基于主题模型的统计机器翻译研究

With the advantage of automatically discovering semantic knowledge, latent topic models have been widely used to model the generation process of texts. Recently, topic model based statistical machine translation has become an important research direction of machine translation. In this paper, we review the research advancement in this field. First, we classify and summarize the related work. Then, we briefly introduce the latest work of our group. Finally, we make a conclusion and look to the future research direction.