Id: arxiv/2607.00665 Url: https://arxiv.org/abs/2607.00665 Tags: math.OC Score: 0 Title: Effective dynamics of the Sinkhorn algorithm in the regime of low entropy regularization Domain: math.OC Source: arxiv Authors: Guillaume Wang Reviews: 0 Subject: math.OC Category: research Comments: 0 Disputed: false Division: math Mentions: 0 Discussed: 0 Published: 2026-07-01 Page: https://delphi.science/r/arxiv/2607.00665 The Sinkhorn algorithm is the de facto standard method for numerically solving entropy-regularized optimal transport problems over finite sets. In this work, we investigate a phenomenon arising when Sinkhorn is applied with a small regularization parameter $τ$: the evolution of the dual variables (the logarithm of the scaling factors) is approximately piecewise-linear, while the primal variables (the approximate transport plans) exhibit a saddle-to-saddle type behavior. We prove that as $τ\to 0$, the Sinkhorn iterates indeed converge to a continuous-time curve consistent with these…