Training Neural Machine Translation using Word Embedding-based Loss

Publication
arXiv:1807.11219