Communication-Efficient Federated Deep Learning With Layerwise Asynchronous Model Update and Temporally Weighted Aggregation | IEEE Journals & Magazine | IEEE Xplore