OUT-OF-DISTRIBUTION DETECTION BASED ON TOTAL VARIATION ESTIMATION
Dabiao Ma, Zhiba Su, Jian Yang, Haojun Fei · 2026
Read OUT-OF-DISTRIBUTION DETECTION BASED ON TOTAL VARIATION ESTIMATION by Dabiao Ma, Zhiba Su, Jian Yang, Haojun Fei (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-2601-15867.txt.
Description
Dabiao Ma, Zhiba Su, Jian Yang, and Haojun Fei detect out-of-distribution inputs in a 2026 arXiv paper. Models deployed in the real world meet data unlike their training sets. Distribution shifts can silently wreck their accuracy. Guards that flag foreign inputs are therefore essential. Their method is TV-OOD, short for total variation out-of-distribution detection. It improves on existing methods, the authors say, by a new measurement.
A total variation network estimator scores each input. It estimates the input's contribution to overall total variation. Inputs that push variation high look out of place. That signal separates them from normal data. For safety-critical deployments, a good OOD alarm is the difference between a graceful refusal and a confident error.
Catalog record
| CREATOR | Dabiao Ma, Zhiba Su, Jian Yang, Haojun Fei |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 18 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-15867