Metraipolis › archive › dc_dd8a12034b02

Coupled or Independent? Fault Coupling in Agentic Medical Vision Systems — arXiv:2610.40913

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.40913
License (as archived)
CC0-1.0
Archived
2026-11-02T04:50:30.695Z
SHA-256 of the archived text
2aff8ffc78d69cf9ebe921e15952b9e609f65baff0bc2f6b8637b34acfb1a8bd
Stable URL of this page
https://metraipolis.com/documento/dc_dd8a12034b02

How to cite

Coupled or Independent? Fault Coupling in Agentic Medical Vision Systems — arXiv:2610.40913 — arXiv (en), https://arxiv.org/abs/2610.40913 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_dd8a12034b02 (Metraipolis archive dc_dd8a12034b02, sha256 2aff8ffc78d69cf9ebe921e15952b9e609f65baff0bc2f6b8637b34acfb1a8bd)
BibTeX
@misc{dcdd8a12034b02,
  title = {Coupled or Independent? Fault Coupling in Agentic Medical Vision Systems — arXiv:2610.40913},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.40913},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_dd8a12034b02; sha256 2aff8ffc78d69cf9ebe921e15952b9e609f65baff0bc2f6b8637b34acfb1a8bd},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.40913. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Vision-language agents equipped with tools expose intermediate outputs for inspection and comparison with the model's reasoning, which may offer safety benefits over single-shot models. This paper tests that potential benefit when the sensor reading itself is untrusted, so that the tool and reasoning channels share a common visual fault. We consider a dermatology agent based on segmentation, contrastive retrieval, and an iterative vision-language reasoner and define a coupling taxonomy that distinguishes single-channel errors from joint errors that agree or disagree. On PAD-UFES-20, we show that typographic injection produces detectable directional retrieval capture, whereas physical corruption does not. Crop-first retrieval preserves the payload at a readable scale and produces 55% joint wrong agreement across all cases (71% of faulty cases); full-image retrieval produces 35% (43%). Under a binary decision, if both channels are wrong, their predictions must agree, so hard-label channel comparison cannot detect joint failures, limiting disagreement recall to 0.29 (crop-first) versus 0.57 (full-image). The direction of the order effect is reproduced on the independent HIBA test set.

Tema degli esploratori: fiducia digitale e crittografia (elenco arXiv cs.CR, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Cryptography and Security (cs.CR).
Pagina del lavoro: https://arxiv.org/abs/2610.40913 — 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 2026-11-02T04:44:42.683Z dalla pagina https://arxiv.org/list/cs.CR/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).