The Department of Computer Science at the University of Cyprus cordially invites you to the Colloquium entitled:
Preference-based Big Data Exploration
Speaker: Prof. Panos K. Chrysanthis
As the amount of data being generated every day increases exponentially, the term “Big Data” used to represent the challenge of large-scale data processing, is being mentioned more and more frequently in everyday life. This reflects the fact that people are increasingly relying on using data to drive their daily activities and decisions. Given the volumes of data, the challenge is how to avoid overwhelming the users with irrelevant results. Query personalization is a well-known technique in dealing with this challenge by utilizing user preferences with the goal of providing relevant results to the users. Along with preferences, diversity is another important aspect of query personalization which reduces the amount of redundant information included in the results. In this talk, we will present two new personalization techniques that significantly improve big data exploration by utilizing all types of user preferences in ranking and diversification. We will first present the HYPRE graph model and prototype system that integrate qualitative and quantitative preferences by means of preference strength or intensity. In the HYPRE model, users submit both qualitative and quantitative preferences along with an intensity value, both of which are used to filter and rank the query results. Then we will introduce a new framework called Preferential Diversity (PrefDiv), which is capable of generating results that are not only relevant to users' preference but are also diverse. Our framework provides users with a fine control over the trade-off between relevancy and diversity through intuitive tunable parameters. We design and implement a prototype of a real system for PrefDiv and design algorithms to work with the HYPER hybrid preferences model. Our experimental evaluations show that PrefDiv can successfully increase coverage of the result set compared to other alternatives, and achieves a significantly better Relevancy-Diversity trade-off ratio than other models. This work was in collaboration with Roxana Gheorghiu, Xiaoyu Ge and Alexandros Labrinidis.
Panos K. Chrysanthis is a Professor of Computer Science and the founder and director of the Advanced Data Management Technologies Laboratory at the University of Pittsburgh. Among his research interests are big database systems, data stream processing, mobile and pervasive data management, and distributed computing. He has fostered interdisciplinary collaborations between computer science, medicine, astronomy and mechanical engineering, both within and outside the University of Pittsburgh. His research contributions in principles, algorithms and prototypes to data management have been documented in more than 150 papers in top journals and prestigious, peer-reviewed conferences and workshops. In 1995, he received one of the first NSF CAREER Awards for his pioneer work on mobile data management and in 2010, he was recognized as a Distinguished Scientist by ACM. In 2007, he was also elevated to the level of a Senior Member of IEEE. The impact of his work is also evident in his appointment to the editorial board of several journals, his selection as a General and Program Chair of conferences and workshops and his invitations as a keynote speaker in various meetings. He was invited to offer tutorials, contribute book chapters, and organize and participate in NSF and Dagstuhl planning meetings. He has repeatedly served as a Program Committee member in all major data management conferences and his work has appeared in textbooks. For more information, please see http://db.cs.pitt.edu.
Slides available at the following URL: https://www.cs.ucy.ac.cy/colloquium/slides/2015-chrysanthis-talk-slides.pdf
|Mailing List: https://listserv.cs.ucy.ac.cy/mailman/listinfo/cs-colloquium|
|Sponsor: The CS Colloquium Series is supported by a generous donation from