1,227 followers
Saturday classical paper series: (reinforcement learning) *learning to predict by methods of temporal differences* - Sutton 1988. Assigning credit by means of the difference between temporally successive predictions. https://t.co/A1xynaYmgX