Average of Recentered Parallel MCMC for Big Data

TL;DR

This paper introduces a novel parallel MCMC method for big data that operates on recentered and rescaled subposteriors, improving scalability and performance over existing divide-and-conquer approaches.

Contribution

It proposes a new average of recentered parallel MCMC method within the divide-and-conquer framework, with mathematical justification and applicability to non-parametric models.

Findings

Method performs well on several models.

Shares computation cost with existing methods.

Applicable to non-parametric cases.

Abstract

In big data context, traditional MCMC methods, such as Metropolis-Hastings algorithms and hybrid Monte Carlo, scale poorly because of their need to evaluate the likelihood over the whole data set at each iteration. In order to resurrect MCMC methods, numerous approaches belonging to two categories: divide-and-conquer and subsampling, are proposed. In this article, we study the parallel MCMC and propose a new combination method in the divide-and-conquer framework. Compared with some parallel MCMC methods, such as consensus Monte Carlo, Weierstrass Sampler, instead of sampling from subposteriors, our method runs MCMC on rescaled subposteriors, but share the same computation cost in the parallel stage. We also give the mathematical justification of our method and show its performance in several models. Besides, even though our new methods is proposed in parametric framework, it can been applied to non-parametric cases without difficulty.