Google's AlphaGo can now teach itself how to win at Go

Researchers from Google's London-based DeepMind have unveiled the new AlphaGo Zero system in a study published in Nature today.