Spring 2015

Strong Data Processing Inequalities: Applications to MCMC and Graphical Models

Wednesday, Mar. 18, 2015 1:55 pm2:20 pm PDT

Add to Calendar


Calvin Lab Auditorium

Strong (or quantitative) data processing inequalities provide sharp estimates of the rate at which a noisy channel “destroys” information. In this talk, I will present some recent results on strong data processing inequalities in discrete settings, with a focus on their use for quantifying the mixing behavior of Markov Chain Monte Carlo (MCMC) algorithms and the decay of correlations in graphical models.