arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.01081cs.CV
收藏

MedAD-R1: 通过一致性强化策略优化实现可解释医学异常检测中的一致推理

MedAD-R1: Consistency-Reinforced Policy Optimization for Interpretable Medical Anomaly Detection

School of Informatics, Xiamen University(厦门大学信息学院);Institute of Artificial Intelligence, Xiamen University(厦门大学人工智能研究院);Zhejiang Expressway Co., Ltd.(浙江高速公路有限公司);School of Science and Engineering, Chinese University of Hong Kong(香港中文大学科学与工程学院);Shanghai Artificial Intelligence Laboratory(上海人工智能实验室)

Haitao Zhang, Yingying Wang, Jiaxiang Wang, Haote Xu, Hongyang Zhang, Yirong Chen, Yue Huang, Xinghao Ding

AI总结:

MedAD-R1通过一致性强化策略优化实现医学异常检测中的可解释推理,提升模型的可信度和可解释性。

摘要:

医学异常检测(MedAD)提供了一个显著的机会,利用大型多模态模型(LMMs)来解释和回答基于医学图像的问题,以提高诊断准确性。然而,依赖于监督微调(SFT)在简单且碎片化的数据集上的依赖性阻碍了能够进行合理推理和稳健多模态泛化的模型的发展。为了解决这个问题,我们引入了MedAD-38K,这是第一个大规模、多模态、多中心的MedAD基准,其特征是具有诊断链式思维(CoT)注释的结构化视觉问答(VQA)对。在此基础上,我们提出了一种两阶段训练框架。第一阶段,认知注入,使用SFT来灌输基础医学知识,并使模型与结构化的思考-然后回答范式对齐。鉴于标准策略优化可以产生与最终答案不连贯的推理,第二阶段引入了一种新的算法,即一致性组相对策略优化(Con-GRPO)。这种新算法纳入了一个关键的一致性奖励,以确保生成的推理过程与最终诊断相关且逻辑一致。我们提出的模型MedAD-R1在MedAD-38K基准上实现了最先进的(SOTA)性能,比强大的基线高出超过10%。这种优越的性能源于其能够生成透明且逻辑一致的推理路径,为提高AI在临床决策支持中的可信度和可解释性提供了有前景的方法。

英文摘要:

Medical Anomaly Detection (MedAD) offers a promising direction for medical image analysis with Large Multimodal Models (LMMs). However, progress is limited by fragmented datasets and the tendency of Supervised Fine-Tuning (SFT) to learn superficial image-text correlations rather than verifiable diagnostic reasoning. Consequently, current models often generate fluent explanations that are either insufficiently grounded in the image or inconsistent with their final answers, limiting their reliability in high-stakes medical applications. To address these issues, we introduce MedAD-38K, a large-scale, multimodal, and multicenter benchmark containing structured Visual Question Answering pairs and quality-controlled diagnostic Chain-of-Thought annotations across five core MedAD tasks. Based on this benchmark, we propose a two-stage framework. Cognitive Injection first uses SFT to inject domain-specific medical knowledge and establish a structured think-then-answer format. Consistency Group Relative Policy Optimization (Con-GRPO) then employs an Evidence-Aware Consistency Reward to reinforce reasoning that remains grounded in the image and logically supports the final answer. The resulting MedAD-R1 achieves state-of-the-art performance on MedAD-38K and consistently outperforms all evaluated baselines across five source-disjoint external datasets spanning diverse imaging modalities. Beyond accuracy, it achieves higher reasoning-answer consistency and visual-grounding scores across multiple backbones. With only 0.8B parameters, MedAD-R1 offers practical potential for resource-constrained deployment. These results demonstrate the generality of evidence-aware consistency optimization for interpretable MedAD. Project resources are available at https://github.com/zhtstar/MedAD-R1.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑