In this paper, we propose a new approach which automatically acquires translation templates from the unannotated bilingual spoken language corpora.

• 摘要本文提出了一种从未经深层次处理的双语口语语料库中自动获取机器翻译模板的方法。