Search |  Contact |  SRI Home Do not follow this link, or your host will be blocked from this site. This is a spider trap. Do not follow this link, or your host will be blocked from this site. This is a spider trap. Do not follow this link, or your host will be blocked from this site. This is a spider trap.A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A ASRI International.  333 Ravenswood Avenue.  Menlo Park, CA 94025-3493. SRI International is a nonprofit corporation.

AIC Seminar Series

A Contextual-Bandit Approach to Personalized News Article Recommendation

Lihong LiYahoo! Research[Home Page]

Notice:  Hosted by Omid Madani

Date:  Thursday, April 22nd 2010 at 4:00pm

Location:  EJ228 (SRI E building); WebEx 1-888-355-1249, 749045 (sound), (slides via Web)  (Directions)


Personalized web services strive to adapt their services (advertisements, news articles, etc.) to individual users by making use of both content and user information. Despite a few recent advances, this problem remains challenging for a few reasons. First, web service is featured with dynamically changing pools of content, rendering traditional collaborative filtering methods inapplicable. Second, the scale of most web services of practical interest calls for solutions that are both fast in learning and computation. Third, such applications normally require solving the so-called exploration/exploitation dilemma that are common in many other problems like reinforcement learning, optimal control, and experimental design, etc.

In this work, we model personalized recommendation of news articles as a contextual-bandit problem, a principled approach in which a learning algorithm sequentially selects articles to serve users based on contextual information about the users and articles, while simultaneously adapting its article-selection strategy based on historical user-click feedback in order to maximize total user clicks over time. We report some of our recent processes in this direction, including (1) a general contextual bandit algorithm that is computationally efficient and well motivated from machine-learning theory, and (2) an unbiased method for evaluating bandit algorithms offline with a static dataset (which is not as straightforward to do as in supervised learning). We successfully applied our new algorithm to a Yahoo! Front Page dataset containing over 33 million data, demonstrating a 12.5% click lift compared to a standard non-contextual bandit algorithm.

Joint work with Wei Chu, John Langford, and Rob Schapire.

   Bio for Lihong Li

Lihong Li has been a Postdoctoral Scientist in the machine learning group at Yahoo! Research since 2009. He obtained a PhD degree in Computer Science from the Rutgers University, advised by Prof. Michael Littman. Before that, he obtained a MSc degree from the University of Alberta, advised by Profs Vadim Bulitko and Russell Greiner, and BE from the Tsinghua University. His main research interests are in reinforcement learning and machine learning, including: exploration/exploitation tradeoff, value-function approximation, online learning, computational learning theory, and decision-theoretic planning. In the summers of 2006-2008, he interned at Google, Yahoo! Research, and AT&T Shannon Labs. He is a co-winner of an ICML Best Student Paper Award in 2008.

   Note for Visitors to SRI

Please arrive at least 10 minutes early as you will need to sign in by following instructions by the lobby phone at Building E (or call Wilma Lenz at 650 859 4904, or Eunice Tseng at 650 859 2799). SRI is located at 333 Ravenswood Avenue in Menlo Park. Visitors may park in the parking lots off Fourth Street. Detailed directions to SRI, as well as maps, are available from the Visiting AIC web page. There are two entrances to SRI International located on Ravenswood Ave. Please check the Building E entrance signage.

SRI International
©2018 SRI International 333 Ravenswood Avenue, Menlo Park, CA 94025-3493
SRI International is an independent, nonprofit corporation. Privacy policy