We developed a two-stage MT system. The first stage consists of an automatically created pattern-based machine translation system. The second stage consists of an phrase-based SMT system. Our goal with this system is to obtain fewer ungrammatical sentences. We performed ABX tests between the output of a standard SMT system (Moses) and the output of the proposed system for 100 sentences. The results indicated that 30 sentences output by the proposed system were evaluated as better than those output by the standard SMT system. In contrast, 9 sentences output by the standard SMT system were thought to be better than those output by the proposed system. This means that our proposed system functioned effectively in the Japanese-English simple sentence task.
We need to overcome several difficulties in order to improve the proposed methods. Moreover, there were many bugs in our system. We will focus on how to solve such difficulties in the future.