@article{batchpolicygradientmethodsforimproving,
  title = {Batch Policy Gradient Methods for Improving Neural Conversation Models},
  author = {Kirthevasan Kandasamy and Yoram Bachrach and Ryota Tomioka and Daniel Tarlow and David Carter},
  year = {2017},
  eprint = {1702.03334},
  archivePrefix = {arXiv},
  url = {http://arxiv.org/abs/1702.03334v1},
}