ICML
2010
IEEE
14 years 19 days ago
2010
IEEE
We consider approximate policy evaluation for finite state and action Markov decision processes (MDP) in the off-policy learning context and with the simulation-based least square...