Monthly Archives: 十月 2018

  • 0

讲座信息:On the Generalizability of Results from Interactive Information Retrieval Research

时间: 10月16日 周二 上午9:30-10:30
题目: On the Generalizability of Results from Interactive Information Retrieval Research
报告人:  Professor Diane Kelly


The idealized model of conducting empirical research starts with a theory, which is then used to derive one or more hypotheses, which are then evaluated using an appropriate method. In this idealized model, statistical methods are used to evaluate the hypothesis. Very often, especially in fields that do not have a strong tradition of theory-building and testing, such as interactive information retrieval, researchers deduce hypotheses from past empirical research reports. But what if these research reports cannot be trusted or generalized?  What if the findings are strictly a function of the time at which the studies were conducted, or the environments in which they were conducted? What if the measures themselves produce findings that are unlikely to be observed in another study context even when the same instruments are used? Concerns about the generalizability, replicability and reproducibility of research is of growing interest to those working in many research specialties, including information retrieval.  This talk focuses on one aspect of generalizability – the extent to which the findings from one research study can be used to make predictions about what will happen in another research study – and  considers how different community practices with respect replicability and reproducibility might help us address result generalizability so we can begin to construct more lasting and useful theories about information search behaviors.


Diane Kelly is Professor and Director of the School of Information Sciences at the University of Tennessee.  Prior to this, she was a professor at the University of North Carolina at Chapel Hill. Her research and teaching interests are in interactive information search and retrieval, information search behavior, and research methods. She is the recipient of the 2014 ASIST Research Award, the 2013 British Computer Society’s IRSG Karen Spärck Jones Award, the 2009 ASIST/Thomson Reuters Outstanding Information Science Teacher Award and the 2007 SILS Outstanding Teacher of the Year Award.  She is the current chair of ACM SIGIR, associate editor of ACM Transactions on Information Systems and serves on the editorial boards of several journals including, Information Processing & Management, and Information Retrieval Journal.  Kelly received a PhD, MLS and a graduate certificate in cognitive science from Rutgers University and an undergraduate degree from the University of Alabama.


  • 0

讲座信息: 搜索引擎的技术趋势和精准度提高

Category : 人工智能 , 信息检索

时间: 10月10日 周三 上午10:00-11:30
题目: 搜索引擎的技术趋势和精准度提高
报告人: 常毅


互联网搜索行业具有后发劣势,而搜索精准度是互联网搜索引擎的核心竞争力。 基于机器学习的排序算法是提高搜索精准度的高效手段,然而,提高搜索引擎的精准度不仅仅只是排序学习, 而是一个系统工程。我们将介绍该领域的历史,现状及其最新发展,并展示提高搜索精准度的实践经验和最新研究成果: 从排序学习算法、多种语义匹配特征、查询重构这三个角度出发,提高搜索引擎的精准度(该研究成果获得KDD 2016最佳论文奖)。


常毅教授,现任吉林大学人工智能学院院长,第十三批“千人计划”教授,美国计算机学会杰出科学家。曾任华为美国研究所技术副总裁,负责华为公司智能搜索、问答系统、知识图谱的技术研发。之前曾就职于雅虎研究院,任研究总监,负责其搜索研究部门, 主持雅虎公司的几乎所有重要搜索产品的应用研究和精准度提高。他在国际一流会议和期刊上发表100多篇论文,并发表了一本专著,30余项专利,并荣获了WSDM 2016最佳论文奖, KDD 2016最佳论文奖。他现担任TKDE副主编, WSDM 2018大会主席, SIGIR 2020大会主席。