Pathway to Robot Intelligence: Learning Complex Sensorimotor Skills

Explore a 34-minute lecture on developing complex and generalizable sensorimotor skills for robotic agents, presented by Pulkit Agrawal from MIT at the Simons Institute. Delve into the challenges of executing accurate and reliable sensorimotor skills in robotics, such as locomotion, door opening, and object manipulation. Learn about a framework for scaling the learning of new and complex tasks, bridging the gap between natural language processing capabilities and physical task execution in robotic systems. Gain insights into the intersection of AI, psychology, and neuroscience perspectives in understanding lower-level intelligence and its application to robotics.