61
Voted
ICML
2007
IEEE
15 years 11 months ago
2007
IEEE
This paper investigates the problem of automatically learning how to restructure the reward function of a Markov decision process so as to speed up reinforcement learning. We begi...