汉语机器理解与汉英机器翻译

Machine Understanding of Chinese and Chinese-English Machine Translation

自然语言的理解是机器翻译(简称M T)的基础。机器理解汉语所面临的主要困难,包括自动分词,歧义消解,意义的多层次解读等。自然语言处理中的这些难题使得计算机难以对汉语语句做出正确的描述。基于规则的汉英机器翻译的质量还有待提高。基于语料库的方法为机器翻译研究提供了新的途径。将基于规则的理性主义方法同基于语料库的经验主义方法结合起来,必将极大地提高汉英机器的翻译的质量。

Natural language understanding is essential to Machine Translation （abbreviated as MT）. For the purpose of the research of Chinese-English Machine Translation （abbreviated as CEMT）, the present paper explores an in-depth analysis of the main difficulties in machine understanding of Chinese. They are automatic word segmentation, disambiguation, the multi-layer interpretation of meanings respectiveiy. These unsolved problems in natural language processing make it difficult for computers to give out correct description of Chinese sentences. Some examples cited in the paper prove that the quality of the rule-based CEMT is still to be improved. The corpus-based approach opens up a new way to MT research. Combining rationalist and empiricist approaches to CEMT will improve the quality of CEMT greatly.