Optimizing training delay of federated learning with reinforcement learning-based pruning
