Harvard Theory of Computation Seminar

Graph Covers and Quadratic Minimization

Justin Thaler, Harvard



Wednesday, Sept 23, 2009 at 12 noon.  Room TBA.

ABSTRACT

Belief propagation (BP) is a widely-used message-passing algorithm that approximately computes marginal distributions in graphical models. A wide variety of computational problems can be formulated as inference problems in such models, and BP and its variants such as the min-sum algorithm perform empirically well in application areas including coding theory, statistical physics, and linear programming. However, rigorously characterizing their behavior outside of a few well-structured instances has proven challenging.

 

In this talk, I will present a new approach to understanding the behavior of the min-sum algorithm by exploiting the properties of graph covers.  I will first motivate this approach by demonstrating that the standard sufficient conditions for convergence of min-sum in the case of quadratic minimization have a natural analog in terms of graph covers.  I will then explain how to use graph covers to understand the frequently-observed periodic behavior of the min-sum algorithm in some important special cases. Finally, I will show how to use these results to recover useful estimates from the min-sum algorithm even in the presence of periodic behavior. Some of our techniques apply more broadly, and we believe that by capturing the notion of indistinguishability, graph covers represent a valuable tool for understanding the abilities and limitations of general message-passing algorithms.

 

Joint work with Nicholas Ruozzi and Sekhar Tatikonda of Yale University.