Engineers at the Toyota Research Institute (TRI) are using generative AI techniques to teach robots new behaviors, working toward their goal of building Large Behavior Models (LBMs). These robots successfully learned 60 difficult, dexterous skills, including pouring liquids, using tools, and manipulating deformable objects.
This proves that robots can be taught to function in new scenarios and perform a wide range of behaviors using Diffusion Policy, without being limited to just simply picking up objects or putting them down in new locations. The generative-AI approach to behavior learning, developed in collaboration with Columbia University, allows for easy and rapid behavior teaching from demonstration. TRI hopes that robots will be able to one day support people in everyday situations and unpredictable, ever-changing environments.
- Authentically detailed LEGO replica of the legendary Toyota GR Supra, plus racing driver minifigure
- With its wider 8-stud chassis, the collectible model car toy has room for 2 minifigures and lots of authentic detailing
- Driver minifigure has a Toyota racing suit, safety helmet and wrench, providing plenty of inspiration for imaginative, high-speed play
Our research in robotics is aimed at amplifying people rather than replacing them. This new teaching technique is both very efficient and produces very high performing behaviors, enabling robots to much more effectively amplify people in many ways,” said Gill Pratt, CEO of TRI and Chief Scientist for Toyota Motor Corporation.
