Reinforcement learning (RL) is widely regarded as a promising frontier in artificial intelligence and a significant advancement in machine learning. This view is supported by the highly publicized successes of deep reinforcement learning in domains such as video games and classical strategy games like chess and Go. However, many of these successes rely on heuristic-based algorithmic implementations. While these heuristics often perform remarkably well, their effectiveness and underlying mechanisms remain only partially understood, even by their developers. Moreover, the majority of current RL applications focus on robotic systems, including self-driving cars, and primarily address single-agent control problems. Recently, there has been growing interest in multi-agent reinforcement learning, which involves coordinating systems such as fleets of autonomous vehicles, presenting heightened mathematical and computational challenges. These challenges become even more pronounced in systems involving a large number of agents, where direct modeling and control are computationally prohibitive. In such cases, mean-field approximations have emerged as a powerful tool to scale reinforcement learning algorithms by approximating the interactions among agents through aggregate effects. This workshop will focus on multi-agent RL and mean-field RL. Besides the talks, we will organize one or two tutorials on Markov Decision Processes and Mean Field Games to help students be more familiar with the topic of the workshop. We will also organize a special lecture and a session of lightning talks for the long program attendees.
IMSI is committed to making all of our programs and events inclusive and accessible. Contact [email protected] to request disability-related accommodations.
In order to register for this workshop, you must have an IMSI account and be logged in. Please use one of the buttons below to login or create an account.