NeFut Logo NeFut
Admin Login

[CS.AI] VeriDx: Earning the Right to Diagnose with Disease-Centric Verification

Published at: 2026-09-15 22:00 Last updated: 2026-09-16 00:22
#Machine Learning #LLM #Artificial Intelligence

In clinical reasoning, every disease hypothesis generates obligations: key evidence must be examined, alternatives ruled out, contradictions resolved, useful tests considered, and diagnostic closure justified. Current evaluations of medical LLMs focus on final answers, local steps, or isolated facts, thus missing these hypothesis‑driven commitments.

We introduce VeriDx, a disease‑centric verification framework that maps free‑form diagnostic reasoning to structured disease profiles. VeriDx tracks each hypothesis’s obligations, labeling them as satisfied, unresolved, or violated. This exposes failures such as missing critical tests, unresolved differentials, ignored contradictions, unsupported claims, and premature closure.

We instantiate VeriDx for complex respiratory diagnosis using guideline‑derived disease profiles and expert‑annotated longitudinal cases. Results show that many diagnostic errors are not isolated mistakes but broken commitments made earlier in the reasoning process.

Review

Original Source: https://arxiv.org/abs/2609.14018

[h] Back to Home