Efficient distributed learning in 5G Fog Computing Environments

The ability of performing tasks of distributed knowledge extraction fromdata collected and stored at, and in perspective, beyond the Edge of the network inFederated settings is considered one of the cornerstones of the future evolutions ofapplications and systems based on 5G. One of the main enabling factors, amongothers, is the pervasive diffusion of personal mobile and IoT devices, that collect andpossibly process the data not only through cloud facilities but also directly at the Edgeof the network, coupled with the pervasiveness and high speed connectivity of 5Gnetworks. Coupling wireless 5G networks and distributed AI is a fascinating researchtopic, which poses many challenges also in terms of wise use of network and devices'resources. In this chapter, we discuss the challenges connected to optimizing networkresources during a distributed learning process, and we present selected results onthe topic.