Title: Text Understanding using a Probabilistic Knowledge-base

Date: 9 February 2012

Time: 5:15 pm – 5:45 pm

Venue: Room G22, L1 Lecture Theater, Institute of Chinese Studies

Speaker:  Dr. Haixun Wang

               Senior Researcher in the Data Management Team

               Microsoft Research Asia

Abstract:

Integrating, representing, and reasoning over human knowledge is a computational grand challenge for the 21st century. In this talk, I will introduce the Probase project at Microsoft Research Asia. The goal of the Probase project is to enable machines to understand human communications. Much interest has been devoted to building universal ontologies, either automatically constructed or built by community effort, but these have limited scope. Freebase, the best-known community-built taxonomy, contains approximately 1,500 concepts, a far cry from covering everything that exists. Probase is a universal, probabilistic taxonomy more comprehensive than any current taxonomy. It contains more than 2 million concepts, harnessed automatically from a corpus of 1.68 billion web pages and two years’ worth of search-log data. It enables probabilistic interpretations of this information. The probabilistic nature enables it to incorporate heterogeneous information naturally. I will explain how the core taxonomy, which contains hypernym-hyponym relationships, is constructed and how it models knowledge's inherent uncertainty, ambiguity, and inconsistency. 

Biography:

Haixun Wang is a senior researcher at Microsoft Research Asia in Beijing, China, where he leads the data management team. Before joining Microsoft, he had been a research staff member at IBM T. J. Watson Research Center for 9 years. Haixun Wang has published more than 120 research papers in referred international journals and conference proceedings. He is associate editor of IEEE Transactions of Knowledge and Data Engineering (TKDE), Knowledge and Information Systems (KAIS), Journal of Computer Science and Technology (JCST). He is PC co-Chair of CIKM 2012, ICMLA 2011, WAIM 2011. Haixun Wang got the ER 2008 Conference best paper award (DKE 25 year award), and ICDM 2009 Best Student Paper run-up award.