federated learning is a communication-efficient training process that
alternates between local training at the edge devices and averaging the updated
local model at the central server. Nevertheless, it is impractical to achieve a
perfect acquisition of the local models in wireless comm