Key points are not available for this paper at this time.
. Our work allows solving challenging control problems without extensive experimentation, making reinforcement learning broadly applicable.
Hafner et al. (Wed,) studied this question.