A Hybrid Approach For Spoken Language Machine Translation

In this paper, we propose a hybrid approach, which is a statistical machine translation (SMT), while using an example-based decoder. In this way, it will solve efficiently the re-ordering problem in SMT and the problems for spoken language MT, such as lots of omissions, idioms etc. We present a novel re-ordering model for SMT firstly and then an example-based decoder. Through experiments, we show that this approach obtains significant improvements over the baseline on a Chinese-English spoken language translation task.