Local Linearity: the Key for No-regret Reinforcement Learning in Continuous MDPs

Open in new window