Boston Dynamics Wednesday announced a partnership Developed to bring an improved increase in reinforcement Electric Atlas Humanoid robot. The connection is with that Robotics & AI Institute (RAI Institute)Earlier known as Boston Dynamics Ai Institute.
Both organizations were founded by Marc Raibert, a former professor with a former professor who acted as CEO of Boston Dynamics for 30 years. The institute, founded in 2022, enables Raibert to continue research, which served as the basis for the dynamics of Boston.
Both have connections to Hyundai. The Korean automobile manufacturer acquired the Boston dynamics in 2021; Hyundai also finances the institute and gives Raibert free relief to examine experimental and bleeding technologies than is possible in a commercial company. The institute reflects Toyota’s creation of Tri or Toyota Research Institute announced its own partnership With Boston Dynamics in October, the use of large behavior models focused. (LBMS).
The Twin partnerships should improve the way Boston Dynamics’ Electric Atlas Humanoid learns new tasks. The Deal Robotics & AI Institute focuses specifically on the learning to reinforcement, a method that works through experiments and errors, similar to the way both humans and animals learn. Learning for reinforcement traditionally was extremely time -consuming, although the creation of an effective simulation made it possible for many processes to be carried out immediately in a virtual environment.
The Boston Dynamics/Rai Institute Union started in Massachusett in the early this month. It is the latest in a series of cooperation between the couple, including a joint effort to develop a research research kit for reinforcements, for the Quadrupedal Spot Robot from Boston Dynamics (the familiar robot “dog”). The new work focuses on the transfer of simulation -based learning into real attitudes as well as on the improvement of the environments of the company through the company’s humanoid atlas.
With regard to the latter, Boston Dynamics indicates “dynamic running and full body manipulation of heavy objects”. Both are examples of actions that require the synchronization of the legs and arms. The Bipedal form factor of the humanoid places a number of unique challenges -and opportunities -compared to the spot. Every activity is also subject to a wide range of forces, including balance, strength, resistance and movement.
Raisbert states in an explanation: “Our goal at Rai is to develop technology that enables future generations of intelligent machines. Through working on Atlas with Boston Dynamics, we can make progress when learning reinforcement on the most sophisticated humanoid robot. This work will play a crucial role in the further development of the skills of humanoids, not only by expanding the skills, but also by tightening the process in order to achieve new skills. “
The news of the partnership comes one day after the AI founder and CEO of Illustration AI, Brett Adcock announced That the Bay Area Company gives up a partnership with Openaai to develop its own internal models.
“We have found that you have to solve the size of the size in the real world and have to integrate the robot -KI vertically,” the managing director told Techcrunch. “For the same reason, we cannot outsource AI why we cannot outsource our hardware.”
Illustration showed that the best AI models for his humanoid are those who are specially developed for its robots. Openais approach to embodied intelligence – which means AI in physical form – was understandably less concentrated in view of the size and scope of the chatt maker. This news also came after rumors that Openai researched the creation of his creation Own humanoid robots.
Most companies involved in the humanoid space work on their own tailor -made AI models. This certainly applies to Boston Dynamics, which offers decades of experience in the development of software in order to operate on its own unique robotics systems. While the RAI institute is technically a separate organization, both a parent company, founder and presumably common goals share.
