English-Chinese Translation using Neural Machine Models

During the last few years, neural machine translation (NMT) as a notable branch of machine translation has been increasing its popularity both in research and in practice. In particular, neural machine translation between English and Chinese, which are two of the most widely used languages all over the world, is receiving more and more attention. In this review, we discuss current mainstream models for neural machine translation, including recurrent neural network (RNN), convolution neural network (CNN), and self-attention network (SAN) or transformer. The mechanisms of these models are illustrated. Moreover, examples of studies and applications of them are analyzed. In addition, comparisons on the performance of different models are implemented.