paper-with-me

Papers

Direct Preference Optimization for Suppressing Hallucinated Prior Exams in Radiology Report Generation

2024-06-10 · Oishi Banerjee, Hong-Yu Zhou, Subathra Adithan, Stephen Kwak, Kay Wu, Pranav Rajpurkar

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.

📄 PDF Abstract BibTeX arXiv:2406.06496

Code (0)

등록된 구현이 없습니다.

Methods 이 논문이 사용한 방법론

DPO 설명 없음

Similar Papers 제목 키워드 기반

CHiP: Cross-modal Hierarchical Direct Preference Optimization for Multimodal LLMs

2025-01-28 · Jinlan Fu, Shenzhen Huangfu, Hao Fei, Xiaoyu Shen 외

Multimodal Large Language Models (MLLMs) still struggle with hallucinations despite their impressive capabilities. Recent studies have attempted to mitigate this by applying Direct Preference Optimization (DPO) to multim…

Hallucination

ISR-DPO: Aligning Large Multimodal Models for Videos by Iterative Self-Retrospective DPO

2024-06-17 · Daechul Ahn, Yura Choi, San Kim, Youngjae Yu 외

Iterative self-improvement, a concept extending beyond personal growth, has found powerful applications in machine learning, particularly in transforming weak models into strong ones. While recent advances in natural lan…

Language ModellingQuestion AnsweringResponse GenerationVideo Question Answering

Mitigating Hallucinations in Multimodal LLMs via Object-aware Preference Optimization

2025-08-27 · Alberto Compagnoni, Davide Caffagni, Nicholas Moratelli, Lorenzo Baraldi 외 arxiv

Multimodal Large Language Models (MLLMs) emerge as a unified interface to address a multitude of tasks, ranging from NLP to computer vision. Despite showcasing state-of-the-art results in many benchmarks, a long-standing…

Image Captioning

HIVE: Understanding Post-Hallucination Reasoning in Vision Language Models

2026-07-08 · Feng He, Zhenting Wang, Qifan Wang, Qiang Guan 외 arxiv

Hallucinations in vision language models (VLMs) are commonly treated as semantic errors, yet they often arise from partial or ambiguous visual evidence. Prior work mainly focuses on detecting or suppressing hallucination…

Multimodal Reasoning

Mitigating Hallucinated Translations in Large Language Models with Hallucination-focused Preference Optimization

2025-01-28 · Zilu Tang, Rajen Chatterjee, Sarthak Garg

Machine Translation (MT) is undergoing a paradigm shift, with systems based on fine-tuned large language models (LLM) becoming increasingly competitive with traditional encoder-decoder models trained specifically for tra…

DecoderHallucinationMachine TranslationTranslation