主讲人：Dr. Zhaochun Ren
题目 Multi-viewpoint summarization and social collaborative viewpoint regression
Abstract: A viewpoint is a triple consisting of an entity, a topic related to this entity and sentiment towards this topic. In multi-viewpoint summarization one monitors viewpoints for a running topic and selects a small set of informative documents. We first focus on multi-viewpoint summarization of multilingual social text streams. Our approach includes three core ingredients: dynamic viewpoint modeling, cross-language viewpoint alignment, and, finally, multi-viewpoint summarization. Specifically, we propose a dynamic latent factor model to explicitly characterize a set of viewpoints through which entities, topics and sentiment labels during a time interval are derived jointly; we connect viewpoints in different languages by using an entity-based semantic similarity measure; and we employ an update viewpoint summarization strategy to generate a time-aware summary to reflect viewpoints. Moreover, we utilize viewpoints from user reviews and trusted social relations for generating explainable results in recommender systems. We propose a latent variable model, called social collaborative viewpoint regression, for predicting item ratings based on user opinions and social relations. Experiments conducted on real-world benchmark datasets demonstrate the effectiveness of our proposed methods for summarizing multi-viewpoints and for predicting item ratings with explanations.
Bio: Dr. Zhaochun Ren is currently a research associate in the department of computer science, University College London (UCL), UK. Prior to that, he received his Ph.D. from the University of Amsterdam, The Netherlands, in 2016. Zhaochun received his B.E and M.Sc from Shandong University in 2009 and 2012 respectively, and worked as a short-term visiting scholar in Max-Planck-Institut für Informatik, Germany, 2012. His research interests and expertise lie in social media search and analysis, healthcare data mining, web mining, content analysis and semantic search in e-discovery. Zhaochun Ren has published over 30 papers in journals and conferences. Various parts of his work have been published in top-tier venues including SIGIR, KDD, AAAI, WSDM, CIKM, and TKDE. He has served as PC member or reviewer for reputable conferences and journals, including SIGIR, ACL, TOIS, IP&M, CIKM, EACL, ECIR, and AIRS. Dr. Ren is now serving as the information director of ACM Transactions on Information Systems (TOIS, CCF A), and an editor for Information Processing and Management Journal (IP&M, CCF B).