Resolving complex information needs by means of natural language processing and interaction with user

用户，尤其是那些对某一主题进行专业研究的用户，经常有复杂的信息需求。我们所说的复杂信息需求是指用户要求获取与两个或更多实体之间的特定类型的关系有关的信息，例如，“通货膨胀的原因是什么？”“什么可以预防骨质疏松症？”“氨基葡萄糖对关节炎有什么作用？”“人们对干细胞研究有什么看法？”可能有复杂信息需求的用户包括业务分析师、情报分析师、调查性记者、医生、警察侦探、律师和科学家。例如，业务分析师可能想要找到有关信用卡欺诈对加拿大银行的影响的信息。当前可用的信息检索技术不能为具有这种复杂信息需求的用户提供足够的支持：第一，界面不鼓励或帮助用户制定复杂的查询；第二，搜索算法不试图找到实体之间特定类型的关系的实例。

Users, in particular those doing professional research on a certain subject, frequently have complex information needs. By a complex information need we mean a user's requirement to obtain information pertaining to a specific type of relationship between two or more entities, such as "What are the causes of inflation?", "What can prevent osteoporosis?", "What effect does glucosamine have on arthritis?, "What do people think about stem cell research?". Some examples of users who might have complex information needs are business analysts, intelligence analysts, investigative journalists, medical doctors, police detectives, lawyers, and scientists. For instance, a business analyst might want to find information about the effects of credit card fraud on the Canadian banks. Currently available information retrieval technologies do not provide adequate support to users with such complex information needs: firstly, interfaces do not encourage or assist users in formulating complex queries, and secondly, search algorithms do not attempt to find instances of specific types of relationships between entities.