NeRD: Neuro-Symbolic Rule Distillation for Efficient Ontology-Grounded Chain-of-Thought in Medical Image Diagnosis
Researchers have introduced NeRD (Neuro-Symbolic Rule Distillation), a novel framework designed to enhance interpretability and efficiency in medical image diagnosis. NeRD addresses limitations in existing methods by generating concise, ontology-grounded reasoning chains without requiring manual rule crafting. Experiments on skin datasets show that NeRD achieves strong diagnostic performance and clinical plausibility, enabling effective concept-level intervention through expert-in-the-loop studies. AI
IMPACT Enhances AI interpretability in medical diagnosis by providing efficient and clinically plausible reasoning chains.