Distortion Model Considering Rich Context for Statistical Machine Translation

Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics, pages 155–165, Sofia, Bulgaria, August 4-9 2013. c⃝2013 Association for Computational Linguistics Distortion Model Considering Rich Context for Statistical Machine Translation Isao Goto†,‡ Masao Utiyama† Eiichiro Sumita† Akihiro Tamura† Sadao Kurohashi‡ †National Institute of Information and Communications Technology ‡Kyoto University goto.i-es@nhk.or.jp {mutiyama, eiichiro.sumita, akihiro.tamura}@nict.go.jp kuro@i.kyoto-u.ac.jp Abstract This paper proposes new distortion mod- els for phrase-based SMT. In decoding, a distortion model estimates the source word position to be translated next (NP) given the last translated source word position (CP). We propose a distortion model that can consider the word at the CP, a word at an NP candidate, and the context of the CP and the NP candidate simultaneously. Moreover, we propose a further improved model that considers richer context by dis- criminating label sequences that specify spans from the CP to NP candidates. It enables our model to learn the effect of relative word order among NP candidates as well as to learn the effect of distances from the training data. In our experiments, our model improved 2.9 BLEU points for Japanese-English and 2.6 BLEU points for Chinese-English translation compared to the lexical reordering models. 1 Introduction Estimating appropriate word order in a target lan- guage is one of the most difficult problems for statistical machine translation (SMT). This is par- ticularly true when translating between languages with widely different word orders. To address this problem, there has been a lot of research done into word reordering: lexical reordering model (Tillman, 2004), which is one of the distortion models, reordering constraint (Zens et al., 2004), pre-ordering (Xia and Mc- Cord, 2004), hierarchical phrase-based SMT (Chi- ang, 2007), and syntax-based SMT (Yamada and Knight, 2001). In general, source language syntax is useful for handling long distance word reordering. However, obtaining syntax requires a syntactic parser, which is not available for many languages. Phrase-based SMT (Koehn et al., 2007) is a widely used SMT method that does not use a parser. Phrase-based SMT mainly1 estimates word re- ordering using distortion models2. Therefore, dis- tortion models are one of the most important com- ponents for phrase-based SMT. On the other hand, there are methods other than distortion models for improving word reordering for phrase-based SMT, such as pre-ordering or reordering constraints. However, these methods also use distortion mod- els when translating by phrase-based SMT. There- fore, distortion models do not compete against these methods and are commonly used with them. If there is a good distortion model, it will improve the translation quality of phrase-based SMT and benefit to the methods using distortion models. In this paper, we propose two distortion mod- els for phrase-based SMT. In decoding, a distor- tion model estimates the source word position to be translated next (NP) given the last translated source word position (CP). The proposed models are the pair model and the sequence model. The pair model utilizes the word at the CP, a word at an NP candidate site, and the words …

Content truncated to 500 words