Improving Neural Machine Translation with AMR Semantic Graphs

The Seq2Seq model and its variants (ConvSeq2Seq Transformer) emerge as a promising novel solution to the machine translation problem. However, these models only focus on exploiting knowledge from bilingual sentences without paying much attention utilizing external linguistic sources such semantic representations. Not do representations can help preserve meaning but they also minimize data sparsity date, information remains rarely integrated into models. In this study, we examine effect of abstract representation (AMR) graphs in different Experimental results IWSLT15 English-Vietnamese dataset have proven efficiency proposed model, expanding use language significantly improve performance models, especially application low-resource pairs.