Local-Global MCMC kernels: the best of both worlds
Preprint
Abstract
Recent works leveraging learning to enhance sampling have shown promising results, in particular by designing effective non-local moves and global proposals. However, learning accuracy is inevitably limited in regions where little data is available such as in the tails of distributions as well as in high-dimensional problems. In the present paper we study an Explore-Exploit Markov chain Monte Carlo strategy ($Ex^2MCMC$) that combines local and global samplers showing that it enjoys the advantages of both approaches. We prove $V$-uniform geometric ergodicity of $Ex^2MCMC$ without requiring a uniform adaptation of the global sampler to the target distribution. We also compute explicit bounds on the mixing rate of the Explore-Exploit strategy under realistic conditions. Moreover, we also analyze an adaptive version of the strategy ($FlEx^2MCMC$) where a normalizing flow is trained while sampling to serve as a proposal for global moves. We illustrate the efficiency of $Ex^2MCMC$ and its adaptive version on classical sampling benchmarks as well as in sampling high-dimensional distributions defined by Generative Adversarial Networks seen as Energy Based Models. We provide the code to reproduce the experiments at the link: https://github.com/svsamsonov/ex2mcmc_new.
Keywords
All Related Versions
This publication has 41 references indexed in Scilit:
- Accurate Sampling of Macromolecular Conformations Using Adaptive Deep Learning and Coarse-Grained RepresentationJournal of Chemical Information and Modeling, 2022
- Adaptive Monte Carlo augmented with normalizing flowsProceedings of the National Academy of Sciences, 2022
- Reflection couplings and contraction rates for diffusionsProbability Theory and Related Fields, 2015
- Importance sampling: a reviewWIREs Computational Statistics, 2009
- A tutorial on adaptive MCMCStatistics and Computing, 2008
- Adaptive proposal distribution for random walk Metropolis algorithmComputational Statistics, 1999
- Théorèmes de convergence presque sure pour une classe d'algorithmes stochastiques à pas décroissantProbability Theory and Related Fields, 1987
- Correlation functions and computer simulationsNuclear Physics B, 1981
- Brownian dynamics as smart Monte Carlo simulationThe Journal of Chemical Physics, 1978
- A computer simulation of charged particles in solution. I. Technique and equilibrium propertiesThe Journal of Chemical Physics, 1975