随着互联网的迅速发展及在线数据的急剧上升，用户从海量数据中搜索有价值的信息变得越来越困难，并且随着数据库、自然语言处理和数据挖掘等领域的相关技术不断应用到信息检索当中，搜索引擎从关键字搜索发展到智能搜索成为可能。而实现智能搜索的一项关键任务就是基于Web的实体知识库的构建。本项目基于丰富的网络资源，以建立实体知识库来为Web智能搜索服务作为基本应用目标，以实体及关系提取、实体名字排歧、实体知识库链接整合、以及实体知识库索引等关键技术作为研究重点，拟解决以下三类研究问题： 第一，基于大规模Web网页以及现有知识库，实现实体的准确提取，设计实体名字排歧方法，进行实体与知识库链接、整合；第二，对实体及其相关属性进行建模，提取实体关系；第三，开展实体知识库索引技术研究，以支持快速查询和动态更新。 最后将建立一个基于实体的知识库系统，以验证成果的正确性和可行性，为开发下一代智能搜索系统奠定基础。

With the rapid development of the Internet and the sharp rise of online data, discovering valuable information from the massive data becomes increasingly difficult. Meanwhile, with the application of the technologies from the fields of database, natural language processing, and data mining to information retrieval, it becomes feasible for keyword search to evolve to intelligent search. One of the most important steps in Web intelligent search service is to build a solid Web-based knowledge base, which turns out to be the aim of this project. With the rich Internet data, we will mainly study in this project some key techniques regarding entity and relationship extraction, entity name disambiguation, entity linking with knowledge bases, as well as the storage and index system for an entity-centric knowledge base. Specifically, we intend to explore the following three research problems. First, based on the large-scale Web data and existing knowledge bases, we will design some methods for accurately extracting entities, disambiguating entity names, and linking entities with existing knowledge bases. Second, we plan to build the entity-relation model based on all the entities and entity attributes. Third, we will develop an index scheme for the entity-centric knowledge base to support fast queries and dynamic updates. Finally, we construct a real entity-centric knowledge base system to verify the correctness and feasibility, which forms the solid foundations for the next-generation intelligent search system.