Machine Learning Seminar: Belief Propogation Algorithms
Wednesday, March 30, 2016
3:30 pm - 5:00 pm
Gross Hall 330
Christian Borgs - Microsoft Research
In this talk, I describe how the theory of graph limits and graphons developed in the last ten years can help to address these questions, reviewing both the classical theory for dense graphs and newer results for sparse graphs including power law graphs. Time permitting, I will also discuss how consistent estimation can be done in a way which respects privacy, in the sense of differential privacy, i.e., so that the removal or addition of a vertex has only a small influence on the output.Bio: Christian Borgs is the deputy managing director of Microsoft Research New England in Cambridge, Massachusetts. He studied physics at the University of Munich, the University Pierre et Marie Curie in Paris, the Institut des Hautes Etudes in Bures-sur-Yvettes, and the Max-Planck-Institute for Physics in Munich. He received his Ph.D. in mathematical physics from the University of Munich, held a postdoctoral fellowship at the ETH Zurich, and received his Habilitation in mathematical physics from the Free University in Berlin. After his Habilitation he became the C4 Chair for Statistical Mechanics at the University of Leipzig, and in 1997 he joined Microsoft Research to co-found the Theory Group. He was a manager of the Theory group until 2008, when he co-founded Microsoft Research New England.