Enabling Integrated Information Access and Retrieval

实现综合信息访问和检索

• 批准号: RGPIN-2015-04978
• 负责人: Clarke, Charles
• 金额: $ 3.64万
• 依托单位: University of Waterloo
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2018
• 资助国家: 加拿大
• 起止时间: 2018-01-01 至 2019-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=647784
• 关键词: Enabling; Integrated; Information; Access; Retrieval

Enabling access to information remains a complex and challenging task, and the tremendous success of commercial Web search engines has raised user expectations to extraordinarily high levels. Translating this success to domains beyond Web search, while matching these user expectations, forms a central goal of my work. Broadly, my research interests encompass a range of interrelated topics within the field of information retrieval, including search interfaces, evaluation, indexing, ranking, and text data mining. My work is primarily of an applied nature, with the results validated through implementation, experimentation, and performance measurement.****The activities proposed under this application build upon my current research, particularly my work on indexing, ranking, and modeling user behaviour. Over the next five years, I plan to extend this work to topics such as efficient multi-staged ranking, focused retrieval, high-recall retrieval, and search sessions. In addition, I plan to address problems in specific domains, such as city search.****My research has potential for impact at several levels. Basic research on efficient indexing and ranking can increase system performance, reducing requirements for machines and power, which is important for systems operating over thousands of machines. New methods for focused retrieval would facilitate information access across complex collections of books, journals and other documents, helping researchers in science and other areas. Search technology to maximize recall (i.e., the amount of relevant material found) while minimizing human effort has application in receptor domains where near-exhaustive results are required.  Methods for interpreting and modeling search sessions, particularly in the light of personalization and context, can directly translate to improved searcher experience. Finally, methods for point-of-interest recommendation and city search have the potential to enable access to open data and shared experience through the infrastructure of a living lab.****To transfer technology, I work closely with industry, including major commercial search companies, software consultants, as well as a number of local Waterloo-based start-ups. My graduating students have gone on to opportunities in industry and academia, both within Canada and internationally, including the founding of a least one Canadian start-up.  I also regularly interact with receptors outside the software industry who are involved in information access problems across many domains.  I intend to maintain this collaborative approach over the lifetime of this Discovery Grant, with a continuing emphasis on links between graduate student research activities and industry.**

实现信息访问仍然是一项复杂且具有挑战性的任务，商业网络搜索引擎的巨大成功将用户的期望提高到了极高的水平。将这种成功转化为网络搜索以外的领域，同时满足这些用户的期望，构成了我工作的中心目标。广泛而言，我的研究兴趣涵盖信息检索领域内的一系列相关主题，包括搜索界面、评估、索引、排名和文本数据挖掘。我的工作主要是应用性质的，通过实施、实验和性能测量来验证结果。****在此应用程序下提出的活动建立在我当前的研究基础上，特别是我在索引、排名和用户行为建模方面的工作。在接下来的五年里，我计划将这项工作扩展到高效的多阶段排名、聚焦检索、高召回率检索和搜索会话等主题。此外，我计划解决特定领域的问题，例如城市搜索。****我的研究有可能在多个层面产生影响。 对高效索引和排名的基础研究可以提高系统性能，降低对机器和功率的要求，这对于运行数千台机器的系统非常重要。 集中检索的新方法将有助于跨复杂的书籍、期刊和其他文档集合进行信息访问，从而帮助科学和其他领域的研究人员。最大限度地提高召回率（即找到的相关材料的数量）同时最大限度地减少人为努力的搜索技术可应用于需要近乎详尽的结果的受体领域。  解释和建模搜索会话的方法，特别是根据个性化和上下文，可以直接转化为改进的搜索者体验。 最后，兴趣点推荐和城市搜索方法有可能通过生活实验室的基础设施访问开放数据和共享经验。****为了转让技术，我与业界密切合作，包括主要的商业搜索公司、软件顾问以及滑铁卢当地的一些初创企业。 我的毕业生在加拿大国内和国际上获得了工业界和学术界的机会，包括创办了至少一家加拿大初创企业。  我还定期与软件行业之外的接受者进行互动，他们涉及多个领域的信息访问问题。  我打算在该发现补助金的整个生命周期中保持这种合作方式，并继续强调研究生研究活动与行业之间的联系。 **