Improved Discretization Analysis for Underdamped Langevin Monte Carlo

TL;DR

This paper provides a new analysis of Underdamped Langevin Monte Carlo (ULMC), improving sampling guarantees under weaker conditions and translating continuous-time acceleration results into practical algorithms with better condition number dependence.

Contribution

It relaxes assumptions for ULMC analysis by removing Hessian smoothness, achieves state-of-the-art dimension dependence, and extends continuous-time acceleration results to discrete algorithms.

Findings

Achieves state-of-the-art dimension dependence in sampling guarantees.

Provides the first KL divergence guarantees for ULMC without Hessian smoothness.

Translates continuous-time acceleration results into improved discrete-time algorithms.

Abstract

Underdamped Langevin Monte Carlo (ULMC) is an algorithm used to sample from unnormalized densities by leveraging the momentum of a particle moving in a potential well. We provide a novel analysis of ULMC, motivated by two central questions: (1) Can we obtain improved sampling guarantees beyond strong log-concavity? (2) Can we achieve acceleration for sampling? For (1), prior results for ULMC only hold under a log-Sobolev inequality together with a restrictive Hessian smoothness condition. Here, we relax these assumptions by removing the Hessian smoothness condition and by considering distributions satisfying a Poincar\'e inequality. Our analysis achieves the state of art dimension dependence, and is also flexible enough to handle weakly smooth potentials. As a byproduct, we also obtain the first KL divergence guarantees for ULMC without Hessian smoothness under strong log-concavity, which is based on a new result on the log-Sobolev constant along the underdamped Langevin diffusion. For (2), the recent breakthrough of Cao, Lu, and Wang (2020) established the first accelerated result for sampling in continuous time via PDE methods. Our discretization analysis translates their result into an algorithmic guarantee, which indeed enjoys better condition number dependence than prior works on ULMC, although we leave open the question of full acceleration in discrete time. Both (1) and (2) necessitate R\'enyi discretization bounds, which are more challenging than the typically used Wasserstein coupling arguments. We address this using a flexible discretization analysis based on Girsanov's theorem that easily extends to more general settings.