Marcello Restelli is a Professor at the Department of Electronics, Information, and Bioengineering (DEIB) of Politecnico di Milano, affiliated with the School of Industrial and Information Engineering. His research focuses on Reinforcement Learning (RL), Machine Learning, Robotics, and Control Systems. He explores challenges in real-world RL applications, multi-agent systems, and optimization under uncertainty. Key contributions include developing scalable algorithms for sparse-reward environments, POMDPs, and safe robotics. He leads projects like the EU-funded I3LUNG initiative, integrating AI for personalized lung cancer care. His work bridges theory and practice, addressing topics such as exploration strategies, regret minimization, and causal feature selection. He has contributed to frameworks like Gym4ReaL for benchmarking real-world RL systems and Storehouse for warehouse optimization. Recent efforts emphasize energy systems, financial decision-making, and explainability in extreme event prediction. His publications highlight advancements in inverse RL, online learning, and transfer learning. Collaborations span academia and industry, reflecting his impact on both theoretical and applied AI domains.
