{"id":"arxiv/2602.10726","url":"https://arxiv.org/abs/2602.10726","tags":["math.AP","math.OC"],"score":0,"title":"The Wasserstein gradient flow of the Sinkhorn divergence between Gaussian distributions","domain":"math.AP","source":"arxiv","authors":["Mathis Hardion","Théo Lacombe"],"reviews":0,"subject":"math.AP","summary":"We study the Wasserstein gradient flow of the Sinkhorn divergence when both the source and the target are Gaussian distributions. We prove the existence of a flow that stays in the class of Gaussian distributions, and is unique in the larger class of measures with strongly-concave and smooth log-densities. We prove that the flow globally converges toward the target measure when the source's covariance matrix is not singular, and provide counter-examples to global convergence when it is, giving a first answer to an open question raised in [Carlier et al. 2024, \\S4.2]. When the covariance…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-02-11","reactions":{}}