Phrase-Based Chinese-Vietnamese Pseudo-Parallel Sentence Pair Generation

Jiaxin Zhai,Zhengtao Yu,Shengxiang Gao,Zhenhan Wang,Liuqing Pu

DOI: https://doi.org/10.1007/978-981-15-1721-1_6

2019-01-01

Abstract:The lack of Chinese-Vietnamese parallel corpus has resulted in poor translation of Chinese-Vietnamese neural machine translation. In order to solve this problem, we propose a phrase-based Chinese-Vietnamese pseudo-parallel sentence pair generation method. This method expands the corpus of Chinese-Vietnamese neural machine translation and improves the performance of Chinese-Vietnamese neural machine translation. Firstly, based on the small-scale Chinese-Vietnamese parallel corpus, the method selects the phrase module according to the phrase syntactic structure information. Then this method combines word alignment information with replacement rules. Finally, the method achieves the expansion of Chinese-Vietnamese pseudo-parallel corpus. Experiments show that this method can effectively generate Chinese-Vietnamese pseudo-parallel sentence pairs and improve the performance of Chinese-Vietnamese neural machine translation.