Metraipolis › archive › dc_fedb03fb669e

Different Models, Shared Failures: Auditing Frozen-Encoder Pools for Medical Segmentation — arXiv:2610.41790

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.41790
License (as archived): CC0-1.0
Archived: 2026-11-03T08:03:27.665Z
SHA-256 of the archived text: c28aaae8be0e0711823d2ecc6fe6f4a4225ef71074889f413e8948845e3aa70f
Stable URL of this page: https://metraipolis.com/documento/dc_fedb03fb669e

How to cite

Different Models, Shared Failures: Auditing Frozen-Encoder Pools for Medical Segmentation — arXiv:2610.41790 — arXiv (en), https://arxiv.org/abs/2610.41790 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_fedb03fb669e (Metraipolis archive dc_fedb03fb669e, sha256 c28aaae8be0e0711823d2ecc6fe6f4a4225ef71074889f413e8948845e3aa70f)

BibTeX

@misc{dcfedb03fb669e,
  title = {Different Models, Shared Failures: Auditing Frozen-Encoder Pools for Medical Segmentation — arXiv:2610.41790},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.41790},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_fedb03fb669e; sha256 c28aaae8be0e0711823d2ecc6fe6f4a4225ef71074889f413e8948845e3aa70f},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.41790. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Average segmentation accuracy and model count do not reveal whether models fail on the same cases. We present CoFailMed, an audit of correlated failures in pretrained-encoder pools for medical segmentation. Across five 2D tasks, we evaluate 11 public pretrained encoders (a controlled sample, not a census) using frozen features and separately trained $1\times 1$ readouts with four seeds per encoder. Shared-checkpoint pairs have the highest mean per-case Dice correlation. An expanded 17-encoder analysis shows that distinct checkpoints within the same encoder family remain more correlated than cross-family pairs. At the pooled 33rd-percentile Dice threshold, four-member pools drawn from distinct encoders have mean effective sizes of 1.56-2.92, estimated from failure correlations in pools with enough failures and passes per member. In the primary pool, richer readouts and fine-tuning narrow the Dice-correlation gap between pairs sharing a checkpoint (or the DINOv2-B/S release) and all other pairs. Yet dependence between different encoders can increase, and fine-tuning does not consistently increase effective pool size at 33rd-percentile thresholds recomputed for each training recipe. A smaller between-group gap can coexist with greater dependence. Pool evaluation should therefore report absolute dependence and threshold-specific effective size alongside accuracy and nominal pool size. The CoFailMed release provides aligned per-case traces, analysis scripts and study code to reproduce the central audit results.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.41790 — 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 2026-11-03T07:59:38.818Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).