Analysis of Langevin Monte Carlo from Poincaré to Log-Sobolev

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Chewi, Sinho, Erdogdu, Murat A., Li, Mufan Bill, Shen, Ruoqi, Zhang, Matthew
Format: Preprint
Publié: 2021
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866909249788968960
author Chewi, Sinho
Erdogdu, Murat A.
Li, Mufan Bill
Shen, Ruoqi
Zhang, Matthew
author_facet Chewi, Sinho
Erdogdu, Murat A.
Li, Mufan Bill
Shen, Ruoqi
Zhang, Matthew
contents Classically, the continuous-time Langevin diffusion converges exponentially fast to its stationary distribution $π$ under the sole assumption that $π$ satisfies a Poincaré inequality. Using this fact to provide guarantees for the discrete-time Langevin Monte Carlo (LMC) algorithm, however, is considerably more challenging due to the need for working with chi-squared or Rényi divergences, and prior works have largely focused on strongly log-concave targets. In this work, we provide the first convergence guarantees for LMC assuming that $π$ satisfies either a Latała--Oleszkiewicz or modified log-Sobolev inequality, which interpolates between the Poincaré and log-Sobolev settings. Unlike prior works, our results allow for weak smoothness and do not require convexity or dissipativity conditions.
format Preprint
id arxiv_https___arxiv_org_abs_2112_12662
institution arXiv
publishDate 2021
record_format arxiv
spellingShingle Analysis of Langevin Monte Carlo from Poincaré to Log-Sobolev
Chewi, Sinho
Erdogdu, Murat A.
Li, Mufan Bill
Shen, Ruoqi
Zhang, Matthew
Statistics Theory
Machine Learning
Classically, the continuous-time Langevin diffusion converges exponentially fast to its stationary distribution $π$ under the sole assumption that $π$ satisfies a Poincaré inequality. Using this fact to provide guarantees for the discrete-time Langevin Monte Carlo (LMC) algorithm, however, is considerably more challenging due to the need for working with chi-squared or Rényi divergences, and prior works have largely focused on strongly log-concave targets. In this work, we provide the first convergence guarantees for LMC assuming that $π$ satisfies either a Latała--Oleszkiewicz or modified log-Sobolev inequality, which interpolates between the Poincaré and log-Sobolev settings. Unlike prior works, our results allow for weak smoothness and do not require convexity or dissipativity conditions.
title Analysis of Langevin Monte Carlo from Poincaré to Log-Sobolev
topic Statistics Theory
Machine Learning
url https://arxiv.org/abs/2112.12662