ICT: A Translation based Method for Cross-lingual Textual Entailment

First Joint Conference on Lexical and Computational Semantics (*SEM), pages 715–720, Montr´eal, Canada, June 7-8, 2012. c⃝2012 Association for Computational Linguistics ICT: A Translation based Method for Cross-lingual Textual Entailment Fandong Meng, Hao Xiong and Qun Liu Key Lab. of Intelligent Information Processing Institute of Computing Technology Chinese Academy of Sciences P.O. Box 2704, Beijing 100190, China {mengfandong,xionghao,liuqun}@ict.ac.cn Abstract In this paper, we present our system descrip- tion in task of Cross-lingual Textual Entail- ment. The goal of this task is to detect entailment relations between two sentences written in different languages. To accomplish this goal, we first translate sentences written in foreign languages into English. Then, we use EDITS1, an open source package, to rec- ognize entailment relations. Since EDITS only draws monodirectional relations while the task requires bidirectional prediction, thus we ex- change the hypothesis and test to detect en- tailment in another direction. Experimental results show that our method achieves promis- ing results but not perfect results compared to other participants. 1 Introduction In Cross-Lingual Textual Entailment task (CLTE) of 2012, the organizers hold a task for Cross- Lingual Textual Entailment. The Cross-Lingual Textual Entailment task addresses textual entail- ment (TE) recognition under a new dimension (cross-linguality), and within a new challenging application scenario (content synchronization) Readers can refer to M. Negri et al. 2012.s., for more detailed introduction. 1 Textual entailment, on the other hand, recog- nize, generate, or extract pairs of natural language expressions, and infer that if one element is true, whether the other element is also true. Several methods are proposed by previous researchers. There have been some workshops on textual en- tailment in recent years. The recognizing textual entailment challenges (Bar-Haim et al. 2006; Giampiccolo, Magnini, Dagan, & Dolan, 2007; Giampiccolo, Dang, Magnini, Dagan, & Dolan, 2008), currently in the 7th year, provide additional significant thrust. Consequently, there are a large number of published articles, proposed methods, and resources related to textual entailment. A spe- cial issue on textual entailment was also recently published, and its editorial provides a brief over- view of textual entailment methods (Dagan, Dolan, Magnini, & Roth, 2009). Textual entailment recognizers judge whether or not two given language expressions constitute a correct textual entailment pair. Different methods may operate at different levels of representation of the input expressions. For example, they may treat the input expressions simply as surface strings, they may operate on syntactic or semantic repre- sentations of the input expressions, or on represen- tations combining information from different 1http://edits.fbk.eu/ 715 levels. Logic-based approach is to map the lan- guage expressions to logical meaning representa- tions, and then rely on logical entailment checks, possibly by invoking theorem provers (Rinaldi et al., 2003; Bos & Markert, 2005; Tatu & Moldovan, 2005, 2007). An alternative to use logical meaning representations is to start by mapping each word of the input language expressions to a vector that shows how strongly the word co-occurs with par- ticular other words in corpora (Lin, 1998b), possi- bly also taking into account syntactic …

Content truncated to 500 words