Model Pruning Enables Efficient Federated Learning on Edge DevicesYuang JiangShiqiang Wanget al.2021IEEE TNNLS