开放信息抽取技术的现状研究

On the State-of-the-art Technology of Open Information Extraction

如何高效地自动理解网络上出现的海量文本信息,日益成为了个严峻的考验。美国华盛顿大学图灵中心提出的开放信息抽取,是一个有效的解决方法。它具有领域的独立性,无监督抽取,对大量文本的可伸缩性等特点。该论文首先介绍了开放信息抽取系统的主要功能,然后详细论述了三个主要的开放信息抽取系统的特点、组成部分以及优缺点,接着分析了开放信息抽取系统的改进方法和发展趋势。最后对未来进行展望。

How to efficiently and automatically understand the mass text information appearing on the Web is increasingly becoming a se- vere issue. Introduced by Turing Center of University Washington, Open Information Extraction （OIE} is an effective method with char- acteristics of domain-independent, unsupervised extraction, scalability to large amounts of text. This paper firstly introduces the main functions of OIE system, and then discusses in detail three major open information extraction systems about their features, components, and advantages and disadvantages, and then analyzes the improvement methods and trends of OIE. Finally, the future development of the OIE technologies is explored.