Diffusion Approximations to Schr\"{o}dinger Bridges on Manifolds

TL;DR

This paper develops explicit diffusion approximations for small temperature Schr"{o}dinger bridges on manifolds, showing convergence of potentials and applications to Euclidean cases with quadratic cost.

Contribution

It introduces new diffusion approximation techniques for Schr"{o}dinger bridges on manifolds, especially when marginals are identical and reference processes are reversible diffusions.

Findings

Gradient of Schr"{o}dinger potential converges to manifold score function as temperature vanishes.

Approximation of Euclidean Schr"{o}dinger bridge using stationary Mirror Langevin diffusion.

Explicit small temperature diffusion approximations for Schr"{o}dinger bridges on manifolds.

Abstract

We present a collection of explicit diffusion approximations to small temperature Schr\"{o}dinger bridges on manifolds. Our most precise results are when both marginals are the same and the Schr\"{o}dinger bridge is on a manifold with a reference process given by a reversible diffusion. In the special case that the reference process is the manifold Brownian motion, we use the small time heat kernel asymptotics to show that the gradient of the corresponding Schr\"{o}dinger potential converges in $L^2$, as the temperature vanishes, to a manifold analogue of the score function of the marginal. As an application of the previous result we show that the Euclidean Schr\"{o}dinger bridge, computed for the quadratic cost, between two different marginal distributions can be approximated by a transformation of a two point distribution of a stationary Mirror Langevin diffusion.