Báo cáo khoa học: "Statistical Models for Unsupervised Prepositional Phrase Attachment" pdf

... Statistical Models for Unsupervised Prepositional Phrase Attachment Adwait Ratnaparkhi Dept. of Computer and Information Science University of Pennsylvania ... Abstract We present several unsupervised statistical models for the prepositional phrase attachment task that approach the accuracy of the best su- pervised methods for this task. Our unsuper- ... base forms, as opposed to attachment information. It is therefore less resource-intensive and more portable than pre- vious corpus-based algorithm proposed for this task. We present results for...

Báo cáo khoa học: "Minimized Models for Unsupervised Part-of-Speech Tagging" pot

... AFNLPMinimized Models for Unsupervised Part-of-Speech TaggingSujith Ravi and Kevin KnightUniversity of Southern CaliforniaInformation Sciences InstituteMarina del Rey, California 90292{sravi,knight}@isi.eduAbstractWe ... dictionary varies for different models from Figure 5. In the case of theobserved grammar (Figure 6), precision increases5 For all experiments, EM training is allowed to run for 40 iterations ... unlabeleddata used for training the models. Their models are trained on the entire Penn Treebank data (in-stead of using only the 24,115-token test data),and so are the tagging models used by Goldberget...

Báo cáo khoa học: "A Connectionist Approach to Prepositional Phrase Attachment for Real World Texts" docx

... syntactic information does not suffice to make an assignment decision. Prepositional phrase (PP) attachment is, perhaps, the canonical case of structural ambiguity. What kind of information ... sense. As a result, before deciding attachment, it is first nec- essary to determine the correct sense for each word. 3 None of the preceding methods used classes for verbs. 4 For reasons of complexity, ... Bounds for Superposition of a Sigmoidal Function. IEEE Transac- tions on Information Theory, 39:930-945. Brill, E. & Resnik, P. (1994). A Rule-Based Approach to Prepositional Phrase Attachment...

Báo cáo khoa học: "Distortion Models For Statistical Machine Translation" doc

... correctly, and was added to the phrase dic-tionary. However, if the phrase Almmlkp AlmtHdp isnot in the phrase dictionary, it will not be translatedcorrectly by a monotone phrase decoder even if the ... to a phrase- based model.(Och et al., 2004; Tillman, 2004) proposeorientation-based distortion models lexicalized on the phrase level. There are two important distinctions be-tween their models ... not phrases. Therefore, we need to general-ize the distortion cost in order to use it in a phrase- based decoder. This generalization is deﬁned in termsof the internal word alignment within phrases...

Báo cáo khoa học: "Fertility Models for Statistical Natural Language Understanding" pdf

... in- crease in performance of about 2-3% for most models. For General-LM, results increased by 8-10%. The Poisson and general fertility models show a 2- 5% gain in performance over the basic ... into phrases. Each clump is then generated from a sin- gle formal language word using a translation model. The notion of what constitutes a natural clumping depends on the formal language. For ... translation probabilities are shown for the unigram and headword models in table 2. The formal language words have captured reason- able English words for their most likely translation...

Báo cáo khoa học: "Annealing Techniques for Unsupervised Statistical Language Learning" ppt

... how DA can be used for parameter estimation for models of language structure that use dynamicprogramming to compute posteriors over hiddenstructure, such as hidden Markov models (HMMs)and stochastic ... algorithm(like the forward-backward algorithm). The M step for these models normalizes the posterior expectedcounts from the E step to get probabilities.43With an M step modiﬁed for their objective ... θβj. For a given x, the forwardpass of the DP computes (in a dense representation)Pr(y | x,θ) for all y. Each Pr(y | x,θ) is a productof some of the θj(each θjis multiplied in once for each...

Báo cáo khoa học: "Generative Models for Statistical Parsing with Combinatory Categorial Grammar" pptx

... ap-propriate for a formalism like CCG with a large setof lexical categories than one generic token for allunknown words.The performance of the baseline model is shownin the top row of table 3. For ... for passive, bare inﬁnitives and ing-forms.There is a separate level for nouns and noun phrases,but, like the nonterminal NP in the Penn Treebank,noun phrases do not carry any number agreement.The ... probability models for a wide-coverage Combinatory Categorial Gram-mar (CCG) parser. These models aretrained and tested on a corpus obtained bytranslating the Penn Treebank trees intoCCG normal-form...

Tài liệu Báo cáo khoa học: "Topic Models for Dynamic Translation Model Adaptation" pptx

... Meeting of the Association for Computational Linguistics, pages 115–119,Jeju, Republic of Korea, 8-14 July 2012.c2012 Association for Computational LinguisticsTopic Models for Dynamic Translation ... the most useful notion of domain for bettertranslations.We take a ﬁner-grained, ﬂexible, unsupervised ap-proach for lexical weighting by domain. We induce unsupervised domains from large corpora, ... unseen for a given domain s, we are alreadyperforming an implicit form of smoothing (whencomputing the expected counts), since each docu-ment has a distribution over all topics, and thereforewe...

Báo cáo khoa học: "Ensemble Methods for Unsupervised WSD" doc

... Association for Computational LinguisticsEnsemble Methods for Unsupervised WSDSamuel BrodySchool of InformaticsUniversity of Edinburghs.brody@sms.ed.ac.ukRoberto NavigliDipartimento di InformaticaUniversit`a ... about27 person-years of human annotation effort.This paper focuses on unsupervised methodswhich we argue are useful for broad coveragesense disambiguation. Unsupervised WSD algo-rithms fall into ... WSDmethod can act as an arbiter for adjudicating dis-agreements among component systems. It makessense for the adjudicator to have reasonableperformance on its own. We therefore selectedMethod AccpsAccwsd/psSimilarity...

Báo cáo khoa học: "Structured Models for Fine-to-Coarse Sentiment Analysis" pdf

... whichhelped improve performance.2.2 Beyond Two-Level Models To this point, we have focused solely on a model for two-level ﬁne-to-coarse sentiment analysis not only for simplicity, but because ... different applications have different needs. For example, a summarization system for productreviews might require polarity classiﬁcation at thesentence or phrase level; a question answering sys-tem ... systems.1.1 Related WorkThe models in this work fall into the broad class ofglobal structured models, which are typically trainedwith structured learning algorithms. Hidden Markov models (Rabiner, 1989)...

Xem thêm