Reinforcement Learning with Sparse Bellman Error Extrapolation for Infinite-Horizon Approximate Optimal Regulation | IEEE Conference Publication | IEEE Xplore