Enregistré dans:
Détails bibliographiques
Auteurs principaux: Jin, Kexin, Liu, Chenguang, Latz, Jonas
Format: Preprint
Publié: 2023
Sujets:
Accès en ligne:https://arxiv.org/abs/2305.13882
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866911855661809664
author Jin, Kexin
Liu, Chenguang
Latz, Jonas
author_facet Jin, Kexin
Liu, Chenguang
Latz, Jonas
contents The Stochastic Gradient Langevin Dynamics (SGLD) are popularly used to approximate Bayesian posterior distributions in statistical learning procedures with large-scale data. As opposed to many usual Markov chain Monte Carlo (MCMC) algorithms, SGLD is not stationary with respect to the posterior distribution; two sources of error appear: The first error is introduced by an Euler--Maruyama discretisation of a Langevin diffusion process, the second error comes from the data subsampling that enables its use in large-scale data settings. In this work, we consider an idealised version of SGLD to analyse the method's pure subsampling error that we then see as a best-case error for diffusion-based subsampling MCMC methods. Indeed, we introduce and study the Stochastic Gradient Langevin Diffusion (SGLDiff), a continuous-time Markov process that follows the Langevin diffusion corresponding to a data subset and switches this data subset after exponential waiting times. There, we show the exponential ergodicity of SLGDiff and that the Wasserstein distance between the posterior and the limiting distribution of SGLDiff is bounded above by a fractional power of the mean waiting time. We bring our results into context with other analyses of SGLD.
format Preprint
id arxiv_https___arxiv_org_abs_2305_13882
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Subsampling Error in Stochastic Gradient Langevin Diffusions
Jin, Kexin
Liu, Chenguang
Latz, Jonas
Machine Learning
Computation
65C05, 62F15
The Stochastic Gradient Langevin Dynamics (SGLD) are popularly used to approximate Bayesian posterior distributions in statistical learning procedures with large-scale data. As opposed to many usual Markov chain Monte Carlo (MCMC) algorithms, SGLD is not stationary with respect to the posterior distribution; two sources of error appear: The first error is introduced by an Euler--Maruyama discretisation of a Langevin diffusion process, the second error comes from the data subsampling that enables its use in large-scale data settings. In this work, we consider an idealised version of SGLD to analyse the method's pure subsampling error that we then see as a best-case error for diffusion-based subsampling MCMC methods. Indeed, we introduce and study the Stochastic Gradient Langevin Diffusion (SGLDiff), a continuous-time Markov process that follows the Langevin diffusion corresponding to a data subset and switches this data subset after exponential waiting times. There, we show the exponential ergodicity of SLGDiff and that the Wasserstein distance between the posterior and the limiting distribution of SGLDiff is bounded above by a fractional power of the mean waiting time. We bring our results into context with other analyses of SGLD.
title Subsampling Error in Stochastic Gradient Langevin Diffusions
topic Machine Learning
Computation
65C05, 62F15
url https://arxiv.org/abs/2305.13882