A cautionary tale on the efficiency of some adaptive Monte Carlo schemes

DOI10.1214/09-AAP636MaRDI QIDQ988756zbMATH OpenOpenAlexFDO

Authors Yves F. Atchadé

Publication date 18 August 2010

Published in The Annals of Applied Probability (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/0901.1378

zbMATH Keywords

central limit theoremadaptive Markov chain Monte Carloequi-energy samplerlaw of large numbersimportance resamplingmartingale approximation method

Mathematics Subject Classification ID

Computational methods in Markov chains (60J22) Numerical analysis or methods applied to Markov chains (65C40) Randomized algorithms (68W20) Discrete-time Markov processes on general state spaces (60J05)

Abstract: There is a growing interest in the literature for adaptive Markov chain Monte Carlo methods based on sequences of random transition kernels $P_n$ where the kernel $P_n$ is allowed to have an invariant distribution $p{i}_n$ not necessarily equal to the distribution of interest $pi$ (target distribution). These algorithms are designed such that as $noinfty$, $P_n$ converges to $P$, a kernel that has the correct invariant distribution $pi$. Typically, $P$ is a kernel with good convergence properties, but one that cannot be directly implemented. It is then expected that the algorithm will inherit the good convergence properties of $P$. The equi-energy sampler of [Ann. Statist. 34 (2006) 1581--1619] is an example of this type of adaptive MCMC. We show in this paper that the asymptotic variance of this type of adaptive MCMC is always at least as large as the asymptotic variance of the Markov chain with transition kernel $P$. We also show by simulation that the difference can be substantial.

Recommendations

• A framework for adaptive Monte Carlo procedures
• Adaptive schemes for piecewise deterministic Monte Carlo algorithms
• On the efficiency of adaptive MCMC algorithms
• Adaptive schemes for the Monte Carlo method of an enhanced accuracy
• Adaptive optimization of the Monte-Carlo method
• On the convergence of adaptive sequential Monte Carlo methods
• On adaptive Markov chain Monte Carlo algorithms
• On the convergence rates of some adaptive Markov chain Monte Carlo algorithms
• Adaptive adjoint Monte Carlo simulation for the uncertainty
• On adaptive resampling strategies for sequential Monte Carlo methods

Cites work

• scientific article; zbMATH DE number 1713116 (Why is no real title available?)
• scientific article; zbMATH DE number 46153 (Why is no real title available?)
• Adaptive Markov Chain Monte Carlo through Regeneration
• An adaptive Metropolis algorithm
• An adaptive version for the Metropolis adjusted Langevin algorithm with a truncated drift
• Coupling and Ergodicity of Adaptive Markov Chain Monte Carlo Algorithms
• Equi-energy sampler with applications in statistical inference and statistical mechanics
• Fluctuations of interacting Markov chain Monte Carlo methods
• Geometric ergodicity of Metropolis algorithms
• Interacting Markov chain Monte Carlo methods for solving nonlinear measure-valued equations
• Limit theorems for some adaptive MCMC algorithms with subgeometric kernels
• Markov Processes, Gaussian Processes, and Local Times
• Markov chains and stochastic stability
• Non-linear Markov Chain Monte Carlo
• On a Strong Law of Large Numbers for Martingales
• On the efficiency of adaptive MCMC algorithms
• On the ergodicity properties of some adaptive MCMC algorithms
• Renewal theory and computable convergence rates for geometrically erdgodic Markov chains
• Resampling from the past to improve on MCMC algorithms

Cited in

(12)

• A central limit theorem for adaptive and interacting Markov chains
• On the convergence rates of some adaptive Markov chain Monte Carlo algorithms
• Adaptive equi-energy sampler: convergence and illustration
• Efficiency bounds for semiparametric models with singular score functions
• Fluctuations of interacting Markov chain Monte Carlo methods
• Kernel estimators of asymptotic variance for adaptive Markov chain Monte Carlo
• Convergence of adaptive and interacting Markov chain Monte Carlo algorithms
• A Note on Convergence of the Equi-Energy Sampler
• On nonlinear Markov chain Monte Carlo
• Vegas revisited: Adaptive Monte Carlo integration beyond factorization
• The containment condition and AdapFail algorithms
• Parallel tempering with equi-energy moves

This page was built for publication: A cautionary tale on the efficiency of some adaptive Monte Carlo schemes

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q988756)