Scientists from University College London, Intrinsic, and Google’s DeepMind released a new algorithm this week called RoboBallet, designed to help teams of robotic arms work together, in harmony, in obstacle filled spaces. (Like, you know, a factory floor). Currently, these kinds of complex training procedures tend to require painstaking, custom solutions for each new kind of task, and if you don’t get it right, your robot arms slam into one another. RoboBallet could one day use a neural network trained with reinforcement learning to simplify and standardize this process. https://www.ft.com/conten.. #Robotics Revolution 🦾🤖🦿