Convergence properties of pseudo-marginal Markov chain Monte Carlo algorithms

Abstract

We study convergence properties of pseudo-marginal Markov chain Monte Carlo algorithms (Andrieu and Roberts [Ann. Statist. 37 (2009) 697-725]). We find that the asymptotic variance of the pseudo-marginal algorithm is always at least as large as that of the marginal algorithm. We show that if the marginal chain admits a (right) spectral gap and the weights (normalised estimates of the target density) are uniformly bounded, then the pseudo-marginal chain has a spectral gap. In many cases, a similar result holds for the absolute spectral gap, which is equivalent to geometric ergodicity. We consider also unbounded weight distributions and recover polynomial convergence rates in more specific cases, when the marginal algorithm is uniformly ergodic or an independent Metropolis-Hastings or a random-walk Metropolis targeting a super-exponential density with regular contours. Our results on geometric and polynomial convergence rates imply central limit theorems. We also prove that under general conditions, the asymptotic variance of the pseudo-marginal algorithm converges to the asymptotic variance of the marginal algorithm if the accuracy of the estimators is increased.

Published at http://dx.doi.org/10.1214/14-AAP1022 in the Annals of Applied Probability (http://www.imstat.org/aap/) by the Institute of Mathematical Statistics (http://www.imstat.org)

References in corpus (6)

• The pseudo-marginal approach for efficient Monte Carlo computations
• Renewal theory and computable convergence rates for geometrically ergodic Markov chains
• Quantitative bounds on convergence of time-inhomogeneous Markov chains
• Variance bounding and geometric ergodicity of Markov chain Monte Carlo kernels for approximate Bayesian computation
• Variance bounding Markov chains
• Coupled MCMC with a randomized acceptance probability

Cited by in corpus (17)

• Particle Filters and Data Assimilation
• On particle Gibbs sampling
• Probabilistic learning of nonlinear dynamical systems using sequential Monte Carlo
• Comparison of asymptotic variances of inhomogeneous Markov chains with application to Markov chain Monte Carlo methods
• Importance sampling type estimators based on approximate marginal MCMC
• Twisted particle filters
• An Introduction to Twisted Particle Filters and Parameter Estimation in Non-linear State-space Models
• Conditional convex orders and measurable martingale couplings
• Sequential Monte Carlo as Approximate Sampling: bounds, adaptive resampling via $\infty$-ESS, and an application to Particle Gibbs
• Data-Free Likelihood-Informed Dimension Reduction of Bayesian Inverse Problems
• Prior normalization for certified likelihood-informed subspace detection of Bayesian inverse problems
• Comparison of Markov chains via weak Poincaré inequalities with application to pseudo-marginal MCMC
• Importance sampling correction versus standard averages of reversible MCMCs in terms of the asymptotic variance
• A survey of Monte Carlo methods for noisy and costly densities with application to reinforcement learning and ABC
• The Importance Markov Chain
• Convergence Speed and Approximation Accuracy of Numerical MCMC
• Bring the noise: exact inference from noisy simulations in collider physics