Friday, February 26, 2010
10:00 AM
CBS Room 209
Ashish Sabharwal
Cornell University
A Quest for Large-Scale Combinatorial Reasoning Methods
A central goal of Artificial Intelligence is to develop scalable and robust automated reasoning technology that will allow computers to act intelligently in increasingly complex real-world settings and in competitive and uncertain environments. Combinatorial inference, in particular for propositional satisfiability (SAT) and for constraint reasoning and optimization, has been one of the most successful areas in this quest. This talk will present my work in this area during the past few years. We will discuss the first formal study of learning methods used in today's highly engineered inference engines in terms of resolution proof complexity, and then explore new methods for more complex reasoning problems such as maximum satisfiability on multi-core platforms, counting and sampling solutions near-uniformly, and message passing techniques to uncover the clustering structure of solution spaces.

These new methods have pushed the limits of scalability by orders of magnitude, and promise to open up a range of new applications in AI, particularly those involving the integration of logical and probabilistic inference. Specifically, we will visit some exciting applications of such techniques to the newly emerging, highly-interdisciplinary field of Computational Sustainability.

Bio: Ashish Sabharwal is a Research Associate in Computer Science at Cornell University, working both at the Institute for Computational Sustainability(ICS) and the Intelligent Information Systems Institute (IISI). After receiving his M.S. and Ph.D. from the University of Washington, Seattle in 2001/2005, he continued to broaden his research interests as a Postdoctoral Associate at Cornell till 2008. He has (co-)authored nearly 40 publications and surveys, including two Best Paper Awards, one Runner-Up Prize, and four best paper nominations. His research spans combinatorial reasoning, constraints, probabilistic inference, multi-agent and adversarial reasoning, and optimization. Most recently, his focus is on applying these reasoning techniques to applications in Computational Sustainability, ranging from wildlife conservation to sustainable community development and discovery of new materials.

Refreshments will be provided at 9:45AM