This paper proposes three distortion models to explicitly incorporate the word reordering knowledge into attention-based Neural Machine Translation (NMT) for further improving translation performance.
No takes yet. Share an insight, caveat, or question.
Zhang et al. (2017) studied this question.