arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2402.09677cs.CV
收藏

基于提示的个性化联邦学习用于医学视觉问答

Prompt-based Personalized Federated Learning for Medical Visual Question Answering

Graduate School of Information Science and Technology, Hokkaido University(北海道大学信息科学与技术研究生院);Faculty of Information Science and Technology, Hokkaido University(北海道大学信息科学与技术学院)

He Zhu, Ren Togo, Takahiro Ogawa, Miki Haseyama

AI总结:

提出一种基于提示的个性化联邦学习方法,通过小型可学习参数共享信息和引入可靠性参数,有效解决了医学视觉问答中的数据异构性、隐私保护及通信复杂度问题。

摘要:

我们提出了一种新颖的基于提示的个性化联邦学习(pFL)方法,以解决传统医学视觉问答(VQA)方法中的数据异构性和隐私问题。具体而言,我们将来自不同器官的医学数据集视为客户端,并使用pFL为每个客户端训练个性化的基于Transformer的VQA模型。为了解决先前pFL方法中客户端间通信的高计算复杂度问题,我们通过引入作为小型可学习参数的提示,提出了一种简洁的信息共享系统。此外,所提出的方法引入了可靠性参数,以防止低性能和不相关客户端的负面影响。最后,在各种异构医学数据集上的广泛评估证明了我们提出方法的有效性。

英文摘要:

We present a novel prompt-based personalized federated learning (pFL) method to address data heterogeneity and privacy concerns in traditional medical visual question answering (VQA) methods. Specifically, we regard medical datasets from different organs as clients and use pFL to train personalized transformer-based VQA models for each client. To address the high computational complexity of client-to-client communication in previous pFL methods, we propose a succinct information sharing system by introducing prompts that are small learnable parameters. In addition, the proposed method introduces a reliability parameter to prevent the negative effects of low performance and irrelevant clients. Finally, extensive evaluations on various heterogeneous medical datasets attest to the effectiveness of our proposed method.

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。

↑