@article{exploreexploitorlistencombininghuman,
  title = {Explore, Exploit or Listen: Combining Human Feedback and Policy Model to Speed up Deep Reinforcement Learning in 3D Worlds},
  author = {Zhiyu Lin and Brent Harrison and Aaron Keech and Mark O. Riedl},
  year = {2017},
  eprint = {1709.03969},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1709.03969v2},
}