Langevin Monte Carlo Beyond Lipschitz Gradient Continuity

TL;DR

This paper introduces IPLA, an advanced Langevin Monte Carlo method that handles a broader class of potentials with improved convergence and moment bounds, surpassing traditional smoothness limitations.

Contribution

The paper presents IPLA, extending LMC to non-smooth, super-quadratic potentials with better convergence rates and moment bounds, broadening its practical applicability.

Findings

IPLA effectively handles non-L-smooth potentials.

Improved convergence rates over existing LMC algorithms.

Provides comprehensive bounds on Markov chain moments.

Abstract

We present a significant advancement in the field of Langevin Monte Carlo (LMC) methods by introducing the Inexact Proximal Langevin Algorithm (IPLA). This novel algorithm broadens the scope of problems that LMC can effectively address while maintaining controlled computational costs. IPLA extends LMC's applicability to potentials that are convex, strongly convex in the tails, and exhibit polynomial growth, beyond the conventional $L$-smoothness assumption. Moreover, we extend LMC's applicability to super-quadratic potentials and offer improved convergence rates over existing algorithms. Additionally, we provide bounds on all moments of the Markov chain generated by IPLA, enhancing its analytical robustness.