摘抄
Federated learning (aka collaborative learning) is a machine learning technique that trains an algorithm across multiple decentralized edge devices or servers holding local data samples, without exchanging their data samples.
不同于：
传统centralized ML
