Confidence-Evidence Bayesian Gain (CEBaG) provides deterministic hallucination detection for medical VQA without requiring 10-20 stochastic generations.
March 24, 2026
Original Paper
Deterministic Hallucination Detection in Medical VQA via Confidence-Evidence Bayesian Gain
arXiv · 2603.21693
The Takeaway
By analyzing token-level predictive variance and 'evidence magnitude' (visual shift in log-probs), it outperforms sampling-heavy methods while being drastically faster. This makes robust safety-checking feasible for real-time clinical deployment.
From the abstract
Multimodal large language models (MLLMs) have shown strong potential for medical Visual Question Answering (VQA), yet they remain prone to hallucinations, defined as generating responses that contradict the input image, posing serious risks in clinical settings. Current hallucination detection methods, such as Semantic Entropy (SE) and Vision-Amplified Semantic Entropy (VASE), require 10 to 20 stochastic generations per sample together with an external natural language inference model for semant