@misc{indiciaef9ac851abb66,
  title = {Policy Optimization Achieves Data-Dependent Regret Bounds in MDPs with Unknown Transitions},
  author = {Mingyi Li and Taira Tsuchiya and Kenji Yamanishi},
  year = {2026},
  url = {https://arxiv.org/abs/2606.31769},
  note = {Source identifier: 2606.31769}
}