Catalyst: Out-of-Distribution Detection via Elastic Scaling
Abid Hassan, Tuan Ngo, Saad Shafiq, Nenad Medvidovic · 2026
Read Catalyst: Out-of-Distribution Detection via Elastic Scaling by Abid Hassan, Tuan Ngo, Saad Shafiq, Nenad Medvidovic (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-2602-02409.txt.
Description
Abid Hassan and 3 co-authors find a discarded signal that helps neural networks flag unfamiliar inputs, in a 2026 arXiv paper. Detecting out-of-distribution inputs, data unlike anything seen in training, is vital for safe deployment. Most detectors score inputs from final outputs or from features averaged over the whole image, a step that throws away channel-by-channel detail.
Their framework, Catalyst, computes a scaling factor on the fly from those raw pre-pooling statistics, such as channel means and peak activations. The factor multiplies an existing detector's score, elastically stretching in-distribution and unfamiliar inputs further apart. Catalyst plugs into both logit-based and distance-based detectors. False positive rates dropped sharply on CIFAR and ImageNet benchmarks. The lesson is that information averaged away still carries safety value.
Catalog record
| CREATOR | Abid Hassan, Tuan Ngo, Saad Shafiq, Nenad Medvidovic |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 171 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-02409