@misc{indiciae36118153c8c4,
  title = {Why Policy Gradient Algorithms Work for Undiscounted Total-Reward MDPs},
  author = {Jongmin Lee and Ernest K. Ryu},
  year = {2026},
  url = {https://arxiv.org/abs/2510.18340},
  note = {Source identifier: 2510.18340}
}