Development and evaluation of information retrieval methods to support users with complex information needs

解决复杂的现实问题通常需要获取大量的信息。一开始，用户可能对主题领域知之甚少，可能不知道他需要什么信息来解决手头的问题。当前可用的搜索引擎非常擅长根据定义良好的查询检索主题相关的文档。但是，如果用户对问题域没有很好的理解，并且不能制定有效的查询，那么现有的搜索引擎就没有太大的帮助。当前用户有一个复杂的搜索任务，例如：“我需要知道什么才能在加拿大创办自己的公司？”，需要迭代地制定几个查询，浏览大量检索的文档，并确定重要的事实、实体（例如人员、组织）及其相互关系，以便获得解决手头问题所需的必要知识。在本研究中，我将探讨如何在复杂的问题解决过程中为用户提供连贯的信息寻求体验。这包括设计在文档中寻找有用信息、概念和实体的方法，识别它们之间的关系，并以有意义和连贯的方式向用户呈现它们。

Solving complex real world problems usually requires access to large amounts of information. At the start, a user may have little knowledge of the subject domain and may not know what information he needs in order to solve the problem at hand. Currently available search engines are very good at retrieving topically relevant documents in response to a well-defined query. However, if a user does not have a good understanding of the problem domain, and cannot formulate effective queries, then existing search engines are not very helpful. Currently a user with a complex search task, e.g. "What do I need to know to start my own company in Canada?", needs to formulate iteratively several queries, go through large numbers of retrieved documents, and identify important facts, entities (e.g. people, organizations), and their interrelationships in order to obtain the necessary knowledge needed to solve the problem at hand. In the proposed research I will investigate how to provide users with a coherent information seeking experience throughout their complex problem solving processes. This includes the design of methods for finding useful information nuggets, concepts and entities in documents, identifying relationships between them and presenting them in a meaningful and coherent way to the user.