With burgeoning technology, numerous intelligent Internet of Things (IoT) applications are exploited by federated learning that admits Artificial Intelligence (AI) training at distributed IoT devices without data exchanging. Owing to the unprecedented approach, federated learning extend powerful advances in the field of IoT, such as data privacy improvements, low-latency network communication, and enhanced learning quality.
Integration of IoT with federated learning comprises diverse applications, including smart healthcare: Electronic Health Records (EHRs) and healthcare cooperation, smart transportation: vehicular traffic planning and resource management, Unmanned Aerial Vehicles (UAVs): communication and network management, smart city: data management and smart grid, smart industry: robotics and industrial edge-based IoT, smart surveillance, smart banking, and augmented reality.
Many surveys on federated learning-based IoT examine the state of the art concepts and services, implementation areas, research challenges, and future scopes. Sparsification, data heterogeneity, mobility, quantization, interference, adaptive resource allocation, communication, and learning convergence are some of the unsolved research problems needed to be contended for enabling federated learning over IoT networks. Hence, the future development of federated learning requires more research efforts towards realizing IoT networks.