Stochastic cubic-regularized policy gradient method

, which improves upon the current optimal complexity . Finally, experimental results are provided to demonstrate the effectiveness of the method.