@inproceedings{imperfectalsodeservesrewardmultilevel,
  title = {Imperfect also Deserves Reward: Multi-Level and Sequential Reward Modeling for Better Dialog Management},
  author = {Zhengxu Hou and Bang Liu and Ruihui Zhao and Zijing Ou and Yafei Liu and Xi Chen and Yefeng Zheng},
  year = {2021},
  booktitle = {NAACL 2021 4},
  eprint = {2104.04748},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2104.04748v1},
}