Metraipolis › archive › dc_7c6c04deda65
Coverage Hides Churn: Auditing and Preserving Conformal Decisions as VLM Label Registries Grow — arXiv:2611.07502
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.07502
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-06T14:51:11.995Z
- SHA-256 of the archived text
ffd9af0a6ec83d7cba45625cb93e55ba5e6bdbb1ce383b56edc328baac69f66f- Stable URL of this page
https://metraipolis.com/documento/dc_7c6c04deda65
How to cite
Coverage Hides Churn: Auditing and Preserving Conformal Decisions as VLM Label Registries Grow — arXiv:2611.07502 — arXiv (en), https://arxiv.org/abs/2611.07502 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7c6c04deda65 (Metraipolis archive dc_7c6c04deda65, sha256 ffd9af0a6ec83d7cba45625cb93e55ba5e6bdbb1ce383b56edc328baac69f66f)
BibTeX
@misc{dc7c6c04deda65,
title = {Coverage Hides Churn: Auditing and Preserving Conformal Decisions as VLM Label Registries Grow — arXiv:2611.07502},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.07502},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7c6c04deda65; sha256 ffd9af0a6ec83d7cba45625cb93e55ba5e6bdbb1ce383b56edc328baac69f66f},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.07502. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Vision-language classifiers can learn new labels from text alone, so the label lists they serve keep growing. We show that this growth silently changes conformal decisions about old labels. Common softmax-based conformal scores compare each label with an average over all labels; adding labels changes that average, and with it the score of every old label, even when the model, the image, and the old calibration data stay the same. Recalibrating after the update keeps the coverage guarantee, yet across eleven settings it changes at least one old decision on 53-90% of new images. Coverage cannot reveal this, because it checks only the true label: in five controlled settings that keep the old calibration, so that only the normalizer changes, old-class coverage moves by at most 0.01 while 60-67% of images change. We propose a lifecycle audit that measures these changes alongside coverage and set size, and a simple fix, PrefixCP: each class keeps the label registry that existed when it was added. Its old decisions then never change, each class keeps exact class-conditional coverage under exchangeability, and one prefix scan computes all historical normalizers at once. PrefixCP recovers 40.9-58.5% of the set-size gap between never updating the reference and recalibrating, which gives smaller sets but changes old decisions. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2611.07502 — 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 2026-11-06T14:21:42.660Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).