Agent-Aware Dropout DQN for Safe and Efficient On-line Dialogue Policy Learning.

Lu Chen, Xiang Zhou, Cheng Chang, Runzhe Yang, Kai Yu

Browse the full EMNLP paper archive.