
CaBaFL, a novel asynchronous federated learning approach, employs a hierarchical cache-based aggregation mechanism and a feature balance-guided device selection strategy to address the challenges of stragglers and data imbalance in federated learning.


coremsg

asynchronous-federated-learning-with-hierarchical-cache-and-feature-balance-for-efficient-and-accurate-model-training


Asynchronous Federated Learning with Hierarchical Cache and Feature Balance for Efficient and Accurate Model Training
