Metraipolis › archive › dc_53b185aef9dd
Answer or Defer? Testing VLMs' Stated Confidence on Brain MRI — arXiv:2611.11833
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.11833
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-10T18:43:43.316Z
- SHA-256 of the archived text
40292172f12759bb3d40e1bd0337a24d5c334d815c402c043e6c1c2034bb8783- Stable URL of this page
https://metraipolis.com/documento/dc_53b185aef9dd
How to cite
Answer or Defer? Testing VLMs' Stated Confidence on Brain MRI — arXiv:2611.11833 — arXiv (en), https://arxiv.org/abs/2611.11833 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_53b185aef9dd (Metraipolis archive dc_53b185aef9dd, sha256 40292172f12759bb3d40e1bd0337a24d5c334d815c402c043e6c1c2034bb8783)
BibTeX
@misc{dc53b185aef9dd,
title = {Answer or Defer? Testing VLMs' Stated Confidence on Brain MRI — arXiv:2611.11833},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.11833},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_53b185aef9dd; sha256 40292172f12759bb3d40e1bd0337a24d5c334d815c402c043e6c1c2034bb8783},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.11833. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Vision-language models (VLMs) could support medical-image interpretation, but their use requires deciding which interpretations to accept and which images to refer to a clinician. We test whether thresholding stated confidence can retain useful coverage while limiting errors among accepted answers, and how this depends on the population. Eight open-weight VLMs, including two reasoning models, assess 3,888 brain MRI slices from post-treatment glioma patients (BraTS 2024) and healthy volunteers (IXI) for tumor or gross abnormality. BraTS labels are mask-derived; IXI slices are treated as negatives. On BraTS, no threshold selects a nonempty set with at most 5% empirical error for any model, even when chosen retrospectively. This also holds for four models evaluated on a large-lesion, distant-negative subset. At a 10% error budget, Gemma-4-12B retains 24.4% of slices and every other model at most 2.5%. To investigate this limited selection ability, we separate confidence differences between predicted answers from differences within each answer. For four models with available outcome-level summaries, pooled mean confidence differs more between answers than between correct and incorrect responses within each answer. A covariance decomposition explains how this pattern can produce opposite confidence-correctness associations across populations. At a 20% error budget, accepting every "tumor" answer covers more slices than any confidence threshold for two models. These findings caution against using unvalidated stated confidence as a standalone deferral signal and support population-specific validation before deployment. Code is available at this https URL . 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); Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2611.11833 — 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 2026-11-10T18:37:37.663Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).