Geometric Value Iteration: Dynamic Error-Aware KL Regularization for Reinforcement Learning.

Toshinori Kitamura, Lingwei Zhu, Takamitsu Matsubara

Browse the full ACML paper archive.