text
stringlengths 8
3.91k
| label
int64 0
10
|
|---|---|
abstract
to each of the symmetry groups of the platonic solids we adjoin a carefully designed
involution yielding topological generators of pu (2) which have optimal covering properties
as well as efficient navigation. these are a consequence of optimal strong approximation
for integral quadratic forms associated with certain special quaternion algebras and their
arithmetic groups. the generators give super efficient 1-qubit quantum gates and are
natural building blocks for the design of universal quantum gates.
| 4 |
abstract
deep neural network (dnn) models have recently obtained state-of-the-art prediction accuracy for
the transcription factor binding (tfbs) site classification task. however, it remains unclear how
these approaches identify meaningful dna sequence signals and give insights as to why tfs bind
to certain locations. in this paper, we propose a toolkit called the deep motif dashboard (demo
dashboard) which provides a suite of visualization strategies to extract motifs, or sequence patterns
from deep neural network models for tfbs classification. we demonstrate how to visualize and
understand three important dnn models: convolutional, recurrent, and convolutional-recurrent
networks. our first visualization method is finding a test sequence’s saliency map which uses
first-order derivatives to describe the importance of each nucleotide in making the final prediction.
second, considering recurrent models make predictions in a temporal manner (from one end of a
tfbs sequence to the other), we introduce temporal output scores, indicating the prediction score
of a model over time for a sequential input. lastly, a class-specific visualization strategy finds the
optimal input sequence for a given tfbs positive class via stochastic gradient optimization. our
experimental results indicate that a convolutional-recurrent architecture performs the best among
the three architectures. the visualization techniques indicate that cnn-rnn makes predictions by
modeling both motifs as well as dependencies among them.
| 9 |
abstract
for lengths 64 and 66, we construct extremal singly even self-dual
codes with weight enumerators for which no extremal singly even selfdual codes were previously known to exist. we also construct new
40 inequivalent extremal doubly even self-dual [64, 32, 12] codes with
covering radius 12 meeting the delsarte bound.
| 7 |
abstract
| 6 |
abstract
we describe a question answering model that
applies to both images and structured knowledge bases. the model uses natural language strings to automatically assemble neural networks from a collection of composable
modules. parameters for these modules are
learned jointly with network-assembly parameters via reinforcement learning, with only
(world, question, answer) triples as supervision. our approach, which we term a dynamic
neural module network, achieves state-of-theart results on benchmark datasets in both visual and structured domains.
| 9 |
abstract. grammars written as constraint handling rules (chr) can
be executed as efficient and robust bottom-up parsers that provide a
straightforward, non-backtracking treatment of ambiguity. abduction
with integrity constraints as well as other dynamic hypothesis generation techniques fit naturally into such grammars and are exemplified
for anaphora resolution, coordination and text interpretation.
| 6 |
abstract
referential integrity (ri) is an important correctness property of a
shared, distributed object storage system. it is sometimes thought
that enforcing ri requires a strong form of consistency. in this paper, we argue that causal consistency suffices to maintain ri. we
support this argument with pseudocode for a reference crdt data
type that maintains ri under causal consistency. quickcheck has
not found any errors in the model.
| 8 |
abstract
the presence of behind-the-meter rooftop pv and storage in the residential sector is poised
to increase significantly. here we quantify in detail the value of these technologies to consumers
and service providers. we characterize the heterogeneity in household electricity cost savings
under time-varying prices due to consumption behavior differences. the top 15% of consumers
benefit two to three times as much as the remaining 85%. different pricing policies do not
significantly alter how households fare with respect to one another. we define the value of
information as the financial value of improved forecasting capabilities for a household. the
typical value of information is 3.5 cents per hour per kwh reduction of standard deviation of
forecast error. coordination services that combine the resources available at all households can
reduce costs by an additional 15% to 30% of the original total cost. surprisingly, on the basis of
coordinated action alone, service providers will not encourage adoption beyond 50% within a
group. coordinated information, however, enables the providers to generate additional value
with increasing adoption.
| 3 |
abstract
we consider inference in the scalar diffusion model dxt = b(xt ) dt + σ(xt ) dwt with
discrete data (xj∆n )0≤j≤n , n → ∞, ∆n → 0 and periodic coefficients. for σ given, we
prove a general theorem detailing conditions under which bayesian posteriors will contract
in l2 –distance around the true drift function b0 at the frequentist minimax rate (up to
logarithmic factors) over besov smoothness classes. we exhibit natural nonparametric priors
which satisfy our conditions. our results show that the bayesian method adapts both to an
unknown sampling regime and to unknown smoothness.
| 10 |
abstract
expectation maximization (em) has recently been shown to be an efficient algorithm
for learning finite-state controllers (fscs) in large decentralized pomdps (dec-pomdps).
however, current methods use fixed-size fscs and often converge to maxima that are far
from optimal. this paper considers a variable-size fsc to represent the local policy of each
agent. these variable-size fscs are constructed using a stick-breaking prior, leading to a new
framework called decentralized stick-breaking policy representation (dec-sbpr). this approach
learns the controller parameters with a variational bayesian algorithm without having to assume
that the dec-pomdp model is available. the performance of dec-sbpr is demonstrated
on several benchmark problems, showing that the algorithm scales to large problems while
outperforming other state-of-the-art methods.
| 3 |
abstract
deep learning methods have shown great promise in many practical applications,
ranging from speech recognition, visual object recognition, to text processing.
however, most of the current deep learning methods suffer from scalability problems for large-scale applications, forcing researchers or users to focus on smallscale problems with fewer parameters.
in this paper, we consider a well-known machine learning model, deep belief networks (dbns) that have yielded impressive classification performance on a large
number of benchmark machine learning tasks. to scale up dbn, we propose an
approach that can use the computing clusters in a distributed environment to train
large models, while the dense matrix computations within a single machine are
sped up using graphics processors (gpu). when training a dbn, each machine
randomly drops out a portion of neurons in each hidden layer, for each training
case, making the remaining neurons only learn to detect features that are generally
helpful for producing the correct answer. within our approach, we have developed
four methods to combine outcomes from each machine to form a unified model.
our preliminary experiment on the mnist handwritten digit database demonstrates that our approach outperforms the state of the art test error rate.
| 9 |
abstract
a shortcoming of existing reachability approaches for nonlinear systems is the poor scalability
with the number of continuous state variables. to mitigate this problem we present a simulationbased approach where we first sample a number of trajectories of the system and next establish
bounds on the convergence or divergence between the samples and neighboring trajectories. we
compute these bounds using contraction theory and reduce the conservatism by partitioning the
state vector into several components and analyzing contraction properties separately in each
direction. among other benefits this allows us to analyze the effect of constant but uncertain
parameters by treating them as state variables and partitioning them into a separate direction.
we next present a numerical procedure to search for weighted norms that yield a prescribed
contraction rate, which can be incorporated in the reachability algorithm to adjust the weights
to minimize the growth of the reachable set.
| 3 |
abstract: one of the most recent architectures of networks is software-defined networks (sdns) using a controller appliance to control the set of switches on the network. the controlling process includes installing or
uninstalling packet-processing rules on flow tables of switches.
this paper presents a high-level imperative network programming language, called imnet, to facilitate writing
efficient, yet simple, programs executed by controller to manage switches. imnet is simply-structured, expressive,
compositional, and imperative. this paper also introduces an operational semantics to imnet. detailed examples
of programs (with their operational semantics) constructed in imnet are illustrated in the paper as well.
key–words: network programming languages, controller-switch architecture, operational semantics, syntax, imnet.
| 6 |
abstract—advances in de novo synthesis of dna and computational gene design methods make possible the customization of
genes by direct manipulation of features such as codon bias and mrna secondary structure. codon context is another feature
significantly affecting mrna translational efficiency, but existing methods and tools for evaluating and designing novel optimized
protein coding sequences utilize untested heuristics and do not provide quantifiable guarantees on design quality. in this study
we examine statistical properties of codon context measures in an effort to better understand the phenomenon. we analyze the
computational complexity of codon context optimization and design exact and efficient heuristic gene recoding algorithms under
reasonable constraint models. we also present a web-based tool for evaluating codon context bias in the appropriate context.
index terms— computational biology, dynamic programming, simulated annealing, synthetic biology
| 5 |
abstract
initial population plays an important role in heuristic algorithms such as ga as it help to decrease
the time those algorithms need to achieve an acceptable result. furthermore, it may influence the
quality of the final answer given by evolutionary algorithms. in this paper, we shall introduce a
heuristic method to generate a target based initial population which possess two mentioned
characteristics. the efficiency of the proposed method has been shown by presenting the results of
our tests on the benchmarks.
| 9 |
abstract. given a positive integer κ, we investigate the class of numerical semigroups verifying the property that every two subsequent non gaps, smaller than the
conductor, are spaced by at least κ. these semigroups will be called κ-sparse and
generalize the concept of sparse numerical semigroups.
| 0 |
abstract
solvency games, introduced by berger et al., provide an abstract framework for modelling
decisions of a risk-averse investor, whose goal is to avoid ever going broke. we study a new
variant of this model, where, in addition to stochastic environment and fixed increments and
decrements to the investor’s wealth, we introduce interest, which is earned or paid on the current
level of savings or debt, respectively.
we study problems related to the minimum initial wealth sufficient to avoid bankruptcy
(i.e. steady decrease of the wealth) with probability at least p. we present an exponential time
algorithm which approximates this minimum initial wealth, and show that a polynomial time
approximation is not possible unless p = np. for the qualitative case, i.e. p = 1, we show
that the problem whether a given number is larger than or equal to the minimum initial wealth
belongs to np ∩ conp, and show that a polynomial time algorithm would yield a polynomial
time algorithm for mean-payoff games, existence of which is a longstanding open problem. we
also identify some classes of solvency mdps for which this problem is in p. in all above cases the
algorithms also give corresponding bankruptcy avoiding strategies.
1998 acm subject classification g.3 probability and statistics.
keywords and phrases markov decision processes, algorithms, complexity, market models.
| 5 |
abstract
new vision sensors, such as the dynamic and active-pixel vision sensor (davis), incorporate a conventional globalshutter camera and an event-based sensor in the same pixel array. these sensors have great potential for high-speed
robotics and computer vision because they allow us to combine the benefits of conventional cameras with those of
event-based sensors: low latency, high temporal resolution, and very high dynamic range. however, new algorithms
are required to exploit the sensor characteristics and cope with its unconventional output, which consists of a stream of
asynchronous brightness changes (called “events”) and synchronous grayscale frames. for this purpose, we present
and release a collection of datasets captured with a davis in a variety of synthetic and real environments, which we
hope will motivate research on new algorithms for high-speed and high-dynamic-range robotics and computer-vision
applications. in addition to global-shutter intensity images and asynchronous events, we provide inertial measurements
and ground-truth camera poses from a motion-capture system. the latter allows comparing the pose accuracy of egomotion estimation algorithms quantitatively. all the data are released both as standard text files and binary files (i.e.,
rosbag). this paper provides an overview of the available data and describes a simulator that we release open-source
to create synthetic event-camera data.
keywords
event-based cameras, visual odometry, slam, simulation
| 1 |
abstract
health related social media mining is a valuable apparatus for the early recognition of the diverse antagonistic
medicinal conditions. mostly, the existing methods are based on machine learning with knowledge-based learning.
this working note presents the recurrent neural network (rnn) and long short-term memory (lstm) based embedding for automatic health text classification in the social media mining. for each task, two systems are built and that
classify the tweet at the tweet level. rnn and lstm are used for extracting features and non-linear activation function at the last layer facilitates to distinguish the tweets of different categories. the experiments are conducted on 2nd
social media mining for health applications shared task at amia 2017. the experiment results are considerable;
however the proposed method is appropriate for the health text classification. this is primarily due to the reason that,
it doesn’t rely on any feature engineering mechanisms.
introduction
with the expansion of micro blogging platforms such as twitter, the internet is progressively being utilized to spread
health information instead of similarly as a wellspring of data1, 2 . twitter allows users to share their status messages
typically called as tweets, restricted to 140 characters. most of the time, these tweets expresses the opinions about the
topics. thus analysis of tweets has been considered as a significant task in many of the applications, here for health
related applications.
health text classification is taken into account a special case of text classification. the existing methods have used
machine learning methods with feature engineering. most commonly used features are n-grams, parts-of-speech tags,
term frequency-inverse document frequency, semantic features such as mentions of chemical substance and disease,
wordnet synsets, adverse drug reaction lexicon, etc3–6, 16 . in6, 7 proposed ensemble based approach for classifying the
adverse drug reactions tweets.
recently, the deep learning methods have performed well8 and used in many tasks mainly due to that it doesn’t rely on
any feature engineering mechanism. however, the performance of deep learning methods implicitly relies on the large
amount of raw data sets. to make use of unlabeled data,9 proposed semi-supervised approach based on convolutional
neural network for adverse drug event detection. though the data sets of task 1 and task 2 are limited, this paper
proposes rnn and lstm based embedding method.
background and hyper parameter selection
this section discusses the concepts of tweet representation and deep learning algorithms particularly recurrent neural
network (rnn) and long short-term memory (lstm) in a mathematical way.
tweet representation
representation of tweets typically called as tweet encoding. this contains two steps. the tweets are tokenized to
words during the first step. moreover, all words are transformed to lower-case. in second step, a dictionary is formed
by assigning a unique key for each word in a tweet. the unknown words in a tweet are assigned to default key 0.
to retain the word order in a tweet, each word is replaced by a unique number according to a dictionary. each tweet
vector sequence is made to same length by choosing the particular length. the tweet sequences that are too long than
the particular length are discarded and too short are padded by zeros. this type of word vector representation is passed
as input to the word embedding layer. for task 1, the maximum tweet sequence length is 35. thus the train matrix
| 2 |
abstract
in this paper, we address the dataset scarcity issue with the
hyperspectral image classification. as only a few thousands
of pixels are available for training, it is difficult to effectively
learn high-capacity convolutional neural networks (cnns).
to cope with this problem, we propose a novel cross-domain
cnn containing the shared parameters which can co-learn
across multiple hyperspectral datasets. the network also contains the non-shared portions designed to handle the datasetspecific spectral characteristics and the associated classification tasks. our approach is the first attempt to learn a cnn
for multiple hyperspectral datasets, in an end-to-end fashion.
moreover, we have experimentally shown that the proposed
network trained on three of the widely used datasets outperform all the baseline networks which are trained on single
dataset.
index terms— hyperspectral image classification, convolutional neural network (cnn), shared network, cross domain, domain adaptation
1. introduction
the introduction of convolutional neural network (cnn) has
brought forth unprecedented performance increase for classification problems in many different domains including rgb,
rgbd, and hyperspectral images. [1, 2, 3, 4] such performance increase was made possible due to the ability of cnn
being able to learn and express the deep and wide connection
between the input and the output using a huge number of parameters. in order to learn such a huge set of parameters, having a large scale dataset has become a significant requirement.
when the size of the given dataset is insufficient to learn a
network, one may consider using a larger external dataset to
better learn the large set of parameters. for instance, girshick
et al. [2] introduced a domain adaptation approach where the
network is trained on a large scale source domain (imagenet
dataset) and then finetuned on a target domain (object detection dataset).
when applying cnn to hyperspectral image classification problem, we also face the similar issue as there are no
| 1 |
abstract syntax tree (ast), consisting of syntax nodes (corresponding to nonterminals in the programming language’s grammar) and syntax tokens (corresponding to terminals).
we label syntax nodes with the name of the nonterminal from the program’s grammar, whereas
syntax tokens are labeled with the string that they represent. we use child edges to connect nodes
according to the ast. as this does not induce an order on children of a syntax node, we additionally
add nexttoken edges connecting each syntax token to its successor. an example of this is shown in
fig. 2a.
to capture the flow of control and data through a program, we add additional edges connecting
different uses and updates of syntax tokens corresponding to variables. for such a token v, let dr (v)
be the set of syntax tokens at which the variable could have been used last. this set may contain
several nodes (for example, when using a variable after a conditional in which it was used in both
branches), and even syntax tokens that follow in the program code (in the case of loops). similarly,
let dw (v) be the set of syntax tokens at which the variable was last written to. using these, we
add lastread (resp. lastwrite) edges connecting v to all elements of dr (v) (resp. dw (v)).
additionally, whenever we observe an assignment v = expr , we connect v to all variable tokens
occurring in expr using computedfrom edges. an example of such semantic edges is shown in
fig. 2b.
we extend the graph to chain all uses of the same variable using lastlexicaluse edges (independent
of data flow, i.e., in if (...) { ... v ...} else { ... v ...}, we link the two occurrences of v). we also connect return tokens to the method declaration using returnsto edges
(this creates a “shortcut” to its name and type). inspired by rice et al. (2017), we connect arguments
in method calls to the formal parameters that they are matched to with formalargname edges,
i.e., if we observe a call foo(bar) and a method declaration foo(inputstream stream),
we connect the bar token to the stream token. finally, we connect every token corresponding
to a variable to enclosing guard expressions that use the variable with guardedby and guardedbynegation edges. for example, in if (x > y) { ... x ...} else { ... y ...},
we add a guardedby edge from x (resp. a guardedbynegation edge from y) to the ast node
corresponding to x > y.
finally, for all types of edges we introduce their respective backwards edges (transposing the
adjacency matrix), doubling the number of edges and edge types. backwards edges help with
propagating information faster across the ggnn and make the model more expressive.
leveraging variable type information we assume a statically typed language and that the
source code can be compiled, and thus each variable has a (known) type τ (v). to use it, we define
a learnable embedding function r(τ ) for known types and additionally define an “u nk t ype” for
all unknown/unrepresented types. we also leverage the rich type hierarchy that is available in many
object-oriented languages. for this, we map a variable’s type τ (v) to the set of its supertypes, i.e.
τ ∗ (v) = {τ : τ (v) implements type τ } ∪ {τ (v)}. we then compute the type representation r∗ (v)
used for state updates and the number of propagation steps per ggnn layer is fixed to 1. instead, several layers
are used. in our experiments, gcns generalized less well than ggnns.
| 2 |
abstract—wireless fading channels suffer from both channel
fadings and additive white gaussian noise (awgn). as a result,
it is impossible for fading channels to support a constant rate data
stream without using buffers. in this paper, we consider information transmission over an infinite-buffer-aided block rayleigh
fading channel in the low signal-to-noise ratio (snr) regime. we
characterize the transmission capability of the channel in terms of
stationary queue length distribution, packet delay, as well as data
rate. based on the memoryless property of the service provided by
the channel in each block, we formulate the transmission process
as a discrete time discrete state d/g/1 queueing problem. the
obtained results provide a full characterization of block rayleigh
fading channels and can be extended to the finite-buffer-aided
transmissions.
index terms—block rayleigh fading channel, buffer-aided
communication, queueing analysis, queue length distribution,
packet delay.
| 7 |
abstract
the millimeter wave spectra at 71-76ghz (70ghz) and 81-86ghz (80ghz) have the potential
to endow fifth-generation new radio (5g-nr) with mobile connectivity at gigabit rates. however, a
pressing issue is the presence of incumbent systems in these bands, which are primarily point-topoint fixed stations (fss). in this paper, we first identify the key properties of incumbents by parsing
databases of existing stations in major cities to devise several modeling guidelines and characterize their
deployment geometry and antenna specifications. second, we develop a detailed interference framework
to compute the aggregate interference from outdoor 5g-nr users into fss. we then present several case
studies in dense populated areas, using actual incumbent databases and building layouts. our simulation
results demonstrate promising 5g coexistence at 70ghz and 80ghz as the majority of fss experience
interference well below the noise floor thanks to the propagation losses in these bands and the deployment
geometry of the incumbent and 5g systems. for the few fss that may incur higher interference, we
propose several passive interference mitigation techniques such as angular-based exclusion zones and
spatial power control. simulations results show that the techniques can effectively protect fss, without
tangible degradation of the 5g coverage.
| 7 |
abstract. quantum machine learning witnesses an increasing amount of quantum
algorithms for data-driven decision making, a problem with potential applications
ranging from automated image recognition to medical diagnosis. many of those
algorithms are implementations of quantum classifiers, or models for the classification
of data inputs with a quantum computer. following the success of collective decision
making with ensembles in classical machine learning, this paper introduces the concept
of quantum ensembles of quantum classifiers. creating the ensemble corresponds
to a state preparation routine, after which the quantum classifiers are evaluated in
parallel and their combined decision is accessed by a single-qubit measurement. this
framework naturally allows for exponentially large ensembles in which – similar to
bayesian learning – the individual classifiers do not have to be trained. as an example,
we analyse an exponentially large quantum ensemble in which each classifier is weighed
according to its performance in classifying the training data, leading to new results for
quantum as well as classical machine learning.
| 10 |
abstract
crowdsourced 3d cad models are becoming easily accessible online, and can potentially generate an infinite
number of training images for almost any object category.
we show that augmenting the training data of contemporary
deep convolutional neural net (dcnn) models with such
synthetic data can be effective, especially when real training data is limited or not well matched to the target domain.
most freely available cad models capture 3d shape but are
often missing other low level cues, such as realistic object
texture, pose, or background. in a detailed analysis, we
use synthetic cad-rendered images to probe the ability of
dcnn to learn without these cues, with surprising findings.
in particular, we show that when the dcnn is fine-tuned
on the target detection task, it exhibits a large degree of invariance to missing low-level cues, but, when pretrained on
generic imagenet classification, it learns better when the
low-level cues are simulated. we show that our synthetic
dcnn training approach significantly outperforms previous methods on the pascal voc2007 dataset when learning in the few-shot scenario and improves performance in a
domain shift scenario on the office benchmark.
| 9 |
abstract. we show that a countable group is locally virtually cyclic if and only if its
bredon cohomological dimension for the family of virtually cyclic subgroups is at most one.
| 4 |
abstract
an adaptive system for the suppression of vibration transmission using a single piezoelectric
actuator shunted by a negative capacitance circuit is presented. it is known that using negative
capacitance shunt, the spring constant of piezoelectric actuator can be controlled to extreme values
of zero or infinity. since the value of spring constant controls a force transmitted through an elastic
element, it is possible to achieve a reduction of transmissibility of vibrations through a piezoelectric
actuator by reducing its effective spring constant. the narrow frequency range and broad frequency
range vibration isolation systems are analyzed, modeled, and experimentally investigated. the
problem of high sensitivity of the vibration control system to varying operational conditions is
resolved by applying an adaptive control to the circuit parameters of the negative capacitor. a
control law that is based on the estimation of the value of effective spring constant of shunted
piezoelectric actuator is presented. an adaptive system, which achieves a self-adjustment of the
negative capacitor parameters is presented. it is shown that such an arrangement allows a design
of a simple electronic system, which, however, offers a great vibration isolation efficiency in variable
vibration conditions.
keywords: piezoelectric actuator, vibration transmission suppression, piezoelectric shunt damping, negative capacitor, elastic stiffness control, adaptive device
| 5 |
abstract
graphs provide a powerful means for representing complex
interactions between entities. recently, new deep learning approaches have emerged for representing and modeling graphstructured data while the conventional deep learning methods,
such as convolutional neural networks and recurrent neural
networks, have mainly focused on the grid-structured inputs
of image and audio. leveraged by representation learning capabilities, deep learning-based techniques can detect structural characteristics of graphs, giving promising results for
graph applications. in this paper, we attempt to advance deep
learning for graph-structured data by incorporating another
component: transfer learning. by transferring the intrinsic geometric information learned in the source domain, our approach can construct a model for a new but related task in
the target domain without collecting new data and without
training a new model from scratch. we thoroughly tested our
approach with large-scale real-world text data and confirmed
the effectiveness of the proposed transfer learning framework
for deep learning on graphs. according to our experiments,
transfer learning is most effective when the source and target domains bear a high level of structural similarity in their
graph representations.
| 9 |
abstract
in survival analysis it often happens that some subjects under study do not experience the event of interest; they are considered to be ‘cured’. the population is thus
a mixture of two subpopulations : the one of cured subjects, and the one of ‘susceptible’ subjects. when covariates are present, a so-called mixture cure model can be
used to model the conditional survival function of the population. it depends on two
components : the probability of being cured and the conditional survival function of
the susceptible subjects.
in this paper we propose a novel approach to estimate a mixture cure model when
the data are subject to random right censoring. we work with a parametric model for
the cure proportion (like e.g. a logistic model), while the conditional survival function
of the uncured subjects is unspecified. the approach is based on an inversion which
allows to write the survival function as a function of the distribution of the observable
random variables. this leads to a very general class of models, which allows a flexible
and rich modeling of the conditional survival function. we show the identifiability of
the proposed model, as well as the weak consistency and the asymptotic normality of
the model parameters. we also consider in more detail the case where kernel estimators
are used for the nonparametric part of the model. the new estimators are compared
with the estimators from a cox mixture cure model via finite sample simulations.
finally, we apply the new model and estimation procedure on two medical data sets.
| 10 |
abstract
we prove that orthogonal constructor term rewrite systems and lambda-calculus with weak
(i.e., no reduction is allowed under the scope of a lambda-abstraction) call-by-value reduction can simulate each other with a linear overhead. in particular, weak call-by-value betareduction can be simulated by an orthogonal constructor term rewrite system in the same
number of reduction steps. conversely, each reduction in an term rewrite system can be
simulated by a constant number of beta-reduction steps. this is relevant to implicit computational complexity, because the number of beta steps to normal form is polynomially related
to the actual cost (that is, as performed on a turing machine) of normalization, under weak
call-by-value reduction. orthogonal constructor term rewrite systems and lambda-calculus
are thus both polynomially related to turing machines, taking as notion of cost their natural
parameters.
| 6 |
abstract
if every element of a matrix group is similar to a permutation matrix,
then it is called a permutation-like matrix group. references [4], [5] and
[6] showed that, if a permutation-like matrix group contains a maximal
cycle such that the maximal cycle generates a normal subgroup and the
length of the maximal cycle equals to a prime, or a square of a prime,
or a power of an odd prime, then the permutation-like matrix group is
similar to a permutation matrix group. in this paper, we prove that if
a permutation-like matrix group contains a maximal cycle such that the
maximal cycle generates a normal subgroup and the length of the maximal
cycle equals to any power of 2, then it is similar to a permutation matrix
group.
key words: permutation-like matrix group, permutation matrix group,
maximal cycle.
msc2010: 15a18, 15a30, 20h20.
| 4 |
abstract
the main problems of school course timetabling are time, curriculum, and classrooms. in addition there
are other problems that vary from one institution to another. this paper is intended to solve the problem of
satisfying the teachers’ preferred schedule in a way that regards the importance of the teacher to the
supervising institute, i.e. his score according to some criteria. genetic algorithm (ga) has been presented
as an elegant method in solving timetable problem (ttp) in order to produce solutions with no conflict. in
this paper, we consider the analytic hierarchy process (ahp) to efficiently obtain a score for each teacher,
and consequently produce a ga-based ttp solution that satisfies most of the teachers’ preferences.
| 9 |
abstract— over the years, data mining has attracted most of the
attention from the research community. the researchers attempt
to develop faster, more scalable algorithms to navigate over the
ever increasing volumes of spatial gene expression data in search
of meaningful patterns. association rules are a data mining
technique that tries to identify intrinsic patterns in spatial gene
expression data. it has been widely used in different applications,
a lot of algorithms introduced to discover these rules. however
priori-like algorithms has been used to find positive association
rules. in contrast to positive rules, negative rules encapsulate
relationship between the occurrences of one set of items with
absence of the other set of items. in this paper, an algorithm for
mining negative association rules from spatial gene expression
data is introduced. the algorithm intends to discover the negative
association rules which are complementary to the association
rules often generated by priori like algorithm. our study shows
that negative association rules can be discovered efficiently from
spatial gene expression data.
| 5 |
abstract—signal processing played an important role in improving the quality of communications over copper cables in
earlier dsl technologies. even more powerful signal processing
techniques are required to enable a gigabit per second data
rate in the upcoming g.fast standard. this new standard is
different from its predecessors in many respects. in particular,
g.fast will use a significantly higher bandwidth. at such a high
bandwidth, crosstalk between different lines in a binder will reach
unprecedented levels, which are beyond the capabilities of most
efficient techniques for interference mitigation. in this article, we
survey the state of the art and research challenges in the design of
signal processing algorithms for the g.fast system, with a focus on
novel research approaches and design considerations for efficient
interference mitigation in g.fast systems. we also detail relevant
vdsl techniques and points out their strengths and limitations
for the g.fast system.
| 7 |
abstract
this paper presents a model for the simulation of liquid-gas-solid flows by means of the
lattice boltzmann method. the approach is built upon previous works for the simulation of
liquid-solid particle suspensions on the one hand, and on a liquid-gas free surface model on
the other. we show how the two approaches can be unified by a novel set of dynamic cell
conversion rules. for evaluation, we concentrate on the rotational stability of non-spherical
rigid bodies floating on a plane water surface – a classical hydrostatic problem known from
naval architecture. we show the consistency of our method in this kind of flows and obtain
convergence towards the ideal solution for the measured heeling stability of a floating box.
1. introduction
since its establishment the lattice boltzmann method (lbm) has become a popular
alternative in the field of complex flow simulations [33]. its application to particle suspensions
has been propelled to a significant part by the works of ladd et al. [22, 23] and aidun et al.
[3, 4, 1]. based on the approach of the so-called momentum exchange method, it is possible to
calculate the hydromechanical stresses on the surface of fully resolved solid particles directly
from the lattice boltzmann boundary treatment. in this paper, the aforementioned fluidsolid coupling approach is extended to liquid-gas free surface flows, i.e., the problem of solid
bodies moving freely within a flow of two immiscible fluids. we use the free surface model
of [21, 30] to simulate a liquid phase in interaction with a gas by means of a volume of fluid
approach and a special kinematic free surface boundary condition. i.e., the interface of the
two phases is assumed sharp enough to be modeled by a locally defined boundary layer.
this boundary layer is updated dynamically according to the liquid advection by a set of
cell conversion rules.
this paper proposes a unification of the update rules of the free surface model with those
of the particulate flow model, which also requires a dynamical mapping of the respective solid
boundaries to the lattice boltzmann grid. as described in [5], the resulting scheme allows full
freedom of motion of the solid bodies in the flow, which can be calculated according to rigid
body physics as in [20]. we demonstrate the consistency of the combined liquid-gas-solid
method by means of a simple advection test with a floating body in a stratified liquid-gas
channel flow, and discuss the main source of error in the dynamic boundary handling with
particles in motion.
preprint submitted to elsevier
| 5 |
abstract syntax tree according to the provided
grammar and is implemented using recursive descent parsing techniques as described in [3]. the lexer is
a module used by this pass. errors and warnings reported by the first pass are only lexical and syntactic
errors. pass two traverses the syntax tree depth first, replacing all variables and constructs that are purely
elements of purl and are not expressible in the standard notation. these constructs will be discussed when
exploring the individual pattern elements. in the third pass, the syntax tree is again traversed depth first.
all verification occurs in this pass and errors indicate problems in the structure of the knitting pattern. a
global state object is used throughout parsing to track information necessary for error reporting, such as
section name, position in code, and row number in the generated pattern. it is also used in the verification
pass to track the pattern orientation, width, and row index, and to update nodes with these values as
necessary.
the reason for breaking up parsing into three passes is because a syntax tree representation of the
pattern is much easier to manipulate and verify. a main feature of purl is the ability to define modular
and parametrized segments of patterns, through the pattern sample construct introduced by this language
(see 2.15), so a second pass is used for trimming nodes representing sample calls. also, there are some
challenges in verifying a pattern. it is necessary that every row works all of the stitches of the previous
row, but there are some pattern constructs which work a number of stitches that depends on the width of
the current row. since we allow modular pattern definitions and parameterized segments of patterns, this
verification cannot be done in a single pass over the source language.
| 6 |
abstract
counting the number of permutations of a given total displacement is equivalent to counting
weighted motzkin paths of a given area (guay-paquet and petersen [10]). the former combinatorial problem is still open. in this work we show that this connection allows to construct
efficient algorithms for counting and for sampling such permutations. these algorithms provide
a tool to better understand the original combinatorial problem. a by-product of our approach
is a different way of counting based on certain “building sequences” for motzkin paths, which
may be of independent interest.
| 8 |
abstract
a promising technique to provide mobile applications with high computation resources is to offload
the processing task to the cloud. mobile cloud computing enables mobile devices with limited batteries
to run resource hungry applications with the help of abundant processing capabilities of the clouds and
to save power. however, it is not always true that cloud computing consumes less energy compared to
mobile edge computing. it may take more energy for the mobile device to transmit a file to the cloud
than running the task itself at the edge. this paper investigates the power minimization problem for the
mobile devices by data offloading in multi-cell multi-user ofdma mobile cloud computing networks.
we consider the maximum acceptable delay and tolerable interference as qos metrics to be satisfied
in our network. we formulate the problem as a mixed integer nonlinear problem which is converted
into a convex form using d.c. approximation. to solve the optimization problem, we have proposed
centralized and distributed algorithms for joint power allocation and channel assignment together with
decision making. our simulation results illustrate that by utilizing the proposed algorithms, considerable
power saving could be achieved e.g. about 60% for short delays and large bitstream sizes in comparison
with the baselines.
index terms
offloading, resource allocation, mobile cloud computing, mobile edge computing.
| 7 |
abstract. in this preliminary note, we will illustrate our ideas on automated mechanisms for termination and non-termination reasoning.
| 6 |
abstract: despite extensive research and remarkable advancements in the control of complex networks,
time-invariant control schedules (tics) still dominate
the literature. this is both due to their simplicity and
the fact that the potential benefits of time-varying control schedules (tvcs) have remained largely uncharacterized. yet, tvcs have the potential to significantly
enhance network controllability over tics, especially
when applied to large networks. in this paper we study
networks with linear and discrete-time dynamics and
analyze the role of network structure in tvcs. through
the analysis of a new scale-dependent notion of nodal
communicability, we show that optimal tvcs involves
the actuation of the most central nodes at appropriate spatial scales at all times. consequently, we show
that it is the scale-heterogeneity of the central-nodes
in a network that determine whether, and to what extent, tvcs outperforms conventional policies based on
tics. several analytical results and numerical examples support and illustrate this relationship.
| 3 |
abstract
let k be a nonperfect separably closed field. let g be a connected reductive algebraic group defined over k. we study rationality problems for serre’s notion of complete
reducibility of subgroups of g. in particular, we present a new example of subgroup h
of g of type d4 in characteristic 2 such that h is g-completely reducible but not gcompletely reducible over k (or vice versa). this is new: all known such examples are for
g of exceptional type. we also find a new counterexample for külshammer’s question on
representations of finite groups for g of type d4 . a problem concerning the number of
conjugacy classes is also considered. the notion of nonseparable subgroups plays a crucial
role in all our constructions.
| 4 |
abstract
an algebra has the howson property if the intersection of any two finitely generated subalgebras
is finitely generated. a simple necessary and sufficient condition is given for the howson property
to hold on an inverse semigroup with finitely many idempotents. in addition, it is shown that any
monogenic inverse semigroup has the howson property.
keywords: howson property; e-unitary; monogenic
mathematics subject classification: 20m18
| 4 |
abstract. the perfect phylogeny problem is a classic problem in computational biology, where we seek an unrooted phylogeny that is compatible with a set of qualitative characters. such a tree exists precisely
when an intersection graph associated with the character set, called the
partition intersection graph, can be triangulated using a restricted set
of fill edges. semple and steel used the partition intersection graph to
characterize when a character set has a unique perfect phylogeny. bordewich, huber, and semple showed how to use the partition intersection
graph to find a maximum compatible set of characters. in this paper, we
build on these results, characterizing when a unique perfect phylogeny
exists for a subset of partial characters. our characterization is stated in
terms of minimal triangulations of the partition intersection graph that
are uniquely representable, also known as ur-chordal graphs. our characterization is motivated by the structure of ur-chordal graphs, and the
fact that the block structure of minimal triangulations is mirrored in the
graph that has been triangulated.
| 5 |
abstract
we say that a polynomial automorphism φ in n variables is stably
co-tame if the tame subgroup in n variables is contained in the subgroup generated by φ and affine automorphisms in n + 1 variables. in
this paper, we give conditions for stably co-tameness of polynomial
automorphisms.
| 4 |
abstract
in this paper, we introduce a generalized value iteration network (gvin), which is an end-to-end neural network planning
module. gvin emulates the value iteration algorithm by using
a novel graph convolution operator, which enables gvin to
learn and plan on irregular spatial graphs. we propose three
novel differentiable kernels as graph convolution operators
and show that the embedding-based kernel achieves the best
performance. furthermore, we present episodic q-learning,
an improvement upon traditional n-step q-learning that stabilizes training for vin and gvin. lastly, we evaluate gvin
on planning problems in 2d mazes, irregular graphs, and realworld street networks, showing that gvin generalizes well
for both arbitrary graphs and unseen graphs of larger scale
and outperforms a naive generalization of vin (discretizing a
spatial graph into a 2d image).
| 2 |
abstract—this work proposes a new adaptive-robust control
(arc) architecture for a class of uncertain euler-lagrange (el)
systems where the upper bound of the uncertainty satisfies linear
in parameters (lip) structure. conventional arc strategies
either require structural knowledge of the system or presume
that the overall uncertainties or its time derivative is norm
bounded by a constant. due to unmodelled dynamics and
modelling imperfection, true structural knowledge of the system
is not always available. further, for the class of systems under
consideration, prior assumption regarding the uncertainties (or
its time derivative) being upper bounded by a constant, puts
a restriction on states beforehand. conventional arc laws
invite overestimation-underestimation problem of switching gain.
towards this front, adaptive switching-gain based robust control (asrc) is proposed which alleviates the overestimationunderestimation problem of switching gain. moreover, asrc
avoids any presumption of constant upper bound on the overall
uncertainties and can negotiate uncertainties regardless of being
linear or nonlinear in parameters. experimental results of
asrc using a wheeled mobile robot notes improved control
performance in comparison to adaptive sliding mode control.
index terms—adaptive-robust control, euler-lagrange systems, wheeled mobile robot, uncertainty.
| 3 |
abstract
python is a popular dynamic language with a large part of its
appeal coming from powerful libraries and extension modules. these augment the language and make it a productive
environment for a wide variety of tasks, ranging from web
development (django) to numerical analysis (numpy).
unfortunately, python’s performance is quite poor when
compared to modern implementations of languages such as
lua and javascript. why does python lag so far behind
these other languages? as we show, the very same api and
extension libraries that make python a powerful language
also make it very difficult to efficiently execute.
given that we want to retain access to the great extension
libraries that already exist for python, how fast can we make
it? to evaluate this, we designed and implemented falcon, a
high-performance bytecode interpreter fully compatible with
the standard cpython interpreter. falcon applies a number
of well known optimizations and introduces several new
techniques to speed up execution of python bytecode. in our
evaluation, we found falcon an average of 25% faster than
the standard python interpreter on most benchmarks and in
some cases about 2.5x faster.
| 6 |
abstract. this paper investigates how high school students approach computing
through an introductory computer science course situated in the logic programming (lp) paradigm. this study shows how novice students operate within the
lp paradigm while engaging in foundational computing concepts and skills, and
presents a case for lp as a viable paradigm choice for introductory cs courses.
keywords: cs education, high school cs, declarative programming, logic programming, answer set programming
| 2 |
abstract
we prove an explicit formula for the first non-zero entry in the n-th
row of the graded betti table of an n-dimensional projective toric variety
associated to a normal polytope with at least one interior lattice point.
this applies to veronese embeddings of pn . we also prove an explicit
formula for the entire n-th row when the interior of the polytope is onedimensional. all results are valid over an arbitrary field k.
| 0 |
abstract— as autonomous service robots become more affordable and thus available also for the general public, there
is a growing need for user friendly interfaces to control the
robotic system. currently available control modalities typically
expect users to be able to express their desire through either
touch, speech or gesture commands. while this requirement
is fulfilled for the majority of users, paralyzed users may not
be able to use such systems. in this paper, we present a novel
framework, that allows these users to interact with a robotic
service assistant in a closed-loop fashion, using only thoughts.
the brain-computer interface (bci) system is composed of
several interacting components, i.e., non-invasive neuronal signal recording and decoding, high-level task planning, motion
and manipulation planning as well as environment perception.
in various experiments, we demonstrate its applicability and
robustness in real world scenarios, considering fetch-and-carry
tasks and tasks involving human-robot interaction. as our
results demonstrate, our system is capable of adapting to
frequent changes in the environment and reliably completing
given tasks within a reasonable amount of time. combined with
high-level planning and autonomous robotic systems, interesting
new perspectives open up for non-invasive bci-based humanrobot interactions.
| 2 |
abstract
| 2 |
abstraction
a directed graph g (v, e) is strongly connected if and only if, for any pair of vertices x and y from
v, there exists a path from x to y and a path from y to x. in computer science, the partition of a
graph in strongly connected components is represented by the partition of all vertices from the
graph, so that for any two vertices, x and y, from the same partition, there exists a path from x
to y and a path from y to x and for any two vertices, u and v, from different partition, the
property is not met. the algorithm presented below is meant to find the partition of a given graph
in strongly connected components in o (numberofnodes + numberofedges * log*
(numberofnodes)), where log* function stands for iterated logarithm.
| 8 |
abstract
first-order logic (fol) is widely regarded as one of the most important foundations for knowledge representation. nevertheless, in this paper, we argue that
fol has several critical issues for this purpose. instead, we propose an alternative called assertional logic, in which all syntactic objects are categorized as set
theoretic constructs including individuals, concepts and operators, and all kinds of
knowledge are formalized by equality assertions. we first present a primitive form
of assertional logic that uses minimal assumed knowledge and constructs. then,
we show how to extend it by definitions, which are special kinds of knowledge,
i.e., assertions. we argue that assertional logic, although simpler, is more expressive and extensible than fol. as a case study, we show how assertional logic can
be used to unify logic and probability, and more building blocks in ai.
| 2 |
abstract
this article is devoted to study the effects of the s-periodical fractional differencing filter (1 − ls )dt . to put this effect in evidence, we
have derived the periodic auto-covariance functions of two distinct univariate seasonally fractionally differenced periodic models. a multivariate
representation of periodically correlated process is exploited to provide
the exact and approximated expression auto-covariance of each models.
the distinction between the models is clearly obvious through the expression of periodic auto-covariance function. besides producing different
auto-covariance functions, the two models differ in their implications.
in the first model, the seasons of the multivariate series are separately
fractionally integrated. in the second model, however, the seasons for the
univariate series are fractionally co-integrated. on the simulated sample,
for each models, with the same parameters, the empirical periodic autocovariance are calculated and graphically represented for illustrating the
results and support the comparison between the two models.
| 10 |
abstract—cooperative adaptive cruise control (cacc) is
one of the driving applications of vehicular ad-hoc networks
(vanets) and promises to bring more efficient and faster
transportation through cooperative behavior between vehicles.
in cacc, vehicles exchange information, which is relied on to
partially automate driving; however, this reliance on cooperation
requires resilience against attacks and other forms of misbehavior. in this paper, we propose a rigorous attacker model
and an evaluation framework for this resilience by quantifying
the attack impact, providing the necessary tools to compare
controller resilience and attack effectiveness simultaneously. although there are significant differences between the resilience of
the three analyzed controllers, we show that each can be attacked
effectively and easily through either jamming or data injection.
our results suggest a combination of misbehavior detection
and resilient control algorithms with graceful degradation are
necessary ingredients for secure and safe platoons.
| 3 |
abstract. let g and g0 be two right-angled artin groups. we show they
are quasi-isometric if and only if they are isomorphic, under the assumption
that the outer automorphism groups out(g) and out(g0 ) are finite. if we
only assume out(g) is finite, then g0 is quasi-isometric g if and only if g0
is isomorphic to a subgroup of finite index in g. in this case, we give an
algorithm to determine whether g and g0 are quasi-isometric by looking at
their defining graphs.
| 4 |
abstract
we study machine learning formulations of inductive program synthesis; that is,
given input-output examples, synthesize source code that maps inputs to corresponding outputs. our key contribution is t erpre t, a domain-specific language
for expressing program synthesis problems. a t erpre t model is composed of
a specification of a program representation and an interpreter that describes how
programs map inputs to outputs. the inference task is to observe a set of inputoutput examples and infer the underlying program. from a t erpre t model we
automatically perform inference using four different back-ends: gradient descent
(thus each t erpre t model can be seen as defining a differentiable interpreter),
linear program (lp) relaxations for graphical models, discrete satisfiability solving,
and the s ketch program synthesis system. t erpre t has two main benefits. first,
it enables rapid exploration of a range of domains, program representations, and
interpreter models. second, it separates the model specification from the inference
algorithm, allowing proper comparisons between different approaches to inference.
we illustrate the value of t erpre t by developing several interpreter models and
performing an extensive empirical comparison between alternative inference algorithms on a variety of program models. to our knowledge, this is the first work
to compare gradient-based search over program space to traditional search-based
alternatives. our key empirical finding is that constraint solvers dominate the gradient descent and lp-based formulations.
| 9 |
abstract
the automatic coding of clinical documentation according
to diagnosis codes is a useful task in the electronic health
record, but a challenging one due to the large number
of codes and the length of patient notes. we investigate
four models for assigning multiple icd codes to discharge
summaries, and experiment with data from the mimic ii
and iii clinical datasets. we present hierarchical attentionbidirectional gated recurrent unit (ha-gru), a hierarchical approach to tag a document by identifying the sentences
relevant for each label. ha-gru achieves state-of-the art results. furthermore, the learned sentence-level attention layer
highlights the model decision process, allows for easier error
analysis, and suggests future directions for improvement.
| 2 |
abstract. in this article, we introduce a procedure for selecting variables in principal components analysis. it is developed to identify a small
subset of the original variables that best explain the principal components through nonparametric relationships. there are usually some
noisy uninformative variables in a dataset, and some variables that are
strongly related to one another because of their general dependence. the
procedure is designed to be used following the satisfactory initial principal components analysis with all variables, and its aim is to help to
interpret the underlying structures. we analyze the asymptotic behavior
of the method and provide some examples.
| 10 |
abstract
| 5 |
abstract. lyubeznik’s conjecture, ([ly1], remark 3.7) asserts the finiteness
of the set ssociated primes of local cohomology modules for regular rings. but,
in the case of ramified regular local ring, it is open. recently, in theorem 1.2
of [nu], it is proved that in any noetherian regular local ring s and for a fixed
ideal j ⊂ s, associated primes of local cohomology hji (s) for i ≥ 0 is finite, if
it does not contain p. in this paper, we use this result to construct examples
of local cohomology modules for ramified regular local ring so that they have
finitely many associated primes.
| 0 |
abstract. we analyze stable homology over associative rings and obtain results over artin algebras and commutative noetherian rings. our study develops similarly for these classes; for simplicity we only discuss the latter here.
stable homology is a broad generalization of tate homology. vanishing of
stable homology detects classes of rings—among them gorenstein rings, the
original domain of tate homology. closely related to gorensteinness of rings is
auslander’s g-dimension for modules. we show that vanishing of stable homology detects modules of finite g-dimension. this is the first characterization
of such modules in terms of vanishing of (co)homology alone.
stable homology, like absolute homology, tor, is a theory in two variables.
it can be computed from a flat resolution of one module together with an
injective resolution of the other. this betrays that stable homology is not
balanced in the way tor is balanced. in fact, we prove that a ring is gorenstein
if and only if stable homology is balanced.
| 0 |
abstract
| 5 |
abstract
it was recently proven that all free and many virtually free verbally closed subgroups are algebraically closed in any group. we establish sufficient conditions for a group that is an extension
of a free non-abelian group by a group satisfying a non-trivial law to be algebraically closed in any
group in which it is verbally closed. we apply these conditions to prove that the fundamental groups
of all closed surfaces, except the klein bottle, and almost all free products of groups satisfying a
non-trivial law are algebraically closed in any group in which they are verbally closed.
| 4 |
abstract
| 8 |
abstract
this paper studies synchronization of dynamical networks with event-based communication. firstly, two estimators are
introduced into each node, one to estimate its own state, and the other to estimate the average state of its neighbours. then,
with these two estimators, a distributed event-triggering rule (etr) with a dwell time is designed such that the network
achieves synchronization asymptotically with no zeno behaviours. the designed etr only depends on the information that
each node can obtain, and thus can be implemented in a decentralized way.
key words: distributed event-triggered control, asymptotic synchronization, dynamical networks.
| 3 |
abstract
| 3 |
abstract
this paper considers fully dynamic graph algorithms with both faster worst case update
time and sublinear space. the fully dynamic graph connectivity problem is the following: given
a graph on a fixed set of n nodes, process an online sequence of edge insertions, edge deletions,
and queries of the form “is there a path between nodes a and b?” in 2013, the first data structure
was presented with worst case time per operation which was polylogarithmic in n. in this paper,
we shave off a factor of log n from that time, to o(log 4 n) per update. for sequences which are
polynomial in length, our algorithm answers queries in o(log n/ log log n) time correctly with
high probability and using o(n log2 n) words (of size log n). this matches the amount of space
used by the most space-efficient graph connectivity streaming algorithm. we also show that
2-edge connectivity can be maintained using o(n log2 n) words with an amortized update time
of o(log6 n).
| 8 |
abstract— in this paper a binary feature based loop closure
detection (lcd) method is proposed, which for the first time
achieves higher precision-recall (pr) performance compared
with state-of-the-art sift feature based approaches. the proposed system originates from our previous work multi-index
hashing for loop closure detection (mild), which employs
multi-index hashing (mih) [1] for approximate nearest neighbor (ann) search of binary features. as the accuracy of mild
is limited by repeating textures and inaccurate image similarity
measurement, burstiness handling is introduced to solve this
problem and achieves considerable accuracy improvement.
additionally, a comprehensive theoretical analysis on mih
used in mild is conducted to further explore the potentials
of hashing methods for ann search of binary features from
probabilistic perspective. this analysis provides more freedom
on best parameter choosing in mih for different application
scenarios. experiments on popular public datasets show that
the proposed approach achieved the highest accuracy compared
with state-of-the-art while running at 30hz for databases
containing thousands of images.
| 1 |
abstract
we study the complexity of the problem detection pair. a detection pair of a graph
g is a pair (w, l) of sets of detectors with w ⊆ v (g), the watchers, and l ⊆ v (g), the
listeners, such that for every pair u, v of vertices that are not dominated by a watcher of w ,
there is a listener of l whose distances to u and to v are different. the goal is to minimize
|w | + |l|. this problem generalizes the two classic problems dominating set and metric
dimension, that correspond to the restrictions l = ∅ and w = ∅, respectively. detection
pair was recently introduced by finbow, hartnell and young [a. s. finbow, b. l. hartnell
and j. r. young. the complexity of monitoring a network with both watchers and listeners.
networks, accepted], who proved it to be np-complete on trees, a surprising result given that
both dominating set and metric dimension are known to be linear-time solvable on trees.
it follows from an existing reduction by hartung and nichterlein for metric dimension that
even on bipartite subcubic graphs of arbitrarily large girth, detection pair is np-hard to
approximate within a sub-logarithmic factor and w[2]-hard (when parameterized by solution
size). we show, using a reduction to set cover, that detection pair is approximable
within a factor logarithmic in the number of vertices of the input graph. our two main results
are a linear-time 2-approximation algorithm and an fpt algorithm for detection pair on
trees.
keywords: graph theory, detection pair, metric dimension, dominating set, approximation algorithm, parameterized complexity
| 8 |
abstract. let g be the circulant graph cn (s) with s ⊆ {1, 2, . . . , ⌊ n2 ⌋}, and let i(g)
denote its the edge ideal in the ring r = k[x1 , . . . , xn ]. we consider the problem of
determining when g is cohen-macaulay, i.e, r/i(g) is a cohen-macaulay ring. because
a cohen-macaulay graph g must be well-covered, we focus on known families of wellcovered circulant graphs of the form cn (1, 2, . . . , d). we also characterize which cubic
circulant graphs are cohen-macaulay. we end with the observation that even though
the well-covered property is preserved under lexicographical products of graphs, this is
not true of the cohen-macaulay property.
| 0 |
abstractions for optimal checkpointing in
inversion problems
navjot kukreja∗
| 5 |
abstract
the k-co-path set problem asks, given a graph g and a positive
integer k, whether one can delete k edges from g so that the remainder
is a collection of disjoint paths. we give a linear-time fpt algorithm
with complexity o∗ (1.588k ) for deciding k-co-path set, significantly
improving the previously best known o∗ (2.17k ) of feng, zhou, and wang
(2015). our main tool is a new o∗ (4tw(g) ) algorithm for co-path set
using the cut&count framework, where tw(g) denotes treewidth. in
general graphs, we combine this with a branching algorithm which refines
a 6k-kernel into reduced instances, which we prove have bounded treewidth.
| 8 |
abstract. we present a parametric abstract domain for array content
analysis. the method maintains invariants for contiguous regions of the
array, similar to the methods of gopan, reps and sagiv, and of halbwachs and péron. however, it introduces a novel concept of an array
content graph, avoiding the need for an up-front factorial partitioning
step. the resulting analysis can be used with arbitrary numeric relational abstract domains; we evaluate the domain on a range of array
manipulating program fragments.
| 6 |
abstract
satellite radar altimetry is one of the most powerful techniques for measuring sea surface height variations,
with applications ranging from operational oceanography to climate research. over open oceans, altimeter
return waveforms generally correspond to the brown model, and by inversion, estimated shape parameters
provide mean surface height and wind speed. however, in coastal areas or over inland waters, the waveform
shape is often distorted by land influence, resulting in peaks or fast decaying trailing edges. as a result,
derived sea surface heights are then less accurate and waveforms need to be reprocessed by sophisticated
algorithms. to this end, this work suggests a novel spatio-temporal altimetry retracking (star) technique. we show that star enables the derivation of sea surface heights over the open ocean as well as
over coastal regions of at least the same quality as compared to existing retracking methods, but for a
larger number of cycles and thus retaining more useful data. novel elements of our method are (a) integrating information from spatially and temporally neighboring waveforms through a conditional random
field approach, (b) sub-waveform detection, where relevant sub-waveforms are separated from corrupted or
non-relevant parts through a sparse representation approach, and (c) identifying the final best set of sea
surfaces heights from multiple likely heights using dijkstra’s algorithm. we apply star to data from the
jason-1, jason-2 and envisat missions for study sites in the gulf of trieste, italy and in the coastal region
of the ganges-brahmaputra-meghna estuary, bangladesh. we compare to several established and recent
retracking methods, as well as to tide gauge data. our experiments suggest that the obtained sea surface
heights are significantly less affected by outliers when compared to results obtained by other approaches.
keywords: coastal, oceans, altimetry, retracking, sea surface heights, conditional random fields, sparse
representation
| 1 |
abstract
an extremely simple, description of karmarkar’s algorithm with very few
technical terms is given.
| 8 |
abstract. we prove that the word problem is undecidable in functionally recursive groups, and that the order problem is undecidable in automata groups,
even under the assumption that they are contracting.
| 4 |
abstract
consider a process satisfying a stochastic differential equation with unknown drift
parameter, and suppose that discrete observations are given. it is known that a simple
least squares estimator (lse) can be consistent, but numerically unstable in the sense
of large standard deviations under finite samples when the noise process has jumps.
we propose a filter to cut large shocks from data, and construct the same lse from
data selected by the filter. the proposed estimator can be asymptotically equivalent to
the usual lse, whose asymptotic distribution strongly depends on the noise process.
however, in numerical study, it looked asymptotically normal in an example where
filter was choosen suitably, and the noise was a lévy process. we will try to justify
this phenomenon mathematically, under certain restricted assumptions.
key words: stochastic differential equation, semimartingale noise, small noise
asymptotics, drift estimation, threshold estimator, mighty convergence.
msc2010: 62f12, 62m05; 60g52, 60j75
| 10 |
abstract
we study approximations of the partition function of dense graphical models. partition
functions of graphical models play a fundamental role is statistical physics, in statistics and in
machine learning. two of the main methods for approximating the partition function are markov
chain monte carlo and variational methods. an impressive body of work in mathematics,
physics and theoretical computer science provides conditions under which markov chain monte
carlo methods converge in polynomial time. these methods often lead to polynomial time
approximation algorithms for the partition function in cases where the underlying model exhibits
correlation decay. there are very few theoretical guarantees for the performance of variational
methods. one exception is recent results by risteski (2016) who considered dense graphical
models and showed that using variational methods, it is possible to find an o(ǫn) additive
2
approximation to the log partition function in time no(1/ǫ ) even in a regime where correlation
decay does not hold.
we show that under essentially the same conditions, an o(ǫn) additive approximation of the
log partition function can be found in constant time, independent of n. in particular, our results
cover dense ising and potts models as well as dense graphical models with k-wise interaction.
they also apply for low threshold rank models.
to the best of our knowledge, our results are the first to give a constant time approximation
to log partition functions and the first to use the algorithmic regularity lemma for estimating
partition functions. as an application of our results we derive a constant time algorithm for
approximating the magnetization of ising and potts model on dense graphs.
| 8 |
abstract—millimeter-wave (mmwave) communication and
network densification hold great promise for achieving highrate communication in next-generation wireless networks. cloud
radio access network (cran), in which low-complexity remote
radio heads (rrhs) coordinated by a central unit (cu) are
deployed to serve users in a distributed manner, is a costeffective solution to achieve network densification. however, when
operating over a large bandwidth in the mmwave frequencies,
the digital fronthaul links in a cran would be easily saturated
by the large amount of sampled and quantized signals to be
transferred between rrhs and the cu. to tackle this challenge,
we propose in this paper a new architecture for mmwavebased cran with advanced lens antenna arrays at the rrhs.
due to the energy focusing property, lens antenna arrays are
effective in exploiting the angular sparsity of mmwave channels,
and thus help in substantially reducing the fronthaul rate and
simplifying the signal processing at the multi-antenna rrhs
and the cu, even when the channels are frequency-selective.
we consider the uplink transmission in a mmwave cran with
lens antenna arrays and propose a low-complexity quantization
bit allocation scheme for multiple antennas at each rrh to
meet the given fronthaul rate constraint. further, we propose
a channel estimation technique that exploits the energy focusing
property of the lens array and can be implemented at the cu
with low complexity. finally, we compare the proposed mmwave
cran using lens antenna arrays with a conventional cran
using uniform planar arrays at the rrhs, and show that the
proposed design achieves significant throughput gains, yet with
much lower complexity.
index terms—cloud radio access network, millimeter-wave
communication, lens antenna array, channel estimation, fronthaul constraint, antenna selection, quantization bit allocation.
| 7 |
abstract
an oracle is a design for potentially high power
artificial intelligences (ais), where the ai is
made safe by restricting it to only answer questions. unfortunately most designs cause the oracle to be motivated to manipulate humans with
the contents of their answers, and oracles of potentially high intelligence might be very successful at this. solving that problem, without compromising the accuracy of the answer, is tricky.
this paper reduces the issue to a cryptographicstyle problem of alice ensuring that her oracle
answers her questions while not providing key
information to an eavesdropping eve. two oracle designs solve this problem, one counterfactual (the oracle answers as if it expected its answer to never be read) and one on-policy, but limited by the quantity of information it can transmit.
| 2 |
abstract
this paper describes a generic algorithm for concurrent
resizing and on-demand per-bucket rehashing for an extensible
hash table. in contrast to known lock-based hash table algorithms,
the proposed algorithm separates the resizing and rehashing stages
so that they neither invalidate existing buckets nor block any
concurrent operations. instead, the rehashing work is deferred and
split across subsequent operations with the table. the rehashing
operation uses bucket-level synchronization only and therefore
allows a race condition between lookup and moving operations
running in different threads. instead of using explicit
synchronization, the algorithm detects the race condition and
restarts the lookup operation. in comparison with other lock-based
algorithms, the proposed algorithm reduces high-level
synchronization on the hot path, improving performance,
concurrency, and scalability of the table. the response time of the
operations is also more predictable. the algorithm is compatible
with cache friendly data layouts for buckets and does not depend
on any memory reclamation techniques thus potentially achieving
additional performance gain with corresponding implementations.
categories and subject descriptors: d.1.3 [programming
techniques]: concurrent programming – parallel programming;
d.4.1 [operating systems]: process management –
synchronization;
concurrency;
multiprocessing,
multiprogramming,
multitasking;
e.2
[data
storage
representation] – hash-table representations.
general terms:
| 8 |
abstract
most reinforcement learning algorithms are inefficient for learning multiple tasks in complex robotic
systems, where different tasks share a set of actions. in such environments a compound policy
may be learnt with shared neural network parameters, which performs multiple tasks concurrently.
however such compound policy may get biased towards a task or the gradients from different tasks
negate each other, making the learning unstable
and sometimes less data efficient. in this paper,
we propose a new approach for simultaneous training of multiple tasks sharing a set of common actions in continuous action spaces, which we call as
digrad (differential policy gradient). the proposed framework is based on differential policy
gradients and can accommodate multi-task learning
in a single actor-critic network. we also propose
a simple heuristic in the differential policy gradient update to further improve the learning. the
proposed architecture was tested on 8 link planar
manipulator and 27 degrees of freedom(dof) humanoid for learning multi-goal reachability tasks
for 3 and 2 end effectors respectively. we show that
our approach supports efficient multi-task learning
in complex robotic systems, outperforming related
methods in continuous action spaces.
| 2 |
abstract—this paper deals with the certification problem
for robust quadratic stability, robust state convergence, and
robust quadratic performance of linear systems that exhibit
bounded rates of variation in their parameters. we consider both
continuous-time (ct) and discrete-time (dt) parameter-varying
systems. in this paper, we provide a uniform method for this
certification problem in both cases and we show that, contrary to
what was claimed previously, the dt case requires a significantly
different treatment compared to the existing ct results. in the
established uniform approach, quadratic lyapunov functions,
that are affine in the parameter, are used to certify robust
stability, robust convergence rates, and robust performance in
terms of linear matrix inequality feasibility tests. to exemplify
the procedure, we solve the certification problem for l2 -gain
performance both in the ct and the dt cases. a numerical
example is given to show that the proposed approach is less
conservative than a method with slack variables.
index terms—linear parameter-varying systems; parametervarying lyapunov functions; stability of linear systems; lmis.
| 3 |
abstract
blind source separation, i.e. extraction of independent sources from
a mixture, is an important problem for both artificial and natural signal
processing. here, we address a special case of this problem when sources
(but not the mixing matrix) are known to be nonnegative, for example,
due to the physical nature of the sources. we search for the solution to
this problem that can be implemented using biologically plausible neural
networks. specifically, we consider the online setting where the dataset
is streamed to a neural network. the novelty of our approach is that we
formulate blind nonnegative source separation as a similarity matching
problem and derive neural networks from the similarity matching objective.
importantly, synaptic weights in our networks are updated according to
biologically plausible local learning rules.
| 9 |
abstract
ultra-reliable, low latency communications (urllc) are currently attracting significant attention
due to the emergence of mission-critical applications and device-centric communication. urllc will
entail a fundamental paradigm shift from throughput-oriented system design towards holistic designs for
guaranteed and reliable end-to-end latency. a deep understanding of the delay performance of wireless
networks is essential for efficient urllc systems. in this paper, we investigate the network layer
performance of multiple-input, single-output (miso) systems under statistical delay constraints. we
provide closed-form expressions for miso diversity-oriented service process and derive probabilistic
delay bounds using tools from stochastic network calculus. in particular, we analyze transmit beamforming with perfect and imperfect channel knowledge and compare it with orthogonal space-time codes
and antenna selection. the effect of transmit power, number of antennas, and finite blocklength channel
coding on the delay distribution is also investigated. our higher layer performance results reveal key
insights of miso channels and provide useful guidelines for the design of ultra-reliable communication
systems that can guarantee the stringent urllc latency requirements.
| 7 |
abstract explores two extensions to this system, which we explain in the
context of the actor model (although they are equally applicable to a system using locks). rather
than rejecting programs where actors leak internal objects, we allow an actor to bestow its
synchronisation mechanism upon the exposed objects. this allows multiple objects to effectively
construct an actor’s interface. exposing internal operations externally makes concurrency more
fine-grained. to allow external control of the possible interleaving of these operations, we introduce
an atomic block that groups them together. the following section motivates these extensions.
v.t. vasconcelos and p. haller (eds.): workshop on programming language
approaches to concurrency- and communication-centric software (places’17)
eptcs 246, 2017, pp. 10–20, doi:10.4204/eptcs.246.4
| 6 |
abstract: high-dimensional linear regression with interaction effects is
broadly applied in research fields such as bioinformatics and social science.
in this paper, we first investigate the minimax rate of convergence for regression estimation in high-dimensional sparse linear models with two-way
interactions. we derive matching upper and lower bounds under three types
of heredity conditions: strong heredity, weak heredity and no heredity. from
the results: (i) a stronger heredity condition may or may not drastically
improve the minimax rate of convergence. in fact, in some situations, the
minimax rates of convergence are the same under all three heredity conditions; (ii) the minimax rate of convergence is determined by the maximum
of the total price of estimating the main effects and that of estimating the
interaction effects, which goes beyond purely comparing the order of the
number of non-zero main effects r1 and non-zero interaction effects r2 ; (iii)
under any of the three heredity conditions, the estimation of the interaction
terms may be the dominant part in determining the rate of convergence for
two different reasons: 1) there exist more interaction terms than main effect
terms or 2) a large ambient dimension makes it more challenging to estimate
even a small number of interaction terms. second, we construct an adaptive
estimator that achieves the minimax rate of convergence regardless of the
true heredity condition and the sparsity indices r1 , r2 .
msc 2010 subject classifications: primary 62c20; secondary 62j05.
keywords and phrases: minimax rate of convergence, sparsity, highdimensional regression, quadratic model, interaction selection, heredity condition, hierarchical structure, adaptive estimation.
| 10 |
abstract
this report studies data-driven estimation of the directed information (di) measure between
twoem discrete-time and continuous-amplitude random process, based on the k-nearest-neighbors
(k-nn) estimation framework. detailed derivations of two k-nn estimators are provided. the two
estimators differ in the metric based on which the nearest-neighbors are found. to facilitate the
estimation of the di measure, it is assumed that the observed sequences are (jointly) markovian
of order m. as m is generally not known, a data-driven method (that is also based on the k-nn
principle) for estimating m from the observed sequences is presented. an exhaustive numerical
study shows that the discussed k-nn estimators perform well even for relatively small number
of samples (few thousands). moreover, it is shown that the discussed estimators are capable of
accurately detecting linear as well as non-linear causal interactions.
| 7 |
abstract
graphical models use the intuitive and well-studied methods of graph theory to implicitly represent
dependencies between variables in large systems. they can model the global behaviour of a complex
system by specifying only local factors.this thesis studies inference in discrete graphical models from
an “algebraic perspective” and the ways inference can be used to express and approximate np-hard
combinatorial problems.
we investigate the complexity and reducibility of various inference problems, in part by organizing
them in an inference hierarchy. we then investigate tractable approximations for a subset of these
problems using distributive law in the form of message passing. the quality of the resulting message
passing procedure, called belief propagation (bp), depends on the influence of loops in the graphical
model. we contribute to three classes of approximations that improve bp for loopy graphs (i) loop
correction techniques; (ii) survey propagation, another message passing technique that surpasses bp
in some settings; and (iii) hybrid methods that interpolate between deterministic message passing and
markov chain monte carlo inference.
we then review the existing message passing solutions and provide novel graphical models and inference techniques for combinatorial problems under three broad classes: (i) constraint satisfaction problems (csps) such as satisfiability, coloring, packing, set / clique-cover and dominating / independent set
and their optimization counterparts; (ii) clustering problems such as hierarchical clustering, k-median,
k-clustering, k-center and modularity optimization; (iii) problems over permutations including (bottleneck) assignment, graph “morphisms” and alignment, finding symmetries and (bottleneck) traveling
salesman problem. in many cases we show that message passing is able to find solutions that are either
near optimal or favourably compare with today’s state-of-the-art approaches.
| 0 |
abstract
consider the standard nonparametric regression model and take as estimator the penalized least squares function. in this article, we study the trade-off
between closeness to the true function and complexity penalization of the estimator, where complexity is described by a seminorm on a class of functions.
first, we present an exponential concentration inequality revealing the concentration behavior of the trade-off of the penalized least squares estimator around
a nonrandom quantity, where such quantity depends on the problem under consideration. then, under some conditions and for the proper choice of the tuning
parameter, we obtain bounds for this nonrandom quantity. we illustrate our results with some examples that include the smoothing splines estimator.
keywords: concentration inequalities, regularized least squares,
statistical trade-off.
| 10 |
abstract. this is a short survey on existing upper and lower
bounds on the probability of the union of a finite number of events
using partial information given in terms of the individual or pairwise event probabilities (or their sums). new proofs for some of
the existing bounds are provided and new observations regarding
the existing gallot–kounias bound are given.
| 7 |
abstract—we present kleuren, a novel assembly-free
method to reconstruct phylogenetic trees using the colored
de bruijn graph. kleuren works by constructing the colored de bruijn graph and then traversing it, finding bubble
structures in the graph that provide phylogenetic signal.
the bubbles are then aligned and concatenated to form a
supermatrix, from which a phylogenetic tree is inferred. we
introduce the algorithms that kleuren uses to accomplish
this task, and show its performance on reconstructing the
phylogenetic tree of 12 drosophila species. kleuren reconstructed the established phylogenetic tree accurately and
is a viable tool for phylogenetic tree reconstruction using
whole genome sequences. software package available at:
https://github.com/colelyman/kleuren.
keywords-phylogenetics; algorithm; whole genome sequence;
colored de bruijn graph
| 8 |
abstract— we consider an energy harvesting transmitter sending status updates regarding a physical phenomenon it observes
to a receiver. different from the existing literature, we consider
a scenario where the status updates carry information about
an independent message. the transmitter encodes this message
into the timings of the status updates. the receiver needs to
extract this encoded information, as well as update the status
of the observed phenomenon. the timings of the status updates,
therefore, determine both the age of information (aoi) and the
message rate (rate). we study the tradeoff between the achievable
message rate and the achievable average aoi. we propose several
achievable schemes and compare their rate-aoi performances.
| 7 |
abstract
the evolution of the hemagglutinin amino acids sequences of influenza a virus is studied by a method
based on an informational metrics, originally introduced by rohlin for partitions in abstract probability
spaces. this metrics does not require any previous functional or syntactic knowledge about the sequences
and it is sensitive to the correlated variations in the characters disposition. its efficiency is improved by
algorithmic tools, designed to enhance the detection of the novelty and to reduce the noise of useless
mutations. we focus on the usa data from 1993/94 to 2010/2011 for a/h3n2 and on usa data from
2006/07 to 2010/2011 for a/h1n1 . we show that the clusterization of the distance matrix gives strong
evidence to a structure of domains in the sequence space, acting as weak attractors for the evolution,
in very good agreement with the epidemiological history of the virus. the structure proves very robust
with respect to the variations of the clusterization parameters, and extremely coherent when restricting
the observation window. the results suggest an efficient strategy in the vaccine forecast, based on the
presence of ”precursors” (or ”buds”) populating the most recent attractor.
| 5 |
abstract. finding the common structural features of two molecules is
a fundamental task in cheminformatics. most drugs are small molecules,
which can naturally be interpreted as graphs. hence, the task is formalized as maximum common subgraph problem. albeit the vast majority
of molecules yields outerplanar graphs this problem remains np-hard.
we consider a variation of the problem of high practical relevance, where
the rings of molecules must not be broken, i.e., the block and bridge structure of the input graphs must be retained by the common subgraph. we
present an algorithm for finding a maximum common connected induced
subgraph of two given outerplanar graphs subject to this constraint. our
approach runs in time o(∆n2 ) in outerplanar graphs on n vertices with
maximum degree ∆. this leads to a quadratic time complexity in molecular graphs, which have bounded degree. the experimental comparison
on synthetic and real-world datasets shows that our approach is highly
efficient in practice and outperforms comparable state-of-the-art algorithms.
| 8 |
abstract of thesis presented to lncc/mct in partial fulfillment of the requirements for the degree of doctor of sciences (d.sc.)
managing large-scale scientific hypotheses as
uncertain and probabilistic data
bernardo gonçalves
february - 2015
| 5 |
abstract
transversality is a simple and effective method for implementing quantum computation faulttolerantly. however, no quantum error-correcting code (qecc) can transversally implement a
quantum universal gate set (eastin and knill, phys. rev. lett., 102, 110502). since reversible
classical computation is often a dominating part of useful quantum computation, whether or not
it can be implemented transversally is an important open problem. we show that, other than
a small set of non-additive codes that we cannot rule out, no binary qecc can transversally
implement a classical reversible universal gate set. in particular, no such qecc can implement
the toffoli gate transversally.
we prove our result by constructing an information theoretically secure (but inefficient) quantum homomorphic encryption (its-qhe) scheme inspired by ouyang et al. (arxiv:1508.00938).
homomorphic encryption allows the implementation of certain functions directly on encrypted
data, i.e. homomorphically. our scheme builds on almost any qecc, and implements that
code’s transversal gate set homomorphically. we observe a restriction imposed by nayak’s
bound (focs 1999) on its-qhe, implying that any its quantum fully homomorphic scheme
(its-qfhe) implementing the full set of classical reversible functions must be highly inefficient. while our scheme incurs exponential overhead, any such qecc implementing toffoli
transversally would still violate this lower bound through our scheme.
| 7 |
abstract
minwise hashing is a fundamental and one of
the most successful hashing algorithm in the literature. recent advances based on the idea of
densification (shrivastava & li, 2014a;c) have
shown that it is possible to compute k minwise hashes, of a vector with d nonzeros, in
mere (d + k) computations, a significant improvement over the classical o(dk). these advances have led to an algorithmic improvement
in the query complexity of traditional indexing
algorithms based on minwise hashing. unfortunately, the variance of the current densification
techniques is unnecessarily high, which leads to
significantly poor accuracy compared to vanilla
minwise hashing, especially when the data is
sparse. in this paper, we provide a novel densification scheme which relies on carefully tailored
2-universal hashes. we show that the proposed
scheme is variance-optimal, and without losing
the runtime efficiency, it is significantly more accurate than existing densification techniques. as
a result, we obtain a significantly efficient hashing scheme which has the same variance and
collision probability as minwise hashing. experimental evaluations on real sparse and highdimensional datasets validate our claims. we believe that given the significant advantages, our
method will replace minwise hashing implementations in practice.
| 8 |
abstract
suppose that we wish to infer the value of a statistical parameter at a law from which we
sample independent observations. suppose that this parameter is smooth and that we can
define two variation-independent, infinite-dimensional features of the law, its so called q- and
g-components (comp.), such that if we estimate them consistently at a fast enough product of
rates, then we can build a confidence interval (ci) with a given asymptotic level based on a
plain targeted minimum loss estimator (tmle). the estimators of the q- and g-comp. would
typically be by products of machine learning algorithms. we focus on the case that the machine
learning algorithm for the g-comp. is fine-tuned by a real-valued parameter h. then, a plain
tmle with an h chosen by cross-validation would typically not lend itself to the construction
of a ci, because the selection of h would trade-off its empirical bias with something akin to
the empirical variance of the estimator of the g-comp. as opposed to that of the tmle. a
collaborative tmle (c-tmle) might, however, succeed in achieving the relevant trade-off. we
prove that this is the case indeed.
we construct a c-tmle and show that, under high-level empirical processes conditions,
and if there exists an oracle h that makes a bulky remainder term asymptotically gaussian,
then the c-tmle is asymptotically gaussian hence amenable to building a ci provided that
its asymptotic variance can be estimated too. the construction hinges on guaranteeing that
an additional, well chosen estimating equation is solved on top of the estimating equation that
a plain tmle solves. the optimal h is chosen by cross-validating an empirical criterion that
guarantees the wished trade-off between empirical bias and variance.
we illustrate the construction and main result with the inference of the so called average
treatment effect, where the q-comp. consists in a marginal law and a conditional expectation,
and the g-comp. is a propensity score (a conditional probability). we also conduct a multifaceted simulation study to investigate the empirical properties of the collaborative tmle when
the g-comp. is estimated by the lasso. here, h is the bound on the `1 -norm of the candidate
coefficients. the variety of scenarios shed light on small and moderate sample properties, in the
face of low-, moderate- or high-dimensional baseline covariates, and possibly positivity violation.
| 10 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.