Mean-field fun and exponential random graphs

Mean-field fun and exponential random graphs
June, 05, 2018
11:30
Room 861 Electrical Eng. Building Technion City

Probability and Stochastic Processes Seminar

Speaker: Renan Gross

Affiliation: Weizmann Institute of Science

Gibbs distributions are everywhere, whether you’re Ising-modeling a magnet or fitting an exponential random graph to your social network. Alas, explicitly solving Gibbs distributions can sometimes be a bit hard. In this talk, we’ll discuss a mean-field condition on the Hamiltonian and describe how this condition decomposes the distribution into a mixture of product measures. We’ll then apply this decomposition to exponential random graphs to say some (hopefully) interesting things about them, such as their block structure and closeness to G(n,p). Joint work with Ronen Eldan. For a friendly introduction to decomposition of Gibbs measures, see here: https://sarcasticresonance.wordpress.com/2017/09/09/new-paper-on-arxiv-decomposition-of-mean-field-gibbs-distributions-into-product-measures/ For a friendly introduction to exponential random graphs, see here: https://sarcasticresonance.wordpress.com/2017/07/14/random-graphs-the-exponential-random-graph-model/