Abstract
To meet the diverse requirements for intelligent autonomous IoT systems and ensure privacy, the concept of Federated Learning (FL) was proposed and used effectively. An IoT using computing edge technologies, security related technologies and machine learning, to enable smart autonomous systems is envisioned. So, FL has provided a platform in many of these applications to protect the data and reduce latency. These smart services and applications rely on efficient computation and communication resources. But being resource constrained, such IoT devices have introduced a number of new challenges. In this talk, we showcase our research activities to mitigate some of these challenges and contribute to such efforts to advocate possible solutions using these models. Then, we conclude with a list of some research directions along the same lines.