The Machine Learning Center at Georgia Tech presents a seminar by Robert Nowak, a professor at the University of Wisconsin-Madison, where his research focuses on signal processing, machine learning, optimization, and statistics. The event will be held in the GTMI Auditorium from 12:15-1:15 p.m. and is open to the public.
RSVP at http://bit.ly/2WVYf9I
Ranking from Ratings and Comparisons
Ranking items based on ratings or pairwise comparisons is ubiquitous, with applications ranging from product recommendations to teaching evaluations. This talk describes ongoing research in two directions. First, I present a new approach to deriving confidence intervals tailored to rating systems using techniques from information theory, including Sanov and Csiszar inequalities. The new intervals are often considerably tighter than commonly used bounds such as Hoeffding and Bernstein inequalities. The utility of the new confidence intervals is illustrated with applications to recommender systems and multi-armed bandits. Second, I will discuss ranking algorithms based on adaptively collecting ratings or comparisons. Consider an algorithm that can draw samples from K unknown distributions. The goal is to order the distributions according to their (unknown) means using a minimal number of samples. This model encompasses many problems including best-arms identification in multi-armed bandits, noisy sorting and ranking, and outlier detection. Ranking urban scenes based on human-perceived safety motivates our work. A challenge in such applications is that correctly ordering distributions with nearly equal means is an expensive task (in terms of number of required samples). Moreover, it is arguably unnecessary to resolve the order of very similar distributions. This observation motivates the recovery of a partial ordering into clusters of distributions with similar means (e.g., safety ratings). Clustering requires locating large “gaps” in the ordered sequence of means, and I will focus on the fundamental problem of quickly finding the largest gap.
Robert Nowak received the B.S., M.S., and Ph.D. degrees in electrical engineering from the University of Wisconsin-Madison in 1990, 1992, and 1995, respectively. He was a Postdoctoral Fellow at Rice University in 1995-1996, an Assistant Professor at Michigan State University from 1996-1999, held Assistant and Associate Professor positions at Rice University from 1999-2003, and is now the McFarland-Bascom Professor of Engineering at the University of Wisconsin-Madison. Professor Nowak has held visiting positions at INRIA, Sophia-Antipolis (2001), and Trinity College, Cambridge (2010). He has served as an Associate Editor for the IEEE Transactions on Image Processing and the ACM Transactions on Sensor Networks, and as the Secretary of the SIAM Activity Group on Imaging Science. He was General Chair for the 2007 IEEE Statistical Signal Processing workshop and Technical Program Chair for the 2003 IEEE Statistical Signal Processing Workshop and the 2004 IEEE/ACM International Symposium on Information Processing in Sensor Networks. Professor Nowak received the General Electric Genius of Invention Award (1993), the National Science Foundation CAREER Award (1997), the Army Research Office Young Investigator Program Award (1999), the Office of Naval Research Young Investigator Program Award (2000), the IEEE Signal Processing Society Young Author Best Paper Award (2000), the IEEE Signal Processing Society Best Paper Award (2011), and the ASPRS Talbert Abrams Paper Award (2012). He is a Fellow of the Institute of Electrical and Electronics Engineers (IEEE). His research interests include signal processing, machine learning, imaging, and network science, and applications in communications, bioimaging, and systems biology. His Google Scholar page contains further information about his research and publications.
For scheduling information please contact Yao Xie at email@example.com