Boston Dynamics announced a designed partnership on Wednesday Wednesday to provide improved reinforced learning to the electric Atlas Humanoid Robot. This is previously partnered with Robotics & AI Institute (Rai Institute), which was previously known as Boston Dynamics AI Institute.
Both organizations were established by MARC RAIBERT, former MIT professor who worked as CEO of Boston Dynamics for 30 years. Founded in 2022, the Institute allows Raibert to continue research as the basis of Boston Dynamics.
Both have a relationship with modern. Korean automakers acquired Boston epidemiology in 2021. Hyundai also provides funds to the institute so that RAIBERT can explore more experimental and bleeding skills than possible in commercial companies. The institute focused on the use of large behavioral models in October reflecting the production of Toyota’s TOSTA or TOYOTA Research Institute, which published its own partnership in October. (LBMS).
The twin partnerships are designed to improve the way Boston Dynamics’s Electric Atlas Humanoid learns new tasks. Robotics & AI Institute Deal focuses on reinforcement learning, which is a way to work through trial and error, especially with human and animal learning methods. The reinforcement learning was traditionally intensive, but due to the creation of effective simulation, many processes could be performed at once in virtual settings.
Boston Dynamics/Rai Institute Union started in Massachusetts earlier this month. It is the latest in various collaboration between the pairs, including the joint efforts to develop a reinforced learning research kit for the fourth branch robot (familiar robot “dog”) of Boston Dynamics. The new work focuses on transmitting simulation -based learning to the real environment, and focuses on improving the way the company’s Humanoid Atlas interacts with the physical environment.
Regarding the latter, Boston dynamics points out “dynamic running and systemic operation of heavy objects.” Both are examples of movements that require synchronization of legs and arms. Humanoid’s double form factors offer a variety of unique challenges and opportunities compared to the seats. All activities are also the object of extensive force, including balance, force, resistance and exercise.
Raibert said, “Our goal of RAI is to develop technology that enables future generations of intelligent machines. Working in Atlas with Boston Dynamics allows you to develop the development of reinforcement learning for the most sophisticated humanoid robots. This work will not only play an important role in developing the ability of human noide by expanding technology, but also play an important role in simplifying processes to achieve new technologies. ”
The news of the partnership was the founder of the figure AI and the CEO Brett ADCOCK reached the day after the announcement that Bay Area Company was giving up its partnership with Openai in favor of developing its own in -house model.
“We knew we had to integrate the robot AI vertically to solve the embodied AI in the real world,” TechCrunch said. “We cannot outsourcing AI for the same reason as we can’t outsourcing hardware.”
The painting decides that the AI model, which is most suitable for humanoids, is a particularly developed model for in -house robots. The approach to the specified intelligence of Openai (meaning AI in physical form) was incomprehensible given the size and scope of the Chatgpt producer. The news also arrived on a rumor that Openai is exploring the creation of its own humanoid robot.
Most companies related to the humanoid space are working on their own AI models. This is definitely applied to Boston Dynamics, which has developed software to run on its own robot system. RAI Institute is a technically separate organization, but it shares a general goal.
