|
,text,id,label |
|
0,[[ Recognition of proper nouns ]] in Japanese text has been studied as a part of the more general problem of << morphological analysis >> in Japanese text processing -LRB- -LSB- 1 -RSB- -LSB- 2 -RSB- -RRB- .,0,4 |
|
1,Recognition of [[ proper nouns ]] in << Japanese text >> has been studied as a part of the more general problem of morphological analysis in Japanese text processing -LRB- -LSB- 1 -RSB- -LSB- 2 -RSB- -RRB- .,1,4 |
|
2,Recognition of proper nouns in Japanese text has been studied as a part of the more general problem of [[ morphological analysis ]] in << Japanese text processing >> -LRB- -LSB- 1 -RSB- -LSB- 2 -RSB- -RRB- .,2,3 |
|
3,<< It >> has also been studied in the framework of [[ Japanese information extraction ]] -LRB- -LSB- 3 -RSB- -RRB- in recent years .,3,3 |
|
4,Our [[ approach ]] to the << Multi-lingual Evaluation Task -LRB- MET -RRB- >> for Japanese text is to consider the given task as a morphological analysis problem in Japanese .,4,3 |
|
5,Our approach to the [[ Multi-lingual Evaluation Task -LRB- MET -RRB- ]] for << Japanese text >> is to consider the given task as a morphological analysis problem in Japanese .,5,3 |
|
6,Our approach to the Multi-lingual Evaluation Task -LRB- MET -RRB- for Japanese text is to consider the given << task >> as a [[ morphological analysis problem ]] in Japanese .,6,3 |
|
7,Our approach to the Multi-lingual Evaluation Task -LRB- MET -RRB- for Japanese text is to consider the given task as a << morphological analysis problem >> in [[ Japanese ]] .,7,3 |
|
8,"Our [[ morphological analyzer ]] has done all the necessary work for the << recognition and classification of proper names , numerical and temporal expressions >> , i.e. Named Entity -LRB- NE -RRB- items in the Japanese text .",8,3 |
|
9,"Our morphological analyzer has done all the necessary work for the recognition and classification of << proper names , numerical and temporal expressions >> , i.e. [[ Named Entity -LRB- NE -RRB- items ]] in the Japanese text .",9,2 |
|
10,"Our morphological analyzer has done all the necessary work for the recognition and classification of proper names , numerical and temporal expressions , i.e. [[ Named Entity -LRB- NE -RRB- items ]] in the << Japanese text >> .",10,4 |
|
11,[[ Amorph ]] recognizes << NE items >> in two stages : dictionary lookup and rule application .,11,3 |
|
12,<< Amorph >> recognizes NE items in two stages : [[ dictionary lookup ]] and rule application .,12,4 |
|
13,Amorph recognizes NE items in two stages : [[ dictionary lookup ]] and << rule application >> .,13,0 |
|
14,<< Amorph >> recognizes NE items in two stages : dictionary lookup and [[ rule application ]] .,14,4 |
|
15,"First , << it >> uses several kinds of [[ dictionaries ]] to segment and tag Japanese character strings .",15,3 |
|
16,"First , it uses several kinds of [[ dictionaries ]] to segment and tag << Japanese character strings >> .",16,3 |
|
17,"Second , based on the information resulting from the [[ dictionary lookup stage ]] , a set of << rules >> is applied to the segmented strings in order to identify NE items .",17,3 |
|
18,"Second , based on the information resulting from the dictionary lookup stage , a set of [[ rules ]] is applied to the segmented strings in order to identify << NE items >> .",18,3 |
|
19,We propose to incorporate a [[ priori geometric constraints ]] in a << 3 |
|
20,We propose to incorporate a priori geometric constraints in a 3 |
|
21,Our << approach >> is based on the [[ iterative deformation of a 3 |
|
22,Our approach is based on the [[ iterative deformation of a 3 |
|
23,We show that combining [[ anisotropic meshing ]] with a << non-quadratic approach >> to regularization enables us to obtain satisfactory reconstruction results using triangulations with few vertices .,23,0 |
|
24,We show that combining [[ anisotropic meshing ]] with a non-quadratic approach to regularization enables us to obtain satisfactory << reconstruction >> results using triangulations with few vertices .,24,3 |
|
25,We show that combining anisotropic meshing with a [[ non-quadratic approach ]] to << regularization >> enables us to obtain satisfactory reconstruction results using triangulations with few vertices .,25,3 |
|
26,We show that combining anisotropic meshing with a [[ non-quadratic approach ]] to regularization enables us to obtain satisfactory << reconstruction >> results using triangulations with few vertices .,26,3 |
|
27,We show that combining anisotropic meshing with a non-quadratic approach to regularization enables us to obtain satisfactory << reconstruction >> results using [[ triangulations ]] with few vertices .,27,3 |
|
28,[[ Structural or numerical constraints ]] can then be added locally to the << reconstruction process >> through a constrained optimization scheme .,28,3 |
|
29,<< Structural or numerical constraints >> can then be added locally to the reconstruction process through a [[ constrained optimization scheme ]] .,29,3 |
|
30,[[ They ]] improve the << reconstruction >> results and enforce their consistency with a priori knowledge about object shape .,30,3 |
|
31,They improve the reconstruction results and enforce their consistency with a << priori knowledge >> about [[ object shape ]] .,31,1 |
|
32,The strong description and modeling properties of differential features make [[ them ]] useful tools that can be efficiently used as constraints for << 3 |
|
33,It is based on a [[ weakly supervised dependency parser ]] that can model << speech syntax >> without relying on any annotated training corpus .,33,3 |
|
34,Labeled data is replaced by a few [[ hand-crafted rules ]] that encode basic << syntactic knowledge >> .,34,3 |
|
35,"[[ Bayesian inference ]] then samples the << rules >> , disambiguating and combining them to create complex tree structures that maximize a discriminative model 's posterior on a target unlabeled corpus .",35,3 |
|
36,"Bayesian inference then samples the rules , disambiguating and combining [[ them ]] to create << complex tree structures >> that maximize a discriminative model 's posterior on a target unlabeled corpus .",36,3 |
|
37,"Bayesian inference then samples the rules , disambiguating and combining them to create [[ complex tree structures ]] that maximize a << discriminative model 's posterior >> on a target unlabeled corpus .",37,3 |
|
38,"Bayesian inference then samples the rules , disambiguating and combining them to create complex tree structures that maximize a << discriminative model 's posterior >> on a target [[ unlabeled corpus ]] .",38,3 |
|
39,This [[ posterior ]] encodes << sparse se-lectional preferences >> between a head word and its dependents .,39,3 |
|
40,"The << model >> is evaluated on [[ English and Czech newspaper texts ]] , and is then validated on French broadcast news transcriptions .",40,6 |
|
41,"The << model >> is evaluated on English and Czech newspaper texts , and is then validated on [[ French broadcast news transcriptions ]] .",41,6 |
|
42,[[ Listen-Communicate-Show -LRB- LCS -RRB- ]] is a new paradigm for << human interaction with data sources >> .,42,3 |
|
43,We integrate a << spoken language understanding system >> with [[ intelligent mobile agents ]] that mediate between users and information sources .,43,4 |
|
44,We have built and will demonstrate an application of this [[ approach ]] called << LCS-Marine >> .,44,3 |
|
45,A [[ domain independent model ]] is proposed for the << automated interpretation of nominal compounds >> in English .,45,3 |
|
46,A domain independent model is proposed for the automated interpretation of << nominal compounds >> in [[ English ]] .,46,1 |
|
47,This [[ model ]] is meant to account for << productive rules of interpretation >> which are inferred from the morpho-syntactic and semantic characteristics of the nominal constituents .,47,3 |
|
48,This model is meant to account for << productive rules of interpretation >> which are inferred from the [[ morpho-syntactic and semantic characteristics ]] of the nominal constituents .,48,3 |
|
49,This model is meant to account for productive rules of interpretation which are inferred from the [[ morpho-syntactic and semantic characteristics ]] of the << nominal constituents >> .,49,1 |
|
50,"In particular , we make extensive use of Pustejovsky 's principles concerning the << predicative information >> associated with [[ nominals ]] .",50,1 |
|
51,We argue that it is necessary to draw a line between [[ generalizable semantic principles ]] and << domain-specific semantic information >> .,51,5 |
|
52,"We explain this distinction and we show how this [[ model ]] may be applied to the << interpretation of compounds >> in real texts , provided that complementary semantic information are retrieved .",52,3 |
|
53,We present a new [[ method ]] for << detecting interest points >> using histogram information .,53,3 |
|
54,We present a new method for << detecting interest points >> using [[ histogram information ]] .,54,3 |
|
55,"Unlike existing << interest point detectors >> , which measure [[ pixel-wise differences in image intensity ]] , our detectors incorporate histogram-based representations , and thus can find image regions that present a distinct distribution in the neighborhood .",55,6 |
|
56,"Unlike existing interest point detectors , which measure pixel-wise differences in image intensity , our << detectors >> incorporate [[ histogram-based representations ]] , and thus can find image regions that present a distinct distribution in the neighborhood .",56,4 |
|
57,"The proposed [[ detectors ]] are able to capture << large-scale structures >> and distinctive textured patterns , and exhibit strong invariance to rotation , illumination variation , and blur .",57,3 |
|
58,"The proposed [[ detectors ]] are able to capture large-scale structures and << distinctive textured patterns >> , and exhibit strong invariance to rotation , illumination variation , and blur .",58,3 |
|
59,"The proposed [[ detectors ]] are able to capture large-scale structures and distinctive textured patterns , and exhibit strong invariance to << rotation >> , illumination variation , and blur .",59,3 |
|
60,"The proposed [[ detectors ]] are able to capture large-scale structures and distinctive textured patterns , and exhibit strong invariance to rotation , << illumination variation >> , and blur .",60,3 |
|
61,"The proposed [[ detectors ]] are able to capture large-scale structures and distinctive textured patterns , and exhibit strong invariance to rotation , illumination variation , and << blur >> .",61,3 |
|
62,"The proposed detectors are able to capture [[ large-scale structures ]] and << distinctive textured patterns >> , and exhibit strong invariance to rotation , illumination variation , and blur .",62,0 |
|
63,"The proposed detectors are able to capture large-scale structures and distinctive textured patterns , and exhibit strong invariance to [[ rotation ]] , << illumination variation >> , and blur .",63,0 |
|
64,"The proposed detectors are able to capture large-scale structures and distinctive textured patterns , and exhibit strong invariance to rotation , [[ illumination variation ]] , and << blur >> .",64,0 |
|
65,"The experimental results show that the proposed [[ histogram-based interest point detectors ]] perform particularly well for the tasks of << matching textured scenes >> under blur and illumination changes , in terms of repeatability and distinctiveness .",65,3 |
|
66,"The experimental results show that the proposed << histogram-based interest point detectors >> perform particularly well for the tasks of matching textured scenes under blur and illumination changes , in terms of [[ repeatability ]] and distinctiveness .",66,6 |
|
67,"The experimental results show that the proposed histogram-based interest point detectors perform particularly well for the tasks of matching textured scenes under blur and illumination changes , in terms of [[ repeatability ]] and << distinctiveness >> .",67,0 |
|
68,"The experimental results show that the proposed << histogram-based interest point detectors >> perform particularly well for the tasks of matching textured scenes under blur and illumination changes , in terms of repeatability and [[ distinctiveness ]] .",68,6 |
|
69,An extension of our [[ method ]] to << space-time interest point detection >> for action classification is also presented .,69,3 |
|
70,An extension of our method to [[ space-time interest point detection ]] for << action classification >> is also presented .,70,3 |
|
71,We have implemented a << restricted domain parser >> called [[ Plume ]] .,71,2 |
|
72,"Building on previous work at Carnegie-Mellon University e.g. -LSB- 4 , 5 , 8 -RSB- , [[ Plume 's approach ]] to << parsing >> is based on semantic caseframe instantiation .",72,3 |
|
73,"Building on previous work at Carnegie-Mellon University e.g. -LSB- 4 , 5 , 8 -RSB- , << Plume 's approach >> to parsing is based on [[ semantic caseframe instantiation ]] .",73,3 |
|
74,"This has the advantages of efficiency on grammatical input , and << robustness >> in the face of [[ ungrammatical input ]] .",74,1 |
|
75,"While [[ Plume ]] is well adapted to simple << declarative and imperative utterances >> , it handles passives , relative clauses and interrogatives in an ad hoc manner leading to patchy syntactic coverage .",75,3 |
|
76,"While Plume is well adapted to simple declarative and imperative utterances , [[ it ]] handles << passives >> , relative clauses and interrogatives in an ad hoc manner leading to patchy syntactic coverage .",76,3 |
|
77,"While Plume is well adapted to simple declarative and imperative utterances , [[ it ]] handles passives , << relative clauses >> and interrogatives in an ad hoc manner leading to patchy syntactic coverage .",77,3 |
|
78,"While Plume is well adapted to simple declarative and imperative utterances , [[ it ]] handles passives , relative clauses and << interrogatives >> in an ad hoc manner leading to patchy syntactic coverage .",78,3 |
|
79,"While Plume is well adapted to simple declarative and imperative utterances , it handles [[ passives ]] , << relative clauses >> and interrogatives in an ad hoc manner leading to patchy syntactic coverage .",79,0 |
|
80,"While Plume is well adapted to simple declarative and imperative utterances , it handles passives , [[ relative clauses ]] and << interrogatives >> in an ad hoc manner leading to patchy syntactic coverage .",80,0 |
|
81,"This paper outlines Plume as it currently exists and describes our detailed design for extending [[ Plume ]] to handle << passives >> , relative clauses , and interrogatives in a general manner .",81,3 |
|
82,"This paper outlines Plume as it currently exists and describes our detailed design for extending [[ Plume ]] to handle passives , << relative clauses >> , and interrogatives in a general manner .",82,3 |
|
83,"This paper outlines Plume as it currently exists and describes our detailed design for extending [[ Plume ]] to handle passives , relative clauses , and << interrogatives >> in a general manner .",83,3 |
|
84,"This paper outlines Plume as it currently exists and describes our detailed design for extending Plume to handle [[ passives ]] , << relative clauses >> , and interrogatives in a general manner .",84,0 |
|
85,"This paper outlines Plume as it currently exists and describes our detailed design for extending Plume to handle passives , [[ relative clauses ]] , and << interrogatives >> in a general manner .",85,0 |
|
86,"In this paper , we present an [[ unlexicalized parser ]] for << German >> which employs smoothing and suffix analysis to achieve a labelled bracket F-score of 76.2 , higher than previously reported results on the NEGRA corpus .",86,3 |
|
87,"In this paper , we present an << unlexicalized parser >> for German which employs [[ smoothing ]] and suffix analysis to achieve a labelled bracket F-score of 76.2 , higher than previously reported results on the NEGRA corpus .",87,3 |
|
88,"In this paper , we present an unlexicalized parser for German which employs [[ smoothing ]] and << suffix analysis >> to achieve a labelled bracket F-score of 76.2 , higher than previously reported results on the NEGRA corpus .",88,0 |
|
89,"In this paper , we present an << unlexicalized parser >> for German which employs smoothing and [[ suffix analysis ]] to achieve a labelled bracket F-score of 76.2 , higher than previously reported results on the NEGRA corpus .",89,3 |
|
90,"In this paper , we present an << unlexicalized parser >> for German which employs smoothing and suffix analysis to achieve a [[ labelled bracket F-score ]] of 76.2 , higher than previously reported results on the NEGRA corpus .",90,6 |
|
91,"In this paper , we present an << unlexicalized parser >> for German which employs smoothing and suffix analysis to achieve a labelled bracket F-score of 76.2 , higher than previously reported results on the [[ NEGRA corpus ]] .",91,6 |
|
92,"In addition to the high [[ accuracy ]] of the << model >> , the use of smoothing in an unlexicalized parser allows us to better examine the interplay between smoothing and parsing results .",92,6 |
|
93,"In addition to the high accuracy of the model , the use of [[ smoothing ]] in an << unlexicalized parser >> allows us to better examine the interplay between smoothing and parsing results .",93,3 |
|
94,This paper presents an [[ unsupervised learning approach ]] to disambiguate various << relations between named entities >> by use of various lexical and syntactic features from the contexts .,94,3 |
|
95,This paper presents an << unsupervised learning approach >> to disambiguate various relations between named entities by use of various [[ lexical and syntactic features ]] from the contexts .,95,3 |
|
96,[[ It ]] works by calculating eigenvectors of an adjacency graph 's Laplacian to recover a << submanifold >> of data from a high dimensionality space and then performing cluster number estimation on the eigenvectors .,96,3 |
|
97,<< It >> works by calculating [[ eigenvectors ]] of an adjacency graph 's Laplacian to recover a submanifold of data from a high dimensionality space and then performing cluster number estimation on the eigenvectors .,97,3 |
|
98,It works by calculating << eigenvectors >> of an [[ adjacency graph 's Laplacian ]] to recover a submanifold of data from a high dimensionality space and then performing cluster number estimation on the eigenvectors .,98,1 |
|
99,It works by calculating eigenvectors of an adjacency graph 's Laplacian to recover a << submanifold >> of data from a [[ high dimensionality space ]] and then performing cluster number estimation on the eigenvectors .,99,3 |
|
100,<< It >> works by calculating eigenvectors of an adjacency graph 's Laplacian to recover a submanifold of data from a high dimensionality space and then performing [[ cluster number estimation ]] on the eigenvectors .,100,3 |
|
101,It works by calculating eigenvectors of an adjacency graph 's Laplacian to recover a submanifold of data from a high dimensionality space and then performing [[ cluster number estimation ]] on the << eigenvectors >> .,101,3 |
|
102,Experiment results on [[ ACE corpora ]] show that this << spectral clustering based approach >> outperforms the other clustering methods .,102,6 |
|
103,Experiment results on [[ ACE corpora ]] show that this spectral clustering based approach outperforms the other << clustering methods >> .,103,6 |
|
104,Experiment results on ACE corpora show that this [[ spectral clustering based approach ]] outperforms the other << clustering methods >> .,104,5 |
|
105,"This paper proposes a generic mathematical formalism for the combination of various << structures >> : [[ strings ]] , trees , dags , graphs , and products of them .",105,2 |
|
106,"This paper proposes a generic mathematical formalism for the combination of various structures : [[ strings ]] , << trees >> , dags , graphs , and products of them .",106,0 |
|
107,"This paper proposes a generic mathematical formalism for the combination of various << structures >> : strings , [[ trees ]] , dags , graphs , and products of them .",107,2 |
|
108,"This paper proposes a generic mathematical formalism for the combination of various structures : strings , [[ trees ]] , << dags >> , graphs , and products of them .",108,0 |
|
109,"This paper proposes a generic mathematical formalism for the combination of various << structures >> : strings , trees , [[ dags ]] , graphs , and products of them .",109,2 |
|
110,"This paper proposes a generic mathematical formalism for the combination of various structures : strings , trees , [[ dags ]] , << graphs >> , and products of them .",110,0 |
|
111,"This paper proposes a generic mathematical formalism for the combination of various << structures >> : strings , trees , dags , [[ graphs ]] , and products of them .",111,2 |
|
112,"This [[ formalism ]] is both elementary and powerful enough to strongly simulate many << grammar formalisms >> , such as rewriting systems , dependency grammars , TAG , HPSG and LFG .",112,3 |
|
113,"This formalism is both elementary and powerful enough to strongly simulate many << grammar formalisms >> , such as [[ rewriting systems ]] , dependency grammars , TAG , HPSG and LFG .",113,2 |
|
114,"This formalism is both elementary and powerful enough to strongly simulate many grammar formalisms , such as [[ rewriting systems ]] , << dependency grammars >> , TAG , HPSG and LFG .",114,0 |
|
115,"This formalism is both elementary and powerful enough to strongly simulate many << grammar formalisms >> , such as rewriting systems , [[ dependency grammars ]] , TAG , HPSG and LFG .",115,2 |
|
116,"This formalism is both elementary and powerful enough to strongly simulate many grammar formalisms , such as rewriting systems , [[ dependency grammars ]] , << TAG >> , HPSG and LFG .",116,0 |
|
117,"This formalism is both elementary and powerful enough to strongly simulate many << grammar formalisms >> , such as rewriting systems , dependency grammars , [[ TAG ]] , HPSG and LFG .",117,2 |
|
118,"This formalism is both elementary and powerful enough to strongly simulate many grammar formalisms , such as rewriting systems , dependency grammars , [[ TAG ]] , << HPSG >> and LFG .",118,0 |
|
119,"This formalism is both elementary and powerful enough to strongly simulate many << grammar formalisms >> , such as rewriting systems , dependency grammars , TAG , [[ HPSG ]] and LFG .",119,2 |
|
120,"This formalism is both elementary and powerful enough to strongly simulate many grammar formalisms , such as rewriting systems , dependency grammars , TAG , [[ HPSG ]] and << LFG >> .",120,0 |
|
121,"This formalism is both elementary and powerful enough to strongly simulate many << grammar formalisms >> , such as rewriting systems , dependency grammars , TAG , HPSG and [[ LFG ]] .",121,2 |
|
122,"A [[ mixed-signal paradigm ]] is presented for << high-resolution parallel inner-product computation >> in very high dimensions , suitable for efficient implementation of kernels in image processing .",122,3 |
|
123,"A [[ mixed-signal paradigm ]] is presented for high-resolution parallel inner-product computation in very high dimensions , suitable for efficient implementation of << kernels >> in image processing .",123,3 |
|
124,"A mixed-signal paradigm is presented for high-resolution parallel inner-product computation in very high dimensions , suitable for efficient implementation of [[ kernels ]] in << image processing >> .",124,3 |
|
125,"At the core of the << externally digital architecture >> is a [[ high-density , low-power analog array ]] performing binary-binary partial matrix-vector multiplication .",125,4 |
|
126,"At the core of the externally digital architecture is a << high-density , low-power analog array >> performing [[ binary-binary partial matrix-vector multiplication ]] .",126,3 |
|
127,"Full digital resolution is maintained even with low-resolution analog-to-digital conversion , owing to [[ random statistics ]] in the << analog summation of binary products >> .",127,4 |
|
128,A [[ random modulation scheme ]] produces << near-Bernoulli statistics >> even for highly correlated inputs .,128,3 |
|
129,A << random modulation scheme >> produces near-Bernoulli statistics even for [[ highly correlated inputs ]] .,129,3 |
|
130,"The << approach >> is validated with [[ real image data ]] , and with experimental results from a CID/DRAM analog array prototype in 0.5 cents m CMOS .",130,6 |
|
131,"In this paper we specialize the [[ projective unifocal , bifo-cal , and trifocal tensors ]] to the << affine case >> , and show how the tensors obtained relate to the registered tensors encountered in previous work .",131,3 |
|
132,"Finally , we show how the << estimation of the tensors >> from [[ point correspondences ]] is achieved through factorization , and discuss the estimation from line correspondences .",132,3 |
|
133,"Finally , we show how the estimation of the << tensors >> from point correspondences is achieved through [[ factorization ]] , and discuss the estimation from line correspondences .",133,3 |
|
134,"Finally , we show how the estimation of the tensors from point correspondences is achieved through factorization , and discuss the << estimation >> from [[ line correspondences ]] .",134,3 |
|
135,"We propose a [[ corpus-based method ]] -LRB- Biber ,1993 ; Nagao ,1993 ; Smadja ,1993 -RRB- which generates << Noun Classifier Associations -LRB- NCA -RRB- >> to overcome the problems in classifier assignment and semantic construction of noun phrase .",135,3 |
|
136,"We propose a [[ corpus-based method ]] -LRB- Biber ,1993 ; Nagao ,1993 ; Smadja ,1993 -RRB- which generates Noun Classifier Associations -LRB- NCA -RRB- to overcome the problems in << classifier assignment >> and semantic construction of noun phrase .",136,3 |
|
137,"We propose a [[ corpus-based method ]] -LRB- Biber ,1993 ; Nagao ,1993 ; Smadja ,1993 -RRB- which generates Noun Classifier Associations -LRB- NCA -RRB- to overcome the problems in classifier assignment and << semantic construction of noun phrase >> .",137,3 |
|
138,"We propose a corpus-based method -LRB- Biber ,1993 ; Nagao ,1993 ; Smadja ,1993 -RRB- which generates [[ Noun Classifier Associations -LRB- NCA -RRB- ]] to overcome the problems in << classifier assignment >> and semantic construction of noun phrase .",138,3 |
|
139,"We propose a corpus-based method -LRB- Biber ,1993 ; Nagao ,1993 ; Smadja ,1993 -RRB- which generates [[ Noun Classifier Associations -LRB- NCA -RRB- ]] to overcome the problems in classifier assignment and << semantic construction of noun phrase >> .",139,3 |
|
140,"We propose a corpus-based method -LRB- Biber ,1993 ; Nagao ,1993 ; Smadja ,1993 -RRB- which generates Noun Classifier Associations -LRB- NCA -RRB- to overcome the problems in [[ classifier assignment ]] and << semantic construction of noun phrase >> .",140,0 |
|
141,The << NCA >> is created statistically from a large corpus and recomposed under [[ concept hierarchy constraints ]] and frequency of occurrences .,141,3 |
|
142,The << NCA >> is created statistically from a large corpus and recomposed under concept hierarchy constraints and [[ frequency of occurrences ]] .,142,3 |
|
143,The << perception of transparent objects >> from [[ images ]] is known to be a very hard problem in vision .,143,3 |
|
144,We show how << features >> that are imaged through a transparent object behave differently from [[ those ]] that are rigidly attached to the scene .,144,5 |
|
145,We present a novel [[ model-based approach ]] to recover the << shapes and the poses of transparent objects >> from known motion .,145,3 |
|
146,We present a novel model-based approach to recover the << shapes and the poses of transparent objects >> from [[ known motion ]] .,146,3 |
|
147,The objects can be complex in that << they >> may be composed of [[ multiple layers ]] with different refractive indices .,147,4 |
|
148,The objects can be complex in that they may be composed of << multiple layers >> with different [[ refractive indices ]] .,148,1 |
|
149,We have applied [[ it ]] to << real scenes >> that include transparent objects and recovered the shapes of the objects with high accuracy .,149,3 |
|
150,We have applied [[ it ]] to real scenes that include transparent objects and recovered the << shapes of the objects >> with high accuracy .,150,3 |
|
151,We have applied it to << real scenes >> that include [[ transparent objects ]] and recovered the shapes of the objects with high accuracy .,151,4 |
|
152,We have applied it to real scenes that include transparent objects and recovered the << shapes of the objects >> with high [[ accuracy ]] .,152,6 |
|
153,We propose a novel [[ probabilistic framework ]] for learning << visual models of 3D object categories >> by combining appearance information and geometric constraints .,153,3 |
|
154,We propose a novel << probabilistic framework >> for learning visual models of 3D object categories by combining [[ appearance information ]] and geometric constraints .,154,3 |
|
155,We propose a novel probabilistic framework for learning visual models of 3D object categories by combining [[ appearance information ]] and << geometric constraints >> .,155,0 |
|
156,We propose a novel << probabilistic framework >> for learning visual models of 3D object categories by combining appearance information and [[ geometric constraints ]] .,156,3 |
|
157,A [[ generative framework ]] is used for learning a << model >> that captures the relative position of parts within each of the discretized viewpoints .,157,3 |
|
158,"Contrary to most of the existing << mixture of viewpoints models >> , our [[ model ]] establishes explicit correspondences of parts across different viewpoints of the object class .",158,5 |
|
159,"Given a new [[ image ]] , << detection >> and classification are achieved by determining the position and viewpoint of the model that maximize recognition scores of the candidate objects .",159,3 |
|
160,"Given a new [[ image ]] , detection and << classification >> are achieved by determining the position and viewpoint of the model that maximize recognition scores of the candidate objects .",160,3 |
|
161,"Given a new image , [[ detection ]] and << classification >> are achieved by determining the position and viewpoint of the model that maximize recognition scores of the candidate objects .",161,0 |
|
162,"Given a new image , << detection >> and classification are achieved by determining the [[ position ]] and viewpoint of the model that maximize recognition scores of the candidate objects .",162,3 |
|
163,"Given a new image , detection and << classification >> are achieved by determining the [[ position ]] and viewpoint of the model that maximize recognition scores of the candidate objects .",163,3 |
|
164,"Given a new image , detection and classification are achieved by determining the [[ position ]] and << viewpoint >> of the model that maximize recognition scores of the candidate objects .",164,0 |
|
165,"Given a new image , << detection >> and classification are achieved by determining the position and [[ viewpoint ]] of the model that maximize recognition scores of the candidate objects .",165,3 |
|
166,"Given a new image , detection and << classification >> are achieved by determining the position and [[ viewpoint ]] of the model that maximize recognition scores of the candidate objects .",166,3 |
|
167,Our approach is among the first to propose a [[ generative proba-bilistic framework ]] for << 3D object categorization >> .,167,3 |
|
168,We test our [[ algorithm ]] on the << detection task >> and the viewpoint classification task by using '' car '' category from both the Savarese et al. 2007 and PASCAL VOC 2006 datasets .,168,3 |
|
169,We test our [[ algorithm ]] on the detection task and the << viewpoint classification task >> by using '' car '' category from both the Savarese et al. 2007 and PASCAL VOC 2006 datasets .,169,3 |
|
170,We test our algorithm on the [[ detection task ]] and the << viewpoint classification task >> by using '' car '' category from both the Savarese et al. 2007 and PASCAL VOC 2006 datasets .,170,0 |
|
171,We test our << algorithm >> on the detection task and the viewpoint classification task by using '' car '' category from both the Savarese et al. 2007 and [[ PASCAL VOC 2006 datasets ]] .,171,6 |
|
172,We show promising results in both the << detection and viewpoint classification tasks >> on these two challenging [[ datasets ]] .,172,6 |
|
173,We present an application of [[ ambiguity packing and stochastic disambiguation techniques ]] for << Lexical-Functional Grammars -LRB- LFG -RRB- >> to the domain of sentence condensation .,173,3 |
|
174,We present an application of [[ ambiguity packing and stochastic disambiguation techniques ]] for Lexical-Functional Grammars -LRB- LFG -RRB- to the domain of << sentence condensation >> .,174,3 |
|
175,"Our << system >> incorporates a [[ linguistic parser/generator ]] for LFG , a transfer component for parse reduction operating on packed parse forests , and a maximum-entropy model for stochastic output selection .",175,4 |
|
176,"Our system incorporates a [[ linguistic parser/generator ]] for << LFG >> , a transfer component for parse reduction operating on packed parse forests , and a maximum-entropy model for stochastic output selection .",176,3 |
|
177,"Our system incorporates a [[ linguistic parser/generator ]] for LFG , a << transfer component >> for parse reduction operating on packed parse forests , and a maximum-entropy model for stochastic output selection .",177,0 |
|
178,"Our << system >> incorporates a linguistic parser/generator for LFG , a [[ transfer component ]] for parse reduction operating on packed parse forests , and a maximum-entropy model for stochastic output selection .",178,4 |
|
179,"Our system incorporates a linguistic parser/generator for LFG , a [[ transfer component ]] for << parse reduction >> operating on packed parse forests , and a maximum-entropy model for stochastic output selection .",179,3 |
|
180,"Our system incorporates a linguistic parser/generator for LFG , a [[ transfer component ]] for parse reduction operating on packed parse forests , and a << maximum-entropy model >> for stochastic output selection .",180,0 |
|
181,"Our system incorporates a linguistic parser/generator for LFG , a transfer component for << parse reduction >> operating on [[ packed parse forests ]] , and a maximum-entropy model for stochastic output selection .",181,3 |
|
182,"Our << system >> incorporates a linguistic parser/generator for LFG , a transfer component for parse reduction operating on packed parse forests , and a [[ maximum-entropy model ]] for stochastic output selection .",182,4 |
|
183,"Our system incorporates a linguistic parser/generator for LFG , a transfer component for parse reduction operating on packed parse forests , and a [[ maximum-entropy model ]] for << stochastic output selection >> .",183,3 |
|
184,"Furthermore , we propose the use of standard [[ parser evaluation methods ]] for automatically evaluating the << summarization quality >> of sentence condensation systems .",184,6 |
|
185,"Furthermore , we propose the use of standard parser evaluation methods for automatically evaluating the [[ summarization quality ]] of << sentence condensation systems >> .",185,6 |
|
186,An experimental evaluation of [[ summarization quality ]] shows a close correlation between the << automatic parse-based evaluation >> and a manual evaluation of generated strings .,186,6 |
|
187,An experimental evaluation of summarization quality shows a close correlation between the [[ automatic parse-based evaluation ]] and a << manual evaluation >> of generated strings .,187,5 |
|
188,"Overall [[ summarization quality ]] of the proposed << system >> is state-of-the-art , with guaranteed grammaticality of the system output due to the use of a constraint-based parser/generator .",188,6 |
|
189,"Overall summarization quality of the proposed << system >> is state-of-the-art , with guaranteed [[ grammaticality ]] of the system output due to the use of a constraint-based parser/generator .",189,6 |
|
190,"Overall summarization quality of the proposed << system >> is state-of-the-art , with guaranteed grammaticality of the system output due to the use of a [[ constraint-based parser/generator ]] .",190,3 |
|
191,"The [[ robust principal component analysis -LRB- robust PCA -RRB- problem ]] has been considered in many << machine learning applications >> , where the goal is to decompose the data matrix to a low rank part plus a sparse residual .",191,3 |
|
192,"The robust principal component analysis -LRB- robust PCA -RRB- problem has been considered in many machine learning applications , where the goal is to decompose the << data matrix >> to a [[ low rank part ]] plus a sparse residual .",192,4 |
|
193,"The robust principal component analysis -LRB- robust PCA -RRB- problem has been considered in many machine learning applications , where the goal is to decompose the data matrix to a [[ low rank part ]] plus a << sparse residual >> .",193,0 |
|
194,"The robust principal component analysis -LRB- robust PCA -RRB- problem has been considered in many machine learning applications , where the goal is to decompose the << data matrix >> to a low rank part plus a [[ sparse residual ]] .",194,4 |
|
195,"While current << approaches >> are developed by only considering the [[ low rank plus sparse structure ]] , in many applications , side information of row and/or column entities may also be given , and it is still unclear to what extent could such information help robust PCA .",195,3 |
|
196,"While current approaches are developed by only considering the low rank plus sparse structure , in many applications , side information of row and/or column entities may also be given , and it is still unclear to what extent could such [[ information ]] help << robust PCA >> .",196,3 |
|
197,"Thus , in this paper , we study the problem of << robust PCA >> with [[ side information ]] , where both prior structure and features of entities are exploited for recovery .",197,3 |
|
198,"Thus , in this paper , we study the problem of robust PCA with side information , where both [[ prior structure ]] and << features of entities >> are exploited for recovery .",198,0 |
|
199,"Thus , in this paper , we study the problem of robust PCA with side information , where both [[ prior structure ]] and features of entities are exploited for << recovery >> .",199,3 |
|
200,"Thus , in this paper , we study the problem of robust PCA with side information , where both prior structure and [[ features of entities ]] are exploited for << recovery >> .",200,3 |
|
201,We propose a [[ convex problem ]] to incorporate << side information >> in robust PCA and show that the low rank matrix can be exactly recovered via the proposed method under certain conditions .,201,3 |
|
202,We propose a convex problem to incorporate [[ side information ]] in << robust PCA >> and show that the low rank matrix can be exactly recovered via the proposed method under certain conditions .,202,4 |
|
203,We propose a convex problem to incorporate side information in robust PCA and show that the << low rank matrix >> can be exactly recovered via the proposed [[ method ]] under certain conditions .,203,3 |
|
204,"In particular , our guarantee suggests that a substantial amount of << low rank matrices >> , which can not be recovered by standard robust PCA , become re-coverable by our proposed [[ method ]] .",204,3 |
|
205,The result theoretically justifies the effectiveness of [[ features ]] in << robust PCA >> .,205,1 |
|
206,"In addition , we conduct synthetic experiments as well as a real application on [[ noisy image classification ]] to show that our << method >> also improves the performance in practice by exploiting side information .",206,6 |
|
207,"In addition , we conduct synthetic experiments as well as a real application on noisy image classification to show that our << method >> also improves the performance in practice by exploiting [[ side information ]] .",207,3 |
|
208,This paper presents necessary and sufficient conditions for the use of [[ demonstrative expressions ]] in << English >> and discusses implications for current discourse processing algorithms .,208,1 |
|
209,This paper presents necessary and sufficient conditions for the use of demonstrative expressions in English and discusses [[ implications ]] for current << discourse processing algorithms >> .,209,3 |
|
210,"This research is part of a larger study of [[ anaphoric expressions ]] , the results of which will be incorporated into a << natural language generation system >> .",210,3 |
|
211,"Using the [[ IEMOCAP database ]] , << discrete -LRB- categorical -RRB- and continuous -LRB- attribute -RRB- emotional assessments >> evaluated by the actors and na ¨ ıve listeners are compared .",211,3 |
|
212,The problem of << blind separation of underdetermined instantaneous mixtures of independent signals >> is addressed through a [[ method ]] relying on nonstationarity of the original signals .,212,3 |
|
213,The problem of blind separation of underdetermined instantaneous mixtures of independent signals is addressed through a << method >> relying on [[ nonstationarity ]] of the original signals .,213,3 |
|
214,"In comparison with previous works , in this paper it is assumed that the << signals >> are not i.i.d. in each epoch , but obey a [[ first-order autoregressive model ]] .",214,3 |
|
215,This [[ model ]] was shown to be more appropriate for << blind separation of natural speech signals . >>,215,3 |
|
216,"A << separation method >> is proposed that is nearly statistically efficient -LRB- approaching the corresponding [[ Cramér-Rao lower bound -RRB- ]] , if the separated signals obey the assumed model .",216,1 |
|
217,"In the case of << natural speech signals >> , the [[ method ]] is shown to have separation accuracy better than the state-of-the-art methods .",217,3 |
|
218,"In the case of natural speech signals , the [[ method ]] is shown to have separation accuracy better than the state-of-the-art << methods >> .",218,5 |
|
219,"In the case of natural speech signals , the << method >> is shown to have [[ separation accuracy ]] better than the state-of-the-art methods .",219,6 |
|
220,"In the case of natural speech signals , the method is shown to have [[ separation accuracy ]] better than the state-of-the-art << methods >> .",220,6 |
|
221,"In the case of << natural speech signals >> , the method is shown to have separation accuracy better than the state-of-the-art [[ methods ]] .",221,3 |
|
222,This paper proposes to use a [[ convolution kernel over parse trees ]] to model << syntactic structure information >> for relation extraction .,222,3 |
|
223,This paper proposes to use a convolution kernel over parse trees to model [[ syntactic structure information ]] for << relation extraction >> .,223,3 |
|
224,Our study reveals that the [[ syntactic structure features ]] embedded in a << parse tree >> are very effective for relation extraction and these features can be well captured by the convolution tree kernel .,224,1 |
|
225,Our study reveals that the [[ syntactic structure features ]] embedded in a parse tree are very effective for << relation extraction >> and these features can be well captured by the convolution tree kernel .,225,3 |
|
226,Our study reveals that the syntactic structure features embedded in a parse tree are very effective for relation extraction and these << features >> can be well captured by the [[ convolution tree kernel ]] .,226,3 |
|
227,Evaluation on the [[ ACE 2003 corpus ]] shows that the << convolution kernel over parse trees >> can achieve comparable performance with the previous best-reported feature-based methods on the 24 ACE relation subtypes .,227,6 |
|
228,Evaluation on the ACE 2003 corpus shows that the << convolution kernel over parse trees >> can achieve comparable performance with the previous best-reported [[ feature-based methods ]] on the 24 ACE relation subtypes .,228,5 |
|
229,It also shows that our [[ method ]] significantly outperforms the previous two << dependency tree kernels >> on the 5 ACE relation major types .,229,5 |
|
230,This paper presents the results of automatically inducing a [[ Combinatory Categorial Grammar -LRB- CCG -RRB- lexicon ]] from a << Turkish dependency treebank >> .,230,4 |
|
231,The fact that [[ Turkish ]] is an << agglutinating free word order language >> presents a challenge for language theories .,231,2 |
|
232,"We explored possible ways to obtain a [[ compact lexicon ]] , consistent with CCG principles , from a << treebank >> which is an order of magnitude smaller than Penn WSJ .",232,4 |
|
233,"We explored possible ways to obtain a compact lexicon , consistent with CCG principles , from a [[ treebank ]] which is an order of magnitude smaller than << Penn WSJ >> .",233,5 |
|
234,"While [[ sentence extraction ]] as an approach to << summarization >> has been shown to work in documents of certain genres , because of the conversational nature of email communication where utterances are made in relation to one made previously , sentence extraction may not capture the necessary segments of dialogue that would make a summary coherent .",234,3 |
|
235,"In this paper , we present our work on the [[ detection of question-answer pairs ]] in an email conversation for the task of << email summarization >> .",235,3 |
|
236,"In this paper , we present our work on the << detection of question-answer pairs >> in an [[ email conversation ]] for the task of email summarization .",236,3 |
|
237,We show that various [[ features ]] based on the structure of email-threads can be used to improve upon << lexical similarity >> of discourse segments for question-answer pairing .,237,3 |
|
238,We show that various [[ features ]] based on the structure of email-threads can be used to improve upon lexical similarity of discourse segments for << question-answer pairing >> .,238,3 |
|
239,We show that various << features >> based on the [[ structure of email-threads ]] can be used to improve upon lexical similarity of discourse segments for question-answer pairing .,239,3 |
|
240,We show that various features based on the structure of email-threads can be used to improve upon [[ lexical similarity ]] of << discourse segments >> for question-answer pairing .,240,1 |
|
241,"Specifically , we show how to incorporate a simple [[ prior on the distribution of natural images ]] into << support vector machines >> .",241,3 |
|
242,"[[ SVMs ]] are known to be robust to << overfitting >> ; however , a few training examples usually do not represent well the structure of the class .",242,3 |
|
243,"Our experiments on [[ real data sets ]] show that the resulting << detector >> is more robust to the choice of training examples , and substantially improves both linear and kernel SVM when trained on 10 positive and 10 negative examples .",243,6 |
|
244,"Our experiments on real data sets show that the resulting [[ detector ]] is more robust to the choice of training examples , and substantially improves both << linear and kernel SVM >> when trained on 10 positive and 10 negative examples .",244,5 |
|
245,"Although the study of clustering is centered around an intuitively compelling goal , it has been very difficult to develop a [[ unified framework ]] for << reasoning >> about it at a technical level , and profoundly diverse approaches to clustering abound in the research community .",245,3 |
|
246,"Relaxations of these properties expose some of the interesting -LRB- and unavoidable -RRB- trade-offs at work in << well-studied clustering techniques >> such as [[ single-linkage ]] , sum-of-pairs , k-means , and k-median .",246,2 |
|
247,"Relaxations of these properties expose some of the interesting -LRB- and unavoidable -RRB- trade-offs at work in well-studied clustering techniques such as [[ single-linkage ]] , << sum-of-pairs >> , k-means , and k-median .",247,0 |
|
248,"Relaxations of these properties expose some of the interesting -LRB- and unavoidable -RRB- trade-offs at work in << well-studied clustering techniques >> such as single-linkage , [[ sum-of-pairs ]] , k-means , and k-median .",248,2 |
|
249,"Relaxations of these properties expose some of the interesting -LRB- and unavoidable -RRB- trade-offs at work in well-studied clustering techniques such as single-linkage , [[ sum-of-pairs ]] , << k-means >> , and k-median .",249,0 |
|
250,"Relaxations of these properties expose some of the interesting -LRB- and unavoidable -RRB- trade-offs at work in << well-studied clustering techniques >> such as single-linkage , sum-of-pairs , [[ k-means ]] , and k-median .",250,2 |
|
251,"Relaxations of these properties expose some of the interesting -LRB- and unavoidable -RRB- trade-offs at work in well-studied clustering techniques such as single-linkage , sum-of-pairs , [[ k-means ]] , and << k-median >> .",251,0 |
|
252,"Relaxations of these properties expose some of the interesting -LRB- and unavoidable -RRB- trade-offs at work in << well-studied clustering techniques >> such as single-linkage , sum-of-pairs , k-means , and [[ k-median ]] .",252,2 |
|
253,"With << relevant approach >> , we identify important contents by [[ PageRank algorithm ]] on the event map constructed from documents .",253,3 |
|
254,"With relevant approach , we identify important contents by << PageRank algorithm >> on the [[ event map ]] constructed from documents .",254,3 |
|
255,"With relevant approach , we identify important contents by PageRank algorithm on the << event map >> constructed from [[ documents ]] .",255,3 |
|
256,We present a [[ scanning method ]] that recovers << dense sub-pixel camera-projector correspondence >> without requiring any photometric calibration nor preliminary knowledge of their relative geometry .,256,3 |
|
257,<< Subpixel accuracy >> is achieved by considering several [[ zero-crossings ]] defined by the difference between pairs of unstructured patterns .,257,3 |
|
258,We use << gray-level band-pass white noise patterns >> that increase [[ robustness ]] to indirect lighting and scene discontinuities .,258,6 |
|
259,We use gray-level band-pass white noise patterns that increase << robustness >> to [[ indirect lighting ]] and scene discontinuities .,259,1 |
|
260,We use gray-level band-pass white noise patterns that increase robustness to [[ indirect lighting ]] and << scene discontinuities >> .,260,0 |
|
261,We use gray-level band-pass white noise patterns that increase << robustness >> to indirect lighting and [[ scene discontinuities ]] .,261,1 |
|
262,"Simulated and experimental results show that our [[ method ]] recovers << scene geometry >> with high subpixel precision , and that it can handle many challenges of active reconstruction systems .",262,3 |
|
263,"Simulated and experimental results show that our method recovers << scene geometry >> with high [[ subpixel precision ]] , and that it can handle many challenges of active reconstruction systems .",263,1 |
|
264,"Simulated and experimental results show that our method recovers scene geometry with high subpixel precision , and that [[ it ]] can handle many challenges of << active reconstruction systems >> .",264,3 |
|
265,We compare our results to << state of the art methods >> such as [[ mi-cro phase shifting ]] and modulated phase shifting .,265,2 |
|
266,We compare our results to state of the art methods such as [[ mi-cro phase shifting ]] and << modulated phase shifting >> .,266,0 |
|
267,We compare our results to << state of the art methods >> such as mi-cro phase shifting and [[ modulated phase shifting ]] .,267,2 |
|
268,This paper describes a novel [[ system ]] for << acquiring adjectival subcategorization frames -LRB- scfs -RRB- >> and associated frequency information from English corpus data .,268,3 |
|
269,The << system >> incorporates a [[ decision-tree classifier ]] for 30 scf types which tests for the presence of grammatical relations -LRB- grs -RRB- in the output of a robust statistical parser .,269,4 |
|
270,The system incorporates a [[ decision-tree classifier ]] for 30 scf types which tests for the presence of << grammatical relations -LRB- grs -RRB- >> in the output of a robust statistical parser .,270,3 |
|
271,<< It >> uses a powerful [[ pattern-matching language ]] to classify grs into frames hierarchically in a way that mirrors inheritance-based lexica .,271,3 |
|
272,It uses a powerful [[ pattern-matching language ]] to classify << grs >> into frames hierarchically in a way that mirrors inheritance-based lexica .,272,3 |
|
273,The experiments show that the << system >> is able to detect scf types with 70 % [[ precision ]] and 66 % recall rate .,273,6 |
|
274,The experiments show that the system is able to detect scf types with 70 % [[ precision ]] and 66 % << recall >> rate .,274,0 |
|
275,The experiments show that the << system >> is able to detect scf types with 70 % precision and 66 % [[ recall ]] rate .,275,6 |
|
276,A new [[ tool ]] for << linguistic annotation of scfs >> in corpus data is also introduced which can considerably alleviate the process of obtaining training and test data for subcategorization acquisition .,276,3 |
|
277,A new tool for linguistic annotation of scfs in corpus data is also introduced which can considerably alleviate the process of obtaining [[ training and test data ]] for << subcategorization acquisition >> .,277,3 |
|
278,[[ Machine transliteration/back-transliteration ]] plays an important role in many << multilingual speech and language applications >> .,278,3 |
|
279,"In this paper , a novel [[ framework ]] for << machine transliteration/backtransliteration >> that allows us to carry out direct orthographical mapping -LRB- DOM -RRB- between two different languages is presented .",279,3 |
|
280,"In this paper , a novel framework for [[ machine transliteration/backtransliteration ]] that allows us to carry out << direct orthographical mapping -LRB- DOM -RRB- >> between two different languages is presented .",280,3 |
|
281,"Under this [[ framework ]] , a << joint source-channel transliteration model >> , also called n-gram transliteration model -LRB- n-gram TM -RRB- , is further proposed to model the transliteration process .",281,3 |
|
282,"Under this framework , a joint source-channel transliteration model , also called [[ n-gram transliteration model -LRB- n-gram TM -RRB- ]] , is further proposed to model the << transliteration process >> .",282,3 |
|
283,We evaluate the proposed << methods >> through several [[ transliteration/backtransliteration ]] experiments for English/Chinese and English/Japanese language pairs .,283,6 |
|
284,We evaluate the proposed methods through several [[ transliteration/backtransliteration ]] experiments for << English/Chinese and English/Japanese language pairs >> .,284,3 |
|
285,Our study reveals that the proposed << method >> not only reduces an extensive system development effort but also improves the [[ transliteration accuracy ]] significantly .,285,6 |
|
286,"A [[ bio-inspired model ]] for an << analog programmable array processor -LRB- APAP -RRB- >> , based on studies on the vertebrate retina , has permitted the realization of complex programmable spatio-temporal dynamics in VLSI .",286,3 |
|
287,"A [[ bio-inspired model ]] for an analog programmable array processor -LRB- APAP -RRB- , based on studies on the vertebrate retina , has permitted the realization of << complex programmable spatio-temporal dynamics >> in VLSI .",287,3 |
|
288,"A << bio-inspired model >> for an analog programmable array processor -LRB- APAP -RRB- , based on studies on the [[ vertebrate retina ]] , has permitted the realization of complex programmable spatio-temporal dynamics in VLSI .",288,3 |
|
289,"A bio-inspired model for an analog programmable array processor -LRB- APAP -RRB- , based on studies on the vertebrate retina , has permitted the realization of [[ complex programmable spatio-temporal dynamics ]] in << VLSI >> .",289,1 |
|
290,"This model mimics the way in which << images >> are processed in the [[ visual pathway ]] , rendering a feasible alternative for the implementation of early vision applications in standard technologies .",290,3 |
|
291,[[ Computing power per area ]] and << power consumption >> is amongst the highest reported for a single chip .,291,0 |
|
292,Another problem with << determiners >> is their inherent [[ ambiguity ]] .,292,1 |
|
293,"In this paper we propose a [[ logical formalism ]] , which , among other things , is suitable for representing << determiners >> without forcing a particular interpretation when their meaning is still not clear .",293,3 |
|
294,"We investigate the [[ verbal and nonverbal means ]] for << grounding >> , and propose a design for embodied conversational agents that relies on both kinds of signals to establish common ground in human-computer interaction .",294,3 |
|
295,"We investigate the verbal and nonverbal means for grounding , and propose a [[ design ]] for << embodied conversational agents >> that relies on both kinds of signals to establish common ground in human-computer interaction .",295,3 |
|
296,"We investigate the verbal and nonverbal means for grounding , and propose a design for embodied conversational agents that relies on both kinds of signals to establish [[ common ground ]] in << human-computer interaction >> .",296,3 |
|
297,"We analyzed [[ eye gaze ]] , << head nods >> and attentional focus in the context of a direction-giving task .",297,0 |
|
298,"We analyzed [[ eye gaze ]] , head nods and attentional focus in the context of a << direction-giving task >> .",298,4 |
|
299,"We analyzed eye gaze , [[ head nods ]] and << attentional focus >> in the context of a direction-giving task .",299,0 |
|
300,"We analyzed eye gaze , [[ head nods ]] and attentional focus in the context of a << direction-giving task >> .",300,4 |
|
301,"We analyzed eye gaze , head nods and [[ attentional focus ]] in the context of a << direction-giving task >> .",301,4 |
|
302,"Based on these results , we present an << ECA >> that uses [[ verbal and nonverbal grounding acts ]] to update dialogue state .",302,3 |
|
303,"Based on these results , we present an ECA that uses [[ verbal and nonverbal grounding acts ]] to update << dialogue state >> .",303,3 |
|
304,"[[ Sentence boundary detection ]] in speech is important for enriching << speech recognition output >> , making it easier for humans to read and downstream modules to process .",304,3 |
|
305,"<< Sentence boundary detection >> in [[ speech ]] is important for enriching speech recognition output , making it easier for humans to read and downstream modules to process .",305,3 |
|
306,"In previous work , we have developed [[ hidden Markov model -LRB- HMM -RRB- and maximum entropy -LRB- Maxent -RRB- classifiers ]] that integrate textual and prosodic knowledge sources for << detecting sentence boundaries >> .",306,3 |
|
307,"In previous work , we have developed << hidden Markov model -LRB- HMM -RRB- and maximum entropy -LRB- Maxent -RRB- classifiers >> that integrate [[ textual and prosodic knowledge sources ]] for detecting sentence boundaries .",307,3 |
|
308,"In this paper , we evaluate the use of a [[ conditional random field -LRB- CRF -RRB- ]] for this << task >> and relate results with this model to our prior work .",308,3 |
|
309,We evaluate across two [[ corpora ]] -LRB- conversational telephone speech and broadcast news speech -RRB- on both << human transcriptions >> and speech recognition output .,309,6 |
|
310,We evaluate across two [[ corpora ]] -LRB- conversational telephone speech and broadcast news speech -RRB- on both human transcriptions and << speech recognition output >> .,310,6 |
|
311,We evaluate across two << corpora >> -LRB- [[ conversational telephone speech ]] and broadcast news speech -RRB- on both human transcriptions and speech recognition output .,311,2 |
|
312,We evaluate across two corpora -LRB- [[ conversational telephone speech ]] and << broadcast news speech >> -RRB- on both human transcriptions and speech recognition output .,312,0 |
|
313,We evaluate across two << corpora >> -LRB- conversational telephone speech and [[ broadcast news speech ]] -RRB- on both human transcriptions and speech recognition output .,313,2 |
|
314,We evaluate across two corpora -LRB- conversational telephone speech and broadcast news speech -RRB- on both [[ human transcriptions ]] and << speech recognition output >> .,314,0 |
|
315,"In general , our [[ CRF model ]] yields a lower error rate than the << HMM and Max-ent models >> on the NIST sentence boundary detection task in speech , although it is interesting to note that the best results are achieved by three-way voting among the classifiers .",315,5 |
|
316,"In general , our << CRF model >> yields a lower [[ error rate ]] than the HMM and Max-ent models on the NIST sentence boundary detection task in speech , although it is interesting to note that the best results are achieved by three-way voting among the classifiers .",316,6 |
|
317,"In general , our CRF model yields a lower [[ error rate ]] than the << HMM and Max-ent models >> on the NIST sentence boundary detection task in speech , although it is interesting to note that the best results are achieved by three-way voting among the classifiers .",317,6 |
|
318,"In general , our << CRF model >> yields a lower error rate than the HMM and Max-ent models on the [[ NIST sentence boundary detection task ]] in speech , although it is interesting to note that the best results are achieved by three-way voting among the classifiers .",318,6 |
|
319,"In general , our CRF model yields a lower error rate than the << HMM and Max-ent models >> on the [[ NIST sentence boundary detection task ]] in speech , although it is interesting to note that the best results are achieved by three-way voting among the classifiers .",319,6 |
|
320,"In general , our CRF model yields a lower error rate than the HMM and Max-ent models on the << NIST sentence boundary detection task >> in [[ speech ]] , although it is interesting to note that the best results are achieved by three-way voting among the classifiers .",320,1 |
|
321,"In general , our CRF model yields a lower error rate than the HMM and Max-ent models on the NIST sentence boundary detection task in speech , although it is interesting to note that the best results are achieved by << three-way voting >> among the [[ classifiers ]] .",321,3 |
|
322,This probably occurs because each [[ model ]] has different strengths and weaknesses for modeling the << knowledge sources >> .,322,3 |
|
323,We propose a novel [[ approach ]] to associate objects across multiple PTZ cameras that can be used to perform << camera handoff in wide-area surveillance scenarios >> .,323,3 |
|
324,"While previous << approaches >> relied on [[ geometric , appearance , or correlation-based information ]] for establishing correspondences between static cameras , they each have well-known limitations and are not extendable to wide-area settings with PTZ cameras .",324,3 |
|
325,"Towards this goal , we also propose a novel << Multiple Instance Learning -LRB- MIL -RRB- formulation >> for the problem based on the [[ logistic softmax function of covariance-based region features ]] within a MAP estimation framework .",325,3 |
|
326,"Towards this goal , we also propose a novel << Multiple Instance Learning -LRB- MIL -RRB- formulation >> for the problem based on the logistic softmax function of covariance-based region features within a [[ MAP estimation framework ]] .",326,3 |
|
327,We demonstrate our [[ approach ]] with multiple PTZ camera sequences in typical << outdoor surveillance settings >> and show a comparison with state-of-the-art approaches .,327,3 |
|
328,We demonstrate our [[ approach ]] with multiple PTZ camera sequences in typical outdoor surveillance settings and show a comparison with << state-of-the-art approaches >> .,328,5 |
|
329,We demonstrate our << approach >> with [[ multiple PTZ camera sequences ]] in typical outdoor surveillance settings and show a comparison with state-of-the-art approaches .,329,3 |
|
330,This paper solves a [[ specialized regression problem ]] to obtain << sampling probabilities >> for records in databases .,330,3 |
|
331,This paper solves a specialized regression problem to obtain [[ sampling probabilities ]] for << records >> in databases .,331,3 |
|
332,This paper solves a specialized regression problem to obtain sampling probabilities for [[ records ]] in << databases >> .,332,4 |
|
333,The goal is to sample a small set of << records >> over which evaluating [[ aggregate queries ]] can be done both efficiently and accurately .,333,6 |
|
334,We provide a [[ principled and provable solution ]] for this << problem >> ; it is parameterless and requires no data insights .,334,3 |
|
335,"Moreover , a << cost zero solution >> always exists and can only be excluded by [[ hard budget constraints ]] .",335,3 |
|
336,Our extensive experimental results significantly improve over both [[ uniform sampling ]] and standard << stratified sampling >> which are de-facto the industry standards .,336,0 |
|
337,"We consider the problem of computing the Kullback-Leibler distance , also called the relative entropy , between a [[ probabilistic context-free grammar ]] and a << probabilistic finite automaton >> .",337,5 |
|
338,"We show that there is a [[ closed-form -LRB- analytical -RRB- solution ]] for one part of the << Kullback-Leibler distance >> , viz the cross-entropy .",338,3 |
|
339,"We show that there is a [[ closed-form -LRB- analytical -RRB- solution ]] for one part of the Kullback-Leibler distance , viz the << cross-entropy >> .",339,3 |
|
340,"We show that there is a closed-form -LRB- analytical -RRB- solution for one part of the << Kullback-Leibler distance >> , viz the [[ cross-entropy ]] .",340,4 |
|
341,We discuss several applications of the result to the problem of [[ distributional approximation ]] of << probabilistic context-free grammars >> by means of probabilistic finite automata .,341,1 |
|
342,We discuss several applications of the result to the problem of << distributional approximation >> of probabilistic context-free grammars by means of [[ probabilistic finite automata ]] .,342,3 |
|
343,"In spite of over two decades of intense research , [[ illumination ]] and << pose invariance >> remain prohibitively challenging aspects of face recognition for most practical applications .",343,0 |
|
344,"In spite of over two decades of intense research , [[ illumination ]] and pose invariance remain prohibitively challenging aspects of << face recognition >> for most practical applications .",344,4 |
|
345,"In spite of over two decades of intense research , illumination and [[ pose invariance ]] remain prohibitively challenging aspects of << face recognition >> for most practical applications .",345,4 |
|
346,"The objective of this work is to recognize faces using video sequences both for training and recognition input , in a realistic , unconstrained setup in which [[ lighting ]] , << pose >> and user motion pattern have a wide variability and face images are of low resolution .",346,0 |
|
347,"The objective of this work is to recognize faces using video sequences both for training and recognition input , in a realistic , unconstrained setup in which lighting , [[ pose ]] and << user motion pattern >> have a wide variability and face images are of low resolution .",347,0 |
|
348,"The objective of this work is to recognize faces using video sequences both for training and recognition input , in a realistic , unconstrained setup in which lighting , pose and user motion pattern have a wide variability and << face images >> are of low [[ resolution ]] .",348,1 |
|
349,"In particular there are three areas of novelty : -LRB- i -RRB- we show how a [[ photometric model ]] of << image formation >> can be combined with a statistical model of generic face appearance variation , learnt offline , to generalize in the presence of extreme illumination changes ; -LRB- ii -RRB- we use the smoothness of geodesically local appearance manifold structure and a robust same-identity likelihood to achieve invariance to unseen head poses ; and -LRB- iii -RRB- we introduce an accurate video sequence '' reillumination '' algorithm to achieve robustness to face motion patterns in video .",349,3 |
|
350,"In particular there are three areas of novelty : -LRB- i -RRB- we show how a [[ photometric model ]] of image formation can be combined with a << statistical model >> of generic face appearance variation , learnt offline , to generalize in the presence of extreme illumination changes ; -LRB- ii -RRB- we use the smoothness of geodesically local appearance manifold structure and a robust same-identity likelihood to achieve invariance to unseen head poses ; and -LRB- iii -RRB- we introduce an accurate video sequence '' reillumination '' algorithm to achieve robustness to face motion patterns in video .",350,0 |
|
351,"In particular there are three areas of novelty : -LRB- i -RRB- we show how a photometric model of image formation can be combined with a [[ statistical model ]] of << generic face appearance variation >> , learnt offline , to generalize in the presence of extreme illumination changes ; -LRB- ii -RRB- we use the smoothness of geodesically local appearance manifold structure and a robust same-identity likelihood to achieve invariance to unseen head poses ; and -LRB- iii -RRB- we introduce an accurate video sequence '' reillumination '' algorithm to achieve robustness to face motion patterns in video .",351,3 |
|
352,"In particular there are three areas of novelty : -LRB- i -RRB- we show how a photometric model of image formation can be combined with a [[ statistical model ]] of generic face appearance variation , learnt offline , to generalize in the presence of << extreme illumination changes >> ; -LRB- ii -RRB- we use the smoothness of geodesically local appearance manifold structure and a robust same-identity likelihood to achieve invariance to unseen head poses ; and -LRB- iii -RRB- we introduce an accurate video sequence '' reillumination '' algorithm to achieve robustness to face motion patterns in video .",352,3 |
|
353,"In particular there are three areas of novelty : -LRB- i -RRB- we show how a photometric model of image formation can be combined with a statistical model of generic face appearance variation , learnt offline , to generalize in the presence of extreme illumination changes ; -LRB- ii -RRB- we use the [[ smoothness ]] of << geodesically local appearance manifold structure >> and a robust same-identity likelihood to achieve invariance to unseen head poses ; and -LRB- iii -RRB- we introduce an accurate video sequence '' reillumination '' algorithm to achieve robustness to face motion patterns in video .",353,1 |
|
354,"In particular there are three areas of novelty : -LRB- i -RRB- we show how a photometric model of image formation can be combined with a statistical model of generic face appearance variation , learnt offline , to generalize in the presence of extreme illumination changes ; -LRB- ii -RRB- we use the smoothness of [[ geodesically local appearance manifold structure ]] and a << robust same-identity likelihood >> to achieve invariance to unseen head poses ; and -LRB- iii -RRB- we introduce an accurate video sequence '' reillumination '' algorithm to achieve robustness to face motion patterns in video .",354,0 |
|
355,"In particular there are three areas of novelty : -LRB- i -RRB- we show how a photometric model of image formation can be combined with a statistical model of generic face appearance variation , learnt offline , to generalize in the presence of extreme illumination changes ; -LRB- ii -RRB- we use the smoothness of geodesically local appearance manifold structure and a robust same-identity likelihood to achieve invariance to unseen head poses ; and -LRB- iii -RRB- we introduce an accurate << video sequence '' reillumination '' algorithm >> to achieve [[ robustness ]] to face motion patterns in video .",355,6 |
|
356,"In particular there are three areas of novelty : -LRB- i -RRB- we show how a photometric model of image formation can be combined with a statistical model of generic face appearance variation , learnt offline , to generalize in the presence of extreme illumination changes ; -LRB- ii -RRB- we use the smoothness of geodesically local appearance manifold structure and a robust same-identity likelihood to achieve invariance to unseen head poses ; and -LRB- iii -RRB- we introduce an accurate video sequence '' reillumination '' algorithm to achieve << robustness >> to [[ face motion patterns ]] in video .",356,1 |
|
357,"In particular there are three areas of novelty : -LRB- i -RRB- we show how a photometric model of image formation can be combined with a statistical model of generic face appearance variation , learnt offline , to generalize in the presence of extreme illumination changes ; -LRB- ii -RRB- we use the smoothness of geodesically local appearance manifold structure and a robust same-identity likelihood to achieve invariance to unseen head poses ; and -LRB- iii -RRB- we introduce an accurate video sequence '' reillumination '' algorithm to achieve robustness to [[ face motion patterns ]] in << video >> .",357,4 |
|
358,"We describe a << fully automatic recognition system >> based on the proposed [[ method ]] and an extensive evaluation on 171 individuals and over 1300 video sequences with extreme illumination , pose and head motion variation .",358,3 |
|
359,"We describe a << fully automatic recognition system >> based on the proposed method and an extensive evaluation on 171 individuals and over 1300 [[ video sequences ]] with extreme illumination , pose and head motion variation .",359,6 |
|
360,"We describe a fully automatic recognition system based on the proposed method and an extensive evaluation on 171 individuals and over 1300 << video sequences >> with extreme [[ illumination ]] , pose and head motion variation .",360,1 |
|
361,"We describe a fully automatic recognition system based on the proposed method and an extensive evaluation on 171 individuals and over 1300 video sequences with extreme [[ illumination ]] , << pose >> and head motion variation .",361,0 |
|
362,"We describe a fully automatic recognition system based on the proposed method and an extensive evaluation on 171 individuals and over 1300 << video sequences >> with extreme illumination , [[ pose ]] and head motion variation .",362,1 |
|
363,"We describe a fully automatic recognition system based on the proposed method and an extensive evaluation on 171 individuals and over 1300 video sequences with extreme illumination , [[ pose ]] and << head motion variation >> .",363,0 |
|
364,"We describe a fully automatic recognition system based on the proposed method and an extensive evaluation on 171 individuals and over 1300 << video sequences >> with extreme illumination , pose and [[ head motion variation ]] .",364,1 |
|
365,"On this challenging [[ data set ]] our << system >> consistently demonstrated a nearly perfect recognition rate -LRB- over 99.7 % on all three databases -RRB- , significantly out-performing state-of-the-art commercial software and methods from the literature .",365,6 |
|
366,"On this challenging data set our [[ system ]] consistently demonstrated a nearly perfect recognition rate -LRB- over 99.7 % on all three databases -RRB- , significantly out-performing state-of-the-art << commercial software >> and methods from the literature .",366,5 |
|
367,"On this challenging data set our [[ system ]] consistently demonstrated a nearly perfect recognition rate -LRB- over 99.7 % on all three databases -RRB- , significantly out-performing state-of-the-art commercial software and << methods >> from the literature .",367,5 |
|
368,"On this challenging data set our << system >> consistently demonstrated a nearly perfect [[ recognition rate ]] -LRB- over 99.7 % on all three databases -RRB- , significantly out-performing state-of-the-art commercial software and methods from the literature .",368,6 |
|
369,"On this challenging data set our system consistently demonstrated a nearly perfect recognition rate -LRB- over 99.7 % on all three databases -RRB- , significantly out-performing state-of-the-art [[ commercial software ]] and << methods >> from the literature .",369,0 |
|
370,We present [[ Minimum Bayes-Risk -LRB- MBR -RRB- decoding ]] for << statistical machine translation >> .,370,3 |
|
371,This statistical approach aims to minimize expected loss of translation errors under [[ loss functions ]] that measure << translation >> performance .,371,6 |
|
372,"We describe a hierarchy of << loss functions >> that incorporate different levels of [[ linguistic information ]] from word strings , word-to-word alignments from an MT system , and syntactic structure from parse-trees of source and target language sentences .",372,3 |
|
373,"We describe a hierarchy of << loss functions >> that incorporate different levels of linguistic information from word strings , [[ word-to-word alignments ]] from an MT system , and syntactic structure from parse-trees of source and target language sentences .",373,3 |
|
374,"We describe a hierarchy of loss functions that incorporate different levels of linguistic information from word strings , [[ word-to-word alignments ]] from an << MT system >> , and syntactic structure from parse-trees of source and target language sentences .",374,4 |
|
375,"We describe a hierarchy of << loss functions >> that incorporate different levels of linguistic information from word strings , word-to-word alignments from an MT system , and [[ syntactic structure ]] from parse-trees of source and target language sentences .",375,3 |
|
376,"We describe a hierarchy of loss functions that incorporate different levels of linguistic information from word strings , word-to-word alignments from an MT system , and << syntactic structure >> from [[ parse-trees ]] of source and target language sentences .",376,4 |
|
377,We report the performance of the [[ MBR decoders ]] on a << Chinese-to-English translation task >> .,377,3 |
|
378,Our results show that [[ MBR decoding ]] can be used to tune << statistical MT >> performance for specific loss functions .,378,3 |
|
379,Our results show that [[ MBR decoding ]] can be used to tune statistical MT performance for specific << loss functions >> .,379,3 |
|
380,This paper presents a critical discussion of the various [[ approaches ]] that have been used in the << evaluation of Natural Language systems >> .,380,3 |
|
381,"We conclude that previous [[ approaches ]] have neglected to evaluate << systems >> in the context of their use , e.g. solving a task requiring data retrieval .",381,6 |
|
382,"We conclude that previous approaches have neglected to evaluate [[ systems ]] in the context of their use , e.g. solving a << task >> requiring data retrieval .",382,3 |
|
383,"We conclude that previous approaches have neglected to evaluate systems in the context of their use , e.g. solving a << task >> requiring [[ data retrieval ]] .",383,4 |
|
384,"In the second half of the paper , we report a laboratory study using the [[ Wizard of Oz technique ]] to identify << NL requirements >> for carrying out this task .",384,3 |
|
385,"In the second half of the paper , we report a laboratory study using the [[ Wizard of Oz technique ]] to identify NL requirements for carrying out this << task >> .",385,3 |
|
386,"We evaluate the demands that [[ task dialogues ]] collected using this technique , place upon a << prototype Natural Language system >> .",386,3 |
|
387,"We evaluate the demands that << task dialogues >> collected using this [[ technique ]] , place upon a prototype Natural Language system .",387,3 |
|
388,We present results on << addressee identification in four-participants face-to-face meetings >> using [[ Bayesian Network ]] and Naive Bayes classifiers .,388,3 |
|
389,We present results on << addressee identification in four-participants face-to-face meetings >> using Bayesian Network and [[ Naive Bayes classifiers ]] .,389,3 |
|
390,We present results on addressee identification in four-participants face-to-face meetings using << Bayesian Network >> and [[ Naive Bayes classifiers ]] .,390,0 |
|
391,"First , we investigate how well the << addressee of a dialogue act >> can be predicted based on [[ gaze ]] , utterance and conversational context features .",391,3 |
|
392,"First , we investigate how well the addressee of a dialogue act can be predicted based on [[ gaze ]] , << utterance >> and conversational context features .",392,0 |
|
393,"First , we investigate how well the << addressee of a dialogue act >> can be predicted based on gaze , [[ utterance ]] and conversational context features .",393,3 |
|
394,"First , we investigate how well the addressee of a dialogue act can be predicted based on gaze , [[ utterance ]] and << conversational context features >> .",394,0 |
|
395,"First , we investigate how well the << addressee of a dialogue act >> can be predicted based on gaze , utterance and [[ conversational context features ]] .",395,3 |
|
396,Both << classifiers >> perform the best when [[ conversational context ]] and utterance features are combined with speaker 's gaze information .,396,3 |
|
397,Both classifiers perform the best when [[ conversational context ]] and << utterance features >> are combined with speaker 's gaze information .,397,0 |
|
398,Both << classifiers >> perform the best when conversational context and [[ utterance features ]] are combined with speaker 's gaze information .,398,3 |
|
399,Both << classifiers >> perform the best when conversational context and utterance features are combined with [[ speaker 's gaze information ]] .,399,3 |
|
400,Both classifiers perform the best when conversational context and << utterance features >> are combined with [[ speaker 's gaze information ]] .,400,0 |
|
401,"Towards deep analysis of << compositional classes of paraphrases >> , we have examined a [[ class-oriented framework ]] for collecting paraphrase examples , in which sentential paraphrases are collected for each paraphrase class separately by means of automatic candidate generation and manual judgement .",401,3 |
|
402,"Towards deep analysis of compositional classes of paraphrases , we have examined a [[ class-oriented framework ]] for collecting << paraphrase examples >> , in which sentential paraphrases are collected for each paraphrase class separately by means of automatic candidate generation and manual judgement .",402,3 |
|
403,"Towards deep analysis of compositional classes of paraphrases , we have examined a class-oriented framework for collecting paraphrase examples , in which << sentential paraphrases >> are collected for each paraphrase class separately by means of [[ automatic candidate generation ]] and manual judgement .",403,3 |
|
404,"Towards deep analysis of compositional classes of paraphrases , we have examined a class-oriented framework for collecting paraphrase examples , in which sentential paraphrases are collected for each paraphrase class separately by means of [[ automatic candidate generation ]] and << manual judgement >> .",404,0 |
|
405,"Towards deep analysis of compositional classes of paraphrases , we have examined a class-oriented framework for collecting paraphrase examples , in which << sentential paraphrases >> are collected for each paraphrase class separately by means of automatic candidate generation and [[ manual judgement ]] .",405,3 |
|
406,"The purpose of this research is to test the efficacy of applying [[ automated evaluation techniques ]] , originally devised for the << evaluation of human language learners >> , to the output of machine translation -LRB- MT -RRB- systems .",406,3 |
|
407,"We believe that these [[ evaluation techniques ]] will provide information about both the << human language learning process >> , the translation process and the development of machine translation systems .",407,3 |
|
408,"We believe that these [[ evaluation techniques ]] will provide information about both the human language learning process , the << translation process >> and the development of machine translation systems .",408,3 |
|
409,"We believe that these [[ evaluation techniques ]] will provide information about both the human language learning process , the translation process and the development of << machine translation systems >> .",409,3 |
|
410,"We believe that these evaluation techniques will provide information about both the [[ human language learning process ]] , the << translation process >> and the development of machine translation systems .",410,0 |
|
411,"We believe that these evaluation techniques will provide information about both the human language learning process , the [[ translation process ]] and the development of << machine translation systems >> .",411,0 |
|
412,A [[ language learning ]] experiment showed that << assessors >> can differentiate native from non-native language essays in less than 100 words .,412,6 |
|
413,"Some of the extracts were << expert human translations >> , others were [[ machine translation outputs ]] .",413,0 |
|
414,The subjects were given three minutes per extract to determine whether they believed the sample output to be an [[ expert human translation ]] or a << machine translation >> .,414,5 |
|
415,This paper presents a [[ machine learning approach ]] to << bare slice disambiguation >> in dialogue .,415,3 |
|
416,This paper presents a machine learning approach to << bare slice disambiguation >> in [[ dialogue ]] .,416,3 |
|
417,We extract a set of << heuristic principles >> from a [[ corpus-based sample ]] and formulate them as probabilistic Horn clauses .,417,3 |
|
418,We extract a set of << heuristic principles >> from a corpus-based sample and formulate them as [[ probabilistic Horn clauses ]] .,418,1 |
|
419,"We then use the predicates of such clauses to create a set of domain independent features to annotate an input dataset , and run two different machine learning algorithms : [[ SLIPPER ]] , a << rule-based learning algorithm >> , and TiMBL , a memory-based system .",419,2 |
|
420,"We then use the predicates of such clauses to create a set of domain independent features to annotate an input dataset , and run two different << machine learning algorithms >> : SLIPPER , a [[ rule-based learning algorithm ]] , and TiMBL , a memory-based system .",420,4 |
|
421,"We then use the predicates of such clauses to create a set of domain independent features to annotate an input dataset , and run two different machine learning algorithms : SLIPPER , a [[ rule-based learning algorithm ]] , and TiMBL , a << memory-based system >> .",421,5 |
|
422,"We then use the predicates of such clauses to create a set of domain independent features to annotate an input dataset , and run two different machine learning algorithms : SLIPPER , a rule-based learning algorithm , and [[ TiMBL ]] , a << memory-based system >> .",422,2 |
|
423,"We then use the predicates of such clauses to create a set of domain independent features to annotate an input dataset , and run two different << machine learning algorithms >> : SLIPPER , a rule-based learning algorithm , and TiMBL , a [[ memory-based system ]] .",423,4 |
|
424,"The results show that the [[ features ]] in terms of which we formulate our << heuristic principles >> have significant predictive power , and that rules that closely resemble our Horn clauses can be learnt automatically from these features .",424,1 |
|
425,We suggest a new goal and [[ evaluation criterion ]] for << word similarity measures >> .,425,3 |
|
426,The new criterion -- [[ meaning-entailing substitutability ]] -- fits the needs of << semantic-oriented NLP applications >> and can be evaluated directly -LRB- independent of an application -RRB- at a good level of human agreement .,426,3 |
|
427,The new criterion -- << meaning-entailing substitutability >> -- fits the needs of semantic-oriented NLP applications and can be evaluated directly -LRB- independent of an application -RRB- at a good level of [[ human agreement ]] .,427,6 |
|
428,"Motivated by this [[ semantic criterion ]] we analyze the empirical quality of << distributional word feature vectors >> and its impact on word similarity results , proposing an objective measure for evaluating feature vector quality .",428,6 |
|
429,"Motivated by this semantic criterion we analyze the empirical quality of [[ distributional word feature vectors ]] and its impact on << word similarity >> results , proposing an objective measure for evaluating feature vector quality .",429,3 |
|
430,"Motivated by this semantic criterion we analyze the empirical quality of distributional word feature vectors and its impact on word similarity results , proposing an objective [[ measure ]] for evaluating << feature vector quality >> .",430,6 |
|
431,"Finally , a novel [[ feature weighting and selection function ]] is presented , which yields superior << feature vectors >> and better word similarity performance .",431,3 |
|
432,"Finally , a novel [[ feature weighting and selection function ]] is presented , which yields superior feature vectors and better << word similarity >> performance .",432,3 |
|
433,"Finally , a novel feature weighting and selection function is presented , which yields superior [[ feature vectors ]] and better << word similarity >> performance .",433,0 |
|
434,This phenomenon causes many image processing techniques to fail as they assume the presence of only one layer at each examined site e.g. [[ motion estimation ]] and << object recognition >> .,434,0 |
|
435,This work presents an automated [[ technique ]] for << detecting reflections in image sequences >> by analyzing motion trajectories of feature points .,435,3 |
|
436,This work presents an automated << technique >> for detecting reflections in image sequences by analyzing [[ motion trajectories ]] of feature points .,436,3 |
|
437,This work presents an automated technique for detecting reflections in image sequences by analyzing << motion trajectories >> of [[ feature points ]] .,437,1 |
|
438,[[ It ]] models << reflection >> as regions containing two different layers moving over each other .,438,3 |
|
439,We present a strong << detector >> based on combining a set of weak [[ detectors ]] .,439,3 |
|
440,"We use novel [[ priors ]] , generate << sparse and dense detection maps >> and our results show high detection rate with rejection to pathological motion and occlusion .",440,3 |
|
441,"We use novel priors , generate sparse and dense detection maps and our results show high detection rate with rejection to [[ pathological motion ]] and << occlusion >> .",441,0 |
|
442,This paper considers the problem of << reconstructing the motion of a 3D articulated tree >> from [[ 2D point correspondences ]] subject to some temporal prior .,442,3 |
|
443,"Hitherto , << smooth motion >> has been encouraged using a [[ trajectory basis ]] , yielding a hard combinatorial problem with time complexity growing exponentially in the number of frames .",443,3 |
|
444,"Hitherto , smooth motion has been encouraged using a trajectory basis , yielding a << hard combinatorial problem >> with [[ time complexity ]] growing exponentially in the number of frames .",444,6 |
|
445,[[ Branch and bound strategies ]] have previously attempted to curb this << complexity >> whilst maintaining global optimality .,445,3 |
|
446,<< Branch and bound strategies >> have previously attempted to curb this complexity whilst maintaining [[ global optimality ]] .,446,1 |
|
447,"However , [[ they ]] provide no guarantee of being more efficient than << exhaustive search >> .",447,5 |
|
448,Extension to [[ affine projection ]] enables << reconstruction >> without estimating cameras .,448,3 |
|
449,[[ Topical blog post retrieval ]] is the task of << ranking blog posts >> with respect to their relevance for a given topic .,449,2 |
|
450,Topical blog post retrieval is the task of ranking << blog posts >> with respect to their [[ relevance ]] for a given topic .,450,1 |
|
451,To improve << topical blog post retrieval >> we incorporate [[ textual credibility indicators ]] in the retrieval process .,451,3 |
|
452,To improve topical blog post retrieval we incorporate [[ textual credibility indicators ]] in the << retrieval process >> .,452,4 |
|
453,We describe how to estimate these indicators and how to integrate [[ them ]] into a << retrieval approach >> based on language models .,453,4 |
|
454,We describe how to estimate these indicators and how to integrate << them >> into a retrieval approach based on [[ language models ]] .,454,3 |
|
455,Experiments on the [[ TREC Blog track test set ]] show that both groups of << credibility indicators >> significantly improve retrieval effectiveness ; the best performance is achieved when combining them .,455,6 |
|
456,Experiments on the TREC Blog track test set show that both groups of << credibility indicators >> significantly improve [[ retrieval effectiveness ]] ; the best performance is achieved when combining them .,456,6 |
|
457,We investigate the problem of learning to predict moves in the << board game of Go >> from [[ game records of expert players ]] .,457,3 |
|
458,"This [[ distribution ]] has numerous applications in << computer Go >> , including serving as an efficient stand-alone Go player .",458,3 |
|
459,[[ It ]] would also be effective as a << move selector >> and move sorter for game tree search and as a training tool for Go players .,459,3 |
|
460,[[ It ]] would also be effective as a move selector and << move sorter >> for game tree search and as a training tool for Go players .,460,3 |
|
461,[[ It ]] would also be effective as a move selector and move sorter for game tree search and as a << training tool >> for Go players .,461,3 |
|
462,It would also be effective as a [[ move selector ]] and << move sorter >> for game tree search and as a training tool for Go players .,462,0 |
|
463,It would also be effective as a [[ move selector ]] and move sorter for << game tree search >> and as a training tool for Go players .,463,3 |
|
464,It would also be effective as a move selector and [[ move sorter ]] for << game tree search >> and as a training tool for Go players .,464,3 |
|
465,It would also be effective as a move selector and move sorter for game tree search and as a [[ training tool ]] for << Go players >> .,465,3 |
|
466,Our << method >> has two major components : a -RRB- a [[ pattern extraction scheme ]] for efficiently harvesting patterns of given size and shape from expert game records and b -RRB- a Bayesian learning algorithm -LRB- in two variants -RRB- that learns a distribution over the values of a move given a board position based on the local pattern context .,466,4 |
|
467,Our method has two major components : a -RRB- a [[ pattern extraction scheme ]] for efficiently harvesting patterns of given size and shape from expert game records and b -RRB- a << Bayesian learning algorithm >> -LRB- in two variants -RRB- that learns a distribution over the values of a move given a board position based on the local pattern context .,467,0 |
|
468,Our << method >> has two major components : a -RRB- a pattern extraction scheme for efficiently harvesting patterns of given size and shape from expert game records and b -RRB- a [[ Bayesian learning algorithm ]] -LRB- in two variants -RRB- that learns a distribution over the values of a move given a board position based on the local pattern context .,468,4 |
|
469,"The << system >> is trained on 181,000 [[ expert games ]] and shows excellent prediction performance as indicated by its ability to perfectly predict the moves made by professional Go players in 34 % of test positions .",469,3 |
|
470,We present a novel [[ approach ]] for << automatically acquiring English topic signatures >> .,470,3 |
|
471,"[[ Topic signatures ]] can be useful in a number of << Natural Language Processing -LRB- NLP -RRB- applications >> , such as Word Sense Disambiguation -LRB- WSD -RRB- and Text Summarisation .",471,3 |
|
472,"[[ Topic signatures ]] can be useful in a number of Natural Language Processing -LRB- NLP -RRB- applications , such as << Word Sense Disambiguation -LRB- WSD -RRB- >> and Text Summarisation .",472,3 |
|
473,"[[ Topic signatures ]] can be useful in a number of Natural Language Processing -LRB- NLP -RRB- applications , such as Word Sense Disambiguation -LRB- WSD -RRB- and << Text Summarisation >> .",473,3 |
|
474,"Topic signatures can be useful in a number of << Natural Language Processing -LRB- NLP -RRB- applications >> , such as [[ Word Sense Disambiguation -LRB- WSD -RRB- ]] and Text Summarisation .",474,2 |
|
475,"Topic signatures can be useful in a number of Natural Language Processing -LRB- NLP -RRB- applications , such as [[ Word Sense Disambiguation -LRB- WSD -RRB- ]] and << Text Summarisation >> .",475,0 |
|
476,"Topic signatures can be useful in a number of << Natural Language Processing -LRB- NLP -RRB- applications >> , such as Word Sense Disambiguation -LRB- WSD -RRB- and [[ Text Summarisation ]] .",476,2 |
|
477,"Our method takes advantage of the different way in which word senses are lexicalised in English and Chinese , and also exploits the large amount of [[ Chinese text ]] available in << corpora >> and on the Web .",477,4 |
|
478,"Our method takes advantage of the different way in which word senses are lexicalised in English and Chinese , and also exploits the large amount of [[ Chinese text ]] available in corpora and on the << Web >> .",478,4 |
|
479,"Our method takes advantage of the different way in which word senses are lexicalised in English and Chinese , and also exploits the large amount of Chinese text available in [[ corpora ]] and on the << Web >> .",479,0 |
|
480,"We evaluated the << topic signatures >> on a [[ WSD task ]] , where we trained a second-order vector cooccurrence algorithm on standard WSD datasets , with promising results .",480,6 |
|
481,"We evaluated the topic signatures on a WSD task , where we trained a << second-order vector cooccurrence algorithm >> on standard [[ WSD datasets ]] , with promising results .",481,3 |
|
482,"[[ Joint matrix triangularization ]] is often used for estimating the << joint eigenstructure >> of a set M of matrices , with applications in signal processing and machine learning .",482,3 |
|
483,"Joint matrix triangularization is often used for estimating the [[ joint eigenstructure ]] of a set M of matrices , with applications in << signal processing >> and machine learning .",483,3 |
|
484,"Joint matrix triangularization is often used for estimating the [[ joint eigenstructure ]] of a set M of matrices , with applications in signal processing and << machine learning >> .",484,3 |
|
485,"Joint matrix triangularization is often used for estimating the joint eigenstructure of a set M of matrices , with applications in [[ signal processing ]] and << machine learning >> .",485,0 |
|
486,Our main result is a first-order upper bound on the distance between any [[ approximate joint triangularizer ]] of the matrices in M ' and any << exact joint triangularizer >> of the matrices in M .,486,0 |
|
487,"To our knowledge , this is the first a [[ posteriori bound ]] for << joint matrix decomposition >> .",487,3 |
|
488,"The [[ psycholinguistic literature ]] provides evidence for << syntactic priming >> , i.e. , the tendency to repeat structures .",488,3 |
|
489,This paper describes a [[ method ]] for incorporating << priming >> into an incremental probabilistic parser .,489,3 |
|
490,This paper describes a method for incorporating [[ priming ]] into an << incremental probabilistic parser >> .,490,3 |
|
491,"These models simulate the reading time advantage for [[ parallel structures ]] found in << human data >> , and also yield a small increase in overall parsing accuracy .",491,4 |
|
492,[[ Learned confidence measures ]] gain increasing importance for << outlier removal >> and quality improvement in stereo vision .,492,3 |
|
493,[[ Learned confidence measures ]] gain increasing importance for outlier removal and << quality improvement >> in stereo vision .,493,3 |
|
494,Learned confidence measures gain increasing importance for [[ outlier removal ]] and << quality improvement >> in stereo vision .,494,0 |
|
495,Learned confidence measures gain increasing importance for [[ outlier removal ]] and quality improvement in << stereo vision >> .,495,4 |
|
496,Learned confidence measures gain increasing importance for outlier removal and [[ quality improvement ]] in << stereo vision >> .,496,4 |
|
497,"However , acquiring the necessary training data is typically a tedious and time consuming << task >> that involves [[ manual interaction ]] , active sensing devices and/or synthetic scenes .",497,3 |
|
498,"However , acquiring the necessary training data is typically a tedious and time consuming task that involves [[ manual interaction ]] , << active sensing devices >> and/or synthetic scenes .",498,0 |
|
499,"However , acquiring the necessary training data is typically a tedious and time consuming << task >> that involves manual interaction , [[ active sensing devices ]] and/or synthetic scenes .",499,3 |
|
500,"However , acquiring the necessary training data is typically a tedious and time consuming task that involves manual interaction , [[ active sensing devices ]] and/or << synthetic scenes >> .",500,0 |
|
501,"However , acquiring the necessary training data is typically a tedious and time consuming << task >> that involves manual interaction , active sensing devices and/or [[ synthetic scenes ]] .",501,3 |
|
502,The key idea of our << approach >> is to use different [[ view points ]] for reasoning about contradictions and consistencies between multiple depth maps generated with the same stereo algorithm .,502,3 |
|
503,"Among other experiments , we demonstrate the potential of our [[ approach ]] by boosting the performance of three << learned confidence measures >> on the KITTI2012 dataset by simply training them on a vast amount of automatically generated training data rather than a limited amount of laser ground truth data .",503,3 |
|
504,"Among other experiments , we demonstrate the potential of our approach by boosting the performance of three << learned confidence measures >> on the [[ KITTI2012 dataset ]] by simply training them on a vast amount of automatically generated training data rather than a limited amount of laser ground truth data .",504,6 |
|
505,"Among other experiments , we demonstrate the potential of our approach by boosting the performance of three learned confidence measures on the KITTI2012 dataset by simply training << them >> on a vast amount of [[ automatically generated training data ]] rather than a limited amount of laser ground truth data .",505,3 |
|
506,"Among other experiments , we demonstrate the potential of our approach by boosting the performance of three learned confidence measures on the KITTI2012 dataset by simply training them on a vast amount of << automatically generated training data >> rather than a limited amount of [[ laser ground truth data ]] .",506,5 |
|
507,An important area of [[ learning in autonomous agents ]] is the ability to learn << domain-speciic models of actions >> to be used by planning systems .,507,3 |
|
508,An important area of learning in autonomous agents is the ability to learn << domain-speciic models of actions >> to be used by [[ planning systems ]] .,508,3 |
|
509,"These << methods >> diier from previous work in the area in two ways : the use of an [[ action model formalism ]] which is better suited to the needs of a re-active agent , and successful implementation of noise-handling mechanisms .",509,3 |
|
510,"These methods diier from previous work in the area in two ways : the use of an [[ action model formalism ]] which is better suited to the needs of a << re-active agent >> , and successful implementation of noise-handling mechanisms .",510,3 |
|
511,"These << methods >> diier from previous work in the area in two ways : the use of an action model formalism which is better suited to the needs of a re-active agent , and successful implementation of [[ noise-handling mechanisms ]] .",511,3 |
|
512,"Training instances are generated from experience and observation , and a variant of [[ GOLEM ]] is used to learn << action models >> from these instances .",512,3 |
|
513,The << integrated learning system >> has been experimentally validated in [[ simulated construction ]] and ooce domains .,513,6 |
|
514,The integrated learning system has been experimentally validated in [[ simulated construction ]] and << ooce domains >> .,514,0 |
|
515,The << integrated learning system >> has been experimentally validated in simulated construction and [[ ooce domains ]] .,515,6 |
|
516,"This paper describes [[ FERRET ]] , an << interactive question-answering -LRB- Q/A -RRB- system >> designed to address the challenges of integrating automatic Q/A applications into real-world environments .",516,2 |
|
517,"This paper describes [[ FERRET ]] , an interactive question-answering -LRB- Q/A -RRB- system designed to address the challenges of << integrating automatic Q/A applications into real-world environments >> .",517,3 |
|
518,<< FERRET >> utilizes a novel [[ approach ]] to Q/A known as predictive questioning which attempts to identify the questions -LRB- and answers -RRB- that users need by analyzing how a user interacts with a system while gathering information related to a particular scenario .,518,3 |
|
519,FERRET utilizes a novel [[ approach ]] to << Q/A >> known as predictive questioning which attempts to identify the questions -LRB- and answers -RRB- that users need by analyzing how a user interacts with a system while gathering information related to a particular scenario .,519,3 |
|
520,"In order to build robust << automatic abstracting systems >> , there is a need for better [[ training resources ]] than are currently available .",520,3 |
|
521,"In this paper , we introduce an [[ annotation scheme ]] for << scientific articles >> which can be used to build such a resource in a consistent way .",521,3 |
|
522,"In this paper , we introduce an [[ annotation scheme ]] for scientific articles which can be used to build such a << resource >> in a consistent way .",522,3 |
|
523,The seven categories of the << scheme >> are based on [[ rhetorical moves of argumentation ]] .,523,3 |
|
524,The << automated segmentation >> of [[ images ]] into semantically meaningful parts requires shape information since low-level feature analysis alone often fails to reach this goal .,524,3 |
|
525,We introduce a novel [[ method ]] of << shape constrained image segmentation >> which is based on mixtures of feature distributions for color and texture as well as probabilistic shape knowledge .,525,3 |
|
526,We introduce a novel << method >> of shape constrained image segmentation which is based on [[ mixtures of feature distributions ]] for color and texture as well as probabilistic shape knowledge .,526,3 |
|
527,We introduce a novel method of shape constrained image segmentation which is based on [[ mixtures of feature distributions ]] for << color >> and texture as well as probabilistic shape knowledge .,527,3 |
|
528,We introduce a novel method of shape constrained image segmentation which is based on [[ mixtures of feature distributions ]] for color and << texture >> as well as probabilistic shape knowledge .,528,3 |
|
529,We introduce a novel method of shape constrained image segmentation which is based on [[ mixtures of feature distributions ]] for color and texture as well as << probabilistic shape knowledge >> .,529,3 |
|
530,We introduce a novel method of shape constrained image segmentation which is based on mixtures of feature distributions for [[ color ]] and << texture >> as well as probabilistic shape knowledge .,530,0 |
|
531,We introduce a novel method of shape constrained image segmentation which is based on mixtures of feature distributions for color and [[ texture ]] as well as << probabilistic shape knowledge >> .,531,0 |
|
532,The combined [[ approach ]] is formulated in the framework of Bayesian statistics to account for the << robust-ness requirement in image understanding >> .,532,3 |
|
533,The combined << approach >> is formulated in the framework of [[ Bayesian statistics ]] to account for the robust-ness requirement in image understanding .,533,3 |
|
534,The goal of this work is the enrichment of << human-machine interactions >> in a [[ natural language environment ]] .,534,1 |
|
535,This paper highlights a particular class of << miscommunication >> |
|
536,This paper highlights a particular class of miscommunication |
|
537,This paper examines the benefits of [[ system combination ]] for << unsupervised WSD >> .,537,3 |
|
538,We investigate several [[ voting - and arbiter-based combination strategies ]] over a diverse pool of << unsupervised WSD systems >> .,538,3 |
|
539,Our << combination methods >> rely on [[ predominant senses ]] which are derived automatically from raw text .,539,3 |
|
540,Our combination methods rely on << predominant senses >> which are derived automatically from [[ raw text ]] .,540,3 |
|
541,Experiments using the [[ SemCor and Senseval-3 data sets ]] demonstrate that our << ensembles >> yield significantly better results when compared with state-of-the-art .,541,6 |
|
542,Experiments using the [[ SemCor and Senseval-3 data sets ]] demonstrate that our ensembles yield significantly better results when compared with << state-of-the-art >> .,542,6 |
|
543,The applicability of many current << information extraction techniques >> is severely limited by the need for [[ supervised training data ]] .,543,3 |
|
544,"We demonstrate that for certain << field structured extraction tasks >> , such as [[ classified advertisements ]] and bibliographic citations , small amounts of prior knowledge can be used to learn effective models in a primarily unsupervised fashion .",544,2 |
|
545,"We demonstrate that for certain field structured extraction tasks , such as [[ classified advertisements ]] and << bibliographic citations >> , small amounts of prior knowledge can be used to learn effective models in a primarily unsupervised fashion .",545,0 |
|
546,"We demonstrate that for certain << field structured extraction tasks >> , such as classified advertisements and [[ bibliographic citations ]] , small amounts of prior knowledge can be used to learn effective models in a primarily unsupervised fashion .",546,2 |
|
547,"We demonstrate that for certain << field structured extraction tasks >> , such as classified advertisements and bibliographic citations , small amounts of [[ prior knowledge ]] can be used to learn effective models in a primarily unsupervised fashion .",547,3 |
|
548,"Although [[ hidden Markov models -LRB- HMMs -RRB- ]] provide a suitable << generative model >> for field structured text , general unsupervised HMM learning fails to learn useful structure in either of our domains .",548,3 |
|
549,"Although hidden Markov models -LRB- HMMs -RRB- provide a suitable [[ generative model ]] for << field structured text >> , general unsupervised HMM learning fails to learn useful structure in either of our domains .",549,3 |
|
550,"In both domains , we found that [[ unsupervised methods ]] can attain accuracies with 400 unlabeled examples comparable to those attained by << supervised methods >> on 50 labeled examples , and that semi-supervised methods can make good use of small amounts of labeled data .",550,5 |
|
551,"In both domains , we found that << unsupervised methods >> can attain [[ accuracies ]] with 400 unlabeled examples comparable to those attained by supervised methods on 50 labeled examples , and that semi-supervised methods can make good use of small amounts of labeled data .",551,6 |
|
552,"In both domains , we found that unsupervised methods can attain [[ accuracies ]] with 400 unlabeled examples comparable to those attained by << supervised methods >> on 50 labeled examples , and that semi-supervised methods can make good use of small amounts of labeled data .",552,6 |
|
553,"In both domains , we found that << unsupervised methods >> can attain accuracies with 400 [[ unlabeled examples ]] comparable to those attained by supervised methods on 50 labeled examples , and that semi-supervised methods can make good use of small amounts of labeled data .",553,3 |
|
554,"In both domains , we found that unsupervised methods can attain accuracies with 400 unlabeled examples comparable to those attained by << supervised methods >> on 50 [[ labeled examples ]] , and that semi-supervised methods can make good use of small amounts of labeled data .",554,3 |
|
555,"In both domains , we found that unsupervised methods can attain accuracies with 400 unlabeled examples comparable to those attained by supervised methods on 50 labeled examples , and that << semi-supervised methods >> can make good use of small amounts of [[ labeled data ]] .",555,3 |
|
556,"This paper gives an overall account of a prototype << natural language question answering system >> , called [[ Chat-80 ]] .",556,2 |
|
557,"The << system >> is implemented entirely in [[ Prolog ]] , a programming language based on logic .",557,3 |
|
558,"The system is implemented entirely in [[ Prolog ]] , a << programming language >> based on logic .",558,2 |
|
559,"The system is implemented entirely in Prolog , a << programming language >> based on [[ logic ]] .",559,3 |
|
560,"With the aid of a << logic-based grammar formalism >> called [[ extraposition grammars ]] , Chat-80 translates English questions into the Prolog subset of logic .",560,2 |
|
561,"With the aid of a logic-based grammar formalism called [[ extraposition grammars ]] , << Chat-80 >> translates English questions into the Prolog subset of logic .",561,3 |
|
562,"The resulting << logical expression >> is then transformed by a [[ planning algorithm ]] into efficient Prolog , cf. query optimisation in a relational database .",562,3 |
|
563,"The resulting logical expression is then transformed by a planning algorithm into efficient Prolog , cf. << query optimisation >> in a [[ relational database ]] .",563,3 |
|
564,<< Human action recognition >> from [[ well-segmented 3D skeleton data ]] has been intensively studied and attracting an increasing attention .,564,3 |
|
565,"[[ Online action detection ]] goes one step further and is more challenging , which identifies the << action type >> and localizes the action positions on the fly from the untrimmed stream .",565,3 |
|
566,"[[ Online action detection ]] goes one step further and is more challenging , which identifies the action type and localizes the << action positions >> on the fly from the untrimmed stream .",566,3 |
|
567,"Online action detection goes one step further and is more challenging , which identifies the [[ action type ]] and localizes the << action positions >> on the fly from the untrimmed stream .",567,0 |
|
568,"<< Online action detection >> goes one step further and is more challenging , which identifies the action type and localizes the action positions on the fly from the [[ untrimmed stream ]] .",568,3 |
|
569,"In this paper , we study the problem of << online action detection >> from the [[ streaming skeleton data ]] .",569,3 |
|
570,We propose a [[ multi-task end-to-end Joint Classification-Regression Recurrent Neural Network ]] to better explore the << action type >> and temporal localiza-tion information .,570,3 |
|
571,We propose a [[ multi-task end-to-end Joint Classification-Regression Recurrent Neural Network ]] to better explore the action type and << temporal localiza-tion information >> .,571,3 |
|
572,We propose a multi-task end-to-end Joint Classification-Regression Recurrent Neural Network to better explore the [[ action type ]] and << temporal localiza-tion information >> .,572,0 |
|
573,"By employing a [[ joint classification and regression optimization objective ]] , this << network >> is capable of automatically localizing the start and end points of actions more accurately .",573,3 |
|
574,"Specifically , by leveraging the merits of the [[ deep Long Short-Term Memory -LRB- LSTM -RRB- subnetwork ]] , the proposed << model >> automatically captures the complex long-range temporal dynamics , which naturally avoids the typical sliding window design and thus ensures high computational efficiency .",574,3 |
|
575,"Specifically , by leveraging the merits of the deep Long Short-Term Memory -LRB- LSTM -RRB- subnetwork , the proposed << model >> automatically captures the complex [[ long-range temporal dynamics ]] , which naturally avoids the typical sliding window design and thus ensures high computational efficiency .",575,1 |
|
576,"To evaluate our proposed << model >> , we build a large [[ streaming video dataset ]] with annotations .",576,6 |
|
577,Experimental results on our [[ dataset ]] and the public << G3D dataset >> both demonstrate very promising performance of our scheme .,577,0 |
|
578,The task of [[ machine translation -LRB- MT -RRB- evaluation ]] is closely related to the task of << sentence-level semantic equivalence classification >> .,578,0 |
|
579,"This paper investigates the utility of applying standard [[ MT evaluation methods ]] -LRB- BLEU , NIST , WER and PER -RRB- to building << classifiers >> to predict semantic equivalence and entailment .",579,3 |
|
580,"This paper investigates the utility of applying standard << MT evaluation methods >> -LRB- [[ BLEU ]] , NIST , WER and PER -RRB- to building classifiers to predict semantic equivalence and entailment .",580,2 |
|
581,"This paper investigates the utility of applying standard MT evaluation methods -LRB- [[ BLEU ]] , << NIST >> , WER and PER -RRB- to building classifiers to predict semantic equivalence and entailment .",581,0 |
|
582,"This paper investigates the utility of applying standard << MT evaluation methods >> -LRB- BLEU , [[ NIST ]] , WER and PER -RRB- to building classifiers to predict semantic equivalence and entailment .",582,2 |
|
583,"This paper investigates the utility of applying standard MT evaluation methods -LRB- BLEU , [[ NIST ]] , << WER >> and PER -RRB- to building classifiers to predict semantic equivalence and entailment .",583,0 |
|
584,"This paper investigates the utility of applying standard << MT evaluation methods >> -LRB- BLEU , NIST , [[ WER ]] and PER -RRB- to building classifiers to predict semantic equivalence and entailment .",584,2 |
|
585,"This paper investigates the utility of applying standard MT evaluation methods -LRB- BLEU , NIST , [[ WER ]] and << PER >> -RRB- to building classifiers to predict semantic equivalence and entailment .",585,0 |
|
586,"This paper investigates the utility of applying standard << MT evaluation methods >> -LRB- BLEU , NIST , WER and [[ PER ]] -RRB- to building classifiers to predict semantic equivalence and entailment .",586,2 |
|
587,"This paper investigates the utility of applying standard MT evaluation methods -LRB- BLEU , NIST , WER and PER -RRB- to building [[ classifiers ]] to predict << semantic equivalence >> and entailment .",587,3 |
|
588,"This paper investigates the utility of applying standard MT evaluation methods -LRB- BLEU , NIST , WER and PER -RRB- to building [[ classifiers ]] to predict semantic equivalence and << entailment >> .",588,3 |
|
589,"This paper investigates the utility of applying standard MT evaluation methods -LRB- BLEU , NIST , WER and PER -RRB- to building classifiers to predict [[ semantic equivalence ]] and << entailment >> .",589,0 |
|
590,We also introduce a novel << classification method >> based on [[ PER ]] which leverages part of speech information of the words contributing to the word matches and non-matches in the sentence .,590,3 |
|
591,We also introduce a novel classification method based on [[ PER ]] which leverages << part of speech information >> of the words contributing to the word matches and non-matches in the sentence .,591,3 |
|
592,We also introduce a novel classification method based on PER which leverages [[ part of speech information ]] of the words contributing to the << word matches and non-matches >> in the sentence .,592,3 |
|
593,Our results show that [[ MT evaluation techniques ]] are able to produce useful << features >> for paraphrase classification and to a lesser extent entailment .,593,3 |
|
594,Our results show that [[ MT evaluation techniques ]] are able to produce useful features for << paraphrase classification >> and to a lesser extent entailment .,594,3 |
|
595,Our results show that [[ MT evaluation techniques ]] are able to produce useful features for paraphrase classification and to a lesser extent << entailment >> .,595,3 |
|
596,Our results show that MT evaluation techniques are able to produce useful features for [[ paraphrase classification ]] and to a lesser extent << entailment >> .,596,0 |
|
597,Our [[ technique ]] gives a substantial improvement in paraphrase classification accuracy over all of the other << models >> used in the experiments .,597,5 |
|
598,Our << technique >> gives a substantial improvement in [[ paraphrase classification accuracy ]] over all of the other models used in the experiments .,598,6 |
|
599,Our technique gives a substantial improvement in [[ paraphrase classification accuracy ]] over all of the other << models >> used in the experiments .,599,6 |
|
600,"Given an object model and a black-box measure of similarity between the model and candidate targets , we consider << visual object tracking >> as a [[ numerical optimization problem ]] .",600,3 |
|
601,"During normal tracking conditions when the object is visible from frame to frame , [[ local optimization ]] is used to track the << local mode of the similarity measure >> in a parameter space of translation , rotation and scale .",601,3 |
|
602,"During normal tracking conditions when the object is visible from frame to frame , local optimization is used to track the << local mode of the similarity measure >> in a [[ parameter space of translation , rotation and scale ]] .",602,3 |
|
603,"However , when the object becomes partially or totally occluded , such local tracking is prone to failure , especially when common << prediction techniques >> like the [[ Kalman filter ]] do not provide a good estimate of object parameters in future frames .",603,4 |
|
604,"To recover from these inevitable tracking failures , we consider << object detection >> as a [[ global optimization problem ]] and solve it via Adaptive Simulated Annealing -LRB- ASA -RRB- , a method that avoids becoming trapped at local modes and is much faster than exhaustive search .",604,3 |
|
605,"To recover from these inevitable tracking failures , we consider object detection as a global optimization problem and solve << it >> via [[ Adaptive Simulated Annealing -LRB- ASA -RRB- ]] , a method that avoids becoming trapped at local modes and is much faster than exhaustive search .",605,3 |
|
606,"To recover from these inevitable tracking failures , we consider object detection as a global optimization problem and solve it via Adaptive Simulated Annealing -LRB- ASA -RRB- , a [[ method ]] that avoids becoming trapped at local modes and is much faster than << exhaustive search >> .",606,5 |
|
607,"As a << Monte Carlo approach >> , [[ ASA ]] stochastically samples the parameter space , in contrast to local deterministic search .",607,2 |
|
608,"As a Monte Carlo approach , [[ ASA ]] stochastically samples the parameter space , in contrast to << local deterministic search >> .",608,5 |
|
609,We apply [[ cluster analysis ]] on the << sampled parameter space >> to redetect the object and renew the local tracker .,609,3 |
|
610,We apply [[ cluster analysis ]] on the sampled parameter space to redetect the object and renew the << local tracker >> .,610,3 |
|
611,Our << numerical hybrid local and global mode-seeking tracker >> is validated on challenging [[ airborne videos ]] with heavy occlusion and large camera motions .,611,6 |
|
612,Our numerical hybrid local and global mode-seeking tracker is validated on challenging << airborne videos >> with [[ heavy occlusion ]] and large camera motions .,612,1 |
|
613,Our numerical hybrid local and global mode-seeking tracker is validated on challenging airborne videos with [[ heavy occlusion ]] and large << camera motions >> .,613,0 |
|
614,Our numerical hybrid local and global mode-seeking tracker is validated on challenging << airborne videos >> with heavy occlusion and large [[ camera motions ]] .,614,1 |
|
615,Our << approach >> outperforms [[ state-of-the-art trackers ]] on the VIVID benchmark datasets .,615,5 |
|
616,Our << approach >> outperforms state-of-the-art trackers on the [[ VIVID benchmark datasets ]] .,616,6 |
|
617,Our approach outperforms << state-of-the-art trackers >> on the [[ VIVID benchmark datasets ]] .,617,6 |
|
618,"[[ Techniques ]] for << automatically training modules >> of a natural language generator have recently been proposed , but a fundamental concern is whether the quality of utterances produced with trainable components can compete with hand-crafted template-based or rule-based approaches .",618,3 |
|
619,"Techniques for [[ automatically training modules ]] of a << natural language generator >> have recently been proposed , but a fundamental concern is whether the quality of utterances produced with trainable components can compete with hand-crafted template-based or rule-based approaches .",619,4 |
|
620,"Techniques for automatically training modules of a natural language generator have recently been proposed , but a fundamental concern is whether the quality of [[ utterances ]] produced with << trainable components >> can compete with hand-crafted template-based or rule-based approaches .",620,6 |
|
621,"Techniques for automatically training modules of a natural language generator have recently been proposed , but a fundamental concern is whether the quality of [[ utterances ]] produced with trainable components can compete with << hand-crafted template-based or rule-based approaches >> .",621,6 |
|
622,"Techniques for automatically training modules of a natural language generator have recently been proposed , but a fundamental concern is whether the quality of utterances produced with [[ trainable components ]] can compete with << hand-crafted template-based or rule-based approaches >> .",622,5 |
|
623,In this paper We experimentally evaluate a [[ trainable sentence planner ]] for a << spoken dialogue system >> by eliciting subjective human judgments .,623,3 |
|
624,In this paper We experimentally evaluate a << trainable sentence planner >> for a spoken dialogue system by eliciting [[ subjective human judgments ]] .,624,6 |
|
625,"In order to perform an exhaustive comparison , we also evaluate a [[ hand-crafted template-based generation component ]] , two << rule-based sentence planners >> , and two baseline sentence planners .",625,0 |
|
626,"In order to perform an exhaustive comparison , we also evaluate a hand-crafted template-based generation component , two [[ rule-based sentence planners ]] , and two << baseline sentence planners >> .",626,0 |
|
627,"We show that the [[ trainable sentence planner ]] performs better than the << rule-based systems >> and the baselines , and as well as the hand-crafted system .",627,5 |
|
628,"We show that the [[ trainable sentence planner ]] performs better than the rule-based systems and the << baselines >> , and as well as the hand-crafted system .",628,5 |
|
629,"We show that the [[ trainable sentence planner ]] performs better than the rule-based systems and the baselines , and as well as the << hand-crafted system >> .",629,5 |
|
630,"We show that the trainable sentence planner performs better than the [[ rule-based systems ]] and the << baselines >> , and as well as the hand-crafted system .",630,0 |
|
631,"We show that the trainable sentence planner performs better than the rule-based systems and the [[ baselines ]] , and as well as the << hand-crafted system >> .",631,0 |
|
632,A new [[ algorithm ]] is proposed for << novel view generation >> in one-to-one teleconferencing applications .,632,3 |
|
633,A new algorithm is proposed for [[ novel view generation ]] in << one-to-one teleconferencing applications >> .,633,3 |
|
634,"Given the << video streams >> acquired by two [[ cameras ]] placed on either side of a computer monitor , the proposed algorithm synthesises images from a virtual camera in arbitrary position -LRB- typically located within the monitor -RRB- to facilitate eye contact .",634,3 |
|
635,"Given the video streams acquired by two cameras placed on either side of a computer monitor , the proposed [[ algorithm ]] synthesises images from a virtual camera in arbitrary position -LRB- typically located within the monitor -RRB- to facilitate << eye contact >> .",635,3 |
|
636,"Given the video streams acquired by two cameras placed on either side of a computer monitor , the proposed algorithm synthesises << images >> from a [[ virtual camera ]] in arbitrary position -LRB- typically located within the monitor -RRB- to facilitate eye contact .",636,3 |
|
637,"Given the video streams acquired by two cameras placed on either side of a computer monitor , the proposed algorithm synthesises images from a << virtual camera >> in [[ arbitrary position ]] -LRB- typically located within the monitor -RRB- to facilitate eye contact .",637,1 |
|
638,"Our [[ technique ]] is based on an improved , dynamic-programming , stereo algorithm for efficient << novel-view generation >> .",638,3 |
|
639,"Our << technique >> is based on an improved , [[ dynamic-programming , stereo algorithm ]] for efficient novel-view generation .",639,3 |
|
640,"The two main contributions of this paper are : i -RRB- a new type of [[ three-plane graph ]] for << dense-stereo dynamic-programming >> , that encourages correct occlusion labeling ; ii -RRB- a compact geometric derivation for novel-view synthesis by direct projection of the minimum-cost surface .",640,3 |
|
641,"The two main contributions of this paper are : i -RRB- a new type of three-plane graph for [[ dense-stereo dynamic-programming ]] , that encourages correct << occlusion labeling >> ; ii -RRB- a compact geometric derivation for novel-view synthesis by direct projection of the minimum-cost surface .",641,3 |
|
642,"The two main contributions of this paper are : i -RRB- a new type of three-plane graph for dense-stereo dynamic-programming , that encourages correct occlusion labeling ; ii -RRB- a [[ compact geometric derivation ]] for << novel-view synthesis >> by direct projection of the minimum-cost surface .",642,3 |
|
643,"The two main contributions of this paper are : i -RRB- a new type of three-plane graph for dense-stereo dynamic-programming , that encourages correct occlusion labeling ; ii -RRB- a << compact geometric derivation >> for novel-view synthesis by [[ direct projection of the minimum-cost surface ]] .",643,3 |
|
644,"Furthermore , this paper presents a novel [[ algorithm ]] for the << temporal maintenance of a background model >> to enhance the rendering of occlusions and reduce temporal artefacts -LRB- flicker -RRB- ; and a cost aggregation algorithm that acts directly on our three-dimensional matching cost space .",644,3 |
|
645,"Furthermore , this paper presents a novel [[ algorithm ]] for the temporal maintenance of a background model to enhance the << rendering of occlusions >> and reduce temporal artefacts -LRB- flicker -RRB- ; and a cost aggregation algorithm that acts directly on our three-dimensional matching cost space .",645,3 |
|
646,"Furthermore , this paper presents a novel [[ algorithm ]] for the temporal maintenance of a background model to enhance the rendering of occlusions and reduce << temporal artefacts -LRB- flicker -RRB- >> ; and a cost aggregation algorithm that acts directly on our three-dimensional matching cost space .",646,3 |
|
647,"Furthermore , this paper presents a novel << algorithm >> for the temporal maintenance of a background model to enhance the rendering of occlusions and reduce temporal artefacts -LRB- flicker -RRB- ; and a [[ cost aggregation algorithm ]] that acts directly on our three-dimensional matching cost space .",647,0 |
|
648,"Furthermore , this paper presents a novel algorithm for the temporal maintenance of a background model to enhance the rendering of occlusions and reduce temporal artefacts -LRB- flicker -RRB- ; and a [[ cost aggregation algorithm ]] that acts directly on our << three-dimensional matching cost space >> .",648,3 |
|
649,Examples are given that demonstrate the [[ robustness ]] of the new << algorithm >> to spatial and temporal artefacts for long stereo video streams .,649,6 |
|
650,Examples are given that demonstrate the robustness of the new [[ algorithm ]] to << spatial and temporal artefacts >> for long stereo video streams .,650,3 |
|
651,Examples are given that demonstrate the robustness of the new algorithm to [[ spatial and temporal artefacts ]] for << long stereo video streams >> .,651,3 |
|
652,We further demonstrate << synthesis >> from a freely [[ translating virtual camera ]] .,652,3 |
|
653,"To a large extent , these statistics reflect [[ semantic constraints ]] and thus are used to disambiguate << anaphora references >> and syntactic ambiguities .",653,3 |
|
654,"To a large extent , these statistics reflect [[ semantic constraints ]] and thus are used to disambiguate anaphora references and << syntactic ambiguities >> .",654,3 |
|
655,"To a large extent , these statistics reflect semantic constraints and thus are used to disambiguate [[ anaphora references ]] and << syntactic ambiguities >> .",655,0 |
|
656,The results of the experiment show that in most of the cases the [[ cooccurrence statistics ]] indeed reflect the semantic constraints and thus provide a basis for a useful << disambiguation tool >> .,656,3 |
|
657,"We present a novel [[ method ]] for << discovering parallel sentences >> in comparable , non-parallel corpora .",657,3 |
|
658,"We present a novel method for << discovering parallel sentences >> in [[ comparable , non-parallel corpora ]] .",658,3 |
|
659,"Using this [[ approach ]] , we extract << parallel data >> from large Chinese , Arabic , and English non-parallel newspaper corpora .",659,3 |
|
660,"Using this approach , we extract [[ parallel data ]] from large << Chinese , Arabic , and English non-parallel newspaper corpora >> .",660,4 |
|
661,We evaluate the quality of the extracted data by showing that [[ it ]] improves the performance of a state-of-the-art << statistical machine translation system >> .,661,3 |
|
662,"We also show that a good-quality << MT system >> can be built from scratch by starting with a very small [[ parallel corpus ]] -LRB- 100,000 words -RRB- and exploiting a large non-parallel corpus .",662,3 |
|
663,"We also show that a good-quality MT system can be built from scratch by starting with a very small [[ parallel corpus ]] -LRB- 100,000 words -RRB- and exploiting a large << non-parallel corpus >> .",663,0 |
|
664,"We also show that a good-quality << MT system >> can be built from scratch by starting with a very small parallel corpus -LRB- 100,000 words -RRB- and exploiting a large [[ non-parallel corpus ]] .",664,3 |
|
665,"Thus , our << method >> can be applied with great benefit to language pairs for which only [[ scarce resources ]] are available .",665,3 |
|
666,"In this paper , we describe a [[ search procedure ]] for << statistical machine translation -LRB- MT -RRB- >> based on dynamic programming -LRB- DP -RRB- .",666,3 |
|
667,"In this paper , we describe a search procedure for << statistical machine translation -LRB- MT -RRB- >> based on [[ dynamic programming -LRB- DP -RRB- ]] .",667,3 |
|
668,"Starting from a DP-based solution to the traveling salesman problem , we present a novel [[ technique ]] to restrict the possible word reordering between source and target language in order to achieve an efficient << search algorithm >> .",668,3 |
|
669,"The experimental tests are carried out on the [[ Verbmobil task ]] -LRB- German-English , 8000-word vocabulary -RRB- , which is a << limited-domain spoken-language task >> .",669,2 |
|
670,"A purely functional implementation of << LR-parsers >> is given , together with a simple [[ correctness proof ]] .",670,0 |
|
671,<< It >> is presented as a generalization of the [[ recursive descent parser ]] .,671,3 |
|
672,"For non-LR grammars the [[ time-complexity ]] of our << parser >> is cubic if the functions that constitute the parser are implemented as memo-functions , i.e. functions that memorize the results of previous invocations .",672,6 |
|
673,"For << non-LR grammars >> the time-complexity of our [[ parser ]] is cubic if the functions that constitute the parser are implemented as memo-functions , i.e. functions that memorize the results of previous invocations .",673,3 |
|
674,"For non-LR grammars the time-complexity of our parser is cubic if the functions that constitute the << parser >> are implemented as [[ memo-functions ]] , i.e. functions that memorize the results of previous invocations .",674,3 |
|
675,[[ Memo-functions ]] also facilitate a simple way to construct a very compact representation of the << parse forest >> .,675,3 |
|
676,"For << LR -LRB- 0 -RRB- grammars >> , our [[ algorithm ]] is closely related to the recursive ascent parsers recently discovered by Kruse-man Aretz -LSB- 1 -RSB- and Roberts -LSB- 2 -RSB- .",676,3 |
|
677,"For LR -LRB- 0 -RRB- grammars , our [[ algorithm ]] is closely related to the << recursive ascent parsers >> recently discovered by Kruse-man Aretz -LSB- 1 -RSB- and Roberts -LSB- 2 -RSB- .",677,0 |
|
678,Extended CF grammars -LRB- << grammars >> with [[ regular expressions ]] at the right hand side -RRB- can be parsed with a simple modification of the LR-parser for normal CF grammars .,678,1 |
|
679,<< Extended CF grammars >> -LRB- grammars with regular expressions at the right hand side -RRB- can be parsed with a simple modification of the [[ LR-parser ]] for normal CF grammars .,679,3 |
|
680,Extended CF grammars -LRB- grammars with regular expressions at the right hand side -RRB- can be parsed with a simple modification of the [[ LR-parser ]] for normal << CF grammars >> .,680,3 |
|
681,"In this theory , << discourse structure >> is composed of three separate but interrelated [[ components ]] : the structure of the sequence of utterances -LRB- called the linguistic structure -RRB- , a structure of purposes -LRB- called the intentional structure -RRB- , and the state of focus of attention -LRB- called the attentional state -RRB- .",681,4 |
|
682,"In this theory , discourse structure is composed of three separate but interrelated << components >> : the structure of the sequence of utterances -LRB- called the [[ linguistic structure ]] -RRB- , a structure of purposes -LRB- called the intentional structure -RRB- , and the state of focus of attention -LRB- called the attentional state -RRB- .",682,4 |
|
683,"In this theory , discourse structure is composed of three separate but interrelated components : the structure of the sequence of utterances -LRB- called the [[ linguistic structure ]] -RRB- , a structure of purposes -LRB- called the << intentional structure >> -RRB- , and the state of focus of attention -LRB- called the attentional state -RRB- .",683,0 |
|
684,"In this theory , discourse structure is composed of three separate but interrelated << components >> : the structure of the sequence of utterances -LRB- called the linguistic structure -RRB- , a structure of purposes -LRB- called the [[ intentional structure ]] -RRB- , and the state of focus of attention -LRB- called the attentional state -RRB- .",684,4 |
|
685,"In this theory , discourse structure is composed of three separate but interrelated components : the structure of the sequence of utterances -LRB- called the linguistic structure -RRB- , a structure of purposes -LRB- called the [[ intentional structure ]] -RRB- , and the state of focus of attention -LRB- called the << attentional state >> -RRB- .",685,0 |
|
686,"In this theory , discourse structure is composed of three separate but interrelated << components >> : the structure of the sequence of utterances -LRB- called the linguistic structure -RRB- , a structure of purposes -LRB- called the intentional structure -RRB- , and the state of focus of attention -LRB- called the [[ attentional state ]] -RRB- .",686,4 |
|
687,"The [[ intentional structure ]] captures the << discourse-relevant purposes >> , expressed in each of the linguistic segments as well as relationships among them .",687,3 |
|
688,"The distinction among these components is essential to provide an adequate explanation of such << discourse phenomena >> as [[ cue phrases ]] , referring expressions , and interruptions .",688,2 |
|
689,"The distinction among these components is essential to provide an adequate explanation of such discourse phenomena as [[ cue phrases ]] , << referring expressions >> , and interruptions .",689,0 |
|
690,"The distinction among these components is essential to provide an adequate explanation of such << discourse phenomena >> as cue phrases , [[ referring expressions ]] , and interruptions .",690,2 |
|
691,"The distinction among these components is essential to provide an adequate explanation of such discourse phenomena as cue phrases , [[ referring expressions ]] , and << interruptions >> .",691,0 |
|
692,"The distinction among these components is essential to provide an adequate explanation of such << discourse phenomena >> as cue phrases , referring expressions , and [[ interruptions ]] .",692,2 |
|
693,We examine the relationship between the two << grammatical formalisms >> : [[ Tree Adjoining Grammars ]] and Head Grammars .,693,2 |
|
694,We examine the relationship between the two grammatical formalisms : [[ Tree Adjoining Grammars ]] and << Head Grammars >> .,694,5 |
|
695,We examine the relationship between the two << grammatical formalisms >> : Tree Adjoining Grammars and [[ Head Grammars ]] .,695,2 |
|
696,We then turn to a discussion comparing the [[ linguistic expressiveness ]] of the two << formalisms >> .,696,1 |
|
697,We provide a unified account of << sentence-level and text-level anaphora >> within the framework of a [[ dependency-based grammar model ]] .,697,3 |
|
698,"[[ Criteria ]] for << anaphora resolution within sentence boundaries >> rephrase major concepts from GB 's binding theory , while those for text-level anaphora incorporate an adapted version of a Grosz-Sidner-style focus model .",698,3 |
|
699,"<< Criteria >> for anaphora resolution within sentence boundaries rephrase major concepts from [[ GB 's binding theory ]] , while those for text-level anaphora incorporate an adapted version of a Grosz-Sidner-style focus model .",699,3 |
|
700,"Criteria for anaphora resolution within sentence boundaries rephrase major concepts from GB 's binding theory , while [[ those ]] for << text-level anaphora >> incorporate an adapted version of a Grosz-Sidner-style focus model .",700,3 |
|
701,"Criteria for anaphora resolution within sentence boundaries rephrase major concepts from GB 's binding theory , while << those >> for text-level anaphora incorporate an adapted version of a [[ Grosz-Sidner-style focus model ]] .",701,4 |
|
702,[[ Coedition ]] of a natural language text and its representation in some interlingual form seems the best and simplest way to share << text revision >> across languages .,702,3 |
|
703,<< Coedition >> of a [[ natural language text ]] and its representation in some interlingual form seems the best and simplest way to share text revision across languages .,703,3 |
|
704,The modified [[ graph ]] is then sent to the << UNL-L0 deconverter >> and the result shown .,704,3 |
|
705,"On the internal side , << liaisons >> are established between elements of the text and the graph by using broadly available [[ resources ]] such as a LO-English or better a L0-UNL dictionary , a morphosyntactic parser of L0 , and a canonical graph2tree transformation .",705,3 |
|
706,"On the internal side , liaisons are established between elements of the text and the graph by using broadly available << resources >> such as a [[ LO-English or better a L0-UNL dictionary ]] , a morphosyntactic parser of L0 , and a canonical graph2tree transformation .",706,2 |
|
707,"On the internal side , liaisons are established between elements of the text and the graph by using broadly available resources such as a [[ LO-English or better a L0-UNL dictionary ]] , a << morphosyntactic parser of L0 >> , and a canonical graph2tree transformation .",707,0 |
|
708,"On the internal side , liaisons are established between elements of the text and the graph by using broadly available << resources >> such as a LO-English or better a L0-UNL dictionary , a [[ morphosyntactic parser of L0 ]] , and a canonical graph2tree transformation .",708,2 |
|
709,"On the internal side , liaisons are established between elements of the text and the graph by using broadly available resources such as a LO-English or better a L0-UNL dictionary , a [[ morphosyntactic parser of L0 ]] , and a << canonical graph2tree transformation >> .",709,0 |
|
710,"On the internal side , liaisons are established between elements of the text and the graph by using broadly available << resources >> such as a LO-English or better a L0-UNL dictionary , a morphosyntactic parser of L0 , and a [[ canonical graph2tree transformation ]] .",710,2 |
|
711,"Establishing a `` best '' correspondence between the '' [[ UNL-tree + L0 ]] '' and the '' << MS-L0 structure >> '' , a lattice , may be done using the dictionary and trying to align the tree and the selected trajectory with as few crossing liaisons as possible .",711,0 |
|
712,"Establishing a `` best '' correspondence between the '' UNL-tree + L0 '' and the '' MS-L0 structure '' , a << lattice >> , may be done using the [[ dictionary ]] and trying to align the tree and the selected trajectory with as few crossing liaisons as possible .",712,3 |
|
713,"A central goal of this research is to merge approaches from [[ pivot MT ]] , << interactive MT >> , and multilingual text authoring .",713,0 |
|
714,"A central goal of this research is to merge approaches from pivot MT , [[ interactive MT ]] , and << multilingual text authoring >> .",714,0 |
|
715,We report experiments conducted on a [[ multilingual corpus ]] to estimate the number of << analogies >> among the sentences that it contains .,715,6 |
|
716,Our goal is to learn a << Mahalanobis distance >> by minimizing a [[ loss ]] defined on the weighted sum of the precision at different ranks .,716,3 |
|
717,Our goal is to learn a Mahalanobis distance by minimizing a loss defined on the [[ weighted sum ]] of the << precision >> at different ranks .,717,1 |
|
718,Our core motivation is that minimizing a [[ weighted rank loss ]] is a natural criterion for many problems in << computer vision >> such as person re-identification .,718,3 |
|
719,Our core motivation is that minimizing a [[ weighted rank loss ]] is a natural criterion for many problems in computer vision such as << person re-identification >> .,719,3 |
|
720,Our core motivation is that minimizing a weighted rank loss is a natural criterion for many problems in << computer vision >> such as [[ person re-identification ]] .,720,2 |
|
721,We propose a novel << metric learning formulation >> called [[ Weighted Approximate Rank Component Analysis -LRB- WARCA -RRB- ]] .,721,2 |
|
722,We then derive a scalable [[ stochastic gradient descent algorithm ]] for the resulting << learning problem >> .,722,3 |
|
723,We also derive an efficient << non-linear extension of WARCA >> by using the [[ kernel trick ]] .,723,3 |
|
724,[[ Kernel space embedding ]] decouples the training and prediction costs from the data dimension and enables us to plug << inarbitrary distance measures >> which are more natural for the features .,724,3 |
|
725,We also address a more general problem of [[ matrix rank degeneration ]] & << non-isolated minima >> in the low-rank matrix optimization by using new type of regularizer which approximately enforces the or-thonormality of the learned matrix very efficiently .,725,0 |
|
726,We also address a more general problem of [[ matrix rank degeneration ]] & non-isolated minima in the << low-rank matrix optimization >> by using new type of regularizer which approximately enforces the or-thonormality of the learned matrix very efficiently .,726,1 |
|
727,We also address a more general problem of matrix rank degeneration & [[ non-isolated minima ]] in the << low-rank matrix optimization >> by using new type of regularizer which approximately enforces the or-thonormality of the learned matrix very efficiently .,727,1 |
|
728,We also address a more general problem of matrix rank degeneration & non-isolated minima in the << low-rank matrix optimization >> by using new type of [[ regularizer ]] which approximately enforces the or-thonormality of the learned matrix very efficiently .,728,3 |
|
729,We also address a more general problem of matrix rank degeneration & non-isolated minima in the low-rank matrix optimization by using new type of [[ regularizer ]] which approximately enforces the << or-thonormality >> of the learned matrix very efficiently .,729,3 |
|
730,We also address a more general problem of matrix rank degeneration & non-isolated minima in the low-rank matrix optimization by using new type of regularizer which approximately enforces the [[ or-thonormality ]] of the << learned matrix >> very efficiently .,730,1 |
|
731,We validate this new << method >> on nine standard [[ person re-identification datasets ]] including two large scale Market-1501 and CUHK03 datasets and show that we improve upon the current state-of-the-art methods on all of them .,731,6 |
|
732,We validate this new method on nine standard << person re-identification datasets >> including two large [[ scale Market-1501 ]] and CUHK03 datasets and show that we improve upon the current state-of-the-art methods on all of them .,732,2 |
|
733,We validate this new method on nine standard << person re-identification datasets >> including two large scale Market-1501 and [[ CUHK03 datasets ]] and show that we improve upon the current state-of-the-art methods on all of them .,733,2 |
|
734,We validate this new method on nine standard person re-identification datasets including two large << scale Market-1501 >> and [[ CUHK03 datasets ]] and show that we improve upon the current state-of-the-art methods on all of them .,734,0 |
|
735,"In this paper , we discuss << language model adaptation methods >> given a [[ word list ]] and a raw corpus .",735,3 |
|
736,"In this paper , we discuss language model adaptation methods given a [[ word list ]] and a << raw corpus >> .",736,0 |
|
737,"In this paper , we discuss << language model adaptation methods >> given a word list and a [[ raw corpus ]] .",737,3 |
|
738,"In this situation , the general [[ method ]] is to segment the << raw corpus >> automatically using a word list , correct the output sentences by hand , and build a model from the segmented corpus .",738,3 |
|
739,"In this situation , the general << method >> is to segment the raw corpus automatically using a [[ word list ]] , correct the output sentences by hand , and build a model from the segmented corpus .",739,3 |
|
740,"In this situation , the general method is to segment the raw corpus automatically using a word list , correct the output sentences by hand , and build a << model >> from the [[ segmented corpus ]] .",740,3 |
|
741,"In the experiments , we used a variety of [[ methods ]] for << preparing a segmented corpus >> and compared the language models by their speech recognition accuracies .",741,3 |
|
742,"In the experiments , we used a variety of methods for preparing a segmented corpus and compared the << language models >> by their [[ speech recognition accuracies ]] .",742,6 |
|
743,Many practical << modeling problems >> involve [[ discrete data ]] that are best represented as draws from multinomial or categorical distributions .,743,3 |
|
744,Many practical << modeling problems >> involve discrete data that are best represented as draws from [[ multinomial or categorical distributions ]] .,744,3 |
|
745,"For example , << nucleotides in a DNA sequence >> , children 's names in a given state and year , and text documents are all commonly modeled with [[ multinomial distributions ]] .",745,3 |
|
746,"For example , nucleotides in a DNA sequence , children 's names in a given state and year , and << text documents >> are all commonly modeled with [[ multinomial distributions ]] .",746,3 |
|
747,"Here , we leverage a [[ logistic stick-breaking representation ]] and recent innovations in Pólya-gamma augmentation to reformu-late the << multinomial distribution >> in terms of latent variables with jointly Gaussian likelihoods , enabling us to take advantage of a host of Bayesian inference techniques for Gaussian models with minimal overhead .",747,3 |
|
748,"Here , we leverage a logistic stick-breaking representation and recent innovations in [[ Pólya-gamma augmentation ]] to reformu-late the << multinomial distribution >> in terms of latent variables with jointly Gaussian likelihoods , enabling us to take advantage of a host of Bayesian inference techniques for Gaussian models with minimal overhead .",748,3 |
|
749,"Here , we leverage a logistic stick-breaking representation and recent innovations in Pólya-gamma augmentation to reformu-late the << multinomial distribution >> in terms of [[ latent variables ]] with jointly Gaussian likelihoods , enabling us to take advantage of a host of Bayesian inference techniques for Gaussian models with minimal overhead .",749,4 |
|
750,"Here , we leverage a logistic stick-breaking representation and recent innovations in Pólya-gamma augmentation to reformu-late the multinomial distribution in terms of << latent variables >> with [[ jointly Gaussian likelihoods ]] , enabling us to take advantage of a host of Bayesian inference techniques for Gaussian models with minimal overhead .",750,1 |
|
751,"Here , we leverage a logistic stick-breaking representation and recent innovations in Pólya-gamma augmentation to reformu-late the multinomial distribution in terms of latent variables with jointly Gaussian likelihoods , enabling us to take advantage of a host of [[ Bayesian inference techniques ]] for << Gaussian models >> with minimal overhead .",751,3 |
|
752,"Here , we leverage a logistic stick-breaking representation and recent innovations in Pólya-gamma augmentation to reformu-late the multinomial distribution in terms of latent variables with jointly Gaussian likelihoods , enabling us to take advantage of a host of Bayesian inference techniques for << Gaussian models >> with [[ minimal overhead ]] .",752,1 |
|
753,"[[ MINPRAN ]] , a new << robust operator >> , nds good ts in data sets where more than 50 % of the points are outliers .",753,2 |
|
754,"Unlike other [[ techniques ]] that handle << large outlier percentages >> , MINPRAN does not rely on a known error bound for the good data .",754,3 |
|
755,"Unlike other [[ techniques ]] that handle large outlier percentages , << MINPRAN >> does not rely on a known error bound for the good data .",755,5 |
|
756,"Based on this , << MINPRAN >> uses [[ random sampling ]] to search for the t and the number of inliers to the t that are least likely to have occurred randomly .",756,3 |
|
757,<< MINPRAN >> 's properties are connrmed experimentally on [[ synthetic data ]] and compare favorably to least median of squares .,757,6 |
|
758,<< MINPRAN >> 's properties are connrmed experimentally on synthetic data and compare favorably to [[ least median of squares ]] .,758,5 |
|
759,Related work applies [[ MINPRAN ]] to << complex range >> and intensity data 23 -RSB- .,759,3 |
|
760,Related work applies [[ MINPRAN ]] to complex range and << intensity data >> 23 -RSB- .,760,3 |
|
761,<< Metagrammatical formalisms >> that combine [[ context-free phrase structure rules ]] and metarules -LRB- MPS grammars -RRB- allow concise statement of generalizations about the syntax of natural languages .,761,4 |
|
762,Metagrammatical formalisms that combine [[ context-free phrase structure rules ]] and << metarules -LRB- MPS grammars -RRB- >> allow concise statement of generalizations about the syntax of natural languages .,762,0 |
|
763,<< Metagrammatical formalisms >> that combine context-free phrase structure rules and [[ metarules -LRB- MPS grammars -RRB- ]] allow concise statement of generalizations about the syntax of natural languages .,763,4 |
|
764,"We evaluate several proposals for constraining << them >> , basing our assessment on [[ computational tractability and explanatory adequacy ]] .",764,6 |
|
765,"The unique properties of tree-adjoining grammars -LRB- TAG -RRB- present a challenge for the application of [[ TAGs ]] beyond the limited confines of syntax , for instance , to the task of << semantic interpretation >> or automatic translation of natural language .",765,3 |
|
766,"The unique properties of tree-adjoining grammars -LRB- TAG -RRB- present a challenge for the application of [[ TAGs ]] beyond the limited confines of syntax , for instance , to the task of semantic interpretation or << automatic translation of natural language >> .",766,3 |
|
767,"The unique properties of tree-adjoining grammars -LRB- TAG -RRB- present a challenge for the application of TAGs beyond the limited confines of syntax , for instance , to the task of [[ semantic interpretation ]] or << automatic translation of natural language >> .",767,0 |
|
768,"The formalism 's intended usage is to relate expressions of natural languages to their associated << semantics >> represented in a [[ logical form language ]] , or to their translates in another natural language ; in summary , we intend it to allow TAGs to be used beyond their role in syntax proper .",768,3 |
|
769,"The formalism 's intended usage is to relate expressions of natural languages to their associated semantics represented in a logical form language , or to their translates in another natural language ; in summary , we intend it to allow [[ TAGs ]] to be used beyond their role in << syntax proper >> .",769,3 |
|
770,A [[ model-based approach ]] to << on-line cursive handwriting analysis and recognition >> is presented and evaluated .,770,3 |
|
771,"In this [[ model ]] , << on-line handwriting >> is considered as a modulation of a simple cycloidal pen motion , described by two coupled oscillations with a constant linear drift along the line of the writing .",771,3 |
|
772,"In this model , [[ on-line handwriting ]] is considered as a modulation of a simple << cycloidal pen motion >> , described by two coupled oscillations with a constant linear drift along the line of the writing .",772,4 |
|
773,A general procedure for the estimation and quantization of these [[ cycloidal motion parameters ]] for << arbitrary handwriting >> is presented .,773,3 |
|
774,"The result is a [[ discrete motor control representation ]] of the << continuous pen motion >> , via the quantized levels of the model parameters .",774,3 |
|
775,This [[ motor control representation ]] enables successful << word spotting >> and matching of cursive scripts .,775,3 |
|
776,This [[ motor control representation ]] enables successful word spotting and << matching of cursive scripts >> .,776,3 |
|
777,This motor control representation enables successful [[ word spotting ]] and << matching of cursive scripts >> .,777,0 |
|
778,Our experiments clearly indicate the potential of this [[ dynamic representation ]] for complete << cursive handwriting recognition >> .,778,3 |
|
779,"In the << Object Recognition task >> , there exists a di-chotomy between the [[ categorization of objects ]] and estimating object pose , where the former necessitates a view-invariant representation , while the latter requires a representation capable of capturing pose information over different categories of objects .",779,4 |
|
780,"In the Object Recognition task , there exists a di-chotomy between the [[ categorization of objects ]] and << estimating object pose >> , where the former necessitates a view-invariant representation , while the latter requires a representation capable of capturing pose information over different categories of objects .",780,0 |
|
781,"In the << Object Recognition task >> , there exists a di-chotomy between the categorization of objects and [[ estimating object pose ]] , where the former necessitates a view-invariant representation , while the latter requires a representation capable of capturing pose information over different categories of objects .",781,4 |
|
782,"In the Object Recognition task , there exists a di-chotomy between the categorization of objects and estimating object pose , where the << former >> necessitates a [[ view-invariant representation ]] , while the latter requires a representation capable of capturing pose information over different categories of objects .",782,3 |
|
783,"In the Object Recognition task , there exists a di-chotomy between the categorization of objects and estimating object pose , where the former necessitates a view-invariant representation , while the << latter >> requires a [[ representation ]] capable of capturing pose information over different categories of objects .",783,3 |
|
784,"In the Object Recognition task , there exists a di-chotomy between the categorization of objects and estimating object pose , where the former necessitates a view-invariant representation , while the latter requires a [[ representation ]] capable of capturing << pose information >> over different categories of objects .",784,3 |
|
785,"With the rise of [[ deep archi-tectures ]] , the prime focus has been on << object category recognition >> .",785,3 |
|
786,"In contrast , << object pose estimation >> using these [[ approaches ]] has received relatively less attention .",786,3 |
|
787,"In this work , we study how [[ Convolutional Neural Networks -LRB- CNN -RRB- architectures ]] can be adapted to the task of simultaneous << object recognition >> and pose estimation .",787,3 |
|
788,"In this work , we study how [[ Convolutional Neural Networks -LRB- CNN -RRB- architectures ]] can be adapted to the task of simultaneous object recognition and << pose estimation >> .",788,3 |
|
789,"In this work , we study how Convolutional Neural Networks -LRB- CNN -RRB- architectures can be adapted to the task of simultaneous [[ object recognition ]] and << pose estimation >> .",789,0 |
|
790,We investigate and analyze the [[ layers ]] of various << CNN models >> and extensively compare between them with the goal of discovering how the layers of distributed representations within CNNs represent object pose information and how this contradicts with object category representations .,790,4 |
|
791,We investigate and analyze the layers of various CNN models and extensively compare between them with the goal of discovering how the [[ layers of distributed representations ]] within << CNNs >> represent object pose information and how this contradicts with object category representations .,791,4 |
|
792,We investigate and analyze the layers of various CNN models and extensively compare between them with the goal of discovering how the [[ layers of distributed representations ]] within CNNs represent << object pose information >> and how this contradicts with object category representations .,792,3 |
|
793,We investigate and analyze the layers of various CNN models and extensively compare between them with the goal of discovering how the layers of distributed representations within CNNs represent object pose information and how [[ this ]] contradicts with << object category representations >> .,793,5 |
|
794,[[ It ]] is particularly valuable to << empirical MT research >> .,794,3 |
|
795,"In this paper , we explore [[ geometric structures of 3D lines ]] in ray space for improving << light field triangulation >> and stereo matching .",795,3 |
|
796,"In this paper , we explore [[ geometric structures of 3D lines ]] in ray space for improving light field triangulation and << stereo matching >> .",796,3 |
|
797,"In this paper , we explore << geometric structures of 3D lines >> in [[ ray space ]] for improving light field triangulation and stereo matching .",797,1 |
|
798,"In this paper , we explore geometric structures of 3D lines in ray space for improving [[ light field triangulation ]] and << stereo matching >> .",798,0 |
|
799,Such a [[ triangulation ]] provides a << piecewise-linear interpolant >> useful for light field super-resolution .,799,3 |
|
800,Such a triangulation provides a [[ piecewise-linear interpolant ]] useful for << light field super-resolution >> .,800,3 |
|
801,Experiments on [[ synthetic and real data ]] show that both our << triangulation and LAGC algorithms >> outperform state-of-the-art solutions in accuracy and visual quality .,801,6 |
|
802,Experiments on [[ synthetic and real data ]] show that both our triangulation and LAGC algorithms outperform << state-of-the-art solutions >> in accuracy and visual quality .,802,6 |
|
803,Experiments on synthetic and real data show that both our [[ triangulation and LAGC algorithms ]] outperform << state-of-the-art solutions >> in accuracy and visual quality .,803,5 |
|
804,Experiments on synthetic and real data show that both our << triangulation and LAGC algorithms >> outperform state-of-the-art solutions in [[ accuracy ]] and visual quality .,804,6 |
|
805,Experiments on synthetic and real data show that both our triangulation and LAGC algorithms outperform << state-of-the-art solutions >> in [[ accuracy ]] and visual quality .,805,6 |
|
806,Experiments on synthetic and real data show that both our << triangulation and LAGC algorithms >> outperform state-of-the-art solutions in accuracy and [[ visual quality ]] .,806,6 |
|
807,Experiments on synthetic and real data show that both our triangulation and LAGC algorithms outperform << state-of-the-art solutions >> in accuracy and [[ visual quality ]] .,807,6 |
|
808,"This paper presents a << phrase-based statistical machine translation method >> , based on [[ non-contiguous phrases ]] , i.e. phrases with gaps .",808,3 |
|
809,A [[ method ]] for producing such << phrases >> from a word-aligned corpora is proposed .,809,3 |
|
810,A << method >> for producing such phrases from a [[ word-aligned corpora ]] is proposed .,810,6 |
|
811,"A [[ statistical translation model ]] is also presented that deals such << phrases >> , as well as a training method based on the maximization of translation accuracy , as measured with the NIST evaluation metric .",811,3 |
|
812,"A statistical translation model is also presented that deals such phrases , as well as a << training method >> based on the [[ maximization of translation accuracy ]] , as measured with the NIST evaluation metric .",812,3 |
|
813,"A << statistical translation model >> is also presented that deals such phrases , as well as a training method based on the maximization of translation accuracy , as measured with the [[ NIST evaluation metric ]] .",813,6 |
|
814,<< Translations >> are produced by means of a [[ beam-search decoder ]] .,814,3 |
|
815,[[ GLOSSER ]] is designed to support << reading and learning >> to read in a foreign language .,815,3 |
|
816,"There are four [[ language pairs ]] currently supported by << GLOSSER >> : English-Bulgarian , English-Estonian , English-Hungarian and French-Dutch .",816,3 |
|
817,"There are four << language pairs >> currently supported by GLOSSER : [[ English-Bulgarian ]] , English-Estonian , English-Hungarian and French-Dutch .",817,2 |
|
818,"There are four language pairs currently supported by GLOSSER : [[ English-Bulgarian ]] , << English-Estonian >> , English-Hungarian and French-Dutch .",818,0 |
|
819,"There are four << language pairs >> currently supported by GLOSSER : English-Bulgarian , [[ English-Estonian ]] , English-Hungarian and French-Dutch .",819,2 |
|
820,"There are four language pairs currently supported by GLOSSER : English-Bulgarian , [[ English-Estonian ]] , << English-Hungarian >> and French-Dutch .",820,0 |
|
821,"There are four << language pairs >> currently supported by GLOSSER : English-Bulgarian , English-Estonian , [[ English-Hungarian ]] and French-Dutch .",821,2 |
|
822,"There are four language pairs currently supported by GLOSSER : English-Bulgarian , English-Estonian , [[ English-Hungarian ]] and << French-Dutch >> .",822,0 |
|
823,"There are four << language pairs >> currently supported by GLOSSER : English-Bulgarian , English-Estonian , English-Hungarian and [[ French-Dutch ]] .",823,2 |
|
824,"A demonstration -LRB- in UNIX -RRB- for Applied Natural Language Processing emphasizes [[ components ]] put to novel technical uses in << intelligent computer-assisted morphological analysis -LRB- ICALL -RRB- >> , including disambiguated morphological analysis and lemmatized indexing for an aligned bilingual corpus of word examples .",824,3 |
|
825,"A demonstration -LRB- in UNIX -RRB- for Applied Natural Language Processing emphasizes << components >> put to novel technical uses in intelligent computer-assisted morphological analysis -LRB- ICALL -RRB- , including [[ disambiguated morphological analysis ]] and lemmatized indexing for an aligned bilingual corpus of word examples .",825,2 |
|
826,"A demonstration -LRB- in UNIX -RRB- for Applied Natural Language Processing emphasizes components put to novel technical uses in intelligent computer-assisted morphological analysis -LRB- ICALL -RRB- , including [[ disambiguated morphological analysis ]] and << lemmatized indexing >> for an aligned bilingual corpus of word examples .",826,0 |
|
827,"A demonstration -LRB- in UNIX -RRB- for Applied Natural Language Processing emphasizes components put to novel technical uses in intelligent computer-assisted morphological analysis -LRB- ICALL -RRB- , including [[ disambiguated morphological analysis ]] and lemmatized indexing for an << aligned bilingual corpus >> of word examples .",827,3 |
|
828,"A demonstration -LRB- in UNIX -RRB- for Applied Natural Language Processing emphasizes << components >> put to novel technical uses in intelligent computer-assisted morphological analysis -LRB- ICALL -RRB- , including disambiguated morphological analysis and [[ lemmatized indexing ]] for an aligned bilingual corpus of word examples .",828,2 |
|
829,"A demonstration -LRB- in UNIX -RRB- for Applied Natural Language Processing emphasizes components put to novel technical uses in intelligent computer-assisted morphological analysis -LRB- ICALL -RRB- , including disambiguated morphological analysis and [[ lemmatized indexing ]] for an << aligned bilingual corpus >> of word examples .",829,3 |
|
830,"We present a new << part-of-speech tagger >> that demonstrates the following ideas : -LRB- i -RRB- explicit use of both preceding and following [[ tag contexts ]] via a dependency network representation , -LRB- ii -RRB- broad use of lexical features , including jointly conditioning on multiple consecutive words , -LRB- iii -RRB- effective use of priors in conditional loglinear models , and -LRB- iv -RRB- fine-grained modeling of unknown word features .",830,3 |
|
831,"We present a new part-of-speech tagger that demonstrates the following ideas : -LRB- i -RRB- explicit use of both preceding and following << tag contexts >> via a [[ dependency network representation ]] , -LRB- ii -RRB- broad use of lexical features , including jointly conditioning on multiple consecutive words , -LRB- iii -RRB- effective use of priors in conditional loglinear models , and -LRB- iv -RRB- fine-grained modeling of unknown word features .",831,3 |
|
832,"We present a new << part-of-speech tagger >> that demonstrates the following ideas : -LRB- i -RRB- explicit use of both preceding and following tag contexts via a dependency network representation , -LRB- ii -RRB- broad use of [[ lexical features ]] , including jointly conditioning on multiple consecutive words , -LRB- iii -RRB- effective use of priors in conditional loglinear models , and -LRB- iv -RRB- fine-grained modeling of unknown word features .",832,3 |
|
833,"We present a new << part-of-speech tagger >> that demonstrates the following ideas : -LRB- i -RRB- explicit use of both preceding and following tag contexts via a dependency network representation , -LRB- ii -RRB- broad use of lexical features , including jointly conditioning on multiple consecutive words , -LRB- iii -RRB- effective use of [[ priors in conditional loglinear models ]] , and -LRB- iv -RRB- fine-grained modeling of unknown word features .",833,3 |
|
834,"We present a new << part-of-speech tagger >> that demonstrates the following ideas : -LRB- i -RRB- explicit use of both preceding and following tag contexts via a dependency network representation , -LRB- ii -RRB- broad use of lexical features , including jointly conditioning on multiple consecutive words , -LRB- iii -RRB- effective use of priors in conditional loglinear models , and -LRB- iv -RRB- [[ fine-grained modeling of unknown word features ]] .",834,3 |
|
835,"Using these ideas together , the resulting << tagger >> gives a 97.24 % [[ accuracy ]] on the Penn Treebank WSJ , an error reduction of 4.4 % on the best previous single automatically learned tagging result .",835,6 |
|
836,"Using these ideas together , the resulting << tagger >> gives a 97.24 % accuracy on the [[ Penn Treebank WSJ ]] , an error reduction of 4.4 % on the best previous single automatically learned tagging result .",836,6 |
|
837,"Using these ideas together , the resulting << tagger >> gives a 97.24 % accuracy on the Penn Treebank WSJ , an [[ error ]] reduction of 4.4 % on the best previous single automatically learned tagging result .",837,6 |
|
838,"Owing to these variations , the << pedestrian data >> is distributed as [[ highly-curved manifolds ]] in the feature space , despite the current convolutional neural networks -LRB- CNN -RRB- 's capability of feature extraction .",838,3 |
|
839,"Owing to these variations , the pedestrian data is distributed as << highly-curved manifolds >> in the [[ feature space ]] , despite the current convolutional neural networks -LRB- CNN -RRB- 's capability of feature extraction .",839,1 |
|
840,"Owing to these variations , the pedestrian data is distributed as highly-curved manifolds in the feature space , despite the current [[ convolutional neural networks -LRB- CNN -RRB- ]] 's capability of << feature extraction >> .",840,3 |
|
841,"In practice , the current << deep embedding methods >> use the [[ Euclidean distance ]] for the training and test .",841,3 |
|
842,"On the other hand , the << manifold learning methods >> suggest to use the [[ Euclidean distance ]] in the local range , combining with the graphical relationship between samples , for approximating the geodesic distance .",842,3 |
|
843,"On the other hand , the manifold learning methods suggest to use the [[ Euclidean distance ]] in the local range , combining with the << graphical relationship >> between samples , for approximating the geodesic distance .",843,0 |
|
844,"On the other hand , the manifold learning methods suggest to use the [[ Euclidean distance ]] in the local range , combining with the graphical relationship between samples , for approximating the << geodesic distance >> .",844,3 |
|
845,"On the other hand , the manifold learning methods suggest to use the << Euclidean distance >> in the [[ local range ]] , combining with the graphical relationship between samples , for approximating the geodesic distance .",845,1 |
|
846,"On the other hand , the manifold learning methods suggest to use the Euclidean distance in the local range , combining with the [[ graphical relationship ]] between samples , for approximating the << geodesic distance >> .",846,3 |
|
847,"From this point of view , selecting suitable positive -LRB- i.e. intra-class -RRB- training samples within a local range is critical for training the CNN embedding , especially when the << data >> has large [[ intra-class variations ]] .",847,1 |
|
848,"In this paper , we propose a novel [[ moderate positive sample mining method ]] to train << robust CNN >> for person re-identification , dealing with the problem of large variation .",848,3 |
|
849,"In this paper , we propose a novel moderate positive sample mining method to train [[ robust CNN ]] for << person re-identification >> , dealing with the problem of large variation .",849,3 |
|
850,"In addition , we improve the << learning >> by a [[ metric weight constraint ]] , so that the learned metric has a better generalization ability .",850,3 |
|
851,"In addition , we improve the learning by a metric weight constraint , so that the << learned metric >> has a better [[ generalization ability ]] .",851,1 |
|
852,"Experiments show that these two strategies are effective in learning [[ robust deep metrics ]] for << person re-identification >> , and accordingly our deep model significantly outperforms the state-of-the-art methods on several benchmarks of person re-identification .",852,3 |
|
853,"Experiments show that these two strategies are effective in learning robust deep metrics for person re-identification , and accordingly our [[ deep model ]] significantly outperforms the << state-of-the-art methods >> on several benchmarks of person re-identification .",853,5 |
|
854,"Experiments show that these two strategies are effective in learning robust deep metrics for person re-identification , and accordingly our [[ deep model ]] significantly outperforms the state-of-the-art methods on several benchmarks of << person re-identification >> .",854,3 |
|
855,"Experiments show that these two strategies are effective in learning robust deep metrics for person re-identification , and accordingly our deep model significantly outperforms the [[ state-of-the-art methods ]] on several benchmarks of << person re-identification >> .",855,3 |
|
856,"Therefore , the study presented in this paper may be useful in inspiring new designs of [[ deep models ]] for << person re-identification >> .",856,3 |
|
857,"[[ Utterance Verification -LRB- UV -RRB- ]] is a critical function of an << Automatic Speech Recognition -LRB- ASR -RRB- System >> working on real applications where spontaneous speech , out-of-vocabulary -LRB- OOV -RRB- words and acoustic noises are present .",857,2 |
|
858,In this paper we present a new UV procedure with two major features : a -RRB- [[ Confidence tests ]] are applied to << decoded string hypotheses >> obtained from using word and garbage models that represent OOV words and noises .,858,3 |
|
859,In this paper we present a new UV procedure with two major features : a -RRB- Confidence tests are applied to decoded string hypotheses obtained from using word and garbage models that represent << OOV words >> and [[ noises ]] .,859,0 |
|
860,Thus the [[ ASR system ]] is designed to deal with what we refer to as << Word Spotting >> and Noise Spotting capabilities .,860,3 |
|
861,Thus the [[ ASR system ]] is designed to deal with what we refer to as Word Spotting and << Noise Spotting capabilities >> .,861,3 |
|
862,"b -RRB- The << UV procedure >> is based on three different [[ confidence tests ]] , two based on acoustic measures and one founded on linguistic information , applied in a hierarchical structure .",862,3 |
|
863,"b -RRB- The UV procedure is based on three different [[ confidence tests ]] , two based on acoustic measures and one founded on linguistic information , applied in a << hierarchical structure >> .",863,3 |
|
864,"b -RRB- The UV procedure is based on three different << confidence tests >> , [[ two ]] based on acoustic measures and one founded on linguistic information , applied in a hierarchical structure .",864,2 |
|
865,"b -RRB- The UV procedure is based on three different confidence tests , << two >> based on [[ acoustic measures ]] and one founded on linguistic information , applied in a hierarchical structure .",865,3 |
|
866,"b -RRB- The UV procedure is based on three different << confidence tests >> , two based on acoustic measures and [[ one ]] founded on linguistic information , applied in a hierarchical structure .",866,2 |
|
867,"b -RRB- The UV procedure is based on three different confidence tests , two based on acoustic measures and << one >> founded on [[ linguistic information ]] , applied in a hierarchical structure .",867,3 |
|
868,Experimental results from a real << telephone application >> on a [[ natural number recognition task ]] show an 50 % reduction in recognition errors with a moderate 12 % rejection rate of correct utterances and a low 1.5 % rate of false acceptance .,868,1 |
|
869,Experimental results from a real telephone application on a << natural number recognition task >> show an 50 % reduction in [[ recognition errors ]] with a moderate 12 % rejection rate of correct utterances and a low 1.5 % rate of false acceptance .,869,6 |
|
870,A critical step in [[ encoding sound ]] for << neuronal processing >> occurs when the analog pressure wave is coded into discrete nerve-action potentials .,870,3 |
|
871,A critical step in encoding sound for neuronal processing occurs when the << analog pressure wave >> is coded into [[ discrete nerve-action potentials ]] .,871,3 |
|
872,"Recent [[ pool models ]] of the << inner hair cell synapse >> do not reproduce the dead time period after an intense stimulus , so we used visual inspection and automatic speech recognition -LRB- ASR -RRB- to investigate an offset adaptation -LRB- OA -RRB- model proposed by Zhang et al. -LSB- 1 -RSB- .",872,3 |
|
873,"Recent pool models of the inner hair cell synapse do not reproduce the dead time period after an intense stimulus , so we used [[ visual inspection ]] and << automatic speech recognition -LRB- ASR -RRB- >> to investigate an offset adaptation -LRB- OA -RRB- model proposed by Zhang et al. -LSB- 1 -RSB- .",873,0 |
|
874,"Recent pool models of the inner hair cell synapse do not reproduce the dead time period after an intense stimulus , so we used [[ visual inspection ]] and automatic speech recognition -LRB- ASR -RRB- to investigate an << offset adaptation -LRB- OA -RRB- model >> proposed by Zhang et al. -LSB- 1 -RSB- .",874,3 |
|
875,"Recent pool models of the inner hair cell synapse do not reproduce the dead time period after an intense stimulus , so we used visual inspection and [[ automatic speech recognition -LRB- ASR -RRB- ]] to investigate an << offset adaptation -LRB- OA -RRB- model >> proposed by Zhang et al. -LSB- 1 -RSB- .",875,3 |
|
876,[[ OA ]] improved << phase locking in the auditory nerve -LRB- AN -RRB- >> and raised ASR accuracy for features derived from AN fibers -LRB- ANFs -RRB- .,876,3 |
|
877,[[ OA ]] improved phase locking in the auditory nerve -LRB- AN -RRB- and raised ASR accuracy for << features >> derived from AN fibers -LRB- ANFs -RRB- .,877,3 |
|
878,OA improved phase locking in the auditory nerve -LRB- AN -RRB- and raised [[ ASR accuracy ]] for << features >> derived from AN fibers -LRB- ANFs -RRB- .,878,6 |
|
879,OA improved phase locking in the auditory nerve -LRB- AN -RRB- and raised ASR accuracy for << features >> derived from [[ AN fibers -LRB- ANFs -RRB- ]] .,879,3 |
|
880,"We also found that [[ OA ]] is crucial for << auditory processing >> by onset neurons -LRB- ONs -RRB- in the next neuronal stage , the auditory brainstem .",880,3 |
|
881,"We also found that << OA >> is crucial for auditory processing by [[ onset neurons -LRB- ONs -RRB- ]] in the next neuronal stage , the auditory brainstem .",881,3 |
|
882,[[ Multi-layer perceptrons -LRB- MLPs -RRB- ]] performed much better than standard << Gaussian mixture models -LRB- GMMs -RRB- >> for both our ANF-based and ON-based auditory features .,882,5 |
|
883,[[ Multi-layer perceptrons -LRB- MLPs -RRB- ]] performed much better than standard Gaussian mixture models -LRB- GMMs -RRB- for both our << ANF-based and ON-based auditory features >> .,883,3 |
|
884,Multi-layer perceptrons -LRB- MLPs -RRB- performed much better than standard [[ Gaussian mixture models -LRB- GMMs -RRB- ]] for both our << ANF-based and ON-based auditory features >> .,884,3 |
|
885,Recent progress in << computer vision >> has been driven by [[ high-capacity models ]] trained on large datasets .,885,3 |
|
886,Recent progress in computer vision has been driven by << high-capacity models >> trained on [[ large datasets ]] .,886,3 |
|
887,"Unfortunately , creating << large datasets >> with [[ pixel-level labels ]] has been extremely costly due to the amount of human effort required .",887,1 |
|
888,"In this paper , we present an [[ approach ]] to rapidly creating << pixel-accurate semantic label maps >> for images extracted from modern computer games .",888,3 |
|
889,"In this paper , we present an approach to rapidly creating [[ pixel-accurate semantic label maps ]] for << images >> extracted from modern computer games .",889,3 |
|
890,"In this paper , we present an approach to rapidly creating pixel-accurate semantic label maps for [[ images ]] extracted from << modern computer games >> .",890,4 |
|
891,"We propose a novel step toward the << unsupervised seg-mentation of whole objects >> by combining '' hints '' of [[ partial scene segmentation ]] offered by multiple soft , binary mattes .",891,3 |
|
892,"We propose a novel step toward the unsupervised seg-mentation of whole objects by combining '' hints '' of << partial scene segmentation >> offered by multiple [[ soft , binary mattes ]] .",892,3 |
|
893,These << mattes >> are implied by a set of [[ hypothesized object boundary fragments ]] in the scene .,893,3 |
|
894,"This reflects [[ contemporary methods ]] for << unsupervised object discovery >> from groups of images , and it allows us to define intuitive evaluation met-rics for our sets of segmentations based on the accurate and parsimonious delineation of scene objects .",894,3 |
|
895,"Our proposed << approach >> builds on recent advances in [[ spectral clustering ]] , image matting , and boundary detection .",895,3 |
|
896,"Our proposed approach builds on recent advances in [[ spectral clustering ]] , << image matting >> , and boundary detection .",896,0 |
|
897,"Our proposed << approach >> builds on recent advances in spectral clustering , [[ image matting ]] , and boundary detection .",897,3 |
|
898,"Our proposed approach builds on recent advances in spectral clustering , [[ image matting ]] , and << boundary detection >> .",898,0 |
|
899,"Our proposed << approach >> builds on recent advances in spectral clustering , image matting , and [[ boundary detection ]] .",899,3 |
|
900,[[ It ]] is demonstrated qualitatively and quantitatively on a dataset of scenes and is suitable for current work in << unsupervised object discovery >> without top-down knowledge .,900,3 |
|
901,<< It >> is demonstrated qualitatively and quantitatively on a [[ dataset of scenes ]] and is suitable for current work in unsupervised object discovery without top-down knowledge .,901,6 |
|
902,[[ Language resource quality ]] is crucial in << NLP >> .,902,1 |
|
903,"Many of the resources used are derived from data created by human beings out of an << NLP >> context , especially regarding [[ MT ]] and reference translations .",903,2 |
|
904,"Many of the resources used are derived from data created by human beings out of an NLP context , especially regarding [[ MT ]] and << reference translations >> .",904,0 |
|
905,"Many of the resources used are derived from data created by human beings out of an << NLP >> context , especially regarding MT and [[ reference translations ]] .",905,2 |
|
906,"Indeed , << automatic evaluations >> need [[ high-quality data ]] that allow the comparison of both automatic and human translations .",906,6 |
|
907,This paper describes the impact of using [[ different-quality references ]] on << evaluation >> .,907,3 |
|
908,"Thus , the limitations of the [[ automatic metrics ]] used within << MT >> are also discussed in this regard .",908,6 |
|
909,"This poster paper describes a [[ full scale two-level morphological description ]] -LRB- Karttunen , 1983 ; Koskenniemi , 1983 -RRB- of << Turkish word structures >> .",909,3 |
|
910,"The << description >> has been implemented using the [[ PC-KIMMO environment ]] -LRB- Antworth , 1990 -RRB- and is based on a root word lexicon of about 23,000 roots words .",910,3 |
|
911,"The << description >> has been implemented using the PC-KIMMO environment -LRB- Antworth , 1990 -RRB- and is based on a [[ root word lexicon ]] of about 23,000 roots words .",911,3 |
|
912,[[ Turkish ]] is an << agglutinative language >> with word structures formed by productive affixations of derivational and inflectional suffixes to root words .,912,2 |
|
913,Turkish is an << agglutinative language >> with [[ word structures ]] formed by productive affixations of derivational and inflectional suffixes to root words .,913,1 |
|
914,Turkish is an agglutinative language with << word structures >> formed by [[ productive affixations of derivational and inflectional suffixes ]] to root words .,914,4 |
|
915,The << surface realizations of morphological constructions >> are constrained and modified by a number of [[ phonetic rules ]] such as vowel harmony .,915,3 |
|
916,The surface realizations of morphological constructions are constrained and modified by a number of << phonetic rules >> such as [[ vowel harmony ]] .,916,2 |
|
917,This paper deals with the problem of generating the [[ fundamental frequency -LRB- F0 -RRB- contour of speech ]] from a text input for << text-to-speech synthesis >> .,917,3 |
|
918,This paper deals with the problem of generating the << fundamental frequency -LRB- F0 -RRB- contour of speech >> from a [[ text input ]] for text-to-speech synthesis .,918,3 |
|
919,"We have previously introduced a [[ statistical model ]] describing the generating process of << speech F0 contours >> , based on the discrete-time version of the Fujisaki model .",919,3 |
|
920,"We have previously introduced a << statistical model >> describing the generating process of speech F0 contours , based on the discrete-time version of the [[ Fujisaki model ]] .",920,3 |
|
921,One [[ remarkable feature ]] of this << model >> is that it has allowed us to derive an efficient algorithm based on powerful statistical methods for estimating the Fujisaki-model parameters from raw F0 contours .,921,1 |
|
922,One remarkable feature of this model is that it has allowed us to derive an efficient [[ algorithm ]] based on powerful statistical methods for estimating the << Fujisaki-model parameters >> from raw F0 contours .,922,3 |
|
923,One remarkable feature of this model is that it has allowed us to derive an efficient << algorithm >> based on powerful [[ statistical methods ]] for estimating the Fujisaki-model parameters from raw F0 contours .,923,3 |
|
924,One remarkable feature of this model is that it has allowed us to derive an efficient algorithm based on powerful statistical methods for estimating the << Fujisaki-model parameters >> from [[ raw F0 contours ]] .,924,3 |
|
925,"To associate a sequence of the << Fujisaki-model parameters >> with a [[ text input ]] based on statistical learning , this paper proposes extending this model to a context-dependent one .",925,3 |
|
926,"To associate a sequence of the << Fujisaki-model parameters >> with a text input based on [[ statistical learning ]] , this paper proposes extending this model to a context-dependent one .",926,3 |
|
927,We further propose a [[ parameter training algorithm ]] for the present << model >> based on a decision tree-based context clustering .,927,3 |
|
928,We further propose a << parameter training algorithm >> for the present model based on a [[ decision tree-based context clustering ]] .,928,3 |
|
929,We introduce a [[ method ]] to accelerate the << evaluation of object detection cascades >> with the help of a divide-and-conquer procedure in the space of candidate regions .,929,3 |
|
930,We introduce a << method >> to accelerate the evaluation of object detection cascades with the help of a [[ divide-and-conquer procedure ]] in the space of candidate regions .,930,3 |
|
931,We introduce a method to accelerate the evaluation of object detection cascades with the help of a << divide-and-conquer procedure >> in the [[ space of candidate regions ]] .,931,1 |
|
932,"Compared to the [[ exhaustive procedure ]] that thus far is the state-of-the-art for << cascade evaluation >> , the proposed method requires fewer evaluations of the classifier functions , thereby speeding up the search .",932,3 |
|
933,"Compared to the [[ exhaustive procedure ]] that thus far is the state-of-the-art for cascade evaluation , the proposed << method >> requires fewer evaluations of the classifier functions , thereby speeding up the search .",933,5 |
|
934,"Compared to the exhaustive procedure that thus far is the state-of-the-art for cascade evaluation , the proposed [[ method ]] requires fewer evaluations of the classifier functions , thereby speeding up the << search >> .",934,3 |
|
935,"Furthermore , we show how the recently developed efficient [[ subwindow search -LRB- ESS -RRB- procedure ]] -LSB- 11 -RSB- can be integrated into the last stage of our << method >> .",935,4 |
|
936,"This allows us to use our [[ method ]] to act not only as a faster procedure for << cascade evaluation >> , but also as a tool to perform efficient branch-and-bound object detection with nonlinear quality functions , in particular kernel-ized support vector machines .",936,3 |
|
937,"This allows us to use our [[ method ]] to act not only as a faster procedure for cascade evaluation , but also as a tool to perform efficient << branch-and-bound object detection >> with nonlinear quality functions , in particular kernel-ized support vector machines .",937,3 |
|
938,"This allows us to use our method to act not only as a faster procedure for cascade evaluation , but also as a tool to perform efficient << branch-and-bound object detection >> with [[ nonlinear quality functions ]] , in particular kernel-ized support vector machines .",938,3 |
|
939,"This allows us to use our method to act not only as a faster procedure for cascade evaluation , but also as a tool to perform efficient branch-and-bound object detection with << nonlinear quality functions >> , in particular [[ kernel-ized support vector machines ]] .",939,2 |
|
940,Experiments on the [[ PASCAL VOC 2006 dataset ]] show an acceleration of more than 50 % by our << method >> compared to standard cascade evaluation .,940,6 |
|
941,Experiments on the [[ PASCAL VOC 2006 dataset ]] show an acceleration of more than 50 % by our method compared to standard << cascade evaluation >> .,941,6 |
|
942,Experiments on the PASCAL VOC 2006 dataset show an acceleration of more than 50 % by our << method >> compared to standard [[ cascade evaluation ]] .,942,5 |
|
943,[[ Background modeling ]] is an important component of many << vision systems >> .,943,4 |
|
944,"When the << scene >> exhibits a [[ persistent dynamic behavior ]] in time , such an assumption is violated and detection performance deteriorates .",944,1 |
|
945,"In this paper , we propose a new [[ method ]] for the << modeling and subtraction of such scenes >> .",945,3 |
|
946,"Towards the << modeling of the dynamic characteristics >> , [[ optical flow ]] is computed and utilized as a feature in a higher dimensional space .",946,3 |
|
947,"Towards the modeling of the dynamic characteristics , [[ optical flow ]] is computed and utilized as a << feature >> in a higher dimensional space .",947,3 |
|
948,"Towards the << modeling of the dynamic characteristics >> , optical flow is computed and utilized as a [[ feature ]] in a higher dimensional space .",948,3 |
|
949,"Towards the modeling of the dynamic characteristics , optical flow is computed and utilized as a << feature >> in a [[ higher dimensional space ]] .",949,1 |
|
950,Inherent [[ ambiguities ]] in the << computation of features >> are addressed by using a data-dependent bandwidth for density estimation using kernels .,950,1 |
|
951,Inherent << ambiguities >> in the computation of features are addressed by using a [[ data-dependent bandwidth ]] for density estimation using kernels .,951,3 |
|
952,Inherent ambiguities in the computation of features are addressed by using a [[ data-dependent bandwidth ]] for << density estimation >> using kernels .,952,3 |
|
953,Inherent ambiguities in the computation of features are addressed by using a data-dependent bandwidth for << density estimation >> using [[ kernels ]] .,953,3 |
|
954,"In this paper , we present our approach for using [[ information extraction annotations ]] to augment << document retrieval for distillation >> .",954,3 |
|
955,This paper presents a novel [[ representation ]] for << three-dimensional objects >> in terms of affine-invariant image patches and their spatial relationships .,955,3 |
|
956,This paper presents a novel representation for << three-dimensional objects >> in terms of [[ affine-invariant image patches ]] and their spatial relationships .,956,1 |
|
957,This paper presents a novel representation for three-dimensional objects in terms of << affine-invariant image patches >> and their [[ spatial relationships ]] .,957,1 |
|
958,"[[ Multi-view constraints ]] associated with groups of patches are combined with a << normalized representation >> of their appearance to guide matching and reconstruction , allowing the acquisition of true three-dimensional affine and Euclidean models from multiple images and their recognition in a single photograph taken from an arbitrary viewpoint .",958,0 |
|
959,"[[ Multi-view constraints ]] associated with groups of patches are combined with a normalized representation of their appearance to guide << matching >> and reconstruction , allowing the acquisition of true three-dimensional affine and Euclidean models from multiple images and their recognition in a single photograph taken from an arbitrary viewpoint .",959,3 |
|
960,"[[ Multi-view constraints ]] associated with groups of patches are combined with a normalized representation of their appearance to guide matching and << reconstruction >> , allowing the acquisition of true three-dimensional affine and Euclidean models from multiple images and their recognition in a single photograph taken from an arbitrary viewpoint .",960,3 |
|
961,"Multi-view constraints associated with groups of patches are combined with a [[ normalized representation ]] of their appearance to guide << matching >> and reconstruction , allowing the acquisition of true three-dimensional affine and Euclidean models from multiple images and their recognition in a single photograph taken from an arbitrary viewpoint .",961,3 |
|
962,"Multi-view constraints associated with groups of patches are combined with a [[ normalized representation ]] of their appearance to guide matching and << reconstruction >> , allowing the acquisition of true three-dimensional affine and Euclidean models from multiple images and their recognition in a single photograph taken from an arbitrary viewpoint .",962,3 |
|
963,"Multi-view constraints associated with groups of patches are combined with a normalized representation of their appearance to guide [[ matching ]] and << reconstruction >> , allowing the acquisition of true three-dimensional affine and Euclidean models from multiple images and their recognition in a single photograph taken from an arbitrary viewpoint .",963,0 |
|
964,"Multi-view constraints associated with groups of patches are combined with a normalized representation of their appearance to guide matching and reconstruction , allowing the << acquisition of true three-dimensional affine and Euclidean models >> from multiple [[ images ]] and their recognition in a single photograph taken from an arbitrary viewpoint .",964,3 |
|
965,The proposed [[ approach ]] does not require a separate segmentation stage and is applicable to << cluttered scenes >> .,965,3 |
|
966,[[ Fast algorithms ]] for << nearest neighbor -LRB- NN -RRB- search >> have in large part focused on 2 distance .,966,3 |
|
967,Here we develop an [[ approach ]] for << 1 distance >> that begins with an explicit and exactly distance-preserving embedding of the points into 2 2 .,967,3 |
|
968,"We show how [[ this ]] can efficiently be combined with << random-projection based methods >> for 2 NN search , such as locality-sensitive hashing -LRB- LSH -RRB- or random projection trees .",968,0 |
|
969,"We show how this can efficiently be combined with [[ random-projection based methods ]] for 2 << NN search >> , such as locality-sensitive hashing -LRB- LSH -RRB- or random projection trees .",969,3 |
|
970,"We show how this can efficiently be combined with << random-projection based methods >> for 2 NN search , such as [[ locality-sensitive hashing -LRB- LSH -RRB- ]] or random projection trees .",970,2 |
|
971,"We show how this can efficiently be combined with random-projection based methods for 2 NN search , such as [[ locality-sensitive hashing -LRB- LSH -RRB- ]] or << random projection trees >> .",971,0 |
|
972,"We show how this can efficiently be combined with << random-projection based methods >> for 2 NN search , such as locality-sensitive hashing -LRB- LSH -RRB- or [[ random projection trees ]] .",972,2 |
|
973,We rigorously establish the correctness of the methodology and show by experimentation using LSH that [[ it ]] is competitive in practice with available << alternatives >> .,973,5 |
|
|
