Toyota Research Institute (TRI) reveals a new AI technique, named Diffusion Policy, that efficiently teaches robots dexterous skills, moving closer to building Large Behavior Models (LBMs) for robots. This new method improves robot utility and allows robots to perform complex tasks more efficiently.
TRI has already taught robots over 60 complex skills using Diffusion Policy, without writing any new code. The institute aims to teach hundreds of new skills by the end of the year and 1,000 by the end of 2024, proving that robots can perform a wide range of behaviors and function in new scenarios.