Báo cáo khoa học: "Maximum Expected BLEU Training of Phrase and Lexicon Translation Models" pptx

... phrase- based system usually include LM, reordering model, word and phrase counts, and phrase and lexicon translation models. Given the focus of this paper, we review only the phrase and lexicon ... with the baseline, training phrase or lexicon models alone gives a gain of 0.7 and 0.5 BLEU points, respectively, on the test set. For a full training of both phrase and lexicon models, we ... Republic of Korea, 8-14 July 2012.c2012 Association for Computational LinguisticsMaximum Expected BLEU Training of Phrase and Lexicon Translation Models Xiaodong He Li Deng Microsoft Research...

Tài liệu Báo cáo khoa học: "Online Large-Margin Training of Dependency Parsers" docx

... parent of xj. T = {(xt, yt)}Tt=1denotes the training data.We follow the edge based factorization method of Eisner (1996) and deﬁne the score of a dependencytree as the sum of the score of ... between a child and its parent. These features take the form of a POStrigram: the POS of the parent, of the child, and of a word in between, for all words linearly betweenthe parent and the child. ... al-gorithm (Censor and Zenios, 1997). Crammer and Singer (2003) and Crammer et al. (2003) providean analysis of both the online generalization error and convergence properties of MIRA. In equation(1),...

Báo cáo khoa học: "Simple semi-supervised training of part-of-speech taggers" pptx

... tri -training and self -training is near-signiﬁcant (p <0.0150). It seems that tri -training with disagreement is a competitive technique interms of accuracy. The main advantage of tri- training ... until none of cichanges16: apply majority vote over ciFigure 1: Tri -training (Li and Zhou, 2005).3.1 Tri -training with disagreementWe introduce a possible improvement of the tri- training ... obtained by training a SVM on top of the predictions of SVMTool and the word clusters of Unsupos. SVMs performed better than deci-sion trees, bagging and random forests on our de-velopment...

Báo cáo khoa học: "Stochastic Gradient Descent Training for L1-regularized Log-linear Models with Cumulative Penalty" potx

... for sec-tions 0-18 of the WSJ corpus).For the features, we used unigrams and bigrams of neighboring words, preﬁxes and sufﬁxes of the current word, and some characteristics of theword. We also ... fromsubsets of the training data and updatesthe parameters in an online fashion. Thislearning framework is attractive becauseit often requires much less training timein practice than batch training ... uni-grams and bigrams of neighboring words, and un-igrams, bigrams and trigrams of neighboring POStags.To avoid giving any advantage to our SGD al-gorithms over the OWL-QN algorithm in terms of the...

Báo cáo khoa học: "Maximum Entropy Based Restoration of Arabic Diacritics" ppt

... integrate and make eﬀective use of diverse types of information; the model we propose inte-grates a wide array of lexical, segment-based and part -of- speech tag features. Thecombination of these ... based on the context and their knowledge of thegrammar and the lexicon of Arabic. However, atext without diacritics becomes a source of confu-sion for beginning readers and people with learningdisabilities. ... shadda.state -of- the-art system evaluated on Arabic Tree-bank Part 2: in (Nelken and Shieber, 2005) a DER of 12.79% and a WER of 23.61% are reported.The system they described in (Nelken and Shieber,2005)...

Báo cáo khoa học: "Maximum Entropy Model Learning of the Translation Rules" pot

... properties of a language model. The expected value of f with respected to iS(x, y) is defined such as: p(f) = p(x,y)f(x,y) (z) x,y Thus training data are summarized as the expected value of feature ... part -of- speech for English and 49 part -of- speech for Japanese). We tried to learn the translation rules from English to Japanese. We had two experiments: one of model 1 as the set of fea- ... value of Ai according to: Ai ~- A~ + AAi 4 Maximum Entropy Model Learning of the Translation Rules The art of modeling with the maximum entropy method is to define an informative set of...

Tài liệu Báo cáo khoa học: "Cross-Domain Co-Extraction of Sentiment and Topic Lexicons" pdf

... set of nodesrepresents topic words, including new topic candi-dates and words in the lexicon C, and the other set of nodes represents sentiment words, including newsentiment candidates and ... topic wordswith the ﬁnal scores, and add them to lexicons B and C.UpdateS1(wi) and S3(wj) accordingly.8: end for9: return Expanded lexicons B and C.5.2 Graph ConstructionBased ... foreach edge of the graph.6: Reﬁne the scoresS1 and S3 of the k2sentiment and topic word candidates using Eqs. (5) and (6) iteratively.7: Select k1new sentiment words and k1new...

Tài liệu Báo cáo khoa học: Mapping the functional domains of human transcobalamin using monoclonal antibodies pptx

... involved in the binding of Cbl [13,14]. Assembly of two domains achieved thecomposite structure of the ligand-binding site and built the compatible interface between IF and itsreceptor [14]. ... 3C4 and 2-2 ⁄ 3C4, where binding of the second mAb of thepair was inhibited by < 20%.Interaction of mAbs with the C-terminal domain of human TC was determined using peptides TCp11 and TCp12, ... fragments and theirreactivity with mAbs, epitope clusters 2, 4 and 5Fig. 3. Binding of the antibodies to CNBrcleavage peptides of TCy. CNBr peptides of TCywere fractionated by HPLC and subjec-ted...

Tài liệu Báo cáo khoa học: Unraveling the catalytic mechanism of lactoperoxidase and myeloperoxidase A reﬂection on some controversial features Elena Ghibaudi and Enzo Laurenti docx

... features of the LPO and MPO catalytic cycle, such as theexistence of Compound I and Compound II isomers and the identiﬁcation of their spectroscopic properties. Afteraddressing each of these ... each of these questions, we will proposea new hypothesis that describes an integrated vision of thecatalytic mechanism of MPO and LPO.A brief survey of the absorption features of MPO, LPO and ... them [19]. On the other hand, this factdoes not represent proof of the presence of one or twooxidizing equivalents on the decay product of Cpd I.The EPR spectrum of a spin-trapped radical...

Tài liệu Báo cáo khoa học: "Bilingual Terminology Acquisition from Comparable Corpora and Phrasal Translation to Cross-Language Information Retrieval" pptx

... associated to each translation alternative. Selection of a phrase willmodify the ranked list of phrases and will providean access to documents related to the phrase. 3 Experiments and Evaluations ... to re-score the set of translation candidatesrelated to the source terms.Sequences of all possible combinations are con-structed between elements of sets of highly ranked translation alternatives. ... values and rank counts are assigned in increasing order. A ﬁxednumber of top-ranked translation alternatives are se-lected and misleading candidates are discarded.2.3 Phrasal Translation Query translation...

Xem thêm