Federated AI Inference for Scalable and Secure Real-Time IoT Dataflows

The proliferation of IoT devices creates massive real-time data, which creates challenges for centralized AI in terms of latency, scalability and privacy. Federated AI inference supports distributed model training and inference on the edge devices as well as keeping data confidential while using real time analytics. Architectures, Privacy-reserving mechanisms and Edge-cloud collaboration strategies including use cases in healthcare, smart cities, and industrial systems are considered in this chapter and presented with a variety of real-world applications. Important challenges and future trends for secure and scalable federated intelligence are also presented for IoT environments.