Parallel Chromatic MCMC with Spatial Partitioning

TL;DR

This paper presents a novel parallel MCMC method leveraging spatial partitioning and graph coloring to efficiently perform inference in models with spatially dependent structures, exemplified by seismic event detection.

Contribution

It introduces a new parallelization technique for MCMC based on spatial graph coloring, applicable to models where traditional graphical model methods fail.

Findings

Achieves significant speedups over serial MCMC

Maintains inference quality despite parallelization

Applicable to models with spatially dependent conditional independence

Abstract

We introduce a novel approach for parallelizing MCMC inference in models with spatially determined conditional independence relationships, for which existing techniques exploiting graphical model structure are not applicable. Our approach is motivated by a model of seismic events and signals, where events detected in distant regions are approximately independent given those in intermediate regions. We perform parallel inference by coloring a factor graph defined over regions of latent space, rather than individual model variables. Evaluating on a model of seismic event detection, we achieve significant speedups over serial MCMC with no degradation in inference quality.