Báo cáo khoa học: "Using Reinforcement Learning to Build a Better Model of Dialogue State" pdf

... complexity of tasksthat spoken dialogue systems are trying to handle, Reinforcement Learning (RL) hasbeen increasingly used as a way of au-tomatically learning the best policy for a system to make. ... known immediately, in fact,usually only the ﬁnal reward is known at the end of the dialogue. As long as we have a reward func-tion, Reinforcement Learning allows one to auto-matically compute ... user’s last dialogue act,which was found to produce a 52% increase in av-erage reward. Williams et al. (2003) used Super-vised Learning to select good state and action fea-tures as an initial...

Báo cáo khoa học: "Using Machine-Learning to Assign Function Labels to Parser Output for Spanish" ppt

... output by a prob-abilistic parser trained on Cast3LB. To our knowl-edge, this paper is the ﬁrst to describe applying a data-driven approach to function-tag assignment to a language other than English.Our ... complex labels asatomic categories the number of unique labels in-creases and parse quality can deteriorate due to sparse data problems. Secondly, this approach, byrelying on the parser to assign ... of target node can have negative impact, greater re-liance on context means greater susceptibility to parse errors. Another factor to consider is the factthat we trained and adjusted parameters...

Báo cáo khoa học: "Using Machine Learning to Explore Human Multimodal Clariﬁcation Strategies" ppt

... information to a classiﬁer, we compared the ac-curacy achieved in a 10-fold cross validation by a Na¨ıve Bayes classiﬁer (as a standard) on thesedata sets against the majority class baseline, ... driving). All fea-tures are automatically extracted from the XMLlog-ﬁles (and are available at runtime in ISU-based dialogue systems). From these features wewant to learn whether to generate a screen ... op-timisation as an integral part of the training pro-cess). All experiments are carried out using 10-fold cross-validation. We take an approach similar to (Daelemans et al., 2003) where parameters of the...

Báo cáo khoa học: "Using Machine Learning to Maintain Rule-based Named-Entity Recognition and Classification Systems" pdf

... due to the lack of experts. The exploitation of learning techniques to support this adaptation task hasattracted the attention of researchers in languageengineering.However, the adaptation of ... consists of a lexiconand a grammar. The lexicon is a set of NEs thatare known beforehand and have been classifiedinto semantic classes. The grammar is used to recognize and classify NEs that are ... importantadvantage of this approach is that no manualtagging of training data is needed, despite theuse of a supervised learning algorithm.Our method bears some similarities with sys-tems based...

Tài liệu Báo cáo khoa học: "Using Multiple Sources to Construct a Sentiment Sensitive Thesaurus for Cross-Domain Sentiment Classiﬁcation" doc

... cheaper to collect compared to labeled data and is often available in large quan-tities. The use of unlabeled data enables us to ac-curately estimate the distribution of words in sourceand target ... UKj .a. carroll@sussex.ac.ukAbstractWe describe a sentiment classiﬁcation methodthat is applicable when we do not have any la-beled data for a target domain but have somelabeled data for ... cross-domain sentiment classiﬁcation as the problem of learning a binary classiﬁer (i.e. positive or negativesentiment) given a small set of labeled data for thesource domain, and unlabeled data for...

Tài liệu Báo cáo khoa học: "USING BRACKETED PARSES TO EVALUATE A GRAMMAR CHECKING APPLICATION" ppt

... the hand parse), Recall (the percentage of phrases in the hand parse that are also in the ma- chine parse), and Precision (the percentage of phrases in the machine parse that are in the hand ... utilizing hand-bracketed corpora (such as the University of Pennsylvania Treebank) as a basis for evaluation metrics have been discussed in Black et al. (1991), Harrison et al. (1991), and Black ... hand parse). We have developed a methodology for using hand-bracketed parses to examine both the inter- nal and external performance of a grammar checker. The internal performance refers to...

Báo cáo khoa học: "Using Decision Trees to Construct a Practical Parser" pdf

... against training data, it not only adapts to any application domain but. also may avoid over- fitting to the data. In the late 80s and early 90s, the induction and parameter estimation of probabilis- ... Research Laboratories 2-2 Hikaridai, Seika-cho, Soraku-gun, Kyoto 619-02, Japan. tNTT Communication Science Laboratories 2-4 Hikaridai, Seika-cho, Soraku-gun, Kyoto 619-02, Japan. Abstract ... Trees to Construct a Practical Parser Masahiko Haruno* Satoshi Shirai t Yoshifumi Ooyama t mharuno ~hlp.atr.co.jp shirai,~cslab.kecl.ntt.co.jp oovama~cslal).kecl.nt t.co.j p *ATR Human Information...

Báo cáo khoa học: "Using Machine Learning Techniques to Build a Comma Checker for Basque" pdf

... for BasqueIñaki Alegria Bertol Arrieta Arantza Diaz de Ilarraza Eli Izagirre Montse MaritxalarComputer Engineering Faculty. University of the Basque Country.Manuel de Lardizabal Pasealekua, ... International Proofing Tools and Language Technologies. University of Patras (Greece).Aranzabe M., Arriola J.M., Díaz de Ilarraza A. 2004.Towards a Dependency Parser of Basque.Proceedings of ... Re-cent Advances in Dependency Grammar. Geneva (Switzerland).Bayraktar M., Say B., Akman V. 1998. An Analysis of English Punctuation: the special case of comma. International Journal of Corpus...

Báo cáo khoa học: "Using Machine Learning Techniques to Interpret WH-questions" potx

... Cover-age Asked in the above example); Additionalrefers to a level of detail characterized by a one-paragraph answer (this is the value of CoverageWould Give in the above example); and Extendedindicates ... traditional QAresearch in its use of statistical models to pre-dict variables that represent a user’s informationalgoals. The variables under consideration includethe type of the information ... effect of various train-ing and testing factors on predictive performance,and examines the relationships among the targetvariables.5.1 Training FactorsWe examine how the quality of the training...

Tài liệu Báo cáo khoa học: "Coreference for Learning to Extract Relations: Yes, Virginia, Coreference Matters" ppt

... words of detailed, exhaustive annotation in Automatic Content Extraction (ACE) evaluations1) many have explored bootstrapping relation extraction from a few (~20) seed instances of a relation. ... is in accordance with DoDI 5230.29, January 8, 2009. Abstract As an alternative to requiring substantial su-pervised relation training data, many have ex-plored bootstrapping relation ... +Coref has access to much more training data at the outset because it can find more matching seed instances,9 potentially leading to better and more stable training. 4 Evaluation Framework...

Xem thêm