AI Chat AI Image Generator AI Video Text to Speech

Quantitative Uniform Stability of the Iterative Proportional Fitting Procedure

08/18/2021

∙

by George Deligiannidis, et al.

∙

∙

We establish the uniform in time stability, w.r.t. the marginals, of the Iterative Proportional Fitting Procedure, also known as Sinkhorn algorithm, used to solve entropy-regularised Optimal Transport problems. Our result is quantitative and stated in terms of the 1-Wasserstein metric. As a corollary we establish a quantitative stability result for Schrödinger bridges.

George Deligiannidis
31 publications
Valentin De Bortoli
34 publications
Arnaud Doucet
106 publications

page 1

page 2

page 3

page 4

research

∙ 07/26/2022

The derivatives of Sinkhorn-Knopp converge

We show that the derivatives of the Sinkhorn-Knopp algorithm, or iterati...

0 Edouard Pauwels, et al. ∙

research

∙ 07/09/2023

Copula-like inference for discrete bivariate distributions with rectangular support

After reviewing a large body of literature on the modeling of bivariate ...

0 Ivan Kojadinovic, et al. ∙

research

∙ 12/31/2019

Schrödinger Bridge Samplers

Consider a reference Markov process with initial distribution π_0 and tr...

16 Espen Bernton, et al. ∙

research

∙ 07/31/2023

Wasserstein Mirror Gradient Flow as the limit of the Sinkhorn Algorithm

We prove that the sequence of marginals obtained from the iterations of ...

0 Nabarun Deb, et al. ∙

research

∙ 10/17/2018

The Wasserstein transform

We introduce the Wasserstein transform, a method for enhancing and denoi...

0 Facundo Mémoli, et al. ∙

research

∙ 10/12/2020

Quantitative stability and numerical analysis of Markovian quadratic BSDEs with reflection

We study quantitative stability of the solutions to Markovian quadratic ...

0 Dingqian Sun, et al. ∙

research

∙ 06/16/2021

On the stability of the L^2 projection and the quasiinterpolant in the space of smooth periodic splines

In this paper we derive stability estimates in L^2- and L^∞- based Sobol...

0 D. C. Antonopoulos V. A. Dougalis, et al. ∙