I was going to ask how this is different than a Reinforcement Learning algorithm but then they called out Deep Minds Alpha-Go