MCMC with strings and branes: the suburban algorithm (extended version)

DOI10.1142/S0217751X17501330MaRDI QIDQ5367510zbMATHFDO

Authors Jonathan J. Heckman, Jeffrey G. Bernstein, Ben Vigoda

Publication date 13 October 2017

Published in International Journal of Modern Physics A (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1605.05334

zbMATH Keywords

Markov chain Monte Carlostatistical inferencestring theorybranes

Mathematics Subject Classification ID

Monte Carlo methods (65C05) Numerical analysis or methods applied to Markov chains (65C40) String and superstring theories; other extended objects (e.g., branes) in quantum field theory (81T30)

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.

Recommendations

• Markov chain Monte Carlo algorithms allowing parallel processing. II
• MCMC using Hamiltonian dynamics
• Markov Chain Monte Carlo Algorithms: Theory and Practice
• Multipoint Metropolis method with application to hybrid Monte Carlo
• scientific article; zbMATH DE number 1246228

Cites work

• A First Course in String Theory
• Conformal field theory
• Ensemble samplers with affine invariance
• Equation of state calculations by fast computing machines
• Equi-energy sampler with applications in statistical inference and statistical mechanics
• MCMC with strings and branes: the suburban algorithm (extended version)
• Monte Carlo sampling methods using Markov chains and their applications
• Slice sampling. (With discussions and rejoinder)
• Statistical approach to quantum field theory. An introduction
• Weak convergence and optimal scaling of random walk Metropolis algorithms

Cited in

(1)

• MCMC with strings and branes: the suburban algorithm (extended version)

Describes a project that uses

Uses Software

• Dimple

This page was built for publication: MCMC with strings and branes: the suburban algorithm (extended version)

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5367510)