Analysis of Measure-Valued Derivatives in a Reinforcement Learning Actor-Critic Framework | IEEE Conference Publication | IEEE Xplore