An $L^2$ Analysis of Reinforcement Learning in High Dimensions With Kernel and Neural Network Approximation. CSIAM-AM 2022, 3 (2), 191-220. https://doi.org/10.4208/csiam-am.SO-2021-0026.