Báo cáo khoa học: "Learning to Transform and Select Elementary Trees for Improved Syntax-based Machine Translations" ppt

... Association for Computational Linguistics, pages 846–855, Portland, Oregon, June 19-24, 2011. c 2011 Association for Computational Linguistics Learning to Transform and Select Elementary Trees for Improved Syntax-based ... Discussions and Conclusions We proposed a framework to learn models to predict how to transform an elementary tree into its simplified for...

Ngày tải lên: 17/03/2014, 00:20

... engine, and automatically annotating the snippets with tags and features for training a conditional random field model. At run- time, the model is used to extracting translation candidates for ... describe and evaluate a multi-stage machine translation method for back transliterating English names into Japanese, while Bian and Chen (2000) describe cross-language informa...

Ngày tải lên: 19/02/2014, 19:20

... decoder has to store in each hypothesis the source and target root labels of the partial derivation, but these la- bels are used for calculating feature vectors only and not for checking well-formedness ... bitexts, and one on two billion words of English. These were smoothed using modified Kneser-Ney (Chen and Goodman, 1998) and stored using randomized data structures similar...

Ngày tải lên: 07/03/2014, 22:20

... instructor answer and the student answer. In all, the application of the three transforma- tions leads to eight different transform combina- tions, and therefore eight different alignment scores. For ... Support Vec- tor Machines (SVM) to produce a combined real- number grade. Finally, we build an Isotonic Regres- sion (IR) model to transform our output scores onto the origina...

Ngày tải lên: 07/03/2014, 22:20

... per word form token. An excerpt of the annotated data with all generated symbolic and numeric 1 features is presented in Table 1. Word forms (Wrd) – The word form tokens form the central unit to which ... terms of feature selection and classifier parameters gives rise to significant improvements in performance. In this paper, therefore, we optimize both CART and MBL individually and...

Ngày tải lên: 17/03/2014, 06:20

... weight vector w, takes in a foreign sen- tence f and returns a translated hypothesis h. The argmax operates in vector space and our goal is to find w leading to hypotheses on the Pareto Frontier. In ... features. Machine Trans- lation, 23(2-3). Kishore Papineni, Salim Roukos, Todd Ward, and Wei- Jing Zhu. 2002. BLEU: A method for automatic eval- uation of machine translation. I...

Ngày tải lên: 19/02/2014, 19:20

... to the first author and award IIS-0811974 to the second author and by the Air Force Research Laboratory (AFRL), under prime contract no. FA8750-09-C-0181. Thanks to Michael Levit and Deb Roy for ... respect to two metrics: the order in which we visit landmarks and the side we pass them on. To determine the order P e visits landmarks we compute the minimum distance from P e...

Ngày tải lên: 20/02/2014, 04:20

... work. For example, the wh- determiners (WDT) split into one class for that and an- other for which, while the wh-adverbs align by refer- ence type: event-based how and why vs. entity-based when and ... independent. In fact, extra subsymbols may need to be added to several nonterminals before they can cooperate to pass information along the parse tree. Therefore, we go in the...

Ngày tải lên: 20/02/2014, 12:20

... Walker, 2000; Scheffler and Young, 2002; Peek and Chickering, 2005; Frampton and Lemon, 2006), for selecting initiative and con- firmation strategies (Singh et al., 2002); for detect- ing speech ... effective dialogue strategies for information seeking spo- ken dialogue systems, and in particular the use of reinforcement learning methods to learn policies for action selectio...

Ngày tải lên: 20/02/2014, 12:20

... misleading and at worst may lead to erroneous analysis of the text. As tables contain important data and information, it is critical for an information extraction system to be able to extract ... conducted ten random trials and then averaged the accuracy over the ten trials. In each random trial, 20% of the texts are randomly chosen to serve as the texts for testing,...

Ngày tải lên: 20/02/2014, 19:20