Automatic Discovery of Subgoals in Reinforcement Learning Using Unique-Dreiction Value | IEEE Conference Publication | IEEE Xplore