A review of matrix scaling and Sinkhorn's normal form for matrices and positive maps

Abstract

Given a nonnegative matrix $A$, can you find diagonal matrices $D_1,D_2$ such that $D_{1}A{D}_2$ is doubly stochastic? The answer to this question is known as Sinkhorn's theorem. It has been proved with a wide variety of methods, each presenting a variety of possible generalisations. Recently, generalisations such as to positive maps between matrix algebras have become more and more interesting for applications. This text gives a review of over 70 years of matrix scaling. The focus lies on the mathematical landscape surrounding the problem and its solution as well as the generalisation to positive maps and contains hardly any nontrivial unpublished results.

Maybe this helps somebody. Comments, corrections or suggestions for improving the review most welcome

References in corpus (1)

• Unifying several separability conditions using the covariance matrix criterion

Cited by in corpus (15)

• A mathematical theory of cooperative communication
• Biwhitening Reveals the Rank of a Count Matrix
• Optimal Transport losses and Sinkhorn algorithm with general convex regularization
• Optimal Lockdown for Pandemic Control
• On the Variance of Unbiased Online Recurrent Optimization
• The unreasonable effectiveness of optimal transport in economics
• Matching Function Equilibria with Partial Assignment: Existence, Uniqueness and Estimation
• Local Two-Sample Testing over Graphs and Point-Clouds by Random-Walk Distributions
• An essay on copula modelling for discrete random vectors; or how to pour new wine into old bottles
• Optimal Approximation of Doubly Stochastic Matrices
• Scaling positive random matrices: concentration and asymptotic convergence
• Quantum algorithms for matrix scaling and matrix balancing
• Mosaic: A Sample-Based Database System for Open World Query Processing
• Matrix scaling, explicit Sinkhorn limits, and arithmetic
• Sample Debiasing in the Themis Open World Database System (Extended Version)