Asymptotic distribution and convergence rates of stochastic algorithms for entropic optimal transportation between probability measures (Q2039796)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Asymptotic distribution and convergence rates of stochastic algorithms for entropic optimal transportation between probability measures |
scientific article |
Statements
Asymptotic distribution and convergence rates of stochastic algorithms for entropic optimal transportation between probability measures (English)
0 references
5 July 2021
0 references
Consider the estimation of a regularized Wasserstein distance \(W_\varepsilon(\mu,\nu)\) (also known as a Sinkhorn divergence) between two probability measures \(\mu\) and \(\nu\) in the case where \(\nu\) is discrete with finite support. This can be tackled with a stochastic Robbins-Monro algorithm using a sequence of observations from \(\mu\). In this paper, the authors establish asymptotic properties of such algorithms. In particular, their main results are almost sure convergence and asymptotic normality of the resulting estimator of the Sinkhorn divergence, as well as analysis of the expected excess risk of the estimator. Further, the authors also discuss the unregularized case (corresponding to the choice \(\varepsilon=0\)) in which case the Sinkhorn divergence becomes a Wasserstein distance. Numerical experiments illustrate the results of the paper.
0 references
convergence of random variables
0 references
entropic regularization
0 references
optimal transport
0 references
Sinkhorn divergence
0 references
stochastic optimisation
0 references
Wasserstein distance
0 references
0 references
0 references
0 references
0 references