Chapter 17: Federated Learning over Wireless Channels

Learning Objectives

• Integrate AirComp aggregation (Chapter 16) into an end-to-end federated-learning pipeline
• Derive the wireless-FL convergence rate as a function of per-round MSE and round count
• Design device scheduling and resource allocation rules balancing fairness, energy, and convergence
• Compare digital and analog aggregation on the same FL workload and identify when each dominates
• Recognize the CommIT contribution on information-theoretically secure federated representation learning