Federated Learning in Cellular Wireless Networks

Join us for the fourth session of the exciting webinar series on “New Frontiers in Signal Processing in 6G Wireless Networks”, a collaboration between IEEE Signal Processing, IEEE Communications Society chapters in Ottawa, and IEEE ComSoC Young <a href="http://Professionals.

Register” target=”_blank” title=”Professionals.

Register”>Professionals.

Register now and stay tuned for updates on upcoming speakers and topics!

Speaker(s): Prof. Ekram Hossain

Agenda:
Federated learning (FL) is a distributed machine learning setting where a centralized
server trains a learning model by using remote devices. FL algorithms cannot be employed in
wireless networks unless the unreliable and resource-constrained nature of the wireless medium
is taken into account. In this talk, I shall present an FL algorithm that is suitable for cellular
wireless networks in a real-world scenario. I shall discuss its convergence properties and the
effects of local computation steps and communication steps on its convergence. Through
experiments on real and synthetic datasets, I shall demonstrate the performance of the proposed
algorithm. Also, I will present several applications of FL in wireless communications <a href="http://scenarios.

Virtual:” target=”_blank” title=”scenarios.

Virtual:”>scenarios.

Virtual: https://events.vtools.ieee.org/m/507457