Metraipolis › archive › dc_2993e2ee34ed

ProME: Prototype-Margin Environments with Repair-Aware Selection for Group-Robust Learning — arXiv:2608.13190

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.13190
License (as archived)
CC0-1.0
Archived
2026-10-19T04:27:35.191Z
SHA-256 of the archived text
b4c3bbe292c56b19bf62b78880f6a7e219046c79b53e11a4e0bd58d85f5c53e6
Stable URL of this page
https://metraipolis.com/documento/dc_2993e2ee34ed

How to cite

ProME: Prototype-Margin Environments with Repair-Aware Selection for Group-Robust Learning — arXiv:2608.13190 — arXiv (en), https://arxiv.org/abs/2608.13190 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_2993e2ee34ed (Metraipolis archive dc_2993e2ee34ed, sha256 b4c3bbe292c56b19bf62b78880f6a7e219046c79b53e11a4e0bd58d85f5c53e6)
BibTeX
@misc{dc2993e2ee34ed,
  title = {ProME: Prototype-Margin Environments with Repair-Aware Selection for Group-Robust Learning — arXiv:2608.13190},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.13190},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_2993e2ee34ed; sha256 b4c3bbe292c56b19bf62b78880f6a7e219046c79b53e11a4e0bd58d85f5c53e6},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.13190. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Group-robust learning is crucial for maintaining accuracy on rare subpopulations when training-group labels are unavailable. However, existing methods often infer environments from a separate reference model and select representations before fitting the classifier used at deployment, leaving both decisions misaligned with the deployed predictor. In this work, we formulate group robustness without training-group labels as the endogenous environments with repair-aware selection (ERAS) problem, and propose ProME (Prototype-Margin Environments) to align both decisions with the deployed predictor. ProME splits prototype margins at their median to construct approximately balanced environments along the training trajectory, and fits a group-balanced linear head on group-annotated validation data to rank the resulting predictors by validation worst-group accuracy. We theoretically bound the worst risk across the inferred environments for a fixed predictor and partition, showing that this bound transfers to the oracle groups under an explicit alignment condition. Extensive experiments show that prototype margins enrich shortcut-conflicting examples, classifier repair reshapes candidate evaluation, and ProME achieves the highest average worst-group accuracy among the compared methods with the same group-label access.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-14 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG) ; Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2608.13190 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-19T04:20:35.849Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-08-14?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).