一种联邦学习模型训练方法及系统
By leveraging the collaborative work of centralized and distributed nodes and employing a parallel approach to acquiring and updating the loss function, the problems of slow training and convergence speeds in federated learning are addressed, resulting in faster training speeds and higher utilization of computational resources.
CN116484944BActive Publication Date: 2026-07-17BEIJING TOPSEC NETWORK SECURITY TECH +2
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- BEIJING TOPSEC NETWORK SECURITY TECH
- Filing Date
- 2023-04-26
- Publication Date
- 2026-07-17
Smart Images
Figure CN116484944B_ABST
Abstract
本申请提供一种联邦学习模型训练方法及系统，该方法包括：集中式节点获取当前训练轮次对应的全局参数；集中式节点将全局参数下发至分布式节点；分布式节点根据全局参数和预设训练时间对本地模型进行训练，并在对本地模型进行训练时并行获取本地模型新损失函数；分布式节点根据本地模型新损失函数计算当前训练轮次对应的本地模型新参数；分布式节点将当前训练轮次对应的本地模型新参数发送至集中式节点。可见，该方法及系统能够提高全局模型的训练速度和收敛速度，优化全局模型的性能表现。
Need to check novelty before this filing date? Find Prior Art