A unified Personalized Federated Learning framework ensuring Domain Generalization

Yuan Liu, Zhe Qu, Shu Wang, Chengchao Shen, Yixiong Liang, Jianxin Wang

Abstract: Highlights•We develop a significant multi-objective problem in federated learning which not only protects personalization but also enables generalization.•We propose a principled federated learning framework to address this problem through a novel loss function with a multi-objective structure.•Our method enhances the applicability of federated learning in real-world scenarios.•Experiments demonstrate that our method outperforms state-of-the-art methods.