Applicability of Deep Reinforcement Learning for Efficient Federated Learning in Massive IoT Communications

To build intelligent model learning in conventional architecture, the local data are required to be transmitted toward cloud server, which causes heavy backhaul congestion, leakage of personalization, and insufficient use network resources. address these issues, federated (FL) is introduced by offering a systematical framework that converges distributed modeling process between participants parameter server. However, challenging issues participant scheduling, aggregation policies, offloading, resource management still remain within FL architecture. In this survey article, state-of-the-art solutions for optimizing orchestration communications presented, primarily querying deep reinforcement (DRL)-based autonomy approaches. The correlations DRL mechanisms described optimized system architectures selected literature observable states, configurable actions, target rewards inquired into illustrate applicability DRL-assisted control self-organizing systems. Various deployment strategies Internet Things applications discussed. Furthermore, article offers review challenges future research perspectives advancing practical performances. Advanced aspects will drive converged autonomous communication-efficient privacy-aware learning.