Exponential Convergence Guarantees for Iterative Markovian Fitting

Marta Gentiloni Silveri, Giovanni Conforti, Alain Durmus · 2025

Read Exponential Convergence Guarantees for Iterative Markovian Fitting by Marta Gentiloni Silveri, Giovanni Conforti, Alain Durmus (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2510-20871.txt.

Description

Marta Gentiloni Silveri and two co-authors give the first quantitative exponential convergence guarantees for iterative Markovian fitting, a method for the Schrödinger bridge problem in transport and generative modeling. Earlier work proved only that the method settles in the limit, with no rate for finite runs.

Under mild structural demands on the reference measure and the target margins, with a strong enough starting fit, the team shows the error shrinks by a fixed factor at each round. That yields a clear time estimate for reaching a chosen accuracy, rather than a promise with no clock. Practical bridge matching methods inherit firmer ground, since their idealized core now has a proven speed.

Catalog record

CREATORMarta Gentiloni Silveri, Giovanni Conforti, Alain Durmus
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE103 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2510-20871

Browse the full library · About the Tower of Babel