Fast Online Distributionally Robust Optimization via Data Compression
Irina Wang, Bartolomeo Stellato · 2025
Read Fast Online Distributionally Robust Optimization via Data Compression by Irina Wang, Bartolomeo Stellato (2025) 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-2504-08097.txt.
Description
Irina Wang and Bartolomeo Stellato wrote this 2025 paper on making robust optimization fast enough for streaming data. Distributionally robust optimization guards a decision against shifts in the data it was built on. The Wasserstein version of the method is strong but slow, and it grows as data pile up.
Their fix is online compression. Incoming data are clustered on the fly, and the robust problem is built around that small, clustered summary. The problem size then stays fixed while data keep arriving. The authors prove a dynamic regret bound against an oracle that sees the full, uncompressed stream. With the right clustering rules, the regret shrinks over time toward a gap set by the clustering error. Faster gradient updates replace full solves, and tests bear out the speed gains.
Catalog record
| CREATOR | Irina Wang, Bartolomeo Stellato |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 102 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-08097