The Phrase-Based Statistical Machine Translation (PB-SMT) model has recently begun to include source context modeling, under the assumption that the proper lexical choice of an ambiguous word can be determined from the context in which it appears. Various types of lexical and syntactic features such as words, parts-of-speech, and supertags have been explored as effective source context in SMT. In this paper, we show that position-independent syntactic dependency relations of the head of a source phrase can be modeled as useful source context to improve target phrase selection and thereby improve overall performance of PB-SMT. On a Dutch—English translation task, by combining dependency relations and syntactic contextual features (part-of-speech), we achieved a 1.0 BLEU (Papineni et al., 2002) point improvement (3.1% relative) over the baseline.
|Title of host publication||PACLIC 23|
|Subtitle of host publication||the 23rd Pacific Asia Conference on Language, Information and Computation|
|Editors||B. T'sou, C.-R. Huang|
|Place of Publication||Hong Kong, China|
|Publication status||Published - 2009|
Haque, R., Naskar, S. K., van den Bosch, A., & Way, A. (2009). Dependency relations as source context in phrase-based SMT. In B. T'sou, & C-R. Huang (Eds.), PACLIC 23: the 23rd Pacific Asia Conference on Language, Information and Computation (pp. 170-179). PACLIC.