An experiment on Japanese-Uighur statistical machine translation with increased corpus

摘要

In this paper, we present the results of our work on the development of a phrase-based statistical machine translation system from Japanese to Uyghur — an agglutinative language with productive inflectional and derivational morphology. We experiment with different morpheme-level representations for Japanese-Uyghur parallel texts. We improve from our baseline 2.61 BLEU points for our phrase-based baseline model to 7.03 BLEU points for an improvement of 4.42 points.