Lack of Fluency is Hurting Your Translation Model

Many machine translation models are trained on bilingual corpus, which consist of aligned sentence pairs from two different languages with same semantic. However, there is a qualitative discrepancy between train and test set in bilingual corpus. While the most train sentences are created via automatic techniques such as crawling and sentence-alignment methods, the test sentences are annotated with the consideration of fluency by human. We suppose this discrepancy in training corpus will yield performance drop of translation model. In this work, we define textitfluency noise to determine which parts of train sentences cause them to seem unnatural. We show that textitfluency noise can be detected by simple gradient-based method with pre-trained classifier. By removing textitfluency noise in train sentences, our final model outperforms the baseline on WMT-14 DErightarrowEN and RUrightarrowEN. We also show the compatibility with back-translation augmentation, which has been commonly used to improve the fluency of the translation model. At last, the qualitative analysis of textitfluency noise provides the insight of what points we should focus on.