Rapid Convergence of Informed Importance Tempering

TL;DR

This paper introduces informed importance tempering (IIT), a class of MCMC methods that combine importance sampling and informed proposals, with theoretical spectral gap bounds showing scalability in high-dimensional Bayesian posterior computations.

Contribution

The paper proposes IIT, providing the first spectral gap bounds for this class and analyzing how proposal choices affect performance in high-dimensional settings.

Findings

IIT exhibits remarkable scalability when the posterior concentrates on small sets.

The performance of informed proposals is sensitive to the target distribution's shape.

Square-root proposal weighting generally performs well across various scenarios.

Abstract

Informed Markov chain Monte Carlo (MCMC) methods have been proposed as scalable solutions to Bayesian posterior computation on high-dimensional discrete state spaces, but theoretical results about their convergence behavior in general settings are lacking. In this article, we propose a class of MCMC schemes called informed importance tempering (IIT), which combine importance sampling and informed local proposals, and derive generally applicable spectral gap bounds for IIT estimators. Our theory shows that IIT samplers have remarkable scalability when the target posterior distribution concentrates on a small set. Further, both our theory and numerical experiments demonstrate that the informed proposal should be chosen with caution: the performance of some proposals may be very sensitive to the shape of the target distribution. We find that the "square-root proposal weighting" scheme tends to perform well in most settings.