FCC: Modeling Probabilities with GIZA++ for Task 2 and 3 of SemEval-2

Proceedings of the 5th International Workshop on Semantic Evaluation, ACL 2010, pages 112–116, Uppsala, Sweden, 15-16 July 2010. c⃝2010 Association for Computational Linguistics FCC: Modeling Probabilities with GIZA++ for Task #2 and #3 of SemEval-2 Darnes Vilari˜no, Carlos Balderas, David Pinto, Miguel Rodr´ıguez, Saul Le´on Faculty of Computer Science, BUAP Puebla, Mexico {darnes,mrodriguez,dpinto}@cs.buap.mx Abstract In this paper we present a na¨ıve approach to tackle the problem of cross-lingual WSD and cross-lingual lexical substitu- tion which correspond to the Task #2 and #3 of the SemEval-2 competition. We used a bilingual statistical dictionary, which is calculated with Giza++ by using the EU- ROPARL parallel corpus, in order to cal- culate the probability of a source word to be translated to a target word (which is as- sumed to be the correct sense of the source word but in a different language). Two ver- sions of the probabilistic model are tested: unweighted and weighted. The obtained values show that the unweighted version performs better thant the weighted one. 1 Introduction Word Sense Disambiguation (WSD) is con- sidered one of the most important prob- lems in Natural Language Processing (Agirre and Edmonds, 2006). It is claimed that WSD is essential for those applications that require of language comprehension modules such as search engines, machine translation systems, automatic answer machines, second life agents, etc. Moreover, with the huge amounts of information in Internet and the fact that this information is continuosly growing in different languages, we are encourage to deal with cross- lingual scenarios where WSD systems are also needed. Despite the WSD task has been studied for a long time, the expected feeling is that WSD should be integrated into real applications such as mono and multi-lingual search engines, machine translation systems, automatic answer machines, etc (Agirre and Edmonds, 2006). Different stud- ies on this issue have demonstrated that those applications benefit from WSD, such as in the case of machine translation (Chan et al., 2007; Carpuat and Wu., 2007). On the other hand, Lexical Substitution (LS) refers to the process of finding a substitute word for a source word in a given sentence. The LS task needs to be approached by firstly disambiguating the source word, therefore, these two tasks (WSD and LS) are somehow related. Since we are describing the modules of our system, we did not provide information of the datasets used. For details about the corpora, see the task description paper for both tasks (#2 and #3) in this volume (Mihalcea et al., 2010; Lefever and Hoste, 2010). Description about the other teams are also described in the same papers. 2 A Na¨ıve Approach to WSD and LS In this section it is presented an overview of the presented system, but also we further discuss the particularities of the general approach for each task evaluated. We will start this section by explaining the manner we deal with the Cross- Lingual Word Sense Disambiguation (C-WSD) problem. 2.1 Cross-Lingual Word Sense Disambiguation We have approached the cross-lingual word sense disambiguation task …

Content truncated to 500 words