Watching artificial intelligence teach itself how to walk is weirdly captivating | Ideas from and for MAKERS | Scoop.it
Researchers from Singapore and British Columbia have used reinforcement learning to create an AI which learns to walk through trial and error.