Reinforcement Learning Generalization by Dr Mirco Mutti

Abstract

In the last few years, Reinforcement Learning (RL) has played a central role to in turning foundation models into capable conversational agents and robotic manipulation systems. Despite these successes, an open challenge remains: How can RL enable artificial agents to generalize to a broad range of tasks, environments, and conditions beyond their training distributions? In this talk, I will explore this question through the lens of my research, presenting fundamental pillars, recent advancements, and open problems in RL generalization.

Biography

Mirco Mutti is a postdoctoral researcher at the Technion, working with Aviv Tamar in the Robots Learning Lab. Formerly, he obtained his PhD at Politecnico di Milano, jointly with Università di Bologna, under the supervision of Marcello Restelli. His research focuses on the theoretical and methodological foundations of reinforcement learning, including unsupervised exploration, partial observability, and generalization.