Jose Blanchet

Title: Optimal Transport Distributionally Robust Optimization

Abstract: In this talk, we discuss recent advances in modeling, statistics, and computational methods for optimal transport based distributionally robust optimization. We discuss, for example, motivating formulations which include distributionally pricing and hedging with martingale constraints, portfolio optimization and Bayesian non-parametric inverse problems. In these settings, we discuss optimal sample complexity approximations and iterative optimization algorithms for these types of problems.