Advancing Robotic Arm Training and Human Cell Study with AI and Transformer Models

Graduate students at Carnegie Mellon University's Mechanical Engineering Department have developed a hands-off approach to training robotic arms using artificial intelligence. By recreating simple tasks in a virtual reality simulation and augmenting human-like examples of movements, the researchers were able to significantly decrease the robot's learning time. The method has the potential to produce promising results with minimal human input, and future research aims to teach robots how to interact with malleable materials.