Metraipolis › archive › dc_a3ea43487ef6

Separating Capability from Confidence: Grounded Dual-State Calibration for GRPO-Trained Medical Vision-Language Models — arXiv:2609.06419

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.06419
License (as archived)
CC0-1.0
Archived
2026-10-16T13:58:58.281Z
SHA-256 of the archived text
0cae94fd71cc6efe59bbbcff3f175dd3d0ca7846461983a8b60b78b92b75d536
Stable URL of this page
https://metraipolis.com/documento/dc_a3ea43487ef6

How to cite

Separating Capability from Confidence: Grounded Dual-State Calibration for GRPO-Trained Medical Vision-Language Models — arXiv:2609.06419 — arXiv (en), https://arxiv.org/abs/2609.06419 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_a3ea43487ef6 (Metraipolis archive dc_a3ea43487ef6, sha256 0cae94fd71cc6efe59bbbcff3f175dd3d0ca7846461983a8b60b78b92b75d536)
BibTeX
@misc{dca3ea43487ef6,
  title = {Separating Capability from Confidence: Grounded Dual-State Calibration for GRPO-Trained Medical Vision-Language Models — arXiv:2609.06419},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.06419},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_a3ea43487ef6; sha256 0cae94fd71cc6efe59bbbcff3f175dd3d0ca7846461983a8b60b78b92b75d536},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.06419. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Medical vision-language models (VLMs) require confidence that reflects both answer correctness and patient-specific visual evidence. Recent GRPO-based methods optimize verbalized confidence together with answer generation. However, this joint optimization may interfere with answer learning and drive confidence toward near-binary values. Verbalized confidence also provides no explicit assessment of visual support. We therefore separate capability learning from confidence estimation and propose \textbf{DualRead}. DualRead builds on the insight that reliability can be read from the actor's internal states at critical moments in the answering process. It freezes the GRPO-trained actor and combines pre-answer solvability with a post-answer assessment of the generated answer and its visual support. To further assess whether confidence reflects visual grounding, we introduce \textbf{Counterfactual Confidence Grounding AUC} (CCG-AUC). It measures whether confidence decreases when real-image substitution changes the actor from correct to incorrect. Across two VLM backbones and both in- and out-of-distribution medical VQA benchmarks, DualRead improves correctness discrimination and calibration over verbalized confidence while preserving answer accuracy. CCG-AUC reveals whether confidence responds to answer-relevant visual evidence rather than primarily to non-visual cues.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-09 in cs.LG).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.06419 — 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 2026-10-16T13:39:43.172Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-09?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).