Physical Intelligence: Science and Systems

Catalog description

This course provides an in-depth exploration of advanced topics in robot learning, an area of artificial intelligence that focuses on the agentic aspects of real-world decision making. Students will learn what are the challenges in applying foundational concepts in imitation and reinforcement learning, including policy optimization, value-based methods, model-based RL, and exploration-exploitation trade-offs, when applied to real robots. In the final part of the course, we will cover recent advancements in the field, with a particular focus on learning from high-dimensional sensory observations, e.g., vision, sound, or touch. Alongside a theoretical understanding of the foundations, students will gain experience with practical implementations of robot learning algorithms via assignments and a final project. Recommended prerequisite: ESE 6500