Multi-Granularity Self-Attention for Neural Machine Translation.

Jie Hao, Xing Wang, Shuming Shi, Jinfeng Zhang, Zhaopeng Tu

Browse the full EMNLP paper archive.