The Correlated Pseudo-Marginal Method

TL;DR

The paper introduces the correlated pseudo-marginal method, improving sampling efficiency in Bayesian inference by reducing variance in likelihood estimations, especially beneficial for large datasets, leading to significant computational gains.

Contribution

It proposes a novel correlated estimator for the likelihood ratio, enabling variance control as data size grows, and analyzes its performance using the Bernstein-von Mises theorem.

Findings

Efficiency increased by over 20 times for datasets with hundreds of data points.

Efficiency increased by over 100 times for datasets with tens of thousands of data points.

The method maintains variance control as data size increases, improving scalability.

Abstract

The pseudo-marginal algorithm is a popular variant of the Metropolis--Hastings scheme which allows us to sample asymptotically from a target probability density $\pi$, when we are only able to estimate an unnormalized version of $\pi$ pointwise unbiasedly. It has found numerous applications in Bayesian statistics as there are many scenarios where the likelihood function is intractable but can be estimated unbiasedly using Monte Carlo samples. Using many samples will typically result in averages computed under this chain with lower asymptotic variances than the corresponding averages that use fewer samples. For a fixed computing time, it has been shown in several recent contributions that an efficient implementation of the pseudo-marginal method requires the variance of the log-likelihood ratio estimator appearing in the acceptance probability of the algorithm to be of order 1, which in turn usually requires scaling the number $N$ of Monte Carlo samples linearly with the number $T$ of data points. We propose a modification of the pseudo-marginal algorithm, termed the correlated pseudo-marginal algorithm, which is based on a novel log-likelihood ratio estimator computed using the difference of two positively correlated log-likelihood estimators. We show that the parameters of this scheme can be selected such that the variance of this estimator is order $1$ as $N,T\rightarrow\infty$ whenever $N/T\rightarrow 0$. By combining these results with the Bernstein-von Mises theorem, we provide an analysis of the performance of the correlated pseudo-marginal algorithm in the large $T$ regime. In our numerical examples, the efficiency of computations is increased relative to the standard pseudo-marginal algorithm by more than 20 fold for values of $T$ of a few hundreds to more than 100 fold for values of $T$ of around 10,000-20,000.