@inproceedings{directpolicygradientsdirectoptimization,
  title = {Direct Policy Gradients: Direct Optimization of Policies in Discrete Action Spaces},
  author = {Guy Lorberbom and Chris J. Maddison and Nicolas Heess and Tamir Hazan and Daniel Tarlow},
  year = {2019},
  booktitle = {NeurIPS 2020 12},
  eprint = {1906.06062},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1906.06062v2},
}