Improved Regret Bounds for Undiscounted Continuous Reinforcement Learning.

K. Lakshmanan, Ronald Ortner, Daniil Ryabko

Browse the full ICML paper archive.