75
Voted
JMLR
2010
14 years 3 months ago
2010
We consider reinforcement learning as solving a Markov decision process with unknown transition distribution. Based on interaction with the environment, an estimate of the transit...