# Interacting Langevin Diffusions: Gradient Structure And Ensemble Kalman Sampler

## TL;DR

This paper introduces a derivative-free ensemble Kalman sampler based on interacting Langevin diffusions with a novel gradient flow structure, demonstrating its theoretical properties and practical effectiveness in Bayesian inverse problems.

## Contribution

It develops a new gradient flow framework for interacting Langevin diffusions and proposes a derivative-free ensemble Kalman sampler with proven convergence properties.

## Key findings

- The invariant measure matches that of a single Langevin diffusion.
- Exponential convergence to the invariant measure is demonstrated.
- Numerical experiments show the sampler's effectiveness in Bayesian inverse problems.

## Abstract

Solving inverse problems without the use of derivatives or adjoints of the forward model is highly desirable in many applications arising in science and engineering. In this paper, we propose a new version of such a methodology, a framework for its analysis, and numerical evidence of the practicality of the method proposed. Our starting point is an ensemble of over-damped Langevin diffusions which interact through a single preconditioner computed as the empirical ensemble covariance. We demonstrate that the nonlinear Fokker-Planck equation arising from the mean-field limit of the associated stochastic differential equation (SDE) has a novel gradient flow structure, built on the Wasserstein metric and the covariance matrix of the noisy flow. Using this structure, we investigate large time properties of the Fokker-Planck equation, showing that its invariant measure coincides with that of a single Langevin diffusion, and demonstrating exponential convergence to the invariant measure in a number of settings. We introduce a new noisy variant on ensemble Kalman inversion (EKI) algorithms found from the original SDE by replacing exact gradients with ensemble differences; this defines the ensemble Kalman sampler (EKS). Numerical results are presented which demonstrate its efficacy as a derivative-free approximate sampler for the Bayesian posterior arising from inverse problems.