Schrödinger bridge problem via empirical risk minimization
Denis Belomestny, Alexey Naumov, Nikita Puchkin, Denis Suchkov · 2026
Read Schrödinger bridge problem via empirical risk minimization by Denis Belomestny, Alexey Naumov, Nikita Puchkin, Denis Suchkov (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-2602-08374.txt.
Description
The Schrödinger bridge problem asks for the most likely path between two distributions, but in practice those endpoints are known only through samples. Denis Belomestny and three co-authors take a learning-theory route instead of the usual Sinkhorn iterations. Their arXiv paper rewrites the bridge system around a single positive transformed potential that obeys a nonlinear fixed-point equation.
The authors then fit that potential to data, choosing from a broad family of candidate functions, with uniform concentration guarantees under sub-Gaussian assumptions. Plugging the learned potential into a stochastic control picture yields samples from the bridge. Numerical trials illustrate the method in action. Readers get a fresh, theory-backed way to bridge distributions from data alone.
Catalog record
| CREATOR | Denis Belomestny, Alexey Naumov, Nikita Puchkin, Denis Suchkov |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 82 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-08374