PEPR: Privileged Event-based Predictive Regularization for Domain Generalization
Gabriele Magrini, Federico Becattini, Niccolò Biondi, Pietro Pala · 2026
Read PEPR: Privileged Event-based Predictive Regularization for Domain Generalization by Gabriele Magrini, Federico Becattini, Niccolò Biondi, Pietro Pala (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-2602-04583.txt.
Description
Gabriele Magrini and 3 co-authors use event cameras to make ordinary image models tougher in a 2026 paper on domain generalization. Event cameras report pixel changes rather than full frames. Their sparse output changes less between day and night than regular images do.
Training the image model to copy event features directly would erase rich semantic detail. Instead, their method, PEPR, asks the image encoder to predict event-based features in a shared space. Event data serve as privileged information, available only during training. The final model uses ordinary images alone. It improved robustness on detection and segmentation tasks under domain shifts, beating methods based on direct feature alignment.
Catalog record
| CREATOR | Gabriele Magrini, Federico Becattini, Niccolò Biondi, Pietro Pala |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 47 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-04583