Efficient and Stable Information Directed Exploration for Continuous Reinforcement Learning | IEEE Conference Publication | IEEE Xplore