Direct Preference Optimization for Suppressing Hallucinated Prior Exams in Radiology Report Generation
This addresses a critical issue for clinicians and patients by reducing harmful hallucinations in medical AI, though it is an incremental application of an existing method to a new domain.
The paper tackled the problem of hallucinations in generative vision-language models for radiology report generation, specifically suppressing prior exam hallucinations, and achieved a 3.2-4.8x reduction in such hallucinations while maintaining clinical accuracy.
Recent advances in generative vision-language models (VLMs) have exciting potential implications for AI in radiology, yet VLMs are also known to produce hallucinations, nonsensical text, and other unwanted behaviors that can waste clinicians' time and cause patient harm. Drawing on recent work on direct preference optimization (DPO), we propose a simple method for modifying the behavior of pretrained VLMs performing radiology report generation by suppressing unwanted types of generations. We apply our method to the prevention of hallucinations of prior exams, addressing a long-established problem behavior in models performing chest X-ray report generation. Across our experiments, we find that DPO fine-tuning achieves a 3.2-4.8x reduction in lines hallucinating prior exams while maintaining model performance on clinical accuracy metrics. Our work is, to the best of our knowledge, the first work to apply DPO to medical VLMs, providing a data- and compute- efficient way to suppress problem behaviors while maintaining overall clinical accuracy.