论文标题
最佳的MIMO合并用于盲目联合边缘学习与梯度稀疏
Optimal MIMO Combining for Blind Federated Edge Learning with Gradient Sparsification
论文作者
论文摘要
我们为多输入多输出（MIMO）系统中联合学习提供了最佳的接收组合策略。我们提出的算法允许客户执行单个梯度稀疏，从而在具有异质（非I.I.D.）培训数据的情况下大大提高了性能。所提出的方法击败了基准。
We provide the optimal receive combining strategy for federated learning in multiple-input multiple-output (MIMO) systems. Our proposed algorithm allows the clients to perform individual gradient sparsification which greatly improves performance in scenarios with heterogeneous (non i.i.d.) training data. The proposed method beats the benchmark by a wide margin.
