FedAgg: Adaptive Federated Learning with Aggregated Gradients

no code implementations • 28 Mar 2023 • Wenhao Yuan, Xuehe Wang

To surmount the obstacle that acquiring other clients' local information, we introduce the mean-field approach by leveraging two mean-field terms to approximately estimate the average local parameters and gradients over time in a manner that precludes the need for local information exchange among clients and design the decentralized adaptive learning rate for each client.

Federated Learning

Adaptive Federated Learning via New Entropy Approach

no code implementations • 27 Mar 2023 • Shensheng Zheng, Wenhao Yuan, Xuehe Wang, Lingjie Duan

In this paper, by leveraging entropy as a new metric for assessing the degree of system disorder, we propose an adaptive FEDerated learning algorithm based on ENTropy theory (FedEnt) to alleviate the parameter deviation among heterogeneous clients and achieve fast convergence.

Federated Learning