1 paper
Hepeng Li, Nicholas Clavette, Haibo He
We present an analytical policy update rule that is independent of parametric function approximators. The policy update rule is suitable for optimizing general stochastic policies…