复杂自然语言的简化处理

Concise Processing of Complex Natural Languages

目前自然语言处理系统难以正确解释部分复杂句子,其中的知识关系只能由操作者简化后再输入,如何使复杂的句子直接被计算机理解呢?本文针对这一问题而提出了自动识别关键字词的新算法。与人的大脑类似,知识处理机也可以对简化后的不完整信息(甚至缺少大多数语言处理机所必需的链接谓词)进行准确理解,而且这种理解对复杂句子要比逐字逐句地理解更为精确和快速。

It is hard for most knowledge extraction machines to correctly construct a complex sentence that can contain several clauses and connections. For this reason, many knowledge systems need to input complex sentences by human. In this paper, the way without human inputting by using concise processing has been presented and analyzed. In this way, the time-consuming and misunderstanding by computer can be largely decreased and the accuracy can be enhanced. As the same as the brain in human being, the incomplete sentence can also be understood by machine in a way of extracting the key information in the sentence.