I discuss the family of methods known under the names of BeliefPropagation (BP) in artificial intelligence, Message Passing(MP) in computer science/information theory, and Bethe-Peierls(still BP) approximation in statistical physics. These BP/MPmethods offer efficient and distributed algorithmic solutions,along with rich theoretical guarantees, for variety of problemsin statistical inference and optimization. The power of BP/MPwill be illustrated on examples from error-correction, particletracking in fluid mechanics and switching control over powergrids.
Bio:Dr. Chertkov's areas of interest include statistical and mathematicalphysics applied to information theory, computer science, hydrodynamics,optics, communication and infrastructure networks. Dr. Chertkov receivedhis Ph.D. in physics from the Weizmann Institute of Science in 1996,and his M.Sc. in physics from Novosibirsk State University in 1990. Afterhis Ph.D., Dr. Chertkov spent three years at Princeton University as aR.H. Dicke Fellow in the Department of Physics. He joined Los AlamosNational Lab in 1999, initially as a J.R. Oppenheimer Fellow in theTheoretical Division. He is now a staff member in the same division. Dr.Chertkov has published 90 papers in these research areas and is currentlyleading "Physics of Algorithms" and "Optimization and Control Theory forSmart Grids" projects at LANL.