Pattern-based Machine Translation for English-Thai

PATTERN-BASED MACHINE TRANSLATION FOR ENGLISH-THAI Kaewchai Chancharoen, Nisanad Tannin and Booncharoen Sirinaovakul Artificial Intelligence Center, King Mongkut's University of Technology Thonburi 91 Pracha-uthit Thungkhru District, Bangkok 10140, Thailand Tel & Fax (662) 4284204, E-mail : ikaeroen@cc.kmutt.ac.th ABSTRACT This paper proposes a new model of machine translation system in which rule-based and example-based approaches are applied for English-to-Thai sentence translation. The proposed method has 4 steps :1) analyze an English sentence into a string of grammatical nodes, based on Phrase Structure Grammar, 2) map the input pattern with a table of English-Thai sentence patterns, 3) look up the bilingual dictionary for the equivalent Thai words, reorder and then generate output sentences and 4) rank the possible combinations and eliminate the ambiguous output sentences by using a statistical method. The translated sentences will then be stored in a bilingual corpus to serve as a guide or template for imitating the translation, i.e., the example-based approach. The future work will focus on disambiguation by using semantic features to make a more precise translation. 1. INTRODUCTION Various efforts have been made in developing machine translation (MT) systems for practical use. Historically, there are many approaches on MT research: transfer-based, interlingua-based, and etc. Among these approaches, the most distinctive are rule-based and corpus-based methods. Research on the corpus-based approach has emphasized on the importance of text corpora used as a source for linguistic and knowledge databases. There have been two major approaches among the corpus-based MT known as statistics-based and example-based. It might be said that all approaches have their own pros and cons. Therefore some MT researchers have selected and combined them together for creating a new effective model. We also combine two potential approaches to produce our own strategy; namely, rule-based and example-based. 1.1 Rule-Based and Example-Based Approaches The rule-based translation mostly consists of (1) a process of analyzing input sentences of a source language morphologically, syntactically and/or semantically and (2) a process of generating output sentences of a target language based on an internal structure or interlingua. Each process is controlled by the dictionary and the rules. Meanwhile, the basic idea of example-based method is to translate a sentence by using translation examples of similar sentences [1]. The primary steps of example-based method are 1) collect examples in a database, 2) given an input, retrieve similar examples from the database, and 3) adapt the results of the similar examples to the current input and obtain the output. Utsuro et al. [2] propose an example retrieval method for avoiding full retrieval of examples. The proposed method generates retrieval queries from similarities, retrieving examples through the tree structure of a thesaurus and then using binary search along subsumption ordering of retrieval queries. Cranias et al. [I] introduce a matching method that measures similarity according to both surface structure and content. Another contribution involves the use of a clustering procedure to make the best matching example from the database. This method relies on the segmentation of sentences into coherent segments and their alignment at the …

Content truncated to 500 words