Standard Test Collection for English-Persian Cross-Lingual Word Sense Disambiguation

Standard Test Collection for English-Persian Cross-Lingual Word Sense Disambiguation Navid Rekabsaz, Serwah Sabetghadam, Mihai Lupu, Linda Andersson, Allan Hanbury Vienna University of Technology Favoritenstrasse 9 1040 Vienna Austria surname@ifs.tuwien.ac.at Abstract In this paper, we address the shortage of evaluation benchmarks on Persian (Farsi) language by creating and making available a new benchmark for English to Persian Cross Lingual Word Sense Disambiguation (CL-WSD). In creating the benchmark, we follow the format of the SemEval 2013 CL-WSD task, such that the introduced tools of the task can also be applied on the benchmark. In fact, the new benchmark extends the SemEval-2013 CL-WSD task to Persian language. Keywords: Persian, Farsi, Cross Lingual Word Sense Disambiguation, test collection 1. Introduction Word Sense Disambiguation (WSD)—the task of automat- ically selecting the most related sense for a word occurring in a context—is considered as a main step in the course of approaching language understanding beyond the surface of the words. WSD has been intensively studied in Natu- ral Language Processing (NLP) (Navigli, 2012), Machine Translation (Chan et al., 2007; Costa-Juss`a and Farr´us, 2014), and Information Retrieval (Zhong and Ng, 2012). As a paradigm, Cross-lingual Word Sense Disambiguation (CL-WSD) focuses on lexical substitution in a target lan- guage such that it targets disambiguation of one word in a source language while translating to a target language. SemEval-2010 (Lefever and Hoste, 2010) and SemEval- 2013 (Lefever and Hoste, 2013) provide an evaluation plat- form for term disambiguation from English to Dutch, Ger- man, Italian, Spanish, and French. In recent years, several tools and libraries for Persian language have been developed. For example, Seraji et al. (2012) provides a set of tools for preprocessing (Pre- Per), sentence segmentation and tokenization (SeTPer), and POS tagging (TagPer). More recently, Samvelian et al. (2014) introduces PersPred, focusing on processing of compounding verbs. Finally, Feely et al. (2014) provides a front-end and new tools for language processing. While these tools mostly target the complexities of the language, there is no standard evaluation framework for Persian lan- guage in the WSD and CL-WSD domain. In this paper, we address this shortage by creating and mak- ing available a new benchmark for English to Persian CL- WSD. In creating the benchmark, we follow the format of the SemEval 2013 CL-WSD task, such that the introduced tools of the task can also be applied on the benchmark. In fact, the new benchmark extends the SemEval-2013 CL- WSD task to Persian language. In the following, we review the related work and resources in Persian language, followed by the description of the col- lection in Section 3.. 2. Resources in Persian Language Persian is a member of the Indo-European language fam- ily, and uses Arabic letters for writing. Seraji et al. (2012) provide a comprehensive overview on the main characteris- tics of the language. For instance, the diacritic signs are not written—it is expected that the reader can read the text in the absence of the short vowels. This characteristic causes a special kind of word ambiguity in …

Content truncated to 500 words