|Date: Tuesday, September 15, 2015
Title: Discovering hidden structures in large networks
Abstract: Many real-world networks -- social, technological, biological -- have wonderful structures. Some of the structures are apparent (e.g. trees); other structures are well hidden (for example, clusters or communities). When and how can hidden structures be discovered? Known approaches to "structure mining" in networks come from a variety of disciplines, including probability, combinatorics, physics, statistics, optimization, signal processing and information theory. We will focus on probabilistic approaches to structure mining, where a network is regarded as a random graph. This brings together insights from random matrices, random graphs, and semidefinite programming.
Speaker: Roman Vershynin
Institution: University of Michigan