@inproceedings{onlinemarkovdecisionprocessesunderbandit,
  title = {Online Markov Decision Processes under Bandit Feedback},
  author = {Gergely Neu and Andras Antos and András György and Csaba Szepesvári},
  year = {2010},
  booktitle = {NeurIPS 2010 12},
  url = {http://papers.nips.cc/paper/4048-online-markov-decision-processes-under-bandit-feedback},
}