Active Policy Iteration: Efficient Exploration through Active Learning for Value Function Approximation in Reinforcement Learning.

Takayuki Akiyama, Hirotaka Hachiya, Masashi Sugiyama

Browse the full IJCAI paper archive.