Work

BGEFL: Enabling Communication-Efficient Federated Learning via Bandit Gradient Estimation in Resource-Constrained Networks