text stringlengths 49 577 | label stringclasses 7
values | metadata list |
|---|---|---|
We show that the newly proposed concept-distance measures outperform traditional [[ distributional word-distance measures ]] in the << tasks >> of -LRB- 1 -RRB- ranking word pairs in order of semantic distance , and -LRB- 2 -RRB- correcting real-word spelling errors . | USED-FOR | [
10,
12,
15,
15
] |
We show that the newly proposed concept-distance measures outperform traditional distributional word-distance measures in the << tasks >> of -LRB- 1 -RRB- [[ ranking word pairs in order of semantic distance ]] , and -LRB- 2 -RRB- correcting real-word spelling errors . | HYPONYM-OF | [
20,
27,
15,
15
] |
We show that the newly proposed concept-distance measures outperform traditional distributional word-distance measures in the << tasks >> of -LRB- 1 -RRB- ranking word pairs in order of semantic distance , and -LRB- 2 -RRB- [[ correcting real-word spelling errors ]] . | HYPONYM-OF | [
33,
36,
15,
15
] |
We show that the newly proposed concept-distance measures outperform traditional distributional word-distance measures in the tasks of -LRB- 1 -RRB- << ranking word pairs in order of semantic distance >> , and -LRB- 2 -RRB- [[ correcting real-word spelling errors ]] . | CONJUNCTION | [
33,
36,
20,
27
] |
In the latter [[ task ]] , of all the << WordNet-based measures >> , only that proposed by Jiang and Conrath outperforms the best distributional concept-distance measures . | EVALUATE-FOR | [
3,
3,
8,
9
] |
In the latter [[ task ]] , of all the WordNet-based measures , only that proposed by Jiang and Conrath outperforms the best << distributional concept-distance measures >> . | EVALUATE-FOR | [
3,
3,
21,
23
] |
In the latter task , of all the << WordNet-based measures >> , only that proposed by Jiang and Conrath outperforms the best [[ distributional concept-distance measures ]] . | COMPARE | [
21,
23,
8,
9
] |
One of the main results of this work is the definition of a relation between [[ broad semantic classes ]] and << LCS meaning components >> . | CONJUNCTION | [
15,
17,
19,
21
] |
Our [[ acquisition program - LEXICALL - ]] takes , as input , the result of previous work on verb classification and thematic grid tagging , and outputs << LCS representations >> for different languages . | USED-FOR | [
1,
5,
26,
27
] |
Our << acquisition program - LEXICALL - >> takes , as input , the result of previous work on [[ verb classification ]] and thematic grid tagging , and outputs LCS representations for different languages . | USED-FOR | [
17,
18,
1,
5
] |
Our acquisition program - LEXICALL - takes , as input , the result of previous work on [[ verb classification ]] and << thematic grid tagging >> , and outputs LCS representations for different languages . | CONJUNCTION | [
17,
18,
20,
22
] |
Our << acquisition program - LEXICALL - >> takes , as input , the result of previous work on verb classification and [[ thematic grid tagging ]] , and outputs LCS representations for different languages . | USED-FOR | [
20,
22,
1,
5
] |
These [[ representations ]] have been ported into << English , Arabic and Spanish lexicons >> , each containing approximately 9000 verbs . | USED-FOR | [
1,
1,
6,
11
] |
We are currently using these [[ lexicons ]] in an << operational foreign language tutoring >> and machine translation . | USED-FOR | [
5,
5,
8,
11
] |
We are currently using these [[ lexicons ]] in an operational foreign language tutoring and << machine translation >> . | USED-FOR | [
5,
5,
13,
14
] |
We are currently using these lexicons in an [[ operational foreign language tutoring ]] and << machine translation >> . | CONJUNCTION | [
8,
11,
13,
14
] |
The theoretical study of the [[ range concatenation grammar -LSB- RCG -RSB- formalism ]] has revealed many attractive properties which may be used in << NLP >> . | USED-FOR | [
5,
11,
22,
22
] |
In particular , << range concatenation languages -LSB- RCL -RSB- >> can be parsed in [[ polynomial time ]] and many classical grammatical formalisms can be translated into equivalent RCGs without increasing their worst-case parsing time complexity . | FEATURE-OF | [
13,
14,
3,
8
] |
In particular , range concatenation languages -LSB- RCL -RSB- can be parsed in polynomial time and many classical << grammatical formalisms >> can be translated into equivalent RCGs without increasing their [[ worst-case parsing time complexity ]] . | EVALUATE-FOR | [
29,
32,
18,
19
] |
For example , after translation into an equivalent RCG , any << tree adjoining grammar >> can be parsed in [[ O -LRB- n6 -RRB- time ]] . | FEATURE-OF | [
18,
22,
11,
13
] |
In this paper , we study a [[ parsing technique ]] whose purpose is to improve the practical efficiency of << RCL parsers >> . | USED-FOR | [
7,
8,
18,
19
] |
The non-deterministic parsing choices of the [[ main parser ]] for a << language L >> are directed by a guide which uses the shared derivation forest output by a prior RCL parser for a suitable superset of L . | USED-FOR | [
6,
7,
10,
11
] |
The non-deterministic parsing choices of the main parser for a language L are directed by a guide which uses the << shared derivation forest >> output by a prior [[ RCL parser ]] for a suitable superset of L . | USED-FOR | [
27,
28,
20,
22
] |
The results of a practical evaluation of this << method >> on a [[ wide coverage English grammar ]] are given . | EVALUATE-FOR | [
11,
14,
8,
8
] |
In this paper we introduce [[ Ant-Q ]] , a family of algorithms which present many similarities with Q-learning -LRB- Watkins , 1989 -RRB- , and which we apply to the solution of << symmetric and asym-metric instances of the traveling salesman problem -LRB- TSP -RRB- >> . | USED-FOR | [
5,
5,
31,
42
] |
<< Ant-Q algorithms >> were inspired by work on the [[ ant system -LRB- AS -RRB- ]] , a distributed algorithm for combinatorial optimization based on the metaphor of ant colonies which was recently proposed in -LRB- Dorigo , 1992 ; Dorigo , Maniezzo and Colorni , 1996 -RRB- . | USED-FOR | [
8,
12,
0,
1
] |
Ant-Q algorithms were inspired by work on the [[ ant system -LRB- AS -RRB- ]] , a << distributed algorithm >> for combinatorial optimization based on the metaphor of ant colonies which was recently proposed in -LRB- Dorigo , 1992 ; Dorigo , Maniezzo and Colorni , 1996 -RRB- . | HYPONYM-OF | [
8,
12,
15,
16
] |
Ant-Q algorithms were inspired by work on the ant system -LRB- AS -RRB- , a [[ distributed algorithm ]] for << combinatorial optimization >> based on the metaphor of ant colonies which was recently proposed in -LRB- Dorigo , 1992 ; Dorigo , Maniezzo and Colorni , 1996 -RRB- . | USED-FOR | [
15,
16,
18,
19
] |
We show that [[ AS ]] is a particular instance of the << Ant-Q family >> , and that there are instances of this family which perform better than AS . | HYPONYM-OF | [
3,
3,
10,
11
] |
We show that AS is a particular instance of the Ant-Q family , and that there are [[ instances ]] of this << family >> which perform better than AS . | PART-OF | [
17,
17,
20,
20
] |
We show that AS is a particular instance of the Ant-Q family , and that there are [[ instances ]] of this family which perform better than << AS >> . | COMPARE | [
17,
17,
25,
25
] |
We experimentally investigate the functioning of Ant-Q and we show that the results obtained by [[ Ant-Q ]] on << symmetric TSP >> 's are competitive with those obtained by other heuristic approaches based on neural networks or local search . | USED-FOR | [
15,
15,
17,
18
] |
We experimentally investigate the functioning of Ant-Q and we show that the results obtained by [[ Ant-Q ]] on symmetric TSP 's are competitive with those obtained by other << heuristic approaches >> based on neural networks or local search . | COMPARE | [
15,
15,
27,
28
] |
We experimentally investigate the functioning of Ant-Q and we show that the results obtained by Ant-Q on symmetric TSP 's are competitive with those obtained by other << heuristic approaches >> based on [[ neural networks ]] or local search . | USED-FOR | [
31,
32,
27,
28
] |
We experimentally investigate the functioning of Ant-Q and we show that the results obtained by Ant-Q on symmetric TSP 's are competitive with those obtained by other heuristic approaches based on [[ neural networks ]] or << local search >> . | CONJUNCTION | [
31,
32,
34,
35
] |
We experimentally investigate the functioning of Ant-Q and we show that the results obtained by Ant-Q on symmetric TSP 's are competitive with those obtained by other << heuristic approaches >> based on neural networks or [[ local search ]] . | USED-FOR | [
34,
35,
27,
28
] |
Finally , we apply [[ Ant-Q ]] to some difficult << asymmetric TSP >> 's obtaining very good results : Ant-Q was able to find solutions of a quality which usually can be found only by very specialized algorithms . | USED-FOR | [
4,
4,
8,
9
] |
In this paper , we develop a [[ geometric framework ]] for << linear or nonlinear discriminant subspace learning and classification >> . | USED-FOR | [
7,
8,
10,
17
] |
In our framework , the << structures of classes >> are conceptualized as a [[ semi-Riemannian manifold ]] which is considered as a submanifold embedded in an ambient semi-Riemannian space . | USED-FOR | [
12,
13,
5,
7
] |
In our framework , the structures of classes are conceptualized as a semi-Riemannian manifold which is considered as a [[ submanifold ]] embedded in an << ambient semi-Riemannian space >> . | PART-OF | [
19,
19,
23,
25
] |
The << class structures >> of original samples can be characterized and deformed by [[ local metrics of the semi-Riemannian space ]] . | USED-FOR | [
12,
17,
1,
2
] |
<< Semi-Riemannian metrics >> are uniquely determined by the [[ smoothing of discrete functions ]] and the nullity of the semi-Riemannian space . | USED-FOR | [
7,
10,
0,
1
] |
Semi-Riemannian metrics are uniquely determined by the [[ smoothing of discrete functions ]] and the << nullity of the semi-Riemannian space >> . | CONJUNCTION | [
7,
10,
13,
17
] |
<< Semi-Riemannian metrics >> are uniquely determined by the smoothing of discrete functions and the [[ nullity of the semi-Riemannian space ]] . | USED-FOR | [
13,
17,
0,
1
] |
Based on the geometrization of class structures , optimizing << class structures >> in the [[ feature space ]] is equivalent to maximizing the quadratic quantities of metric tensors in the semi-Riemannian space . | FEATURE-OF | [
13,
14,
9,
10
] |
Based on the geometrization of class structures , optimizing class structures in the feature space is equivalent to maximizing the << quadratic quantities of metric tensors >> in the [[ semi-Riemannian space ]] . | FEATURE-OF | [
27,
28,
20,
24
] |
Based on the proposed [[ framework ]] , a novel << algorithm >> , dubbed as Semi-Riemannian Discriminant Analysis -LRB- SRDA -RRB- , is presented for subspace-based classification . | USED-FOR | [
4,
4,
8,
8
] |
Based on the proposed framework , a novel [[ algorithm ]] , dubbed as Semi-Riemannian Discriminant Analysis -LRB- SRDA -RRB- , is presented for << subspace-based classification >> . | USED-FOR | [
8,
8,
22,
23
] |
The performance of [[ SRDA ]] is tested on face recognition -LRB- singular case -RRB- and handwritten capital letter classification -LRB- nonsingular case -RRB- against existing << algorithms >> . | COMPARE | [
3,
3,
24,
24
] |
The performance of << SRDA >> is tested on [[ face recognition -LRB- singular case ]] -RRB- and handwritten capital letter classification -LRB- nonsingular case -RRB- against existing algorithms . | EVALUATE-FOR | [
7,
11,
3,
3
] |
The performance of SRDA is tested on [[ face recognition -LRB- singular case ]] -RRB- and << handwritten capital letter classification -LRB- nonsingular case -RRB- >> against existing algorithms . | CONJUNCTION | [
7,
11,
14,
21
] |
The performance of SRDA is tested on [[ face recognition -LRB- singular case ]] -RRB- and handwritten capital letter classification -LRB- nonsingular case -RRB- against existing << algorithms >> . | EVALUATE-FOR | [
7,
11,
24,
24
] |
The performance of << SRDA >> is tested on face recognition -LRB- singular case -RRB- and [[ handwritten capital letter classification -LRB- nonsingular case -RRB- ]] against existing algorithms . | EVALUATE-FOR | [
14,
21,
3,
3
] |
The performance of SRDA is tested on face recognition -LRB- singular case -RRB- and [[ handwritten capital letter classification -LRB- nonsingular case -RRB- ]] against existing << algorithms >> . | EVALUATE-FOR | [
14,
21,
24,
24
] |
The experimental results show that [[ SRDA ]] works well on << recognition >> and classification , implying that semi-Riemannian geometry is a promising new tool for pattern recognition and machine learning . | USED-FOR | [
5,
5,
9,
9
] |
The experimental results show that [[ SRDA ]] works well on recognition and << classification >> , implying that semi-Riemannian geometry is a promising new tool for pattern recognition and machine learning . | USED-FOR | [
5,
5,
11,
11
] |
The experimental results show that SRDA works well on [[ recognition ]] and << classification >> , implying that semi-Riemannian geometry is a promising new tool for pattern recognition and machine learning . | CONJUNCTION | [
9,
9,
11,
11
] |
The experimental results show that SRDA works well on recognition and classification , implying that [[ semi-Riemannian geometry ]] is a promising new tool for << pattern recognition >> and machine learning . | USED-FOR | [
15,
16,
23,
24
] |
The experimental results show that SRDA works well on recognition and classification , implying that [[ semi-Riemannian geometry ]] is a promising new tool for pattern recognition and << machine learning >> . | USED-FOR | [
15,
16,
26,
27
] |
The experimental results show that SRDA works well on recognition and classification , implying that semi-Riemannian geometry is a promising new tool for [[ pattern recognition ]] and << machine learning >> . | CONJUNCTION | [
23,
24,
26,
27
] |
A [[ deterministic parser ]] is under development which represents a departure from traditional << deterministic parsers >> in that it combines both symbolic and connectionist components . | COMPARE | [
1,
2,
12,
13
] |
A deterministic parser is under development which represents a departure from traditional deterministic parsers in that << it >> combines both [[ symbolic and connectionist components ]] . | PART-OF | [
19,
22,
16,
16
] |
The << connectionist component >> is trained either from [[ patterns ]] derived from the rules of a deterministic grammar . | USED-FOR | [
7,
7,
1,
2
] |
The connectionist component is trained either from << patterns >> derived from the [[ rules of a deterministic grammar ]] . | USED-FOR | [
11,
15,
7,
7
] |
The development and evolution of such a [[ hybrid architecture ]] has lead to a << parser >> which is superior to any known deterministic parser . | USED-FOR | [
7,
8,
13,
13
] |
The development and evolution of such a hybrid architecture has lead to a [[ parser ]] which is superior to any known << deterministic parser >> . | COMPARE | [
13,
13,
20,
21
] |
Experiments are described and powerful [[ training techniques ]] are demonstrated that permit << decision-making >> by the connectionist component in the parsing process . | USED-FOR | [
5,
6,
11,
11
] |
Experiments are described and powerful training techniques are demonstrated that permit << decision-making >> by the [[ connectionist component ]] in the parsing process . | USED-FOR | [
14,
15,
11,
11
] |
Experiments are described and powerful training techniques are demonstrated that permit decision-making by the [[ connectionist component ]] in the << parsing process >> . | PART-OF | [
14,
15,
18,
19
] |
Data are presented which show how a [[ connectionist -LRB- neural -RRB- network ]] trained with linguistic rules can parse both << expected -LRB- grammatical -RRB- sentences >> as well as some novel -LRB- ungrammatical or lexically ambiguous -RRB- sentences . | USED-FOR | [
7,
11,
19,
23
] |
Data are presented which show how a [[ connectionist -LRB- neural -RRB- network ]] trained with linguistic rules can parse both expected -LRB- grammatical -RRB- sentences as well as some novel << -LRB- ungrammatical or lexically ambiguous -RRB- sentences >> . | USED-FOR | [
7,
11,
29,
35
] |
Data are presented which show how a << connectionist -LRB- neural -RRB- network >> trained with [[ linguistic rules ]] can parse both expected -LRB- grammatical -RRB- sentences as well as some novel -LRB- ungrammatical or lexically ambiguous -RRB- sentences . | USED-FOR | [
14,
15,
7,
11
] |
Data are presented which show how a connectionist -LRB- neural -RRB- network trained with linguistic rules can parse both [[ expected -LRB- grammatical -RRB- sentences ]] as well as some novel << -LRB- ungrammatical or lexically ambiguous -RRB- sentences >> . | CONJUNCTION | [
19,
23,
29,
35
] |
Robust << natural language interpretation >> requires strong [[ semantic domain models ]] , fail-soft recovery heuristics , and very flexible control structures . | USED-FOR | [
6,
8,
1,
3
] |
Robust natural language interpretation requires strong [[ semantic domain models ]] , << fail-soft recovery heuristics >> , and very flexible control structures . | CONJUNCTION | [
6,
8,
10,
12
] |
Robust << natural language interpretation >> requires strong semantic domain models , [[ fail-soft recovery heuristics ]] , and very flexible control structures . | USED-FOR | [
10,
12,
1,
3
] |
Robust natural language interpretation requires strong semantic domain models , [[ fail-soft recovery heuristics ]] , and very flexible << control structures >> . | CONJUNCTION | [
10,
12,
17,
18
] |
Robust << natural language interpretation >> requires strong semantic domain models , fail-soft recovery heuristics , and very flexible [[ control structures ]] . | USED-FOR | [
17,
18,
1,
3
] |
Although [[ single-strategy parsers ]] have met with a measure of success , a << multi-strategy approach >> is shown to provide a much higher degree of flexibility , redundancy , and ability to bring task-specific domain knowledge -LRB- in addition to general linguistic knowledge -RRB- to bear on both grammatical and u... | COMPARE | [
1,
2,
12,
13
] |
Although single-strategy parsers have met with a measure of success , a multi-strategy approach is shown to provide a much higher degree of flexibility , redundancy , and ability to bring [[ task-specific domain knowledge ]] -LRB- in addition to << general linguistic knowledge >> -RRB- to bear on both grammatical and u... | CONJUNCTION | [
31,
33,
38,
40
] |
A << parsing algorithm >> is presented that integrates several different [[ parsing strategies ]] , with case-frame instantiation dominating . | PART-OF | [
9,
10,
1,
2
] |
A parsing algorithm is presented that integrates several different << parsing strategies >> , with [[ case-frame instantiation ]] dominating . | HYPONYM-OF | [
13,
14,
9,
10
] |
Each of these [[ parsing strategies ]] exploits different types of knowledge ; and their combination provides a strong framework in which to process << conjunctions >> , fragmentary input , and ungrammatical structures , as well as less exotic , grammatically correct input . | USED-FOR | [
3,
4,
22,
22
] |
Each of these [[ parsing strategies ]] exploits different types of knowledge ; and their combination provides a strong framework in which to process conjunctions , << fragmentary input >> , and ungrammatical structures , as well as less exotic , grammatically correct input . | USED-FOR | [
3,
4,
24,
25
] |
Each of these [[ parsing strategies ]] exploits different types of knowledge ; and their combination provides a strong framework in which to process conjunctions , fragmentary input , and << ungrammatical structures >> , as well as less exotic , grammatically correct input . | USED-FOR | [
3,
4,
28,
29
] |
Each of these [[ parsing strategies ]] exploits different types of knowledge ; and their combination provides a strong framework in which to process conjunctions , fragmentary input , and ungrammatical structures , as well as less << exotic , grammatically correct input >> . | USED-FOR | [
3,
4,
35,
39
] |
Each of these parsing strategies exploits different types of knowledge ; and their combination provides a strong framework in which to process [[ conjunctions ]] , << fragmentary input >> , and ungrammatical structures , as well as less exotic , grammatically correct input . | CONJUNCTION | [
22,
22,
24,
25
] |
Each of these parsing strategies exploits different types of knowledge ; and their combination provides a strong framework in which to process conjunctions , [[ fragmentary input ]] , and << ungrammatical structures >> , as well as less exotic , grammatically correct input . | CONJUNCTION | [
24,
25,
28,
29
] |
Each of these parsing strategies exploits different types of knowledge ; and their combination provides a strong framework in which to process conjunctions , fragmentary input , and [[ ungrammatical structures ]] , as well as less << exotic , grammatically correct input >> . | CONJUNCTION | [
28,
29,
35,
39
] |
Several [[ specific heuristics ]] for handling << ungrammatical input >> are presented within this multi-strategy framework . | USED-FOR | [
1,
2,
5,
6
] |
Several [[ specific heuristics ]] for handling ungrammatical input are presented within this << multi-strategy framework >> . | PART-OF | [
1,
2,
11,
12
] |
Recently , [[ Stacked Auto-Encoders -LRB- SAE -RRB- ]] have been successfully used for << learning imbalanced datasets >> . | USED-FOR | [
2,
6,
12,
14
] |
In this paper , for the first time , we propose to use a [[ Neural Network classifier ]] furnished by an SAE structure for detecting the errors made by a strong << Automatic Speech Recognition -LRB- ASR -RRB- system >> . | USED-FOR | [
14,
16,
30,
36
] |
In this paper , for the first time , we propose to use a << Neural Network classifier >> furnished by an [[ SAE structure ]] for detecting the errors made by a strong Automatic Speech Recognition -LRB- ASR -RRB- system . | USED-FOR | [
20,
21,
14,
16
] |
[[ Error detection ]] on an << automatic transcription >> provided by a '' strong '' ASR system , i.e. exhibiting a small word error rate , is difficult due to the limited number of '' positive '' examples -LRB- i.e. words erroneously recognized -RRB- available for training a binary classi-fier . | USED-FOR | [
0,
1,
4,
5
] |
In this paper we investigate and compare different types of [[ classifiers ]] for << automatically detecting ASR errors >> , including the one based on a stacked auto-encoder architecture . | USED-FOR | [
10,
10,
12,
15
] |
In this paper we investigate and compare different types of << classifiers >> for automatically detecting ASR errors , including the [[ one ]] based on a stacked auto-encoder architecture . | HYPONYM-OF | [
19,
19,
10,
10
] |
In this paper we investigate and compare different types of classifiers for automatically detecting ASR errors , including the << one >> based on a [[ stacked auto-encoder architecture ]] . | USED-FOR | [
23,
25,
19,
19
] |
We show the effectiveness of the latter by measuring and comparing performance on the << automatic transcriptions >> of an [[ English corpus ]] collected from TED talks . | FEATURE-OF | [
18,
19,
14,
15
] |
We show the effectiveness of the latter by measuring and comparing performance on the automatic transcriptions of an << English corpus >> collected from [[ TED talks ]] . | USED-FOR | [
22,
23,
18,
19
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.