Truthfulness Analysis Of Fact Statements

2011-06-14 11:27 Share:

Lecturer: Professor Weiyi Meng

Lecture Time: June 15 10:00 AM– 12:00 AM

Lecture Place: Yi Fu Conference Center Room 2

Introduction of Speaker:

Weiyi Meng is currently a professor in the Department of Computer Science of the State University of New York at Binghamton. He received his Ph.D. from the Department of Computer Science at University of Illinois at Chicago in 1992. He is the co-author of two books “Advanced Metasearch Engine Technology” and “Principles of Database Query Processing for Advanced Applications”. He has published over 100 papers and many of them in top conferences and journals, including SIGMOD, VLDB, ICDE, WWW, KDD, SIGIR, ACM TODS, ACM TOIS, and IEEE TKDE. He has served as general chair/program chair/program committee member of many international conferences, his research has been focused on information truthfulness, opinion retrieval, metasearch engine, Web data integration, Internet-based Information Retrieval, and information extraction.

Abstract:

The Web has become the most popular place for people to acquire information. Unfortunately, it is widely recognized that the Web contains a significant amount of untruthful information. As a result, good tools are needed to help Web users determine the truthfulness of certain information. In this talk, I will introduce a two-step method that aims to determine whether a given statement is truthful, and if it is not, find out the truthful statement most related to the given statement. In the first step, we try to find a small number of alternative statements of the same topic as the given statement and make sure that one of these statements is truthful. In the second step, we identify the truthful statement from the given statement and the alternative statements. Both steps heavily rely on analyzing various features extracted from the search results returned by a popular search engine for appropriate queries. Our experimental results show that the proposed method very promising. I will also briefly introduce some ongoing work and future directions in this topic.