DIMACS Workshop: On-Line Decision Making

July 12 - 15, 1999
Rutgers University, Busch Student Center

Principal Organizers:

Presented as a DIMACS Exploratory Workshop and also presented under the auspices of the DIMACS Special Year on Networks.

Workshop Program:

Monday July 12, 1999

8:15 - 8:55
Registration

8:55 - 9:00
Welcome and Greeting:

9:00 - 10:00
Introduction and Overview
Yoav Freund, AT&T Labs-Research

10:15 - 11:15
Learning in Games
Rakesh Vohra, Northwestern University and David Levine, University of California - Los Angeles

11:30 - 12:30
An overview of the competitive analysis of online algorithms and its relation to game theory
Allan Borodin, University of Toronto and Ran El-Yaniv, The Technion, Israel

12:30 - 1:30
Lunch

1:30 - 2:30
The Prequential Approach to Probability and Statistics
Philip Dawid, University College London

2:45 - 3:45
Decision Theory of Regret for Universal Coding, Gambling, and Prediction
Andrew Barron, Yale University

4:00 - 5:00
Reinforcement learning as a useful apporximation: accuracy of prediction in experimental games.
Alvin Roth, Harvard University

Tuesday July 13, 1999

8:30 - 9:00
Registration

9:00 - 10:30
Adaptive Srategies, Regret, and Approachability
Sergiu Hart, Hebrew University, Jerusalem

10:45 - 11:45
Information Geometry and its Applications
Imre Csiszar, Hungarian Academy of Sciences

11:45 - 1:00
Lunch

1:00 - 2:30
Relative Loss Bounds for On-Line Learning Using Divergence Measures
Manfred Warmuth, University of California - Santa Cruz

2:45 - 3:45
Learning to Play Games Using Multiplicative Weights
Robert Schapire, AT&T Labs

4:00 - 4:40
Worst-case analyses of the exploration/exploitation tradeoff
Phil Long, National University of Singapore

4:50 - 5:30
Tracking the Best Predictor
Mark Herbster, Royal Holloway University

8:00
Reception - Holiday Inn, South Plainfield

Wednesday July 14, 1999

8:30 - 9:00
Registration

9:00 - 10:00
Rational Learning
Ehud Kalai, Northwestern University

10:15 - 10:55
Rational Bayesian Learning in Repeated Games
John Nachbar, Wash University, St. Louis

11:15 - 11:55
"When rational learning fails"
Dean Foster, Wharton School, University of Pennsylvania and Peyton Young,
John Hopkins University

11:55 - 1:00
Lunch

1:00 - 1:40
`Impossibility' of Absolute Continuity
Chris Sanchirrico, Columbia University

1:55 - 2:35
Minimizing regret: the general case
Aldo Rustichini, Tilburg University

2:50 - 3:30
Reasonable Learning in Non-cooperative Games and the Internet
Eric Friedman, Rutgers University, Department of Economics

3:45 - 4:25
Bayesian representations of stochastic processes under learning - De-Finetti revisited
Matt Jackson, CalTech; Rann Smorodnitsky, Technion, and Ehud Kalai, Northwestern University

4:35 - 5:35
Fast Universal Portfolios For Parameterized Target Classes
Jason Cross, Yale University

Thursday July 15, 1999

8:30 - 9:00
Registration

9:00 - 10:00
On-Line Learning and the Metrical Task System Problem
Avrim Blum, Carnegie Mellon University

10:15 - 10:55
On-Line load balancing and routing
Yossi Azar, Tel Aviv University

11:10 - 11:50
Average case and worst case sample complexity of learning distributions
Manfred Opper, Sheffield University

11:50 - 1:00
Lunch

1:00 - 1:40
On prediction of individual sequences relative to a set of experts.
N. Cesa-Bianchi, University of Milan, Italy and Gabor Lugosi,
Pompeu Fabra University, Spain

1:55 - 2:55
Approachability in Infinite Dimensional Spaces and an Application:
A Universal Algorithm for Generating Extended Normal Numbers
Ehud Lehrer, Tel-Aviv University

3:10 - 3:50
Calibration with Many Checking Rules
Alvaro Sandroni, Northwestern University

3:50 - 4:50
Closing Comments

Previous: Participation
Next: Registration
Workshop Index
DIMACS Homepage
Contacting the Center
Document last modified on July 12, 1999.