Sample Efficient Reinforcement Learning with Double Importance Sampling Weight Clipping | IEEE Conference Publication | IEEE Xplore