Regularized Optimal Transport and the Rot Mover's Distance

TL;DR

This paper introduces a unified framework for smooth convex regularization of discrete optimal transport, generalizing existing methods and proposing efficient algorithms for computing regularized optimal plans, with applications in machine learning and pattern recognition.

Contribution

The paper develops a general framework for regularized optimal transport using Bregman divergences, introduces new algorithms for efficient computation, and demonstrates their effectiveness in practical applications.

Findings

Efficient algorithms for regularized optimal transport plans.

Generalization of entropy-based regularization to other divergences.

Successful application to audio scene classification.

Abstract

This paper presents a unified framework for smooth convex regularization of discrete optimal transport problems. In this context, the regularized optimal transport turns out to be equivalent to a matrix nearness problem with respect to Bregman divergences. Our framework thus naturally generalizes a previously proposed regularization based on the Boltzmann-Shannon entropy related to the Kullback-Leibler divergence, and solved with the Sinkhorn-Knopp algorithm. We call the regularized optimal transport distance the rot mover's distance in reference to the classical earth mover's distance. We develop two generic schemes that we respectively call the alternate scaling algorithm and the non-negative alternate scaling algorithm, to compute efficiently the regularized optimal plans depending on whether the domain of the regularizer lies within the non-negative orthant or not. These schemes are based on Dykstra's algorithm with alternate Bregman projections, and further exploit the Newton-Raphson method when applied to separable divergences. We enhance the separable case with a sparse extension to deal with high data dimensions. We also instantiate our proposed framework and discuss the inherent specificities for well-known regularizers and statistical divergences in the machine learning and information geometry communities. Finally, we demonstrate the merits of our methods with experiments using synthetic data to illustrate the effect of different regularizers and penalties on the solutions, as well as real-world data for a pattern recognition application to audio scene classification.