Adaptive Personalized Federated Learning
This work addresses the need for personalized models in federated learning, offering an incremental improvement over existing methods.
The paper tackles the trade-off between global model performance and local personalization in federated learning by proposing an adaptive personalized federated learning (APFL) algorithm, which finds the optimal mixing parameter and demonstrates effectiveness through extensive experiments.
Investigation of the degree of personalization in federated learning algorithms has shown that only maximizing the performance of the global model will confine the capacity of the local models to personalize. In this paper, we advocate an adaptive personalized federated learning (APFL) algorithm, where each client will train their local models while contributing to the global model. We derive the generalization bound of mixture of local and global models, and find the optimal mixing parameter. We also propose a communication-efficient optimization method to collaboratively learn the personalized models and analyze its convergence in both smooth strongly convex and nonconvex settings. The extensive experiments demonstrate the effectiveness of our personalization schema, as well as the correctness of established generalization theories.