@misc{indiciae6749d6007314,
  title = {Beyond Value-Function Gaps: Improved Instance-Dependent Regret Bounds for Episodic Reinforcement Learning},
  author = {Christoph Dann and Teodor V. Marinov and Mehryar Mohri and Julian Zimmert},
  year = {2021},
  url = {https://arxiv.org/abs/2107.01264},
  note = {Source identifier: 2107.01264}
}