Optimal transport problems regularized by generic convex functions: a geometric and algorithmic approach

DOI10.1007/S41884-021-00062-6zbMATH Open1495.90109arXiv2011.13683OpenAlexW3217042118MaRDI QIDQ2154656FDOQ2154656

Daiji Tsutsui

Publication date: 20 July 2022

Published in: Information Geometry (Search for Journal in Brave)

Abstract: In order to circumvent the difficulties in solving numerically the discrete optimal transport problem, in which one minimizes the linear target function $PmapstolangleC,Pangle:=su{m}_{i,j}{C}_{ij}{P}_{ij}$, Cuturi introduced a variant of the problem in which the target function is altered by a convex one $Phi(P)=langleC,Pangle-lambdamathcalH(P)$, where $mathcalH$ is the Shannon entropy and $lambda$ is a positive constant. We herein generalize their formulation to a target function of the form $Phi(P)=langleC,Pangle+lambdaf(P)$, where $f$ is a generic strictly convex smooth function. We also propose an iterative method for finding a numerical solution, and clarify that the proposed method is particularly efficient when $f(P)=frac12|P{|}^2$.

Full work available at URL: https://arxiv.org/abs/2011.13683

zbMATH Keywords

discrete optimal transportconvex optimizationinformation geometryWasserstein barycenterentropic regularization

Mathematics Subject Classification ID

Convex programming (90C25) Special problems of linear programming (transportation, multi-index, data envelopment analysis, etc.) (90C08)

Cites Work

• Title not available (Why is that?)
• Title not available (Why is that?)
• Computational Optimal Transport: With Applications to Data Science
• Optimal transport for applied mathematicians. Calculus of variations, PDEs, and modeling
• Barycenters in the Wasserstein Space
• Iterative Bregman Projections for Regularized Transportation Problems
• On the scaling of multidimensional matrices
• The geometry of proper scoring rules
• A smoothed dual approach for variational Wasserstein problems
• Information geometry connecting Wasserstein distance and Kullback-Leibler divergence via the entropy-relaxed transportation problem
• Quadratically Regularized Optimal Transport on Graphs

Cited In (6)

• A geometric variational framework for computing spherical optimal transportation maps. II
• Minimizing optimal transport for functions with fixed-size nodal sets
• Generalized unnormalized optimal transport and its fast algorithms
• The GenCol Algorithm for High-Dimensional Optimal Transport: General Formulation and Application to Barycenters and Wasserstein Splines
• Quadratically regularized optimal transport
• When optimal transport meets information geometry

This page was built for publication: Optimal transport problems regularized by generic convex functions: a geometric and algorithmic approach

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2154656)