Transport Elliptical Slice Sampling

TL;DR

This paper introduces Transport Elliptical Slice Sampling (TESS), a novel method combining normalizing flows and elliptical slice sampling to efficiently generate low-autocorrelation samples from complex distributions, optimized for modern parallel architectures.

Contribution

The paper presents TESS, a new framework that learns a diffeomorphism to transform complex distributions into Gaussian form for efficient sampling with elliptical slice sampling.

Findings

TESS produces samples with lower autocorrelation than non-transformed methods.

TESS significantly improves efficiency over gradient-based proposals on parallel architectures.

Numerical results demonstrate TESS's effectiveness on complex target distributions.

Abstract

We propose a new framework for efficiently sampling from complex probability distributions using a combination of normalizing flows and elliptical slice sampling (Murray et al., 2010). The central idea is to learn a diffeomorphism, through normalizing flows, that maps the non-Gaussian structure of the target distribution to an approximately Gaussian distribution. We then use the elliptical slice sampler, an efficient and tuning-free Markov chain Monte Carlo (MCMC) algorithm, to sample from the transformed distribution. The samples are then pulled back using the inverse normalizing flow, yielding samples that approximate the stationary target distribution of interest. Our transport elliptical slice sampler (TESS) is optimized for modern computer architectures, where its adaptation mechanism utilizes parallel cores to rapidly run multiple Markov chains for a few iterations. Numerical demonstrations show that TESS produces Monte Carlo samples from the target distribution with lower autocorrelation compared to non-transformed samplers, and demonstrates significant improvements in efficiency when compared to gradient-based proposals designed for parallel computer architectures, given a flexible enough diffeomorphism.