A0209
Title: Parallelized midpoint randomization for Langevin Monte Carlo
Authors: Lu Yu - City University of Hong Kong (Hong Kong) [presenting]
Abstract: The purpose is to study the problem of sampling from a target probability density function in frameworks where parallel evaluations of the log-density gradient are feasible. Focusing on smooth and strongly log-concave densities, the parallelized randomized midpoint method is revisited, and its properties are investigated using recently developed techniques for analyzing its sequential version. Through these techniques, upper bounds are derived on the Wasserstein distance between sampling and target densities. These bounds quantify the substantial runtime improvements achieved through parallel processing.
