信息工程大学学报2024，Vol.25Issue(4)：404-410,7.DOI:10.3969/j.issn.1671-0673.2024.04.005
分布差异感知的联邦学习方法
Distribution Divergence-Aware Federated Learning
摘要
Abstract
To address the problem of slow convergence and unstable training of the model caused by non-independent identically distribution data in federated learning,the Jensen-Shannon(JS)diver-gence is calculated to evaluate the divergence,the minimization of distribution difference for the feder-ated learning is modeled,and a federated learning method with awareness of the distribution difference is proposed.Numerical experiments are conducted to verify the effectiveness of this method.Experi-ment results show that optimized data distribution can effectively accelerate the training accuracy and make the model converge to a stable state.关键词
联邦学习/数据共享/非独立同分布/分布差异/梯度下降Key words
federated learning/data sharing/non-independent identically distribution/distribution discrepancy分类
信息技术与安全科学引用本文复制引用
胡智尧,于淼..分布差异感知的联邦学习方法[J].信息工程大学学报,2024,25(4):404-410,7.基金项目
国家自然科学基金青年科学基金(62025208) （62025208）
