Current research and prospects of federated language large models in the medical field

Yumeng Tang; Yuanpeng Deng

doi:10.1117/12.3036891

13 July 2024 Current research and prospects of federated language large models in the medical field

Yumeng Tang, Yuanpeng Deng

Proceedings Volume 13208, Third International Conference on Biomedical and Intelligent Systems (IC-BIS 2024); 1320838 (2024) https://doi.org/10.1117/12.3036891
Event: 3rd International Conference on Biomedical and Intelligent Systems (IC-BIS 2024), 2024, Nanchang, China

Abstract

This paper delves into the current state and future prospects of federated learning (FL) in the medical field, with a focus on its key advantages in privacy protection, data efficiency, and model performance improvement. The main innovations of the paper lie in the introduction of three strategies: federated pre-training, fine-tuning, and prompt engineering, aimed at balancing data privacy and model performance optimization. Additionally, the paper discusses the major technical challenges faced by federated learning, including security threats, privacy breaches, and issues with non-independent and identically distributed (Non-IID) data, proposing the use of differential privacy and secure multi-party computation techniques to enhance data privacy and model security. Finally, the paper outlines future research directions, including designing more comprehensive security and privacy protection mechanisms for federated learning, optimizing communication efficiency, and developing more effective algorithms for Non-IID data, all of which are crucial factors in promoting the widespread application of FL in the medical field.

(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Yumeng Tang and Yuanpeng Deng "Current research and prospects of federated language large models in the medical field", Proc. SPIE 13208, Third International Conference on Biomedical and Intelligent Systems (IC-BIS 2024), 1320838 (13 July 2024); https://doi.org/10.1117/12.3036891

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available