Modeling Discourse Structure For Document-Level Neural Machine Translation
WORKSHOP ON AUTOMATIC SIMULTANEOUS TRANSLATION CHALLENGES, RECENT ADVANCES, AND FUTURE DIRECTIONS(2020)
摘要
Recently, document-level neural machine translation (NMT) has become a hot topic in the community of machine translation. De spite its success, most of existing studies ignored the discourse structure information of the input document to be translated, which has shown effective in other tasks. In this paper, we propose to improve document-level NMT with the aid of discourse structure in formation. Our encoder is based on a hierarchical attention network (HAN) (Miculicich et al., 2018). Specifically, we first parse the input document to obtain its discourse structure. Then, we introduce a Transformer-based path encoder to embed the discourse structure information of each word. Finally, we combine the discourse structure information with the word embedding before it is fed into the encoder. Experimental results on the English to-German dataset show that our model can significantly outperform both Transformer and Transformer+HAN.
更多查看译文
关键词
discourse structure,translation,document-level
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络