text
stringlengths
49
577
label
stringclasses
7 values
metadata
list
Finally , a [[ cross-corpus -LRB- and cross-language -RRB- experiment ]] reveals better noise and reverberation robustness for DOCCs than for << MFCCs >> .
EVALUATE-FOR
[ 3, 8, 19, 19 ]
Finally , a cross-corpus -LRB- and cross-language -RRB- experiment reveals better [[ noise and reverberation robustness ]] for << DOCCs >> than for MFCCs .
EVALUATE-FOR
[ 11, 14, 16, 16 ]
Finally , a cross-corpus -LRB- and cross-language -RRB- experiment reveals better [[ noise and reverberation robustness ]] for DOCCs than for << MFCCs >> .
EVALUATE-FOR
[ 11, 14, 19, 19 ]
Finally , a cross-corpus -LRB- and cross-language -RRB- experiment reveals better noise and reverberation robustness for [[ DOCCs ]] than for << MFCCs >> .
COMPARE
[ 16, 16, 19, 19 ]
This paper proposes [[ document oriented preference sets -LRB- DoPS -RRB- ]] for the << disambiguation of the dependency structure >> of sentences .
USED-FOR
[ 3, 9, 12, 16 ]
<< Sentence ambiguities >> can be resolved by using [[ domain targeted preference knowledge ]] without using complicated large knowledgebases .
USED-FOR
[ 7, 10, 0, 1 ]
Sentence ambiguities can be resolved by using [[ domain targeted preference knowledge ]] without using complicated large << knowledgebases >> .
COMPARE
[ 7, 10, 15, 15 ]
Implementation and empirical results are described for the the analysis of [[ dependency structures ]] of << Japanese patent claim sentences >> .
FEATURE-OF
[ 11, 12, 14, 17 ]
<< Multimodal interfaces >> require effective [[ parsing ]] and understanding of utterances whose content is distributed across multiple input modes .
USED-FOR
[ 4, 4, 0, 1 ]
Johnston 1998 presents an [[ approach ]] in which strategies for << multimodal integration >> are stated declaratively using a unification-based grammar that is used by a multidimensional chart parser to compose inputs .
USED-FOR
[ 4, 4, 9, 10 ]
Johnston 1998 presents an approach in which strategies for << multimodal integration >> are stated declaratively using a [[ unification-based grammar ]] that is used by a multidimensional chart parser to compose inputs .
USED-FOR
[ 16, 17, 9, 10 ]
Johnston 1998 presents an approach in which strategies for multimodal integration are stated declaratively using a [[ unification-based grammar ]] that is used by a << multidimensional chart parser >> to compose inputs .
USED-FOR
[ 16, 17, 23, 25 ]
In this paper , we present an alternative [[ approach ]] in which << multimodal parsing and understanding >> are achieved using a weighted finite-state device which takes speech and gesture streams as inputs and outputs their joint interpretation .
USED-FOR
[ 8, 8, 11, 14 ]
In this paper , we present an alternative approach in which << multimodal parsing and understanding >> are achieved using a [[ weighted finite-state device ]] which takes speech and gesture streams as inputs and outputs their joint interpretation .
USED-FOR
[ 19, 21, 11, 14 ]
In this paper , we present an alternative approach in which multimodal parsing and understanding are achieved using a << weighted finite-state device >> which takes [[ speech and gesture streams ]] as inputs and outputs their joint interpretation .
USED-FOR
[ 24, 27, 19, 21 ]
This [[ approach ]] is significantly more efficient , enables tight-coupling of multimodal understanding with speech recognition , and provides a general probabilistic framework for << multimodal ambiguity resolution >> .
USED-FOR
[ 1, 1, 23, 25 ]
This approach is significantly more efficient , enables tight-coupling of << multimodal understanding >> with [[ speech recognition ]] , and provides a general probabilistic framework for multimodal ambiguity resolution .
CONJUNCTION
[ 13, 14, 10, 11 ]
Recently , we initiated a project to develop a << phonetically-based spoken language understanding system >> called [[ SUMMIT ]] .
HYPONYM-OF
[ 15, 15, 9, 13 ]
In contrast to many of the past efforts that make use of << heuristic rules >> whose development requires intense [[ knowledge engineering ]] , our approach attempts to express the speech knowledge within a formal framework using well-defined mathematical tools .
USED-FOR
[ 18, 19, 12, 13 ]
In contrast to many of the past efforts that make use of heuristic rules whose development requires intense knowledge engineering , our [[ approach ]] attempts to express the << speech knowledge >> within a formal framework using well-defined mathematical tools .
USED-FOR
[ 22, 22, 27, 28 ]
In contrast to many of the past efforts that make use of heuristic rules whose development requires intense knowledge engineering , our approach attempts to express the << speech knowledge >> within a formal framework using well-defined [[ mathematical tools ]] .
USED-FOR
[ 35, 36, 27, 28 ]
In our system , [[ features ]] and << decision strategies >> are discovered and trained automatically , using a large body of speech data .
CONJUNCTION
[ 4, 4, 6, 7 ]
In our system , features and << decision strategies >> are discovered and trained automatically , using a large body of [[ speech data ]] .
USED-FOR
[ 19, 20, 6, 7 ]
This paper describes an implemented << program >> that takes a [[ tagged text corpus ]] and generates a partial list of the subcategorization frames in which each verb occurs .
EVALUATE-FOR
[ 9, 11, 5, 5 ]
We present a [[ method ]] for estimating the << relative pose of two calibrated or uncalibrated non-overlapping surveillance cameras >> from observing a moving object .
USED-FOR
[ 3, 3, 7, 16 ]
We show how to tackle the problem of << missing point correspondences >> heavily required by [[ SfM pipelines ]] and how to go beyond this basic paradigm .
USED-FOR
[ 14, 15, 8, 10 ]
We relax the [[ non-linear nature ]] of the << problem >> by accepting two assumptions which surveillance scenarios offer , ie .
FEATURE-OF
[ 3, 4, 7, 7 ]
By those assumptions we cast the << problem >> as a [[ Quadratic Eigenvalue Problem ]] offering an elegant way of treating nonlinear monomials and delivering a quasi closed-form solution as a reliable starting point for a further bundle adjustment .
USED-FOR
[ 9, 11, 6, 6 ]
By those assumptions we cast the problem as a [[ Quadratic Eigenvalue Problem ]] offering an elegant way of treating << nonlinear monomials >> and delivering a quasi closed-form solution as a reliable starting point for a further bundle adjustment .
USED-FOR
[ 9, 11, 18, 19 ]
By those assumptions we cast the problem as a [[ Quadratic Eigenvalue Problem ]] offering an elegant way of treating nonlinear monomials and delivering a << quasi closed-form solution >> as a reliable starting point for a further bundle adjustment .
USED-FOR
[ 9, 11, 23, 25 ]
By those assumptions we cast the problem as a Quadratic Eigenvalue Problem offering an elegant way of treating nonlinear monomials and delivering a [[ quasi closed-form solution ]] as a reliable starting point for a further << bundle adjustment >> .
USED-FOR
[ 23, 25, 34, 35 ]
We are the first to bring the [[ closed form solution ]] to such a very practical << problem >> arising in video surveillance .
USED-FOR
[ 7, 9, 15, 15 ]
We are the first to bring the closed form solution to such a very practical << problem >> arising in [[ video surveillance ]] .
FEATURE-OF
[ 18, 19, 15, 15 ]
In this paper , we propose a [[ human action recognition system ]] suitable for << embedded computer vision applications >> in security systems , human-computer interaction and intelligent environments .
USED-FOR
[ 7, 10, 13, 16 ]
In this paper , we propose a human action recognition system suitable for [[ embedded computer vision applications ]] in << security systems >> , human-computer interaction and intelligent environments .
USED-FOR
[ 13, 16, 18, 19 ]
In this paper , we propose a human action recognition system suitable for [[ embedded computer vision applications ]] in security systems , << human-computer interaction >> and intelligent environments .
USED-FOR
[ 13, 16, 21, 22 ]
In this paper , we propose a human action recognition system suitable for [[ embedded computer vision applications ]] in security systems , human-computer interaction and << intelligent environments >> .
USED-FOR
[ 13, 16, 24, 25 ]
In this paper , we propose a human action recognition system suitable for embedded computer vision applications in [[ security systems ]] , << human-computer interaction >> and intelligent environments .
CONJUNCTION
[ 18, 19, 21, 22 ]
In this paper , we propose a human action recognition system suitable for embedded computer vision applications in security systems , [[ human-computer interaction ]] and << intelligent environments >> .
CONJUNCTION
[ 21, 22, 24, 25 ]
Our [[ system ]] is suitable for << embedded computer vision application >> based on three reasons .
USED-FOR
[ 1, 1, 5, 8 ]
Firstly , the << system >> was based on a [[ linear Support Vector Machine -LRB- SVM -RRB- classifier ]] where classification progress can be implemented easily and quickly in embedded hardware .
USED-FOR
[ 8, 15, 3, 3 ]
Firstly , the system was based on a linear Support Vector Machine -LRB- SVM -RRB- classifier where << classification progress >> can be implemented easily and quickly in [[ embedded hardware ]] .
USED-FOR
[ 26, 27, 17, 18 ]
Secondly , we use << compacted motion features >> easily obtained from [[ videos ]] .
USED-FOR
[ 10, 10, 4, 6 ]
We address the limitations of the well known Motion History Image -LRB- MHI -RRB- and propose a new [[ Hierarchical Motion History Histogram -LRB- HMHH -RRB- feature ]] to represent the << motion information >> .
USED-FOR
[ 18, 25, 29, 30 ]
[[ HMHH ]] not only provides << rich motion information >> , but also remains computationally inexpensive .
USED-FOR
[ 0, 0, 4, 6 ]
Finally , we combine [[ MHI ]] and << HMHH >> together and extract a low dimension feature vector to be used in the SVM classifiers .
CONJUNCTION
[ 4, 4, 6, 6 ]
Finally , we combine [[ MHI ]] and HMHH together and extract a << low dimension feature vector >> to be used in the SVM classifiers .
USED-FOR
[ 4, 4, 11, 14 ]
Finally , we combine MHI and [[ HMHH ]] together and extract a << low dimension feature vector >> to be used in the SVM classifiers .
USED-FOR
[ 6, 6, 11, 14 ]
Finally , we combine MHI and HMHH together and extract a [[ low dimension feature vector ]] to be used in the << SVM classifiers >> .
USED-FOR
[ 11, 14, 20, 21 ]
Experimental results show that our << system >> achieves significant improvement on the [[ recognition ]] performance .
EVALUATE-FOR
[ 11, 11, 5, 5 ]
In this paper I will argue for a << model of grammatical processing >> that is based on [[ uniform processing ]] and knowledge sources .
USED-FOR
[ 16, 17, 8, 11 ]
In this paper I will argue for a << model of grammatical processing >> that is based on uniform processing and [[ knowledge sources ]] .
USED-FOR
[ 19, 20, 8, 11 ]
In this paper I will argue for a model of grammatical processing that is based on << uniform processing >> and [[ knowledge sources ]] .
CONJUNCTION
[ 19, 20, 16, 17 ]
The main feature of this model is to view [[ parsing ]] and << generation >> as two strongly interleaved tasks performed by a single parametrized deduction process .
CONJUNCTION
[ 9, 9, 11, 11 ]
The main feature of this model is to view [[ parsing ]] and generation as two strongly interleaved << tasks >> performed by a single parametrized deduction process .
HYPONYM-OF
[ 9, 9, 16, 16 ]
The main feature of this model is to view parsing and [[ generation ]] as two strongly interleaved << tasks >> performed by a single parametrized deduction process .
HYPONYM-OF
[ 11, 11, 16, 16 ]
The main feature of this model is to view parsing and generation as two strongly interleaved << tasks >> performed by a single [[ parametrized deduction process ]] .
USED-FOR
[ 21, 23, 16, 16 ]
[[ Link detection ]] has been regarded as a core technology for the << Topic Detection and Tracking tasks of new event detection >> .
USED-FOR
[ 0, 1, 11, 19 ]
In this paper we formulate [[ story link detection ]] and << new event detection >> as information retrieval task and hypothesize on the impact of precision and recall on both systems .
CONJUNCTION
[ 5, 7, 9, 11 ]
In this paper we formulate [[ story link detection ]] and new event detection as information retrieval task and hypothesize on the impact of precision and recall on both << systems >> .
HYPONYM-OF
[ 5, 7, 27, 27 ]
In this paper we formulate story link detection and [[ new event detection ]] as information retrieval task and hypothesize on the impact of precision and recall on both << systems >> .
HYPONYM-OF
[ 9, 11, 27, 27 ]
In this paper we formulate << story link detection >> and new event detection as [[ information retrieval task ]] and hypothesize on the impact of precision and recall on both systems .
USED-FOR
[ 13, 15, 5, 7 ]
In this paper we formulate story link detection and << new event detection >> as [[ information retrieval task ]] and hypothesize on the impact of precision and recall on both systems .
USED-FOR
[ 13, 15, 9, 11 ]
In this paper we formulate story link detection and new event detection as information retrieval task and hypothesize on the impact of [[ precision ]] and << recall >> on both systems .
CONJUNCTION
[ 22, 22, 24, 24 ]
In this paper we formulate story link detection and new event detection as information retrieval task and hypothesize on the impact of [[ precision ]] and recall on both << systems >> .
EVALUATE-FOR
[ 22, 22, 27, 27 ]
In this paper we formulate story link detection and new event detection as information retrieval task and hypothesize on the impact of precision and [[ recall ]] on both << systems >> .
EVALUATE-FOR
[ 24, 24, 27, 27 ]
Motivated by these arguments , we introduce a number of new << performance enhancing techniques >> including [[ part of speech tagging ]] , new similarity measures and expanded stop lists .
PART-OF
[ 15, 18, 11, 13 ]
Motivated by these arguments , we introduce a number of new performance enhancing techniques including [[ part of speech tagging ]] , new << similarity measures >> and expanded stop lists .
CONJUNCTION
[ 15, 18, 21, 22 ]
Motivated by these arguments , we introduce a number of new << performance enhancing techniques >> including part of speech tagging , new [[ similarity measures ]] and expanded stop lists .
PART-OF
[ 21, 22, 11, 13 ]
Motivated by these arguments , we introduce a number of new performance enhancing techniques including part of speech tagging , new [[ similarity measures ]] and << expanded stop lists >> .
CONJUNCTION
[ 21, 22, 24, 26 ]
Motivated by these arguments , we introduce a number of new << performance enhancing techniques >> including part of speech tagging , new similarity measures and [[ expanded stop lists ]] .
PART-OF
[ 24, 26, 11, 13 ]
We attempt to understand << visual classification >> in humans using both [[ psy-chophysical and machine learning techniques ]] .
USED-FOR
[ 10, 14, 4, 5 ]
[[ Frontal views of human faces ]] were used for a << gender classification task >> .
USED-FOR
[ 0, 4, 9, 11 ]
Several [[ hyperplane learning algorithms ]] were used on the same << classification task >> using the Principal Components of the texture and flowfield representation of the faces .
USED-FOR
[ 1, 3, 9, 10 ]
Several << hyperplane learning algorithms >> were used on the same classification task using the [[ Principal Components of the texture ]] and flowfield representation of the faces .
USED-FOR
[ 13, 17, 1, 3 ]
Several << hyperplane learning algorithms >> were used on the same classification task using the Principal Components of the texture and [[ flowfield representation of the faces ]] .
USED-FOR
[ 19, 23, 1, 3 ]
Several hyperplane learning algorithms were used on the same classification task using the << Principal Components of the texture >> and [[ flowfield representation of the faces ]] .
CONJUNCTION
[ 19, 23, 13, 17 ]
The << classification >> performance of the [[ learning algorithms ]] was estimated using the face database with the true gender of the faces as labels , and also with the gender estimated by the subjects .
USED-FOR
[ 5, 6, 1, 1 ]
The classification performance of the << learning algorithms >> was estimated using the [[ face database ]] with the true gender of the faces as labels , and also with the gender estimated by the subjects .
EVALUATE-FOR
[ 11, 12, 5, 6 ]
Our results suggest that << human classification >> can be modeled by some [[ hyperplane algorithms ]] in the feature space we used .
USED-FOR
[ 11, 12, 4, 5 ]
Our results suggest that human classification can be modeled by some << hyperplane algorithms >> in the [[ feature space ]] we used .
FEATURE-OF
[ 15, 16, 11, 12 ]
For classification , the brain needs more processing for stimuli close to that [[ hyperplane ]] than for << those >> further away .
COMPARE
[ 13, 13, 16, 16 ]
In this paper , we present a [[ corpus-based supervised word sense disambiguation -LRB- WSD -RRB- system ]] for << Dutch >> which combines statistical classification -LRB- maximum entropy -RRB- with linguistic information .
USED-FOR
[ 7, 15, 17, 17 ]
In this paper , we present a << corpus-based supervised word sense disambiguation -LRB- WSD -RRB- system >> for Dutch which combines [[ statistical classification ]] -LRB- maximum entropy -RRB- with linguistic information .
PART-OF
[ 20, 21, 7, 15 ]
In this paper , we present a << corpus-based supervised word sense disambiguation -LRB- WSD -RRB- system >> for Dutch which combines statistical classification -LRB- [[ maximum entropy ]] -RRB- with linguistic information .
PART-OF
[ 23, 24, 7, 15 ]
In this paper , we present a << corpus-based supervised word sense disambiguation -LRB- WSD -RRB- system >> for Dutch which combines statistical classification -LRB- maximum entropy -RRB- with [[ linguistic information ]] .
PART-OF
[ 27, 28, 7, 15 ]
In this paper , we present a corpus-based supervised word sense disambiguation -LRB- WSD -RRB- system for Dutch which combines statistical classification -LRB- << maximum entropy >> -RRB- with [[ linguistic information ]] .
CONJUNCTION
[ 27, 28, 23, 24 ]
Instead of building individual [[ classifiers ]] per ambiguous wordform , we introduce a << lemma-based approach >> .
COMPARE
[ 4, 4, 12, 13 ]
Instead of building individual << classifiers >> per [[ ambiguous wordform ]] , we introduce a lemma-based approach .
USED-FOR
[ 6, 7, 4, 4 ]
The advantage of this novel method is that it clusters all [[ inflected forms ]] of an << ambiguous word >> in one classifier , therefore augmenting the training material available to the algorithm .
FEATURE-OF
[ 11, 12, 15, 16 ]
Testing the [[ lemma-based model ]] on the Dutch Senseval-2 test data , we achieve a significant increase in accuracy over the << wordform model >> .
COMPARE
[ 2, 3, 20, 21 ]
Testing the << lemma-based model >> on the [[ Dutch Senseval-2 test data ]] , we achieve a significant increase in accuracy over the wordform model .
EVALUATE-FOR
[ 6, 9, 2, 3 ]
We propose an exact , general and efficient [[ coarse-to-fine energy minimization strategy ]] for << semantic video segmenta-tion >> .
USED-FOR
[ 8, 11, 13, 15 ]
Our << strategy >> is based on a [[ hierarchical abstraction of the supervoxel graph ]] that allows us to minimize an energy defined at the finest level of the hierarchy by minimizing a series of simpler energies defined over coarser graphs .
USED-FOR
[ 6, 11, 1, 1 ]
It is general , i.e. , [[ it ]] can be used to minimize any << energy function >> -LRB- e.g. , unary , pairwise , and higher-order terms -RRB- with any existing energy minimization algorithm -LRB- e.g. , graph cuts and belief propagation -RRB- .
USED-FOR
[ 6, 6, 13, 14 ]
It is general , i.e. , [[ it ]] can be used to minimize any energy function -LRB- e.g. , unary , pairwise , and higher-order terms -RRB- with any existing << energy minimization algorithm >> -LRB- e.g. , graph cuts and belief propagation -RRB- .
CONJUNCTION
[ 6, 6, 29, 31 ]
It is general , i.e. , it can be used to minimize any << energy function >> -LRB- e.g. , unary , pairwise , and higher-order terms -RRB- with any existing [[ energy minimization algorithm ]] -LRB- e.g. , graph cuts and belief propagation -RRB- .
USED-FOR
[ 29, 31, 13, 14 ]
It is general , i.e. , it can be used to minimize any energy function -LRB- e.g. , unary , pairwise , and higher-order terms -RRB- with any existing << energy minimization algorithm >> -LRB- e.g. , [[ graph cuts ]] and belief propagation -RRB- .
HYPONYM-OF
[ 35, 36, 29, 31 ]
It is general , i.e. , it can be used to minimize any energy function -LRB- e.g. , unary , pairwise , and higher-order terms -RRB- with any existing energy minimization algorithm -LRB- e.g. , [[ graph cuts ]] and << belief propagation >> -RRB- .
CONJUNCTION
[ 35, 36, 38, 39 ]
It is general , i.e. , it can be used to minimize any energy function -LRB- e.g. , unary , pairwise , and higher-order terms -RRB- with any existing << energy minimization algorithm >> -LRB- e.g. , graph cuts and [[ belief propagation ]] -RRB- .
HYPONYM-OF
[ 38, 39, 29, 31 ]