What type of study is this?

This is a Quantitative Study study.

October 20, 2025Open Access

Sampling via Gaussian Mixture Approximations

Key Points

GMA sampling achieves accurate results by fitting mixtures to target densities via optimisation and resampling.
The method employs projected gradient descent and other techniques, yielding consistent approximations under mild conditions.
Empirical results demonstrate that GMA is computationally efficient, outpacing traditional methods in speed and accuracy.
Gradient-free sampling avoids complex gradients, maintaining robust performance across diverse unnormalised target densities.

Abstract

We present a family of Gaussian Mixture Approximation (GMA) samplers for sampling unnormalised target densities, encompassing weights-only GMA (W-GMA), Laplace Mixture Approximation (LMA), expectation-maximization GMA (EM-GMA), and further variants. GMA adopts a simple two-stage paradigm: (i) initialise a finite set of Gaussian components and draw samples from a proposal mixture; (ii) fit the mixture to the target by optimising either only the component weights or also the means and variances, via a sample-based KL divergence objective that requires only evaluations of the unnormalised density, followed by stratified resampling. The method is gradient-free, and computationally efficient: it leverages the ease of sampling from Gaussians, efficient optimisation methods (projected gradient descent, mirror descent, and EM), and the robustness of stratified resampling to produce samples faithful to the target. We show that this optimisation-resampling scheme yields consistent approximations under mild conditions, and we validate this methodology with empirical results demonstrating accuracy and speed across diverse densities.

Read Full Paperexternally

Mark Helpful

Bookmark

Relay

View Full Paper