Faster reinforcement learning after pretraining deep networks to predict state dynamics | IEEE Conference Publication | IEEE Xplore