Non-Gaussian feedback policies

This paper introduces a new perspective on stochastic optimal control in dynamical systems, a key challenge in sequential decision-making. The authors propose a method of control as Markovian score climbing…