Embodied AI at MIT CSAIL's Living Lab: Simulation for Safer Robots

Description

At the CSAIL Living Lab, researchers are developing reliable robotic AI through an iterative process of refining sensors and hardware, building machine learning models, and using simulation training before testing on real machines. The ultimate goal is robust, generalizable AI ready for real-world deployment. Hear from CSAIL Professor Leslie Kaelbling as well as Research Scientist Joseph DelPreto and graduate students Francisco Ramirez Serrano, Jared Boyer, and Wil Norton. In this video, you will learn: Why simulation training is critical for teaching robots to fail safely and improve faster How simulation may unlock emergent, generalizable robot behavior Why general intelligence is the ultimate goal for real-world deployment Visit our website to learn more about the Living Lab and other MIT CSAIL research: https://cap.csail.mit.edu/