跳到正文
arXiv cs.AI· Runxun Zhang, Onofrio Antonio Catalano, Jinqiao Wang, Jingwei Wei·· 4 小时前AI 评分34

并非所有证据都可采信:工具增强医疗智能体的证据治理框架

Not All Evidence Is Admissible: Evidence Governance for Tool-Augmented Medical Agents

AI 导读

研究提出 context-independent evidence governance(CIEG),将“判定哪些工具输出可作为证据”从推理模型中分离出来,以解决工具输出被默认采信导致的“证据劫持”问题。

正文 · 原文

View PDF HTML (experimental)

Abstract:Tool-augmented medical agents treat whatever their tools return as evidence. We show that this default is costly: medical tools fail silently, and a frozen reasoner cannot ignore what enters its context, so many cases it would have answered correctly are flipped to wrong once tool outputs are added. We call this evidence hijacking. $\mathcal{V}$-information explains the phenomenon: the reasoner has lost the option to ignore, so a tool output can carry negative usable information, and moving the decision of what to keep outside the reasoner restores that option. The theory further shows that part of any filtering gain comes merely from letting the reasoner see less, and only selection by content lets it see better. Building on this, we recast tool-augmented reasoning as evidence governance, which separates deciding what counts as evidence from the reasoner that answers, and propose context-independent evidence governance (CIEG), the first medical agent framework that rules, after acquisition, on whether each individual tool output may enter as evidence. Every tool output is written as a card that can be checked against the raw input; an independent checker that sees only the input and the cards decides which cards become evidence; and the reasoner answers from the admitted cards alone, with no training anywhere. On medical question answering, fundus glaucoma screening, and chest X-ray pneumonia classification, it achieves the highest accuracy among models of comparable scale, and significantly outperforms state-of-the-art tool-using agents and multi-agent systems built on the same backbone at lower cost.
Comments: 11 pages
Subjects: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
Cite as: arXiv:2611.22169 [cs.AI]
  (or arXiv:2611.22169v1 [cs.AI] for this version)
  https://doi.org/10.48550/arXiv.2611.22169

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Runxun Zhang [view email]
[v1] Tue, 17 Nov 2026 04:01:28 UTC (491 KB)

来源:arXiv cs.AI · arxiv.org