AI RESEARCH
Dialectic-Med: Mitigating Diagnostic Hallucinations via Counterfactual Adversarial Multi-Agent Debate
arXiv CS.CL
•
ArXi:2604.11258v1 Announce Type: new Multimodal Large Language Models (MLLMs) in healthcare suffer from severe confirmation bias, often hallucinating visual details to initial, potentially erroneous diagnostic hypotheses. Existing Chain-of-Thought (CoT) approaches lack intrinsic correction mechanisms, rendering them vulnerable to error propagation. To bridge this gap, we propose Dialectic-Med, a multi-agent framework that enforces diagnostic rigor through adversarial dialectics.