, domain In this consider adaptation.","In this paper, we consider the problem of domain adaptation.",37845
"In using on results and data. simulations real-measured addition, demonstrate experimental benefits these we","In addition, we demonstrate these benefits using experimental results on simulations and real-measured data.",37846
make for on explore option two based contributions. methods We and algorithmic discovery variational inference,We explore methods for option discovery based on variational inference and make two algorithmic contributions.,37847
in representations task action frame in sequences role the play important recognition. Spatio-temporal an of,Spatio-temporal representations in frame sequences play an important role in the task of action recognition.,37848
"expensive flow) However, requires framework. and an has two-stream (RGB optical computational and it cost","However, it has an expensive computational cost and requires two-stream (RGB and optical flow) framework.",37849
"downstream performance improvements. addition, deep integrated SSNs into easily resulting can in networks In be","In addition, SSNs can be easily integrated into downstream deep networks resulting in performance improvements.",37850
"artifacts to approach a these of address propose sources two-stage, all deep-learning of simultaneously. We","We propose a two-stage, deep-learning approach to address all of these sources of artifacts simultaneously.",37851
"their a systems size, challenge. laser-based communication integrating high-throughput however, remains with Due to them","Due to their size, however, integrating them with high-throughput laser-based communication systems remains a challenge.",37852
method proposed demonstrate experimental state-of-the-art the outperforms that matching our graph results The algorithms.,The experimental results demonstrate that our proposed method outperforms the state-of-the-art graph matching algorithms.,37853
signature technique verification considers a handwritten biometric. authentication that is an Signature as,Signature verification is an authentication technique that considers handwritten signature as a biometric.,37854
to dimension the Furthermore reduction is feature vectors. technique applied large,Furthermore dimension reduction technique is applied to the large feature vectors.,37855
system is the support and vector classifier model using machine evaluated trained algorithm. (SVM) A,A system model is trained and evaluated using the support vector machine (SVM) classifier algorithm.,37856
"introduced is of optical EEG. In flow variant information represent the to addition,","In addition, optical flow is introduced to represent the variant information of EEG.",37857
generated available. directly The proportional the power is wind to speed,The power generated is directly proportional to the wind speed available.,37858
been applied Neural core have to Convolutional Networks the age-related (CNN) research as framework.,Convolutional Neural Networks (CNN) have been applied to age-related research as the core framework.,37859
CNN Modern are based based implementations. algebra typically point linear floating on,Modern CNN are typically based on floating point linear algebra based implementations.,37860
constrained in environments. important power particularly is compute This,This is particularly important in power constrained compute environments.,37861
"formulate Then, we representation quantitatively hardware between data the and efficiency. relationship","Then, we quantitatively formulate the relationship between data representation and hardware efficiency.",37862
that are the It models responses same. of appeared the almost both,It appeared that the responses of both models are almost the same.,37863
"this to we issues. present a address In model novel these paper,","In this paper, we present a novel model to address these issues.",37864
"gaze the and particular, handles represents In model explicitly gaze direction out-of-frame our targets.","In particular, our model explicitly represents the gaze direction and handles out-of-frame gaze targets.",37865
different multi-task using leverage a datasets three We approach. learning,We leverage three different datasets using a multi-task learning approach.,37866
methods the count use of to learn regression-based interest. explicitly optimization where State-of-the-art they objects,State-of-the-art methods use regression-based optimization where they explicitly learn to count the objects of interest.,37867
"capture linearity, Due non-linear they can extreme and to not expressions. this deformations","Due to this linearity, they can not capture extreme deformations and non-linear expressions.",37868
localization activity literature the in Most methods annotation from suffer frame-wise burden requirement. of the,Most activity localization methods in the literature suffer from the burden of frame-wise annotation requirement.,37869
a may labels effort. labeling reducing such from weak solution be manual potential towards Learning,Learning from weak labels may be a potential solution towards reducing such manual labeling effort.,37870
basic signal. the frequently set smallest Noise on limit the level,Noise level frequently set the basic limit on the smallest signal.,37871
always RF technology. in Low play noise amplifiers role a significant,Low noise amplifiers always play a significant role in RF technology.,37872
with The been the single-gate LNA. has done comparison,The comparison has been done with the single-gate LNA.,37873
understanding deep networks computations. features in their can neural (DNNs) help Visualizing,Visualizing features in deep neural networks (DNNs) can help understanding their computations.,37874
are statistical widely among most and used simplest Permutation tests the tools.,Permutation tests are among the simplest and most widely used statistical tools.,37875
be p-values computed Their permutations. sampling by a can straightforward of,Their p-values can be computed by a straightforward sampling of permutations.,37876
The asymptotically than algorithm practically sampling. is straightforward and faster,The algorithm is practically and asymptotically faster than straightforward sampling.,37877
"we the tackle In of work, counting images. crowd this in problem","In this work, we tackle the problem of crowd counting in images.",37878
based density Convolutional to estimation (CNN) We Neural a this approach problem. Network present solve,We present a Convolutional Neural Network (CNN) based density estimation approach to solve this problem.,37879
high resolution in is map go a challenging task. one density a Predicting,Predicting a high resolution density map in one go is a challenging task.,37880
We present matting. the (GAN) first natural image adversarial for generative network,We present the first generative adversarial network (GAN) for natural image matting.,37881
"features to obtain performance. multiple better natural it's Therefore, together integrate to","Therefore, it's natural to integrate multiple features together to obtain better performance.",37882
its utilizing sample nearest neighbors. by k MRPE each reconstructs,MRPE reconstructs each sample by utilizing its k nearest neighbors.,37883
"segmentation offers quality). improved ratio, slightly quantitative results signal-to-noise Furthermore, also it (peak","Furthermore, it also offers slightly improved quantitative results (peak signal-to-noise ratio, segmentation quality).",37884
of number the Another advantage presented of is technique the hyperparameters. low,Another advantage of the presented technique is the low number of hyperparameters.,37885
effectiveness tasks results Superior methods the on demonstrate approach. of over state-of-the-art our both the,Superior results over the state-of-the-art methods on both tasks demonstrate the effectiveness of our approach.,37886
group by clustering a of The up such method a constructs non-local spatial-spectrum cubes. series,The method constructs such a group by clustering up a series of non-local spatial-spectrum cubes.,37887
"adopted NLCTF solve is the strategy the to Finally, model. split-Bregman","Finally, the split-Bregman strategy is adopted to solve the NLCTF model.",37888
The that outperforms method NLCTF results show competitors. other the the,The results show that the NLCTF method outperforms the other competitors.,37889
in networks. paper traffic deep for road a proposes approach complex flow prediction learning This,This paper proposes a deep learning approach for traffic flow prediction in complex road networks.,37890
"by a pooling replaces neural we that routing. network max dynamic Thus, propose capsules with","Thus, we propose a neural network with capsules that replaces max pooling by dynamic routing.",37891
gaze only human from a images natural task. Estimating eye is challenging,Estimating human gaze from natural eye images only is a challenging task.,37892
"problem. gaze ill-posed achieving is an highly Hence, accurate estimates","Hence, achieving highly accurate gaze estimates is an ill-posed problem.",37893
"to localization reduce and and the errors. strategies size, relative the instance absolute","the instance absolute and relative size, and strategies to reduce the localization errors.",37894
AVA for localization simple dataset. action baseline the on We introduce a,We introduce a simple baseline for action localization on the AVA dataset.,37895
contents a images. vision very the powerful from method for the understanding Computer is,Computer vision is a very powerful method for understanding the contents from the images.,37896
sheep. weight to measuring the enabled us of human has involvement in the minimize It,It has enabled us to minimize the human involvement in measuring weight of the sheep.,37897
starting model. (up-down) modular point is bottom-up Our re-implementation of the a top-down,Our starting point is a modular re-implementation of the bottom-up top-down (up-down) model.,37898
"performs however It dataset. to are particular that supervised fully ones, comparably a fine-tuned to","It performs comparably to fully supervised ones, that however are fine-tuned to a particular dataset.",37899
"well Our even hand, method, other datasets. transferred when performs on between the","Our method, on the other hand, performs well even when transferred between datasets.",37900
The estimate to task between visual person person images. re-identification similarities robustly requires,The person re-identification task requires to robustly estimate visual similarities between person images.,37901
"not similarity be hard of samples the estimation a As some accurate. might result,","As a result, the similarity estimation of some hard samples might not be accurate.",37902
relation features for similarity prediction. such be achieved probe-gallery estimation can by using updated Accurate,Accurate similarity estimation can be achieved by using such updated probe-gallery relation features for prediction.,37903
method validated The datasets. person effectiveness three on re-identification proposed public of our is,The effectiveness of our proposed method is validated on three public person re-identification datasets.,37904
is in people Automated counting of task. crowd a challenging images,Automated counting of people in crowd images is a challenging task.,37905
large in in appear major diversity the the difficulty from The stems crowds. people way,The major difficulty stems from the large diversity in the way people appear in crowds.,37906
"creating experts recursively, two every networks The split at are again division. child","The child networks are again split recursively, creating two experts at every division.",37907
major count model datasets. achieves on higher proposed The crowd accuracy,The proposed model achieves higher count accuracy on major crowd datasets.,37908
"the we analyse characteristics mined Further, specialties automatically our method. by of","Further, we analyse the characteristics of specialties mined automatically by our method.",37909
Electric is electric energy important meter in An component an power service.,An Electric power meter is an important component in electric energy service.,37910
"complained have meter. about the reading the In past, inaccurate consumers electric of many","In the past, many consumers have complained about reading inaccurate of the electric meter.",37911
of an This RFID development electrical reader. presents with research power the meter equipped,This research presents the development of an electrical power meter equipped with RFID reader.,37912
dominated are RCNN. region-based segmentation by such Object as and detection instance methods Mask,Object detection and instance segmentation are dominated by region-based methods such as Mask RCNN.,37913
a the We rate real propose world detection improves in conditions. algorithm new which,We propose a new algorithm which improves the detection rate in real world conditions.,37914
which has scheme speed. been added processing the tracking also A simple improves,A simple tracking scheme has also been added which improves the processing speed.,37915
proposed the algorithms outperformed performance. of real-time while the achieving state art algorithm The,The proposed algorithm outperformed the state of the art algorithms while achieving real-time performance.,37916
"(MAC) input. are billions operations intensive, compute multiply-accumulate CNNs of per requiring However,","However, CNNs are compute intensive, requiring billions of multiply-accumulate (MAC) operations per input.",37917
the inherent damages This correlations among the data. consequently,This consequently damages the inherent correlations among the data.,37918
"for affine Besides, is used practical enough not the applications. methods these in constraint flexible","Besides, the affine constraint used in these methods is not flexible enough for practical applications.",37919
literature We to references code. also relevant of provide and the most,We also provide references to most of the relevant literature and code.,37920
computational solutions their in presented complexity. performance and vary The,The solutions presented vary in their performance and computational complexity.,37921
performance simulations. is these compared of The evaluated via solutions and,The performance of these solutions is evaluated and compared via simulations.,37922
"work, sharp In make post-processing. boundaries is this produce to the without aim CNNs","In this work, the aim is to make CNNs produce sharp boundaries without post-processing.",37923
"which tackle bottom-up/top-down to an the adopts network we architecture propose end-to-end task. the Moreover,","Moreover, we propose an end-to-end network which adopts the bottom-up/top-down architecture to tackle the task.",37924
We results. following the on our find based closed-form solution,We find the closed-form solution based on our following results.,37925
same the The is neighboring eigenfunction points. the a of sign on low-frequency,The sign of a low-frequency eigenfunction is the same on the neighboring points.,37926
knowledge natural in discover trained further The networks scenes. visual are to applied,The trained networks are further applied to discover visual knowledge in natural scenes.,37927
"model then can and lightweight, sophisticated be This for editing. video used image","This model can then be used for lightweight, sophisticated video and image editing.",37928
our editing. for video We of show using framework also face examples,We also show examples of using our framework for video face editing.,37929
"each are the by corresponding availability such limited scans of However, modality. of techniques","However, such techniques are limited by the availability of corresponding scans of each modality.",37930
method classification. evaluated Alzheimer's of of efficacy is the in disease The context proposed the,The efficacy of the proposed method is evaluated in the context of Alzheimer's disease classification.,37931
"features residual attenuation. First, without is to network employed hierarchical explore adapted an","First, an adapted residual network is employed to explore hierarchical features without attenuation.",37932
usually works Previous video clips. video captioning on focus on,Previous works on video captioning usually focus on video clips.,37933
regions. some The Hungarian algorithm matching performed unmatched is permitting via,The matching is performed via Hungarian algorithm permitting some unmatched regions.,37934
The similarity context is proposed a clustering the set of employed shapes. of measure in,The proposed similarity measure is employed in the context of clustering a set of shapes.,37935
degradation common a is source. Image blur,Image blur is a common degradation source.,37936
"motion presence the In with of this focus work, super-resolution task the on we blur.","In this work, we focus on the super-resolution task with the presence of motion blur.",37937
very role image deblurring applications. multimedia important and Blind vision a plays many in,Blind image deblurring plays a very important role in many vision and multimedia applications.,37938
"is established framework In the above to work, issues. a this address collaborative learning","In this work, a collaborative learning framework is established to address the above issues.",37939
to similarities person images. robustly between measure re-identification aims Person,Person re-identification aims to robustly measure similarities between person images.,37940
variation The significant challenges poses and accurate viewing of re-identification. person angles person for,The significant variation of person poses and viewing angles challenges for accurate person re-identification.,37941
"depends dramatic ads climax particular, In changes sentiment. of in on structure the in","In the dramatic structure of ads in particular, climax depends on changes in sentiment.",37942
to of structure ads study video the dynamic We automatically. dedicate our understand,We dedicate our study to understand the dynamic structure of video ads automatically.,37943
use supervised to unsupervised both and predict We then methods the climax.,We then use both unsupervised and supervised methods to predict the climax.,37944
complex has problem six solutions (including We that the generically show ones).,We show that the problem generically has six solutions (including complex ones).,37945
"stable that experiments correct, is numerically robust. demonstrate and method The the","The experiments demonstrate that the method is correct, numerically stable and robust.",37946
modalities a different from task. challenging Learning is,Learning from different modalities is a challenging task.,37947
person. have exaggerations a facial features of of Caricature,Caricature have exaggerations of facial features of a person.,37948
the subbands. smoothing The of the of highest coefficients method is a on based,The method is based on a smoothing of the coefficients of the highest subbands.,37949
the methods favorably most Denoising results the of use moment. compare at to in,Denoising results compare favorably to the most of methods in use at the moment.,37950
"such Gram matrix, statistics, to these second-order used textures. In describe models, as are","In these models, second-order statistics, such as Gram matrix, are used to describe textures.",37951
"further Style where to apply We create our Neural Transfer, can styles. we mixed scheme","We further apply our scheme to Neural Style Transfer, where we can create mixed styles.",37952
experiments demonstrate results. our achieve method that The can state-of-the-art,The experiments demonstrate that our method can achieve state-of-the-art results.,37953
robots. paper aspect explores of This visual mobile odometry the for methods deep,This paper explores the aspect of deep visual odometry methods for mobile robots.,37954
odometry making successful. navigation this been in instrumental Visual has,Visual odometry has been instrumental in making this navigation successful.,37955
can accurate mobile in facilitate findings. map more A robots localization for and dense,A dense map in mobile robots can facilitate for localization and more accurate findings.,37956
proposed show benchmark datasets FGVC our efficacy the on of method. Experiments,Experiments on benchmark FGVC datasets show the efficacy of our proposed method.,37957
"is the greatly limited deep methods. metric performance of However, by sampling traditional learning","However, the performance of deep metric learning is greatly limited by traditional sampling methods.",37958
"Due the video emergence Generative exceptional witnessed Networks, has breakthroughs. to of synthesis Adversarial","Due to the emergence of Generative Adversarial Networks, video synthesis has witnessed exceptional breakthroughs.",37959
"methods proper However, in control dynamics representation videos. to lack the existing explicitly a","However, existing methods lack a proper representation to explicitly control the dynamics in videos.",37960
standardized A individual templates. central task data to volumes to computational map is,A central computational task is to map individual data volumes to standardized templates.,37961
the of diffeomorphic template. the subgroup acting group form stabilizer the on These,These form the stabilizer subgroup of the diffeomorphic group acting on the template.,37962
of provides estimates form. volume the reduced-variance This,This provides reduced-variance estimates of the volume form.,37963
derive the for variances the deformations. in bounds small We of limit Cramer-Rao,We derive Cramer-Rao bounds for the variances in the limit of small deformations.,37964
selected each public the detailed are The for provided use. annotations of categories for,The detailed annotations for each of the selected categories are provided for public use.,37965
"Specifically, learns field the generative i.e. that propose a network we correspondence to predict deep","Specifically, we propose a deep generative network that learns to predict the correspondence field i.e.",37966
map a image views camera between stereo in setting. disparity the calibrated two,the disparity map between two image views in a calibrated stereo camera setting.,37967
capability long been has inferences processes desired. regarding physical explainable making of The,The capability of making explainable inferences regarding physical processes has long been desired.,37968
fundamental motion. object One physical is process,One fundamental physical process is object motion.,37969
"However, object qualitative motion three-dimensional has been suitable theory no space. proposed there for in","However, there has been no suitable qualitative theory proposed for object motion in three-dimensional space.",37970
to objects. what causes action infer the We of these want movement,We want to infer what action causes the movement of these objects.,37971
multi-center studies acquisition a variety protocols. necessarily wide of initiatives are Modern with neuroimaging,Modern neuroimaging studies are necessarily multi-center initiatives with a wide variety of acquisition protocols.,37972
paradigm the of developments restoration. learning image revolutionized have in deep Recent,Recent developments in deep learning have revolutionized the paradigm of image restoration.,37973
fully hence differentiable end-to-end trained. proposed The is can pipeline be and restoration,The proposed restoration pipeline is fully differentiable and hence can be end-to-end trained.,37974
given results Experimental less prone is to our data. training demonstrate work that small overfitting,Experimental results demonstrate that our work is less prone to overfitting given small training data.,37975
is biomarker linked Cell function. that an shape important directly is to cell,Cell shape is an important biomarker that is directly linked to cell function.,37976
"temporal much less cell cell understood. morphodynamics, is namely fluctuation the shape However, of","However, cell morphodynamics, namely the temporal fluctuation of cell shape is much less understood.",37977
"cell evolution a of temporal into As mapped is result, morphodynamics morphological phenotypes.","As a result, cell morphodynamics is mapped into temporal evolution of morphological phenotypes.",37978
Rho-signaling enhances and that the manipulating transitions. find phenotype We morphodynamics,We find that manipulating Rho-signaling enhances the morphodynamics and phenotype transitions.,37979
vehicles. perception and control of networks have components become neural Deep key self-driving,Deep neural perception and control networks have become key components of self-driving vehicles.,37980
consists a introspective approach explanations two parts. We of to new propose which,We propose a new approach to introspective explanations which consists of two parts.,37981
The controller's that the output. image identifies attention regions influence network's potentially,The controller's attention identifies image regions that potentially influence the network's output.,37982
"model textual actions. of video-to-text we explanations Second, to produce use attention-based model an","Second, we use an attention-based video-to-text model to produce textual explanations of model actions.",37983
"weak-alignment. strong- attention explore alignment, two to We and approaches","We explore two approaches to attention alignment, strong- and weak-alignment.",37984
"However, and pleasing required. a explanation is visually class-discriminative more","However, a more class-discriminative and visually pleasing explanation is required.",37985
method multi-scale incorporate difference this that prediction is model-agnostic. We a regional concept into,We incorporate this regional multi-scale concept into a prediction difference method that is model-agnostic.,37986
regression and suppression rely Modern objects. detectors object non-maximum bounding CNN-based localize box on to,Modern CNN-based object detectors rely on bounding box regression and non-maximum suppression to localize objects.,37987
"labels reflect classification class While confidence absent. the probabilities is confidence, for localization naturally","While the probabilities for class labels naturally reflect classification confidence, localization confidence is absent.",37988
makes during regression or suppressed bounding iterative during even NMS. boxes localized properly This degenerate,This makes properly localized bounding boxes degenerate during iterative regression or even suppressed during NMS.,37989
A is to of a kernel used separation. between better version class MMC obtain,A kernel version of MMC is used to obtain a better between class separation.,37990
on sidekick introduce learning of policy capitalize imbalance to We this observability.,We introduce sidekick policy learning to capitalize on this imbalance of observability.,37991
"of visualization resulting policies, interpretation a develop To technique. support the policy novel we also","To support interpretation of the resulting policies, we also develop a novel policy visualization technique.",37992
images applications. binocular stereo a Disparity finds estimation range for of wide,Disparity estimation for binocular stereo images finds a wide range of applications.,37993
and work supervised semantic in The both cues unsupervised manners. well,The semantic cues work well in both unsupervised and supervised manners.,37994
"scheme, data generate active In users sparsity. transmit packets the time-slotted random and RSL-MUD with","In the time-slotted RSL-MUD scheme, active users generate and transmit data packets with random sparsity.",37995
location vision not require does and sensors than additional SafeDrive data. other,SafeDrive does not require additional sensors other than vision and location data.,37996
non-invasive technique diseases. is used The the detection electrocardiogram a (ECG) for cardiac widely of,The electrocardiogram (ECG) is a non-invasive technique widely used for the detection of cardiac diseases.,37997
is face This super-resolution. on paper image and,This paper is on image and face super-resolution.,37998
"mean both In squared error network's variance. has our the cases, lowest","In both cases, our network's mean squared error has the lowest variance.",37999
validated using and real The solution is event synthetic proposed sequences.,The proposed solution is validated using synthetic and real event sequences.,38000
studied promising to leverage model shrink and pruning size. Network widely is the a,Network pruning is a promising and widely studied leverage to shrink the model size.,38001
estimation Uncertainty recently. in becomes more deep important learning,Uncertainty estimation in deep learning becomes more important recently.,38002
theories domain first existing tasks. re-ID classification to unsupervised We extend adaptive,We first extend existing unsupervised domain adaptive classification theories to re-ID tasks.,38003
"self-training novel a tasks for To adaptive is unsupervised domain re-ID them, optimize scheme proposed.","To optimize them, a novel self-training scheme for unsupervised domain adaptive re-ID tasks is proposed.",38004
computer inverse vision. a in typical image rain removal is Single problem,Single image rain removal is a typical inverse problem in computer vision.,38005
"To toward paradigm this paper learning this this proposes a firstly task. semi-supervised issue,","To this issue, this paper firstly proposes a semi-supervised learning paradigm toward this task.",38006
is biomarker an important count for Mitosis various of prognosis cancers.,Mitosis count is an important biomarker for prognosis of various cancers.,38007
"This subjective. time-consuming, and very is tedious task","This task is very time-consuming, tedious and subjective.",38008
made Automated mitosis advances detection methods recent great years. in have,Automated mitosis detection methods have made great advances in recent years.,38009
"of labeling of a regions number However, these large exhaustive selected methods require interest. of","However, these methods require exhaustive labeling of a large number of selected regions of interest.",38010
is task because needed accurate annotations. are expert for reliable This pathologists expensive very and,This task is very expensive because expert pathologists are needed for reliable and accurate annotations.,38011
platforms used coverage and provide cellular capacity. Networked can to flying be,Networked flying platforms can be used to provide cellular coverage and capacity.,38012
features networks their influence the introduces of This that specific paper resilience. airborne,This paper introduces the specific features of airborne networks that influence their resilience.,38013
paper the problem challenging image and fingerprint inpainting. of This denoising studies,This paper studies the challenging problem of fingerprint image denoising and inpainting.,38014
the shape in image characteristics. having with the components corresponding similar,with the corresponding components in the image having similar shape characteristics.,38015
propose elegant kinematics relative We solutions the to network. closed the estimate recursively of form,We propose elegant closed form solutions to recursively estimate the relative kinematics of the network.,38016
conventional demonstrate significant datasets over ex- different Our metrics. two on motion human improvements periments,Our ex- periments on two different datasets demonstrate significant improvements over conventional human motion metrics.,38017
"the time-based effective and a challenge. are authoring remains that viewer for However, media personalized","However, authoring time-based media that are effective and personalized for the viewer remains a challenge.",38018
"introduce for time-based AniCode, and authoring consuming novel a media. We framework","We introduce AniCode, a novel framework for authoring and consuming time-based media.",38019
"AniCode in the visual personalized consumer's the Importantly, video presents context.","Importantly, AniCode presents the video personalized in the consumer's visual context.",38020
to low easy to designed and Our system is use. cost be,Our system is designed to be low cost and easy to use.,38021
user use system effectiveness with evaluate the and ease a We study. of of our,We evaluate the ease of use and effectiveness of our system with a user study.,38022
"few labeled contain benchmark leaf only segmentation images. Currently, a datasets for the training hundred","Currently, the benchmark datasets for leaf segmentation contain only a few hundred labeled training images.",38023
of on problem. based the proposed the The parameter is estimation Rao-Blackwellization technique,The proposed technique is based on the Rao-Blackwellization of the parameter estimation problem.,38024
successfully technique structures. covariance The proposed error model different under performs,The proposed technique performs successfully under different model error covariance structures.,38025
on coding is based along with over-complete dictionary. approach second learned a sparse The,The second approach is based on sparse coding along with a learned over-complete dictionary.,38026
effect We Convolutional of injecting equivariance Networks. Neural local into the scale study,We study the effect of injecting local scale equivariance into Convolutional Neural Networks.,38027
applying is convolutional each multiple scales. filter at done This by,This is done by applying each convolutional filter at multiple scales.,38028
intermediate allows to scale This representations locally be all the equivariant.,This allows all the intermediate representations to be locally scale equivariant.,38029
"mitigate DCF-based these we tracking challenges, new propose a method. To","To mitigate these challenges, we propose a new DCF-based tracking method.",38030
"structured multi-channel filers. to spatial constraints specifically, apply sparsity More we","More specifically, we apply structured spatial sparsity constraints to multi-channel filers.",38031
"be approximated can spatial by filters learning lasso process the Consequently, regularisation. the of","Consequently, the process of learning spatial filters can be approximated by the lasso regularisation.",38032
results the method the proposed of approaches. experimental state-of-the-art over superiority demonstrate the The,The experimental results demonstrate the superiority of the proposed method over the state-of-the-art approaches.,38033
"been attracting classification these reasons, has attention. For image multi-label aerial increasing","For these reasons, aerial image multi-label classification has been attracting increasing attention.",38034
because always collaboratively. parts work our body This is,This is because our body parts always work collaboratively.,38035
source and make available. publicly our model will codes We,We will make our model and source codes publicly available.,38036
security. a the study maritime of effectiveness on We framework in our case discuss,We discuss the effectiveness of our framework on a case study in maritime security.,38037
satellite Change multitemporal an sensing. detection emerging heterogeneous is in remote in topic images,Change detection in heterogeneous multitemporal satellite images is an emerging topic in remote sensing.,38038
for optimal connectivity simple rules are the Such unlikely architecture to yield given the problem.,Such simple connectivity rules are unlikely to yield the optimal architecture for the given problem.,38039
"it leads savings significant in to of number Furthermore, certain parameters. in settings","Furthermore, in certain settings it leads to significant savings in number of parameters.",38040
segmentations. demonstrated network abdominal performance based on deep approaches multi-organ Recent promising have,Recent deep network based approaches have demonstrated promising performance on abdominal multi-organ segmentations.,38041
priors. propose based atlas weighting two-stage schemes spatial with a approach on and auto-context We,We propose a two-stage approach with weighting schemes based on auto-context and spatial atlas priors.,38042
"especially methods Learning-based suffer \begin{abstract} clean deficiency from segmentation. biomedical in annotations, a of","\begin{abstract} Learning-based methods suffer from a deficiency of clean annotations, especially in biomedical segmentation.",38043
in from image  fundamental  Depth estimation computer vision. is problem single a a,  Depth estimation from a single image is a fundamental problem in computer vision.,38044
"depth We i.e. benchmarks the estimation, experimented popular CSPN proposed for over two","We experimented the proposed CSPN over two popular benchmarks for depth estimation, i.e.",38045
to be schemes. may any trellis based coding It applied,It may be applied to any trellis based coding schemes.,38046
"computational components. cognition, science Cognitive models human performance has computational decomposing of developed task into","Cognitive science has developed computational models of human cognition, decomposing task performance into computational components.",38047
"in algorithms short human not still are its neurobiology. of intelligence and However, fall grounded","However, its algorithms still fall short of human intelligence and are not grounded in neurobiology.",38048
implement has brain interacting how investigated Computational can neuroscience of component computation. neurons functions,Computational neuroscience has investigated how interacting neurons can implement component functions of brain computation.,38049
"yield test insight only models. experiments will when theoretical However, brain-computational to employed","However, experiments will yield theoretical insight only when employed to test brain-computational models.",38050
It of assemble is of puzzle time computation. the to pieces the brain,It is time to assemble the pieces of the puzzle of brain computation.,38051
Exten- experiments (IIIT-D datasets. (E-PRIP) composite performed sive are semi-forensic) on and semi-forensic,Exten- sive experiments are performed on composite (E-PRIP) and semi-forensic (IIIT-D semi-forensic) datasets.,38052
Our validity without mask. proposal learns of an sparse features efficiently additional the need,Our proposal efficiently learns sparse features without the need of an additional validity mask.,38053
network to We ensure varying show sparsities. to how robustness input,We show how to ensure network robustness to varying input sparsities.,38054
"instrument pixel-wise are forest using bounding potential detected First, segmentation. random boxes a","First, potential instrument bounding boxes are detected using a pixel-wise random forest segmentation.",38055
boxes then of using bounding of a cascade these Each classified forest. random is,Each of these bounding boxes is then classified using a cascade of random forest.,38056
videos different types of approach from evaluated five different procedures. on our two We,We evaluated our approach on five different videos from two different types of procedures.,38057
of Weather the conditions systems. transportation disrupt functioning proper often,Weather conditions often disrupt the proper functioning of transportation systems.,38058
incremental and in These scope. solutions limited resulted cost have,These solutions have resulted in incremental cost and limited scope.,38059
task methods recognition. Our proposed saliency allows high-level of pipeline the object for evaluate to,Our proposed pipeline allows to evaluate saliency methods for the high-level task of object recognition.,38060
and assessment between runtime control. a results shows design Experimental gap phase small,Experimental results shows a small gap between design phase assessment and runtime control.,38061
temporal influences the Such fusion. of performance dramatically offset sensor,Such temporal offset dramatically influences the performance of sensor fusion.,38062
"project, integrated code VINS-Mono. is The of into temporal calibration public our source","The source code of temporal calibration is integrated into our public project, VINS-Mono.",38063
"simple final is partition process a Finally, get results instance to during employed inference.","Finally, a simple instance partition process is employed to get final results during inference.",38064
on all and state-of-the-art dataset We experiments PASCAL-Person-Part methods. conducted our PGN outperforms,We conducted experiments on PASCAL-Person-Part dataset and our PGN outperforms all state-of-the-art methods.,38065
neural with We proposal soft the layers. use architectures examine of coupled convolutional network,We examine the use of convolutional neural network architectures coupled with soft proposal layers.,38066
resulting The tasks. simultaneously localisation (classification) region and detection network performs anatomical,The resulting network simultaneously performs anatomical region detection (classification) and localisation tasks.,38067
attention of the network describing a class. map particular a generate the We proposal for,We generate a proposal map describing the attention of the network for a particular class.,38068
"to and limbs, regions: placenta. spine, Labels anatomical thorax, head, six correspond abdomen,","Labels correspond to six anatomical regions: head, spine, thorax, abdomen, limbs, and placenta.",38069
decode-and-forward reduce (IDF) power Incremental transmit to is consumption. relying exploited the,Incremental decode-and-forward (IDF) relying is exploited to reduce the transmit power consumption.,38070
with The assumed channel to PLC experience fading noise. log-normal impulsive is,The PLC channel is assumed to experience log-normal fading with impulsive noise.,38071
"is algorithms specific video. on the our to applied Consequently, image approach original processing agnostic","Consequently, our approach is agnostic to specific image processing algorithms applied on the original video.",38072
segmentation developed overcome convolutional the fully to Methods: problem. volumetric was A image network,Methods: A fully convolutional network was developed to overcome the volumetric image segmentation problem.,38073
"internally The shape, features results. for deep merged were the segmentation and discriminative contour,","The discriminative contour, shape, and deep features were internally merged for the segmentation results.",38074
an performed evaluation quantitative proposed network eight-fold through of cross-validation. the was The,The quantitative evaluation of the proposed network was performed through an eight-fold cross-validation.,38075
works services. The approach on composite also,The approach also works on composite services.,38076
report We results. translation context of implement text the interesting it and in,We implement it in the context of text translation and report interesting results.,38077
cells of crowded in numerous in are environments. There scenarios which populations migrate,There are numerous scenarios in which populations of cells migrate in crowded environments.,38078
"development. include embryo and examples growth cancer healing, Typical wound","Typical examples include wound healing, cancer growth and embryo development.",38079
"cell adhesion, and interactions, pushing include pulling. repulsion, These volume signalling, excluded","These include excluded volume interactions, adhesion, repulsion, cell signalling, pushing and pulling.",38080
appropriate employ interactions. ABM that is the This when it representing pulling important cell-cell hints,This hints that it is important employ the appropriate ABM when representing pulling cell-cell interactions.,38081
"work domain in adaptation has i.e. the context Prior shifts, of domain examined stationary","Prior work has examined domain adaptation in the context of stationary domain shifts, i.e.",38082
"become data effectively instance, in a function time. dataset of streaming scenario, a For statistics","For instance, in a streaming data scenario, dataset statistics effectively become a function of time.",38083
The data sequentially adapted batches. over incoming streaming is model unsupervised,The model is adapted sequentially over incoming unsupervised streaming data batches.,38084
This annotated improvements need enables for additionally the any batches several over without data.,This enables improvements over several batches without the need for any additionally annotated data.,38085
accuracy Neglecting and cameras degrades of effects severely visual for odometry (VO) the robustness. rolling-shutter,Neglecting the effects of rolling-shutter cameras for visual odometry (VO) severely degrades accuracy and robustness.,38086
the at the a estimate We optimization. velocity each and prior constant-velocity keyframe impose for,We estimate the velocity at each keyframe and impose a constant-velocity prior for the optimization.,38087
"In VO we accurate near this a method. direct obtain way, real-time,","In this way, we obtain a near real-time, accurate direct VO method.",38088
improved VO. Our on global-shutter achieves sequences over challenging approach results state-of-the-art rolling-shutter,Our approach achieves improved results on challenging rolling-shutter sequences over state-of-the-art global-shutter VO.,38089
segmentation real shown of have images segmentation accurate that performance is results Additional achieved.,Additional segmentation results of real images have shown that accurate segmentation performance is achieved.,38090
activity-class in video-level using model labels. a is trained weakly raw supervised Our setting,Our model is trained in a weakly supervised setting using raw video-level activity-class labels.,38091
We discuss ablation choices. regarding extensive analysis design an also the,We also discuss an extensive ablation analysis regarding the design choices.,38092
in signal Recovering presence noise challenging the of degraded problem. is nonlinearly a,Recovering nonlinearly degraded signal in the presence of noise is a challenging problem.,38093
semi-algebraic class broad to The composite function of the resulting cost functions. belongs,The resulting composite cost function belongs to the broad class of semi-algebraic functions.,38094
that Extensive performs favorably algorithms. against the state-of-the-art experiments our demonstrate method,Extensive experiments demonstrate that our method performs favorably against the state-of-the-art algorithms.,38095
the procedure and labeling result. further subsequent segmentation with multi-label A graph refines cuts post-process,A subsequent post-process procedure with multi-label graph cuts further refines the segmentation and labeling result.,38096
on We our two validate sketch proposed datasets. method,We validate our proposed method on two sketch datasets.,38097
of quality texture images. show We that generation the loss alone allows perceptually high,We show that texture loss alone allows the generation of perceptually high quality images.,38098
long-term memory? is What the physiological of basis,What is the physiological basis of long-term memory?,38099
efficacy in in The attributes view acquisition. synaptic memory to changes neuroscience prevailing,The prevailing view in neuroscience attributes changes in synaptic efficacy to memory acquisition.,38100
view to stable This memories correspond implies connectivity stable that patterns.,This view implies that stable memories correspond to stable connectivity patterns.,38101
"significant, activity-independent However, of to body an evidence increasing synaptic in dynamics strengths. experimental points","However, an increasing body of experimental evidence points to significant, activity-independent dynamics in synaptic strengths.",38102
"memories more plasticity are imaginary-coded to noise Specifically, resilient and real-coded than their counterparts. homeostatic","Specifically, imaginary-coded memories are more resilient to noise and homeostatic plasticity than their real-coded counterparts.",38103
"on complex In using networks. hierarchical work, neural natural more images focus this unsupervised we","In this work, we focus on more complex natural images using unsupervised hierarchical neural networks.",38104
problematic when in conditions. segmentation imaging outdoor is there of Semantic are variations scenes,Semantic segmentation of outdoor scenes is problematic when there are variations in imaging conditions.,38105
known is (reflectance) of that effects. albedo illumination to is invariant It all kinds,It is known that albedo (reflectance) is invariant to all kinds of illumination effects.,38106
"images task be can Thus, segmentation reflectance using favorable. semantic for","Thus, using reflectance images for semantic segmentation task can be favorable.",38107
We of two those analyze the addressing gains jointly. problems,We analyze the gains of addressing those two problems jointly.,38108
"single tasks. architectures cascade and new for segmentation-for-intrinsic as CNN proposed are intrinsic-for-segmentation Moreover,","Moreover, new cascade CNN architectures for intrinsic-for-segmentation and segmentation-for-intrinsic are proposed as single tasks.",38109
"detail and these going statistical mathematical Here, the by through framework. we steps","Here, we detail these steps by going through the mathematical and statistical framework.",38110
understanding be We the can of explain it applied \beta-arrestin-dependent networks. to signaling how,We explain how it can be applied to the understanding of \beta-arrestin-dependent signaling networks.,38111
graph matches. cost Edges the of these the of represent,Edges of the graph represent the cost of these matches.,38112
is essential protocols for networks. and evaluation Channel in design modeling numerous performance of vehicular,Channel modeling is essential for design and performance evaluation of numerous protocols in vehicular networks.,38113
An performance ultra-wideband to optical in model estimate presented. analytical transmission networks nonlinear is,An analytical model to estimate nonlinear performance in ultra-wideband optical transmission networks is presented.,38114
to Sequential previously search neural for its architectures classification. capabilities demonstrated has find image state-of-the-art,Sequential search has previously demonstrated its capabilities to find state-of-the-art neural architectures for image classification.,38115
"computational remains even computational under unreachable high, its modest cost settings. However,","However, its computational cost remains high, even unreachable under modest computational settings.",38116
weight-sharing Affording SMBO with problem. this alleviates,Affording SMBO with weight-sharing alleviates this problem.,38117
rank used is sampled to prediction directly the neural architectures. This,This prediction is directly used to rank the sampled neural architectures.,38118
"is SEM problem, solve an mitigation this To described. scheme","To solve this problem, an SEM mitigation scheme is described.",38119
"thus However, pilot accurate estimations, better can performance. more achieve the system","However, the pilot system can achieve more accurate estimations, thus better performance.",38120
"intermediate an view, introduced large is view, gap. First, i.e., the homography to bridge","First, an intermediate view, i.e., homography view, is introduced to bridge the large gap.",38121
discussions its cases. studies some in robustness and reliability show challenging Ablation and,Ablation studies and discussions show its reliability and robustness in some challenging cases.,38122
brain lesion are for learning Deep used methods actively segmentation.,Deep learning methods are actively used for brain lesion segmentation.,38123
patch-sampling We new propose small increase lesions. procedure for performance a to network,We propose a new patch-sampling procedure to increase network performance for small lesions.,38124
show both solid cases. results The experimental in improvements,The experimental results show solid improvements in both cases.,38125
"and feature has scalability learning engineering on issues. machine directly graphs, needs However,","However, machine learning directly on graphs, needs feature engineering and has scalability issues.",38126
proposes on based novel This image. paper image single method fusion a pseudo a multi-exposure,This paper proposes a novel pseudo multi-exposure image fusion method based on a single image.,38127
proposed enables The method pseudo from images to produce multi-exposure single us image. a,The proposed method enables us to produce pseudo multi-exposure images from a single image.,38128
conventional multi-exposure proposed to multi-exposure images. image Most methods applicable the are fusion also of,Most of conventional multi-exposure image fusion methods are also applicable to the proposed multi-exposure images.,38129
synthetic and Experiments the proposed both promising on performance the of validate architecture real datasets.,Experiments validate the promising performance of the proposed architecture on both synthetic and real datasets.,38130
method deconvolution we (FBD). call blind Hence the focused,Hence we call the method focused blind deconvolution (FBD).,38131
relaxed constraints The progress. iterations the focusing are as,The focusing constraints are relaxed as the iterations progress.,38132
"Learning their classification more classification including power elaborate competes methods. Deep approaches Furthermore, with","Furthermore, their classification power competes with more elaborate classification approaches including Deep Learning methods.",38133
"to methods. no evaluate However, the and endoscopy aforementioned compare objective criteria are there","However, there are no objective criteria to evaluate and compare the aforementioned endoscopy methods.",38134
consistent studies. is the observation clinical This with,This observation is consistent with the clinical studies.,38135
"the effectiveness of proposed demonstrated. is the criterion Therefore,","Therefore, the effectiveness of the proposed criterion is demonstrated.",38136
"features visual object-agnostic Therefore, models. we learning relationship generalizable more in are interested for","Therefore, we are interested in learning object-agnostic visual features for more generalizable relationship models.",38137
"propose the novel strategy. \texttt{Shuffle-Then-Assemble} To pre-training alleviate a we bias,","To alleviate the bias, we propose a novel \texttt{Shuffle-Then-Assemble} pre-training strategy.",38138
will ecological to nonlinear biological parameters models an Identifiability consider. of be issue in,Identifiability of biological parameters in nonlinear ecological models will be an issue to consider.,38139
reconstruction a We propose (cGANs). direct conditional based on adversarial fingerprint model networks latent generative,We propose a direct latent fingerprint reconstruction model based on conditional generative adversarial networks (cGANs).,38140
been solver top ASP the of WASP. on has The approach implemented,The approach has been implemented on top of the ASP solver WASP.,38141
(TPLP). and of Under consideration Programming Logic Practice in Theory,Under consideration in Theory and Practice of Logic Programming (TPLP).,38142
Here discuss nature the of models network neuroscience. in we,Here we discuss the nature of network models in neuroscience.,38143
base artificial an Knowledge important (KB) intelligence. is  aspect in ,  Knowledge base (KB) is an important aspect in artificial intelligence.,38144
bases and proposed approaches that efficiently results cleanse the Experimental effectively. knowledge the demonstrate could,Experimental results demonstrate that the proposed approaches could cleanse the knowledge bases efficiently and effectively.,38145
"network architectures, attacks these weak transferability neural Additionally, e.g. across demonstrate cross-network","Additionally, these attacks demonstrate weak cross-network transferability across neural network architectures, e.g.",38146
the semantic of addresses dense problem work This scene under understanding fog.,This work addresses the problem of semantic scene understanding under dense fog.,38147
methods outdoor recognition weather conditions to Extending crucial such fog as for applications. is adverse,Extending recognition methods to adverse weather conditions such as fog is crucial for outdoor applications.,38148
The code and publicly are datasets available.,The datasets and code are publicly available.,38149
for to correct it difficult These radiologists render diagnostic artifacts make decisions.,These artifacts make it difficult for radiologists to render correct diagnostic decisions.,38150
videos This visual an in from learning unsupervised necessitates features setting. of,This necessitates learning of visual features from videos in an unsupervised setting.,38151
The following framework. also step in charts is unnecessary our packing,The following charts packing step is also unnecessary in our framework.,38152
more map efficient. technique the and practical normal makes is and robust Our method,Our method is practical and makes the normal map technique more robust and efficient.,38153
individual arrangement. their appears parts Structure typically of the and form in,Structure typically appears in the form of individual parts and their arrangement.,38154
incomplete information scans. structural templates and onto transfer objects new us the allow to The,The templates allow us to transfer the structural information onto new objects and incomplete scans.,38155
of other method state-of-the-art our showcase the We by comparing approaches. effectiveness it to,We showcase the effectiveness of our method by comparing it to other state-of-the-art approaches.,38156
Diverse utilized improve been information the auxiliary feature has visual learning. to,Diverse auxiliary information has been utilized to improve the visual feature learning.,38157
"birth. induced within the When crest pups at die neural lineage,","When induced within the neural crest lineage, pups die at birth.",38158
derivatives crest appear the unaffected. Neural body in,Neural crest derivatives in the body appear unaffected.,38159
propagation may Its unmixing of in significant occurrence the the process. estimation to errors lead,Its occurrence may lead to the propagation of significant estimation errors in the unmixing process.,38160
a from present human avatar method novel for monocular We creation detail-preserving video. high,We present a novel method for high detail-preserving human avatar creation from monocular video.,38161
"garment wrinkles, Our avatars clothes a high-resolution and texture. with face, hairstyle, feature natural","Our avatars feature a natural face, hairstyle, clothes with garment wrinkles, and high-resolution texture.",38162
the Numerous of results versatility robustness show method. our and,Numerous results show the robustness and versatility of our method.,38163
electroencephalography of important systems. the The is brain-computer component classifier based most interface,The electroencephalography classifier is the most important component of brain-computer interface based systems.,38164
it. of hindering are improvement problems the two There major,There are two major problems hindering the improvement of it.,38165
"multimodal fully do methods not traditional First, exploit information.","First, traditional methods do not fully exploit multimodal information.",38166
"learning two these we a transfer problems. propose novel solve deep to Herein, approach","Herein, we propose a novel deep transfer learning approach to solve these two problems.",38167
paper problem This of environments. addresses the head in crowded detection,This paper addresses the problem of head detection in crowded environments.,38168
pedestrian levels. the an method setting density with We in outdoor validate varying,We validate the method in an outdoor setting with varying pedestrian density levels.,38169
field. requires Semantic sizeable information receptive and segmentation both rich spatial,Semantic segmentation requires both rich spatial information and sizeable receptive field.,38170
"address this (BiSeNet). In paper, novel Bilateral we Network this with a dilemma Segmentation","In this paper, we address this dilemma with a novel Bilateral Segmentation Network (BiSeNet).",38171
to others' shape affective Social may and empathy relations affective states.,Social and affective relations may shape empathy to others' affective states.,38172
stimuli could equally face identifiable. sizes that the of Results suggested two be,Results suggested that face stimuli of the two sizes could be equally identifiable.,38173
uncertainty quantification enable for random density. the normal Rogers-Szego inputs polynomials described by wrapped,Rogers-Szego polynomials enable uncertainty quantification for random inputs described by the wrapped normal density.,38174
Different been to methods MOD. have proposed of categories solve,Different categories of methods have been proposed to solve MOD.,38175
by is unsupervised stochastic descent end-to-end an Optimization in performed and gradient fashion.,Optimization is performed by stochastic gradient descent in an end-to-end and unsupervised fashion.,38176
Our modes. can work batch in online and both algorithm,Our algorithm can work in both batch and online modes.,38177
"all uses like the other In optimizer methods, the images. batch batch mode,","In the batch mode, like other batch methods, optimizer uses all the images.",38178
"into can In online fed be optimizer. mode, the images incrementally","In online mode, images can be incrementally fed into the optimizer.",38179
"for relevant estimation. motion segmentation Moreover, and scene flow are occlusions","Moreover, occlusions are relevant for motion segmentation and scene flow estimation.",38180
estimated The state-of-the-art. the boundaries motion improve over and occlusions clearly,The estimated occlusions and motion boundaries clearly improve over the state-of-the-art.,38181
Merr.] [Glycine step max pod is Loss (L.) soybean a key of during dehiscence,Loss of pod dehiscence is a key step during soybean [Glycine max (L.) Merr.],38182
Humidity the distribution shapes alleles. of indehiscent,Humidity shapes the distribution of indehiscent alleles.,38183
model the is for benchmark SALSA method. state-of-the-art the The applied dataset against challenging to,The model is applied to the challenging SALSA dataset for benchmark against the state-of-the-art method.,38184
"particular this setting, outperforms presented the e.g. formulation in Our state-of-the-art significantly","Our presented formulation outperforms the state-of-the-art significantly in this particular setting, e.g.",38185
Such existing using correspondences. supervised datasets are truth with often derived ground on learning descriptors,Such descriptors are often derived using supervised learning on existing datasets with ground truth correspondences.,38186
"prohibitive. usually datasets the is cost creating such However, of","However, the cost of creating such datasets is usually prohibitive.",38187
weakly Our is supervised aligned. because be need images method to roughly input currently the,Our method is currently weakly supervised because the input images need to be roughly aligned.,38188
also images. method We on our RGB evaluated,We also evaluated our method on RGB images.,38189
"is cluster and each Challenging respectively. firstly, of landmarks optimized are focused further landmarks","Challenging landmarks are focused firstly, and each cluster of landmarks is further optimized respectively.",38190
main compare the biological of to algorithmic paper search techniques sequences. The and surveys evolution,The paper surveys the evolution of main algorithmic techniques to compare and search biological sequences.,38191
replace discuss alignment-based expansion coming in to algorithms alignment-free We of analyses. techniques large-scale the,We discuss the expansion of alignment-free techniques coming to replace alignment-based algorithms in large-scale analyses.,38192
"challenges. on genomics Finally, algorithmic focus associated the outline and to transition population we","Finally, we focus on the transition to population genomics and outline associated algorithmic challenges.",38193
datasets. viewpoint estimation enables iSPA-Net This on state-of-the-art real image to achieve performance various,This enables iSPA-Net to achieve state-of-the-art performance on various real image viewpoint estimation datasets.,38194
"multiple of effectiveness the applications. demonstrate we for Further, approach","Further, we demonstrate effectiveness of the approach for multiple applications.",38195
"codes proposed algorithm. To the encourage reproducible research, for our released we have","To encourage reproducible research, we have released the codes for our proposed algorithm.",38196
"limits or using performance reconstruction usually approach alone applications. However, recovery image either in","However, using either approach alone usually limits performance in image reconstruction or recovery applications.",38197
promising competing state-of-the-art to performance methods. Results compared show,Results show promising performance compared to state-of-the-art competing methods.,38198
effective phase information to and % retrieval. phase the missing compensate achieve for,% to compensate for the missing phase information and achieve effective phase retrieval.,38199
"is stochastic attractor of network responses, experimentally. not observed drift grid which in models, predict","in attractor network models, predict stochastic drift of grid responses, which is not observed experimentally.",38200
of reset network in phase the drifting cells grids. learned Border this,Border cells in this learned network reset the phase of drifting grids.,38201
on U-Net the based well known Our is architecture. network,Our network is based on the well known U-Net architecture.,38202
"can suffer throughout core organisms lives, behaviors. performance their impair Most damage of neuronal which","Most organisms suffer neuronal damage throughout their lives, which can impair performance of core behaviors.",38203
"routines. into producing clinical outputs deterministic hinders DL adoption Moreover,","Moreover, producing deterministic outputs hinders DL adoption into clinical routines.",38204
estimates Uncertainty the subsequent permit by predictions revision clinicians. for would,Uncertainty estimates for the predictions would permit subsequent revision by clinicians.,38205
"variations time manual consuming is inter-observer to delineation dermatologists. subject and However, intra- and among","However, manual delineation is time consuming and subject to intra- and inter-observer variations among dermatologists.",38206
underlines skin need This and segmentation. accurate lesion approach for the automatic to an,This underlines the need for an accurate and automatic approach to skin lesion segmentation.,38207
proposed classifier a regions using a regressor. bounding-box The softmax and subsequently refined are,The proposed regions are subsequently refined using a softmax classifier and a bounding-box regressor.,38208
with two propose hierarchical network a We neural branches.,We propose a hierarchical neural network with two branches.,38209
colors. is branch branch object's the what other the learns the One while object learns,One branch learns what the object is while the other branch learns the object's colors.,38210
and The colorization semantic optimizes loss a a segmentation jointly network loss.,The network jointly optimizes a semantic segmentation loss and a colorization loss.,38211
typical maps stereo by artifacts. depth allows affected obtaining strategy not This,This strategy allows obtaining depth maps not affected by typical stereo artifacts.,38212
to Traditional learned classes. already only image consider objects recognition methods belonging,Traditional image recognition methods only consider objects belonging to already learned classes.,38213
"radar. we aperture graph paper, based for In imaging this propose processing signal synthetic","In this paper, we propose graph signal processing based imaging for synthetic aperture radar.",38214
the method solve function via of direction We multipliers. cost alternating,We solve the cost function via alternating direction method of multipliers.,38215
improved method enhancing capabilities. Our and provides resolution denoising,Our method provides improved denoising and resolution enhancing capabilities.,38216
sensing easily. can accommodate the quite framework compressed It also,It can also accommodate the compressed sensing framework quite easily.,38217
proposed methodology. the Experimental of our results validity corroborate,Experimental results corroborate the validity of our proposed methodology.,38218
"shortage. such food to development, poor are and weather during Organisms challenges frequently exposed as","Organisms are frequently exposed to challenges during development, such as poor weather and food shortage.",38219
"hormonal of glucocorticoids. initiate response, Such secretion involves the stress which challenges can","Such challenges can initiate the hormonal stress response, which involves secretion of glucocorticoids.",38220
are reduced shortening. Glucocorticoids and also telomere survival associated with,Glucocorticoids are also associated with reduced survival and telomere shortening.,38221
detect been (ECG). classification Electrocardiogram automatic method has and to studied effectively recognize An,An automatic classification method has been studied to effectively detect and recognize Electrocardiogram (ECG).,38222
are Cardiovascular Chinese Disease Database the conducted on Experiments (CCDD).,Experiments are conducted on the Chinese Cardiovascular Disease Database (CCDD).,38223
"rarely model complexity has However, observable data. the additional of context studied patient-specific in been","However, additional model complexity has rarely been studied in the context of patient-specific observable data.",38224
developed find the help Many contents want been have people to video they efficiently. methods,Many methods have been developed to help people find the video contents they want efficiently.,38225
"problems However, some unsolved this there area. still in are","However, there are still some unsolved problems in this area.",38226
"task, re-localization}, namely address this to distinctively define new \textbf{video a scenario. We","We define a distinctively new task, namely \textbf{video re-localization}, to address this scenario.",38227
the methods. the show proposed competing that experimental method results outperforms Extensive,Extensive experimental results show that the proposed method outperforms the competing methods.,38228
employ this We ensemble for an of networks neural task. convolutional,We employ an ensemble of convolutional neural networks for this task.,38229
"and as such In pretrained particular, ResNeXt. state-of-the-art fine-tune deep SENet we models Densenet, learning","In particular, we fine-tune pretrained state-of-the-art deep learning models such as Densenet, SENet and ResNeXt.",38230
"the Last, consider for we meta final predictions. learning approaches","Last, we consider meta learning approaches for the final predictions.",38231
human multi-camera approach the motion a We body. propose for capture markerless CNN-based of,We propose a CNN-based approach for multi-camera markerless motion capture of the human body.,38232
deep different semantics. contain maps Feature network in generally neural,Feature maps in deep neural network generally contain different semantics.,38233
to may methods characteristics Existing their omit sub-optimal lead often results. that,Existing methods often omit their characteristics that may lead to sub-optimal results.,38234
"often deep layers semantics. have layers Shallow and advantages in local contain global more information,","Shallow layers often contain more local information, and deep layers have advantages in global semantics.",38235
on datasets. and Experiments synthetic both are performed real,Experiments are performed on both real and synthetic datasets.,38236
"performance. appropriate That of lossless achieve values is, necessary finding the combination to is good","That is, finding appropriate combination of the values is necessary to achieve good lossless performance.",38237
heatmaps to layer-wise local capture better We keypoint contexts. using multi-scale detection refine supervision,We refine keypoint detection heatmaps using layer-wise multi-scale supervision to better capture local contexts.,38238
the is a globally inference using association keypoint network via Pose end. optimized at regression,Pose inference via keypoint association is optimized globally using a regression network at the end.,38239
Placing blockage. height greater the reduces a at of likelihood the BS,Placing the BS at a greater height reduces the likelihood of blockage.,38240
to identify locations. at times person Person specific re-identification a and aims distinct,Person re-identification aims to identify a specific person at distinct times and locations.,38241
"camera is and challenging of illumination, in It change occlusion, views. because viewpoint","It is challenging because of occlusion, illumination, and viewpoint change in camera views.",38242
"is real-world multi-shot Recently, receives closer since to it task application. attention person re-id more","Recently, multi-shot person re-id task receives more attention since it is closer to real-world application.",38243
constructs function. loss a methodology self-defined Our,Our methodology constructs a self-defined loss function.,38244
"equalization step. Consequently, channel can demodulator coupled and be processing in the one","Consequently, the demodulator and channel equalization can be coupled in one processing step.",38245
aims two learn domains. translation the mapping to visual Image-to-image between,Image-to-image translation aims to learn the mapping between two visual domains.,38246
is fundamental problems. in detection different Recognizing visual scales instances at a challenge simultaneously,Recognizing instances at different scales simultaneously is a fundamental challenge in visual detection problems.,38247
achieves performance the proposed show challenging ActvityNet on dataset. that DTPN experiments Extensive state-of-the-art the,Extensive experiments show that the proposed DTPN achieves state-of-the-art performance on the challenging ActvityNet dataset.,38248
for systems driven The wireless to performance. high the need has communication communication,The need for wireless communication has driven the communication systems to high performance.,38249
of multiplier-less pipeline This presents implementation study for structure. butterfly on-chip FFT digit-slicing an,This study presents an on-chip implementation of pipeline digit-slicing multiplier-less butterfly for FFT structure.,38250
software. simulated investigated architecture was with The and new MATLAB,The new architecture was investigated and simulated with MATLAB software.,38251
cameras. visual odometry We omnidirectional novel direct propose real-time a for monocular,We propose a novel real-time direct monocular visual odometry for omnidirectional cameras.,38252
registration minimal approach popular We methods how augment computational overhead. with can demonstrate many our,We demonstrate how our approach can augment many popular registration methods with minimal computational overhead.,38253
"semantic retrieval, provides summarization. Segmenting and events indexing, for content structures into video","Segmenting video content into events provides semantic structures for indexing, retrieval, and summarization.",38254
"events, ineffective to heterogeneous are methodologies deal e.g. However, such with","However, such methodologies are ineffective to deal with heterogeneous events, e.g.",38255
"direction, taking e.g. a temporary in the sight and walk, changes","taking a walk, and temporary changes in the sight direction, e.g.",38256
important communication Digital Decimation plays task system. process a an in,Digital Decimation process plays an important task in a communication system.,38257
reduction mostly the applied required. is transceiver when frequency It is in,It mostly is applied in transceiver when the frequency reduction is required.,38258
"process However, sigma-delta considered decimation the is modulator research this in work. for","However, the decimation process for sigma-delta modulator is considered in this research work.",38259
"arithmetic system the Furthermore, improve to the efficiency. unit is proposed advance applied","Furthermore, the proposed advance arithmetic unit is applied to improve the system efficiency.",38260
"with (i.e. expensive, problem fundamental a for presents or private This data applications limited,","This presents a fundamental problem for applications with limited, expensive, or private data (i.e.",38261
"data), could and personalized. be estimation/tracking which human such as highly pose small behavior","small data), such as human pose and behavior estimation/tracking which could be highly personalized.",38262
"MR accelerate deep learning-based image the reconstruction techniques suggested Recently, acquisition. to MRI were","Recently, the deep learning-based MRI reconstruction techniques were suggested to accelerate MR image acquisition.",38263
most reconstruction MRI deep generalizability common and learning-based transferability. in are issues approaches The any,The most common issues in any deep learning-based MRI reconstruction approaches are generalizability and transferability.,38264
paper want this fact. We this to rewrite avoiding for,We want to rewrite this paper for avoiding this fact.,38265
"security image systems, registrations. be seen could This facial identity in for recognition","This could be seen in facial image recognition for security systems, identity registrations.",38266
of bottleneck is Hence processing. registration identity a image,Hence a bottleneck of identity registration is image processing.,38267
"dual application. increases of high-performance a advance the Furthermore, its database efficiency","Furthermore, its advance dual database increases the efficiency of a high-performance application.",38268
"recognition an input into tool. fed Hence, the authenticated being image before be will","Hence, an input image will be authenticated before being fed into the recognition tool.",38269
used the for filters The decimators commonly modulators. are comb as sigma-delta non-recursive and recursive,The recursive and non-recursive comb filters are commonly used as decimators for the sigma-delta modulators.,38270
low of comb and the high-speed paper analysis presents filters. and power design This,This paper presents the analysis and design of low power and high-speed comb filters.,38271
size accuracy to and building respect with establish We density. also scene trends,We also establish accuracy trends with respect to building size and scene density.,38272
of basis responses remains The mechanistic mostly unknown. differential such,The mechanistic basis of such differential responses remains mostly unknown.,38273
link supported further knockdown by a experiments. Such was,Such a link was further supported by knockdown experiments.,38274
agonists. and These rhCG natural differences act rhLH as that pharmacological biased demonstrate,These pharmacological differences demonstrate that rhLH and rhCG act as natural biased agonists.,38275
ones. for classes classifier their coarse-grained fine-grained with overlap the that respective Each votes,Each classifier votes for the fine-grained classes that overlap with their respective coarse-grained ones.,38276
achieves in Our benchmark performance on recognition algorithm location state-of-the-art datasets. multiple the,Our algorithm achieves the state-of-the-art performance in location recognition on multiple benchmark datasets.,38277
organs based is a analysis computing. imaging on in key of task Morphological medical images,Morphological analysis of organs based on images is a key task in medical imaging computing.,38278
"In work, this network deep predicting diabetes shapes. abdominal for we a on neural propose","In this work, we propose a deep neural network for predicting diabetes on abdominal shapes.",38279
shape operates or requiring clouds processing The point directly without network raw alignment. mesh on,The network directly operates on raw point clouds without requiring mesh processing or shape alignment.,38280
"descriptor to For shape extend the the we state-of-the-art AbdomenPrint. comparison, BrainPrint","For comparison, we extend the state-of-the-art shape descriptor BrainPrint to the AbdomenPrint.",38281
hormone reproduction. and in for the Follicle-stimulating produced pituitary essential is (FSH) is,Follicle-stimulating hormone (FSH) is produced in the pituitary and is essential for reproduction.,38282
somatic receptor membrane It of the binds to in (FSHR) gonads. a cells specifically expressed,It specifically binds to a membrane receptor (FSHR) expressed in somatic cells of the gonads.,38283
presents The system compared FSH/FSHR to many receptors (GPCRs). protein-coupled classical peculiarities G,The FSH/FSHR system presents many peculiarities compared to classical G protein-coupled receptors (GPCRs).,38284
heterodimeric naturally is heterogeneous a large FSH glycoprotein.,FSH is a large naturally heterogeneous heterodimeric glycoprotein.,38285
"Once instances, subunits. couples and, to some other to G{\alpha} in activated, G{\alpha}s FSHR the","Once activated, the FSHR couples to G{\alpha}s and, in some instances, to other G{\alpha} subunits.",38286
designed is evaluated Experimental performance controller the environment. of testing implemented laboratory and under for,Experimental testing of the designed controller is implemented and evaluated for performance under laboratory environment.,38287
"released R In a calibration. present for we recently article, Bayesian package this","In this article, we present a recently released R package for Bayesian calibration.",38288
Many experiments. industrial are unfeasible fields facing costly field or,Many industrial fields are facing unfeasible or costly field experiments.,38289
by numerical/computer experiments These with experiments are realized replaced a numerical running which code. are,These experiments are replaced with numerical/computer experiments which are realized by running a numerical code.,38290
"example Eventually, detailed using CaliCo. is toy a","Eventually, a toy example is detailed using CaliCo.",38291
"of neuroimaging information. important other contain statistics the However, data may","However, other statistics of the neuroimaging data may contain important information.",38292
This the paper wireless access two-tier heterogeneous a efficacy in of decoupled analyzes network.,This paper analyzes the efficacy of decoupled wireless access in a two-tier heterogeneous network.,38293
different benefits been scenarios access The in have decoupled studied wireless performance recently. and its,The decoupled wireless access and its performance benefits have been studied in different scenarios recently.,38294
"In perspective analysis an efficiency this from efficacy on paper, its provided. is spectral in-depth","In this paper, an in-depth analysis on its efficacy from spectral efficiency perspective is provided.",38295
"are which aforementioned the results. validate simulation Furthermore, provided rigorous results analytical","Furthermore, rigorous simulation results are provided which validate the aforementioned analytical results.",38296
computations required perform the to matrix available in is SAS/IML for system. SAS the DRA,SAS/IML is available to perform the required matrix computations for DRA in the SAS system.,38297
"viewing imagery at single a by time. Currently, band examine analysts multiband a","Currently, analysts examine multiband imagery by viewing a single band at a time.",38298
combine form then and We MSAS the the final image. measures backscatter to saliency acoustic,We then combine the saliency and acoustic backscatter measures to form the final MSAS image.,38299
accurate and does use approximations. Our any is not computation,Our computation is accurate and does not use any approximations.,38300
optimization a a Our calibration close-form includes re-projection errors. by non-linear and method minimizing solution,Our calibration method includes a close-form solution and a non-linear optimization by minimizing re-projection errors.,38301
spatially pathological such slides. interface unique pathology to digitised provides a on opportunity dissect Computational,Computational pathology provides a unique opportunity to spatially dissect such interface on digitised pathological slides.,38302
GM accuracy between and Segmenting is point up a trade-off a precision. to,Segmenting GM is up to a point a trade-off between accuracy and precision.,38303
resonance segmentation. for GM-WM (MR) We a and new image acquisition pipeline magnetic propose,We propose a new pipeline for GM-WM magnetic resonance (MR) image acquisition and segmentation.,38304
"articulated evaluated with variations. figures, which We human are ISP-GPM highly reposing on versatile","We evaluated ISP-GPM on reposing human figures, which are highly articulated with versatile variations.",38305
this and study attempt In paper it. to we process this invert sketching,In this paper we study this sketching process and attempt to invert it.,38306
"quality manual often However, procedures time-consuming error-prone. and are control","However, manual quality control procedures are often time-consuming and error-prone.",38307
GDXray Welds on The well dataset. also performs the system,The system also performs well on the GDXray Welds dataset.,38308
pattern systematically exceeds how To simulations extended terminates current using computational spatially this resources. address,To systematically address how this pattern terminates using spatially extended simulations exceeds current computational resources.,38309
learning for the This problem the dataset alleviates simplifies bias motions. and camera significantly,This significantly simplifies the learning problem and alleviates the dataset bias for camera motions.,38310
"estimate. we volume a accumulate cost For depth information in mapping, the at centered current","For mapping, we accumulate information in a cost volume centered at the current depth estimate.",38311
favorably dense powered deep and compare learning against depth algorithms. classic strong We,We compare favorably against strong classic and deep learning powered dense depth algorithms.,38312
consumption data management. to able power monitor for Homeowners reliable are power efficient,Homeowners are able to monitor reliable power consumption data for efficient power management.,38313
"Most architecture, with same quantizing on though precision. the uncritically previous CNNs works assume reduced","Most previous works on quantizing CNNs uncritically assume the same architecture, though with reduced precision.",38314
"chronic hand, stress of patients. oxidative the the for one On is HIV responsible infection","On one hand, HIV infection is responsible for the chronic oxidative stress of the patients.",38315
"the contributions the other HIV On hand, oxidative disease the pathogenesis. stress to","On the other hand, the oxidative stress contributions to the HIV disease pathogenesis.",38316
and Numerical bifurcation viral presented stress. simulations oxidative illustrate are analysis to the under dynamics,Numerical simulations and bifurcation analysis are presented to illustrate the viral dynamics under oxidative stress.,38317
"Nakagami-m Weibull investigated, namely, of \alpha-\mu special and fading. the model cases Three are Rayleigh,","Three special cases of the \alpha-\mu model are investigated, namely, Rayleigh, Nakagami-m and Weibull fading.",38318
accuracy Carlo the our Monte provided The validate of analysis. simulations,The provided Monte Carlo simulations validate the accuracy of our analysis.,38319
FD relaying better the is It systems. shown performs that HD than the scenario also,It is also shown that the FD scenario performs better than the HD relaying systems.,38320
the systems. for all of functioning Protein-DNA are successful critical natural interactions,Protein-DNA interactions are critical for the successful functioning of all natural systems.,38321
Patlak-Keller-Segel a as community. model the use of type We system of a,We use a system of Patlak-Keller-Segel type as a model of the community.,38322
distributions solutions translationally invariant) the basic of all systems where homogeneous with species. are equilibria,translationally invariant) systems where the basic solutions are equilibria with homogeneous distributions of all species.,38323
effect inhomogeneity. short-wavelength the of Our goal is to examine,Our goal is to examine the effect of short-wavelength inhomogeneity.,38324
"employ we procedure. homogenization this, certain do To a","To do this, we employ a certain homogenization procedure.",38325
show data of on experiments proposed patient the the method. effectiveness The,The experiments on the patient data show effectiveness of the proposed method.,38326
environments. deep a for present indoor We scene technique modeling generative,We present a deep generative scene modeling technique for indoor environments.,38327
interpolation scene also model scene applications completion. generative this the We of and in show,We also show the applications of this generative model in scene interpolation and scene completion.,38328
"re-id, metric and learning. discriminative There namely are person two steps to crucial learning feature","There are two steps crucial to person re-id, namely discriminative feature learning and metric learning.",38329
develop sketch this universal In grouper. we aim a to work,In this work we aim to develop a universal sketch grouper.,38330
The large-scale first to with the lack is datasets annotation. grouping goal of this obstacle,The first obstacle to this goal is the lack of large-scale datasets with grouping annotation.,38331
"a novel deep perceptual propose model. Furthermore, universal grouping we","Furthermore, we propose a novel deep universal perceptual grouping model.",38332
The both is with generative and model losses. learned discriminative,The model is learned with both generative and discriminative losses.,38333
the show groupers. the that state-of-the-art We significantly proposed outperforms model,We show that the proposed model significantly outperforms the state-of-the-art groupers.,38334
to achieve estimation utilizes direct module (CNN) end-to-end quantification. Neural The Convolutional Network,The direct estimation module utilizes Convolutional Neural Network (CNN) to achieve end-to-end quantification.,38335
or camera health levels illumination services. patients' at varying mental viewpoints) the at homes and,varying illumination levels and camera viewpoints) at the patients' homes or at mental health services.,38336
prediction evaluate patient-independent SchiNet schizophrenia. the of symptoms proposed We for of,We evaluate the proposed SchiNet for patient-independent prediction of symptoms of schizophrenia.,38337
matrix perform the DRA the to available SAS computations is for SAS/IML required system. in,SAS/IML is available to perform the required matrix computations for DRA in the SAS system.,38338
(FSH) Follicle-stimulating luteinizing in essential (LH) hormone hormone reproduction. and animal roles play,Follicle-stimulating hormone (FSH) and luteinizing hormone (LH) play essential roles in animal reproduction.,38339
vision. active in the research fundamental computer most Feature is matching areas one of and,Feature matching is one of the most fundamental and active research areas in computer vision.,38340
"investigated matching performances their different different scenes types. in Also, are in and","Also, their performances are investigated in different scenes and in different matching types.",38341
"role crucial extracting a category and Therefore, accurate in play localizing features local prediction. discriminative","Therefore, localizing and extracting discriminative local features play a crucial role in accurate category prediction.",38342
"solve Bilinear this propose Supervised In (WS-BAN) Weakly to Attention these issues. Network we paper,","In this paper, we propose Weakly Supervised Bilinear Attention Network (WS-BAN) to solve these issues.",38343
"paper, propose efficient quantized U-Nets we localization. In densely for connected landmark visual this","In this paper, we propose quantized densely connected U-Nets for efficient visual landmark localization.",38344
"localization information the yielding improved connectivity largely accuracy. dense flow, improves This","This dense connectivity largely improves the information flow, yielding improved localization accuracy.",38345
face tasks: validate in two alignment. human and estimation approach our pose We,We validate our approach in two tasks: human pose estimation and face alignment.,38346
video language sentence). and a a clip,a video clip and a language sentence).,38347
of key matching multimodal network consists components. two Our,Our multimodal matching network consists of two key components.,38348
particular the in have a Most in methods application. mind literature found,Most methods found in the literature have in mind a particular application.,38349
"the proposed to methods. traditional variation network is more communication channel robust than system Furthermore,","Furthermore, the proposed system network is more robust to channel variation than traditional communication methods.",38350
communication. laser-based pulse performance characterize the and for the signal We achievable,We characterize the signal and achievable performance for the pulse laser-based communication.,38351
solutions Such level AI free the AIQ hands ultimate its elevate to position.,Such hands free AI solutions elevate the AIQ level to its ultimate position.,38352
interest are There of benign among objects much seafloor. present few,There are few objects of interest present among much benign seafloor.,38353
"large collected of these problems, portion discard a the techniques current classification information. Despite SAS","Despite these problems, current classification techniques discard a large portion of the collected SAS information.",38354
"the the classification process information. void Consequently, completely phase of is","Consequently, the classification process is completely void of the phase information.",38355
results. of the validity to statistical our demonstrate We perform testing,We perform statistical testing to demonstrate the validity of our results.,38356
sensing. therefore the challenges remote one of fusion is Data key in,Data fusion is therefore one of the key challenges in remote sensing.,38357
"multi-modal on during inference. fusion, always modalities prior available assumes Most work focusing that are","Most prior work focusing on multi-modal fusion, assumes that modalities are always available during inference.",38358
"this learning approach zero-shot In introduce simple, observation. efficient a based this work we on","In this work we introduce a simple, efficient zero-shot learning approach based on this observation.",38359
"challenges discuss we and open some research Finally, promising directions. research future","Finally, we discuss some open research challenges and promising future research directions.",38360
prediction. the approximator ConvNet-LSTM for is direct A function adopted frame-to-action,A ConvNet-LSTM function approximator is adopted for the direct frame-to-action prediction.,38361
"AU architecture, Without the our be network easily intensity can changing framework estimation. for extended","Without changing the network architecture, our framework can be easily extended for AU intensity estimation.",38362
for have problem large input. methods interactive of Many that of ask number the user,Many interactive methods have the problem of that ask for large number of user input.,38363
application. intuitive user To is input produce limited efficient important under for industrial segmentation,To efficient produce intuitive segmentation under limited user input is important for industrial application.,38364
applying iterative higher by segmentation random walks performance be algorithm. boundary And can obtained the,And higher segmentation performance can be obtained by applying the iterative boundary random walks algorithm.,38365
large data reconstruct an present from sets. algorithm jointly We also measurements to,We also present an algorithm to jointly reconstruct measurements from large data sets.,38366
preventable of over the main is DR for all the blindness the world. one reasons,DR is one of the main reasons for the preventable blindness all over the world.,38367
early severe may not detected irreversible blindness. of to If the stages patient progress,If not detected early the patient may progress to severe stages of irreversible blindness.,38368
poses the serious Ophthalmologists diabetes Lack a patients. for of growing problem,Lack of Ophthalmologists poses a serious problem for the growing diabetes patients.,38369
when develop present. Hard is DR exudates,Hard exudates develop when DR is present.,38370
"success. methods have (GANs) Recently, Generative great based Adversarial Nets achieved several","Recently, several Generative Adversarial Nets (GANs) based methods have achieved great success.",38371
conditioned line They on illustrations color generate given hints. art can colorized and,They can generate colorized illustrations conditioned on given line art and color hints.,38372
We data. independent introduce features network local is synthetic a also that of,We also introduce a local features network that is independent of synthetic data.,38373
problem in for devices. inertial these on handheld odometry pose a Low-grade IMUs smart-devices,Low-grade IMUs in handheld smart-devices pose a problem for inertial odometry on these devices.,38374
structure-from-motion. global large-scale outlier problem This work the removal in addresses,This work addresses the outlier removal problem in large-scale global structure-from-motion.,38375
in speedup the results algorithms. significant new of The a dimension reduction,The dimension reduction results in a significant speedup of the new algorithms.,38376
"the can the method iteratively reweighted reduce true possibility of Further, significantly removing inliers.","Further, the iteratively reweighted method can significantly reduce the possibility of removing true inliers.",38377
and are of controlled lengths expressions types users. the The by,The lengths and types of the expressions are controlled by users.,38378
our user Experiments approach. and studies of effectiveness verify the,Experiments and user studies verify the effectiveness of our approach.,38379
the filters graph core processing. tools are in signal of one Graph,Graph filters are one of the core tools in graph signal processing.,38380
direct is aspect them distributed their central A of implementation.,A central aspect of them is their direct distributed implementation.,38381
"savings. However, filtering computational the performance distributed and is often traded with communication","However, the filtering performance is often traded with distributed communication and computational savings.",38382
the era have We multidimensional of entered increasingly datasets complex large by represented structures. data,We have entered the era of large multidimensional datasets represented by increasingly complex data structures.,38383
and methods train cohorts. test our independent on We,We train and test our methods on independent cohorts.,38384
"Indeed, was it in any competitive algorithms used not our knowledge. recent in MAX-SAT","Indeed, it was not used in any recent competitive MAX-SAT algorithms in our knowledge.",38385
"In look self-adaptive closer systems. a of taken is the paper, at this structure","In this paper, a closer look is taken at the structure of self-adaptive systems.",38386
in that controllable. FOMQA are settings not finitely relevant study The is of,The study of FOMQA is relevant in settings that are not finitely controllable.,38387
"SOI, decidability proper of and We FOMQA show fragments for SOF, of SOIF: SIF. three","We show decidability of FOMQA for three proper fragments of SOIF: SOI, SOF, and SIF.",38388
entailment. characterise Our relevant for to finite is approach models query deciding,Our approach is to characterise models relevant for deciding finite query entailment.,38389
"is to related the problem i.e. This imaging applications, frequently","This problem is frequently related to the imaging applications, i.e.",38390
within detection the a by camera. scene acquired,detection within the scene acquired by a camera.,38391
scenario. detection we and perform in evaluate these a hyperspectral Finally experiments methods,Finally we perform experiments and evaluate these methods in a hyperspectral detection scenario.,38392
The the with simulated investigated was and new architecture software. MATLAB,The new architecture was investigated and simulated with the MATLAB software.,38393
"introduction i.e. amount large antenna With (BS) Base the of of Station with ports,","With the introduction of Base Station (BS) with large amount of antenna ports, i.e.",38394
quantitatively. compare the We qualitatively and methods,We compare the methods qualitatively and quantitatively.,38395
"generating Conditional allows variant conditions. GANs, GAN, samples given in of a","Conditional GAN, a variant of GANs, allows generating samples in given conditions.",38396
challenging The due reconstruction projections. is inversion image potentially defective to Radon the or radon,The Radon inversion or image reconstruction is challenging due to the potentially defective radon projections.,38397
"the stages contains process several reconstruction the corresponding Radon hoc ad to inversion. Conventionally, approximate","Conventionally, the reconstruction process contains several ad hoc stages to approximate the corresponding Radon inversion.",38398
is stage. of stages of results the on dependent the highly previous Each the,Each of the stages is highly dependent on the results of the previous stage.,38399
"(inverse For the transform Radon as framework proposed is iRadonMap approximation). simplicity, short","For simplicity, the proposed framework is short as iRadonMap (inverse Radon transform approximation).",38400
"the clinical iRadonMap, evaluate data patient of performance is used. the To","To evaluate the performance of the iRadonMap, clinical patient data is used.",38401
of iRadonMap. promising the performance reconstruction Qualitative show results,Qualitative results show promising reconstruction performance of the iRadonMap.,38402
obtaining non-monotonic conclusions systems. systems are that argumentation formal frameworks facilitate Abstract knowledge from,Abstract argumentation frameworks are formal systems that facilitate obtaining conclusions from non-monotonic knowledge systems.,38403
Splitting up framework semantics. the an speed argumentation argumentation efficiently computation can of,Splitting an argumentation framework can efficiently speed up the computation of argumentation semantics.,38404
makes a Bidirectional splitting of use minimum cut.,Bidirectional splitting makes use of a minimum cut.,38405
(AI) of domains. a Intelligence variety in achieved super-human performance broad Artificial,Artificial Intelligence (AI) achieved super-human performance in a broad variety of domains.,38406
quality Our a experiments show high with original model. flexible results,Our experiments show original high quality results with a flexible model.,38407
is data-driven possible. claim to We wrinkle approach cloth generation an entirely realistic,We claim an entirely data-driven approach to realistic cloth wrinkle generation is possible.,38408
approach motor. a The on induction is demonstrated three-phase,The approach is demonstrated on a three-phase induction motor.,38409
"weakly-supervised classes are with ""Thing"" boxes, and with image-level bounding tags. ""stuff""","""Thing"" classes are weakly-supervised with bounding boxes, and ""stuff"" with image-level tags.",38410
"for semantic- on results present first Cityscapes instance-segmentation. both and we the Furthermore, weakly-supervised","Furthermore, we present the first weakly-supervised results on Cityscapes for both semantic- and instance-segmentation.",38411
"generated the were sine data wave For a methods, a known validation sEMG and evaluated.","For the validation methods, a generated sine wave and a known sEMG data were evaluated.",38412
paper satellite imagery. method presents detection from object efficient an This,This paper presents an efficient object detection method from satellite imagery.,38413
target validated objects. We models using our as courses golf,We validated our models using golf courses as target objects.,38414
methods. object deep demonstrated accuracy previous higher The identification learning method than proposed,The proposed deep learning method demonstrated higher accuracy than previous object identification methods.,38415
"temporally. then and descriptors, colors propagated Using are both spatially appearance","Using appearance descriptors, colors are then propagated both spatially and temporally.",38416
We framework learning in videos. propose segmenting images and both end-to-end generic objects an for,We propose an end-to-end learning framework for segmenting generic objects in both images and videos.,38417
propose ways dense segmentation. Thus foreground for we weakly exploit data learning labeled to,Thus we propose ways to exploit weakly labeled data for learning dense foreground segmentation.,38418
and presents This DOA method. estimation paper this the corresponding relationship,This paper presents this relationship and the corresponding DOA estimation method.,38419
"using (GSV) Google View focus Here, measuring cover we Street images. tree on urban","Here, we focus on measuring urban tree cover using Google Street View (GSV) images.",38420
GSV on perform images. consisting a dataset of then We small training additional training,We then perform additional training on a small training dataset consisting of GSV images.,38421
outperform that unsupervised benchmark significantly We deep models learning method. the find,We find that deep learning models significantly outperform the unsupervised benchmark method.,38422
any is conditions. and require Our proposed platform system does lighting specific independent not,Our proposed system is platform independent and does not require any specific lighting conditions.,38423
"merge motion counts we across utilize the camera views. multiple Additionally, to","Additionally, we utilize camera motion to merge the counts across multiple views.",38424
"rows. of from merged both counts tree fruit sides Additionally, report the we","Additionally, we report merged fruit counts from both sides of the tree rows.",38425
improvement two mainly The in embodied aspects. was,The improvement was mainly embodied in two aspects.,38426
"aspect, redefine the In repairing priority. formula the order we the calculation of","In the repairing order aspect, we redefine the calculation formula of the priority.",38427
that the can the the algorithm. overcome confirm algorithm insufficiencies original of Experiments improved,Experiments confirm that the improved algorithm can overcome the insufficiencies of the original algorithm.,38428
"has The the been and a have better repairing results appearance. improved, visual effect","The repairing effect has been improved, and the results have a better visual appearance.",38429
"article, of on focus we movement In the cyclist this detecting initial devices. using smart","In this article, we focus on detecting the initial movement of cyclist using smart devices.",38430
"movement detection the a classification Therefore, initial as problem. model we","Therefore, we model the initial movement detection as a classification problem.",38431
hierarchy Images a parts. object of as are composed,Images are composed as a hierarchy of object parts.,38432
"to inference leads to graph. in intractable the loops this due Typically,","Typically, this leads to intractable inference due to loops in the graph.",38433
hierarchical successfully its images latent of verify states. by a visualising We learned model having,We verify having successfully learned a hierarchical model of images by visualising its latent states.,38434
in most methodology were Convolutional the networks challenge. successful BACH the neuronal,Convolutional neuronal networks were the most successful methodology in the BACH challenge.,38435
"flow same This describe the is thing. since disparity odd, optic and essentially binocular","This is odd, since binocular disparity and optic flow describe essentially the same thing.",38436
be inference has in achieved. Previous can shown how LPMLN work,Previous work has shown how inference in LPMLN can be achieved.,38437
abductive actions. about apply reasoning for to We probabilistic also the the method learn parameters,We also apply the method to learn the parameters for probabilistic abductive reasoning about actions.,38438
Which facial commonly many people use? and cross-cultural and cultural-specific do expressions how,Which and how many cross-cultural and cultural-specific facial expressions do people commonly use?,38439
"to each what And, does communicate variables expression affect observers?","And, what affect variables does each expression communicate to observers?",38440
effective Retinex image is model tool low-light an for enhancement.,Retinex model is an effective tool for low-light image enhancement.,38441
into illumination. the reflectance that and assumes can It be observed decomposed images,It assumes that observed images can be decomposed into the reflectance and illumination.,38442
formulas also control SINR targets. for to We power meet derive pre-determined,We also derive formulas for power control to meet pre-determined SINR targets.,38443
usage the A the formulas provided. of demonstrating is example numerical derived,A numerical example demonstrating the usage of the derived formulas is provided.,38444
"challenging segmentation application remains still of a task. airways to However, the its","However, its application to the segmentation of airways still remains a challenging task.",38445
learned is network and detection jointly unified The semantic by tasks. optimizing pedestrian segmentation,The unified network is learned by jointly optimizing pedestrian detection and semantic segmentation tasks.,38446
outperforms KAIST The approach state-of-the-art fast. dataset the methods while significantly remain on,The approach significantly outperforms state-of-the-art methods on the KAIST dataset while remain fast.,38447
"an as this possible. is that process generate Moreover, realistic manipulation as should image","Moreover, this manipulation process should generate an image that is as realistic as possible.",38448
Experiments environment. in well-known RoboCup@Work challenge the are performed,Experiments are performed in the well-known RoboCup@Work challenge environment.,38449
points. A pickup on network segmentation trained convolutional objects' detecting is fully,A fully convolutional segmentation network is trained on detecting objects' pickup points.,38450
sequences qualitative videos. on challenging YouTube evaluations standard ground-truth with reconstructions and We provide on,We provide evaluations on standard sequences with ground-truth and qualitative reconstructions on challenging YouTube videos.,38451
important roles play methods for pre-procedural in planning increasingly complex and surgeries learning Machine interventions.,Machine learning methods play increasingly important roles in pre-procedural planning for complex surgeries and interventions.,38452
visually pathophysiologically generate only but not patients virtual demonstrates authentic also can This GIN interpretable.,This demonstrates GIN can generate virtual patients not only visually authentic but also pathophysiologically interpretable.,38453
modulation The is to noncoherent transmission. keying on-off achieve adopted differential,The on-off keying differential modulation is adopted to achieve noncoherent transmission.,38454
The predict (CI) performance. detection Information is to the also Chernoff derived,The Chernoff Information (CI) is also derived to predict the detection performance.,38455
"address detection we problem Next, the reader. at the signal","Next, we address the signal detection problem at the reader.",38456
Extensive and verify results the of channel numerical method. proposed detectors estimation the effectiveness the,Extensive numerical results verify the effectiveness of the proposed detectors and the channel estimation method.,38457
presented. branch-cut-and-price this integer A solving problem to programming approach (IP) is,A branch-cut-and-price integer programming (IP) approach to solving this problem is presented.,38458
by high-level input DenseRAN employing as encoder. DenseNet first an features encodes image visual to,DenseRAN first encodes input image to high-level visual features by employing DenseNet as an encoder.,38459
primary consistent the of with finding and secondary This concepts bone healing. is,This finding is consistent with the concepts of primary and secondary bone healing.,38460
most difficult vision computer today. learning in poses challenges one of the Unsupervised,Unsupervised learning poses one of the most difficult challenges in computer vision today.,38461
on approach methods object is published from unsupervised Our discovery. different,Our approach is different from published methods on unsupervised object discovery.,38462
student-teacher of over several training. Our algorithm run generations can learning unsupervised,Our unsupervised learning algorithm can run over several generations of student-teacher training.,38463
disease despite heart of a being cause preventable. common death is Coronary,Coronary heart disease is a common cause of death despite being preventable.,38464
"approaches. study we different transfer learning improved performance, For","For improved performance, we study different transfer learning approaches.",38465
fuse effective a both We multi-path more architecture feature in representations exploitation. for,We fuse both representations in a multi-path architecture for more effective feature exploitation.,38466
"differentiation to challenge Last, of addition the we plaque in address detection.","Last, we address the challenge of plaque differentiation in addition to detection.",38467
that one is accidents. the glare of sun traffic The major cause environmental hazards,The sun glare is one of the major environmental hazards that cause traffic accidents.,38468
"traffic died injured in Every to people many accidents related year, glare. and sun","Every year, many people died and injured in traffic accidents related to sun glare.",38469
"Cambridge, in USA. study conducted Massachusetts, case a We","We conducted a case study in Cambridge, Massachusetts, USA.",38470
presence precisely. the of that method glare predict sun the show Results can,Results show that the method can predict the presence of sun glare precisely.,38471
"method that scenarios. different extensive experiments, between we seamlessly adapts By demonstrate proposed the","By extensive experiments, we demonstrate that the proposed method seamlessly adapts between different scenarios.",38472
by This algorithm discusses paper Metropolis-Hastings developed \citeA{MarsmanIsing}. a,This paper discusses a Metropolis-Hastings algorithm developed by \citeA{MarsmanIsing}.,38473
design. Beta-turn prediction in useful protein and experimental studies function is,Beta-turn prediction is useful in protein function studies and experimental design.,38474
these address provide neural to issues. new networks a Deep opportunity,Deep neural networks provide a new opportunity to address these issues.,38475
analysis has We complexity the complexity. that proposed algorithm perform to polynomial a show,We perform a complexity analysis to show that the proposed algorithm has polynomial complexity.,38476
algorithms our then weight correction. multimodal We mode to using tempering apply simulated targets,We then apply simulated tempering algorithms to multimodal targets using our mode weight correction.,38477
"Markov hidden Chinese handwritten (HMMs) Recently, promising recognition. for achieved offline models results text have","Recently, hidden Markov models (HMMs) have achieved promising results for offline handwritten Chinese text recognition.",38478
morphological extracellular sense mechanical changes. respond intracellular can and Plants forces can and both through,Plants can sense both intracellular and extracellular mechanical forces and can respond through morphological changes.,38479
are signaling mechanotransduction components largely unknown. the response The responsible for of touch,The signaling components responsible for mechanotransduction of the touch response are largely unknown.,38480
both model Future the ends of at might link. mobility this extensions of incorporate,Future extensions of this model might incorporate mobility at both ends of the link.,38481
to efficient be solution. The promises sum-of-sinusoids an model,The sum-of-sinusoids model promises to be an efficient solution.,38482
into training labels embedding by function randomly is subsets. bagging learned the small Each,Each embedding function is learned by randomly bagging the training labels into small subsets.,38483
functions. experimentally that create show embedding these embedding We ensembles effective,We show experimentally that these embedding ensembles create effective embedding functions.,38484
for NPCs. randomly Motion minor plans are often generated game,Motion plans are often randomly generated for minor game NPCs.,38485
"a movements, or pathing Repetitive with and/or regular non-trivial programming require however, system. integration effort","Repetitive or regular movements, however, require non-trivial programming effort and/or integration with a pathing system.",38486
across different space generally assume Existing domains. networks label identical domain adversarial,Existing domain adversarial networks generally assume identical label space across different domains.,38487
state-of-the-art datasets. for several partial that PADA tasks results Experiments domain adaptation show on exceeds,Experiments show that PADA exceeds state-of-the-art results for partial domain adaptation tasks on several datasets.,38488
Moving is a task Segmentation challenging Object for videos. jittery/wobbly,Moving Object Segmentation is a challenging task for jittery/wobbly videos.,38489
the represent means trajectories average of motion clusters. Frechet The the,The Frechet means represent the average trajectories of the motion clusters.,38490
"accuracy stabilized of clusters groups iteratively motion trajectories. are refined, of The distinct producing the","The accuracy of the motion clusters are iteratively refined, producing distinct groups of stabilized trajectories.",38491
quantitative proposed Qualitative method. results superiority indicate of and the the,Qualitative and quantitative results indicate the superiority of the proposed method.,38492
radiologists simulated to are that We between tissue. real and experienced distinguish scar show unable,We show that experienced radiologists are unable to distinguish between real and simulated scar tissue.,38493
recent in great received fine-grained Computer based years. attention recognition has vision,Computer vision based fine-grained recognition has received great attention in recent years.,38494
discriminative localization focus and learning. on part feature works Existing,Existing works focus on discriminative part localization and feature learning.,38495
affect applications. several ad Automatic has useful recognition,Automatic ad affect recognition has several useful applications.,38496
inform on how ad information approaches understanding. such humans visual do Neither prioritize for us,Neither do such approaches inform us on how humans prioritize visual information for ad understanding.,38497
and machine challenging a vision. learning and is task crucial in computer It,It is a crucial and challenging task in machine learning and computer vision.,38498
"the effect image we a clustering. features, feature-weak-relevant method propose To for clustering of reduce","To reduce the effect of clustering features, we propose a feature-weak-relevant method for image clustering.",38499
"clustering In iteratively optimized. this manner, image is","In this manner, image clustering is iteratively optimized.",38500
Regenerative the by train braking is energy produced deceleration. a energy during,Regenerative braking energy is the energy produced by a train during deceleration.,38501
"and train a as produce motors decelerates, generators electricity. act the When","When a train decelerates, the motors act as generators and produce electricity.",38502
paper of to regenerative aim braking energy. and is this simulate The model,The aim of this paper is to model and simulate regenerative braking energy.,38503
examples parameters the influence network These to when trained can them. correct strongly is,These examples can strongly influence parameters when the network is trained to correct them.,38504
"often in training and they obtain. sparse Unfortunately, data, expensive to the are are","Unfortunately, they are often sparse in the training data, and are expensive to obtain.",38505
improves experiments on obtained automatically show these significantly performance. examples that Our retraining detectors often,Our experiments show that retraining detectors on these automatically obtained examples often significantly improves performance.,38506
such influences movement. We greatly future path of prediction that information show common,We show that such common path information greatly influences prediction of future movement.,38507
that further filter information. movement scene non-linear a design holds We important data,We further design a scene data filter that holds important non-linear movement information.,38508
We datasets. model on evaluate two challenging our,We evaluate our model on two challenging datasets.,38509
"memory this However, RAM-like and is objects. unstructured does not matrix naturally thus structured encode","However, this RAM-like memory matrix is unstructured and thus does not naturally encode structured objects.",38510
queries objects programming. about to learns without a of graph-structured explicit set answer RMDN Thus,Thus RMDN learns to answer queries about a set of graph-structured objects without explicit programming.,38511
demonstrate model. of proposed Results the efficacy the,Results demonstrate the efficacy of the proposed model.,38512
step transition Detection of shot video crucial video is analysis. pre-processing in a,Detection of video shot transition is a crucial pre-processing step in video analysis.,38513
"smart trade-offs, framework. Considering performance a we speed design","Considering speed performance trade-offs, we design a smart framework.",38514
"Our autoencoder the branches, for consists structure two for and one geometry. of the one","Our autoencoder consists of two branches, one for the structure and one for the geometry.",38515
evaluate study. an and performance conduct method the ablation of We our,We evaluate the performance of our method and conduct an ablation study.,38516
We geometry. that similarities of explicitly for structure shapes show encoding both in and accounts,We explicitly show that encoding of shapes accounts for both similarities in structure and geometry.,38517
generated variety of results presented. A SAGNet quality are by,A variety of quality results generated by SAGNet are presented.,38518
multibox SSD---single improvements shot propose two We detector. to the,We propose two improvements to the SSD---single shot multibox detector.,38519
"approach in SSD. First, default propose an selection we adaptive box for","First, we propose an adaptive approach for default box selection in SSD.",38520
environments in ways cannot. fluidly Humans their machines adapt interest in can complex that,Humans can fluidly adapt their interest in complex environments in ways that machines cannot.,38521
twofold. of The work are this contributions,The contributions of this work are twofold.,38522
"issues, a To signaling we antenna regularized counteract selection implementation practical scheme. based propose","To counteract practical implementation issues, we propose a regularized antenna selection based signaling scheme.",38523
rates analyzed rate error the is The symbol/bit and derived. are data possible,The possible data rate is analyzed and the symbol/bit error rates are derived.,38524
pairwise discriminative PRN The among to is capture identities. and relations unique trained different,The PRN is trained to capture unique and discriminative pairwise relations among different identities.,38525
recent rapidly Service growing in has (VA been a Analytics Video as years. Software SaaS),Video Analytics Software as a Service (VA SaaS) has been rapidly growing in recent years.,38526
accessed a lightweight is by users VA using typically SaaS client.,VA SaaS is typically accessed by users using a lightweight client.,38527
proposed TSD advantages. three The appealing has,The proposed TSD has three appealing advantages.,38528
"Firstly, TSD the architecture and deployed eg. lightweight can be a client, in has","Firstly, TSD has a lightweight architecture and can be deployed in the client, eg.",38529
"save transmission frames mobile bandwidth. devices, to to compressed representative produce","mobile devices, to produce compressed representative frames to save transmission bandwidth.",38530
Extensive show experiments characteristics and the effectiveness TSD. of,Extensive experiments show the effectiveness and characteristics of TSD.,38531
"scheme no such the effects. partition statistical nonlinearity to Furthermore, devised biodiversity can under be","Furthermore, under such nonlinearity no statistical scheme can be devised to partition the biodiversity effects.",38532
alternative more We estimate present a metric also biodiversity providing of effect. reasonable an,We also present an alternative metric providing a more reasonable estimate of biodiversity effect.,38533
features its effectiveness. has for recently shown Aggregating different image retrieval image,Aggregating different image features for image retrieval has recently shown its effectiveness.,38534
"we a the end, distribution. entropy To exploit membership-score of cluster this","To this end, we exploit the entropy of a cluster membership-score distribution.",38535
"NN addition, voting a bypasses scheme. heuristic In final the set","In addition, the final NN set bypasses a heuristic voting scheme.",38536
on Experiments several benchmark result. show datasets our retrieval method state-of-the-art the that can improve,Experiments on several retrieval benchmark datasets show that our method can improve the state-of-the-art result.,38537
of in effect transfer learning the transfer of both settings. beneficial Results indicate experiments,Results of experiments indicate the beneficial effect of transfer learning in both transfer settings.,38538
using in increase sensitivity. that gives an additional Results exam level indicate annotations,Results indicate that using exam level annotations gives an additional increase in sensitivity.,38539
the in are decision-making belief reviewed. to uncertainty function under Approaches framework,Approaches to decision-making under uncertainty in the belief function framework are reviewed.,38540
"However, bit-widths drastically degraded with quantization yields generally accuracy. decreasing","However, decreasing bit-widths with quantization generally yields drastically degraded accuracy.",38541
an essential and inherent of part Response is human actions. delay,Response delay is an inherent and essential part of human actions.,38542
"To simple the data this, support task. balancing a virtual present experimental stick from we","To support this, we present the experimental data from a simple virtual stick balancing task.",38543
architecture. problem in utilizing address mechanism routing-by We developed this our recently the agreement by,We address this problem by utilizing the recently developed routing-by agreement mechanism in our architecture.,38544
is problem important robotics. an relocalization Image-based in camera vision computer and,Image-based camera relocalization is an important problem in computer vision and robotics.,38545
final coordinates. then optimization scheme RANSAC-based a the predicted solved is via The using pose,The final pose is then solved via a RANSAC-based optimization scheme using the predicted coordinates.,38546
"to The multi-view loss which enables available also new us utilize further constraints, improve performance.","The new loss also enables us to utilize available multi-view constraints, which further improve performance.",38547
"predictions also refined. three blocks the of can Meanwhile, nearest the be","Meanwhile, the predictions of the three nearest blocks can also be refined.",38548
based optimization are cost on The weighted landmark in errors. function points these the,The landmark points are weighted in the optimization cost function based on these errors.,38549
"different Obviously, locate on semantics landmarks facial these because components. have different they","Obviously, these landmarks have different semantics because they locate on different facial components.",38550
"necessary each consider individually. is to it landmark Thus,","Thus, it is necessary to consider each landmark individually.",38551
landmarks. The weight the adaptive residuals is with estimation of,The weight is adaptive with the estimation residuals of landmarks.,38552
applications of robotic important Rotation manipulation. as dexterous is for such rigid known estimation objects,Rotation estimation of known rigid objects is important for robotic applications such as dexterous manipulation.,38553
a rank ordering Elo across method broad The dataset. plausible the produced,The Elo method produced a plausible rank ordering across the broad dataset.,38554
"of to information, those relating correlation with degree highest spatial e.g. The the were metrics","The metrics with the highest degree of correlation were those relating to spatial information, e.g.",38555
"spatial this on the variation was measured. agreement scale However, was dependent from the which","However, this agreement was dependent on the scale from which the spatial variation was measured.",38556
"lacunarity, will and many for presented including image measures entropy. also other Results be compression,","Results will also be presented for many other measures including lacunarity, image compression, and entropy.",38557
on We and method PASCAL COCO MS dataset. evaluate VOC our,We evaluate our method on VOC PASCAL and MS COCO dataset.,38558
on and parameters. structures We by conducting demonstrate several experiments SAN,We demonstrate SAN by conducting several experiments on structures and parameters.,38559
objects with smooth a interaction in for virtual environments experience. user AR/VR makes Natural,Natural interaction with virtual objects in AR/VR environments makes for a smooth user experience.,38560
achieve world to space these a to are extension interactions. augmented Gestures natural from real,Gestures are a natural extension from real world to augmented space to achieve these interactions.,38561
The of the and is amenable timescales. averaging to theory synapse assumption slow method multiple,The slow synapse assumption is amenable to averaging theory and the method of multiple timescales.,38562
demonstrate We examples. theory our two then using,We then demonstrate our theory using two examples.,38563
we In framework a modeling propose deformable unified work for this material.,In this work we propose a unified framework for modeling deformable material.,38564
explicitly network a encapsulates complex to The is neural hard function that model.,The neural network encapsulates a complex function that is hard to explicitly model.,38565
synthetic a of with set is examples. demonstrated method Our,Our method is demonstrated with a set of synthetic examples.,38566
"Generative Adversarial Network Gesture hand this a (GestureGAN). we tackle propose novel problem, To","To tackle this problem, we propose a novel hand Gesture Generative Adversarial Network (GestureGAN).",38567
"loss back-propagating while of issue proposed gradients. the color the ""channel handles pollution"" The","The proposed color loss handles the issue of ""channel pollution"" while back-propagating the gradients.",38568
experiments the FIMA. verify real on superiority of applications Extensive,Extensive experiments on real applications verify the superiority of FIMA.,38569
load with models. simulated power approach systems tested different in increment is various Our,Our approach is tested in various simulated power systems with different load increment models.,38570
collapse for approach speed cases our the of and point. show Other estimating accuracy voltage,Other cases show the accuracy and speed of our approach for estimating voltage collapse point.,38571
and approach. successfully effectiveness These testing cases the accuracy of demonstrate our,These testing cases successfully demonstrate the accuracy and effectiveness of our approach.,38572
images processing pipelines. image medical of essential for is of automation Quality assessment complete,Quality assessment of medical images is essential for complete automation of image processing pipelines.,38573
"need image an assessment there techniques. for quality urgent is automatic Therefore,","Therefore, there is an urgent need for automatic image quality assessment techniques.",38574
to methods. compare state-of-the-art range of quality We assessment our a approach,We compare our approach to a range of state-of-the-art quality assessment methods.,38575
Comparisons presented. with from ImageNet and scratch classifiers classifiers are pre-trained trained,Comparisons with ImageNet pre-trained classifiers and classifiers trained from scratch are presented.,38576
are on factors that Covariates debilitating face have influence a verification performance.,Covariates are factors that have a debilitating influence on face verification performance.,38577
both and of faces. characteristics intrinsic covariates cover factors extrinsic These subject-specific,These covariates cover both intrinsic subject-specific characteristics and extrinsic factors of faces.,38578
"more patches representative. At every smaller extracted step, the the and levels are than previous","At every step, the patches extracted are smaller than the previous levels and more representative.",38579
is to detect therefore crucial images out-of-distribution The applications. capability real-world for many,The capability to detect out-of-distribution images is therefore crucial for many real-world applications.,38580
"concept, new the RPIfield. carefully collected multi-shot dataset To person called re-id illustrate a we","To carefully illustrate the concept, we collected a new multi-shot person re-id dataset called RPIfield.",38581
is systems simulation validated in using DFRC of examples. utilization The sparse arrays,The utilization of sparse arrays in DFRC systems is validated using simulation examples.,38582
and sensing applications. task a remote challenging aerial Vehicle is detection significant in,Vehicle detection is a significant and challenging task in aerial remote sensing applications.,38583
"be our R-RPN In can model, R-DN trained and jointly.","In our model, R-RPN and R-DN can be trained jointly.",38584
with is a newly-recorded original the lip task. Manual tedious the movements of alignment speech,Manual alignment of the newly-recorded speech with the original lip movements is a tedious task.,38585
"successful well a human quantitatively, alignment both as using We perception-inspired as demonstrate metric, qualitatively.","We demonstrate successful alignment both quantitatively, using a human perception-inspired metric, as well as qualitatively.",38586
state-of-the-art methods will cases these In fail.,In these cases state-of-the-art methods will fail.,38587
the the could of system. off rate pulse be improved further trading These performances by,These performances could be further improved by trading off the pulse rate of the system.,38588
work focused cameras. has forward-facing exclusively Recent on vision automotive processing almost,Recent automotive vision work has focused almost exclusively on processing forward-facing cameras.,38589
which powerful. architecture image simple fully-convolutional yet a for propose denoising We neural-network is,We propose a fully-convolutional neural-network architecture for image denoising which is simple yet powerful.,38590
"collected code the the publish in analysis. dataset, We InstaBarcelona used and the","We publish the collected dataset, InstaBarcelona and the code used in the analysis.",38591
"In a (Neural novel this paper, propose Fitting (NBF)). Body approach we","In this paper, we propose a novel approach (Neural Body Fitting (NBF)).",38592
core-shell diffusion a modeled. Fickian geometry is into,Fickian diffusion into a core-shell geometry is modeled.,38593
shell core Langerhan interior The mimics acts as exterior islets pancreatic and inert protection. the,The interior core mimics pancreatic Langerhan islets and the exterior shell acts as inert protection.,38594
the oxygen The using consumption is diffusing approximated Michaelis-Menten cells into kinetics. of,The consumption of oxygen diffusing into the cells is approximated using Michaelis-Menten kinetics.,38595
problem to units and numerically. is transformed The solved dimensionless,The problem is transformed to dimensionless units and solved numerically.,38596
"diffusion are other limited. is regimes one Two the consumption limited that and identified,","Two regimes are identified, one that is diffusion limited and the other consumption limited.",38597
"the and hand, separating good identification scalable has On property. other loss","On the other hand, identification loss has good separating and scalable property.",38598
of material for representation the ubiquitous BRDF tools are models appearance.,BRDF models are ubiquitous tools for the representation of material appearance.,38599
"large an different However, is there number practical of models in now use. astonishingly","However, there is now an astonishingly large number of different models in practical use.",38600
"The semantic trackers. of Siamese as distractors, are the considered robustness which backgrounds hinders always","The semantic backgrounds are always considered as distractors, which hinders the robustness of Siamese trackers.",38601
"long-term we tracking. learning this focus and Siamese networks accurate In distractor-aware for paper, on","In this paper, we focus on learning distractor-aware Siamese networks for accurate and long-term tracking.",38602
"in at are end, traditional Siamese trackers this features used first. analyzed To","To this end, features used in traditional Siamese trackers are analyzed at first.",38603
the learned features makes imbalanced data observe the less distribution that of training discriminative. We,We observe that the imbalanced distribution of training data makes the learned features less discriminative.,38604
asset digital in are management power for development system. grid need Chinese of enterprises,Chinese power grid enterprises are in need for development of digital asset management system.,38605
"the have training the a priori all in known However, operations to be phase.","However, all the operations have to be known a priori in the training phase.",38606
transit consumers rail large Electric systems are energy. of,Electric rail transit systems are large consumers of energy.,38607
"substations. reversible and technologies storage systems optimization, (onboard include: and timetable energy These train wayside),","These technologies include: train timetable optimization, energy storage systems (onboard and wayside), and reversible substations.",38608
generation. approach uses planning It to a story,It uses a planning approach to story generation.,38609
pre- the the raw sensor need process data. eliminates or The post network to,The network eliminates the need to pre- or post process the raw sensor data.,38610
great supervised unsupervised challenges. and methods face Existing,Existing supervised and unsupervised methods face great challenges.,38611
this novel data be Cross-domain in utilized framework. synthetic fully could,Cross-domain synthetic data could be fully utilized in this novel framework.,38612
ensure well are transferred perception capability domains. strategies across to depth proposed Different different learned,Different strategies are proposed to ensure learned depth perception capability well transferred across different domains.,38613
dataset. show results depth state-of-the-art of on KITTI extensive experiments Our monocular estimation,Our extensive experiments show state-of-the-art results of monocular depth estimation on KITTI dataset.,38614
architecture. connectivity U-Net a We the pattern design for new,We design a new connectivity pattern for the U-Net architecture.,38615
the connections flow across efficiently information could U-Nets. coupling The more make,The coupling connections could make the information flow more efficiently across U-Nets.,38616
parameter feature makes across reuse U-Net each very U-Nets The efficient.,The feature reuse across U-Nets makes each U-Net very parameter efficient.,38617
estimation. evaluate U-Nets of datasets pose the coupled human on benchmark We two,We evaluate the coupled U-Nets on two benchmark datasets of human pose estimation.,38618
are number accuracy parameter and model compared. the Both,Both the accuracy and model parameter number are compared.,38619
accuracy CU-Net comparable obtains The state-of-the-art as methods.,The CU-Net obtains comparable accuracy as state-of-the-art methods.,38620
"the retrospective most not existing datasets consist of sequences. FLAIR do However, of","However, most of the existing retrospective datasets do not consist of FLAIR sequences.",38621
"process methods the and process imputation, of modality the segmentation. of imputation Existing missing separate","Existing missing modality imputation methods separate the process of imputation, and the process of segmentation.",38622
are segmentation. Iso-depth via regions determined superpixel,Iso-depth regions are determined via superpixel segmentation.,38623
"model. a dissertation, probabilistic two index solutions provide we In this to fit","In this dissertation, we provide two solutions to fit a probabilistic index model.",38624
entire unique into of set partitions. data consists splitting the The first algorithm,The first algorithm consists of splitting the entire data set into unique partitions.,38625
"fit the aggregate model these, then On each of and we the estimates.","On each of these, we fit the model and then aggregate the estimates.",38626
"we in the work, this address variation. weather of that is One challenges, particular","One of the particular challenges, that we address in this work, is weather variation.",38627
"low-level can not model properly. do produce that drastic Seasonal changes, features classic changes appearance","Seasonal changes can produce drastic appearance changes, that classic low-level features do not model properly.",38628
Our contributions paper in are twofold. this,Our contributions in this paper are twofold.,38629
the outperform the best Our art of field. the of results state,Our best results outperform the state of the art of the field.,38630
"procedure, time-consuming and error-prone. This is however,","This procedure, however, is time-consuming and error-prone.",38631
Using mitosis ensembling combined detector. stain invariant with a network it during in resulted training,Using it during training combined with network ensembling resulted in a stain invariant mitosis detector.,38632
issues two metric important are Feature person for and learning re-identification. representation,Feature representation and metric learning are two important issues for person re-identification.,38633
applications technology medium can find in link power reconfigurable ac-dc distribution. interesting Parallel voltage,Parallel ac-dc reconfigurable link technology can find interesting applications in medium voltage power distribution.,38634
suitable time makes applications. it for The of fast model real-time execution the,The fast execution time of the model makes it suitable for real-time applications.,38635
"overcome. nanoscale Powering with challenge major such wireless communications is supply, however, a energy to","Powering such wireless communications with nanoscale energy supply, however, is a major challenge to overcome.",38636
text-to-image heterogeneous synthesis in homogeneous lie main of and two The gaps: issues the gaps.,The main issues of text-to-image synthesis lie in two gaps: the heterogeneous and homogeneous gaps.,38637
"discriminative exploit these of we (e.g. problems, generic addressing models excellent capability the For","For addressing these problems, we exploit the excellent capability of generic discriminative models (e.g.",38638
datasets two Experiments conducted our are the to verify SDN. proposed of widely-used on effectiveness,Experiments on two widely-used datasets are conducted to verify the effectiveness of our proposed SDN.,38639
This them applications. makes for preferable real-world,This makes them preferable for real-world applications.,38640
"training require datasets. large networks such However, very","However, such networks require very large training datasets.",38641
laborious. difficult is data training Engineering the and/or,Engineering the training data is difficult and/or laborious.,38642
"As on results the obtain consequence, KITTI we benchmarks. a state-of-the-art","As a consequence, we obtain state-of-the-art results on the KITTI benchmarks.",38643
code manuscript. this The matlab available is in,The matlab code is available in this manuscript.,38644
"any bias. in and terms we As for logic approximation, efficiency subjective computational evaluate of","As for any approximation, we evaluate subjective logic in terms of computational efficiency and bias.",38645
IHC to co-registered H&E slides the were The slides.,The H&E slides were co-registered to the IHC slides.,38646
imagery airplane a Automatic in detection variety of applications. aerial has,Automatic airplane detection in aerial imagery has a variety of applications.,38647
"rotation-and-scale a solve invariant generator. present challenges, these we proposal airplane To","To solve these challenges, we present a rotation-and-scale invariant airplane proposal generator.",38648
"non-maximum suppression remove to detections. Finally, employ we duplicate","Finally, we employ non-maximum suppression to remove duplicate detections.",38649
"annotations as Moreover, an of airplanes box-level using estimate direction the we achievement. the extra","Moreover, we estimate the direction of the airplanes using box-level annotations as an extra achievement.",38650
We images. dermoscopic lesion strategy for superpixel-based on present skin segmenting a,We present a superpixel-based strategy for segmenting skin lesion on dermoscopic images.,38651
as mean superpixel RGB The used color merging criterion. of was each,The mean RGB color of each superpixel was used as merging criterion.,38652
the a framework realize is the challenge. in multi-view fusion How such primary to joint,How to realize the multi-view fusion in such a joint framework is the primary challenge.,38653
clustering DMJC-T KL are objective. a Both under optimized and DMJC-S like divergence,Both DMJC-S and DMJC-T are optimized under a KL divergence like clustering objective.,38654
learn network in an rectification We the network and end-to-end recognition manner.,We learn the rectification network and recognition network in an end-to-end manner.,38655
graph. Our model consists one four tensorflow final aggregated models single into of,Our final model consists of four single models aggregated into one tensorflow graph.,38656
a problem. two-stage this tacking We deep for framework learning present,We present a two-stage deep learning framework for tacking this problem.,38657
new such and image Our portrait extrapolation. removal portrait editing applications method enables as occlusion,Our method enables new portrait image editing applications such as occlusion removal and portrait extrapolation.,38658
several performance datasets. state-of-the-art on in results a This,This results in a state-of-the-art performance on several datasets.,38659
The knowledge representation ASP presents a extends language paper with aggregates. $\mathcal{A}log$ which,The paper presents a knowledge representation language $\mathcal{A}log$ which extends ASP with aggregates.,38660
capture high imaging frame to ultrasound Cardiac rapid motion. requires order a in rate,Cardiac ultrasound imaging requires a high frame rate in order to capture rapid motion.,38661
expense acquisition introducing block This shortens artifacts. the of at time the,This shortens the acquisition time at the expense of introducing block artifacts.,38662
"improve quality. MLA we the propose In data-driven image this learning-based to a paper, approach","In this paper, we propose a data-driven learning-based approach to improve the MLA image quality.",38663
endoscopy capsule diagnosis Wireless (WCE) gastrointestinal mean for effective an disorders. of is,Wireless capsule endoscopy (WCE) is an effective mean for diagnosis of gastrointestinal disorders.,38664
of the simplified Both MLP and computational CNN number reduce are operations. to structures,Both CNN and MLP structures are simplified to reduce the number of computational operations.,38665
convolutional regularization spatial for strategies neural (e.g. with networks),convolutional neural networks) with strategies for spatial regularization (e.g.,38666
models conditional as such fields). graphical random,graphical models such as conditional random fields).,38667
dominate the to started re-identification Deep video-based progress have methods person learning (re-id). research of,Deep learning methods have started to dominate the research progress of video-based person re-identification (re-id).,38668
"lack and real-world scalability video surveillance Therefore, they in severely practicality applications.","Therefore, they severely lack scalability and practicality in real-world video surveillance applications.",38669
employed can scheme. our readily standard CNNs DAL Existing be within,Existing standard CNNs can be readily employed within our DAL scheme.,38670
"promising in Convolutional Recently, Neural have (SR). super-resolution performance (CNNs) Networks shown","Recently, Convolutional Neural Networks (CNNs) have shown promising performance in super-resolution (SR).",38671
"branch. IRL, as master pre-trained first typical select we a In SR network a","In IRL, first we select a typical SR pre-trained network as a master branch.",38672
"comprehensive, for a for Finally, implications cost-effective malaria discuss strategy we the control.","Finally, we discuss the implications for a comprehensive, cost-effective strategy for malaria control.",38673
It component software critical based has and systems. is vision applications spread for wide a,It has wide spread applications and is a critical component for vision based software systems.,38674
"plane causes in and the printed phenomenon, misregistration, image. halo gap color artifacts This called","This phenomenon, called color plane misregistration, causes gap and halo artifacts in the printed image.",38675
algorithms embedded are for Our or designed firmware implementation. software,Our algorithms are designed for software or embedded firmware implementation.,38676
based the are first of use The algorithms two look-up tables on (LUTs).,The first two algorithms are based on the use of look-up tables (LUTs).,38677
in firmware particularly printers. low-cost the attractive of for algorithm embedded This is formatter-based implementation,This algorithm is particularly attractive for implementation in the embedded firmware of low-cost formatter-based printers.,38678
business. is dataset applied model open real the to an for And,And the model is applied to an open dataset for real business.,38679
"without with any labelled data. supervised been facial attributes, having","facial attributes, without having been supervised with any labelled data.",38680
index data of pseudo-Rsquared proposed. to new is goodness-of-fit VBGF A the models LFD for,A new pseudo-Rsquared index for the goodness-of-fit of VBGF models to LFD data is proposed.,38681
"curve and robust, bootstrap-based are for presented New, fitting methods discussed.","New, robust, bootstrap-based methods for curve fitting are presented and discussed.",38682
regularized proposed interval-valued for neural (RANN) artificial is prediction. network A data,A regularized artificial neural network (RANN) is proposed for interval-valued data prediction.,38683
"in to introduce reduce paper, method we arising MLT. this In artifacts a data-driven the","In this paper, we introduce a data-driven method to reduce the artifacts arising in MLT.",38684
lane-level accuracy for autonomous localization is very vehicles. important The,The lane-level localization accuracy is very important for autonomous vehicles.,38685
"Satellite The e.g. Global (GNSS), System Navigation","The Global Navigation Satellite System (GNSS), e.g.",38686
classification with was compared the method. The SVM performance CNN the of based method,The performance of the SVM based classification method was compared with the CNN method.,38687
from SFP scratch prune enables capacity the and Large model train to simultaneously.,Large capacity enables SFP to train from scratch and prune the model simultaneously.,38688
"pruning outperforms filter scratch SFP Empirically, the previous methods. from","Empirically, SFP from scratch outperforms the previous filter pruning methods.",38689
"effective CNN many Moreover, advanced for approach demonstrated has architectures. been our","Moreover, our approach has been demonstrated effective for many advanced CNN architectures.",38690
"separates This from avoids recovering high frequencies, denoising over-smoothing. which features low-frequency","This separates recovering low-frequency features from denoising high frequencies, which avoids over-smoothing.",38691
signed (inside/outside) The distances the samples. output balances to the curve,The output curve balances the signed distances (inside/outside) to the samples.,38692
connectivity. on parameter-free is Our fast determined the algorithm local the neighborhoods operates by and,Our algorithm is parameter-free and operates fast on the local neighborhoods determined by the connectivity.,38693
solver system by handle to bounds. We augment a least-squares linear a also constrained,We augment a least-squares solver constrained by a linear system to also handle bounds.,38694
samplers MCMC Carlo computation. an (SMC) attractive Sequential Bayesian for form alternative Monte to,Sequential Monte Carlo (SMC) samplers form an attractive alternative to MCMC for Bayesian computation.,38695
"on rejuvenate used their kernels strongly depends Markov particles. performance to However, the","However, their performance depends strongly on the Markov kernels used to rejuvenate particles.",38696
Learning problem data important vision. transformation an is visual representations of computer in invariant,Learning transformation invariant representations of visual data is an important problem in computer vision.,38697
classification for results demonstrated Deep networks have and image tasks. convolutional video remarkable,Deep convolutional networks have demonstrated remarkable results for image and video classification tasks.,38698
computation Neural achieve superior performance wider bring Deeper (CNNs) and expensive Convolutional but Networks cost.,Deeper and wider Convolutional Neural Networks (CNNs) achieve superior performance but bring expensive computation cost.,38699
Accelerating network such received increased has neural attention. over-parameterized,Accelerating such over-parameterized neural network has received increased attention.,38700
"a three-stage A retraining. algorithm pruning pipeline, and typical i.e., pruning, training, is","A typical pruning algorithm is a three-stage pipeline, i.e., training, pruning, and retraining.",38701
"update filters stage. during pruned the the retraining First, we","First, we update the pruned filters during the retraining stage.",38702
"has model this data. learn to way, In training the capacity the from enough","In this way, the model has enough capacity to learn from the training data.",38703
during at We filters training few and first the prune more filters prune procedure. asymptotically,We prune few filters at first and asymptotically prune more filters during the training procedure.,38704
ASFP classification the Experiments show on benchmarks. image our of effectiveness,Experiments show the effectiveness of our ASFP on image classification benchmarks.,38705
are latency challenges reducing LTE the when Design shown. faced in,Design challenges faced when reducing the latency in LTE are shown.,38706
short structure time of analyzed. and The frame enhancements are performance processing,The performance of short processing time and frame structure enhancements are analyzed.,38707
LTE DCI evaluated. increase Our to and channel Duplication proposed is reliability method control presented,Our proposed DCI Duplication method to increase LTE control channel reliability is presented and evaluated.,38708
models. statistical based on The is three proposed method,The proposed method is based on three statistical models.,38709
All scans data computed tomography automatically from head scans. (CT) and face derived optical are,All data are automatically derived from computed tomography (CT) head scans and optical face scans.,38710
"fit spheres. of the head model all Finally, union to we a","Finally, we fit a head model to the union of all spheres.",38711
our face CT head process scan. an anonymized recovery using validate We,We validate our face recovery process using an anonymized head CT scan.,38712
a the time cloudy of or day (e.g. in day),in a cloudy day) or time of the day (e.g.,38713
Practice Under of consideration Programming Theory (TPLP). and in Logic,Under consideration in Theory and Practice of Logic Programming (TPLP).,38714
leaps? always or can evolution make it gradual Is,Is evolution always gradual or can it make leaps?,38715
that an adaptive mutagenesis hypothesize evolvable is strategy. We stress-induced,We hypothesize that stress-induced mutagenesis is an evolvable adaptive strategy.,38716
asymptotic determine behavior and carrier-phase exact both the We positioning. for pseudo-range,We determine the exact asymptotic behavior for both pseudo-range and carrier-phase positioning.,38717
"a semantic work, we present manipulation. In this hierarchical image novel for framework","In this work, we present a novel hierarchical framework for semantic image manipulation.",38718
generator fills the in the layout. by our textures guided Then semantic image pixel-level,Then our image generator fills in the pixel-level textures guided by the semantic layout.,38719
new component a Hilbert introduced is Real-time A transformer as microwave Processing for (RAP). Analog,A microwave Hilbert transformer is introduced as a new component for Real-time Analog Processing (RAP).,38720
an inference basically an not problem optimization This problem. is,This is basically an inference problem not an optimization problem.,38721
bucket time DES supply chains. the a of and accelerate present approximate method We to,We present a time bucket method to approximate and accelerate the DES of supply chains.,38722
by mutation Synonymous is and choice mainly selection. affected Background codon,Background Synonymous codon choice is mainly affected by mutation and selection.,38723
demonstrate framework results achieve and performance. superior Experimental can methods with comparisons our state-of-the-art that,Experimental results and comparisons with state-of-the-art methods demonstrate that our framework can achieve superior performance.,38724
benefits There and potential these are applications many ideas. of,There are many potential benefits and applications of these ideas.,38725
"unnoticeable often implies failures and such unavoidable of in presence (OOD) input. However, assumption out-of-distribution","However, such assumption implies unavoidable and often unnoticeable failures in presence of out-of-distribution (OOD) input.",38726
detection discriminative this issue propose input to OOD by of We address pixels data. in,We propose to address this issue by discriminative detection of OOD pixels in input data.,38727
prediction setup. dense resolution images experiments our perform on a natural high We in,We perform our experiments on high resolution natural images in a dense prediction setup.,38728
coprime constructed arrays quadratic from The lattices proposed ideal CRT embedded integers. are by,The proposed CRT arrays are constructed by ideal lattices embedded from coprime quadratic integers.,38729
are provided of the Examples verify to feasibility methods. the proposed,Examples are provided to verify the feasibility of the proposed methods.,38730
Recognition sixth challenge. in This details paper (EmotiW) Wild Emotion the the,This paper details the sixth Emotion Recognition in the Wild (EmotiW) challenge.,38731
"detail. and discussed baselines The are databases, in protocols","The databases, protocols and baselines are discussed in detail.",38732
are language Wikipedia syllables Statistical on French of and models n-gram trained English corpora.,Statistical n-gram language models of syllables are trained on French and English Wikipedia corpora.,38733
also and We lexicon n-gram with models. syllable models compare character the,We also compare the syllable models with lexicon and character n-gram models.,38734
demonstrates images for task. of carefully-rendered synthetic intrinsic suprising work data Our the effectiveness the,Our work demonstrates the suprising effectiveness of carefully-rendered synthetic data for the intrinsic images task.,38735
the efficacy our demonstrate method. Extensive of experiments,Extensive experiments demonstrate the efficacy of our method.,38736
"passing iteratively framework, Within a and computes labels. GPNN message inference node matrices the adjacency","Within a message passing inference framework, GPNN iteratively computes the adjacency matrices and node labels.",38737
used Lens been aberrations image. provenance determine of to an previously the have,Lens aberrations have previously been used to determine the provenance of an image.,38738
via only Markov considers saliency chain nodes. detection boundaries Traditional,Traditional saliency detection via Markov chain only considers boundaries nodes.,38739
Markov boundaries The bidirectional chain. both proposed approach combines prior through foreground and cues,The proposed approach combines both boundaries and foreground prior cues through bidirectional Markov chain.,38740
"SE we an that blocks of Hence, can F-CNN part integral be architectures. believe future","Hence, we believe that SE blocks can be an integral part of future F-CNN architectures.",38741
Neural Networks Convolutional (SISR) have super-resolution problem. great (CNNs) demonstrated single-image results for the,Convolutional Neural Networks (CNNs) have demonstrated great results for the single-image super-resolution (SISR) problem.,38742
"CNN and solve to algorithms models computationally promote deep Currently, expensive SISR. most","Currently, most CNN algorithms promote deep and computationally expensive models to solve SISR.",38743
"novel we relatively propose SISR However, a of computations. number less method that uses","However, we propose a novel SISR method that uses relatively less number of computations.",38744
"connections unused get that removed. On have we training, group convolutions","On training, we get group convolutions that have unused connections removed.",38745
computations testing steps these reduce time. the significantly of number at required All,All these steps significantly reduce the number of computations required at testing time.,38746
"is input for the learning. output easier this, network with Along bicubic upsampled added to","Along with this, bicubic upsampled input is added to the network output for easier learning.",38747
step foreground video Extracting accurate from essential for an objects is a applications. scene many,Extracting accurate foreground objects from a scene is an essential step for many video applications.,38748
is novel innovation our key method of information The technique. fusion a,The key innovation of our method is a novel information fusion technique.,38749
handling complex imperfect important Such particularly background. when with integration is scenes,Such integration is particularly important when handling complex scenes with imperfect background.,38750
"built. how the framework are individual is how and We the developed, components show","We show how the framework is developed, and how the individual components are built.",38751
and experiments studies Extensive evaluate proposed are method. to conducted the ablation,Extensive experiments and ablation studies are conducted to evaluate the proposed method.,38752
"this paper, tackle framework new we to propose the above In problems. a","In this paper, we propose a new framework to tackle the above problems.",38753
MTV B-spline registration characterize the to shrinkage. used large regularized diffeomorphic was,B-spline regularized diffeomorphic registration was used to characterize the large MTV shrinkage.,38754
MTV measure Jacobian changes. to the transformation of local was resulting the computed,Jacobian of the resulting transformation was computed to measure the local MTV changes.,38755
using to HAR on focuses research due smartphones use. Recent widespread their on,Recent research on HAR focuses on using smartphones due to their widespread use.,38756
can training online first inference. both the framework that This paper presents HAR and perform,This paper presents the first HAR framework that can perform both online training and inference.,38757
broad deal vision must challenging an extremely of vehicle with set and scenes. Self-driving systems,Self-driving vehicle vision systems must deal with an extremely broad and challenging set of scenes.,38758
high that This in labeled objects with images approach are assumes accuracy.,This approach assumes that objects in images are labeled with high accuracy.,38759
"generate vehicle in video. of coherence labels the accurate we exploit the field, To spatio-temporal","To generate accurate labels in the field, we exploit the spatio-temporal coherence of vehicle video.",38760
of several the we feasibility paper In steps. approach this in demonstrate this,In this paper we demonstrate the feasibility of this approach in several steps.,38761
"we describe self-labeling scheme a using tracking. object Finally,","Finally, we describe a self-labeling scheme using object tracking.",38762
complexity generate and geometry Evolution similar in ways.,Evolution and geometry generate complexity in similar ways.,38763
The order. shape successively random pairs in were presented,The shape pairs were presented successively in random order.,38764
joint (HMC) within Monte Bayesian We Gibbs using Hamiltonian sampling. develop Carlo inference,We develop joint Bayesian inference using Hamiltonian Monte Carlo (HMC) within Gibbs sampling.,38765
hormones plant cardinal considered in Auxin is growth of and one development. the,Auxin is considered one of the cardinal hormones in plant growth and development.,38766
throughout of the a plant. range It processes wide regulates,It regulates a wide range of processes throughout the plant.,38767
pathway agrochemicals. valuable the Synthetic are auxin-signalling and herbicidal auxins exploit as,Synthetic auxins exploit the auxin-signalling pathway and are valuable as herbicidal agrochemicals.,38768
work auxin of of chemistry for the auxins. design receptor-selective This expands the range synthetic,This work expands the range of auxin chemistry for the design of receptor-selective synthetic auxins.,38769
"Logic received non-monotonic interest. In years, Programming Inductive growing recent has","In recent years, non-monotonic Inductive Logic Programming has received growing interest.",38770
"noise-tolerant of learning from we present a the In framework. paper, this sets generalisation answer","In this paper, we present a noise-tolerant generalisation of the learning from answer sets framework.",38771
AG models source the is available. code proposed publicly for The,The source code for the proposed AG models is publicly available.,38772
is of technology Artificial and transformation. the technological intelligence (AI) core industrial revolution,Artificial intelligence (AI) is the core technology of technological revolution and industrial transformation.,38773
"work, this survey intelligence. financial we In on studies existing","In this work, we survey existing studies on financial intelligence.",38774
with much spatial fewer effective CapsNet parameters. has been proved modeling features in,CapsNet has been proved effective in modeling spatial features with much fewer parameters.,38775
number is misery found The of has routing be to optimal which procedure manually.,The optimal number of routing procedure is misery which has to be found manually.,38776
"implement We ""convolutional"" of version ""full-connected"" and both G-CapsNet of version G-CapsNet.","We implement both ""full-connected"" version of G-CapsNet and ""convolutional"" version of G-CapsNet.",38777
achieves similar MNIST the as papers. dataset a G-CapsNet performance in the in original,G-CapsNet achieves a similar performance in the dataset MNIST as in the original papers.,38778
"layers. capsule stacking possibility we multiple of also Besides, explored","Besides, we also explored possibility of stacking multiple capsule layers.",38779
on modular detailed a ontology We ontology patterns. based modeling for provide example design,We provide a detailed example for modular ontology modeling based on ontology design patterns.,38780
that results outperforms the Experimental method methods. proposed conventional reveal demosaicking,Experimental results reveal that the proposed method outperforms conventional demosaicking methods.,38781
get feature better employing representations by learn pyramids. usually multi-scale object detectors State-of-the-art results to,State-of-the-art object detectors usually learn multi-scale representations to get better results by employing feature pyramids.,38782
"in-place, end-to-end reconfiguration and the local global attention lightweight, and Both trainable. are","Both the global attention and local reconfiguration are lightweight, in-place, and end-to-end trainable.",38783
using and build learning detect classify model heart deep a We disease to $X-ray$.,We build a deep learning model to detect and classify heart disease using $X-ray$.,38784
datasets. data We several from collect and hospitals public,We collect data from several hospitals and public datasets.,38785
management current conservation. tourism biodiversity achieving approaches Our are not,Our current tourism management approaches are not achieving biodiversity conservation.,38786
insulin and sensitivity. of Peroxisome gamma proliferator-activated controllers ligand-activated (PPAR{\gamma}) metabolic are various actions receptors,Peroxisome proliferator-activated receptors gamma (PPAR{\gamma}) are ligand-activated controllers of various metabolic actions and insulin sensitivity.,38787
is (S-AOG) indoor And-Or to spatial attributed An graph proposed scenes. represent,An attributed spatial And-Or graph (S-AOG) is proposed to represent indoor scenes.,38788
"in are nodes which grammar the a The terminal S-AOG model, entities. is object probabilistic","The S-AOG is a probabilistic grammar model, in which the terminal nodes are object entities.",38789
"required. So, pixel pixel no is calculation by complex","So, no complex pixel by pixel calculation is required.",38790
through different Potential tested is features medical dataset. and established of proposed,Potential of proposed features is tested and established through different medical dataset.,38791
in and vision applications tasks. wide have Dense cues depth important computer various are,Dense depth cues are important and have wide applications in various computer vision tasks.,38792
"maps obtained hardware its sparse depth generally limitation. of However, by LIDAR are because","However, depth maps obtained by LIDAR are generally sparse because of its hardware limitation.",38793
depth incorporated the improve could be to performance. RGB features further completion Additional,Additional RGB features could be incorporated to further improve the depth completion performance.,38794
"there growing for to is a moving automatically need Accordingly, techniques images. classify","Accordingly, there is a growing need for techniques to automatically classify moving images.",38795
processes hybridization. Phylogenetic such networks for to trees modeling allow as phylogenetic reticulate extend evolutionary,Phylogenetic networks extend phylogenetic trees to allow for modeling reticulate evolutionary processes such as hybridization.,38796
parallel in discuss structures. population the genetics efforts inferring community at aimed similar also We,We also discuss parallel efforts in the population genetics community aimed at inferring similar structures.,38797
"we research Finally, highlight major for directions area. in future this","Finally, we highlight major directions for future research in this area.",38798
and important recently extensively. investigated issue is an been has removal Rain streak,Rain streak removal is an important issue and has recently been investigated extensively.,38799
streak its blur first the rain interference with model motion mechanism. We,We first model the rain streak interference with its motion blur mechanism.,38800
image This a dehazing. single for proposes technique paper novel,This paper proposes a novel technique for single image dehazing.,38801
method approaches. The combines two the different proposed,The proposed method combines the two different approaches.,38802
"i.e., sequence of a the Given vectors. representation, pair, matrix a a generates video-description network","Given a video-description pair, the network generates a matrix representation, i.e., a sequence of vectors.",38803
vector a description. video conditioned represents in Each by frame matrix the the,Each vector in the matrix represents a video frame conditioned by the description.,38804
robust yields the a for representation multi-level multiple Integration video-language abstraction. granularities of,Integration of the multiple granularities yields a robust representation for multi-level video-language abstraction.,38805
of validate large-scale effectiveness two We ASST datasets. on our the,We validate the effectiveness of our ASST on two large-scale datasets.,38806
pose to key a intelligent assistance is Person context-aware and systems. detection develop requirement estimation,Person detection and pose estimation is a key requirement to develop intelligent context-aware assistance systems.,38807
"occlusions also and challenges includes in as such visual environments, It present clutter. the such","It also includes the visual challenges present in such environments, such as occlusions and clutter.",38808
have be convolutional architectures. been shown model a recent competitive to among DenseNets network,DenseNets have been shown to be a competitive model among recent convolutional network architectures.,38809
"introduce early We new a three inference. constructs allow which that ParaNet, architecture for pipelines","We introduce ParaNet, a new architecture that constructs three pipelines which allow for early inference.",38810
leading among of death are Cardiovascular causes the globally. diseases,Cardiovascular diseases are among the leading causes of death globally.,38811
We area measures. introduced the on novel distortion scores based spiculation spiculation metric and,We introduced novel spiculation scores based on the area distortion metric and spiculation measures.,38812
model malignancy introduced. prediction also pathological A is simple,A simple pathological malignancy prediction model is also introduced.,38813
"quantifying We parameter-free interpretable, spiculations nodules. technique developed on a and reproducible for","We developed a reproducible and interpretable, parameter-free technique for quantifying spiculations on nodules.",38814
location-discriminating We method having learning representation propose higher a novel power.,We propose a novel representation learning method having higher location-discriminating power.,38815
is by gradients. analyzing better why insights SARE It provides into,It provides insights into why SARE is better by analyzing gradients.,38816
CNN. any to SARE easy is and implement Our pluggable to loss,Our SARE loss is easy to implement and pluggable to any CNN.,38817
aims navigation at Indoor navigation performing within buildings.,Indoor navigation aims at performing navigation within buildings.,38818
"we itself scenarios, In an in unseen environment. intelligent most expected navigating of capable device","In most scenarios, we expected an intelligent device capable of navigating itself in unseen environment.",38819
sort traces a fingerprint. Such represent of camera,Such traces represent a sort of camera fingerprint.,38820
tasks. breakthrough techniques perceptive are Deep results image-based obtaining in Recent these (DL) Learning,Recent image-based Deep Learning (DL) techniques are obtaining breakthrough results in these perceptive tasks.,38821
lot Training data. good of deep a a learning annotated requires often model,Training a good deep learning model often requires a lot of annotated data.,38822
images. can panoramic on detected be ACPs,ACPs can be detected on panoramic images.,38823
"methods, These used barely dentistry. however in been have","These methods, however have been barely used in dentistry.",38824
used study R-CNN) we Network In Faster the deep this Region-based Convolutional learning. (Faster for,In this study we used the Faster Region-based Convolutional Network (Faster R-CNN) for deep learning.,38825
Receiver calculate was performed accuracy detection. Characteristic the Operating to (ROC) analysis of,Receiver Operating Characteristic (ROC) analysis was performed to calculate the accuracy of detection.,38826
on growing. of capturing dynamic scenes environments rapidly are underwater Demands,Demands on capturing dynamic scenes of underwater environments are rapidly growing.,38827
complex on The state-of-the-art the colorectal dataset. show results performance experimental adenocarcinoma,The experimental results show the state-of-the-art performance on complex colorectal adenocarcinoma dataset.,38828
during are imperfections unknown or process. calibration measurement there when cases Such parameters the exist,Such cases exist when there are imperfections or unknown calibration parameters during the measurement process.,38829
version of The is performance numerically. robust VAMP of the demonstrated,The performance of the robust version of VAMP is demonstrated numerically.,38830
attack for are introduced this paper. Three in discriminative detection face representations presentation,Three discriminative representations for face presentation attack detection are introduced in this paper.,38831
"two also of design kinds detection, attack accurate presentation we combinations. representation For","For accurate presentation attack detection, we also design two kinds of representation combinations.",38832
"Firstly, SSD. cascade combine decision-level we a with to strategy SPMT propose","Firstly, we propose a decision-level cascade strategy to combine SPMT with SSD.",38833
weaker interferers. in suppres- the detection strong of signals a of This improves presence sion,This suppres- sion improves detection of a weaker signals in the presence of strong interferers.,38834
"deep Iizuka present based a we on Accordingly, et approach learning al.","Accordingly, we present a deep learning approach based on Iizuka et al.",38835
network boundaries. adversarially training to for a hallucinate image past,for adversarially training a network to hallucinate past image boundaries.,38836
both outpainting Our that results feasible to deep show approaches are promising. image and learning,Our results show that deep learning approaches to image outpainting are both feasible and promising.,38837
in strategy training. This considering more testing in than helps when classes,This strategy helps when considering more classes in testing than in training.,38838
show environments. also not sacrificed is rich the We in performance scattering that,We also show that the performance is not sacrificed in rich scattering environments.,38839
phenomenon. available enough to generalize The intelligent not are complex this techniques,The available techniques are not intelligent enough to generalize this complex phenomenon.,38840
learning mass of the in mammograms. for segmentation deep explore We use breast,We explore the use of deep learning for breast mass segmentation in mammograms.,38841
"tedious techniques nor Moreover, required are our pre-processing post-processing not neither algorithm. in","Moreover, neither tedious pre-processing nor post-processing techniques are not required in our algorithm.",38842
is food preferences events. of the people One interaction with their of,One of the preferences of people is their interaction with food events.,38843
"The an proposed dataset is in-house private on evaluated ""EgoFoodPlaces"". called model","The proposed model is evaluated on an in-house private dataset called ""EgoFoodPlaces"".",38844
places of promising food recognition photo-streams. shows Experimental results egocentric in results classification,Experimental results shows promising results of food places classification recognition in egocentric photo-streams.,38845
are to Humans movements. innately others' emotional from prepared subtle arguably comprehend body expressions,Humans are arguably innately prepared to comprehend others' emotional expressions from subtle body movements.,38846
insights. dataset the revealed analysis interesting statistical many of Comprehensive,Comprehensive statistical analysis of the dataset revealed many interesting insights.,38847
"far, features, based on So analysis of is mostly this kind radiomics i.e.","So far, this kind of analysis is mostly based on radiomics features, i.e.",38848
features priori. a defined that of mathematically fixed a set is,a fixed set of features that is mathematically defined a priori.,38849
has improved knowledge domain text. Incorporating understanding machine of,Incorporating domain knowledge has improved machine understanding of text.,38850
"architecture this i.e. estimate a to end, CNN is employed To parametric representations","To this end, a CNN architecture is employed to estimate parametric representations i.e.",38851
"pose, hand shape parameters. bone and scales complex","hand pose, bone scales and complex shape parameters.",38852
"data manually real annotating is Also, suboptimal.","Also, manually annotating real data is suboptimal.",38853
The last section. formulae given in are updating the,The updating formulae are given in the last section.,38854
integrating the stereo We factor constraints SfM camera exploiting by results graphs. initial refine,We refine the initial SfM results by integrating stereo camera constraints exploiting factor graphs.,38855
object compute by camera pose background trajectory We information. combining the object and,We compute the object trajectory by combining object and background camera pose information.,38856
"method monocular As our trajectory opposed reconstruction degenerated to no approaches, shows cases.","As opposed to monocular trajectory reconstruction approaches, our method shows no degenerated cases.",38857
approach data available vehicles of We scenes. using in publicly video our evaluate urban,We evaluate our approach using publicly available video data of vehicles in urban scenes.,38858
"(MC-DART), In deals multi-channel Multi-Channel with paper effectively we this DART present which data.","In this paper we present Multi-Channel DART (MC-DART), which deals effectively with multi-channel data.",38859
is communication. wireless resource an frequency The essential for spectrum,The frequency spectrum is an essential resource for wireless communication.,38860
Cognitive the approach Radio solution new comes with (CR) scarcity. to The spectrum feasible,The new Cognitive Radio (CR) approach comes with the feasible solution to spectrum scarcity.,38861
task networks estimating aerial tackles images. of of from road paper the the topology This,This paper tackles the task of estimating the topology of road networks from aerial images.,38862
"AI. field light We of direction assess omics and of the the Biobanks/PheWAS, atlases, in","We assess the direction of the field in light of Biobanks/PheWAS, omics atlases, and AI.",38863
detection self-driving Lane vehicles. is for important very,Lane detection is very important for self-driving vehicles.,38864
"algorithm minimal and was for operation very Furthermore, the run-time. fast designed","Furthermore, the algorithm was designed for very fast operation and minimal run-time.",38865
"of global mechanics, stationarity the the minimum. with In corresponds condition unlike particular, action the","In particular, unlike the action of mechanics, the stationarity condition corresponds with the global minimum.",38866
"in problem is of experimented theory computer extraction the Finally, feature the vision. for","Finally, the theory is experimented for the problem of feature extraction in computer vision.",38867
for representation a Boltzmann (RBM) learning. restricted investigate discriminative the Machine We of potential,We investigate the potential of a restricted Boltzmann Machine (RBM) for discriminative representation learning.,38868
Experimental effectiveness show the on set benchmark of results method. the our data,Experimental results on the benchmark data set show the effectiveness of our method.,38869
in novel context. is proposed here method set active The this,The active set method proposed here is novel in this context.,38870
"artifacts main and described characterized. the are this aim, mathematically recurrent To four","To this aim, the four main recurrent artifacts are described and mathematically characterized.",38871
we and formula. design artifacts measurement attached a these its each For of test-pattern,For each of these artifacts we design a test-pattern and its attached measurement formula.,38872
"contrast, coverage. In accurate yields the interval bootstrap reversed-tails","In contrast, the reversed-tails bootstrap interval yields accurate coverage.",38873
the also We recursive-design bootstrap with theoretically the bootstrap. compare analyzed fixed-design,We also compare the theoretically analyzed fixed-design bootstrap with the recursive-design bootstrap.,38874
the empirical estimation. An application illustrates interval,An empirical application illustrates the interval estimation.,38875
Camera of a image are fingerprints forensics number for precious tasks. tools,Camera fingerprints are precious tools for a number of image forensics tasks.,38876
"well-known example photo the noise pattern, (PRNU) fingerprint. is A a response device non-uniformity powerful","A well-known example is the photo response non-uniformity (PRNU) noise pattern, a powerful device fingerprint.",38877
"affected only by mildly of is residuals content. the However, it high-level PRNU, unlike scene","However, unlike the PRNU, it is only mildly affected by residuals of high-level scene content.",38878
proposed the The experiments localization the method improves over reference. PRNU-based noiseprint-based that show forgery,The experiments show that the proposed noiseprint-based forgery localization method improves over the PRNU-based reference.,38879
a of multilingual handwriting for address the system recognition. design We unified,We address the design of a unified multilingual system for handwriting recognition.,38880
complexity. size reduces the the Dealing with decreases thus model language multigrams strongly lexicon and,Dealing with multigrams strongly reduces the lexicon size and thus decreases the language model complexity.,38881
identity been applied economic in widely Kotlarski's used research. has,Kotlarski's identity has been widely used in applied economic research.,38882
for motorcycle Neural system a classification This using paper Convolutional Network urban presents scenarios (CNN).,This paper presents a motorcycle classification system for urban scenarios using Convolutional Neural Network (CNN).,38883
"Nevertheless, extracted trained. be CNNs already features can from","Nevertheless, features can be extracted from CNNs already trained.",38884
classification. on images performed satisfactory validation set are a showing Further experiments of a,Further experiments are performed on a validation set of images showing a satisfactory classification.,38885
network is The weak supervision. entire with trainable end-to-end only,The entire network is end-to-end trainable with only weak supervision.,38886
dosage tablets. forms benefits can Multi-particulate over oral conventional and significant capsules offer,Multi-particulate oral dosage forms can offer significant benefits over conventional capsules and tablets.,38887
and vitro reported in swallow Both for smaller easier multi-particulates. vivo in tests,Both in vivo and in vitro tests reported easier swallow for smaller multi-particulates.,38888
"oral solids. for optimal appeared of water complete clearance Besides, not thin the liquids","Besides, water thin liquids appeared not optimal for complete oral clearance of the solids.",38889
A convolutional network is these trained attributes. (CNN) learn to neural,A convolutional neural network (CNN) is trained to learn these attributes.,38890
the This Autonomous Over paper Architecture describes project. Domains Restricted,This paper describes the Autonomous Architecture Over Restricted Domains project.,38891
state-of-the-art results of networks. examine and the models deep existing modified neural We,We examine the results of existing state-of-the-art models and modified deep neural networks.,38892
"this the for solving competition we first task. Finally, explain in published detail","Finally, we explain in detail the first published competition for solving this task.",38893
"its scene a stage. in However, to is still preliminary fully social comprehend the ability","However, the ability to fully comprehend a social scene is still in its preliminary stage.",38894
in guide gaze contextual human the to weakly-supervised a also regions It utilizes attended setting.,It also utilizes human gaze to guide the attended contextual regions in a weakly-supervised setting.,38895
"strategy In mis-grouping. we to hard sampling address negative of problem propose a addition, the","In addition, we propose a hard negative sampling strategy to address the problem of mis-grouping.",38896
"on extensive namely We datasets, perform experiments V-COCO benchmark HICO-DET. two and","We perform extensive experiments on two benchmark datasets, namely V-COCO and HICO-DET.",38897
component efficacy each proposed The been has validated. also of,The efficacy of each proposed component has also been validated.,38898
is receiving link only. State Information available side Channel each the (CSI) of at,Channel State Information (CSI) is available at the receiving side of each link only.,38899
"during Time transmission Multiplexing Division each is time-slot. a assumed, occurs where orthogonal","Time Division Multiplexing is assumed, where each orthogonal transmission occurs during a time-slot.",38900
transmit first turns phase. transmission Sources time in in the during slots consecutive,Sources transmit in turns in consecutive time slots during the first transmission phase.,38901
"HARQ, Incremental the two approach. the three Among proposed follow (IR) Redundancy","Among the three proposed HARQ, two follow the Incremental Redundancy (IR) approach.",38902
"localizing face in However, attributes. different these corresponding difficulties methods to areas the","However, these methods face difficulties in localizing the areas corresponding to different attributes.",38903
datasets. Our on surpasses previous the in all five metrics state-of-the-art result both,Our result surpasses the previous state-of-the-art in all five metrics on both datasets.,38904
Traditionally a defined are fixed axiomatic universe. tests for,Traditionally axiomatic tests are defined for a fixed universe.,38905
index thus potentially family the all outline large tests. that We can satisfy a,We thus outline a large index family that potentially can satisfy all the tests.,38906
some techniques errors. inevitably these But run into,But these techniques inevitably run into some errors.,38907
of among comparisons LSTM extensive also multiple models. made variants We,We also made extensive comparisons among multiple variants of LSTM models.,38908
"paper, hybrid system In this FSO multi-hop / multi-user presented. RF is a","In this paper, a multi-user multi-hop hybrid FSO / RF system is presented.",38909
of consisted structure parts. is This two main,This structure is consisted of two main parts.,38910
part proposed This the the is scenario first exactly implemented structure. of at,This scenario is exactly implemented at the first part of the proposed structure.,38911
users number that number of on it low Results indicate has and of dependence relays.,Results indicate that it has low dependence on number of users and number of relays.,38912
"it long-range suitable for population links. is Therefore, and with varying areas","Therefore, it is suitable for areas with varying population and long-range links.",38913
"processing offers performance additional independent latency, without This and power consumption, structure complexity.","This structure offers independent performance without additional power consumption, processing latency, and complexity.",38914
of three sequential sists It modules. con-,It con- sists of three sequential modules.,38915
Traffic systems reduce Recognizing can Signs happening intelligent number of world-wide. the using drastically accidents,Recognizing Traffic Signs using intelligent systems can drastically reduce the number of accidents happening world-wide.,38916
per-user account maximum power takes model the and Our power limitation the control. of,Our model takes account of the maximum power limitation and the per-user power control.,38917
effect have and predict to commonsense of a ability collisions. the physical use Humans remarkable,Humans have a remarkable ability to use physical commonsense and predict the effect of collisions.,38918
underlying factors? do But the understand they,But do they understand the underlying factors?,38919
underlying factors have the if Can predict changed? they,Can they predict if the underlying factors have changed?,38920
test of and of unseen combinations shapes. collisions on,and test on collisions of unseen combinations of shapes.,38921
imaging the is photoacoustic upon an Photoacoustic modality based is emerging imaging that effect.,Photoacoustic imaging is an emerging imaging modality that is based upon the photoacoustic effect.,38922
"methods recovery respect with performance target enhanced tensor-OMP. completion show tensor Furthermore, to in off-grid","Furthermore, tensor completion methods show enhanced performance in off-grid target recovery with respect to tensor-OMP.",38923
our provided correctness analysis. Monte the throughout of to are Carlo verify simulations,Monte Carlo simulations are provided throughout to verify the correctness of our analysis.,38924
apnea abnormal caused by sleep sleep a common Obstructive disorder is breathing. (OSA),Obstructive sleep apnea (OSA) is a common sleep disorder caused by abnormal breathing.,38925
sudden to symptoms death lead OSA (SCD). cardiac The as severity many such can of,The severity of OSA can lead to many symptoms such as sudden cardiac death (SCD).,38926
a diagnosis. standard is gold Polysomnography (PSG) for OSA,Polysomnography (PSG) is a gold standard for OSA diagnosis.,38927
patients value then into to used is OSA This severity levels. classify,This value is then used to classify patients into OSA severity levels.,38928
"and limitations. many has However, disadvantages it","However, it has many disadvantages and limitations.",38929
"of a classification we a novel OSA using proposed Learning severity approach. Consequently, Deep methodology","Consequently, we proposed a novel methodology of OSA severity classification using a Deep Learning approach.",38930
extensively the validate the on method dataset. proposed INbreast widely-used We,We extensively validate the proposed method on the widely-used INbreast dataset.,38931
arena efficiently how of probability compute Emblem in Fire demonstrate to We battles. the victory,We demonstrate how to efficiently compute the probability of victory in Fire Emblem arena battles.,38932
"are and issues implementation a code. implementation full provided is in Some addressed,","Some implementation issues are addressed, and a full implementation is provided in code.",38933
"present a mapping. augmentation technique novel on we feature geometry a Further, parametric based spherical","Further, we present a novel geometry feature augmentation technique based on a parametric spherical mapping.",38934
all and VGGNet of outperforms Inception ResNet for our predictive tasks.,ResNet outperforms VGGNet and Inception for all of our predictive tasks.,38935
years detecting in Recent object recognizing instances. individual witnessed rapid progress and have,Recent years have witnessed rapid progress in detecting and recognizing individual object instances.,38936
"the detecting task this paper, In we of challenging (HOI). interactions human-object tackle","In this paper, we tackle the challenging task of detecting human-object interactions (HOI).",38937
an recognizing allows Such HOIs. us to attention-based for aggregate relevant selectively features network,Such an attention-based network allows us to selectively aggregate features relevant for recognizing HOIs.,38938
certain non-circular problem encountered blind signal separation. Such including analysis in can be applications source,Such problem can be encountered in certain non-circular signal analysis applications including blind source separation.,38939
algorithms of Givens a using and Givens rotations. We new introduce Jacobi-like or combination hyperbolic,We introduce new Jacobi-like algorithms using Givens or a combination of Givens and hyperbolic rotations.,38940
driving many algorithms autonomous the for functions. basis automated or are Such state-of-the-art,Such algorithms are the basis for many state-of-the-art autonomous or automated driving functions.,38941
of filter. the (SG) different in used widely One domains denoising the methods Savitzky-Golay is,One of the widely used denoising methods in different domains is the Savitzky-Golay (SG) filter.,38942
window filter filter length and has SG parameters: The design order. two the,The SG filter has two design parameters: window length and the filter order.,38943
includes and criteria. square error Mean (MSE) both variance measure bias,Mean square error (MSE) measure includes both bias and variance criteria.,38944
model. conventional inversely formation algorithms photometric a BRDF-based stereo image solve Most,Most conventional photometric stereo algorithms inversely solve a BRDF-based image formation model.,38945
"we need so, primary In analyze to network do order the carefully. to","In order to do so, we need to analyze the primary network carefully.",38946
model to adversarials. a CNN introduce We based on MILP create,We introduce a MILP model based on CNN to create adversarials.,38947
image and network compare for recognition evaluate each We tasks.,We compare and evaluate each network for image recognition tasks.,38948
This dataset speech for visual multi-modal paper and new recognition. audio-visual introduces a,This paper introduces a new multi-modal dataset for visual and audio-visual speech recognition.,38949
is unusable These after iris decease. a conclusions contradict shortly promulgated rumor that,These conclusions contradict a promulgated rumor that iris is unusable shortly after decease.,38950
images a during The near light and routine practice. in infrared visible ophthalmological were collected,The images were collected in near infrared and visible light during a routine ophthalmological practice.,38951
The study shows experimental valuable four results.,The experimental study shows four valuable results.,38952
recently. attentions in environments caught high-mobility a have lot Communications of,Communications in high-mobility environments have caught a lot of attentions recently.,38953
"are paper, multiple-input (MIMO) massive addressed. fast multiple-output channels for In this time-varying systems","In this paper, fast time-varying channels for massive multiple-input multiple-output (MIMO) systems are addressed.",38954
spread parameter a the closed derived Doppler is in CCAP form. The optimal minimizing,The optimal CCAP parameter minimizing the Doppler spread is derived in a closed form.,38955
"details accompanied the unpleasant hallucinated However, often artifacts. with are","However, the hallucinated details are often accompanied with unpleasant artifacts.",38956
"logic base the to non-monotonic. However, not does be need","However, the base logic does not need to be non-monotonic.",38957
"for structured Indeed, (e.g. logic proposals a base argumentation monotonic use most","Indeed, most proposals for structured argumentation use a monotonic base logic (e.g.",38958
"form language, logic). with modus rule-based or ponens of classical a some","some form of modus ponens with a rule-based language, or classical logic).",38959
tools to logics. research and there are non-monontonic harnessed And insights be for in,And there are insights and tools to be harnessed for research in non-monontonic logics.,38960
this some issues in these paper. of We consider,We consider some of these issues in this paper.,38961
in collected The routine light were infrared examination. visible ophthalmological images during and near,The images were collected in near infrared and visible light during routine ophthalmological examination.,38962
prediction. This approach saliency novel perspective provides a achieving for,This approach provides a novel perspective for achieving saliency prediction.,38963
finding architectures works Most creatively models designing parameters. aim and fine-tuning at such by,Most works aim at finding such models by creatively designing architectures and fine-tuning parameters.,38964
This Interferometry HyperSpectral introduces (HS) Fourier framework imaging (FTI). a paper based single-pixel Transform on,This paper introduces a single-pixel HyperSpectral (HS) imaging framework based on Fourier Transform Interferometry (FTI).,38965
space-time us time-modulated to coding light using patterns. allows optimize the Hadamard This,This allows us to optimize the space-time light coding using time-modulated Hadamard patterns.,38966
by a our the of approach few confirm effectiveness numerical We experiments.,We confirm the effectiveness of our approach by a few numerical experiments.,38967
open closed discussed. are systems Both and,Both closed and open systems are discussed.,38968
is and illustrated The trans-membrane reaction transporter. using a a simple method enzyme-catalysed,The method is illustrated using a simple enzyme-catalysed reaction and a trans-membrane transporter.,38969
addresses problem fragments. reassembling paper the of This from disjointed images,This paper addresses the problem of reassembling images from disjointed fragments.,38970
variable. We projected each build Vietoris-Rips complex onto from the data-cloud the,We build the Vietoris-Rips complex from the data-cloud projected onto each variable.,38971
package TopSA made an Sensitivy own Analysis). the (Topological We called with analysis,We made the analysis with an own package called TopSA (Topological Sensitivy Analysis).,38972
will researchers. This publicly be all available to database,This database will be publicly available to all researchers.,38973
"very data sets. for Therefore, well large scales it","Therefore, it scales well for very large data sets.",38974
the proposed of Experimental potential approach. the demonstrate results,Experimental results demonstrate the potential of the proposed approach.,38975
Our the in from aspects. SISR main differs optical general two problem super-resolution(OFSR) flow problem,Our optical flow super-resolution(OFSR) problem differs from the general SISR problem in two main aspects.,38976
"rainforests suffering widespread and both damage, natural means. have been The via Amazon artificial","The Amazon rainforests have been suffering widespread damage, both via natural and artificial means.",38977
of of subset exponential This random studies paper problem signals from samples. the recovering a,This paper studies the problem of recovering exponential signals from a random subset of samples.,38978
data. HVaF spectroscopy biological resonance magnetic further The is effectiveness on validated of,The effectiveness of HVaF is further validated on biological magnetic resonance spectroscopy data.,38979
"error (MSE) square estimation performance inter-subarray mean Next, under analyzed mismatches. is its","Next, its estimation mean square error (MSE) performance is analyzed under inter-subarray mismatches.",38980
"is Cramer-Rao of derived. the lower Moreover, bound CFO estimation","Moreover, the Cramer-Rao lower bound of CFO estimation is derived.",38981
to the corroborate method. results and proposed theoretical provided numerical are Both,Both theoretical and numerical results are provided to corroborate the proposed method.,38982
in affected non-match false two genuine final This increase matchers. in the rate scores,This increase in genuine scores affected the final false non-match rate in two matchers.,38983
prior identity racing. for is horses to of Identification crucial confirmation race animal,Identification of race horses is crucial for animal identity confirmation prior to racing.,38984
"approach. this (CNN) human trajectory based we convolutional work, propose a In prediction neural network","In this work, we propose a convolutional neural network (CNN) based human trajectory prediction approach.",38985
"Finally, fusion used is optimization iterative an model. to the algorithm solve","Finally, an iterative optimization algorithm is used to solve the fusion model.",38986
may similarity factors contribute and quality geometrical of Image degradation the to score.,Image quality and geometrical factors may contribute to the degradation of similarity score.,38987
"by blurring. input super-resolution and low-resolution formulation, Different noises further is degraded the traditional from","Different from traditional super-resolution formulation, the low-resolution input is further degraded by noises and blurring.",38988
makes learning estimation kernel setting supervised complicated accurate impossible. This and,This complicated setting makes supervised learning and accurate kernel estimation impossible.,38989
"First, the space. noisy noise-free to and blurry input mapped a is low-resolution","First, the noisy and blurry input is mapped to a noise-free low-resolution space.",38990
intermediate pre-trained is deep up-sampled with model. image a the Then,Then the intermediate image is up-sampled with a pre-trained deep model.,38991
"we an in modules end-to-end to fine-tune the output. manner high-resolution two Finally, get the","Finally, we fine-tune the two modules in an end-to-end manner to get the high-resolution output.",38992
Word many images in visual assistance. scene spotting natural understanding has applications and scene in,Word spotting in natural scene images has many applications in scene understanding and visual assistance.,38993
by team system present We a large-scale PFDet. object detection,We present a large-scale object detection system by team PFDet.,38994
features spatial-temporal video analysis are tasks. for Learning many critical long-term,Learning long-term spatial-temporal features are critical for many video analysis tasks.,38995
dense and requires quadratic Considering that consumption matrix w.r.t. estimation the relation computation overhead memory,Considering that the dense relation matrix estimation requires quadratic computation overhead and memory consumption w.r.t.,38996
finding is of discriminative to features. difficulty the challenging classification due Fine-grained,Fine-grained classification is challenging due to the difficulty of finding discriminative features.,38997
object fully characterize traits Finding not is the subtle that those straightforward.,Finding those subtle traits that fully characterize the object is not straightforward.,38998
"scrutinizes After makes Scrutinizer proposed the the regions from predictions. Navigator and that,","After that, the Scrutinizer scrutinizes the proposed regions from Navigator and makes predictions.",38999
We benchmark performance state-of-the-art datasets. in achieve extensive,We achieve state-of-the-art performance in extensive benchmark datasets.,39000
iris recognition. of paper post-mortem This analysis presents unique a human,This paper presents a unique analysis of post-mortem human iris recognition.,39001
"tracking a In using a optimally we as approach. practice, path problem solve K-shortest this","In practice, we solve this optimally as a tracking problem using a K-shortest path approach.",39002
these are of relation Insightful all conclusions to aspects. offered three in,Insightful conclusions are offered in relation to all three of these aspects.,39003
"Thus, has size departments. railway from increasing EMU associated fleet the attention attracted prediction of","Thus, the prediction of EMU fleet size has attracted increasing attention from associated railway departments.",39004
in two networks analyzed differences of computation The and performances are these experiments in-depth.,The differences and performances of these two networks in computation experiments are analyzed in-depth.,39005
of is object nearby an among symmetries and among Distinction poses challenging.,Distinction among nearby poses and among symmetries of an object is challenging.,39006
"In we group-theoretic both. a propose unified, this to paper, approach tackle","In this paper, we propose a unified, group-theoretic approach to tackle both.",39007
to We matrices. sample Metropolis-Hastings from algorithm space propose the correlation of uniformly a novel,We propose a novel Metropolis-Hastings algorithm to sample uniformly from the space of correlation matrices.,39008
no from However or other populations. population fully animal human human is isolated,However no human population is fully isolated from other human or animal populations.,39009
"neural improvements deep object by to have Driven pose brought networks, methods recent estimation. significant","Driven by deep neural networks, recent methods have brought significant improvements to object pose estimation.",39010
"the have approach. encourage research, To we reproducible proposed for released codes our","To encourage reproducible research, we have released the codes for our proposed approach.",39011
features. Convolutional starting learning neural networks are local to able with hierarchical process perform a,Convolutional neural networks are able to perform a hierarchical learning process starting with local features.,39012
"features However, attention paid a to is like limited level elementary enhancing edges. such","However, a limited attention is paid to enhancing such elementary level features like edges.",39013
transmission. of misclassification community of community-onset role underplays the and cases Underreporting systematic,Underreporting of community-onset cases and systematic misclassification underplays the role of community transmission.,39014
"Harmonic, by Heronian Geometric, means have been Arithmetic, and Contraharmonic mathematicians. many studied","Harmonic, Geometric, Arithmetic, Heronian and Contraharmonic means have been studied by many mathematicians.",39015
on image superior Convolutional performance recent Neural depth single have in Networks years. estimation demonstrated,Convolutional Neural Networks have demonstrated superior performance on single image depth estimation in recent years.,39016
The combines dilated ResNet and DFE convolutions.,The DFE combines ResNet and dilated convolutions.,39017
input from multi-scale extracts keeping while information maps feature image It dense. the,It extracts multi-scale information from input image while keeping the feature maps dense.,39018
"attention DMG, we features multi-scale DFE. As to for in mechanism fuse use produced","As for DMG, we use attention mechanism to fuse multi-scale features produced in DFE.",39019
is any does Network need Our and trained post-processing. not end-to-end,Our Network is trained end-to-end and does not need any post-processing.,39020
"this an domain brings option. appealing adaptation setting, In","In this setting, domain adaptation brings an appealing option.",39021
give Conservative studies Ablation the insights more properties into Loss. of,Ablation studies give more insights into properties of the Conservative Loss.,39022
method a images introduce learning perturbations for adversarial to or individual targeted We videos.,We introduce a method for learning adversarial perturbations targeted to individual images or videos.,39023
data This discussed in prospective of augmentation terms a is scheme.,This is discussed in terms of a prospective data augmentation scheme.,39024
for yet high-quality perturbations also sparse The or be video leveraged compression. image may,The sparse yet high-quality perturbations may also be leveraged for image or video compression.,39025
this generalizing learn discriminant The to views. representations across is view-invariant to well key problem,The key to this problem is to learn discriminant view-invariant representations generalizing well across views.,39026
on results that datasets the show outperforms four Experimental our approach state-ofthe-art multiview approaches.,Experimental results on four multiview datasets show that our approach outperforms the state-ofthe-art approaches.,39027
and detail procedure. in paper evaluation the This presents dataset,This paper presents in detail the dataset and evaluation procedure.,39028
The with for directions paper concludes improvements. future,The paper concludes with directions for future improvements.,39029
can Sensing scenario during  medical the the  safety operations. ensure the surgical,  Sensing the medical scenario can ensure the safety during the surgical operations.,39030
follows. work contributions as our of are The key,The key contributions of our work are as follows.,39031
Systems have convolutional perform remarkable manipulation achieved realism. networks deep image that using,Systems that perform image manipulation using deep convolutional networks have achieved remarkable realism.,39032
despite applications pose large require changing precise Many environments. and estimates operating in robotics,Many robotics applications require precise pose estimates despite operating in large and changing environments.,39033
biology Diffusion important is cell and in complicated.,Diffusion in cell biology is important and complicated.,39034
"to additional as leverage signals. consistency geometric supervisory In propose we this paper,","In this paper, we propose to leverage geometric consistency as additional supervisory signals.",39035
favorably methods. that our compare unsupervised and depth state-of-the-art Extensive demonstrate flow experiments with models,Extensive experiments demonstrate that our depth and flow models compare favorably with state-of-the-art unsupervised methods.,39036
It this roadmap issue. tackle general presented a to,It presented a general roadmap to tackle this issue.,39037
well The samples the of when cGAN even works training are limited. number,The cGAN works well even when the number of training samples are limited.,39038
"Therefore, the approaches. several outperforms proposed method state-of-the-art","Therefore, the proposed method outperforms several state-of-the-art approaches.",39039
shown capability intelligence artificial tasks. various of Convolutional networks (CNNs) have solving great neural,Convolutional neural networks (CNNs) have shown great capability of solving various artificial intelligence tasks.,39040
"in the However, applications. raised in resource-limited has model challenges employing them increasing size","However, the increasing model size has raised challenges in employing them in resource-limited applications.",39041
"this Based light-weight operation, we ChannelNets. CNNs build known as novel on","Based on this novel operation, we build light-weight CNNs known as ChannelNets.",39042
fibrillation. and arrhythmias dangerous synchronisation underlie fatal regime of abnormal self-organised Cardiac re-entry,Cardiac re-entry regime of self-organised abnormal synchronisation underlie dangerous arrhythmias and fatal fibrillation.,39043
models Re-entry cardiac prescribed of the is locations mono-domain spatially tissue. initiated homogeneous at in,Re-entry is initiated at prescribed locations in the spatially homogeneous mono-domain models of cardiac tissue.,39044
"and consequently, they some to However, occur are automatically. traffic rarely recognize difficult signs","However, some traffic signs occur rarely and consequently, they are difficult to recognize automatically.",39045
"keypoints has in presence the an detecting be of unresolved to blur remained issue. However,","However, detecting keypoints in the presence of blur has remained to be an unresolved issue.",39046
"limited performance. a result, have those methods As","As a result, those methods have limited performance.",39047
"make scale-robust, it also to scale it To space. extend we","To make it scale-robust, we also extend it to scale space.",39048
"acceptance.} the will any our publicly our Without parallelization, upon available make code implementation\footnote{We","Without any parallelization, our implementation\footnote{We will make our code publicly available upon the acceptance.}",39049
"of results process, co-decision by acquired the classification learners. testing are During two","During testing process, classification results are acquired by co-decision of the two learners.",39050
"However, is label an process. expensive and time-consuming collection","However, label collection is an expensive and time-consuming process.",39051
"addition, view-adversarial propose In to training method view-invariant a we of learning features. enhance","In addition, we propose a view-adversarial training method to enhance learning of view-invariant features.",39052
of recognition the We representations datasets. on demonstrate the action learned effectiveness for multiple,We demonstrate the effectiveness of the learned representations for action recognition on multiple datasets.,39053
"method we generic, images. segmentation blood vessels our for Though apply of in retinal fundus","Though generic, we apply our method for segmentation of blood vessels in retinal fundus images.",39054
"extraction. framework, for architecture CNN the spectrograms single In is a onto applied feature multiple","In the framework, a single CNN architecture is applied onto multiple spectrograms for feature extraction.",39055
then fused spectrograms The acoustic extracted from features the to scenes. multiple discriminate are deep,The deep features extracted from multiple spectrograms are then fused to discriminate the acoustic scenes.,39056
Experimental accuracies proposed that achieve show the on can results method both promising datasets.,Experimental results show that the proposed method can achieve promising accuracies on both datasets.,39057
structures enhanced can introduce into acid functionalities and base Incorporating metal-mediated stabilities. pairs nucleic new,Incorporating metal-mediated base pairs into nucleic acid structures can introduce new functionalities and enhanced stabilities.,39058
G-duplexes are linear rigid B-DNA. more IMS than indicates that and,IMS indicates that G-duplexes are linear and more rigid than B-DNA.,39059
used to with propose the experiments. were IMS compatible structures calculations DFT,DFT calculations were used to propose structures compatible with the IMS experiments.,39060
and Extensive performance experiments of demonstrate MDCN models. the the over effectiveness state-of-the-art superior,Extensive experiments demonstrate the effectiveness and superior performance of MDCN over the state-of-the-art models.,39061
"the solved CCMDP, The in problems be so approach states efficiently. decouples can large that","The approach decouples states in the CCMDP, so that large problems can be solved efficiently.",39062
estimators empirical assessed of proposed The an usefulness the in example. is,The usefulness of the proposed estimators is assessed in an empirical example.,39063
of is important and images most one operations editing videos. Compositing for the,Compositing is one of the most important editing operations for images and videos.,39064
The called of realism is composite of harmonization. process the improving often results,The process of improving the realism of composite results is often called harmonization.,39065
mainly Previous images. on focus harmonization for approaches,Previous approaches for harmonization mainly focus on images.,39066
"this attack In work, step to the of harmonization. further we video one take problem","In this work, we take one step further to attack the problem of video harmonization.",39067
well show on that composite dataset. synthetic the training our to dataset Experiments real-world generalizes,Experiments show that training on our synthetic dataset generalizes well to the real-world composite dataset.,39068
"and significant pose motion, challenges. However, date, occlusions motion fast to blur","However, to date, fast motion, motion blur and occlusions pose significant challenges.",39069
"to for features depth. represent hint these may predict important Hence, learning an","Hence, these features may represent an important hint for learning to predict depth.",39070
"these blur we depth-prediction studies, out-of-focus that improves From performances. greatly the show network","From these studies, we show that out-of-focus blur greatly improves the depth-prediction network performances.",39071
visual Contrast information is in processing. crucial a factor,Contrast is a crucial factor in visual information processing.,39072
directions of provide topic. in this emerging future research further possible We,We further provide possible directions of future research in this emerging topic.,39073
"intrinsic with we information an model within-person its identity a variations. face First, and representation","First, we model a face representation with an intrinsic identity information and its within-person variations.",39074
paper access This (MEC) non-orthogonal investigates uplink multiple mobile-edge an network. (NOMA)-based computing,This paper investigates an uplink non-orthogonal multiple access (NOMA)-based mobile-edge computing (MEC) network.,39075
shown be optimal. algorithm The is proposed successfully globally to,The proposed algorithm is successfully shown to be globally optimal.,39076
the results better than show performance methods. algorithm proposed achieves Numerical that the conventional,Numerical results show that the proposed algorithm achieves better performance than the conventional methods.,39077
communication communication the integrated The wireless is in generation networks. promising next radar and system,The integrated radar and communication system is promising in the next generation wireless communication networks.,39078
"energy. However, by confined its performance is limited the","However, its performance is confined by the limited energy.",39079
"proposed. wireless it, and to radar overcome communication system powered order In integrated a is","In order to overcome it, a wireless powered integrated radar and communication system is proposed.",39080
problem subject minimization performances. constraints formulated radar to the and on is communication An energy,An energy minimization problem is formulated subject to constraints on the radar and communication performances.,39081
optimized energy energy. jointly beamforming to The waveform are radar-communication the minimize consumption and,The energy beamforming and radar-communication waveform are jointly optimized to minimize the consumption energy.,39082
solved challenging variable relaxation by is using non-convex The and problem auxiliary methods. semidefinite,The challenging non-convex problem is solved by using semidefinite relaxation and auxiliary variable methods.,39083
It is obtained. proved solution optimal that be can the,It is proved that the optimal solution can be obtained.,39084
proposed results Simulation our benchmark scheme. outperforms optimal the design demonstrate that,Simulation results demonstrate that our proposed optimal design outperforms the benchmark scheme.,39085
lesion results in skin models automated show remarkable analysis. learning Deep,Deep learning models show remarkable results in automated skin lesion analysis.,39086
input expand transforming by dataset training Data images. can the augmentation,Data augmentation can expand the training dataset by transforming input images.,39087
"GO-CNN on stability. amount reduces and examples enhances the Furthermore, adversarial dependence of training","Furthermore, GO-CNN reduces dependence on the amount of training examples and enhances adversarial stability.",39088
addresses paper optimal sampling the set selecting for problem on graphs. This an of signals,This paper addresses the problem of selecting an optimal sampling set for signals on graphs.,39089
through execution our the We performance errors approach evaluate time. of and prediction comparisons of,We evaluate the performance of our approach through comparisons of prediction errors and execution time.,39090
"this from suffers still great limitations. framework success, several Despite the tracking","Despite the great success, this tracking framework still suffers from several limitations.",39091
"it object large cannot First, properly handle rotation.","First, it cannot properly handle large object rotation.",39092
"distracted Second, gets salient the background objects. contains tracking when easily","Second, tracking gets easily distracted when the background contains salient objects.",39093
Increasing existing thumbnail capabilities direction. compression research codecs of an beyond is therefore the active,Increasing thumbnail compression beyond the capabilities of existing codecs is therefore an active research direction.,39094
"triangles. present we the algorithm for Second, novel encoding a","Second, we present a novel algorithm for encoding the triangles.",39095
identifying by for promising research. future survey We finish the directions,We finish the survey by identifying promising directions for future research.,39096
has transfer learning acquired interest. recently Deep research significant,Deep transfer learning recently has acquired significant research interest.,39097
deep Model-based probably used most transfer the is method. frequently learning,Model-based deep transfer learning is probably the most frequently used method.,39098
"approach, learning features. useful help learning more Using can transfer models to deep capture this","Using this approach, transfer learning can help deep learning models to capture more useful features.",39099
"assessment degradation To a is performance these novel two proposed. deficiencies, address method","To address these two deficiencies, a novel performance degradation assessment method is proposed.",39100
of has framework been of kinds effectiveness different our on verified datasets. proposed The,The effectiveness of our proposed framework has been verified on different kinds of datasets.,39101
"new a operations. define suite of data We a new structure, and","We define a new data structure, and a suite of new operations.",39102
"for imputing these missing framework provide and handling, exploring, values. a connected Together,","Together, these provide a connected framework for handling, exploring, and imputing missing values.",39103
methods These are `naniar`. available package in R the,These methods are available in the R package `naniar`.,39104
"accuracy complexity. and sampling between recognition tradeoff Using model can we different rates,","Using different sampling rates, we can tradeoff between recognition accuracy and model complexity.",39105
lectin-activated digital. were showed PBMC Robbins responses long that ago,Robbins long ago showed that lectin-activated PBMC responses were digital.,39106
cells. been confirmed This for T now has pMHC-activated,This has now been confirmed for pMHC-activated T cells.,39107
"general, initial signal individual of of responses independent Hence, are cells in strength. T","Hence, in general, responses of individual T cells are independent of initial signal strength.",39108
differentiable Our is approach neural novel forward-chaining a induction on rule network. based,Our approach is based on a novel neural forward-chaining differentiable rule induction network.,39109
"predicates, learned and which are from The may rules their be compositionally invented. interpretable","The rules are interpretable and learned compositionally from their predicates, which may be invented.",39110
remained Automated the key within field a Game design of challenge game AI. has,Automated game design has remained a key challenge within the field of Game AI.,39111
on learning approach relies machine games. to representations learn of Our instead approximate,Our approach instead relies on machine learning to learn approximate representations of games.,39112
recombines from conceptual representations games via create expansion. to learned knowledge Our new approach these,Our approach recombines knowledge from these learned representations to create new games via conceptual expansion.,39113
to games. by existing the for demonstrating evaluate ability the this system approach We recreate,We evaluate this approach by demonstrating the ability for the system to recreate existing games.,39114
by is through differentiating the camera. network trained The end-to-end,The network is trained end-to-end by differentiating through the camera.,39115
is consisting images. a BubGAN trained The ten on of dataset thousand initially bubble labeled,The BubGAN is trained initially on a labeled bubble dataset consisting of ten thousand images.,39116
million dataset using trained BubGAN. is pre-generated A the bubble,A million bubble dataset is pre-generated using the trained BubGAN.,39117
"contain therefore additional labeling. These do detailed not bubble manual images information, require","These images contain detailed bubble information, therefore do not require additional manual labeling.",39118
with which without is information. useful GAN This conventional generates more the compared images labeling,This is more useful compared with the conventional GAN which generates images without labeling information.,39119
can bring such into the classroom? How we opportunities,How can we bring such opportunities into the classroom?,39120
Visual regions most image. an object the important saliency highlight to in detection aims,Visual saliency detection aims to highlight the most important object regions in an image.,39121
on learning of works forgetting models. in Previous problem the address sequential discriminative,Previous works on sequential learning address the problem of forgetting in discriminative models.,39122
this case generative models. consider the we of paper In,In this paper we consider the case of generative models.,39123
agent used testing be make to can approach efficient. This more,This approach can be used to make agent testing more efficient.,39124
"networks. end, we algorithm, a learning \textbf{\textit{rate-adaptive}} present novel training this for To","To this end, we present a novel training algorithm, for learning \textbf{\textit{rate-adaptive}} networks.",39125
would of Similar related regardless their distances. to each features other be,Similar features would be related to each other regardless of their distances.,39126
very Sum-of-squares functions are objective vision popular computer algorithms. in,Sum-of-squares objective functions are very popular in computer vision algorithms.,39127
"easy However, functions to objective are always not these optimize.","However, these objective functions are not always easy to optimize.",39128
apply problem of to the We our stereo ie. motion method,We apply our method to the problem of motion stereo ie.,39129
motion of geometry from estimating sequence. monocular images the and scene a jointly pairs of,jointly estimating the motion and scene geometry from pairs of images of a monocular sequence.,39130
panoptic network segmentation. single for a We method present,We present a single network method for panoptic segmentation.,39131
neural track features. been have using used Siamese deep networks objects recently to Convolutional,Convolutional Siamese neural networks have been recently used to track objects using deep features.,39132
on per-camera-pair labelled model supervised training manually data. Mostexistingpersonre-identification(re-id)methods learning relyon pairwise,Mostexistingpersonre-identification(re-id)methods relyon supervised model learning on per-camera-pair manually labelled pairwise training data.,39133
"pore image. a generates proposed DeepResPore fingerprint input Specifically, intensity the the model from map","Specifically, the proposed DeepResPore model generates a pore intensity map from the input fingerprint image.",39134
"sets. achieves proposed the performance on both state-of-the-art importantly, test Most approach","Most importantly, the proposed approach achieves state-of-the-art performance on both test sets.",39135
Efficient practical of is and of great volumes images segmentation and easy importance.,Efficient and easy segmentation of images and volumes is of great practical importance.,39136
user may very pixel input. a modest final be The classification obtained from,The final pixel classification may be obtained from a very modest user input.,39137
of our graph ingredient important image method is that encodes a An content.,An important ingredient of our method is a graph that encodes image content.,39138
this classification. to a segmentation can efficient be through very pixel approach We demonstrate how,We demonstrate how this can be a very efficient approach to segmentation through pixel classification.,39139
"different visual task. accounting for inputs trajectories the stimulus-driven taken during n-back Additionally, explains the","Additionally, accounting for stimulus-driven visual inputs explains the different trajectories taken during the n-back task.",39140
"dynamics. cognitively underpinnings the brain developmentally elucidate structural results and of these relevant spatiotemporal Overall,","Overall, these results elucidate the structural underpinnings of cognitively and developmentally relevant spatiotemporal brain dynamics.",39141
spherical experiments segmentation promising a semantic on show task. results Initial,Initial experiments show promising results on a spherical semantic segmentation task.,39142
age classification facial on information. system This the gender consider to re-examination requires,This requires re-examination on the gender classification system to consider facial age information.,39143
We the controlling newborns vaccination of different adults have and find that epidemics. in consequences,We find that the vaccination of newborns and adults have different consequences in controlling epidemics.,39144
of most life. diseases influential human's in ophthalmologic one Strabismus the is,Strabismus is one of the most influential ophthalmologic diseases in human's life.,39145
strabismus and Timely detection to of treatment. contributes its prognosis,Timely detection of strabismus contributes to its prognosis and treatment.,39146
"In dataset established strabismus paper, is tele this a ophthalmologists. by the","In this paper, a tele strabismus dataset is established by the ophthalmologists.",39147
"strategies pruning -- global compared. layer-wise are filter and Two strategies,","Two filter pruning strategies -- global and layer-wise strategies, are compared.",39148
"increasing generalization. improves artificially By number helps and it overfitting examples, of training the reducing","By artificially increasing the number of training examples, it helps reducing overfitting and improves generalization.",39149
face further on Codes studies be to detection facilitate problem. the will released,Codes will be released to facilitate further studies on the face detection problem.,39150
been has ing vision computer for task always recognition Facial a scientists experts. and challeng-,Facial recognition has always been a challeng- ing task for computer vision scientists and experts.,39151
"a facial Along lines, disguised same for (DFR). we recognition architecture the learning propose deep","Along the same lines, we propose a deep learning architecture for disguised facial recognition (DFR).",39152
basic This overall our a system. intelligence architecture to imparts,This overall architecture imparts a basic intelligence to our system.,39153
approach also The of Identification with our efficiency methods. Disguised state-of-the-art is compared Facial the,The efficiency of our approach is also compared with the state-of-the-art Disguised Facial Identification methods.,39154
"a in Hence, early for is there PMs. the need of anomalies detection","Hence, there is a need for early detection of the anomalies in PMs.",39155
"states. undergoes machine degradation usually complete failure, operation to a series of From normal the","From normal operation to complete failure, the machine usually undergoes a series of degradation states.",39156
data candidate because scarcity all with the views. is The of of learning challenging task,The task is challenging because of the scarcity of learning data with all candidate views.,39157
"Internet broadcast freely Meanwhile, are (e.g. videos the on available","Meanwhile, broadcast videos are freely available on the Internet (e.g.",39158
"these record selected views, other videos views. only omitting the However, camera candidate the","However, these videos only record the selected camera views, omitting the other candidate views.",39159
"to Then, heatmap objects (e.g. a propose spatial-appearance we describe foreground","Then, we propose a spatial-appearance heatmap to describe foreground objects (e.g.",39160
this significantly dataset. challenging Our methods outperforms method on state-of-the-art,Our method significantly outperforms state-of-the-art methods on this challenging dataset.,39161
"provability, evidence Conjectandi, of theory. of force probabilities, Ars additive Keywords: proof, non probability","Keywords: force of proof, probability of provability, Ars Conjectandi, non additive probabilities, evidence theory.",39162
One steps is determine to weights. key point associated the,One key point steps is to determine the associated weights.,39163
using weights for we determining by approach propose the RIM a OWA quantifier. Then new,Then we propose a new approach for determining OWA weights by using the RIM quantifier.,39164
datasets. studies tag-sequencing Case for shown transcriptome are and,Case studies are shown for transcriptome and tag-sequencing datasets.,39165
remote processing. sensing image a classical Destriping in is problem,Destriping is a classical problem in remote sensing image processing.,39166
"variation is oriented accomplish approximation. of the stripe model all, designed the orientation First to","First of all, the oriented variation model is designed to accomplish the stripe orientation approximation.",39167
networks the achieve state-of-the-art in classification image Neural tasks.,Neural networks achieve the state-of-the-art in image classification tasks.,39168
in that only are if older wrongly males are males predicting elderly the set). training,wrongly predicting that males are older if only elderly males are in the training set).,39169
"adaptation and such domain our framework standard Finally, validate MNIST. we as SVHN on datasets,","Finally, we validate our framework on standard domain adaptation datasets, such as SVHN and MNIST.",39170
the the of results test proposal. The the systems verify application effectiveness on,The application results on the test systems verify the effectiveness of the proposal.,39171
better of determine combinations revealed than removal single ones. Simulations effects traits that,Simulations revealed that combinations of traits better determine removal effects than single ones.,39172
"frameworks. datasets However, methods deep implemented are these various and through and evaluated learning","However, these methods are implemented and evaluated through various datasets and deep learning frameworks.",39173
covers package methods There that in unifying manner. has these not a a been,There has not been a package that covers these methods in a unifying manner.,39174
"unification, of simplicity, emphasizes Geoseg The flexibility. implementation and","The implementation of Geoseg emphasizes unification, simplicity, and flexibility.",39175
toolboxes thermal RGB of modalities. annotation support and images sequential the Both the in,Both toolboxes support the annotation of sequential images in the RGB and thermal modalities.,39176
"using of the So have toolboxes. been far, frames thousands annotated of tens","So far, tens of thousands of frames have been annotated using the toolboxes.",39177
"an is system. a non-trivial requires Measuring testing, which for infeasible individual exhaustive discrimination","Measuring individual discrimination requires an exhaustive testing, which is infeasible for a non-trivial system.",39178
that record it not Calment's unlikely is has How broken? yet been,How unlikely is it that Calment's record has not yet been broken?,39179
but Calment's impossible. record Jeanne Louise is not exceptional,Jeanne Louise Calment's record is exceptional but not impossible.,39180
that evidence lifespans. lifespan does It to provide is upper her limit an human not,It does not provide evidence that her lifespan is an upper limit to human lifespans.,39181
"to we performance. estimated clinical Moreover, further parameters incorporate enhance","Moreover, we incorporate estimated clinical parameters to further enhance performance.",39182
"AS-OCT screening dataset, previous system methods. through to On a our is clinical validated comparisons","On a clinical AS-OCT dataset, our system is validated through comparisons to previous screening methods.",39183
"action is common representation optical Currently, for most recognition the motion flow.","Currently, the most common motion representation for action recognition is optical flow.",39184
dynamics. a adheres which tracking particle Lagrangian on perspective based to Optical on is flow,Optical flow is based on particle tracking which adheres to a Lagrangian perspective on dynamics.,39185
learn complex in We manner. the filters an end-to-end in domain,We learn filters in the complex domain in an end-to-end manner.,39186
"complex phase-information extraction. Gabor employed typically We complex filters, design filters these for to resemble","We design these complex filters to resemble complex Gabor filters, typically employed for phase-information extraction.",39187
of coordination Scheduling these strategies discussed. briefly are schemes network,Scheduling strategies of these network coordination schemes are briefly discussed.,39188
viability with in of analyzed real-time architectures these is mind. surveillance video applications The,The viability of these architectures is analyzed with real-time video surveillance applications in mind.,39189
"the of are degradation perturbation the to they model. sensitive too However,","However, they are too sensitive to the perturbation of the degradation model.",39190
"happiness experiments HAPPEI. challenging Intensive namely the are on performed group-level database, intensity","Intensive experiments are performed on the challenging group-level happiness intensity database, namely HAPPEI.",39191
the a multiple-komi AlphaGo modification Zero of propose We Zero/Leela paradigm.,We propose a multiple-komi modification of the AlphaGo Zero/Leela Zero paradigm.,39192
"propose (ICNet) Inverse-Consistent this registration. for unsupervised end, we To Network an deformable image deep","To this end, we propose an Inverse-Consistent deep Network (ICNet) for unsupervised deformable image registration.",39193
code publicly be available. Our made will,Our code will be made publicly available.,39194
synthetic its computational The evaluated using is advantages algorithm numerically are established. and data new,The new algorithm is evaluated numerically using synthetic data and its computational advantages are established.,39195
popular a becomes Neural technology Convolutional (CNN) implementing recent FPGA years. Network in for,FPGA becomes a popular technology for implementing Convolutional Neural Network (CNN) in recent years.,39196
propose under a humans control. user real of generating video-realistic We method for animations,We propose a method for generating video-realistic animations of real humans under user control.,39197
synthesis. in image outperform the results human state-of-the-art learning-based Our,Our results outperform the state-of-the-art in learning-based human image synthesis.,39198
have Layer also (LN) (IN) and been Normalization proposed. Instance Normalization,Instance Normalization (IN) and Layer Normalization (LN) have also been proposed.,39199
basic as the adopted U-Net structure. is DCNN,U-Net is adopted as the basic DCNN structure.,39200
"detailed feature show subdivision of results i.e. the The that map,","The results show that detailed subdivision of the feature map, i.e.",39201
"IN, group number a with achieves accuracy. higher GN or large","GN with a large group number or IN, achieves higher accuracy.",39202
This improvement mainly generalization. better from comes accuracy model,This accuracy improvement mainly comes from better model generalization.,39203
are uploaded at maintained Codes and Zhou's Github. Xiao-Yun,Codes are uploaded and maintained at Xiao-Yun Zhou's Github.,39204
modeling these long-term improve innovations demonstrate results Our the that trajectories. motion of,Our results demonstrate that these innovations improve the modeling of long-term motion trajectories.,39205
"show predict connections recurrent small i.e. skip to the objects, ability occluded Novel","Novel recurrent skip connections show the ability to predict small occluded objects, i.e.",39206
"previous Experiments improvements a our also Carlo network, to quantify literature. and Monte approach,","Experiments also quantify improvements to our previous network, a Monte Carlo approach, and literature.",39207
divergence introduce to Kullback-Leibler reject on a also We doubtful criterion based examples.,We also introduce a criterion based on Kullback-Leibler divergence to reject doubtful examples.,39208
is effect visual experiments and performed Extensive included. discussion on enhancement are,Extensive experiments are performed and discussion on visual effect enhancement is included.,39209
segmentation and training optimised with post-process. a for present lesion ensemble method We method two-stage,We present a two-stage method for lesion segmentation with optimised training method and ensemble post-process.,39210
sign microcalcifications of cancer. an be breast early of can Cluster,Cluster of microcalcifications can be an early sign of breast cancer.,39211
data. focus We on estimating video the frequency of human problem from hand-tremor input RGB,We focus on the problem of estimating human hand-tremor frequency from input RGB video data.,39212
"The rate e.g. assumed in constant model,","The constant rate model, assumed in e.g.",39213
"Saleh-Valenzuela to same the reproduce is unable model, effects. the","the Saleh-Valenzuela model, is unable to reproduce the same effects.",39214
datasets and algorithms. survey state-of-the-art associated This the also public reviews,This survey also reviews the public datasets and associated state-of-the-art algorithms.,39215
convergence. processes The performed are iteratively until,The processes are performed iteratively until convergence.,39216
"component distance unary employ Specifically, we to the the build geodesic term.","Specifically, we employ the geodesic distance component to build the unary term.",39217
"the outliers. register aim we images step different patients, to yields matching As many of","As we aim to register images of different patients, the matching step yields many outliers.",39218
algorithm we new which efficiently discards EM-weighting Then a outliers. propose,Then we propose a new EM-weighting algorithm which efficiently discards outliers.,39219
carried optimization descent Global a with algorithm. fast gradient is out,Global optimization is carried out with a fast gradient descent algorithm.,39220
register approach large to our allows datasets. robustly This,This allows our approach to robustly register large datasets.,39221
in and role an pedestrians) cyclists will play future important traffic.,cyclists and pedestrians) will play an important role in future traffic.,39222
"of movement consists detection prediction, The primitive basic e.g. intention","The intention detection consists of basic movement primitive prediction, e.g.",39223
"trajectory. a turning, and future forecast moving, standing, the of","standing, moving, turning, and a forecast of the future trajectory.",39224
"open Heterogeneous, agents and of sets vehicles, infrastructure, interacting (cooperating e.g.","Heterogeneous, open sets of agents (cooperating and interacting vehicles, infrastructure, e.g.",39225
"are inconsistencies resolved the of implausibilities Concealments, intelligence cooperating by collective and agents.","Concealments, implausibilities and inconsistencies are resolved by the collective intelligence of cooperating agents.",39226
obtained compared. both the Results are approaches using,Results obtained using both the approaches are compared.,39227
"identification. signal limitation based MUSIC, the methods on as for subspace known is This otherwise","This is the limitation for methods based on MUSIC, otherwise known as signal subspace identification.",39228
"appearance. from structure structure-from-motion not actually learns and context This motion, from algorithm only","This algorithm actually learns structure-from-motion from motion, and not only structure from context appearance.",39229
"use videos, moving stabilized camera and Two ones. cases considered, are unstabilized","Two use cases are considered, unstabilized moving camera videos, and stabilized ones.",39230
"their negate drop-out online which effectiveness. cost high However, suffers from treatment might rates,","However, online treatment suffers from high drop-out rates, which might negate their cost effectiveness.",39231
varying The camera illumination shown changes. improvements and perspective significant have experimental results presented with,The presented experimental results have shown significant improvements with varying illumination and camera perspective changes.,39232
"some drawbacks, i.e. with compact algorithms However, come also","However, compact algorithms also come with some drawbacks, i.e.",39233
and they to performance problems. tend on premature non-separable poorer show convergence,they tend to premature convergence and show poorer performance on non-separable problems.,39234
Gaussian modeled a The are and time-variant using unknown values safety spatio-temporal process.,The unknown and time-variant safety values are modeled using a spatio-temporal Gaussian process.,39235
Many that and complex plants harbor mechanisms transfer. pollen promote outcrossing efficient,Many plants harbor complex mechanisms that promote outcrossing and efficient pollen transfer.,39236
"(SI) self-incompatibility mechanisms, systems. floral include such genetic These adaptations molecular as as well as","These include floral adaptations as well as genetic mechanisms, such as molecular self-incompatibility (SI) systems.",39237
"with association been repeatedly in self-fertilization to lost, has SI (selfing). transitions often Conversely,","Conversely, SI has repeatedly been lost, often in association with transitions to self-fertilization (selfing).",39238
"complexity, of improve lower performance To the is its convolution a target adopted. dilated with","To improve its performance with a target of lower complexity, the dilated convolution is adopted.",39239
cost. and computational thinner A the to shallower structure is designed decrease,A shallower and thinner structure is designed to decrease the computational cost.,39240
shows results on scenes. captured promising system road the Our,Our system shows promising results on the captured road scenes.,39241
of experts method variable selection frameworks A for in proposed. mixture is implicit,A method for implicit variable selection in mixture of experts frameworks is proposed.,39242
prior We information set covariates. taken where introduce independent of from structure is a a,We introduce a prior structure where information is taken from a set of independent covariates.,39243
Robust membership a are prior. normal predictors identified gamma using class,Robust class membership predictors are identified using a normal gamma prior.,39244
implementation inference the a out Fully of Gibbs is via carried sampler. Bayesian,Fully Bayesian inference is carried out via the implementation of a Gibbs sampler.,39245
one in described as cell were populations Chondrocytes literature. cartilage most,Chondrocytes were described as one cell populations in most cartilage literature.,39246
"training, projections network for and mappings, natural introduced. Bases, Neural are metrics,","Bases, mappings, projections and metrics, natural for Neural network training, are introduced.",39247
"relatively in emerges, simple even naturally datasets. Non-Gaussianity","Non-Gaussianity naturally emerges, even in relatively simple datasets.",39248
"analyzed, of point are Training view. statistics, physics energies from hierarchies and","Training statistics, hierarchies and energies are analyzed, from physics point of view.",39249
"Duality (for observations and between observables pixels) established. example, is","Duality between observables (for example, pixels) and observations is established.",39250
numerical between and exact studied. solutions is Relationship,Relationship between exact and numerical solutions is studied.,39251
offered. of financial key a and mathematics problem are interpretations Physics,Physics and financial mathematics interpretations of a key problem are offered.,39252
"is not practical majority for vast Unfortunately, researchers. approach this of the","Unfortunately, this approach is not practical for the vast majority of researchers.",39253
"semantic years. learning progress has great deep shown recent for in segmentation Concurrently,","Concurrently, deep learning for semantic segmentation has shown great progress in recent years.",39254
"tasks in we are propose Specifically, network structure two a these which coupled. highly","Specifically, we propose a network structure in which these two tasks are highly coupled.",39255
two-stage One the refinement of process. key this is novelty approach,One key novelty of this approach is the two-stage refinement process.,39256
from paper a modeling that This can head the problem is solve. we proposes Internet,This paper proposes that head modeling from the Internet is a problem we can solve.,39257
head. target the rough We of shape of reconstruction the,We target reconstruction of the rough shape of the head.,39258
algorithm. call We boundary-value method our growing,We call our method boundary-value growing algorithm.,39259
presented. photos downloaded of celebrities Results are the on from Internet,Results on photos of celebrities downloaded from the Internet are presented.,39260
networks the dynamics spiking of neurons. collective structured of densely connected We analyze hierarchically,We analyze the collective dynamics of hierarchically structured networks of densely connected spiking neurons.,39261
scenes UAV data. on try We synthetic and real our algorithm both flight,We try our algorithm on both synthetic scenes and real UAV flight data.,39262
more this is thoroughly. that with a article video our results Along present,Along with this article is a video that present our results more thoroughly.,39263
problem. classification plays information supervised Label in image an important role hyperspectral,Label information plays an important role in supervised hyperspectral image classification problem.,39264
"be majority vote propagated algorithm. a can the by final calculated label Therefore,","Therefore, the final propagated label can be calculated by a majority vote algorithm.",39265
also The subtlety discussed. of is interpreting results,The subtlety of interpreting results is also discussed.,39266
"efficiently. And such them exist, if how inputs find to","And if such inputs exist, how to find them efficiently.",39267
"practically use these also by the to implementations. providing We show corresponding solutions, how","We also show how to practically use these solutions, by providing the corresponding implementations.",39268
capabilities All utilization of medicine. are the invasive in of use minimally these,All these capabilities are of use in the utilization of minimally invasive medicine.,39269
"Micro-Dictionary a In Coding Learning and paper, Deep Network propose we this (DDLCN). novel","In this paper, we propose a novel Deep Micro-Dictionary Learning and Coding Network (DDLCN).",39270
"most has standard deep (pooling, the layers DDLCN connected, learning fully, etc.) input/output, of","DDLCN has most of the standard deep learning layers (pooling, fully, connected, input/output, etc.)",39271
input an for learning over-complete the dictionary training dictionary data. layer The learns,The dictionary learning layer learns an over-complete dictionary for the input training data.,39272
several proposed methods learning and learning the with deep We architectures. empirically DDLCN compare dictionary,We empirically compare the proposed DDLCN with several dictionary learning methods and deep learning architectures.,39273
understanding works. of how bootstrap provides It intuitive an the,It provides an intuitive understanding of how the bootstrap works.,39274
in details are Mathematical references cited. that available are,Mathematical details are available in references that are cited.,39275
holds. distributional findings our Our similarity that hypothesis indicate,Our findings indicate that our distributional similarity hypothesis holds.,39276
"domain However, in of received techniques attention only computer maritime recently. vision application","However, application of computer vision techniques in maritime domain received attention only recently.",39277
challenges. requirements own maritime offers and unique environment its The,The maritime environment offers its own unique requirements and challenges.,39278
in need Assessment is detections computer of a quality of the vision. fundamental,Assessment of the quality of detections is a fundamental need in computer vision.,39279
computer defining the vision. of discuss We suitable problem assessment maritime for metrics,We discuss the problem of defining assessment metrics suitable for maritime computer vision.,39280
consider computer metrics maritime for metrics new proximity vision. bottom assessment edge We as,We consider new bottom edge proximity metrics as assessment metrics for maritime computer vision.,39281
ESS paper not the MCMC use does for of algorithms. This cover,This paper does not cover the use of ESS for MCMC algorithms.,39282
local commonly a Existing of matching a global compositions pipelines stage. stage and reassembly consist,Existing reassembly pipelines commonly consist of a local matching stage and a global compositions stage.,39283
time-critical a multi-object applications. tracking in problem video Online is analysis fundamental,Online multi-object tracking is a fundamental problem in time-critical video analysis applications.,39284
qualitatively the The algorithm. and experiments conducted the effectiveness quantitatively of demonstrate,The experiments conducted quantitatively and qualitatively demonstrate the effectiveness of the algorithm.,39285
depth propose We a fully architecture network. end-to-end an convolutional with for also inference,We also propose an architecture for end-to-end depth inference with a fully convolutional network.,39286
This be security to is potential issue. phenomenon a considered,This phenomenon is considered to be a potential security issue.,39287
"unknown. shared harmful characteristics perturbations such However, of the remain the directions of","However, the characteristics of the shared directions of such harmful perturbations remain unknown.",39288
stage architecture. structure recurrent-auto to encoder We RNN with the cascade form decoder a a,We cascade the RNN structure with a decoder stage to form a recurrent-auto encoder architecture.,39289
"is the backpropagation. regarding during prior waveform Moreover, information and structure incorporated initialization","Moreover, prior information regarding the waveform structure is incorporated during initialization and backpropagation.",39290
under robustness the desiderata a the Relevance addition of concern irrelevant consequence information. of relation,Relevance desiderata concern the robustness of a consequence relation under the addition of irrelevant information.,39291
of semantic to (ir)relevance syntactic what considerations. and amounts For we use account an,For an account of what (ir)relevance amounts to we use syntactic and semantic considerations.,39292
known non-monotonic criteria relevance are semantic logic. Some from,Some semantic relevance criteria are known from non-monotonic logic.,39293
investigate of relevance paper warrant properties this structured that In we argumentation desiderata. systems,In this paper we investigate properties of structured argumentation systems that warrant relevance desiderata.,39294
recently success single-image super-resolution great achieved (SISR). Convolutional have neural networks (CNNs) in,Convolutional neural networks (CNNs) have recently achieved great success in single-image super-resolution (SISR).,39295
"textural tend produce details. these miss and outputs over-smoothed some However, to methods","However, these methods tend to produce over-smoothed outputs and miss some textural details.",39296
applied Learning impressive to this successfully been results. Deep achieved problem and methods have have,Deep Learning methods have been successfully applied to this problem and have achieved impressive results.,39297
"amount labeled of data. require However, methods large training a these","However, these methods require a large amount of labeled training data.",39298
Each group adversarial realistic through images is generate used to training. then,Each group is then used to generate realistic images through adversarial training.,39299
feature original the aim to generated between samples. is and Our emphasize similarity samples,Our aim is to emphasize feature similarity between generated samples and the original samples.,39300
image method an medical The is becoming influential synthesis. cycleGAN in,The cycleGAN is becoming an influential method in medical image synthesis.,39301
recent on has two The been for re-identification person focused trends. research,The recent research for person re-identification has been focused on two trends.,39302
is to form more the descriptors. local part-based feature features learning One informative,One is learning the part-based local features to form more informative feature descriptors.,39303
is loss triplet metric as functions learning the other effective such loss family. The designing,The other is designing effective metric learning loss functions such as the triplet loss family.,39304
averaging been has frequencies. the Time mixed traditional sampling to approach handle,Time averaging has been the traditional approach to handle mixed sampling frequencies.,39305
"However, information ignores possibly in high embedded frequency. it","However, it ignores information possibly embedded in high frequency.",39306
"is are only together, however, co-registered allowing coherent This, comparisons. feasible graphs when multiple","This, however, is only feasible when graphs are co-registered together, allowing coherent multiple comparisons.",39307
technology of key The stands registration as therefore group-wise topologies analyses. for robust vascular enabling,The robust registration of vascular topologies stands therefore as key enabling technology for group-wise analyses.,39308
are matching real from vascular results angiographies. graphs synthetic and using found of Promising,Promising results of vascular matching are found using graphs from synthetic and real angiographies.,39309
are designs future and towards discussed potential clinical Observations applications.,Observations and future designs are discussed towards potential clinical applications.,39310
"derivatives can band, higher this Gaussian communication order for energy-efficient In pulses nanodevices. of provide","In this band, higher order derivatives of Gaussian pulses can provide energy-efficient communication for nanodevices.",39311
advantage relatively to simple the also implement The being has proposed of computationally. method,The proposed method also has the advantage of being relatively simple to implement computationally.,39312
a system. real-time paper the anatomy In of we analysis describe this facial,In this paper we describe the anatomy of a real-time facial analysis system.,39313
To quality implement interpolated we with output a BEGAN Slerp. the increase the of,To increase the quality of the interpolated output we implement a BEGAN with Slerp.,39314
"(SPG). introduced: Policy is Sampled algorithm new offline In Gradient paper, learning this actor-critic a","In this paper, a new offline actor-critic learning algorithm is introduced: Sampled Policy Gradient (SPG).",39315
"a for VoxelMorph, registration. medical image learning-based deformable, We fast framework present pairwise","We present VoxelMorph, a fast learning-based framework for deformable, pairwise medical image registration.",39316
"work, training two we explore strategies. In different this","In this work, we explore two different training strategies.",39317
"available the setting, auxiliary we segmentations in the leverage data. second training In","In the second setting, we leverage auxiliary segmentations available in the training data.",39318
voxelmorph.csail.mit.edu. code at available Our is freely,Our code is freely available at voxelmorph.csail.mit.edu.,39319
"simple by In multimodal and contrast, average like operators fusion concatenation. data is achieved","In contrast, multimodal data fusion is achieved by simple operators like average and concatenation.",39320
MFB features with explore classification. and different effectiveness in its combine We video video-level,We combine MFB with different video-level features and explore its effectiveness in video classification.,39321
"and process However, intensive slow, is this unreliable. labor","However, this process is slow, labor intensive and unreliable.",39322
"paper, this objective. propose we meet DocFace+ In this to","In this paper, we propose DocFace+ to meet this objective.",39323
great Fully success years. in (FCN) achieved parsing networks in recent have convolutional human,Fully convolutional networks (FCN) have achieved great success in human parsing in recent years.,39324
to labels. segmentation method We generate learning pseudo keypoint masks part iterative develop from an,We develop an iterative learning method to generate pseudo part segmentation masks from keypoint labels.,39325
"we pseudo output FCN train to With parsing network pixel-level human an masks, predictions. these","With these pseudo masks, we train an FCN network to output pixel-level human parsing predictions.",39326
"Full-CRFs methods expensive. for However, are too traditional","However, traditional methods for Full-CRFs are too expensive.",39327
by CRF we handle Then how to multi-label showing implement expansion moves.,Then we handle multi-label CRF by showing how to implement expansion moves.,39328
also We show approach the our effectiveness on task. segmentation semantic of,We also show the effectiveness of our approach on semantic segmentation task.,39329
"considered its repertoire. Inhibition is pattern shape to activity, neural and broaden","Inhibition is considered to shape neural activity, and broaden its pattern repertoire.",39330
"validate to used Simulation, and data phantom proposed have patient framework. AAA sets the been","Simulation, phantom and patient AAA data sets have been used to validate the proposed framework.",39331
"also achieved accuracy, Performance were time-efficiency. in advantages terms of and robustness","Performance advantages were also achieved in terms of accuracy, robustness and time-efficiency.",39332
All will the source. code be open,All the code will be open source.,39333
and with primarily age. alterations on disease its Investigations focus,Investigations primarily focus on its alterations with disease and age.,39334
are differences significant Rasa between causing links the extracted The pairs statistically.,The significant links causing differences between the Rasa pairs are extracted statistically.,39335
It winning entry in robust the flow challenge. competition of is vision the the optical,It is the winning entry in the optical flow competition of the robust vision challenge.,39336
"gains. our analyze the sources of Next, performance we experimentally","Next, we experimentally analyze the sources of our performance gains.",39337
concludes support AI of these four It that each technology areas. can,It concludes that AI technology can support each of these four areas.,39338
"a be to good approach different there business circumstances. in take some may However, reasons","However, there may be good business reasons to take a different approach in some circumstances.",39339
of Graph have graph become Networks crucial (GCNs) representations vertices. learning a tool on Convolutional,Graph Convolutional Networks (GCNs) have become a crucial tool on learning representations of graph vertices.,39340
in We item graphs problems address this in paper. relocation,We address item relocation problems in graphs in this paper.,39341
item Our (CBS). approach (SMT) combines with modulo conflict-based search theory relocation for satisfiability solving,Our solving approach for item relocation combines satisfiability modulo theory (SMT) with conflict-based search (CBS).,39342
benchmarks evaluation SMT-CBS the CBS. several Experimental that shows outperforms standard the on significantly algorithm,Experimental evaluation on several benchmarks shows that the SMT-CBS algorithm significantly outperforms the standard CBS.,39343
"explain algorithm. then define a and reliability Furthermore, we the of notion","Furthermore, we define a notion of reliability and then explain the algorithm.",39344
"well known we simulations, demonstrate performance With over extended policy. the in gains the MaxWeight","With extended simulations, we demonstrate the gains in performance over the well known MaxWeight policy.",39345
group raters multiple evaluate a reliability who the Inter-Rater of quantifies subjects. between,Inter-Rater quantifies the reliability between multiple raters who evaluate a group of subjects.,39346
through of This pairs. accomplished permutations user is,This is accomplished through permutations of user pairs.,39347
layer propose We segmentation. unsupervised for domain scene adaption in semantic normalization a,We propose a normalization layer for unsupervised domain adaption in semantic scene segmentation.,39348
intractable makes the This power problem for systems. large-scale nonlinearity,This nonlinearity makes the problem intractable for large-scale power systems.,39349
allocation power for reliability approaches system two compares improvement. budget for paper This,This paper compares two approaches for budget allocation for power system reliability improvement.,39350
"and is traditional from which first method suffers non-convexity one, the nonlinearity. the The","The first method is the traditional one, which suffers from the non-convexity and nonlinearity.",39351
modified system. applied are to RTBS the Both models,Both models are applied to the modified RTBS system.,39352
"some cases, results directly data. experimental their can be also compared single-molecule In to","In some cases, their results can also be compared directly to single-molecule experimental data.",39353
Code results in are resulting showing given and samples online supplements. the animations,Code samples and resulting animations showing results are given in the online supplements.,39354
assessed the population. of migration of This dynamics on effects the study species a,This study assessed the effects of migration on the dynamics of a species population.,39355
Experimental the of efficacy demonstrate method. proposed results the,Experimental results demonstrate the efficacy of the proposed method.,39356
"limited have resources However, computational power and capability. constrained these devices portable","However, these portable devices have constrained power resources and limited computational capability.",39357
separately. first abnormality proposed is trained each The network by,The proposed network is first trained by each abnormality separately.,39358
abnormalities. using network trained all the Then is,Then the network is trained using all abnormalities.,39359
of The of targetted network. balance regulates in inhibition and the association this span global,The balance of targetted and global inhibition regulates this span of association in the network.,39360
in nontrivial role and a anti-Hebbian Our memory. inhibitory engrams learning results of suggest associative,Our results suggest a nontrivial role of anti-Hebbian learning and inhibitory engrams in associative memory.,39361
proposed Quantitative other and techniques the have adversarial with illustrated model comparisons performance. qualitative,Quantitative and qualitative comparisons with other adversarial techniques have illustrated the proposed model performance.,39362
image by Local task. homography caused common global in a misalignment stitching is issue,Local misalignment caused by global homography is a common issue in image stitching task.,39363
"a we problem. this propose method (GCPW) Warping paper, to Content-Preserving In Generalized alleviate this","In this paper, we propose a Generalized Content-Preserving Warping (GCPW) method to alleviate this problem.",39364
images our We optimal locally by grid align generated method. mesh an GCPW with,We align images locally with an optimal grid mesh generated by our GCPW method.,39365
is data depth The training of lack first models. the limitation for learning deep,The first limitation is the lack of depth data for training deep learning models.,39366
RGB-D multimodal more from complementary resulting The features. model benefits also,The resulting RGB-D model also benefits from more complementary multimodal features.,39367
motivates uncertainty in for the This Semantic managing Web. the need,This motivates the need for managing uncertainty in the Semantic Web.,39368
"tableau algorithm. for of non-determinism, efficient handling its suitable Prolog, implementing is the with","Prolog, with its efficient handling of non-determinism, is suitable for implementing the tableau algorithm.",39369
of TRILLP implementation tableau and are Prolog algorithm. systems TRILL the offering a,TRILL and TRILLP are systems offering a Prolog implementation of the tableau algorithm.,39370
were outperform reasoners to reasoners. shown Both DL state-of-the-art,Both reasoners were shown to outperform state-of-the-art DL reasoners.,39371
An effectiveness the shows TORNADO. of comparison experimental,An experimental comparison shows the effectiveness of TORNADO.,39372
heart-rate field an Automated is sleep stage active of variability using classification research.,Automated sleep stage classification using heart-rate variability is an active field of research.,39373
"Further, adds - a calibration-free environment both complexity to additional and reconstruction texture recovery.","Further, a calibration-free environment adds additional complexity to both - reconstruction and texture recovery.",39374
processes age similar composed bone assessment was of to the physicians' study The procedures.,The study was composed of processes similar to the physicians' bone age assessment procedures.,39375
"also And which applicable well. adolescents method, we Sauvegrain as is presented the to","And we also presented the Sauvegrain method, which is applicable to adolescents as well.",39376
competition the in We a model. fixation two-species study of stochastic probability,We study the probability of fixation in a stochastic two-species competition model.,39377
an easy-to-use tools issue. to important science support has computational citizen Providing become,Providing easy-to-use computational tools to support citizen science has become an important issue.,39378
future to Code be will released work. facilitate,Code will be released to facilitate future work.,39379
Ornstein-Uhlenbeck Brownian studied for motion of evolution. It was trait models and,It was studied for Brownian motion and Ornstein-Uhlenbeck models of trait evolution.,39380
drift parameter of the Ornstein-Uhlenbeck on The the depends of value limit the process.,The limit depends on the value of the drift parameter of the Ornstein-Uhlenbeck process.,39381
"of Even simplicity because used computational method so, its widely and cost. this low is","Even so, this method is widely used because of its simplicity and low computational cost.",39382
"for speed processing existing image varies the augmentation. tools image Moreover, in","Moreover, the image processing speed varies in existing tools for image augmentation.",39383
"person, classes gait for pathological normal, and walks. considered Five assisted each including different are","Five different gait classes are considered for each person, including normal, pathological and assisted walks.",39384
different styles recognized Five walking are with high accuracy.,Five different walking styles are recognized with high accuracy.,39385
that refers for reuse. freely Open data data to available is,Open data refers to data that is freely available for reuse.,39386
of retrieval. attacks video paper zero-example challenging the problem This,This paper attacks the challenging problem of zero-example video retrieval.,39387
"effective encoding conceptually practically and is simple, end-to-end. Dual","Dual encoding is conceptually simple, practically effective and end-to-end.",39388
hierarchical interpretable has policy and The from an structure. resulting our framework,The policy resulting from our framework has an interpretable and hierarchical structure.,39389
experimentally proposed tasks. a robotic validate of set manipulation on We the method,We validate the proposed method experimentally on a set of robotic manipulation tasks.,39390
it under complementary consider three We angles.,We consider it under three complementary angles.,39391
the time visual of performance. both outperforms and method existing in Our terms method,Our method outperforms the existing method in terms of both visual and time performance.,39392
to challenging. super-resolution ground benchmark truth Capturing is data (SR),Capturing ground truth data to benchmark super-resolution (SR) is challenging.,39393
rigorously from both statistical investigate evaluations between agreements perspective. a also We,We also rigorously investigate agreements between both evaluations from a statistical perspective.,39394
divided constraints can two constraints These and constraints. photometric into categories: roughly be geometric,These constraints can be roughly divided into two categories: geometric constraints and photometric constraints.,39395
framework. an combines warping mesh-based proposed image together method in them efficient therefore The,The proposed method therefore combines them together in an efficient mesh-based image warping framework.,39396
registration ConvNets supervised predefined using Thus of for far was training example registrations.,Thus far training of ConvNets for registration was supervised using predefined example registrations.,39397
"However, trivial. is example not registrations obtaining","However, obtaining example registrations is not trivial.",39398
for image registration registration. image designs We deformable flexible affine and propose for ConvNets,We propose flexible ConvNets designs for affine image registration and for deformable image registration.,39399
important hand gestures. unique The umpire using and signals events signals,The umpire signals important events using unique hand signals and gestures.,39400
classifier. results linear a obtained using The are SVM,The results are obtained using a linear SVM classifier.,39401
"recognition approach a we from In this paper, text two-dimensional perspective. scene","In this paper, we approach scene text recognition from a two-dimensional perspective.",39402
performance. degrade to speech automatic recognition causes in speech Non-native systems,Non-native speech causes automatic speech recognition systems to degrade in performance.,39403
end prototyping end to this LDACS-DME in report. coexistence An presented hardware of also is,An end to end hardware prototyping of LDACS-DME coexistence is also presented in this report.,39404
"multi-scale descriptor, More with is a trained SVM Saliency classifier Curviness Histogram (HCS). precisely, Of","More precisely, SVM classifier is trained with a multi-scale descriptor, Histogram Of Curviness Saliency (HCS).",39405
method trained can with set proposed be small The of images. a classification,The proposed classification method can be trained with a small set of images.,39406
optimizing strategy limitations. two each has individual Separately,Separately optimizing each individual strategy has two limitations.,39407
(actor). treated is scenario agent an as Each,Each scenario is treated as an agent (actor).,39408
platform. is The on settings evaluated a E-commerce model online with large,The model is evaluated with online settings on a large E-commerce platform.,39409
potentials. application received its considerable has interest due parsing Human wide to,Human parsing has received considerable interest due to its wide application potentials.,39410
Cloud-RANs. efficient adaptation the Network operation is of essential for,Network adaptation is essential for the efficient operation of Cloud-RANs.,39411
"Unfortunately, intractable mixed-integer programming to problems. highly nonlinear leads it","Unfortunately, it leads to highly intractable mixed-integer nonlinear programming problems.",39412
problem-size is independent achieved and leveraging by the This features. identifying,This is achieved by identifying and leveraging the problem-size independent features.,39413
approach using cameras. The data includes CCD an technique capture economical inexpensive,The approach includes an economical data capture technique using inexpensive CCD cameras.,39414
simulated and the approach captured We physically-based present our using data results data. of real,We present the results of our approach using real captured data and physically-based simulated data.,39415
paper with method a realistic deals for masses. This labeled generating,This paper deals with a method for generating realistic labeled masses.,39416
"fields, labeled scarcely of bioimaging However, large-size dataset many available. the in is","However, in many bioimaging fields, the large-size of labeled dataset is scarcely available.",39417
learning results show outperforms SCC-rFMQ reinforcement Experimental other algorithms. that,Experimental results show that SCC-rFMQ outperforms other reinforcement learning algorithms.,39418
in has Thus of diseases. brain it wide and diagnosis detection use,Thus it has wide use in detection and diagnosis of brain diseases.,39419
CT reported not segmentation But significant. very is on works unfortunately,But unfortunately reported works on CT segmentation is not very significant.,39420
"completely proposed accurate automatic. and fast, is simple, method The","The proposed method is simple, fast, accurate and completely automatic.",39421
multislice single scan run. in handle It can CT,It can handle multislice CT scan in single run.,39422
in slices handle to masks Two created better a are way. nasal types of,Two types of masks are created to handle nasal slices in a better way.,39423
has brain potential high of Successive demonstrated matter in mask segmentation. propagation reported very,Successive propagation of brain matter mask has demonstrated very high potential in reported segmentation.,39424
any or images analysis. Resulted image segmented disease for can brain diagnosis be used further,Resulted segmented images can be used for any brain disease diagnosis or further image analysis.,39425
"eyelids, lashes, brows eye, to the the sclera, region skin. around refers including and Periocular","Periocular refers to the region around the eye, including sclera, eyelids, lashes, brows and skin.",39426
with of captured a Experiments camera. digital database are a done images periocular with,Experiments are done with a database of periocular images captured with a digital camera.,39427
"systems scheduling In NOMA considered. under this fairness is constraints paper, temporal for","In this paper, scheduling for NOMA systems under temporal fairness constraints is considered.",39428
"accomplish variable work, this elimination this task. a to algorithm In proposed is","In this work, a variable elimination algorithm is proposed to accomplish this task.",39429
such semantic in important practical segmentation applications role robots. plays an and self-driving Real-time as,Real-time semantic segmentation plays an important role in practical applications such as self-driving and robots.,39430
accuracy little Most consideration with improving on efficiency. semantic research segmentation estimation on focuses,Most semantic segmentation research focuses on improving estimation accuracy with little consideration on efficiency.,39431
high-accuracy previous that high-speed Several emphasize often studies fail to inference results. produce segmentation,Several previous studies that emphasize high-speed inference often fail to produce high-accuracy segmentation results.,39432
leveraging arguments. these addresses paper by This limitations modeling Quantum-Statistical,This paper addresses these limitations by leveraging Quantum-Statistical modeling arguments.,39433
"we this end, inference To approximation. (ADM) into of density cast matrix attention an NA","To this end, we cast NA into inference of an attention density matrix (ADM) approximation.",39434
perform datasets. benchmark experiments We with challenging,We perform experiments with challenging benchmark datasets.,39435
"of several in yields our show, evaluation favorable approach As we outcomes terms metrics.","As we show, our approach yields favorable outcomes in terms of several evaluation metrics.",39436
improved deep drastically Performance in has datasets with advances benchmark on learning.,Performance on benchmark datasets has drastically improved with advances in deep learning.,39437
synthetic exaggerated is real and domain especially shift This between datasets.,This domain shift is especially exaggerated between synthetic and real datasets.,39438
"of augmentation we propose composed functions. To learned augmentation this, network physically-based a address","To address this, we propose a learned augmentation network composed of physically-based augmentation functions.",39439
has Performance in the proposed different been evaluated simulations PSCs. of of method,Performance of the proposed method has been evaluated in simulations of different PSCs.,39440
"through selected Also, its with has literature been in two the methods simulations. performance compared","Also, its performance has been compared with two selected methods in the literature through simulations.",39441
highlight method. Comparisons proposed the superiority the of,Comparisons highlight the superiority of the proposed method.,39442
"required. relaxation is NP-hard, it Since models are","Since it is NP-hard, relaxation models are required.",39443
proposed derived model. to simple solve algorithm the update been has SPM effective A yet,A simple yet effective update algorithm has been derived to solve the proposed SPM model.,39444
efficiency on Experimental SPM of tasks matching the results feature effectiveness several method. and demonstrate,Experimental results on several feature matching tasks demonstrate the effectiveness and efficiency of SPM method.,39445
human requires Successful model predictive Human-Robot behavior. a of collaboration,Successful Human-Robot collaboration requires a predictive model of human behavior.,39446
and and Our detection outperforms approaches in model state-of-the-art affordance anticipation. both activity,Our model outperforms state-of-the-art approaches in both activity and affordance detection and anticipation.,39447
"continuous. can discrete Mutation which different leads (""traits""), be genetic or to types","Mutation leads to different genetic types (""traits""), which can be discrete or continuous.",39448
"Niklaus the we paper, al. et this of In replicate work","In this paper, we replicate the work of Niklaus et al.",39449
"on claims which high on the quality Adaptive video frame Convolution, task. Separable results interpolation","on Adaptive Separable Convolution, which claims high quality results on the video frame interpolation task.",39450
"In lengths varying of initial were rigid diFabs double-stranded (dsDNA) by DNA joined considered. work,","In initial work, diFabs joined by varying lengths of rigid double-stranded DNA (dsDNA) were considered.",39451
approach using and our methods We compare existing contrast examples. with three,We compare and contrast our approach with existing methods using three examples.,39452
. R-CNN Faster and two baseline Munich others: the on GoogleEarth,the baseline Faster R-CNN on two others: Munich and GoogleEarth .,39453
"LaSOT, paper, for benchmark a Object this we Single present Large-scale Tracking. high-quality In","In this paper, we present LaSOT, a high-quality benchmark for Large-scale Single Object Tracking.",39454
frameworks and for decision social Weighted offer argumentation tool support media analysis. a bipolar,Weighted bipolar argumentation frameworks offer a tool for decision support and social media analysis.,39455
"graphs. cyclic iterative well in was not recently, very of procedures these convergence understood Until","Until recently, convergence of these iterative procedures was not very well understood in cyclic graphs.",39456
this avoided demonstrate by continuizing semantics. be can problems We tradeoff that divergence without,We demonstrate that divergence problems can be avoided without this tradeoff by continuizing semantics.,39457
"results show detector on adaptation. for we In mined object preliminary tracks using addition, the","In addition, we show preliminary results on using the mined tracks for object detector adaptation.",39458
systems to conditions can . design The synthetic physical identified oscillatory be used,The physical conditions identified can be used to design synthetic oscillatory systems .,39459
antenna efficiency parameter is MIMO important system. radiation for Antenna an,Antenna radiation efficiency is an important parameter for MIMO antenna system.,39460
"of Nevertheless, the on description in generation. evidence not image has support there idea been","Nevertheless, there has not been evidence in support of the idea on image description generation.",39461
single layer. matching antenna into to integrate a is the The circuit designed element,The antenna element is designed to integrate the matching circuit into a single layer.,39462
are real as for presented. Experimental results sequences well synthetic as,Experimental results for synthetic as well as real sequences are presented.,39463
both with range of measurements performance datasets. fraction on achieve improvements We significant a small,We achieve significant performance improvements with a small fraction of range measurements on both datasets.,39464
platform the provide Monstar using qualitative our PMDTec We sensor. also results from,We also provide qualitative results from our platform using the PMDTec Monstar sensor.,39465
Both been to are maturity market from developed GmbH. available Nerian have and systems Vision,Both systems have been developed to market maturity and are available from Nerian Vision GmbH.,39466
"process, scheme During to is improve an the the performance. training also employed adversarial further","During the training process, an adversarial scheme is also employed to further improve the performance.",39467
of We proposed challenging detection evaluation provide face datasets. detector face on results the unconstrained,We provide evaluation results of the proposed face detector on challenging unconstrained face detection datasets.,39468
those compared MUSIC (multiple The test results method. of classification) signal the with are,The test results are compared with those of the MUSIC (multiple signal classification) method.,39469
assignment performance optimal evaluation. used quantitative metric subpattern The for is (OSPA),The optimal subpattern assignment (OSPA) metric is used for quantitative performance evaluation.,39470
NP-hard. We CAM that show is query,We show that CAM query is NP-hard.,39471
"Based novel pruning, efficient dominance-based a propose algorithm on high-quality an which results. generates we","Based on a novel dominance-based pruning, we propose an efficient algorithm which generates high-quality results.",39472
and demonstrate highly that experiments efficient algorithm our is results. quality The produces,The experiments demonstrate that our algorithm is highly efficient and produces quality results.,39473
"little this be In accessed with code. analysis change the framework, different methods can in","In this framework, different methods can be accessed with little change in the analysis code.",39474
"Similarly, methods incorporated package. new easily the be can into","Similarly, new methods can easily be incorporated into the package.",39475
efficiency via be Further improvements estimation obtained can augmentation.,Further estimation improvements can be obtained via efficiency augmentation.,39476
is congenital disease Improving fetal of heart diagnosis therefore critical.,Improving fetal diagnosis of congenital heart disease is therefore critical.,39477
"normal HLHS. We trained to and then TOF, between view by hearts, distinguish models","We then trained models to distinguish by view between normal hearts, TOF, and HLHS.",39478
data provide key with deployment backhaul to A is connections challenge rate. high sufficiently,A key deployment challenge is to provide backhaul connections with sufficiently high data rate.,39479
algorithms suffer camera to positions from set. training overfitting Many the in,Many algorithms suffer from overfitting to camera positions in the training set.,39480
perform image tracking methods vision-based domain. of current the the in tracking Most,Most of the current vision-based tracking methods perform tracking in the image domain.,39481
service. the a web regulates The of expressiveness information semantic,The semantic information regulates the expressiveness of a web service.,39482
method. our of Experimental efficiency results show the proposed,Experimental results show the efficiency of our proposed method.,39483
"age However, for GAN based contribute previous facial parts less synthesis. in these local methods","However, these local facial parts contribute less in previous GAN based methods for age synthesis.",39484
of experiments quantitative superiority the method state-of-the-arts. and proposed show the over other Qualitative,Qualitative and quantitative experiments show the superiority of the proposed method over other state-of-the-arts.,39485
"amounts continuously data. of systems Observation Nowadays, modern generate huge Earth","Nowadays, modern Earth Observation systems continuously generate huge amounts of data.",39486
"to research new allowed Recently, field. this methods in advance deep learning generation significantly","Recently, new generation deep learning methods allowed to significantly advance research in this field.",39487
assumptions distribution distribution learning-based estimation. usually to simplify age methods make Label,Label distribution learning-based methods usually make distribution assumptions to simplify age estimation.,39488
distributions age learns network label way. refines in iteratively learning Evolutionary and an,Evolutionary network learns and refines age label distributions in an iteratively learning way.,39489
implemented ethics into these be how systems? But should,But how ethics should be implemented into these systems?,39490
set logic epistemic An this is a language. of program rules in,An epistemic logic program is a set of rules in this language.,39491
"on accuracy. high hand, system, able A to is still the provide multimodal other","A multimodal system, on the other hand, is still able to provide high accuracy.",39492
cancer HeLa moreover the and exhibit between our simulated similarities cells. We cells,We exhibit moreover the similarities between our simulated cells and HeLa cancer cells.,39493
of The sparse written and create ambiguity descriptions alignment challenges. instructions significant,The sparse descriptions and ambiguity of written instructions create significant alignment challenges.,39494
test site with has man-made objects. buried This several prepared been,This test site has been prepared with several buried man-made objects.,39495
fully buried show be Initial results that detected. targets can,Initial results show that fully buried targets can be detected.,39496
is of Reality topic a nowadays. foremost Augmented interest,Augmented Reality is a topic of foremost interest nowadays.,39497
main real-world in blend seamlessly is Its to goal scenes. virtual content,Its main goal is to seamlessly blend virtual content in real-world scenes.,39498
in industrial be It purposes and for used residential sectors. different can,It can be used for different purposes in residential and industrial sectors.,39499
"an increasing new Thus, developed recent number have in years. of been algorithms","Thus, an increasing number of new algorithms have been developed in recent years.",39500
"of requirements describes platform. structure, this paper the building and This steps,","This paper describes the steps, structure, and requirements of building this platform.",39501
this understand to help is study Case platform. presented,Case study is presented to help understand this platform.,39502
"real data often losses. not while Synthetic unsupervised heuristic to videos, generalize require methods does","Synthetic data often does not generalize to real videos, while unsupervised methods require heuristic losses.",39503
"as for interpolation optical a task flow. investigate frame we Here, proxy","Here, we investigate frame interpolation as a proxy task for optical flow.",39504
"a we train Using interpolation. real for unsupervised movies, CNN temporal","Using real movies, we train a CNN unsupervised for temporal interpolation.",39505
"but Such cannot a implicitly handle untextured network estimates regions. motion,","Such a network implicitly estimates motion, but cannot handle untextured regions.",39506
starting complex a to point model interactions. models are suitable social Multi-agent,Multi-agent models are a suitable starting point to model complex social interactions.,39507
practices agents in practical We argue that standard facilitate interactions. social of social the reasoning,We argue that social practices facilitate the practical reasoning of agents in standard social interactions.,39508
"available many only form. In cases, pixel is however, in data","In many cases, however, data is only available in pixel form.",39509
a domain high-level a method generating We (e.g. directly in images for present,We present a method for generating images directly in a high-level domain (e.g.,39510
"for pairwise data. the real strokes), need brush without","brush strokes), without the need for real pairwise data.",39511
the of embeddings knowledge increases. gradually the RNKN exhibits number as training of epochs interpretation,RNKN gradually exhibits the interpretation of knowledge embeddings as the number of training epochs increases.,39512
"network multilevel (PMA). corresponding pose-guided contents, a attention To propose visual multi-granularity we the exploit","To exploit the multilevel corresponding visual contents, we propose a pose-guided multi-granularity attention network (PMA).",39513
fitting is One of studied of the the main statistics models. in problems,One of the main problems studied in statistics is the fitting of models.,39514
"with explain would Ideally, like we as parameters a few dataset large as possible. to","Ideally, we would like to explain a large dataset with as few parameters as possible.",39515
this automatizing been have numerous process. at There attempts,There have been numerous attempts at automatizing this process.,39516
"angle algorithm, the holonomic HELARS. least We call this extended or regression","We call this the holonomic extended least angle regression algorithm, or HELARS.",39517
are for SINR both compared resulting The experiments and interference discussed.,The resulting SINR for both interference experiments are compared and discussed.,39518
and HCI the these to MAHNOB DEAP applied schemes thus databases. We,We thus applied these schemes to the HCI MAHNOB and DEAP databases.,39519
"we Head model. paper, an Detector, this trainable propose Convolutional head detection end-to-end In FCHD-Fully","In this paper, we propose FCHD-Fully Convolutional Head Detector, an end-to-end trainable head detection model.",39520
This makes memory our low lightweight with and inference model requirements. time,This makes our model lightweight with low inference time and memory requirements.,39521
"are bounding when labeling that still the ambiguities observe introduced However, we boxes.","However, we observe that ambiguities are still introduced when labeling the bounding boxes.",39522
architectures the loss with improves nearly additional no various of greatly computation. Our accuracies localization,Our loss greatly improves the localization accuracies of various architectures with nearly no additional computation.,39523
github.com/yihui-he/KL-Loss code at: Our and are available models .,Our code and models are available at: github.com/yihui-he/KL-Loss .,39524
model the several once fine-tuning. without tasks to This allows explicit solve at,This allows the model to solve several tasks at once without explicit fine-tuning.,39525
"designed approaches, our e.g. When model comparing to are for specifically state-of-the-art which","When comparing our model to state-of-the-art approaches, which are specifically designed for e.g.",39526
augment analyses data. services by can qualitative with gerontologists Such quantitative done,Such services can augment qualitative analyses done by gerontologists with quantitative data.,39527
The system as a Arianna+. mechanism home implemented referred reasoning has to in been smart,The reasoning mechanism has been implemented in a smart home system referred to as Arianna+.,39528
Results processes. indicate and scalability human recognition for activity robustness,Results indicate scalability and robustness for human activity recognition processes.,39529
and important number simple allows analytical a restriction a yet obtaining Such results. of,Such a restriction allows obtaining a number of simple and yet important analytical results.,39530
findings variance illustrate by in-depth evolution. our We of analysis the,We illustrate our findings by in-depth analysis of the variance evolution.,39531
"This understand aims problem how two to (sisters, are brothers, and if people etc.) family-related","This problem aims to understand if two people are family-related and how (sisters, brothers, etc.)",39532
Solving applications. give to can varied problem this and rise tasks,Solving this problem can give rise to varied tasks and applications.,39533
"problem. the article, we this Kinship-Verification solving Deep propose approach a using Learning for In","In this article, we propose using a Deep Learning approach for solving the Kinship-Verification problem.",39534
"mask harmonized multi-site with imaging Even can advantages of studies. these the sequences, site-dependent variability","Even with harmonized imaging sequences, site-dependent variability can mask the advantages of these multi-site studies.",39535
edges increase notable for observed. in A was and ICC functional networks values resting-state also,A notable increase in ICC values for functional edges and resting-state networks was also observed.,39536
in to regression. found identifiability were affected Improvements not by be signal global,Improvements in identifiability were not found to be affected by global signal regression.,39537
"include only the metabolic information models current gene However, level. at","However, current metabolic models only include information at the gene level.",39538
in context license our method the We of plates. automotive demonstrate,We demonstrate our method in the context of automotive license plates.,39539
Estimation of a facial shapes central face plays for animation. transfer and role,Estimation of facial shapes plays a central role for face transfer and animation.,39540
"very tissue this therefore, is, segmentation Brain at challenging. age","Brain tissue segmentation at this age is, therefore, very challenging.",39541
dataset code be source Our open-sourced. will synthetic and,Our source code and synthetic dataset will be open-sourced.,39542
significant SR. in The accurate of a role correspondence plays generation video,The generation of accurate correspondence plays a significant role in video SR.,39543
"deep in generation. are methods flows optical correspondence for LR However, existing based used learning","However, LR optical flows are used in existing deep learning based methods for correspondence generation.",39544
"flows. the Then, optical motion is to performed compensation HR according","Then, motion compensation is performed according to the HR optical flows.",39545
"real perform images. numerical we Finally, and synthetic experiments the on both","Finally, we perform the numerical experiments on both synthetic and real images.",39546
The vessel retinal on out quantitive is carried images. validation,The quantitive validation is carried out on retinal vessel images.,39547
that a the proposed promising The results performance. indeed indicate achieves model,The results indicate that the proposed model indeed achieves a promising performance.,39548
the task. in which vital We plays role this pooling a investigate method global,We investigate the global pooling method which plays a vital role in this task.,39549
the ILSVRC and the on We dataset. evaluate localization ability pixel-level classification benchmark,We evaluate the classification and pixel-level localization ability on the ILSVRC benchmark dataset.,39550
We that multi-task with GWAP a our module combines propose class-specific R-FCN. framework,We propose a multi-task framework that combines our class-specific GWAP module with R-FCN.,39551
with framework truth image-level and bounding The labels. ground few trained boxes is large-scale,The framework is trained with few ground truth bounding boxes and large-scale image-level labels.,39552
We dataset. on this PASCAL evaluate VOC framework,We evaluate this framework on PASCAL VOC dataset.,39553
"accuracy model is of The hierarchical through the ensures which detection. obtained the training,","The model is obtained through hierarchical training, which ensures the accuracy of the detection.",39554
advances recent on An learning. capitalize in alternative approach deep to would be,An alternative approach would be to capitalize on recent advances in deep learning.,39555
vision. estimation an computer in challenging yet pose task essential is Human,Human pose estimation is an essential yet challenging task in computer vision.,39556
"architecture model. convolutional network with work, this In attention the a novel we combined proposed","In this work, we proposed a convolutional network architecture combined with the novel attention model.",39557
attention (ACNN). We network neural convolutional named it,We named it attention convolutional neural network (ACNN).,39558
ACNN to focus regions features. different on of input learns specific,ACNN learns to focus on specific regions of different input features.,39559
"as background the ""nothing-regions"", body filtrate Early and parts. such stages redundant","Early stages filtrate the ""nothing-regions"", such as background and redundant body parts.",39560
State-of-the-art obtained the is MPII on performance benchmarks.,State-of-the-art performance is obtained on the MPII benchmarks.,39561
to model's Adversarial aim images by a a perturbing decision minimally target image. change target,Adversarial images aim to change a target model's decision by minimally perturbing a target image.,39562
"task. person we scheme attention improve paper, present this mechanism In an re-identification to","In this paper, we present an attention mechanism scheme to improve person re-identification task.",39563
"measurement, are oscilloscope digital improving and discussed. impairments system Device resolution","Device measurement, digital oscilloscope impairments and improving system resolution are discussed.",39564
recognition text for paired supervisory a using without any This work data. visual method presents,This work presents a method for visual text recognition without using any paired supervisory data.,39565
"two-stage novel intra-mode. in approach a decision propose HEVC partition Therefore, CU we","Therefore, we propose a novel two-stage CU partition decision approach in HEVC intra-mode.",39566
Most effects apply image. existing the to the whole methods artistic of,Most of the existing methods apply artistic effects to the whole image.,39567
"is filters. the to allow enough new method flexible of our Also, addition","Also, our method is flexible enough to allow the addition of new filters.",39568
We segmentation pre-trained segmentation dataset R-CNN instance on the network. COCO as use the Mask,We use pre-trained Mask R-CNN instance segmentation on the COCO dataset as the segmentation network.,39569
filters performed. use the results different on of are Experimental,Experimental results on the use of different filters are performed.,39570
"of amount target perform propose a to limited we data. Then, with adaptation","Then, we propose to perform adaptation with a limited amount of target data.",39571
three by observations. The is framework motivated proposed,The proposed framework is motivated by three observations.,39572
"between a Finally, evaluate association posteriori mode connectivity longitudinal outcomes. the we weights with neurocognitive","Finally, we evaluate a posteriori the association between connectivity mode weights with longitudinal neurocognitive outcomes.",39573
engineer inspections authorities municipal volunteers. structural These gfare by typically through conducted,These inspections gfare typically conducted by municipal authorities through structural engineer volunteers.,39574
"limited spatial However, phosphenic very of images terms resolution (e.g. are in","However, phosphenic images are very limited in terms of spatial resolution (e.g.",39575
"this of recognition under the indoor environments we study In vision. prosthetic work, simulated","In this work, we study the recognition of indoor environments under simulated prosthetic vision.",39576
"can become easily. more sensors depth prevalent Meanwhile, be so maps depth acquired nowadays,","Meanwhile, depth sensors become prevalent nowadays, so depth maps can be acquired more easily.",39577
"semantic studies are the segmentation task. few focus on there RGB-D However, that","However, there are few studies that focus on the RGB-D semantic segmentation task.",39578
to challenge. effectiveness depth information the a improve Exploiting is performance,Exploiting the depth information effectiveness to improve performance is a challenge.,39579
concept multiple modalities. sensor this We to data from extend incorporate Stixel,We extend this Stixel concept to incorporate data from multiple sensor modalities.,39580
variograms on signals extensions of Graph variograms are defined continuous for Euclidean space. used,Graph variograms are extensions of variograms used for signals defined on continuous Euclidean space.,39581
"datasets However, to large are acquire. costly such labeled","However, such large labeled datasets are costly to acquire.",39582
scratch. SNNs from addresses the neuromorphic learning for temporal tutorial This,This tutorial addresses neuromorphic SNNs for temporal learning from the scratch.,39583
large remains in significant challenge a Detecting over objects areas satellite imagery analytics. small,Detecting small objects over large areas remains a significant challenge in satellite imagery analytics.,39584
scales For (e.g. very different objects of,For objects of very different scales (e.g.,39585
"this In a proposed. enhancement algorithm is low-light paper, image gradient-based","In this paper, a gradient-based low-light image enhancement algorithm is proposed.",39586
"addition, constraints In intensity-range the the for image we integration. involve","In addition, we involve the intensity-range constraints for the image integration.",39587
Experiments effectively proposed the demonstrate image enhancement images. low-light enhance that gradient-based can low-light the,Experiments demonstrate that the proposed gradient-based low-light image enhancement can effectively enhance the low-light images.,39588
"In we of detectors scratch. object learning principles for DSOD, set contribute a design from","In DSOD, we contribute a set of design principles for learning object detectors from scratch.",39589
than models. better consistently more much compact with DSOD methods achieves state-of-the-art results the,DSOD achieves consistently better results than the state-of-the-art methods with much more compact models.,39590
the analytical also verify an based robustness our We of individual model develop to prediction.,We also develop an individual based model to verify the robustness of our analytical prediction.,39591
"coevolutionary population cause unexpected However, prey patterns. small selection can sizes, for","However, for small population sizes, prey selection can cause unexpected coevolutionary patterns.",39592
"Creativity of value, often in interestingness. and novelty, terms formulated is surprise","Creativity is often formulated in terms of novelty, value, surprise and interestingness.",39593
Experiments benefits of the network. proposed benefit show the the gLGCN,Experiments show the benefit of the benefits the proposed gLGCN network.,39594
Play-style learning. video introduce gameplay via problem Let's commentary the We machine of of generating,We introduce the problem of generating Let's Play-style commentary of gameplay video via machine learning.,39595
future commentary Play demonstrate promising generation. We towards Let's results,We demonstrate promising results towards future Let's Play commentary generation.,39596
pair of It removesthe ofdual-lens offline calibration every need individual for cameras.,It removesthe need of individual offline calibration for every pair ofdual-lens cameras.,39597
markers the introduced fiducial are to improve precision. overall Visual,Visual fiducial markers are introduced to improve the overall precision.,39598
segmentation semantic supervision prior detection. objectness an object and Class-agnostic knowledge both provides for segmentation,Class-agnostic segmentation supervision provides an objectness prior knowledge for both semantic segmentation and object detection.,39599
"segmentation framework, can other. the sufficiently and each detection boost In proposed","In the proposed framework, detection and segmentation can sufficiently boost each other.",39600
"introduced costs extra stage. test the at are no computational Therefore,","Therefore, no extra computational costs are introduced at the test stage.",39601
"way we this confidence approachin paper, In an matching. unsupervised stereo inference in present","In this paper, we present confidence inference approachin an unsupervised way in stereo matching.",39602
(DNNs) state-of-the-art recently achieving been Deep Networks Neu-ral performance. have,Deep Neu-ral Networks (DNNs) have recently been achieving state-of-the-art performance.,39603
"dense Weshow the confidence is with identical that newly assumption introduced relaxed. map,the","Weshow that with the newly introduced dense confidence map,the identical assumption is relaxed.",39604
"featureless network In pixels (e.g.,noisy occlusions, regions) andfocusonhigh-confidence pixels. learns toattenuatelow-confidence input, practice,the","In practice,the network learns toattenuatelow-confidence pixels (e.g.,noisy input, occlusions, featureless regions) andfocusonhigh-confidence pixels.",39605
trees (e.g. handle The of various can types package,The package can handle trees of various types (e.g.,39606
to world. the The in the Knowledge uses (KG) facts graph the triples real describe,The Knowledge graph (KG) uses the triples to describe the facts in the real world.,39607
widely in applications. analysis has been used It and intelligent,It has been widely used in intelligent analysis and applications.,39608
"possible However, noises in are the constructing. inevitably conflicts introduced of process and","However, possible noises and conflicts are inevitably introduced in the process of constructing.",39609
is neural a crisscrossing network model structure. The,The model is a crisscrossing neural network structure.,39610
pillars the computing. Deformable one of biomedical registration image of been has,Deformable registration has been one of the pillars of biomedical image computing.,39611
using machines structured is support learned vector (LSSVM). This latent combination,This combination is learned using latent structured support vector machines (LSSVM).,39612
"the purpose, we the univocal problem, exploit that matching of i.e. To nature","To that purpose, we exploit the univocal nature of the matching problem, i.e.",39613
all. or must at a corresponded observation feature with corresponded not single be every,every observation must be corresponded with a single feature or not corresponded at all.,39614
"circles tangent use filling idea first for surrounded the and Afterwards, we our circle.","Afterwards, we use our first idea for filling the tangent circles and surrounded circle.",39615
propose a deal unified nonlinear simple measurements. to framework with we Here and,Here we propose a simple and unified framework to deal with nonlinear measurements.,39616
"In introduce in R written alphastable. the language we called this paper, package a","In this paper, we introduce a package written in the R language called alphastable.",39617
"results are Further, extensive simulation the considerations NR behind to design. the presented PUCCH explain","Further, extensive simulation results are presented to explain the considerations behind the NR PUCCH design.",39618
scalable techniques. localization to use An efficient retrieval to and visual approach is image,An efficient and scalable approach to visual localization is to use image retrieval techniques.,39619
"drastically e.g. different image illumination However, across conditions, retrieval","However, image retrieval across drastically different illumination conditions, e.g.",39620
"many achieved models person in Recently, benchmarks. deep re-identification success learning-based great have","Recently, deep learning-based person re-identification models have achieved great success in many benchmarks.",39621
"To be randomly rectangles use pedestrian to images. block we occlude specific,","To be specific, we use block rectangles to randomly occlude pedestrian images.",39622
"use train model. Finally, we baseline augmented the datasets to","Finally, we use the augmented datasets to train baseline model.",39623
"structured-light-based depth indoor sensors For (e.g. scenes,","For indoor scenes, structured-light-based depth sensors (e.g.",39624
"are short-range, depth albeit provide maps. dense, able Kinect) to","Kinect) are able to provide dense, albeit short-range, depth maps.",39625
captured various camera aspects investigate. to have images The,The camera captured images have various aspects to investigate.,39626
"depends of the emphasis on interesting the research regions. Generally,","Generally, the emphasis of research depends on the interesting regions.",39627
"focus detection be object or could Sometimes color text scene analysis. on the segmentation,","Sometimes the focus could be on color segmentation, object detection or scene text analysis.",39628
always from associated machines learning the to provided samples. properties learn The,The learning machines always learn from the properties associated to provided samples.,39629
in results images. on provided text analysis The reasonable convolutional non-cursive approach natural appeared,The convolutional approach provided reasonable results on non-cursive text analysis appeared in natural images.,39630
empirically kernels. various selected through process image pyramid Each,Each pyramid image process through various empirically selected kernels.,39631
"between We maps the manipulating and of tasks consider analyzing shapes. representing,","We consider the tasks of representing, analyzing and manipulating maps between shapes.",39632
in end-to-end We model a use image compression. entropy propose for optimized context-adaptive,We propose a context-adaptive entropy model for use in end-to-end optimized image compression.,39633
The are independent DL. language and results our any of specific,The language and our results are independent of any specific DL.,39634
of some open and discuss that we state also to be interest. believe We problems,We also state and discuss some open problems that we believe to be of interest.,39635
neural diverse in (SP) signal applications networks found processing problems. (DNNs) Deep have,Deep neural networks (DNNs) have found applications in diverse signal processing (SP) problems.,39636
problem. on The (ATR) radar a automatic hybrid-NN is target proposed tested recognition,The proposed hybrid-NN is tested on a radar automatic target recognition (ATR) problem.,39637
proposed human deep-learning based Conclusion: method outperformed experts. The three CADx,Conclusion: The proposed deep-learning based CADx method outperformed three human experts.,39638
"boundary Meanwhile, proposed is extraction to another generate branch features. information boundary","Meanwhile, another boundary information extraction branch is proposed to generate boundary features.",39639
our verify results. that framework achieves large-scale on and benchmarks state-of-the-art Experiments four analysis new,Experiments and analysis on four large-scale benchmarks verify that our framework achieves new state-of-the-art results.,39640
arrays frequencies. antenna millimeter-wave the (mmWave) at This work demonstrates of operating sensitivity lens,This work demonstrates the sensitivity of lens antenna arrays operating at millimeter-wave (mmWave) frequencies.,39641
and ill-posed inversion problem. a is typical inverse Waveform non-linear,Waveform inversion is a typical non-linear and ill-posed inverse problem.,39642
"years, solve In waveform recent way promising methods data-driven inversion become problem. the to a","In recent years, data-driven methods become a promising way to solve the waveform inversion problem.",39643
"suffer frameworks deep over-fitting and most from issue. learning generalization However,","However, most deep learning frameworks suffer from generalization and over-fitting issue.",39644
issue. We on strategy a alleviate generalization learning to transfer the based develop VelocityGAN further,We further develop a transfer learning strategy based on VelocityGAN to alleviate the generalization issue.,39645
to uncertainty scores way efficient We propose an output better calibrated from neural networks.,We propose an efficient way to output better calibrated uncertainty scores from neural networks.,39646
increase fused performance. be information can segmentation The to harnessed,The fused information can be harnessed to increase segmentation performance.,39647
"for schemes classification. simultaneous propose fracture analyze and localization thoroughly Therefore, several we and","Therefore, we propose and thoroughly analyze several schemes for simultaneous fracture localization and classification.",39648
"localization auxiliary We general, performance. task, an improves in show the that using classification","We show that using an auxiliary localization task, in general, improves the classification performance.",39649
make plan and We available. dataset code both to the,We plan to make both the dataset and code available.,39650
networks. distribution the of angular stone grids represent Electric,Electric grids represent the angular stone of distribution networks.,39651
real meant recognition systems are to world. Visual in the work,Visual recognition systems are meant to work in the real world.,39652
"this i.e. deals realistic Within generalization, context, very domain scenario with a","Within this context, a very realistic scenario deals with domain generalization, i.e.",39653
neighboring consider The distribution pattern. the social interaction and their metrics both,The social metrics consider both the neighboring distribution and their interaction pattern.,39654
"independently. Traditionally, two tackled been tasks the have","Traditionally, the two tasks have been tackled independently.",39655
is randomly varying eventually on lighting conditions based multi-shot strategy A discussed.,A multi-shot strategy based on randomly varying lighting conditions is eventually discussed.,39656
Our major steps. of approach consists two,Our approach consists of two major steps.,39657
for results. to supplementary Please the materials all refer,Please refer to the supplementary materials for all results.,39658
"to features. Specifically, deep adaptive we convolutional the new a exploit learned layer propose","Specifically, we propose a new adaptive convolutional layer to exploit the learned deep features.",39659
independent each other overall are of the and These choices performance. impact substantially not,These choices are not independent of each other and substantially impact the overall performance.,39660
also follows are of datasets predictors misleading. It inherently that different comparisons using,It also follows that comparisons of predictors using different datasets are inherently misleading.,39661
only such prevails in imaging. as fields astronomical noise traditionally specific Poisson,Poisson noise traditionally prevails only in specific fields such as astronomical imaging.,39662
deconvolutional connections. The with and architecture along convolutional hybrid layers symmetric a incorporates of,The architecture incorporates a hybrid of convolutional and deconvolutional layers along with symmetric connections.,39663
"of based SR highly performance depends on representation coeffcients. reconstruction, learned however, The learning","The performance of learning based SR reconstruction, however, highly depends on learned representation coeffcients.",39664
"makes localize position the of to best guess rough Instead, a iris. it the","Instead, it makes a best guess to localize the rough position of the iris.",39665
of to are verify benchmark researchers their Lots for algorithms. released datasets,Lots of benchmark datasets are released for researchers to verify their algorithms.,39666
segmentation many Semantic studied been has years. for,Semantic segmentation has been studied for many years.,39667
"has a made Since (DNN), tremendous the of Network Deep Neural segmentation progress. emergence","Since the emergence of Deep Neural Network (DNN), segmentation has made a tremendous progress.",39668
"is area a conclusion drawn. Finally, this in","Finally, a conclusion in this area is drawn.",39669
"process often information their ability sources. observers are in simultaneous However, human to limited","However, human observers are often limited in their ability to process simultaneous information sources.",39670
tracking-by-detection Convolutional Neural the receives framework integration the The Networks growing attentions through with (CNNs).,The tracking-by-detection framework receives growing attentions through the integration with the Convolutional Neural Networks (CNNs).,39671
"variations. objects track based appearance Existing fail with severe tracking-by-detection to methods, however,","Existing tracking-by-detection based methods, however, fail to track objects with severe appearance variations.",39672
"assigned as which are the non-orthogonal In sequences signatures. their with CDMA, serve users non-orthogonal","In non-orthogonal CDMA, the users are assigned with non-orthogonal sequences which serve as their signatures.",39673
"propose point paper, edge and clouds. this novel unorganized for algorithms detection In corner we","In this paper, we propose novel edge and corner detection algorithms for unorganized point clouds.",39674
various Ras present of superfamily. tokens revealed on GTPases protein the Application in token,Application of protein token on Ras GTPases revealed various tokens present in the superfamily.,39675
to cloud systems image patches morphological and separable This magnitudes boundaries. gradient integrates method,This method integrates morphological image gradient magnitudes to separable cloud systems and patches boundaries.,39676
flexible is extendable to imagery. single- method from The proposed and multi-spectral,The proposed method is flexible and extendable from single- to multi-spectral imagery.,39677
vehicles unmanned system. of considered is and as distributed A manned a mixed aerial group,A mixed group of manned and unmanned aerial vehicles is considered as a distributed system.,39678
of A be the fulfilled lattice which it. by to may matches system tasks,A lattice of tasks which may be fulfilled by the system matches to it.,39679
"external operations. defined multiplication which defines the linear is operation An correspondingly logic at lattice,","An external multiplication operation is defined at the lattice, which defines correspondingly linear logic operations.",39680
"mainly the purpose Thus, system choice the consequence. is","Thus, the choice is mainly the system purpose consequence.",39681
"resources. as Microgrids, islandable viable of energy power as considered small-scale are promotors renewable systems,","Microgrids, as small-scale islandable power systems, are considered as viable promotors of renewable energy resources.",39682
"how limit by a on microgrid. be there hosted is PV However, much a can","However, there is a limit on how much PV can be hosted by a microgrid.",39683
"introduced impacts, potential capacity avoid hosting of the negative concept used. is such and To","To avoid such potential negative impacts, the concept of hosting capacity is introduced and used.",39684
based primarily analyses. steadystate are Hosting studies done capacity on,Hosting capacity studies are primarily done based on steadystate analyses.,39685
"result, An exhibits good commonly a EM set over-complete i.e. unmixing","An over-complete EM set commonly exhibits a good unmixing result, i.e.",39686
are the reliable. unmixed abundances values less,the unmixed abundances values are less reliable.,39687
"greedy condition set scheme i.e. the Our number, EM reduction improves reducing by the","Our greedy reduction scheme improves the EM set by reducing the condition number, i.e.",39688
"possible. the while set's stability, as keeping error low reconstruction the enhancing as","enhancing the set's stability, while keeping the reconstruction error as low as possible.",39689
size. The optimal set hint sequence the its set EM resulting gives and to,The resulting set sequence gives hint to the optimal EM set and its size.,39690
allows periods. time of infectious for short high probability and unrealistically exposed This a,This allows for a high probability of unrealistically short exposed and infectious time periods.,39691
effectively can its improve energy utilization. The RBC adaptive (ARBC),The adaptive RBC (ARBC) can effectively improve its energy utilization.,39692
DeepGlobe This to Challenge. contribution presents Detection Building our paper the,This paper presents our contribution to the DeepGlobe Building Detection Challenge.,39693
enhancement next made the should to buildings separation of be adjacent The solution. the,The separation of adjacent buildings should be the next enhancement made to the solution.,39694
these beyond fully a We by unsupervised go proposing approaches strategy.,We go beyond these approaches by proposing a fully unsupervised strategy.,39695
principal the two subtasks. tackle problem We challenging splitting into by scenario this,We tackle this challenging scenario by splitting the problem into two principal subtasks.,39696
algorithm The Results: proposed OCT successfully denoised single-frame unseen B-scans.,Results: The proposed algorithm successfully denoised unseen single-frame OCT B-scans.,39697
"data. histopathology amounts require labeled of large CNNs However,","However, CNNs require large amounts of labeled histopathology data.",39698
order regression This compared models. consistency adversarial framework conventional enforces higher to when CNN,This adversarial regression framework enforces higher order consistency when compared to conventional CNN models.,39699
performing tasks many vision capable of long its Human for in optimized is evolution. not,Human vision is capable of performing many tasks not optimized for in its long evolution.,39700
proposed network and consists sub-networks: two The of detection alignment.,The proposed network consists of two sub-networks: detection and alignment.,39701
"CAM features FCN to detection. pedestrian combine we deep extract and for First,","First, we combine FCN and CAM to extract deep features for pedestrian detection.",39702
"Then, lost parts. perform body part-level the recall to we CNN","Then, we perform part-level CNN to recall the lost body parts.",39703
paper an presents entropy Shannon for neutrosophic of The extension information.,The paper presents an extension of Shannon entropy for neutrosophic information.,39704
between neutrosophic formula a two triplets. new extension uses for This distance,This extension uses a new formula for distance between two neutrosophic triplets.,39705
"intuitionistic obtained particularized In are addition, for information. the results and bifuzzy, paraconsistent fuzzy","In addition, the obtained results are particularized for bifuzzy, intuitionistic and paraconsistent fuzzy information.",39706
Neural It engined an Convolutional end-to-end is advanced Networks. trainable by framework,It is an end-to-end trainable framework engined by advanced Convolutional Neural Networks.,39707
predicts observations Our CPN and according top-down cues. bottom-up text both objects to the,Our CPN predicts text objects according to both top-down observations and the bottom-up cues.,39708
by assembled Propagation call (CP). are boxes Multiple communication spatial candidate mechanism a Correlation,Multiple candidate boxes are assembled by a spatial communication mechanism call Correlation Propagation (CP).,39709
"keeps Namely, producing which plans. planner, it the modules, uses interpretable refined the","Namely, it keeps the planner, which uses the refined modules, producing interpretable plans.",39710
to settings cross-sectional analysis particularly is well-suited to applies panel Our but models.,Our analysis applies to cross-sectional settings but is particularly well-suited to panel models.,39711
and estimator. identification `well-posed' a results Our nonparametric simple motivate,Our identification results motivate a simple and `well-posed' nonparametric estimator.,39712
apply empirical our settings. We methodology to two,We apply our methodology to two empirical settings.,39713
"models are respectively. family and other The two GRU, Soft-BoF","The other two family models are Soft-BoF and GRU, respectively.",39714
present volumetric Michell the We designing first Trusses. for algorithm,We present the first algorithm for designing volumetric Michell Trusses.,39715
robustness demonstrate physical the structural via posteriori our of a We designs simulation.,We demonstrate the structural robustness of our designs via a posteriori physical simulation.,39716
"delivery conducted were using Experiments mobile and a navigation, robot working dialog, tasks. on","Experiments were conducted using a mobile robot working on dialog, navigation, and delivery tasks.",39717
"show model-based Results existing to significant RL methods. in improvements, comparison","Results show significant improvements, in comparison to existing model-based RL methods.",39718
addressing for emergence of complex possibilities vision new artificial RGB-D sensors problems The offered efficiently.,The emergence of RGB-D sensors offered new possibilities for addressing complex artificial vision problems efficiently.,39719
"end, in the introduce this training segmentation location-aware regime. a we loss Towards","Towards this end, we introduce a location-aware segmentation loss in the training regime.",39720
evaluate restaurant created that our corpus we We present story methodology. to a design and,We present a restaurant story corpus that we created to design and evaluate our methodology.,39721
in Logic and Practice consideration of Programming (TPLP). Under Theory,Under consideration in Theory and Practice of Logic Programming (TPLP).,39722
neural tasks. Recently convolutional great recognition (CNNs) visual networks accuracy achieve in,Recently convolutional neural networks (CNNs) achieve great accuracy in visual recognition tasks.,39723
to models one of DenseNet feature-reuse. most its the popular due effectiveness becomes in CNN,DenseNet becomes one of the most popular CNN models due to its effectiveness in feature-reuse.,39724
"overfitting if other models, DenseNets like problem not also CNN severer. However, face","However, like other CNN models, DenseNets also face overfitting problem if not severer.",39725
"in phase modelled global nor been has suitable local form, estimation. Neither for closed Bayesian","Neither global nor local phase has been modelled in closed form, suitable for Bayesian estimation.",39726
been the for has used image Bag-of-Visual-Words (BoVW) in purposes. widely years classification recent approach,Bag-of-Visual-Words (BoVW) approach has been widely used in the recent years for image classification purposes.,39727
affect stability and the model These performance. the reduce of factors,These factors affect the stability of the model and reduce performance.,39728
the by clustering. codebook method to is k-Means++ avoid visual built poor Thus using,Thus the visual codebook is built by using k-Means++ method to avoid poor clustering.,39729
"introduced. and respectively principles Ant-miner the are of algorithm First, ELM basic","First, the basic principles of ELM and Ant-miner algorithm are respectively introduced.",39730
"what and consensus 'interpretable'. 'explainable' over by However, there no is is meant general","However, there is no general consensus over what is meant by 'explainable' and 'interpretable'.",39731
"dysfunction of cilia. marked human of a the organelle, are diseases class cellular by Ciliopathies","Ciliopathies are a class of human diseases marked by dysfunction of the cellular organelle, cilia.",39732
"initiate own. its that However, accumulating evidence unable on suggests transcription RFX is to","However, accumulating evidence suggests that RFX is unable to initiate transcription on its own.",39733
"other Therefore, cis-regulatory and likely elements factors are required.","Therefore, other factors and cis-regulatory elements are likely required.",39734
"challenging nature. reconstruction ill-posed due problem is its to the Nevertheless, still","Nevertheless, the reconstruction problem is still challenging due to its ill-posed nature.",39735
from limited long reconstruction time or explore either methods iterative existing prior Most knowledge. suffer,Most existing methods either suffer from long iterative reconstruction time or explore limited prior knowledge.,39736
"under an which predefined updated is to compared buffer, candidate adaptable Each a condition. is","Each candidate is compared to an adaptable buffer, which is updated under a predefined condition.",39737
of the nonparametric covariates selectively regression We a observed with consider in framework. problem,We consider the problem of regression with selectively observed covariates in a nonparametric framework.,39738
We provide partial hold. to completeness form assumptions functional for primitive,We provide primitive functional form assumptions for partial completeness to hold.,39739
for result identification estimator. the FPW series The is constructive,The identification result is constructive for the FPW series estimator.,39740
asymptotic the convergence the distribution. rate We derive also and pointwise of,We derive the rate of convergence and also the pointwise asymptotic distribution.,39741
"application, focus we two empirical In the on applications. different","In the empirical application, we focus on two different applications.",39742
based presents This sequence method on paper recognition a images. of a face,This paper presents a face recognition method based on a sequence of images.,39743
is reconstructed shape of combination and using structure-from-motion multi-view from images Face methods. a stereo,Face shape is reconstructed from images using a combination of structure-from-motion and multi-view stereo methods.,39744
proposed of also the is verification included. Experimental method,Experimental verification of the proposed method is also included.,39745
has as for reliable sensor provides autonomous driving Lidar essential become depth it estimation. an,Lidar has become an essential sensor for autonomous driving as it provides reliable depth estimation.,39746
online and are in described. execution and modeling Computation mapping pipeline challenges the,Computation and modeling challenges in the mapping and online execution pipeline are described.,39747
"challenges remain still addressed. However, be to key","However, key challenges still remain to be addressed.",39748
the current surveys in state-of-the-art This area applications. of the wireless for medical paper sensors,This paper surveys the current state-of-the-art in the area of wireless sensors for medical applications.,39749
surfels smoothed to surface a We asynchronously (re)triangulate the mesh. reconstruct,We asynchronously (re)triangulate the smoothed surfels to reconstruct a surface mesh.,39750
possible by is cloud surface remeshing and necessary. This asynchronously the the deforming surfel where,This is possible by deforming the surfel cloud and asynchronously remeshing the surface where necessary.,39751
naturally supports scan resolution. surfel-based also The varying strongly representation,The surfel-based representation also naturally supports strongly varying scan resolution.,39752
"input In camera's it the particular, colors reconstructs resolution. at","In particular, it reconstructs colors at the input camera's resolution.",39753
"characters, CTC assumes models conditional independence attention-based alignments. provide of whereas individual can nonsequential","CTC assumes conditional independence of individual characters, whereas attention-based models can provide nonsequential alignments.",39754
"impacts exclusively, affect, the stocks. broad not have targeted although activities that Fishing","Fishing activities have broad impacts that affect, although not exclusively, the targeted stocks.",39755
constructing of by model EwE Icelandic This waters. was done an,This was done by constructing an EwE model of Icelandic waters.,39756
chosen information criterion. model selecting the by Akaike The was lowest final the model with,The final model was chosen by selecting the model with the lowest Akaike information criterion.,39757
and model previously known when simulating The estimated biomass landings. satisfactorily performed,The model performed satisfactorily when simulating previously estimated biomass and known landings.,39758
part analysis. and essential segmentation is image an reliable Accurate biomedical of image,Accurate and reliable image segmentation is an essential part of biomedical image analysis.,39759
for epithelium tasks. evaluate and tubule two We our datasets on public segmentation segmentation method,We evaluate our method on two public datasets for epithelium segmentation and tubule segmentation tasks.,39760
strongly in by rules. structured are affected game updating dynamics populations Evolutionary,Evolutionary game dynamics in structured populations are strongly affected by updating rules.,39761
"on rules, information imitation-based peers. rely studies of social usually which focus payoff on Previous","Previous studies usually focus on imitation-based rules, which rely on payoff information of social peers.",39762
fate on We the graphs population of or strategies by structures explore networks. represented,We explore the fate of strategies on population structures represented by graphs or networks.,39763
Costas non-linearity. task Analysis hard a of is its because of loop,Analysis of Costas loop is a hard task because of its non-linearity.,39764
noise analysis performed. circuits of the the is and are provided models Simulink MATLAB,MATLAB Simulink models of the circuits are provided and the noise analysis is performed.,39765
of and combinations The them such performance are analyzed discussed. some behavior and of indices,The behavior of such performance indices and some combinations of them are analyzed and discussed.,39766
of doctor back visits. for is most Low the common reasons one pain,Low back pain is one of the most common reasons for doctor visits.,39767
We transfer an depict information procedures. and to carrier its model analytical beam propose,We propose an analytical model to depict its carrier beam and information transfer procedures.,39768
data. diffusion-embedding for graph present a novel structured networks We (GDEN) graph,We present a novel graph diffusion-embedding networks (GDEN) for graph structured data.,39769
by formulation our graph. feature diffusion is motivated regularized on on GDEN closed-form,GDEN is motivated by our closed-form formulation on regularized feature diffusion on graph.,39770
and embedding diffusion low-dimensional both in model. feature GDEN integrates regularized simultaneously a network unified,GDEN integrates both regularized feature diffusion and low-dimensional embedding simultaneously in a unified network model.,39771
"based structured data multiple on we with graph with GDEN, deal can structures. naturally Moreover,","Moreover, based on GDEN, we can naturally deal with structured data with multiple graph structures.",39772
"than Rather provide using architectures. insight new existing on architectures, an creating of ensembles we","Rather than using ensembles of existing architectures, we provide an insight on creating new architectures.",39773
"demonstrate our simple, study. We a approach yet on illustrative, toy case","We demonstrate our approach on a simple, yet illustrative, toy case study.",39774
data used sources. multiple of theory is from for The widely belief functions,The theory of belief functions is widely used for data from multiple sources.,39775
kept as The of conflict. on an will be the the mass indicator emptyset,The mass on the emptyset will be kept as an indicator of the conflict.,39776
datasets. presented multi-view number Performance a of test method on evaluated is of the stereo,Performance of the presented method is evaluated on a number of multi-view stereo test datasets.,39777
system. stereo in reconstruction used a as stage be dense a algorithm Presented can first,Presented algorithm can be used as a first stage in a dense stereo reconstruction system.,39778
SFM This target allows resolve waveform velocities. the to,This allows the SFM waveform to resolve target velocities.,39779
human means is most communicate their expression beings to emotions. for natural the Facial,Facial expression is the most natural means for human beings to communicate their emotions.,39780
consider case acted facial studies analysis expression of expressions. Most the,Most facial expression analysis studies consider the case of acted expressions.,39781
Sparse classifier to an used then the task. are to dedicated recognition SVM codes train,Sparse codes are then used to train an SVM classifier dedicated to the recognition task.,39782
the dictionaries of on relies Learning process heavily the dictionary. sparsifying initialization,Learning sparsifying dictionaries heavily relies on the initialization process of the dictionary.,39783
"this cryptosystems. in encryption-decryption key symmetric study, the resembles the In pair key pair","In this study, the pair resembles the encryption-decryption key pair in symmetric key cryptosystems.",39784
formalize SKE threat encloses and that security We of several attacks. the analyze major model,We formalize and analyze the threat model of SKE that encloses several major security attacks.,39785
 in  hierarchical learning. goal-conditioned reinforcement learning representation problem study We of the,  We study the problem of representation learning in goal-conditioned hierarchical reinforcement learning.,39786
vision oldest in literature. preprocessing perhaps step the augmentation Data is computer,Data augmentation is perhaps the oldest preprocessing step in computer vision literature.,39787
augmentation. on of model some uses Almost vision imaging computer every form data trained,Almost every computer vision model trained on imaging data uses some form of augmentation.,39788
prominent of functions is development and brain. The with HSRS for human associated pattern bypassing,The bypassing pattern is prominent for HSRS associated with development and functions of human brain.,39789
of makes variety lesion task skin the Vast very appearance the challenging. this in,Vast variety in the appearance of the skin lesion makes this task very challenging.,39790
for was the border. black outer removing performed Preprocessing,Preprocessing was performed for removing the outer black border.,39791
method. measured was Jaccard evaluate the of the segmentation Index to performance,Jaccard Index was measured to evaluate the performance of the segmentation method.,39792
injury reduces greatly physical the risk of This gains. maximise and,This greatly reduces the risk of injury and maximise physical gains.,39793
selected activity Often area mitotic the with highest be the is MC. to for the,Often the area with the highest mitotic activity is to be selected for the MC.,39794
estimate. activity second map a mitotic in used to construct is a This step,This map is in a second step used to construct a mitotic activity estimate.,39795
"is model optimal. the linear simple in far regulation However, from","However, regulation in the simple linear model is far from optimal.",39796
matching presented. stereo semi-dense is solution robust for A,A robust solution for semi-dense stereo matching is presented.,39797
"It CNN two utilizes computing respectively. confidence-based performing matching models stereo for filtering, and cost","It utilizes two CNN models for computing stereo matching cost and performing confidence-based filtering, respectively.",39798
"For most visual is objects animal species, identification quick survival. and of for critical reliable","For most animal species, quick and reliable identification of visual objects is critical for survival.",39799
methods. unsupervised analysis and through out carried learning been (classification) The supervised has (clustering),The analysis has been carried out through supervised (classification) and unsupervised (clustering) learning methods.,39800
"leads interactions. non-trivial about to Furthermore, predictions change it also ecological long-term phenotypic and","Furthermore, it also leads to non-trivial predictions about long-term phenotypic change and ecological interactions.",39801
solution optimization problem. a finding typically nonconvex requires This a to,This typically requires finding a solution to a nonconvex optimization problem.,39802
coordination. framework intercell and The allows incorporating uncertainties channel,The framework allows incorporating channel uncertainties and intercell coordination.,39803
show of magnitude w.r.t. several be improvements results orders Numerical can achieved that,Numerical results show that several orders of magnitude improvements can be achieved w.r.t.,39804
"art interference as multiuser such the cancellation. and instantaneous consumption reduction state power techniques, of","state of the art techniques, such as instantaneous power consumption reduction and multiuser interference cancellation.",39805
other for computing be traded the platforms efficiency Performance can . for,Performance can be traded for the computing efficiency for other platforms .,39806
"Humans from environments (e.g., natural their structures mathematics, hierarchies). music, social complex language, learn latent","Humans learn complex latent structures from their environments (e.g., natural language, mathematics, music, social hierarchies).",39807
smartphones. on The enhancement significantly defining for image baseline solutions improved proposed results state-of-the-art the,The proposed solutions significantly improved baseline results defining the state-of-the-art for image enhancement on smartphones.,39808
scheme the into problem The decomposes two proposed levels.,The proposed scheme decomposes the problem into two levels.,39809
the method. the show efficiency results of The computational also,The results also show the computational efficiency of the method.,39810
"in past a decade, the interest Over estimation. growing pose been there has human","Over the past decade, there has been a growing interest in human pose estimation.",39811
clutter challenging. robotic manipulation is Autonomous in,Autonomous robotic manipulation in clutter is challenging.,39812
sets learning possible. features small robotic from Employing annotated pretrained makes data,Employing pretrained features makes learning from small annotated robotic data sets possible.,39813
"accuracy model world hardly is scenarios, factor the to consider. only real In","In real world scenarios, model accuracy is hardly the only factor to consider.",39814
peculiar products challenges. grocery store poses Recognition shelves of in,Recognition of grocery products in store shelves poses peculiar challenges.,39815
into measurements network gyroscope method We neural incorporates deblurring propose convolutional that a (CNN). a,We propose a deblurring method that incorporates gyroscope measurements into a convolutional neural network (CNN).,39816
"handle help extremely With such it the of and blur. motion strong can measurements, spatially-variant","With the help of such measurements, it can handle extremely strong and spatially-variant motion blur.",39817
objects. between mathematical paper presents a Our two model new measured causality of,Our paper presents a new mathematical model of causality between two measured objects.,39818
slightly probability the well-known Kolmogorovian We modified have model.,We have slightly modified the well-known Kolmogorovian probability model.,39819
"use set instead we particular, $\sigma$-algebras direct of their sum of In product. the horizontal","In particular, we use the horizontal sum of set $\sigma$-algebras instead of their direct product.",39820
"Consequently, GSFM shape. Ambiguity an Function (AF) resembles thumbtack the that possesses a","Consequently, the GSFM possesses an Ambiguity Function (AF) that resembles a thumbtack shape.",39821
factors the sonar two system consider AF. beyond Practical design must,Practical sonar system design must consider two factors beyond the AF.,39822
waveform the how Peak-to-Average-Power Ratio quantifies The amplitude. constant is (PAPR) close to,The Peak-to-Average-Power Ratio (PAPR) quantifies how close the waveform is to constant amplitude.,39823
"energy compared SE, PAPR, superior efficiency waveforms. and GSFM when to thumbtack The possesses overall","The GSFM possesses superior SE, PAPR, and overall energy efficiency when compared to thumbtack waveforms.",39824
"the throughout is present, The tutorial organizes that, scattered together literature. material at and brings","The tutorial brings together and organizes material that, at present, is scattered throughout the literature.",39825
"explanations, It implementation some also and novel tips. includes","It also includes some novel explanations, and implementation tips.",39826
We and dataset. Oxford dataset datasets: KITTI evaluate outdoor our with method two Robotcar,We evaluate our method with two outdoor datasets: KITTI dataset and Oxford Robotcar dataset.,39827
and branch-and-price Column to generation algorithms are used these inequalities. solve,Column generation and branch-and-price algorithms are used to solve these inequalities.,39828
"primitive a primitive estimation classes. serve Thus, as of with probability associating hypotheses segments can","Thus, segments can serve as primitive hypotheses with a probability estimation of associating primitive classes.",39829
experiments Quantitative and demonstrated qualitative of the BAGSFit. superiority,Quantitative and qualitative experiments demonstrated the superiority of BAGSFit.,39830
exist in inconsistent base. many there knowledge knowledge the Also may,Also there may exist many inconsistent knowledge in the knowledge base.,39831
"access. of scheme completeness, The non-redundancy, and consistency, ease integrity, ensures proposed","The proposed scheme ensures completeness, consistency, integrity, non-redundancy, and ease of access.",39832
important and vision computer system. is in one problem Tracking surveillance Object,Object Tracking is one important problem in computer vision and surveillance system.,39833
exploit mainly the feature (i.e. The models single-view existing,The existing models mainly exploit the single-view feature (i.e.,39834
"color, the to shape) failing the objects problem, describe comprehensively. solve to texture,","color, texture, shape) to solve the problem, failing to describe the objects comprehensively.",39835
"is one big tracking However, inevitably strategy fusion result can problem into non-efficiency. that multi-view","However, one big problem is that multi-view fusion strategy can inevitably result tracking into non-efficiency.",39836
of model the concept video The to extends complex task this prediction. FutureGAN,The FutureGAN model extends this concept to the complex task of video prediction.,39837
"experiments and Cityscapes. conducted KTH three Action, We datasets, on different MovingMNIST,","We conducted experiments on three different datasets, MovingMNIST, KTH Action, and Cityscapes.",39838
with heritage. craft are blessed Borneo rich communities indigenous,Borneo indigenous communities are blessed with rich craft heritage.,39839
such Iban's is plaited the One craft. mat examples,One such examples is the Iban's plaited mat craft.,39840
summarized for theorem purposes of the practical Applicability end. at is this,Applicability of this theorem for practical purposes is summarized at the end.,39841
The has image mobile in high-performance research into more devices recognition. on-device growth of resulted,The growth of high-performance mobile devices has resulted in more research into on-device image recognition.,39842
"In their by environments species mitigate individual in uncertain can ecology, extinction phenotypes. diversifying risks","In ecology, species can mitigate their extinction risks in uncertain environments by diversifying individual phenotypes.",39843
well of is developed. bet-hedging rather theory populations well-mixed The in,The theory of bet-hedging in well-mixed populations is rather well developed.,39844
"fluctuations of For lead slow spatial rates outcomes. and variation, temporal to different","For slow rates of variation, temporal and spatial fluctuations lead to different outcomes.",39845
"temporally In environments, bet-hedging fluctuating spatially to favored compared fluctuating is environments.","In spatially fluctuating environments, bet-hedging is favored compared to temporally fluctuating environments.",39846
population sizes. stochasticity robust These against induced conclusions are by finite,These conclusions are robust against stochasticity induced by finite population sizes.,39847
"image we Graph model Specifically, Games (GTG). using propose Transduction coin to ancient classification","Specifically, we propose to model ancient coin image classification using Graph Transduction Games (GTG).",39848
"Conventional decision the mainly interactions i.e. fraud considers major of involved two parties, control","Conventional fraud control considers mainly the interactions of two major involved decision parties, i.e.",39849
proposed data-driven purely are The real-time a and in self-adaptive models manner.,The proposed models are purely data-driven and self-adaptive in a real-time manner.,39850
"global parametrization geometry is a operation Seamless key of processing, in e.g. surfaces","Seamless global parametrization of surfaces is a key operation in geometry processing, e.g.",39851
(TOF) because Time-of-flight with latency. cameras appealing obtain are minimal dense they depth measurements,Time-of-flight (TOF) cameras are appealing because they obtain dense depth measurements with minimal latency.,39852
semantic work This nighttime of the addresses of segmentation problem scenes. image,This work addresses the problem of semantic image segmentation of nighttime scenes.,39853
"annotated a subset nighttime for evaluation. of method Also, images densely are the","Also, a subset of the nighttime images are densely annotated for method evaluation.",39854
an motion in SA inter-slice respiratory CMR to correct We propose stacks. image approach automatically,We propose an approach to automatically correct inter-slice respiratory motion in SA CMR image stacks.,39855
the show results proposed and motion approach compensation. accuracy of in robustness The the,The results show the accuracy and robustness of the proposed approach in motion compensation.,39856
the model. reveal obviate and We in shortcomings,We reveal and obviate shortcomings in the model.,39857
a MRI) assessment magnetic (diffusion microstructure technique. Diffusion imaging non-invasive is resonance,Diffusion magnetic resonance imaging (diffusion MRI) is a non-invasive microstructure assessment technique.,39858
"diffusion high data. it consuming However, time and costly to is collect quality","However, it is costly and time consuming to collect high quality diffusion data.",39859
specially-tailored two non-linear problems per-frame are resulting Gauss-Newton data-parallel solvers. solved with optimization The,The two resulting non-linear optimization problems per-frame are solved with specially-tailored data-parallel Gauss-Newton solvers.,39860
monocular capture. method first Our performance approach full-body is real-time for the,Our method is the first real-time monocular approach for full-body performance capture.,39861
suggest design users algorithms outfits (e.g. We which automatically,We design algorithms which automatically suggest users outfits (e.g.,39862
of this challenging use We the deep for explore task. networks,We explore the use of deep networks for this challenging task.,39863
deep through The former network. is convolutional a achieved,The former is achieved through a deep convolutional network.,39864
"network we multi-layer a structure. the And latter, fully-connected for adopt","And for the latter, we adopt a multi-layer fully-connected network structure.",39865
compare for FashionNet. alternative three We and architectures design,We design and compare three alternative architectures for FashionNet.,39866
"produce can unrealistic learning limited animation. and therefore However, capabilities such techniques have","However, such techniques have limited learning capabilities and therefore can produce unrealistic animation.",39867
over accumulation bidirectional iterations. mitigate The helps also loss generation error,The bidirectional generation loss also helps mitigate error accumulation over iterations.,39868
single-channel scheme digital backpropagation. the experimental results that show outperforms The proposed,The experimental results show that the proposed scheme outperforms single-channel digital backpropagation.,39869
proposed backpropagation. results outperforms single-channel that scheme digital show The the experimental,The experimental results show that the proposed scheme outperforms single-channel digital backpropagation.,39870
The backpropagation of neural model parameters by are (BP). convolutional (CNNs) networks determined,The model parameters of convolutional neural networks (CNNs) are determined by backpropagation (BP).,39871
adopts FF approach. data-centric design The a,The FF design adopts a data-centric approach.,39872
multiple Saab layers. transforms yield in cascade Multiple convolutional,Multiple Saab transforms in cascade yield multiple convolutional layers.,39873
"and on between BP we relationship FF Finally, designs. the comment","Finally, we comment on the relationship between BP and FF designs.",39874
"as computational implementation primary xMaMIMO, with aspects efficiency our address We of objective.","We address implementation aspects of xMaMIMO, with computational efficiency as our primary objective.",39875
crowded transceivers scenarios. the evaluate the performance We of xMaMIMO bit-error-rate in,We evaluate the bit-error-rate performance of the transceivers in crowded xMaMIMO scenarios.,39876
a Single challenging problem. is image dehazing restoration ill-posed,Single image dehazing is a challenging ill-posed restoration problem.,39877
proposed. Various prior-based and methods learning-based been have,Various prior-based and learning-based methods have been proposed.,39878
of complicated. environment haze The formulation more is realistic in,The formulation of haze in realistic environment is more complicated.,39879
image two network is dehazing The public benchmarks. on evaluated proposed,The proposed network is evaluated on two public image dehazing benchmarks.,39880
demonstrate it superior when compared popular methods. that with state-of-the-art performance achieve experiments can The,The experiments demonstrate that it can achieve superior performance when compared with popular state-of-the-art methods.,39881
"NMR several of Bloch discrete intrinsic manifold responses, encodes parameters.","discrete manifold of Bloch responses, encodes several intrinsic NMR parameters.",39882
secret What intelligence the unlocked? the is that of has human evolution,What is the secret that the evolution of human intelligence has unlocked?,39883
"but more it. to answer, is is there one Generalization","Generalization is one answer, but there is more to it.",39884
loss beamspace systems. an multiple-output multi-user The millimeter-wave multiple-input is (MIMO) issue for rate important,The rate loss is an important issue for multi-user beamspace multiple-input multiple-output (MIMO) millimeter-wave systems.,39885
"signal design downlink we power level using energy a modulation. First,","First, we design a downlink energy signal using power level modulation.",39886
detectors Deep object annotated bounding learning thousands diversified based and class require box examples. of,Deep learning based object detectors require thousands of diversified bounding box and class annotated examples.,39887
"supervised methods. weakly Also, to achieve compared recent of performance baselines we competitive","Also, we achieve competitive performance compared to recent baselines of weakly supervised methods.",39888
function network. deep a An role activation neural has in crucial,An activation function has crucial role in a deep neural network.,39889
widely (ReLU) used linear rectified function. are for the activation A unit simple,A simple rectified linear unit (ReLU) are widely used for the activation function.,39890
"(WiG) is activation weighted sigmoid In paper, unit gate as this a the proposed function.","In this paper, a weighted sigmoid gate unit (WiG) is proposed as the activation function.",39891
sigmoid of gate. and a consists inputs WiG the proposed weighted of multiplication The,The proposed WiG consists of a multiplication of inputs and the weighted sigmoid gate.,39892
the to functions overcome the activation ReLU. proposed of have performance Many been,Many activation functions have been proposed to overcome the performance of the ReLU.,39893
"an is with literature, task. In the recognition performance the mainly object evaluated","In the literature, the performance is mainly evaluated with an object recognition task.",39894
evaluated recognition object is with task and the task. The image proposed the restoration WiG,The proposed WiG is evaluated with the object recognition task and the image restoration task.,39895
lung Cardiovascular disease in population. cancer of screening (CVD) the cause a is leading death,Cardiovascular disease (CVD) is a leading cause of death in the lung cancer screening population.,39896
"region. heart a used was the autoencoder Thereafter, encode convolutional identified to","Thereafter, a convolutional autoencoder was used to encode the identified heart region.",39897
in of The medicine. the tissue importance malign diagnostics crucial of is,The diagnostics of malign tissue is of the crucial importance in medicine.,39898
error . and factor of the subjective probability,of the subjective factor and error probability .,39899
in active is among area researchers. computer an of Face recognition the interest images vision,Face recognition in images is an active area of interest among the computer vision researchers.,39900
"is human recognizing relatively However, environment, research. a area unconstrained an less-explored of in face","However, recognizing human face in an unconstrained environment, is a relatively less-explored area of research.",39901
makes task occlusion faces more of challenging. recognition the Partial even,Partial occlusion of faces makes the recognition task even more challenging.,39902
Recent extended task to multiple graphs; work i.e. this has,Recent work has extended this task to multiple graphs; i.e.,39903
find to multiple the among is groups goal highly dense vertices graphs. of,the goal is to find groups of vertices highly dense among multiple graphs.,39904
"difference/contrast specifically between patterns the highlight Thus, the graphs. considered these","Thus, these patterns specifically highlight the difference/contrast between the considered graphs.",39905
a contrasting of showcase potential the datasets. patterns and synthetic on variety We real-world of,We showcase the potential of contrasting patterns on a variety of synthetic and real-world datasets.,39906
of We the objects detection weakly method propose for in supervised paintings. a,We propose a method for the weakly supervised detection of objects in paintings.,39907
"training annotations time, image-level needed. are only At","At training time, only image-level annotations are needed.",39908
show the losses. only several yields We dropping annotations on performance that mild databases instance-level,We show on several databases that dropping the instance-level annotations only yields mild performance losses.,39909
prediction fundamental Depth the in computer problems one vision. is of,Depth prediction is one of the fundamental problems in computer vision.,39910
"to depth For of the a.k.a sparse completion. dense, tasks","For the tasks of sparse to dense, a.k.a depth completion.",39911
"information hand this a capture work, the state-of-the-art In by hand estimator. we using pose","In this work, we capture the hand information by using a state-of-the-art hand pose estimator.",39912
quality. infrastructure maintains built-in testing code A,A built-in testing infrastructure maintains code quality.,39913
"modeling paper, In the correlations of features. this convolutional we on between channels focus","In this paper, we focus on modeling the correlations between channels of convolutional features.",39914
recalibrate to is adaptively used SEBlock mappings. channel-wise feature,SEBlock is used to adaptively recalibrate channel-wise feature mappings.,39915
"Further, information between SEBlock short remedy each connections loss. used are to","Further, short connections between each SEBlock are used to remedy information loss.",39916
traffic general phenomenon. with this proposed approaches have managers three dealing to engineers and Urban,Urban traffic engineers and managers have proposed three general approaches to dealing with this phenomenon.,39917
the capacity approach network (UTN). of urban the expands first The traffic,The first approach expands the capacity of the urban traffic network (UTN).,39918
approach The traffic called can the be assignment. second,The second approach can be called the traffic assignment.,39919
"capacity the the involves of UTN. approach Finally, optimizing the third","Finally, the third approach involves optimizing the capacity of the UTN.",39920
"information, four traffic optimization, These prediction. assignment, traffic sections include and traffic traffic","These four sections include traffic information, traffic assignment, traffic optimization, and traffic prediction.",39921
"theoretical the the that results illustrate Finally, two of simulation effectiveness results are presented.","Finally, two simulation results that illustrate the effectiveness of the theoretical results are presented.",39922
"systems, we detection. object such To enable faster innovative need","To enable such innovative systems, we need faster object detection.",39923
"and In accuracy approximations, speed with we trade-off investigate the i.e. this work, between domain-specific","In this work, we investigate the trade-off between accuracy and speed with domain-specific approximations, i.e.",39924
"image state-of-the-art object learning-based detection scaling size and category-aware deep proposals for meta-architectures. scaling, two","category-aware image size scaling and proposals scaling, for two state-of-the-art deep learning-based object detection meta-architectures.",39925
problem. enhancement computer resolution classical is image vision Low a,Low resolution image enhancement is a classical computer vision problem.,39926
employ such and streams multiple approaches textual from data as Multi-modal domains. visual input,Multi-modal approaches employ data from multiple input streams such as textual and visual domains.,39927
neural been for approaches. Deep these have networks employed successfully,Deep neural networks have been successfully employed for these approaches.,39928
text image. embeds proposed image obtain an approach encoded an to an onto information-enriched The,The proposed approach embeds an encoded text onto an image to obtain an information-enriched image.,39929
(AC) consensus algorithms. rules compared four have average of We,We have compared four rules of average consensus (AC) algorithms.,39930
is first rule weights. without AC the simple The,The first rule is the simple AC without weights.,39931
are centralised All each hard methods with other compared combining CSS. the and,All methods are compared each other and with the hard combining centralised CSS.,39932
"the computational of WAC complexity IWAC, and very be similar. proven are to Hence, WAC-AE,","Hence, the computational complexity of IWAC, WAC-AE, and WAC are proven to be very similar.",39933
and method We fusion for multi-modality baseline segmentation. image compared to techniques our,We compared our method to baseline techniques for multi-modality image fusion and segmentation.,39934
on dataset show results state-of- Experiments other KITTI algorithms. our significantly that the outperform the-art,Experiments on the KITTI dataset show that our results significantly outperform other state-of- the-art algorithms.,39935
Source codes . at be https: found can //github.com/baidu-research/UnDepthflow,Source codes can be found at https: //github.com/baidu-research/UnDepthflow .,39936
use local for tracking. optimization Current approaches,Current approaches use local optimization for tracking.,39937
mult-armed as a multi-player up problem set We develop bandit the two distributed and algorithms.,We set up the problem as a multi-player mult-armed bandit and develop two distributed algorithms.,39938
suitable ideally battery for SU make them collisions operated Fewer terminals.,Fewer collisions make them ideally suitable for battery operated SU terminals.,39939
powerful neural widely Convolutional and used networks (CNN's) are tools.,Convolutional neural networks (CNN's) are powerful and widely used tools.,39940
"their interpretability from is far ideal. However,","However, their interpretability is far from ideal.",39941
structures the training. the of demonstrate change the over of We course also simple,We also demonstrate the change of the simple structures over the course of training.,39942
context importance. in great a is of such selection Variable,Variable selection in such a context is of great importance.,39943
"oracle well-known TAlasso the achieves of Accommodating property. the regressors, system heterogeneous","Accommodating the system of heterogeneous regressors, TAlasso achieves the well-known oracle property.",39944
"fall however, models short breakthroughs, in still reaching accuracy. enormous human-level Despite and effort huge","Despite enormous effort and huge breakthroughs, however, models still fall short in reaching human-level accuracy.",39945
is loss quite wireless (WSNs) sensor in networks prevalent. Data,Data loss in wireless sensor networks (WSNs) is quite prevalent.,39946
"existing performance methods on of the dependent factors. is various However,","However, the performance of existing methods is dependent on various factors.",39947
"massive Further, for missing be data of correlation. amount should temporal the not exploiting","Further, the amount of missing data should not be massive for exploiting temporal correlation.",39948
"the To drawbacks, method proposed a in this has overcome novel been PCI-MDR above-mentioned paper.","To overcome the above-mentioned drawbacks, a novel method PCI-MDR has been proposed in this paper.",39949
on evaluate our We semantic experimentally approach segmentation.,We experimentally evaluate our approach on semantic segmentation.,39950
adaptation Our no unsupervised adaptation method and over techniques. state-of-the-art improves domain,Our method improves over no adaptation and state-of-the-art unsupervised domain adaptation techniques.,39951
"gene effects well is on However, understood. direct its less regulation","However, its direct effects on gene regulation is less well understood.",39952
"be image reconstructed field-of-view. From without these a higher-resolution images, sacrificing can computationally","From these images, a higher-resolution image can be computationally reconstructed without sacrificing field-of-view.",39953
biological the way Our in work for imaging high-throughput studies. paves,Our work paves the way for high-throughput imaging in biological studies.,39954
introduces prior total based for a multi-channel paper variation MRI on The super-resolution.,The paper introduces a prior based on multi-channel total variation for MRI super-resolution.,39955
is by resolution the input estimating hyper-parameters Bias-variance from low handled scans. trade-off,Bias-variance trade-off is handled by estimating hyper-parameters from the low resolution input scans.,39956
model of large brain validated was The on a database images.,The model was validated on a large database of brain images.,39957
providing fundamental encoder limits for content the identification. codewords,encoder codewords providing the fundamental limits for content identification.,39958
"proposed an for Also, presented implementation procedure efficient is the algorithm.","Also, an efficient implementation procedure is presented for the proposed algorithm.",39959
area algorithm and enhances proposed The eliminates background clutter. effectively the target,The proposed algorithm effectively enhances the target area and eliminates background clutter.,39960
the effectiveness Simulation infrared on algorithm. significant images prove the of real proposed results,Simulation results on real infrared images prove the significant effectiveness of the proposed algorithm.,39961
in all OpenCV algorithms coded in available and are All Python. were,All algorithms are available in OpenCV and were all coded in Python.,39962
"become of inefficient. classical increasing antennas, the number (BS) With station the base detectors","With the increasing number of base station (BS) antennas, the classical detectors become inefficient.",39963
"performance-complexity the employed for tradeoff improvement. Therefore, LAS is","Therefore, the LAS is employed for performance-complexity tradeoff improvement.",39964
"input from Then, compressed the image. this CNN regresses vector microscopy","Then, CNN regresses this compressed vector from the input microscopy image.",39965
our achieved benchmark have algorithm excellent and validated datasets on performances. We,We have validated our algorithm on benchmark datasets and achieved excellent performances.,39966
"placement network localization on the of accuracy of depends Additionally, the the strongly anchors. the","Additionally, the localization accuracy of the network strongly depends on the placement of the anchors.",39967
part a This master's of thesis is done work project. as,This work is done as part of a master's thesis project.,39968
"it out efficient powerful that It enough. is and generic, turned","It turned out that it is generic, efficient and powerful enough.",39969
"Existing the targeting learn human human-region-based either problem representations. perform or alignment, works","Existing works targeting the problem either perform human alignment, or learn human-region-based representations.",39970
required generally inference. and pose information cost computational for Extra is,Extra pose information and computational cost is generally required for inference.,39971
an consider efficient important task of effective and segmentation. image semantic We,We consider an important task of effective and efficient semantic image segmentation.,39972
"Here, research. the the periocular we state review in art of biometrics","Here, we review the state of the art in periocular biometrics research.",39973
of fonts trained models typesets a unseen previously variety to from sources. and recognize,models trained to recognize a variety of fonts and typesets from previously unseen sources.,39974
"reach introduced. optimization framework a To this is Bayesian goal,","To reach this goal, a Bayesian optimization framework is introduced.",39975
reduces computational Bayesian is fully sample that present more We efficient a approach and budget.,We present a fully Bayesian approach that is more sample efficient and reduces computational budget.,39976
analysis theoretical and framework supported empirical an study. is Our by,Our framework is supported by theoretical analysis and an empirical study.,39977
"research. discuss we for Finally, further directions","Finally, we discuss directions for further research.",39978
generated We estimation. quantitatively gaze by for models images evaluate the training,We quantitatively evaluate the generated images by training models for gaze estimation.,39979
proposes dehazing. This end-to-end for image generative an paper method,This paper proposes an end-to-end generative method for image dehazing.,39980
This pipeline of work reconstruction possible the BVS. for the a provides,This work provides a possible pipeline for the reconstruction of the BVS.,39981
"image reconstructions Using analysis, multimodality within a timeframe. we reliable obtain reasonable","Using multimodality image analysis, we obtain reliable reconstructions within a reasonable timeframe.",39982
problem. a an and is it NPC that proved is game video as puzzle Minesweeper,Minesweeper as a puzzle video game and is proved that it is an NPC problem.,39983
"regularization work, novel this scheme this we present In for effect. a that reduces VQA","In this work, we present a novel regularization scheme for VQA that reduces this effect.",39984
Our training is and agnostic implement. model procedure a simple approach to,Our approach is a model agnostic training procedure and simple to implement.,39985
"images. a In work, for we new framework text-based stylization of binary this present the","In this work, we present a new framework for the stylization of text-based binary images.",39986
"explored. is of composition and shape background a Second, image the stylized geometric the","Second, the composition of the stylized geometric shape and a background image is explored.",39987
"the method our be can of binary According to images, many applied contents to fields.","According to the contents of binary images, our method can be applied to many fields.",39988
to devices. localization assistive approach visual a feasible of improving is positioning precision the The,The visual localization is a feasible approach to improving the positioning precision of assistive devices.,39989
"is passed an canvas upscaling network Finally, the to images. generate through","Finally, the canvas is passed through an upscaling network to generate images.",39990
model's our with Reed generated We compare generated images et. those,We compare our model's generated images with those generated Reed et.,39991
assignment investigate multiple-input massive multi-cell systems. We pilot the multiple-output of in effects,We investigate the effects of pilot assignment in multi-cell massive multiple-input multiple-output systems.,39992
studies This work image of representations. problem the dimensional low appropriate learning,This work studies the problem of learning appropriate low dimensional image representations.,39993
trace Our solutions employing by criterion. quotient images representations disentangle of sparse developed the,Our developed solutions disentangle sparse representations of images by employing the trace quotient criterion.,39994
"initialization. strong observe we a Additionally, robustness towards rough","Additionally, we observe a strong robustness towards rough initialization.",39995
for pose a clutter. estimation autonomous manipulation Object to in robots perform crucial is prerequisite,Object pose estimation is a crucial prerequisite for robots to perform autonomous manipulation in clutter.,39996
"objects added warehouses bin-picking present such as new challenges, e.g., additional settings are constantly. Real-world","Real-world bin-picking settings such as warehouses present additional challenges, e.g., new objects are added constantly.",39997
validation The proposed real-world on is method synthetic evaluated cluttered a and scenes. dataset,The proposed method is evaluated on a synthetic validation dataset and cluttered real-world scenes.,39998
a between modes. statistics survival qualitatively crossover with and is associated decline breakdown different This,This breakdown is associated with a crossover between qualitatively different survival statistics and decline modes.,39999
a image super of resolution great low-level vision Single computer as importance task. is,Single image super resolution is of great importance as a low-level computer vision task.,40000