Neural Machine Translation with FastMoS

Abstract

In recent years, machine translation make great progress especially in neural machine translation \cite{Wu:2016:arXiv,luong2015effective,gehring2017convolutional,sutskever2014sequence,tu2016modeling,bahdanau2014neural,kong2019neural}. In \citep{kong2019fast}, they use a fast mos to obtain great performance on some machine translation corpora. In this project, we investigate this method to improve the performance of other machine translation corpora.