Duality for the Adversarial Total Variation
Leon Bungert, Lucas Schmitt · 2026
Read Duality for the Adversarial Total Variation by Leon Bungert, Lucas Schmitt (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-2604-18540.txt.
Description
Bungert and Schmitt recast hostile training of two-way classifiers as penalized risk cutting with a nonlocal total variation. From that view, they pin down the subdifferential of this quantity with duality tools. They get a dual form of it plus a related parts formula built on a nonlocal slope and spread. These duality claims cover both smooth functions that fade at infinity on fitting metric spaces and bounded functions on flat domains. Under extra terms, they also pin down the subdifferential in both cases.
Catalog record
| CREATOR | Leon Bungert, Lucas Schmitt |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 140690 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.18540