Acquisition and Maintenance of Knowledge Bases for Natural Language Processing Systems

这项研究是卡内基梅隆大学的谢尔盖·尼伦堡和普渡大学的维克多·拉斯金共同完成的。他们的目标是提高自然语言处理知识库的最新水平。他们正在建立关于单词、单词含义和单词关系的知识库，这些知识库超出了“玩具”领域，足以全面地在现实情况下进行自然语言处理。这些类型的“计算词典”知识库是用于分析自然语言和生成自然语言输出的自动计算机系统的基础。本研究的意义在于，自然语言的自动处理是开发计算机系统界面的重要组成部分。自然语言处理系统离不开计算机化的词义知识库和世界知识。这项工作开始建立这种知识库的框架和获取其内容的工具，其方式将对许多研究小组有用，并适用于不同的应用。

This research is joint work of Sergei Nirenburg at Carnegie Mellon University and Victor Raskin at Purdue University. Their goal is to advance the state of the art in knowledge bases for natural language processing. They are building knowledge bases about words, word meanings and word relationships, extended beyond "toy" domains to be comprehensive enough for natural language processing in realistic situations. These kinds of "computational lexicography" knowledge bases are fundamental to automatic computer systems for analyzing natural language and for generating natural language output. The significance of this research is that automatic processing of natural language is a major component in developing computer systems interface. Computerized knowledge bases about word meanings and world knowledge are indispensable to natural language processing systems. This work begins to build the framework for such knowledge bases and the tools for acquiring their content, in a way which will be useful to many research groups and suitable for diverse applications.