Distilled One-Shot Federated Learning

Yanlin Zhou, George Pu, Xiyao Ma +2

Current federated learning algorithms take tens of communication rounds transmitting unwieldy model weights under ideal circumstances and hundreds when data is poorly distributed.…