“Reinforcement Learning has made significant progress in domains with clearly defined goals and reward functions, such as mathematics and the game of Go, as demonstrated by DeepMind's AlphaGo.”