Optimal friction matrix for underdamped Langevin sampling

Abstract

A systematic procedure for optimising the friction coefficient in underdamped Langevin dynamics as a sampling tool is given by taking the gradient of the associated asymptotic variance with respect to friction. We give an expression for this gradient in terms of the solution to an appropriate Poisson equation and show that it can be approximated by short simulations of the associated first variation/tangent process under concavity assumptions on the log density. Our algorithm is applied to the estimation of posterior means in Bayesian inference problems and reduced variance is demonstrated when compared to the original underdamped and overdamped Langevin dynamics in both full and stochastic gradient cases.

44 pages, 6 figures

References in corpus (1)

• Accurate sampling using Langevin dynamics

Cited by in corpus (4)

• Geometric Methods for Sampling, Optimisation, Inference and Adaptive Agents
• High order splitting methods for SDEs satisfying a commutativity condition
• Regularity preservation in Kolmogorov equations for non-Lipschitz coefficients under Lyapunov conditions
• Faster Molecular Dynamics with Neural Network Potentials via Distilled Multiple Time-Stepping and Non-Conservative Forces