TY - RPRT TI - Reinforcement Learning by Value Gradients AU - Michael Fairbank PY - 2008 UR - https://arxiv.org/abs/0803.3539 ID - 0803.3539 ER -