See LionPATH to schedule.
Course Topic: Reinforcement Learning
This graduate special topics course provides an introduction to the theory and practice of reinforcement learning (RL), focusing on sequential decision making and deep learning. Driven by recent transformative applications, ranging from training humanoid robots to fine-tuning Large Language Models, the course surveys both foundational concepts and state-of-the-art techniques through lectures, paper discussion, and project assignments. Topics covered by this course include multi-arm bandits and Markov decision processes; state-of-the-art value-based, policy-based, and actor-critic algorithms; cooperative and adversarial multi-agent RL; intelligent exploration; imitation learning, offline RL, and model-based RL; and example applications of RL, such as LLMs, control problems, and combinatorial optimization.
