Peter Erdos : Sampling bipartite degree sequences - the Markov chain approach

Posted by Gyorgy Turan , part of the Mathematical Computer Science Seminar.

At

Oct. 8, 2018, 3 p.m.

In

427 SEO

Abstract

How to analyze real life networks? There are myriads of them and usually experiments cannot be performed directly. Instead, scientists define models, fix parameters and imagine the dynamics of evolution. Then, they build synthetic networks on this basis (one, several, all) and they want to sample them. However, there are far too many such networks. Therefore, typically, some probabilistic method is used for sampling. We will survey one such approach, the Markov Chain Monte Carlo method, to sample realizations of given degree sequences. Some new results will be discussed.