Báo cáo khoa học: "Latent Variable Models for Semantic Orientations of Phrases" pdf

... type of phrase contains most of interesting properties of phrases, such as emergence or inversion of se-mantic orientations. In order to capture the properties of semantic orientations of phrases, ... latent variable models work well in the classiﬁ-cation of semantic orientations of phrasesand achieved nearly 82% classiﬁcation ac-curacy.1 IntroductionTechnology for affect analysis of texts ... obtainedbasic orientations of phrases for affect analysis of higher linguistic units such as sentences and doc-uments.The semantic orientation of a phrase is not amere sum of its component words. Semantic orientations...

Báo cáo khoa học: "Latent variable models of selectional preference" potx

... data.Latent variable models that use E M for infer-ence can be very sensitive to the number of latentvariables chosen. For example, the performance of ROOTH-EM worsens quickly if the number of clusters ... on semantic applications of topic models. Section 3 describes the models used in our exper-iments. Section 4 provides details of the experi-mental design. Section 5 presents results for our models ... coherent semantic clusters iscommon to many research areas. In the ﬁeld of document modelling, a class of methods knownas “topic models have become a de facto stan-dard for identifying semantic...

Báo cáo khoa học: "Latent Variable Models of Concept-Attribute Attachment" pptx

... Proceedings of the 47th Annual Meeting of the ACL and the 4th IJCNLP of the AFNLP, pages 620–628,Suntec, Singapore, 2-7 August 2009.c2009 ACL and AFNLPLatent Variable Models of Concept-Attribute ... The LDA-based models do not inherently make use of any ranking infor-mation contained in the original extractions. How-ever, such information can be incorporated in theform of a prior. The ... synset(s) for a set of goldattributes, taking into account polysemous usage. For each model, ranked lists of possible conceptassignments C(w) are generated for each attributew, using L(c|w) for...

Tài liệu Báo cáo khoa học: "Incorporating Context Information for the Extraction of Terms" pdf

... information gained from the textual context of the candidate term. 2 Context information for terms The idea of incorporating context information for term extraction came from that "Extended ... examined string, lal the length of a in terms of number of words, f(a) the frequency of a in the corpus, Ta the set of candidate terms that contain a, P(T~) the number of these candidate terms. ... ratio for the extraction of two-word English terms. (Frantzi and Ananiadou, 1996), besides the frequency of occurrence, also consider the frequency of the candidate string as a part of longer...

Tài liệu Báo cáo khoa học: "Incremental Parsing Models for Dialog Task Structure" doc

... These types of infor-mation provide rich clues for building dialog mod-els (Grosz and Sidner, 1986). Dialog models canbe built of ine (for dialog mining and summariza-tion), or online (for dialog ... historyNbestFstart!completeconnection!pathsshift!reduceFigure 5: Performance of parse-based methods for subtask tree buildingFigure 5 shows the performance of the differentmethods for determining the subtask tree of the di-alog. Wider ... However, performance of all themethods for dialog act prediction is too low to beuseful at the moment. In future work, we will ex-plore improved models for this task that make use of global information...

Báo cáo khoa học: "Employing Topic Models for Pattern-based Semantic Class Discovery" doc

... the above formula, nDCG(Gi,j) is the nDCG score of semantic class Gi,j; and ki denotes the number of semantic classes assigned to SSCi. For a list of queries, the MnDCG score of an algo-rithm ... perform preprocess-ing (refer to Section 3.4 for details) before build-ing topic models for CR(q), where some low-frequency items are removed. Determine the number of topics: Most topic models ... have semantically peer or sibling relationship. This paper studies the employment of topic models to automati-cally construct semantic classes, taking as the source data a collection of raw semantic...

Báo cáo khoa học: "Learning Expressive Models for Word Sense Disambiguation" pot

... for this monolingual task, the ILP approach signif-icantly outperforms the majority sense baseline and performs as well as the state -of- the-art system re-porting results for the same set of ... for the disambigua-tion of verbs. We plan to further evaluate our approach for other sets of words, including other parts -of- speech to allow further comparisons with other approach-es. For ... defined as the set of all the possible trans-lations of that verb in the corpus. 80% of the corpus was randomly selected and used for train-ing, with the remainder retained for testing. The...

Báo cáo khoa học: "Re-Ranking Models For Spoken Language Understanding Marco Dinarelli University of Trento Italy" potx

... Proceedings of the 12th Conference of the European Chapter of the ACL, pages 202–210,Athens, Greece, 30 March – 3 April 2009.c2009 Association for Computational LinguisticsRe-Ranking Models For ... representationbased on semantic constituents. These are theunits for meaning representation and are often re-ferred to as concepts. Concepts are instantiated bysequences of words, therefore a Spoken ... SRILM toolsand then converted to an FST. This allows the use of a wide set of stochastic language models (bothback-off and interpolated models with several dis-counting techniques like Good-Turing,...

Báo cáo khoa học: "Rule Markov Models for Fast Tree-to-String Translation" pot

... can be formed out of smaller rules (includ-ing larger phrase pairs that can be formerd out of smaller phrase pairs), as opposed to minimal rules,which are rules that cannot be formed out of otherrules. ... results of our paper. Weused grammars of minimal rules and composed rules of maximum height 3 as our baselines. For decod-ing, we used a beam size of 50. Using the bestbigram rule Markov models ... our highest score of 28.0.14.3 AnalysisTables 2 and 3 show how the various types of ruleMarkov models compare, for bigrams and trigrams,1 For this experiment, a beam size of 100 was used.862parameters...

Báo cáo khoa học: "Flow Network Models for Word Alignment and Terminology Extraction from Bilingual Corpora" docx

... model of the possible alignments in the form of flow net- works. This representation allows to define dif- ferent kinds of alignments and to find the most probable or an approximation of this ... 1993), for example, allow to take this fact into account. We can envisage more complex models, in- cluding distortion parameters, multiword no- tions, or information on part -of- speech, infor- ... associations on a set of aligned sentences comprising the one used for the above experiment. An evaluation on single words, showed a pre- cision of 9870 for the first hundred and 97% for the first...

Xem thêm