Stein Variational Evolution Strategies

TL;DR

This paper introduces a novel gradient-free Stein Variational Gradient Descent method that combines SVGD with evolution strategies, enabling efficient sampling from unnormalized distributions without gradient information, outperforming previous methods.

Contribution

The paper proposes integrating evolution strategies with SVGD to enhance gradient-free sampling, addressing limitations of existing Monte Carlo and surrogate-based approaches.

Findings

The combined method produces high-quality samples from unnormalized densities.

It outperforms prior gradient-free SVGD methods on benchmark problems.

The approach does not require gradient information for the target distribution.

Abstract

Stein Variational Gradient Descent (SVGD) is a highly efficient method to sample from an unnormalized probability distribution. However, the SVGD update relies on gradients of the log-density, which may not always be available. Existing gradient-free versions of SVGD make use of simple Monte Carlo approximations or gradients from surrogate distributions, both with limitations. To improve gradient-free Stein variational inference, we combine SVGD steps with evolution strategy (ES) updates. Our results demonstrate that the resulting algorithm generates high-quality samples from unnormalized target densities without requiring gradient information. Compared to prior gradient-free SVGD methods, we find that the integration of the ES update in SVGD significantly improves the performance on multiple challenging benchmark problems.