Unsupervised, Non-stop Extraction of Information from the World Wide Web

无监督、不间断地从万维网上提取信息

• 批准号: 0535284
• 负责人: Oren Etzioni
• 金额: --
• 依托单位: University of Washington
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2006
• 资助国家: 美国
• 起止时间: 2006-03-01 至 2009-02-28
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0535284&HistoricalAwards=false
• 关键词: Unsupervised; Non; stop; Extraction; Information

A long standing dream of Artificial Intelligence has been to create anautonomous learner that continuously increases its knowledge byreading a wide variety of texts. To date, this kind of knowledgeacquisition has been attempted only rarely, and typically at smallscale. Yet the Web has made a vast library of online text readilyaccessible. In response, this project investigates a family of unsupervised,domain-independent, scalable systems that learn from the Web in anopen-ended fashion. Such systems not only extract information butalso extend their ontology, incorporating new classes and relations.Furthermore, the systems' learning is recursive -- once new relationsare learned, the system builds on these to learn new relations,relations between relations, and so on. The project investigates theautomatic control of this process, and analyzes both the power andlimitations of this form of learning.The project will impact both the natural language processing (NLP) and machine learning (ML) communitiesby investigating fundamental issues in learning from text. Inaddition, the systems developed could lead to a new generation of Websearch engines that improve information access, achieving a broadsocietal and economic impact.

人工智能的一个长期梦想是创造一个自主学习者，通过阅读各种文本来不断增长知识。到目前为止，这种知识获取的尝试很少，而且通常是小规模的。然而，网络已经使庞大的在线文本图书馆变得随时可用。作为回应，这个项目调查了一系列无监督的、独立于领域的、可伸缩的系统，它们以一种开放的方式从Web学习。这样的系统不仅提取信息，而且扩展了它们的本体，加入了新的类和关系。此外，系统的学习是递归的--一旦学习到新的关系，系统就建立在这些基础上来学习新的关系、关系之间的关系等等。该项目调查了这一过程的自动控制，并分析了这种学习形式的力量和局限性。该项目将通过调查从文本中学习的基本问题来影响自然语言处理(NLP)和机器学习(ML)社区。此外，开发的系统可能导致新一代网络搜索引擎，改善信息获取，实现广泛的社会和经济影响。