Probability Seminar: Soumik Pal

  • Date: 07/22/2019
  • Time: 14:00
Soumik Pal, University of Washington

University of British Columbia


On Schrodinger bridges, entropic cost and their limits


Consider the Monge-Kantorovich problem of transporting densities $\rho_0$ to $\rho_1$ on $\rr^d$ with a strictly convex cost function. A popular relaxation of the problem is the one-parameter family called the entropic cost problem. The entropic cost J_h, h>0, is significantly faster to compute and $h J_h$ is known to converge to the optimal transport cost as $h$ goes to zero. We will give an overview of various ideas in this field, including discrete approximations, gamma convergence and particle systems. Finally we will discuss Gaussian approximations to Schrodinger bridges as $h$ approaches zero. As a consequence we obtain ``gradient flows’’ of entropy even in cases where the cost function is not a metric.

Other Information: 

Location: ESB 4133 (PIMS lounge)