Search intelligence as users expect

用户期望的搜索智能

• 批准号: 121418-2013
• 负责人: Nie, JianYun
• 金额: $ 3.21万
• 依托单位: Université de Montréal
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2015
• 资助国家: 加拿大
• 起止时间: 2015-01-01 至 2016-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=569189
• 关键词: Search; intelligence; users; expect

Search engines are part of our everyday life and they serve for different purposes. Although tremendous progress has been made in this area during the recent years, many users (and researchers) would like them to be even more intelligent - they expect that search engines could better understand and anticipate what the user needs in the current and even next search task. For example, they should understand the implicit meaning of a user query and the user's search intent. How far could we go in this direction? This is the question we will address in this research program. There are various resources available to make search engines more intelligent: general knowledge (i.e. relations between terms or concepts) in form of texts (Wikipedia), ontologies and lexical databases (e.g. WordNet), knowledge hidden in document collections (e.g. related terms) and in query logs (user preferences). The goal of this research is to push the understanding of user intents as far as possible by exploiting all these forms of knowledge and let them work in synergy to make search engines more capable of guessing what the user expects.

搜索引擎是我们日常生活的一部分，它们有不同的用途。尽管近年来在这一领域取得了巨大的进步，但许多用户（和研究人员）希望它们更加智能——他们期望搜索引擎能够更好地理解和预测用户在当前甚至下一次搜索中需要什么任务。例如，他们应该理解用户查询的隐含含义和用户的搜索意图。我们在这个方向上能走多远？这是我们将在本研究计划中解决的问题。 有多种资源可以使搜索引擎更加智能：文本形式的常识（即术语或概念之间的关系）（维基百科）、本体和词汇数据库（例如 WordNet）、隐藏在文档集合中的知识（例如相关术语）以及在查询日志中（用户首选项）。这项研究的目标是通过利用所有这些形式的知识来尽可能地推动对用户意图的理解，并让它们协同工作，使搜索引擎更有能力猜测用户的期望。