input
stringlengths
37
565
output
dict
schema
listlengths
0
0
In a motorized vehicle a number of easily measurable signals with frequency components related to the rotational speed of the engine can be found , e.g. , vibrations , electrical system voltage level , and ambient sound .
{ "relations": { "used for": [ { "head": { "text": "frequency components", "start": 66, "end": 86 }, "tail": { "text": "measurable signals", "start": 42, "end": 60 } } ], "feature of": [ { ...
[]
These signals could potentially be used to estimate the speed and related states of the vehicle .
{ "relations": { "used for": [ { "head": { "text": "signals", "start": 6, "end": 13 }, "tail": { "text": "speed and related states of the vehicle", "start": 56, "end": 95 } } ] } }
[]
Unfortunately , such estimates would typically require the relations -LRB- scale factors -RRB- between the frequency components and the speed for different gears to be known .
{ "relations": { "conjunction": [ { "head": { "text": "frequency components", "start": 107, "end": 127 }, "tail": { "text": "speed", "start": 136, "end": 141 } } ], "feature of": [ { "head...
[]
Consequently , in this article we look at the problem of estimating these gear scale factors from training data consisting only of speed measurements and measurements of the signal in question .
{ "relations": { "used for": [ { "head": { "text": "training data", "start": 98, "end": 111 }, "tail": { "text": "gear scale factors", "start": 74, "end": 92 } } ] } }
[]
The estimation problem is formulated as a maximum likelihood estimation problem and heuristics is used to find initial values for a numerical evaluation of the estimator .
{ "relations": { "used for": [ { "head": { "text": "maximum likelihood estimation problem", "start": 42, "end": 79 }, "tail": { "text": "estimation problem", "start": 4, "end": 22 } }, { "head": {...
[]
Finally , a measurement campaign is conducted and the functionality of the estimation method is verified on real data .
{ "relations": { "evaluate for": [ { "head": { "text": "real data", "start": 108, "end": 117 }, "tail": { "text": "estimation method", "start": 75, "end": 92 } } ] } }
[]
LPC based speech coders operating at bit rates below 3.0 kbits/sec are usually associated with buzzy or metallic artefacts in the synthetic speech .
{ "relations": { "feature of": [ { "head": { "text": "bit rates", "start": 37, "end": 46 }, "tail": { "text": "LPC based speech coders", "start": 0, "end": 23 } }, { "head": { "text": "b...
[]
In this paper a new LPC vocoder is presented which splits the LPC excitation into two frequency bands using a variable cutoff frequency .
{ "relations": { "used for": [ { "head": { "text": "variable cutoff frequency", "start": 110, "end": 135 }, "tail": { "text": "LPC excitation", "start": 62, "end": 76 } }, { "head": { "t...
[]
In doing so the coder 's performance during both mixed voicing speech and speech containing acoustic noise is greatly improved , producing soft natural sounding speech .
{ "relations": { "used for": [ { "head": { "text": "coder", "start": 16, "end": 21 }, "tail": { "text": "soft natural sounding speech", "start": 139, "end": 167 } }, { "head": { "text": ...
[]
The paper also describes new parameter determination and quantisation techniques vital to the operation of this coder at such low bit rates .
{ "relations": { "conjunction": [ { "head": { "text": "parameter determination", "start": 29, "end": 52 }, "tail": { "text": "quantisation techniques", "start": 57, "end": 80 } } ], "used for": [ ...
[]
We consider a problem of blind source separation from a set of instantaneous linear mixtures , where the mixing matrix is unknown .
{ "relations": { "used for": [ { "head": { "text": "instantaneous linear mixtures", "start": 63, "end": 92 }, "tail": { "text": "blind source separation", "start": 25, "end": 48 } } ] } }
[]
It was discovered recently , that exploiting the sparsity of sources in an appropriate representation according to some signal dictionary , dramatically improves the quality of separation .
{ "relations": { "used for": [ { "head": { "text": "signal dictionary", "start": 120, "end": 137 }, "tail": { "text": "sparsity of sources", "start": 49, "end": 68 } } ], "evaluate for": [ { ...
[]
In this work we use the property of multi scale transforms , such as wavelet or wavelet packets , to decompose signals into sets of local features with various degrees of sparsity .
{ "relations": { "hyponym of": [ { "head": { "text": "wavelet or wavelet packets", "start": 69, "end": 95 }, "tail": { "text": "multi scale transforms", "start": 36, "end": 58 } } ] } }
[]
The performance of the algorithm is verified on noise-free and noisy data .
{ "relations": { "evaluate for": [ { "head": { "text": "noise-free and noisy data", "start": 48, "end": 73 }, "tail": { "text": "algorithm", "start": 23, "end": 32 } } ] } }
[]
Experiments with simulated signals , musical sounds and images demonstrate significant improvement of separation quality over previously reported results .
{ "relations": { "conjunction": [ { "head": { "text": "simulated signals", "start": 17, "end": 34 }, "tail": { "text": "musical sounds", "start": 37, "end": 51 } }, { "head": { "text": "...
[]
In this paper , we explore multilingual feature-level data sharing via Deep Neural Network -LRB- DNN -RRB- stacked bottleneck features .
{ "relations": { "used for": [ { "head": { "text": "Deep Neural Network -LRB- DNN -RRB- stacked bottleneck features", "start": 71, "end": 134 }, "tail": { "text": "multilingual feature-level data sharing", "start": 27, "end"...
[]
Given a set of available source languages , we apply language identification to pick the language most similar to the target language , for more efficient use of multilingual resources .
{ "relations": { "used for": [ { "head": { "text": "language identification", "start": 53, "end": 76 }, "tail": { "text": "multilingual resources", "start": 162, "end": 184 } } ] } }
[]
Our experiments with IARPA-Babel languages show that bottleneck features trained on the most similar source language perform better than those trained on all available source languages .
{ "relations": { "compare": [ { "head": { "text": "those", "start": 137, "end": 142 }, "tail": { "text": "bottleneck features", "start": 53, "end": 72 } } ] } }
[]
Further analysis suggests that only data similar to the target language is useful for multilingual training .
{ "relations": { "used for": [ { "head": { "text": "data", "start": 36, "end": 40 }, "tail": { "text": "multilingual training", "start": 86, "end": 107 } } ] } }
[]
This article introduces a bidirectional grammar generation system called feature structure-directed generation , developed for a dialogue translation system .
{ "relations": { "used for": [ { "head": { "text": "bidirectional grammar generation system", "start": 26, "end": 65 }, "tail": { "text": "dialogue translation system", "start": 129, "end": 156 } }, { ...
[]
The system utilizes typed feature structures to control the top-down derivation in a declarative way .
{ "relations": { "used for": [ { "head": { "text": "typed feature structures", "start": 20, "end": 44 }, "tail": { "text": "system", "start": 4, "end": 10 } }, { "head": { "text": "typed...
[]
This generation system also uses disjunctive feature structures to reduce the number of copies of the derivation tree .
{ "relations": { "used for": [ { "head": { "text": "disjunctive feature structures", "start": 33, "end": 63 }, "tail": { "text": "generation system", "start": 5, "end": 22 } }, { "head": { ...
[]
The grammar for this generator is designed to properly generate the speaker 's intention in a telephone dialogue .
{ "relations": { "used for": [ { "head": { "text": "grammar", "start": 4, "end": 11 }, "tail": { "text": "generator", "start": 21, "end": 30 } }, { "head": { "text": "grammar", ...
[]
Automatic image annotation is a newly developed and promising technique to provide semantic image retrieval via text descriptions .
{ "relations": { "used for": [ { "head": { "text": "Automatic image annotation", "start": 0, "end": 26 }, "tail": { "text": "semantic image retrieval", "start": 83, "end": 107 } }, { "head": { ...
[]
It concerns a process of automatically labeling the image contents with a pre-defined set of keywords which are exploited to represent the image semantics .
{ "relations": { "used for": [ { "head": { "text": "keywords", "start": 93, "end": 101 }, "tail": { "text": "automatically labeling the image contents", "start": 25, "end": 66 } }, { "head": { ...
[]
A Maximum Entropy Model-based approach to the task of automatic image annotation is proposed in this paper .
{ "relations": { "used for": [ { "head": { "text": "Maximum Entropy Model-based approach", "start": 2, "end": 38 }, "tail": { "text": "automatic image annotation", "start": 54, "end": 80 } } ] } }
[]
In the phase of training , a basic visual vocabulary consisting of blob-tokens to describe the image content is generated at first ; then the statistical relationship is modeled between the blob-tokens and keywords by a Maximum Entropy Model constructed from the training set of labeled images .
{ "relations": { "used for": [ { "head": { "text": "visual vocabulary", "start": 35, "end": 52 }, "tail": { "text": "image content", "start": 95, "end": 108 } }, { "head": { "text": "Max...
[]
In the phase of annotation , for an unlabeled image , the most likely associated keywords are predicted in terms of the blob-token set extracted from the given image .
{ "relations": { "used for": [ { "head": { "text": "blob-token set", "start": 120, "end": 134 }, "tail": { "text": "keywords", "start": 81, "end": 89 } } ] } }
[]
We carried out experiments on a medium-sized image collection with about 5000 images from Corel Photo CDs .
{ "relations": { "used for": [ { "head": { "text": "Corel Photo CDs", "start": 90, "end": 105 }, "tail": { "text": "medium-sized image collection", "start": 32, "end": 61 } } ] } }
[]
The experimental results demonstrated that the annotation performance of this method outperforms some traditional annotation methods by about 8 % in mean precision , showing a potential of the Maximum Entropy Model in the task of automatic image annotation .
{ "relations": { "used for": [ { "head": { "text": "method", "start": 78, "end": 84 }, "tail": { "text": "annotation", "start": 47, "end": 57 } }, { "head": { "text": "annotation methods...
[]
However most of the works found in the literature have focused on identifying and understanding temporal expressions in newswire texts .
{ "relations": { "feature of": [ { "head": { "text": "newswire texts", "start": 120, "end": 134 }, "tail": { "text": "temporal expressions", "start": 96, "end": 116 } } ] } }
[]
In this paper we report our work on anchoring temporal expressions in a novel genre , emails .
{ "relations": { "feature of": [ { "head": { "text": "emails", "start": 86, "end": 92 }, "tail": { "text": "temporal expressions", "start": 46, "end": 66 } } ] } }
[]
The highly under-specified nature of these expressions fits well with our constraint-based representation of time , Time Calculus for Natural Language -LRB- TCNL -RRB- .
{ "relations": { "hyponym of": [ { "head": { "text": "Time Calculus for Natural Language -LRB- TCNL -RRB-", "start": 116, "end": 167 }, "tail": { "text": "constraint-based representation of time", "start": 74, "end": 113 ...
[]
We have developed and evaluated a Temporal Expression Anchoror -LRB- TEA -RRB- , and the result shows that it performs significantly better than the baseline , and compares favorably with some of the closely related work .
{ "relations": { "compare": [ { "head": { "text": "it", "start": 107, "end": 109 }, "tail": { "text": "baseline", "start": 149, "end": 157 } } ] } }
[]
We address the problem of populating object category detection datasets with dense , per-object 3D reconstructions , bootstrapped from class labels , ground truth figure-ground segmentations and a small set of keypoint annotations .
{ "relations": { "used for": [ { "head": { "text": "object category detection datasets", "start": 37, "end": 71 }, "tail": { "text": "per-object 3D reconstructions", "start": 85, "end": 114 } }, { ...
[]
Our proposed algorithm first estimates camera viewpoint using rigid structure-from-motion , then reconstructs object shapes by optimizing over visual hull proposals guided by loose within-class shape similarity assumptions .
{ "relations": { "used for": [ { "head": { "text": "algorithm", "start": 13, "end": 22 }, "tail": { "text": "camera viewpoint", "start": 39, "end": 55 } }, { "head": { "text": "algorithm...
[]
We show that our method is able to produce convincing per-object 3D reconstructions on one of the most challenging existing object-category detection datasets , PASCAL VOC .
{ "relations": { "used for": [ { "head": { "text": "method", "start": 17, "end": 23 }, "tail": { "text": "per-object 3D reconstructions", "start": 54, "end": 83 } }, { "head": { "text": ...
[]
Probabilistic models have been previously shown to be efficient and effective for modeling and recognition of human motion .
{ "relations": { "used for": [ { "head": { "text": "Probabilistic models", "start": 0, "end": 20 }, "tail": { "text": "modeling and recognition of human motion", "start": 82, "end": 122 } } ] } }
[]
In particular we focus on methods which represent the human motion model as a triangulated graph .
{ "relations": { "used for": [ { "head": { "text": "triangulated graph", "start": 78, "end": 96 }, "tail": { "text": "human motion model", "start": 54, "end": 72 } } ] } }
[]
Previous approaches learned models based just on positions and velocities of the body parts while ignoring their appearance .
{ "relations": { "used for": [ { "head": { "text": "positions", "start": 49, "end": 58 }, "tail": { "text": "models", "start": 28, "end": 34 } }, { "head": { "text": "velocities", ...
[]
Moreover , a heuristic approach was commonly used to obtain translation invariance .
{ "relations": { "used for": [ { "head": { "text": "heuristic approach", "start": 13, "end": 31 }, "tail": { "text": "translation invariance", "start": 60, "end": 82 } } ] } }
[]
In this paper we suggest an improved approach for learning such models and using them for human motion recognition .
{ "relations": { "used for": [ { "head": { "text": "approach", "start": 37, "end": 45 }, "tail": { "text": "models", "start": 64, "end": 70 } }, { "head": { "text": "them", "st...
[]
The suggested approach combines multiple cues , i.e. , positions , velocities and appearance into both the learning and detection phases .
{ "relations": { "used for": [ { "head": { "text": "approach", "start": 14, "end": 22 }, "tail": { "text": "learning and detection phases", "start": 107, "end": 136 } } ], "hyponym of": [ { ...
[]
Furthermore , we introduce global variables in the model , which can represent global properties such as translation , scale or viewpoint .
{ "relations": { "used for": [ { "head": { "text": "global variables", "start": 27, "end": 43 }, "tail": { "text": "model", "start": 51, "end": 56 } }, { "head": { "text": "global variab...
[]
The model is learned in an unsupervised manner from un-labelled data .
{ "relations": { "used for": [ { "head": { "text": "unsupervised manner", "start": 27, "end": 46 }, "tail": { "text": "model", "start": 4, "end": 9 } }, { "head": { "text": "un-labelled ...
[]
We show that the suggested hybrid proba-bilistic model -LRB- which combines global variables , like translation , with local variables , like relative positions and appearances of body parts -RRB- , leads to : -LRB- i -RRB- faster convergence of learning phase , -LRB- ii -RRB- robustness to occlusions , and , -LRB- iii...
{ "relations": { "used for": [ { "head": { "text": "global variables", "start": 76, "end": 92 }, "tail": { "text": "hybrid proba-bilistic model", "start": 27, "end": 54 } } ], "hyponym of": [ { ...
[]
Factor analysis and principal components analysis can be used to model linear relationships between observed variables and linearly map high-dimensional data to a lower-dimensional hidden space .
{ "relations": { "conjunction": [ { "head": { "text": "Factor analysis", "start": 0, "end": 15 }, "tail": { "text": "principal components analysis", "start": 20, "end": 49 } } ], "used for": [ { ...
[]
We describe a nonlinear generalization of factor analysis , called `` product analy-sis '' , that models the observed variables as a linear combination of products of normally distributed hidden variables .
{ "relations": { "used for": [ { "head": { "text": "nonlinear generalization of factor analysis", "start": 14, "end": 57 }, "tail": { "text": "observed variables", "start": 109, "end": 127 } }, { ...
[]
Just as factor analysis can be viewed as unsupervised linear regression on unobserved , normally distributed hidden variables , product analysis can be viewed as unsupervised linear regression on products of unobserved , normally distributed hidden variables .
{ "relations": { "used for": [ { "head": { "text": "unsupervised linear regression", "start": 41, "end": 71 }, "tail": { "text": "factor analysis", "start": 8, "end": 23 } }, { "head": { ...
[]
The mapping between the data and the hidden space is nonlinear , so we use an approximate variational technique for inference and learning .
{ "relations": { "used for": [ { "head": { "text": "approximate variational technique", "start": 78, "end": 111 }, "tail": { "text": "inference", "start": 116, "end": 125 } }, { "head": { ...
[]
Since product analysis is a generalization of factor analysis , product analysis always finds a higher data likelihood than factor analysis .
{ "relations": { "hyponym of": [ { "head": { "text": "product analysis", "start": 6, "end": 22 }, "tail": { "text": "generalization of factor analysis", "start": 28, "end": 61 } } ], "compare": [ ...
[]
We give results on pattern recognition and illumination-invariant image clustering .
{ "relations": { "conjunction": [ { "head": { "text": "pattern recognition", "start": 19, "end": 38 }, "tail": { "text": "illumination-invariant image clustering", "start": 43, "end": 82 } } ] } }
[]
This paper describes a domain independent strategy for the multimedia articulation of answers elicited by a natural language interface to database query applications .
{ "relations": { "used for": [ { "head": { "text": "domain independent strategy", "start": 23, "end": 50 }, "tail": { "text": "multimedia articulation of answers", "start": 59, "end": 93 } }, { "h...
[]
Multimedia answers include videodisc images and heuristically-produced complete sentences in text or text-to-speech form .
{ "relations": { "part of": [ { "head": { "text": "videodisc images", "start": 27, "end": 43 }, "tail": { "text": "Multimedia answers", "start": 0, "end": 18 } } ] } }
[]
Deictic reference and feedback about the discourse are enabled .
{ "relations": { "conjunction": [ { "head": { "text": "Deictic reference", "start": 0, "end": 17 }, "tail": { "text": "feedback", "start": 22, "end": 30 } } ], "feature of": [ { "head": { ...
[]
The LOGON MT demonstrator assembles independently valuable general-purpose NLP components into a machine translation pipeline that capitalizes on output quality .
{ "relations": { "used for": [ { "head": { "text": "LOGON MT demonstrator", "start": 4, "end": 25 }, "tail": { "text": "general-purpose NLP components", "start": 59, "end": 89 } } ], "part of": [ ...
[]
The demonstrator embodies an interesting combination of hand-built , symbolic resources and stochastic processes .
{ "relations": { "part of": [ { "head": { "text": "hand-built , symbolic resources", "start": 56, "end": 87 }, "tail": { "text": "demonstrator", "start": 4, "end": 16 } }, { "head": { "t...
[]
We describe both the syntax and semantics of a general propositional language of context , and give a Hilbert style proof system for this language .
{ "relations": { "conjunction": [ { "head": { "text": "syntax", "start": 21, "end": 27 }, "tail": { "text": "semantics", "start": 32, "end": 41 } } ], "feature of": [ { "head": { ...
[]
A propositional logic of context extends classical propositional logic in two ways .
{ "relations": { "used for": [ { "head": { "text": "classical propositional logic", "start": 41, "end": 70 }, "tail": { "text": "propositional logic of context", "start": 2, "end": 32 } } ] } }
[]
Image matching is a fundamental problem in Computer Vision .
{ "relations": { "hyponym of": [ { "head": { "text": "Image matching", "start": 0, "end": 14 }, "tail": { "text": "Computer Vision", "start": 43, "end": 58 } } ] } }
[]
In the context of feature-based matching , SIFT and its variants have long excelled in a wide array of applications .
{ "relations": { "used for": [ { "head": { "text": "SIFT", "start": 43, "end": 47 }, "tail": { "text": "feature-based matching", "start": 18, "end": 40 } } ] } }
[]
However , for ultra-wide baselines , as in the case of aerial images captured under large camera rotations , the appearance variation goes beyond the reach of SIFT and RANSAC .
{ "relations": { "feature of": [ { "head": { "text": "large camera rotations", "start": 84, "end": 106 }, "tail": { "text": "aerial images", "start": 55, "end": 68 } } ], "conjunction": [ { ...
[]
In this paper we propose a data-driven , deep learning-based approach that sidesteps local correspondence by framing the problem as a classification task .
{ "relations": { "used for": [ { "head": { "text": "classification task", "start": 134, "end": 153 }, "tail": { "text": "problem", "start": 121, "end": 128 } } ] } }
[]
We train our models on a dataset of urban aerial imagery consisting of ` same ' and ` different ' pairs , collected for this purpose , and characterize the problem via a human study with annotations from Amazon Mechanical Turk .
{ "relations": { "used for": [ { "head": { "text": "dataset of urban aerial imagery", "start": 25, "end": 56 }, "tail": { "text": "models", "start": 13, "end": 19 } }, { "head": { "text"...
[]
We demonstrate that our models outperform the state-of-the-art on ultra-wide baseline matching and approach human accuracy .
{ "relations": { "compare": [ { "head": { "text": "models", "start": 24, "end": 30 }, "tail": { "text": "state-of-the-art", "start": 46, "end": 62 } }, { "head": { "text": "models", ...
[]
We argue that a more sophisticated and fine-grained annotation in the tree-bank would have very positve effects on stochastic parsers trained on the tree-bank and on grammars induced from the treebank , and it would make the treebank more valuable as a source of data for theoretical linguistic investigations .
{ "relations": { "used for": [ { "head": { "text": "fine-grained annotation", "start": 39, "end": 62 }, "tail": { "text": "stochastic parsers", "start": 115, "end": 133 } }, { "head": { ...
[]
The information gained from corpus research and the analyses that are proposed are realized in the framework of SILVA , a parsing and extraction tool for German text corpora .
{ "relations": { "hyponym of": [ { "head": { "text": "SILVA", "start": 112, "end": 117 }, "tail": { "text": "parsing and extraction tool", "start": 122, "end": 149 } } ], "used for": [ { "...
[]
While paraphrasing is critical both for interpretation and generation of natural language , current systems use manual or semi-automatic methods to collect paraphrases .
{ "relations": { "used for": [ { "head": { "text": "paraphrasing", "start": 6, "end": 18 }, "tail": { "text": "interpretation and generation of natural language", "start": 40, "end": 89 } }, { "he...
[]
We present an unsupervised learning algorithm for identification of paraphrases from a corpus of multiple English translations of the same source text .
{ "relations": { "used for": [ { "head": { "text": "unsupervised learning algorithm", "start": 14, "end": 45 }, "tail": { "text": "identification of paraphrases", "start": 50, "end": 79 } }, { "he...
[]
Our approach yields phrasal and single word lexical paraphrases as well as syntactic paraphrases .
{ "relations": { "used for": [ { "head": { "text": "approach", "start": 4, "end": 12 }, "tail": { "text": "phrasal and single word lexical paraphrases", "start": 20, "end": 63 } }, { "head": { ...
[]
An efficient bit-vector-based CKY-style parser for context-free parsing is presented .
{ "relations": { "used for": [ { "head": { "text": "bit-vector-based CKY-style parser", "start": 13, "end": 46 }, "tail": { "text": "context-free parsing", "start": 51, "end": 71 } } ] } }
[]
The parser computes a compact parse forest representation of the complete set of possible analyses for large treebank grammars and long input sentences .
{ "relations": { "used for": [ { "head": { "text": "parser", "start": 4, "end": 10 }, "tail": { "text": "parse forest representation", "start": 30, "end": 57 } }, { "head": { "text": "pa...
[]
The parser uses bit-vector operations to parallelise the basic parsing operations .
{ "relations": { "used for": [ { "head": { "text": "bit-vector operations", "start": 16, "end": 37 }, "tail": { "text": "parser", "start": 4, "end": 10 } } ] } }
[]
In this paper , we propose a partially-blurred-image classification and analysis framework for automatically detecting images containing blurred regions and recognizing the blur types for those regions without needing to perform blur kernel estimation and image deblurring .
{ "relations": { "used for": [ { "head": { "text": "partially-blurred-image classification and analysis framework", "start": 29, "end": 90 }, "tail": { "text": "automatically detecting images", "start": 95, "end": 125 ...
[]
We develop several blur features modeled by image color , gradient , and spectrum information , and use feature parameter training to robustly classify blurred images .
{ "relations": { "used for": [ { "head": { "text": "image color", "start": 44, "end": 55 }, "tail": { "text": "blur features", "start": 19, "end": 32 } }, { "head": { "text": "gradient",...
[]
Our blur detection is based on image patches , making region-wise training and classification in one image efficient .
{ "relations": { "used for": [ { "head": { "text": "image patches", "start": 31, "end": 44 }, "tail": { "text": "blur detection", "start": 4, "end": 18 } }, { "head": { "text": "region-w...
[]
Extensive experiments show that our method works satisfactorily on challenging image data , which establishes a technical foundation for solving several computer vision problems , such as motion analysis and image restoration , using the blur information .
{ "relations": { "used for": [ { "head": { "text": "method", "start": 36, "end": 42 }, "tail": { "text": "computer vision problems", "start": 153, "end": 177 } }, { "head": { "text": "bl...
[]
We have recently reported on two new word-sense disambiguation systems , one trained on bilingual material -LRB- the Canadian Hansards -RRB- and the other trained on monolingual material -LRB- Roget 's Thesaurus and Grolier 's Encyclopedia -RRB- .
{ "relations": { "hyponym of": [ { "head": { "text": "one", "start": 73, "end": 76 }, "tail": { "text": "word-sense disambiguation systems", "start": 37, "end": 70 } }, { "head": { "text...
[]
In addition , it could also be used to help evaluate disambiguation algorithms that did not make use of the discourse constraint .
{ "relations": { "evaluate for": [ { "head": { "text": "it", "start": 6, "end": 8 }, "tail": { "text": "disambiguation algorithms", "start": 53, "end": 78 } } ] } }
[]
We study and compare two novel embedding methods for segmenting feature points of piece-wise planar structures from two -LRB- uncalibrated -RRB- perspective images .
{ "relations": { "used for": [ { "head": { "text": "embedding methods", "start": 31, "end": 48 }, "tail": { "text": "segmenting feature points of piece-wise planar structures", "start": 53, "end": 110 } } ]...
[]
We show that a set of different homographies can be embedded in different ways to a higher-dimensional real or complex space , so that each homography corresponds to either a complex bilinear form or a real quadratic form .
{ "relations": { "feature of": [ { "head": { "text": "higher-dimensional real or complex space", "start": 84, "end": 124 }, "tail": { "text": "homographies", "start": 32, "end": 44 } }, { "head": ...
[]
We give a closed-form segmentation solution for each case by utilizing these properties based on subspace-segmentation methods .
{ "relations": { "used for": [ { "head": { "text": "subspace-segmentation methods", "start": 97, "end": 126 }, "tail": { "text": "closed-form segmentation solution", "start": 10, "end": 43 } } ] } }
[]
These theoretical results show that one can intrinsically segment a piece-wise planar scene from 2-D images without explicitly performing any 3-D reconstruction .
{ "relations": { "feature of": [ { "head": { "text": "2-D images", "start": 97, "end": 107 }, "tail": { "text": "piece-wise planar scene", "start": 68, "end": 91 } } ] } }
[]
Background maintenance is a frequent element of video surveillance systems .
{ "relations": { "part of": [ { "head": { "text": "Background maintenance", "start": 0, "end": 22 }, "tail": { "text": "video surveillance systems", "start": 48, "end": 74 } } ] } }
[]
We develop Wallflower , a three-component system for background maintenance : the pixel-level component performs Wiener filtering to make probabilistic predictions of the expected background ; the region-level component fills in homogeneous regions of foreground objects ; and the frame-level component detects sudden , ...
{ "relations": { "used for": [ { "head": { "text": "three-component system", "start": 26, "end": 48 }, "tail": { "text": "background maintenance", "start": 53, "end": 75 } }, { "head": { ...
[]
We compare our system with 8 other background subtraction algorithms .
{ "relations": { "compare": [ { "head": { "text": "system", "start": 15, "end": 21 }, "tail": { "text": "background subtraction algorithms", "start": 35, "end": 68 } } ] } }
[]
Wallflower is shown to outperform previous algorithms by handling a greater set of the difficult situations that can occur .
{ "relations": { "compare": [ { "head": { "text": "Wallflower", "start": 0, "end": 10 }, "tail": { "text": "algorithms", "start": 43, "end": 53 } } ] } }
[]
Finally , we analyze the experimental results and propose normative principles for background maintenance .
{ "relations": { "used for": [ { "head": { "text": "normative principles", "start": 58, "end": 78 }, "tail": { "text": "background maintenance", "start": 83, "end": 105 } } ] } }
[]
Is it possible to use out-of-domain acoustic training data to improve a speech recognizer 's performance on a speciic , independent application ?
{ "relations": { "used for": [ { "head": { "text": "out-of-domain acoustic training data", "start": 22, "end": 58 }, "tail": { "text": "speech recognizer", "start": 72, "end": 89 } } ] } }
[]
In our experiments , we use Wallstreet Journal -LRB- WSJ -RRB- data to train a recognizer , which is adapted and evaluated in the Phonebook domain .
{ "relations": { "used for": [ { "head": { "text": "Wallstreet Journal -LRB- WSJ -RRB- data", "start": 28, "end": 67 }, "tail": { "text": "recognizer", "start": 79, "end": 89 } } ], "evaluate for": [ ...
[]
First , starting from the WSJ-trained recognizer , how much adaptation data -LRB- taken from the Phonebook training corpus -RRB- is necessary to achieve a reasonable recognition performance in spite of the high degree of mismatch ?
{ "relations": { "used for": [ { "head": { "text": "WSJ-trained recognizer", "start": 26, "end": 48 }, "tail": { "text": "recognition", "start": 166, "end": 177 } }, { "head": { "text": ...
[]
Second , is it possible to improve the recognition performance of a Phonebook-trained baseline acoustic model by using additional out-of-domain training data ?
{ "relations": { "used for": [ { "head": { "text": "Phonebook-trained baseline acoustic model", "start": 68, "end": 109 }, "tail": { "text": "recognition", "start": 39, "end": 50 } }, { "head": { ...
[]
This paper proposes an approach to full parsing suitable for Information Extraction from texts .
{ "relations": { "used for": [ { "head": { "text": "approach", "start": 23, "end": 31 }, "tail": { "text": "full parsing", "start": 35, "end": 47 } }, { "head": { "text": "full parsing",...
[]
It was implemented in the IE module of FACILE , a EU project for multilingual text classification and IE .
{ "relations": { "used for": [ { "head": { "text": "It", "start": 0, "end": 2 }, "tail": { "text": "IE module", "start": 26, "end": 35 } } ], "part of": [ { "head": { "text": "IE...
[]
It then presents an implemented graphic interpretation system that takes into account a variety of communicative signals , and an evaluation study showing that evidence obtained from shallow processing of the graphic 's caption has a significant impact on the system 's success .
{ "relations": { "used for": [ { "head": { "text": "communicative signals", "start": 99, "end": 120 }, "tail": { "text": "graphic interpretation system", "start": 32, "end": 61 } }, { "head": { ...
[]
Graphical models such as Bayesian Networks -LRB- BNs -RRB- are being increasingly applied to various computer vision problems .
{ "relations": { "used for": [ { "head": { "text": "Graphical models", "start": 0, "end": 16 }, "tail": { "text": "computer vision problems", "start": 101, "end": 125 } } ], "hyponym of": [ { ...
[]
One bottleneck in using BN is that learning the BN model parameters often requires a large amount of reliable and representative training data , which proves to be difficult to acquire for many computer vision tasks .
{ "relations": { "used for": [ { "head": { "text": "representative training data", "start": 114, "end": 142 }, "tail": { "text": "BN model parameters", "start": 48, "end": 67 } }, { "head": { ...
[]
On the other hand , there is often available qualitative prior knowledge about the model .
{ "relations": { "feature of": [ { "head": { "text": "qualitative prior knowledge", "start": 45, "end": 72 }, "tail": { "text": "model", "start": 83, "end": 88 } } ] } }
[]
Such knowledge comes either from domain experts based on their experience or from various physical or geometric constraints that govern the objects we try to model .
{ "relations": { "used for": [ { "head": { "text": "domain experts", "start": 33, "end": 47 }, "tail": { "text": "knowledge", "start": 5, "end": 14 } }, { "head": { "text": "physical or ...
[]
Unlike the quantitative prior , the qualitative prior is often ignored due to the difficulty of incorporating them into the model learning process .
{ "relations": { "compare": [ { "head": { "text": "quantitative prior", "start": 11, "end": 29 }, "tail": { "text": "qualitative prior", "start": 36, "end": 53 } } ], "part of": [ { "head"...
[]