Focus on data-efficient learning for networked systems: contextual bandits, multi-task and federated learning, and decentralized optimization. Methods support communications sustainability and cyber defense through information-flow tracking, stochastic modeling, and rigorous performance guarantees.