Retrieval Methods for Information Complementation and Its Applications to Cross-media Information Retrieval

在这项工作中，为了实现一个名为信息营养师的信息系统，它检测信息中的偏见，为用户提供补充信息，我们研究了信息的关系分析和组织机制，特别是对话题分布、因果关系和利益关系进行了分析。基于这些方法的应用系统已经开发出来。我们利用网络内容、电视节目和新闻文章的数据进行了实验。这些实验结果验证了我们的方法。

In this work, to implement a information system called "Information Nutritionist", which detects bias in information and provides complementary information to users, we are studying on relationship analysis and organization mechanisms of information ; especially, we are working on analyzing topic distribution, causal and interests relations. Application systems based on these methods have been developed. We carried out experiments by using the data of Web contents, TV-programs and news articles. These experimental results have validated our methods.