Exponential Convergence Guarantees for Iterative Markovian Fitting

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Silveri, Marta Gentiloni, Conforti, Giovanni, Durmus, Alain
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909867027988480
author Silveri, Marta Gentiloni
Conforti, Giovanni
Durmus, Alain
author_facet Silveri, Marta Gentiloni
Conforti, Giovanni
Durmus, Alain
contents The Schrödinger Bridge (SB) problem has become a fundamental tool in computational optimal transport and generative modeling. To address this problem, ideal methods such as Iterative Proportional Fitting and Iterative Markovian Fitting (IMF) have been proposed-alongside practical approximations like Diffusion Schrödinger Bridge and its Matching (DSBM) variant. While previous work have established asymptotic convergence guarantees for IMF, a quantitative, non-asymptotic understanding remains unknown. In this paper, we provide the first non-asymptotic exponential convergence guarantees for IMF under mild structural assumptions on the reference measure and marginal distributions, assuming a sufficiently large time horizon. Our results encompass two key regimes: one where the marginals are log-concave, and another where they are weakly log-concave. The analysis relies on new contraction results for the Markovian projection operator and paves the way to theoretical guarantees for DSBM.
format Preprint
id arxiv_https___arxiv_org_abs_2510_20871
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Exponential Convergence Guarantees for Iterative Markovian Fitting
Silveri, Marta Gentiloni
Conforti, Giovanni
Durmus, Alain
Machine Learning
Probability
The Schrödinger Bridge (SB) problem has become a fundamental tool in computational optimal transport and generative modeling. To address this problem, ideal methods such as Iterative Proportional Fitting and Iterative Markovian Fitting (IMF) have been proposed-alongside practical approximations like Diffusion Schrödinger Bridge and its Matching (DSBM) variant. While previous work have established asymptotic convergence guarantees for IMF, a quantitative, non-asymptotic understanding remains unknown. In this paper, we provide the first non-asymptotic exponential convergence guarantees for IMF under mild structural assumptions on the reference measure and marginal distributions, assuming a sufficiently large time horizon. Our results encompass two key regimes: one where the marginals are log-concave, and another where they are weakly log-concave. The analysis relies on new contraction results for the Markovian projection operator and paves the way to theoretical guarantees for DSBM.
title Exponential Convergence Guarantees for Iterative Markovian Fitting
topic Machine Learning
Probability
url https://arxiv.org/abs/2510.20871