Pseudo-marginal Metropolis--Hastings using averages of unbiased estimators

TL;DR

This paper analyzes the efficiency of pseudo-marginal Metropolis-Hastings algorithms that average multiple unbiased estimators, showing that using fewer estimators can often be more effective due to variance considerations.

Contribution

The paper provides a theoretical comparison of asymptotic variances for different averaging schemes in pseudo-marginal MCMC, disproving a conjecture and offering practical insights.

Findings

Averaging more estimators does not always reduce variance.

The bound on asymptotic variance is tight and can be achieved.

Using a single estimator (m=1) can be optimal when computational cost is proportional to m.

Abstract

We consider a pseudo-marginal Metropolis--Hastings kernel $P_m$ that is constructed using an average of $m$ exchangeable random variables, as well as an analogous kernel $P_s$ that averages $s<m$ of these same random variables. Using an embedding technique to facilitate comparisons, we show that the asymptotic variances of ergodic averages associated with $P_m$ are lower bounded in terms of those associated with $P_s$. We show that the bound provided is tight and disprove a conjecture that when the random variables to be averaged are independent, the asymptotic variance under $P_m$ is never less than $s/m$ times the variance under $P_s$. The conjecture does, however, hold when considering continuous-time Markov chains. These results imply that if the computational cost of the algorithm is proportional to $m$, it is often better to set $m=1$. We provide intuition as to why these findings differ so markedly from recent results for pseudo-marginal kernels employing particle filter approximations. Our results are exemplified through two simulation studies; in the first the computational cost is effectively proportional to $m$ and in the second there is a considerable start-up cost at each iteration.