一作ACL 2026Poster· Natural Language Processing
Beyond Surface Features: Advancing Medical Vision-Language Alignment via Dynamic Evidence-Guided Preference Optimization
Zixuan Huang · Zhihong Zhu · Xiaolong Liu · Yanchao Hao · Manman Zhang · Zheng Wei · Bowen Xing · Xian Wu · Ye Li · Fen Miao · Yefeng Zheng
摘要
Medical large Vision-Language Models (Med-LVLMs) have shown strong potential in multimodal clinical applications such as medical visual question answering and report generation. However, Med-LVLMs remain challenged by hallucinations caused by modality misalignment, where models prioritize textual...