Posted on June 15, 2017 by New Scientist - NewsDeepMind now learns from human preferences – just like a toddler A new AI system can learn how to perform complex tasks with a little nudge here and there from a human onlooker
