Finite-sample bounds for regularized optimal transport
Alberto González-Sanz, Marcel Nutz, Austin J. Stromme · 2026
Read Finite-sample bounds for regularized optimal transport by Alberto González-Sanz, Marcel Nutz, Austin J. Stromme (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2606-25947.txt.
Description
Alberto González-Sanz, Marcel Nutz, and Austin Stromme sharpen the statistics of regularized optimal transport. Optimal transport compares probability distributions by the cheapest way to morph one into the other. Real data give only samples, so the empirical cost wobbles around the truth. Regularization smooths the problem, at the price of bias. The team bounds both bias and variance for finite samples.
General convex penalties fall under the results. Kullback-Leibler divergence and power penalties are both covered. Bounds show explicit dependence on the smoothing strength and the data's intrinsic dimension. Entropic transport gets improved state-of-the-art bounds. Power penalties get their first fully quantitative treatment. For quadratic cost, the quadratically regularized estimator hits the fastest known rate. It estimates the true transport cost faster than any rival based on regularized transport. The unified approach replaces a patchwork of case-by-case results.
Catalog record
| CREATOR | Alberto González-Sanz, Marcel Nutz, Austin J. Stromme |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 118 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2606.25947