Metraipolis › archive › dc_ee7f99a13899

When to Unpair: Regulating Pairing Dependence in Medical Visual In-Context Learning — arXiv:2610.10335

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.10335
License (as archived)
CC0-1.0
Archived
2026-10-08T12:48:03.082Z
SHA-256 of the archived text
9fe11171082e6757f995f2e298ce4fb5297f902698edf508ace3a371c8a3a3cb
Stable URL of this page
https://metraipolis.com/documento/dc_ee7f99a13899

How to cite

When to Unpair: Regulating Pairing Dependence in Medical Visual In-Context Learning — arXiv:2610.10335 — arXiv (en), https://arxiv.org/abs/2610.10335 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ee7f99a13899 (Metraipolis archive dc_ee7f99a13899, sha256 9fe11171082e6757f995f2e298ce4fb5297f902698edf508ace3a371c8a3a3cb)
BibTeX
@misc{dcee7f99a13899,
  title = {When to Unpair: Regulating Pairing Dependence in Medical Visual In-Context Learning — arXiv:2610.10335},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.10335},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ee7f99a13899; sha256 9fe11171082e6757f995f2e298ce4fb5297f902698edf508ace3a371c8a3a3cb},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.10335. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Visual in-context learning (ICL), well suited to label-scarce medical imaging, uses support image-label pairs to demonstrate input-output mappings, while the labels collectively indicate the requested task. We diagnose dependence on individual pairings with a test-time derangement that reassigns every support label to another support image while preserving the query, support images, and label multiset. The resulting pairing gap, defined as shuffled-minus-matched performance, shows that all four released models depend on the pairing, to widely varying degrees. Further analysis of a paired-trained model reveals support-associated spurious regions and lesion-size biases even with real, unaltered supports, alongside sensitivity to mis-registered support labels. To regulate this dependence, we introduce a late unpairing curriculum (LUC), which starts with matched training and then applies random unpairing, replacing each support label with that of another support in the same episode. LUC nearly closes the pairing gap on two backbones while maintaining or improving matched-support performance across all evaluated task types, with gains extending to held-out tasks and cross-dataset episodes. It also mitigates these failure modes. On BraTS whole-tumor segmentation, matched-support DSC rises from 0.733 to 0.857 while the gap shrinks from -0.184 to -0.008. In a released model, brief fine-tuning with random unpairing reduces the gap. A reversed curriculum that places the same number of unpairing epochs at the start of training leaves a large gap. This shows that pairing dependence is shaped by the order of training and not only by the amount of unpaired training.

Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.CV, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2610.10335 — 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 2026-10-08T12:18:15.482Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).