Federated Learning for Edge AI in IoT-Enabled Smart Computing Environments

Abstract

The rapid expansion of Internet of Things (IoT)-enabled smart computing environments has led to an unprecedented increase in distributed data generation, necessitating efficient and privacy-preserving learning paradigms. Federated Learning (FL) has emerged as a promising decentralized machine learning approach that enables collaborative model training across edge devices without exposing sensitive data. This paper explores the integration of Federated Learning with Edge AI to enhance real-time decision-making, reduce latency, and improve security in IoT-driven smart systems. The proposed framework leverages heterogeneous edge devices, communication-efficient optimization techniques, and personalized FL strategies to address key challenges such as data heterogeneity, communication constraints, and adversarial security threats. Experimental results demonstrate that FL-based Edge AI significantly enhances learning accuracy, reduces energy consumption, and improves model adaptability in IoT ecosystems. Comparative analysis with traditional centralized and distributed learning paradigms highlights the efficiency and scalability of the proposed approach. This study provides valuable insights into the future of privacy-preserving, AI-driven IoT applications and paves the way for robust and scalable smart computing environments.