| other,27-5-P03-1051,bq | corpus </term> , and re-estimate the <term> | model | parameters </term> with the expanded <term> | #4733 To improve the segmentation accuracy, we use an unsupervised algorithm for automatically acquiring new stems from a 155 million word unsegmented corpus, and re-estimate themodel parameters with the expanded vocabulary and training corpus. |
| model,5-3-C04-1103,bq | <term> joint source-channel transliteration | model | </term> , also called <term> n-gram transliteration | #5781 Under this framework, a joint source-channel transliteration model, also called n-gram transliteration model (ngram TM), is further proposed to model the transliteration process. |
| model,12-3-C04-1103,bq | also called <term> n-gram transliteration | model | ( ngram TM ) </term> , is further proposed | #5787 Under this framework, a joint source-channel transliteration model, also called n-gram transliteration model (ngram TM), is further proposed to model the transliteration process. |
| ngram TM ) </term> , is further proposed to | model | the <term> transliteration process </term> | #5797 Under this framework, a joint source-channel transliteration model, also called n-gram transliteration model (ngram TM), is further proposed to model the transliteration process. | |
| tech,2-4-C04-1112,bq | algorithm . Testing the <term> lemma-based | model | </term> on the <term> Dutch SENSEVAL-2 test | #6058 Testing the lemma-based model on the Dutch SENSEVAL-2 test data, we achieve a significant increase in accuracy over the wordform model. |
| tech,20-4-C04-1112,bq | <term> accuracy </term> over the <term> wordform | model | </term> . Also , the <term> WSD system based | #6076 Testing the lemma-based model on the Dutch SENSEVAL-2 test data, we achieve a significant increase in accuracy over the wordform model. |
| model,20-2-C04-1147,bq | <term> terms </term> , an <term> independence | model | </term> , and a <term> parametric affinity | #6342 The framework is composed of a novel algorithm to efficiently compute the co-occurrence distribution between pairs of terms, an independence model, and a parametric affinity model. |
| model,25-2-C04-1147,bq | model </term> , and a <term> parametric affinity | model | </term> . In comparison with previous <term> | #6348 The framework is composed of a novel algorithm to efficiently compute the co-occurrence distribution between pairs of terms, an independence model, and a parametric affinity model. |
| tech,19-4-N04-4028,bq | field ( CRF ) </term> , a <term> probabilistic | model | </term> which has performed well on <term> | #6831 The information extraction system we evaluate is based on a linear-chain conditional random field (CRF), a probabilistic model which has performed well on information extraction tasks because of its ability to capture arbitrary, overlapping features of the input in a Markov model. |
| model,44-4-N04-4028,bq | features </term> of the input in a <term> Markov | model | </term> . We implement several techniques | #6856 The information extraction system we evaluate is based on a linear-chain conditional random field (CRF), a probabilistic model which has performed well on information extraction tasks because of its ability to capture arbitrary, overlapping features of the input in a Markov model. |
| model,1-3-H05-1012,bq | performance </term> . The <term> probabilistic | model | </term> used in the <term> alignment </term> | #7296 The probabilistic model used in the alignment directly models the link decisions. |
| tech,1-3-H05-1095,bq | proposed . A <term> statistical translation | model | </term> is also presented that deals such | #7373 A statistical translation model is also presented that deals such phrases, as well as a training method based on the maximization of translation accuracy, as measured with the NIST evaluation metric. |
| model,18-1-I05-2021,bq | representative <term> Chinese-to-English SMT | model | </term> directly on <term> word sense disambiguation | #7806 We present the first known empirical test of an increasingly common speculative claim, by evaluating a representative Chinese-to-English SMT model directly on word sense disambiguation performance, using standard WSD evaluation methodology and datasets from the Senseval-3 Chinese lexical sample task. |
| tech,2-3-J05-1003,bq | these <term> parses </term> . A second <term> | model | </term> then attempts to improve upon this | #8691 A secondmodel then attempts to improve upon this initial ranking, using additional features of the tree as evidence. |
| tech,40-4-J05-1003,bq | <term> derivation </term> or a <term> generative | model | </term> which takes these <term> features </term> | #8751 The strength of our approach is that it allows a tree to be represented as an arbitrary set of features, without concerns about how these features interact or overlap and without the need to define a derivation or a generative model which takes these features into account. |
| model,7-7-J05-1003,bq | log-likelihood </term> under a <term> baseline | model | </term> ( that of <term> Collins [ 1999 ] </term> | #8807 The method combined the log-likelihood under a baseline model (that of Collins [1999]) with evidence from an additional 500,000 features over parse trees that were not included in the original model. |
| model,34-7-J05-1003,bq | were not included in the original <term> | model | </term> . The new <term> model </term> achieved | #8833 The method combined the log-likelihood under a baseline model (that of Collins [1999]) with evidence from an additional 500,000 features over parse trees that were not included in the originalmodel. |
| tech,2-8-J05-1003,bq | original <term> model </term> . The new <term> | model | </term> achieved 89.75 % <term> F-measure </term> | #8837 The newmodel achieved 89.75% F-measure, a 13% relative decrease in F-measure error over the baseline model’s score of 88.2%. |
| measure(ment),18-8-J05-1003,bq | F-measure </term> error over the <term> baseline | model | ’s score </term> of 88.2 % . The article | #8854 The new model achieved 89.75% F-measure, a 13% relative decrease in F-measure error over the baseline model’s score of 88.2%. |
| model,25-3-P05-1034,bq | </term> , and train a <term> tree-based ordering | model | </term> . We describe an efficient <term> decoder | #9271 We align a parallel corpus, project the source dependency parse onto the target sentence, extract dependency treelet translation pairs, and train a tree-based ordering model. |
