Scalable and Robust Online Learning for AI-powered Networked Systems

In today's world of pervasive connectivity and intelligent technologies, modern networked systems-ranging from sprawling data centers to large-scale Internet of Things (IoT) systems-have grown by leaps and bounds, unlocking numerous transformative services, like smart cities, immersive mixed reality, and generative artificial intelligence (AI). Traditional network optimization and resource allocation methods-built around static models-are increasingly unable to keep up with the evolving demands of these largescale environments. AI-driven solutions are emerging as game-changers, enabling networked systems to process massive streams of data in real time and adapt seamlessly to ever-changing demands. For instance, AI-powered softwaredefined networking (SDN) controllers can instantly reroute traffic in data centers to prevent congestion, while reinforcement learning (RL)-based task offloading can intelligently allocate resources in massive IoT systems. These advancements make AI central to the next generation of intelligent, adaptive network ecosystems.