A framework for adaptive MCMC targeting multimodal distributions

DOI10.1214/19-AOS1916MaRDI QIDQ2215766zbMATH OpenOpenAlexFDO

Authors Emilia Pompe, Chris Holmes, Krzysztof Łatuszyński

Publication date 14 December 2020

Published in The Annals of Statistics (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1812.02609, https://projecteuclid.org/euclid.aos/1600480937

zbMATH Keywords

ergodicityadaptive MCMCmultimodal distribution

Mathematics Subject Classification ID

Computational methods in Markov chains (60J22) Computational methods for problems pertaining to statistics (62-08) Bayesian inference (62F15) Monte Carlo methods (65C05) Numerical analysis or methods applied to Markov chains (65C40)

Abstract: We propose a new Monte Carlo method for sampling from multimodal distributions. The idea of this technique is based on splitting the task into two: finding the modes of a target distribution $pi$ and sampling, given the knowledge of the locations of the modes. The sampling algorithm relies on steps of two types: local ones, preserving the mode; and jumps to regions associated with different modes. Besides, the method learns the optimal parameters of the algorithm while it runs, without requiring user intervention. Our technique should be considered as a flexible framework, in which the design of moves can follow various strategies known from the broad MCMC literature. In order to design an adaptive scheme that facilitates both local and jump moves, we introduce an auxiliary variable representing each mode and we define a new target distribution $ildepi$ on an augmented state space $mathcalXimesmathcalI$, where $mathcalX$ is the original state space of $pi$ and $mathcalI$ is the set of the modes. As the algorithm runs and updates its parameters, the target distribution $ildepi$ also keeps being modified. This motivates a new class of algorithms, Auxiliary Variable Adaptive MCMC. We prove general ergodic results for the whole class before specialising to the case of our algorithm.

Recommendations

• Learn from thy neighbor: parallel-chain and regional adaptive MCMC
• Mode jumping proposals in MCMC
• On the efficiency of adaptive MCMC algorithms
• Adaptive Markov chain Monte Carlo for auxiliary variable method and its application to parallel tempering
• On the Use of Local Optimizations within Metropolis–Hastings Updates

Cites work

• scientific article; zbMATH DE number 5060482 (Why is no real title available?)
• A Repelling–Attracting Metropolis Algorithm for Multimodality
• An adaptive Metropolis algorithm
• Bayesian computation: a summary of the current state, and samples backwards and forwards
• Convergence of adaptive and interacting Markov chain Monte Carlo algorithms
• Convergence properties of pseudo-marginal Markov chain Monte Carlo algorithms
• Coupling and Ergodicity of Adaptive Markov Chain Monte Carlo Algorithms
• Equi-energy sampler with applications in statistical inference and statistical mechanics
• Error bounds for sequential Monte Carlo samplers for multimodal distributions
• Generalized darting Monte Carlo
• Geometric ergodicity of Metropolis algorithms
• Learn from thy neighbor: parallel-chain and regional adaptive MCMC
• Limit theorems for some adaptive MCMC algorithms with subgeometric kernels
• Markov chain Monte Carlo methods and the label switching problem in Bayesian mixture modeling
• Mode jumping proposals in MCMC
• Multi-domain sampling with applications to structural inference of Bayesian networks
• Multimodality of the likelihood in the bivariate seemingly unrelated regressions model
• On the containment condition for adaptive Markov chain Monte Carlo algorithms
• On the ergodicity properties of some adaptive MCMC algorithms
• On the stability and ergodicity of adaptive scaling Metropolis algorithms
• Optimal proposal distributions and adaptive MCMC
• Optimal scaling for various Metropolis-Hastings algorithms.
• Regeneration in Markov Chain Samplers
• Sufficient conditions for torpid mixing of parallel and simulated tempering
• The pseudo-marginal approach for efficient Monte Carlo computations
• The spectrum of the independent Metropolis-Hastings algorithm
• Towards optimal scaling of Metropolis-coupled Markov chain Monte Carlo
• Weak convergence and optimal scaling of random walk Metropolis algorithms
• Weight-preserving simulated tempering

Cited in

(13)

• Cauchy Markov random field priors for Bayesian inversion
• Polya tree Monte Carlo method
• Metropolized randomized maximum likelihood for improved sampling from multimodal distributions
• Global Likelihood Sampler for Multimodal Distributions
• Sampling using adaptive regenerative processes
• Adaptive schemes for piecewise deterministic Monte Carlo algorithms
• A Metropolis-class sampler for targets with non-convex support
• Hopping between distant basins
• Adaptive random neighbourhood informed Markov chain Monte Carlo for high-dimensional Bayesian variable selection
• Adaptive Markov chain Monte Carlo for auxiliary variable method and its application to parallel tempering
• Latent uniform samplers on multivariate binary spaces
• Emerging directions in Bayesian computation
• On the flexibility of Metropolis-Hastings acceptance probabilities in auxiliary variable proposal generation

Describes a project that uses

Uses Software

• MultiNest

This page was built for publication: A framework for adaptive MCMC targeting multimodal distributions

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2215766)