Marginal Policy Gradients Authors: Carson Eisenach, Haichuan Yang Please send all correspondence to eisenach [AT] princeton.edu. Summary This repository contains A PyTorch implementations of a Directional Policy Gradient algorithm and a baseline policy gradient algorithm, The Platform2D-v1 gym environment. References [1] Carson Eisenach, Haichuan Yang, Ji Liu, and Han Liu. "Marginal Policy Gradients: A Unified Family of Estimators for Bounded Action Spaces with Applications". International Conference on Learning Representations (ICLR'19), New Orleans, USA, 2019.