学术报告(Charles Ling)

Active Learning with Generalized Queries

报告人：  Charles Ling, PhD, Professor
   Director, Data Mining and E-Business Lab，
Department of Computer Science，
University of Western Ontario，Canada

时间：5月18日（周一）下午3点
地点：蒙民伟楼109室

Abstract：
Active learning can actively select or construct examples (queries) and reques

t their labels to reduce the number of labeled examples needed for building an

 accurate classifier. However, previous works of active learning can only ask

specific queries with all attribute values, many of which may be irrelevant. A

 more natural and powerful way is to ask ``generalized queries'' with only rel

evant attributes, such as ``are people over 50 with knee pain likely to have o

steoarthritis?'' (with only two attributes: age and type of pain while omittin

g many other irrelevant ones, such as fever, blood type, etc.). The power of a

sking such generalized queries is that one generalized query may be equivalent

 to many specific ones. However, overly general queries may receive uncertain

labels from the oracle, and this makes learning difficult.
We propose a novel active learning algorithm that asks good generalized querie

s. We demonstrate experimentally that our new method asks significantly fewer

queries compared with the previous works of active learning. Our method can be

 readily deployed in real-world data mining tasks where obtaining labeled exam

ples is costly.

Bio：
Charles X. Ling earned his dual-BSc from Shanghai Jiao Tong Univ in China, and

 both of his MSc and PhD from Computer and Information Science at Univ of Penn

sylvania (Ivy League) within four years.
Since then he has been a faculty member in Computer Science at University of W

estern Ontario, Canada. He is currently a Professor.
His main research areas include machine learning and data mining, cognitive mo

deling, and child education.
He has published over 100 research papers in peer-reviewed journals and confer

ences.
He is an Associate Editor for IEEE TKDE and Computational Intelligence Journal

, and IEEE Senior Member.
He is the Director of Data Mining and E-Business Lab, leading data mining deve

lopment in CRM, Bioinformatics, and the Internet.