报告1: Modeling the Translation of Predicate-Argument Structure for SMT

报告人：Dr. Xiong Deyi, Institute for Infocomm Research, Singapore

时间：2011年12月30日 10:00-11:00

地点：理工楼321会议室

报告摘要: Predicate-argument structure contains rich semantic information of which statistical machine translation hasn"t taken full advantage. In this paper, we propose two discriminative, feature-based models to exploit predicate-argument structures for statistical machine translation: 1) a predicate translation model and 2) an argument reordering model. The predicate translation model explores lexical and semantic contexts surrounding a predicate verb to select desirable translations for the predicate. The argument reordering model automatically predicts the moving direction of an argument relative to its predicate after translation using semantic features. The two models are integrated into a state-of-the-art phrase-based machine translation system and evaluated on Chinese-to-English translation tasks with large-scale training data. Experimental results demonstrate that the two models significantly improve translation accuracy.

报告2: A Topic Similarity Model for Hierarchical Phrase-based Translation

报告人：Dr. Xiong Deyi, Institute for Infocomm Research, Singapore

时间：2011年12月31日 10:00-11:00

地点：理工楼321会议室

报告摘要: Previous work on topic model based approaches for statistical machine translation (SMT) explore topic information on the word level.However, SMT systems have been advanced from word-based translation to phrase/rule-based translation. We therefore propose a topic similarity model to exploit topic information on the synchronous rule level for hierarchical phrase-based translation. We associate each synchronous rule with a topic distribution, and select desirable rules according to the similarity of their topic distributions with given documents. Experimental results show that our model significantly improves the translation performance of hierarchical phrase-based system, and also achieves a better and faster performance than previous approaches that work on the word level.

报告人简历: Deyi received the B.Sc degree from China University of Geosciences (Wuhan, China) in 2002, the Ph.D. degree from the Institute of Computing Technology (Beijing, China) in 2007, both in computer science. He joined the Institute for Infocomm Research, Singapore as a research fellow in 2007. Currently, he is a scientist at the institute. His primary research interests are in the area of natural language processing, particularly statistical machine translation, language modeling and parsing.