MCMC with Strings and Branes: The Suburban Algorithm (Extended Version)

Abstract

Motivated by the physics of strings and branes, we develop a class of Markov chain Monte Carlo (MCMC) algorithms involving extended objects. Starting from a collection of parallel Metropolis-Hastings (MH) samplers, we place them on an auxiliary grid, and couple them together via nearest neighbor interactions. This leads to a class of "suburban samplers" (i.e., spread out Metropolis). Coupling the samplers in this way modifies the mixing rate and speed of convergence for the Markov chain, and can in many cases allow a sampler to more easily overcome free energy barriers in a target distribution. We test these general theoretical considerations by performing several numerical experiments. For suburban samplers with a fluctuating grid topology, performance is strongly correlated with the average number of neighbors. Increasing the average number of neighbors above zero initially leads to an increase in performance, though there is a critical connectivity with effective dimension d_eff ~ 1, above which "groupthink" takes over, and the performance of the sampler declines.

v2: 55 pages, 13 figures, references and clarifications added. Published version. This article is an extended version of "MCMC with Strings and Branes: The Suburban Algorithm"

References in corpus (6)

• emcee: The MCMC Hammer
• Parallel Tempering: Theory, Applications, and New Perspectives
• Equi-energy sampler with applications in statistical inference and statistical mechanics
• MCMC Using Ensembles of States for Problems with Fast and Slow Variables such as Gaussian Process Regression
• Accelerating Inference: towards a full Language, Compiler and Hardware stack
• MCMC with Strings and Branes: The Suburban Algorithm

Cited by in corpus (3)

• Disorder Averaging and its UV (Dis)Contents
• Misanthropic Entropy and Renormalization as a Communication Channel
• MCMC with Strings and Branes: The Suburban Algorithm