English

Overrelaxed Sinkhorn-Knopp Algorithm for Regularized Optimal Transport

Numerical Analysis 2021-04-02 v2 Numerical Analysis Optimization and Control

Abstract

This article describes a set of methods for quickly computing the solution to the regularized optimal transport problem. It generalizes and improves upon the widely-used iterative Bregman projections algorithm (or Sinkhorn--Knopp algorithm). We first propose to rely on regularized nonlinear acceleration schemes. In practice, such approaches lead to fast algorithms, but their global convergence is not ensured. Hence, we next propose a new algorithm with convergence guarantees. The idea is to overrelax the Bregman projection operators, allowing for faster convergence. We propose a simple method for establishing global convergence by ensuring the decrease of a Lyapunov function at each step. An adaptive choice of overrelaxation parameter based on the Lyapunov function is constructed. We also suggest a heuristic to choose a suitable asymptotic overrelaxation parameter, based on a local convergence analysis. Our numerical experiments show a gain in convergence speed by an order of magnitude in certain regimes.

Keywords

Cite

@article{arxiv.1711.01851,
  title  = {Overrelaxed Sinkhorn-Knopp Algorithm for Regularized Optimal Transport},
  author = {Alexis Thibault and Lénaïc Chizat and Charles Dossal and Nicolas Papadakis},
  journal= {arXiv preprint arXiv:1711.01851},
  year   = {2021}
}

Comments

NIPS'17 Workshop on Optimal Transport & Machine Learning

R2 v1 2026-06-22T22:37:05.514Z