Distributed and Multi-Task Learning at the Edge for Energy Efficient Radio Access Networks

The big data availability of Radio Access Network (RAN) statistics suggests using it for improving the network management through machine learning based Self Organized (SON) functionalities. However, this may increase already high energy consumption mobile networks. Multi-access Edge Computing can mitigate problem; however, solutions have to be properly designed efficiently working in a distributed fashion. In work, we propose architectures two RAN SON functionalities on multi-task and gossip learning. We evaluate their accuracy consumed realistic scenarios. Results show that proposed implementations same performance but save with respect correspondent centralized versions benchmark solutions. conclude paper discussing open research issues interesting emerging field.