Fast reinforcement learning of dialog strategies | IEEE Conference Publication | IEEE Xplore