input stringlengths 37 565 | output dict | schema listlengths 0 0 |
|---|---|---|
Many computer vision applications , such as image classification and video indexing , are usually multi-label classification problems in which an instance can be assigned to more than one category . | {
"relations": {
"hyponym of": [
{
"head": {
"text": "image classification",
"start": 44,
"end": 64
},
"tail": {
"text": "computer vision applications",
"start": 5,
"end": 33
}
},
{
"head": {
... | [] |
In this paper , we present a novel multi-label classification approach with hypergraph regu-larization that addresses the correlations among different categories . | {
"relations": {
"feature of": [
{
"head": {
"text": "hypergraph regu-larization",
"start": 76,
"end": 102
},
"tail": {
"text": "multi-label classification approach",
"start": 35,
"end": 70
}
}
]
}
} | [] |
Then , an improved SVM like learning system incorporating the hypergraph regularization , called Rank-HLapSVM , is proposed to handle the multi-label classification problems . | {
"relations": {
"used for": [
{
"head": {
"text": "SVM like learning system",
"start": 19,
"end": 43
},
"tail": {
"text": "multi-label classification problems",
"start": 138,
"end": 173
}
}
],
"part of... | [] |
We find that the corresponding optimization problem can be efficiently solved by the dual coordinate descent method . | {
"relations": {
"used for": [
{
"head": {
"text": "dual coordinate descent method",
"start": 85,
"end": 115
},
"tail": {
"text": "optimization problem",
"start": 31,
"end": 51
}
}
]
}
} | [] |
Many promising experimental results on the real datasets including ImageCLEF and Me-diaMill demonstrate the effectiveness and efficiency of the proposed algorithm . | {
"relations": {
"evaluate for": [
{
"head": {
"text": "real datasets",
"start": 43,
"end": 56
},
"tail": {
"text": "algorithm",
"start": 153,
"end": 162
}
}
],
"hyponym of": [
{
"head": {... | [] |
We derive a convex optimization problem for the task of segmenting sequential data , which explicitly treats presence of outliers . | {
"relations": {
"used for": [
{
"head": {
"text": "convex optimization problem",
"start": 12,
"end": 39
},
"tail": {
"text": "segmenting sequential data",
"start": 56,
"end": 82
}
},
{
"head": {
... | [] |
We describe two algorithms for solving this problem , one exact and one a top-down novel approach , and we derive a consistency results for the case of two segments and no outliers . | {
"relations": {
"used for": [
{
"head": {
"text": "algorithms",
"start": 16,
"end": 26
},
"tail": {
"text": "problem",
"start": 44,
"end": 51
}
}
]
}
} | [] |
Robustness to outliers is evaluated on two real-world tasks related to speech segmentation . | {
"relations": {
"feature of": [
{
"head": {
"text": "outliers",
"start": 14,
"end": 22
},
"tail": {
"text": "Robustness",
"start": 0,
"end": 10
}
},
{
"head": {
"text": "speech segmenta... | [] |
Our algorithms outperform baseline seg-mentation algorithms . | {
"relations": {
"compare": [
{
"head": {
"text": "algorithms",
"start": 4,
"end": 14
},
"tail": {
"text": "baseline seg-mentation algorithms",
"start": 26,
"end": 59
}
}
]
}
} | [] |
This paper examines the properties of feature-based partial descriptions built on top of Halliday 's systemic networks . | {
"relations": {
"used for": [
{
"head": {
"text": "Halliday 's systemic networks",
"start": 89,
"end": 118
},
"tail": {
"text": "feature-based partial descriptions",
"start": 38,
"end": 72
}
}
]
}
} | [] |
We show that the crucial operation of consistency checking for such descriptions is NP-complete , and therefore probably intractable , but proceed to develop algorithms which can sometimes alleviate the unpleasant consequences of this intractability . | {
"relations": {
"used for": [
{
"head": {
"text": "consistency checking",
"start": 38,
"end": 58
},
"tail": {
"text": "descriptions",
"start": 68,
"end": 80
}
}
]
}
} | [] |
We describe Yoopick , a combinatorial sports prediction market that implements a flexible betting language , and in turn facilitates fine-grained probabilistic estimation of outcomes . | {
"relations": {
"hyponym of": [
{
"head": {
"text": "Yoopick",
"start": 12,
"end": 19
},
"tail": {
"text": "combinatorial sports prediction market",
"start": 24,
"end": 62
}
}
],
"used for": [
{
... | [] |
The goal of this paper is to discover a set of discriminative patches which can serve as a fully unsupervised mid-level visual representation . | {
"relations": {
"used for": [
{
"head": {
"text": "discriminative patches",
"start": 47,
"end": 69
},
"tail": {
"text": "unsupervised mid-level visual representation",
"start": 97,
"end": 141
}
}
]
}
} | [] |
We pose this as an unsupervised discriminative clustering problem on a huge dataset of image patches . | {
"relations": {
"used for": [
{
"head": {
"text": "image patches",
"start": 87,
"end": 100
},
"tail": {
"text": "unsupervised discriminative clustering problem",
"start": 19,
"end": 65
}
}
]
}
} | [] |
We use an iterative procedure which alternates between clustering and training discriminative classifiers , while applying careful cross-validation at each step to prevent overfitting . | {
"relations": {
"used for": [
{
"head": {
"text": "cross-validation",
"start": 131,
"end": 147
},
"tail": {
"text": "overfitting",
"start": 172,
"end": 183
}
}
]
}
} | [] |
The paper experimentally demonstrates the effectiveness of discriminative patches as an unsupervised mid-level visual representation , suggesting that it could be used in place of visual words for many tasks . | {
"relations": {
"used for": [
{
"head": {
"text": "discriminative patches",
"start": 59,
"end": 81
},
"tail": {
"text": "unsupervised mid-level visual representation",
"start": 88,
"end": 132
}
},
{
... | [] |
Furthermore , discrim-inative patches can also be used in a supervised regime , such as scene classification , where they demonstrate state-of-the-art performance on the MIT Indoor-67 dataset . | {
"relations": {
"used for": [
{
"head": {
"text": "discrim-inative patches",
"start": 14,
"end": 37
},
"tail": {
"text": "supervised regime",
"start": 60,
"end": 77
}
}
],
"hyponym of": [
{
... | [] |
We investigate the utility of an algorithm for translation lexicon acquisition -LRB- SABLE -RRB- , used previously on a very large corpus to acquire general translation lexicons , when that algorithm is applied to a much smaller corpus to produce candidates for domain-specific translation lexicons . | {
"relations": {
"used for": [
{
"head": {
"text": "algorithm",
"start": 33,
"end": 42
},
"tail": {
"text": "translation lexicon acquisition -LRB- SABLE -RRB-",
"start": 47,
"end": 96
}
},
{
"head... | [] |
This paper describes a computational model of word segmentation and presents simulation results on realistic acquisition . | {
"relations": {
"used for": [
{
"head": {
"text": "computational model",
"start": 23,
"end": 42
},
"tail": {
"text": "word segmentation",
"start": 46,
"end": 63
}
},
{
"head": {
"text":... | [] |
In particular , we explore the capacity and limitations of statistical learning mechanisms that have recently gained prominence in cognitive psychology and linguistics . | {
"relations": {
"used for": [
{
"head": {
"text": "cognitive psychology",
"start": 131,
"end": 151
},
"tail": {
"text": "statistical learning mechanisms",
"start": 59,
"end": 90
}
},
{
"head": {
... | [] |
In the model-based policy search approach to reinforcement learning -LRB- RL -RRB- , policies are found using a model -LRB- or `` simulator '' -RRB- of the Markov decision process . | {
"relations": {
"used for": [
{
"head": {
"text": "model-based policy search approach",
"start": 7,
"end": 41
},
"tail": {
"text": "reinforcement learning -LRB- RL -RRB-",
"start": 45,
"end": 82
}
},
{
... | [] |
However , for high-dimensional continuous-state tasks , it can be extremely difficult to build an accurate model , and thus often the algorithm returns a policy that works in simulation but not in real-life . | {
"relations": {
"used for": [
{
"head": {
"text": "model",
"start": 107,
"end": 112
},
"tail": {
"text": "high-dimensional continuous-state tasks",
"start": 14,
"end": 53
}
},
{
"head": {
... | [] |
The other extreme , model-free RL , tends to require infeasibly large numbers of real-life trials . | {
"relations": {
"used for": [
{
"head": {
"text": "real-life trials",
"start": 81,
"end": 97
},
"tail": {
"text": "model-free RL",
"start": 20,
"end": 33
}
}
]
}
} | [] |
In this paper , we present a hybrid algorithm that requires only an approximate model , and only a small number of real-life trials . | {
"relations": {
"used for": [
{
"head": {
"text": "approximate model",
"start": 68,
"end": 85
},
"tail": {
"text": "hybrid algorithm",
"start": 29,
"end": 45
}
},
{
"head": {
"text": "r... | [] |
The key idea is to successively `` ground '' the policy evaluations using real-life trials , but to rely on the approximate model to suggest local changes . | {
"relations": {
"used for": [
{
"head": {
"text": "real-life trials",
"start": 74,
"end": 90
},
"tail": {
"text": "policy evaluations",
"start": 49,
"end": 67
}
}
]
}
} | [] |
Empirical results also demonstrate that -- when given only a crude model and a small number of real-life trials -- our algorithm can obtain near-optimal performance in the real system . | {
"relations": {
"conjunction": [
{
"head": {
"text": "crude model",
"start": 61,
"end": 72
},
"tail": {
"text": "real-life trials",
"start": 95,
"end": 111
}
}
],
"used for": [
{
"head": ... | [] |
Although every natural language system needs a computational lexicon , each system puts different amounts and types of information into its lexicon according to its individual needs . | {
"relations": {
"used for": [
{
"head": {
"text": "computational lexicon",
"start": 47,
"end": 68
},
"tail": {
"text": "natural language system",
"start": 15,
"end": 38
}
}
]
}
} | [] |
This paper presents our experience in planning and building COMPLEX , a computational lexicon designed to be a repository of shared lexical information for use by Natural Language Processing -LRB- NLP -RRB- systems . | {
"relations": {
"hyponym of": [
{
"head": {
"text": "COMPLEX",
"start": 60,
"end": 67
},
"tail": {
"text": "computational lexicon",
"start": 72,
"end": 93
}
}
],
"used for": [
{
"head": {... | [] |
Sentence planning is a set of inter-related but distinct tasks , one of which is sentence scoping , i.e. the choice of syntactic structure for elementary speech acts and the decision of how to combine them into one or more sentences . | {
"relations": {
"part of": [
{
"head": {
"text": "sentence scoping",
"start": 81,
"end": 97
},
"tail": {
"text": "tasks",
"start": 57,
"end": 62
}
}
],
"used for": [
{
"head": {
... | [] |
In this paper , we present SPoT , a sentence planner , and a new methodology for automatically training SPoT on the basis of feedback provided by human judges . | {
"relations": {
"hyponym of": [
{
"head": {
"text": "SPoT",
"start": 27,
"end": 31
},
"tail": {
"text": "sentence planner",
"start": 36,
"end": 52
}
}
],
"used for": [
{
"head": {
... | [] |
First , a very simple , randomized sentence-plan-generator -LRB- SPG -RRB- generates a potentially large list of possible sentence plans for a given text-plan input . | {
"relations": {
"used for": [
{
"head": {
"text": "randomized sentence-plan-generator -LRB- SPG -RRB-",
"start": 24,
"end": 74
},
"tail": {
"text": "sentence plans",
"start": 122,
"end": 136
}
},
{
... | [] |
Second , the sentence-plan-ranker -LRB- SPR -RRB- ranks the list of output sentence plans , and then selects the top-ranked plan . | {
"relations": {
"used for": [
{
"head": {
"text": "sentence-plan-ranker -LRB- SPR -RRB-",
"start": 13,
"end": 49
},
"tail": {
"text": "sentence plans",
"start": 75,
"end": 89
}
}
]
}
} | [] |
The SPR uses ranking rules automatically learned from training data . | {
"relations": {
"used for": [
{
"head": {
"text": "ranking rules",
"start": 13,
"end": 26
},
"tail": {
"text": "SPR",
"start": 4,
"end": 7
}
}
]
}
} | [] |
We show that the trained SPR learns to select a sentence plan whose rating on average is only 5 % worse than the top human-ranked sentence plan . | {
"relations": {
"used for": [
{
"head": {
"text": "SPR",
"start": 25,
"end": 28
},
"tail": {
"text": "sentence plan",
"start": 48,
"end": 61
}
}
],
"compare": [
{
"head": {
"tex... | [] |
We discuss maximum a posteriori estimation of continuous density hidden Markov models -LRB- CDHMM -RRB- . | {
"relations": {
"used for": [
{
"head": {
"text": "maximum a posteriori estimation",
"start": 11,
"end": 42
},
"tail": {
"text": "continuous density hidden Markov models -LRB- CDHMM -RRB-",
"start": 46,
"end": 103
}... | [] |
The classical MLE reestimation algorithms , namely the forward-backward algorithm and the segmental k-means algorithm , are expanded and reestimation formulas are given for HMM with Gaussian mixture observation densities . | {
"relations": {
"hyponym of": [
{
"head": {
"text": "forward-backward algorithm",
"start": 55,
"end": 81
},
"tail": {
"text": "MLE reestimation algorithms",
"start": 14,
"end": 41
}
},
{
"head": ... | [] |
Because of its adaptive nature , Bayesian learning serves as a unified approach for the following four speech recognition applications , namely parameter smoothing , speaker adaptation , speaker group modeling and corrective training . | {
"relations": {
"used for": [
{
"head": {
"text": "Bayesian learning",
"start": 33,
"end": 50
},
"tail": {
"text": "speech recognition applications",
"start": 103,
"end": 134
}
}
],
"hyponym of": [
... | [] |
New experimental results on all four applications are provided to show the effectiveness of the MAP estimation approach . | {
"relations": {
"evaluate for": [
{
"head": {
"text": "applications",
"start": 37,
"end": 49
},
"tail": {
"text": "MAP estimation approach",
"start": 96,
"end": 119
}
}
]
}
} | [] |
This paper describes a characters-based Chinese collocation system and discusses the advantages of it over a traditional word-based system . | {
"relations": {
"compare": [
{
"head": {
"text": "it",
"start": 99,
"end": 101
},
"tail": {
"text": "word-based system",
"start": 121,
"end": 138
}
}
]
}
} | [] |
Since wordbreaks are not conventionally marked in Chinese text corpora , a character-based collocation system has the dual advantages of avoiding pre-processing distortion and directly accessing sub-lexical information . | {
"relations": {
"feature of": [
{
"head": {
"text": "avoiding pre-processing distortion",
"start": 137,
"end": 171
},
"tail": {
"text": "character-based collocation system",
"start": 75,
"end": 109
}
},
... | [] |
Furthermore , word-based collocational properties can be obtained through an auxiliary module of automatic segmentation . | {
"relations": {
"used for": [
{
"head": {
"text": "auxiliary module of automatic segmentation",
"start": 77,
"end": 119
},
"tail": {
"text": "word-based collocational properties",
"start": 14,
"end": 49
}
}
... | [] |
This paper describes a method for utterance classification that does not require manual transcription of training data . | {
"relations": {
"used for": [
{
"head": {
"text": "method",
"start": 23,
"end": 29
},
"tail": {
"text": "utterance classification",
"start": 34,
"end": 58
}
}
]
}
} | [] |
The method combines domain independent acoustic models with off-the-shelf classifiers to give utterance classification performance that is surprisingly close to what can be achieved using conventional word-trigram recognition requiring manual transcription . | {
"relations": {
"used for": [
{
"head": {
"text": "method",
"start": 4,
"end": 10
},
"tail": {
"text": "utterance classification",
"start": 94,
"end": 118
}
},
{
"head": {
"text": "word... | [] |
In our method , unsupervised training is first used to train a phone n-gram model for a particular domain ; the output of recognition with this model is then passed to a phone-string classifier . | {
"relations": {
"part of": [
{
"head": {
"text": "unsupervised training",
"start": 16,
"end": 37
},
"tail": {
"text": "method",
"start": 7,
"end": 13
}
}
],
"used for": [
{
"head": {
... | [] |
The classification accuracy of the method is evaluated on three different spoken language system domains . | {
"relations": {
"evaluate for": [
{
"head": {
"text": "classification accuracy",
"start": 4,
"end": 27
},
"tail": {
"text": "method",
"start": 35,
"end": 41
}
},
{
"head": {
"text": "sp... | [] |
The Interval Algebra -LRB- IA -RRB- and a subset of the Region Connection Calculus -LRB- RCC -RRB- , namely RCC-8 , are the dominant Artificial Intelligence approaches for representing and reasoning about qualitative temporal and topological relations respectively . | {
"relations": {
"conjunction": [
{
"head": {
"text": "Interval Algebra -LRB- IA -RRB-",
"start": 4,
"end": 35
},
"tail": {
"text": "Region Connection Calculus -LRB- RCC -RRB-",
"start": 56,
"end": 98
}
}
]... | [] |
Such qualitative information can be formulated as a Qualitative Constraint Network -LRB- QCN -RRB- . | {
"relations": {
"used for": [
{
"head": {
"text": "Qualitative Constraint Network -LRB- QCN -RRB-",
"start": 52,
"end": 98
},
"tail": {
"text": "qualitative information",
"start": 5,
"end": 28
}
}
]
}
} | [] |
In this paper , we focus on the minimal labeling problem -LRB- MLP -RRB- and we propose an algorithm to efficiently derive all the feasible base relations of a QCN . | {
"relations": {
"used for": [
{
"head": {
"text": "algorithm",
"start": 91,
"end": 100
},
"tail": {
"text": "QCN",
"start": 160,
"end": 163
}
}
]
}
} | [] |
Our algorithm considers chordal QCNs and a new form of partial consistency which we define as ◆ G-consistency . | {
"relations": {
"part of": [
{
"head": {
"text": "chordal QCNs",
"start": 24,
"end": 36
},
"tail": {
"text": "algorithm",
"start": 4,
"end": 13
}
},
{
"head": {
"text": "partial consist... | [] |
Experi-mentations with QCNs of IA and RCC-8 show the importance and efficiency of this new approach . | {
"relations": {
"evaluate for": [
{
"head": {
"text": "QCNs of IA and RCC-8",
"start": 23,
"end": 43
},
"tail": {
"text": "approach",
"start": 91,
"end": 99
}
}
]
}
} | [] |
In this paper a morphological component with a limited capability to automatically interpret -LRB- and generate -RRB- derived words is presented . | {
"relations": {
"used for": [
{
"head": {
"text": "morphological component",
"start": 16,
"end": 39
},
"tail": {
"text": "derived words",
"start": 118,
"end": 131
}
}
]
}
} | [] |
The system combines an extended two-level morphology -LSB- Trost , 1991a ; Trost , 1991b -RSB- with a feature-based word grammar building on a hierarchical lexicon . | {
"relations": {
"used for": [
{
"head": {
"text": "two-level morphology",
"start": 32,
"end": 52
},
"tail": {
"text": "system",
"start": 4,
"end": 10
}
},
{
"head": {
"text": "hierarchi... | [] |
Polymorphemic stems not explicitly stored in the lexicon are given a compositional interpretation . | {
"relations": {
"feature of": [
{
"head": {
"text": "compositional interpretation",
"start": 69,
"end": 97
},
"tail": {
"text": "Polymorphemic stems",
"start": 0,
"end": 19
}
}
]
}
} | [] |
The system is implemented in CommonLisp and has been tested on examples from German derivation . | {
"relations": {
"used for": [
{
"head": {
"text": "CommonLisp",
"start": 29,
"end": 39
},
"tail": {
"text": "system",
"start": 4,
"end": 10
}
}
],
"evaluate for": [
{
"head": {
... | [] |
Four problems render vector space model -LRB- VSM -RRB- - based text classification approach ineffective : 1 -RRB- Many words within song lyrics actually contribute little to sentiment ; 2 -RRB- Nouns and verbs used to express sentiment are ambiguous ; 3 -RRB- Negations and modifiers around the sentiment keywords make ... | {
"relations": {
"conjunction": [
{
"head": {
"text": "Negations",
"start": 261,
"end": 270
},
"tail": {
"text": "modifiers",
"start": 275,
"end": 284
}
}
],
"used for": [
{
"head": {
... | [] |
To address these problems , the sentiment vector space model -LRB- s-VSM -RRB- is proposed to represent song lyric document . | {
"relations": {
"used for": [
{
"head": {
"text": "sentiment vector space model -LRB- s-VSM -RRB-",
"start": 32,
"end": 78
},
"tail": {
"text": "song lyric document",
"start": 104,
"end": 123
}
}
]
}
} | [] |
The preliminary experiments prove that the s-VSM model outperforms the VSM model in the lyric-based song sentiment classification task . | {
"relations": {
"compare": [
{
"head": {
"text": "s-VSM model",
"start": 43,
"end": 54
},
"tail": {
"text": "VSM model",
"start": 45,
"end": 54
}
}
],
"evaluate for": [
{
"head": {
... | [] |
We present an efficient algorithm for the redundancy elimination problem : Given an underspecified semantic representation -LRB- USR -RRB- of a scope ambiguity , compute an USR with fewer mutually equivalent readings . | {
"relations": {
"used for": [
{
"head": {
"text": "algorithm",
"start": 24,
"end": 33
},
"tail": {
"text": "redundancy elimination problem",
"start": 42,
"end": 72
}
},
{
"head": {
"tex... | [] |
The algorithm operates on underspecified chart representations which are derived from dominance graphs ; it can be applied to the USRs computed by large-scale grammars . | {
"relations": {
"used for": [
{
"head": {
"text": "algorithm",
"start": 4,
"end": 13
},
"tail": {
"text": "underspecified chart representations",
"start": 26,
"end": 62
}
},
{
"head": {
... | [] |
We evaluate the algorithm on a corpus , and show that it reduces the degree of ambiguity significantly while taking negligible runtime . | {
"relations": {
"used for": [
{
"head": {
"text": "it",
"start": 21,
"end": 23
},
"tail": {
"text": "degree of ambiguity",
"start": 69,
"end": 88
}
}
]
}
} | [] |
Currently several grammatical formalisms converge towards being declarative and towards utilizing context-free phrase-structure grammar as a backbone , e.g. LFG and PATR-II . | {
"relations": {
"used for": [
{
"head": {
"text": "context-free phrase-structure grammar",
"start": 98,
"end": 135
},
"tail": {
"text": "grammatical formalisms",
"start": 18,
"end": 40
}
}
],
"hyponym ... | [] |
Typically the processing of these formalisms is organized within a chart-parsing framework . | {
"relations": {
"feature of": [
{
"head": {
"text": "chart-parsing framework",
"start": 67,
"end": 90
},
"tail": {
"text": "formalisms",
"start": 34,
"end": 44
}
}
]
}
} | [] |
The aim of this paper is to provide a survey and a practical comparison of fundamental rule-invocation strategies within context-free chart parsing . | {
"relations": {
"part of": [
{
"head": {
"text": "rule-invocation strategies",
"start": 87,
"end": 113
},
"tail": {
"text": "context-free chart parsing",
"start": 121,
"end": 147
}
}
]
}
} | [] |
The present paper focusses on terminology structuring by lexical methods , which match terms on the basis on their content words , taking morphological variants into account . | {
"relations": {
"used for": [
{
"head": {
"text": "lexical methods",
"start": 57,
"end": 72
},
"tail": {
"text": "terminology structuring",
"start": 30,
"end": 53
}
}
]
}
} | [] |
Experiments are done on a ` flat ' list of terms obtained from an originally hierarchically-structured terminology : the French version of the US National Library of Medicine MeSH thesaurus . | {
"relations": {
"hyponym of": [
{
"head": {
"text": "US National Library of Medicine MeSH thesaurus",
"start": 143,
"end": 189
},
"tail": {
"text": "hierarchically-structured terminology",
"start": 77,
"end": 114
}
... | [] |
We compare the lexically-induced relations with the original MeSH relations : after a quantitative evaluation of their congruence through recall and precision metrics , we perform a qualitative , human analysis ofthe ` new ' relations not present in the MeSH . | {
"relations": {
"compare": [
{
"head": {
"text": "lexically-induced relations",
"start": 15,
"end": 42
},
"tail": {
"text": "MeSH relations",
"start": 61,
"end": 75
}
}
],
"evaluate for": [
{
... | [] |
In order to boost the translation quality of EBMT based on a small-sized bilingual corpus , we use an out-of-domain bilingual corpus and , in addition , the language model of an in-domain monolingual corpus . | {
"relations": {
"used for": [
{
"head": {
"text": "small-sized bilingual corpus",
"start": 61,
"end": 89
},
"tail": {
"text": "EBMT",
"start": 45,
"end": 49
}
},
{
"head": {
"text": "ou... | [] |
The two evaluation measures of the BLEU score and the NIST score demonstrated the effect of using an out-of-domain bilingual corpus and the possibility of using the language model . | {
"relations": {
"used for": [
{
"head": {
"text": "evaluation measures",
"start": 8,
"end": 27
},
"tail": {
"text": "out-of-domain bilingual corpus",
"start": 101,
"end": 131
}
}
],
"evaluate for": [
... | [] |
Diagrams are common tools for representing complex concepts , relationships and events , often when it would be difficult to portray the same information with natural images . | {
"relations": {
"used for": [
{
"head": {
"text": "Diagrams",
"start": 0,
"end": 8
},
"tail": {
"text": "complex concepts",
"start": 43,
"end": 59
}
},
{
"head": {
"text": "Diagrams",
... | [] |
Understanding natural images has been extensively studied in computer vision , while diagram understanding has received little attention . | {
"relations": {
"part of": [
{
"head": {
"text": "Understanding natural images",
"start": 0,
"end": 28
},
"tail": {
"text": "computer vision",
"start": 61,
"end": 76
}
}
],
"compare": [
{
... | [] |
In this paper , we study the problem of diagram interpretation and reasoning , the challenging task of identifying the structure of a diagram and the semantics of its constituents and their relationships . | {
"relations": {
"used for": [
{
"head": {
"text": "task",
"start": 95,
"end": 99
},
"tail": {
"text": "structure of a diagram",
"start": 119,
"end": 141
}
}
]
}
} | [] |
We introduce Diagram Parse Graphs -LRB- DPG -RRB- as our representation to model the structure of diagrams . | {
"relations": {
"used for": [
{
"head": {
"text": "Diagram Parse Graphs -LRB- DPG -RRB-",
"start": 13,
"end": 49
},
"tail": {
"text": "structure of diagrams",
"start": 85,
"end": 106
}
}
]
}
} | [] |
We define syntactic parsing of diagrams as learning to infer DPGs for diagrams and study semantic interpretation and reasoning of diagrams in the context of diagram question answering . | {
"relations": {
"used for": [
{
"head": {
"text": "syntactic parsing of diagrams",
"start": 10,
"end": 39
},
"tail": {
"text": "DPGs",
"start": 61,
"end": 65
}
},
{
"head": {
"text": "s... | [] |
We devise an LSTM-based method for syntactic parsing of diagrams and introduce a DPG-based attention model for diagram question answering . | {
"relations": {
"used for": [
{
"head": {
"text": "LSTM-based method",
"start": 13,
"end": 30
},
"tail": {
"text": "syntactic parsing of diagrams",
"start": 35,
"end": 64
}
},
{
"head": {
... | [] |
We compile a new dataset of diagrams with exhaustive annotations of constituents and relationships for over 5,000 diagrams and 15,000 questions and answers . | {
"relations": {
"feature of": [
{
"head": {
"text": "diagrams",
"start": 28,
"end": 36
},
"tail": {
"text": "dataset",
"start": 17,
"end": 24
}
}
]
}
} | [] |
Our results show the significance of our models for syntactic parsing and question answering in diagrams using DPGs . | {
"relations": {
"used for": [
{
"head": {
"text": "models",
"start": 41,
"end": 47
},
"tail": {
"text": "syntactic parsing and question answering in diagrams",
"start": 52,
"end": 104
}
},
{
"hea... | [] |
Previous change detection methods , focusing on detecting large-scale significant changes , can not do this well . | {
"relations": {
"used for": [
{
"head": {
"text": "change detection methods",
"start": 9,
"end": 33
},
"tail": {
"text": "detecting large-scale significant changes",
"start": 48,
"end": 89
}
}
]
}
} | [] |
This paper proposes a feasible end-to-end approach to this challenging problem . | {
"relations": {
"used for": [
{
"head": {
"text": "end-to-end approach",
"start": 31,
"end": 50
},
"tail": {
"text": "problem",
"start": 71,
"end": 78
}
}
]
}
} | [] |
Given two times observations , we formulate fine-grained change detection as a joint optimization problem of three related factors , i.e. , normal-aware lighting difference , camera geometry correction flow , and real scene change mask . | {
"relations": {
"used for": [
{
"head": {
"text": "joint optimization problem",
"start": 79,
"end": 105
},
"tail": {
"text": "fine-grained change detection",
"start": 44,
"end": 73
}
}
],
"feature of":... | [] |
We solve the three factors in a coarse-to-fine manner and achieve reliable change decision by rank minimization . | {
"relations": {
"used for": [
{
"head": {
"text": "coarse-to-fine manner",
"start": 32,
"end": 53
},
"tail": {
"text": "factors",
"start": 19,
"end": 26
}
},
{
"head": {
"text": "rank m... | [] |
We build three real-world datasets to benchmark fine-grained change detection of misaligned scenes under varied multiple lighting conditions . | {
"relations": {
"evaluate for": [
{
"head": {
"text": "real-world datasets",
"start": 15,
"end": 34
},
"tail": {
"text": "fine-grained change detection of misaligned scenes",
"start": 48,
"end": 98
}
}
],
... | [] |
Extensive experiments show the superior performance of our approach over state-of-the-art change detection methods and its ability to distinguish real scene changes from false ones caused by lighting variations . | {
"relations": {
"compare": [
{
"head": {
"text": "approach",
"start": 59,
"end": 67
},
"tail": {
"text": "change detection methods",
"start": 90,
"end": 114
}
}
],
"used for": [
{
"head":... | [] |
Automatic evaluation metrics for Machine Translation -LRB- MT -RRB- systems , such as BLEU or NIST , are now well established . | {
"relations": {
"evaluate for": [
{
"head": {
"text": "Automatic evaluation metrics",
"start": 0,
"end": 28
},
"tail": {
"text": "Machine Translation -LRB- MT -RRB- systems",
"start": 33,
"end": 75
}
}
],
... | [] |
Yet , they are scarcely used for the assessment of language pairs like English-Chinese or English-Japanese , because of the word segmentation problem . | {
"relations": {
"used for": [
{
"head": {
"text": "they",
"start": 6,
"end": 10
},
"tail": {
"text": "assessment of language pairs",
"start": 37,
"end": 65
}
}
],
"hyponym of": [
{
"head"... | [] |
This study establishes the equivalence between the standard use of BLEU in word n-grams and its application at the character level . | {
"relations": {
"used for": [
{
"head": {
"text": "BLEU",
"start": 67,
"end": 71
},
"tail": {
"text": "word n-grams",
"start": 75,
"end": 87
}
},
{
"head": {
"text": "BLEU",
"... | [] |
The use of BLEU at the character level eliminates the word segmentation problem : it makes it possible to directly compare commercial systems outputting unsegmented texts with , for instance , statistical MT systems which usually segment their outputs . | {
"relations": {
"used for": [
{
"head": {
"text": "BLEU",
"start": 11,
"end": 15
},
"tail": {
"text": "character level",
"start": 23,
"end": 38
}
},
{
"head": {
"text": "BLEU",
... | [] |
This paper proposes a series of modifications to the left corner parsing algorithm for context-free grammars . | {
"relations": {
"used for": [
{
"head": {
"text": "left corner parsing algorithm",
"start": 53,
"end": 82
},
"tail": {
"text": "context-free grammars",
"start": 87,
"end": 108
}
}
]
}
} | [] |
It is argued that the resulting algorithm is both efficient and flexible and is , therefore , a good choice for the parser used in a natural language interface . | {
"relations": {
"used for": [
{
"head": {
"text": "algorithm",
"start": 32,
"end": 41
},
"tail": {
"text": "parser",
"start": 116,
"end": 122
}
},
{
"head": {
"text": "parser",
... | [] |
This paper presents a novel statistical singing voice conversion -LRB- SVC -RRB- technique with direct waveform modification based on the spectrum differential that can convert voice timbre of a source singer into that of a target singer without using a vocoder to generate converted singing voice waveforms . | {
"relations": {
"used for": [
{
"head": {
"text": "direct waveform modification",
"start": 96,
"end": 124
},
"tail": {
"text": "statistical singing voice conversion -LRB- SVC -RRB- technique",
"start": 28,
"end": 90
... | [] |
SVC makes it possible to convert singing voice characteristics of an arbitrary source singer into those of an arbitrary target singer . | {
"relations": {
"used for": [
{
"head": {
"text": "SVC",
"start": 0,
"end": 3
},
"tail": {
"text": "singing voice characteristics",
"start": 33,
"end": 62
}
}
]
}
} | [] |
However , speech quality of the converted singing voice is significantly degraded compared to that of a natural singing voice due to various factors , such as analysis and modeling errors in the vocoder-based framework . | {
"relations": {
"evaluate for": [
{
"head": {
"text": "speech quality",
"start": 10,
"end": 24
},
"tail": {
"text": "converted singing voice",
"start": 32,
"end": 55
}
},
{
"head": {
"t... | [] |
The differential spectral feature is directly estimated using a differential Gaussian mixture model -LRB- GMM -RRB- that is analytically derived from the traditional GMM used as a conversion model in the conventional SVC . | {
"relations": {
"used for": [
{
"head": {
"text": "differential Gaussian mixture model -LRB- GMM -RRB-",
"start": 64,
"end": 115
},
"tail": {
"text": "differential spectral feature",
"start": 4,
"end": 33
}
},... | [] |
The experimental results demonstrate that the proposed method makes it possible to significantly improve speech quality in the converted singing voice while preserving the conversion accuracy of singer identity compared to the conventional SVC . | {
"relations": {
"compare": [
{
"head": {
"text": "method",
"start": 55,
"end": 61
},
"tail": {
"text": "SVC",
"start": 240,
"end": 243
}
}
],
"evaluate for": [
{
"head": {
"text... | [] |
During late-2013 through early-2014 NIST coordinated a special i-vector challenge based on data used in previous NIST Speaker Recognition Evaluations -LRB- SREs -RRB- . | {
"relations": {
"used for": [
{
"head": {
"text": "NIST Speaker Recognition Evaluations -LRB- SREs -RRB-",
"start": 113,
"end": 166
},
"tail": {
"text": "i-vector challenge",
"start": 63,
"end": 81
}
}
]
... | [] |
Unlike evaluations in the SRE series , the i-vector challenge was run entirely online and used fixed-length feature vectors projected into a low-dimensional space -LRB- i-vectors -RRB- rather than audio recordings . | {
"relations": {
"used for": [
{
"head": {
"text": "fixed-length feature vectors",
"start": 95,
"end": 123
},
"tail": {
"text": "low-dimensional space -LRB- i-vectors -RRB-",
"start": 141,
"end": 184
}
},
... | [] |
Compared to the 2012 SRE , the i-vector challenge saw an increase in the number of participants by nearly a factor of two , and a two orders of magnitude increase in the number of systems submitted for evaluation . | {
"relations": {
"compare": [
{
"head": {
"text": "SRE",
"start": 21,
"end": 24
},
"tail": {
"text": "i-vector challenge",
"start": 31,
"end": 49
}
}
]
}
} | [] |
Initial results indicate the leading system achieved an approximate 37 % improvement relative to the baseline system . | {
"relations": {
"compare": [
{
"head": {
"text": "leading system",
"start": 29,
"end": 43
},
"tail": {
"text": "baseline system",
"start": 101,
"end": 116
}
}
]
}
} | [] |
Theoretical research in the area of machine translation usually involves the search for and creation of an appropriate formalism . | {
"relations": {
"used for": [
{
"head": {
"text": "formalism",
"start": 119,
"end": 128
},
"tail": {
"text": "machine translation",
"start": 36,
"end": 55
}
}
]
}
} | [] |
In this paper , we will introduce the anaphoric component of the Mimo formalism . | {
"relations": {
"part of": [
{
"head": {
"text": "anaphoric component",
"start": 38,
"end": 57
},
"tail": {
"text": "Mimo formalism",
"start": 65,
"end": 79
}
}
]
}
} | [] |
In Mimo , the translation of anaphoric relations is compositional . | {
"relations": {
"used for": [
{
"head": {
"text": "Mimo",
"start": 3,
"end": 7
},
"tail": {
"text": "translation of anaphoric relations",
"start": 14,
"end": 48
}
}
]
}
} | [] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.