Báo cáo khoa học: "A Stacked Sub-Word Model for Joint Chinese Word Segmentation and Part-of-Speech Tagging" potx

Báo cáo khoa học: "A Cascaded Linear Model for Joint Chinese Word Segmentation and Part-of-Speech Tagging" pdf

... seg-mentation only and joint segmentation and part-of-speech tagging. On the Penn Chinese Treebank 5.0, we obtain an error reduction of18.5% on segmentation and 12% on joint seg-mentation and part-of-speech ... 897–904,Columbus, Ohio, USA, June 2008.c2008 Association for Computational LinguisticsA Cascaded Linear Model for Joint Chinese Word Segmentation and Part-of-Speech TaggingWenbin Jiang†Liang Huang‡Qun ... tagging (Collins, 2002), Chinese word seg-mentation (Ng and Low, 2004; Zhang and Clark,2007) and so on. We trained a character-based per-ceptron for Chinese Joint S&T, and found that theperceptron...

Báo cáo khoa học: "An Error-Driven Word-Character Hybrid Model for Joint Chinese Word Segmentation and POS Tagging" docx

... discriminative word- character hybrid model for joint Chi-nese word segmentation and POS tagging.Our word- character hybrid model offershigh performance since it can handle bothknown and unknown words. ... ACL and the 4th IJCNLP of the AFNLP, pages 513–521,Suntec, Singapore, 2-7 August 2009.c2009 ACL and AFNLPAn Error-Driven Word- Character Hybrid Model for Joint Chinese Word Segmentation and ... litera-ture.1 IntroductionIn Chinese, word segmentation and part-of-speech (POS) tagging are indispensable steps for higher-level NLP tasks. Word segmentation and POS tag-ging results are...

Báo cáo khoa học: "A Stacked Sub-Word Model for Joint Chinese Word Segmentation and Part-of-Speech Tagging" potx

... Association for Computational LinguisticsA Stacked Sub -Word Model for Joint Chinese Word Segmentation and Part-of-Speech TaggingWeiwei SunDepartment of Computational Linguistics, Saarland UniversityGerman ... inter-mediate sub -word structure for joint segmentation and tagging. Since the sub-words are large enoughin practice, the decoding for POS tagging over sub-words is efﬁcient. Finally, the Chinese ... In this work, stacked learning isused to acquire extended training data for sub -word tagging.3 Method3.1 ArchitectureIn our stacked sub -word model, joint word segmen-tation and POS tagging...

Tài liệu Báo cáo khoa học: "A Hybrid Hierarchical Model for Multi-Document Summarization" ppt

... paper, we formulate ex-tractive summarization as a two step learn-ing problem building a generative model for pattern discovery and a regression model for inference. We calculate scores for sentences ... hierarchical model and re-gression model to score sentences in new docu-ments, eliminating the need for building a genera-tive model for new document clusters.3 Summary-Focused Hierarchical Model Our ... model. Then, using thesescores, we train a regression model basedon the lexical and structural characteris-tics of the sentences, and use the model toscore sentences of new documents to forma...

Tài liệu Báo cáo khoa học: "A Unified Graph Model for Sentence-based Opinion Retrieval" pdf

... The Lexicon of Chinese Positive Words, which consists of 5,054 positive words and the Lexicon of Chinese Negative Words, which consists of 3,493 negative words; (2) The opinion word lexicon ... notion of topic-sentiment word pair, which consists of a topic term and a sentiment word. A word pair maintains the asso-ciative information between the two words, and enables systems to draw ... consists of 2,812 positive words and 8,276 negative words; (3) Sentiment word lexicon and comment word lexicon from Hownet. It contains 1836 posi-tive sentiment words, 3,730 positive com-ments,...

Tài liệu Báo cáo khoa học: "A probabilistic generative model for an intermediate constituency-dependency representation" pptx

... re-ranking model performs rather well for a limited number of candidate structures, and out-performs Charniak’s model when k = 5. In thiscase we observe a small boost in performance for the detection ... structure. It models the eventof ﬁlling B with a content word (cw), given thecontent word of the governing block, the cate-gories (cats) and functional words (f w) of B, and further information ... consistently outper-forms the PCFG model on this metric, as for UAS, and BAS. Concerning the other metrics, as thenumber of k-best candidates increases, the PCFG model outperforms the TDS-reranker...

Tài liệu Báo cáo khoa học: "A Phonotactic Language Model for Spoken Language Identification" pptx

... another. Therefore, one can easily draw the analogy between an acoustic token in bag-of-sounds and a word in bag-of-words. Unlike words in a text document, the phonotactic information that ... n-character slice for text categorization by lan-guage (Cavnar and Trenkle, 1994) and Phone Rec-ognition followed by n-gram Language Modeling, or PRLM (Zissman, 1996) . Orthographic forms of language, ... information from acous-tic model and n-gram LM for language l. We have and {,AM}LLMlllλλλ= ( 1, , )llλ∈Λ =. A maxi-mum-likelihood classifier can be formulated as follows: ()(ˆargmax...

Tài liệu Báo cáo khoa học: "A Localized Prediction Model for Statistical Machine Translation" ppt

... length.Single source and target words are denoted by and respectively, where and .We will also use a special single -word block setwhich contains only blocks for which . For the experiments in ... phrase-based model for SMTsimilar to the models presented in (Koehn et al., 2003;Och et al., 1999; Tillmann and Xia, 2003). In our pa-per, phrase pairs are named blocks and our model is de-signed ... itself. and are single -word blocks, where source and target phrases are of length . is the num-ber of blocks for for which.5604 Online Training of Maximum-entropy Model The local model described...

Tài liệu Báo cáo khoa học: "A SPEECH-FIRST MODEL FOR REPAIR DETECTION AND CORRECTION" docx

... statistical analysis does not sup- 6We performed the same analysis for the last and first syllables in the reparandum and repair, respectively, and for normalized f0 and energy; results did not substantially ... Length of Reparandum Offset Word Frag- ments (N=288) bution of initial phonemes for all words in the corpus of 6,414 ATIS sentences, and for all fragments, single syllable fragments, and single ... Offset (N=288) a clear tendency for fragmentation at the reparandum offset to occur in content words rather than function words. 3In our pilot study of the SRI and TI utterances only, we found...

Báo cáo khoa học: "A Unified Statistical Model for the Identification of English BaseNP" pptx

... describe the two-passstatistical model, parameters training and Viterbialgorithm for the search of the best sequences ofPOS tagging and baseNP identification. Beforedescribing our algorithm, ... iiinnnP and (4) ),|(iiibmtwP . Thefirst and the third parameters are trigrams of T and B respectively. The second and the fourthare lexical generation probabilities. Probabilities(1) and (2) ... calculation formulas are similarwith equations (13) and (14) respectively.Before training trigram model (3), all possiblebaseNP rules should be extracted from thetraining corpus. For instance,...

Xem thêm