Investigating Deep Reinforcement Learning Techniques in Personalized Dialogue Generation
文献类型：会议论文
| 作者 | Min Yang; Qiang Qu; Kai Lei; Jia Zhu; Xiaojun Chen; Zhou Zhao; Zhexue Huang |
| 出版日期 | 2018 |
| 会议日期 | 2018 |
| 会议地点 | 美国圣地亚哥 |
| 英文摘要 | In this paper, we propose a personalized dialogue generation system, which combines reinforcement learning techniques with an attention-based hierarchical recurrent encoderdecoder model. Firstly, we incorporate user-speci c information into the decoder to capture user's background information and speaking style. Secondly, we employ reinforcement learning techniques to maximize future reward in dialogue, which enables our system to generate topic-coherent, informative and grammatical responses. Moreover, we propose three types of rewards to characterize good conversations. Finally, we compare the performance of the following reinforcement learning methods in dialogue generation: policy gradient, Q-learning, and actor-critic algorithms. We conduct experiments to verify the e ectiveness of the proposed model on two dialogue datasets. Experimental results demonstrate that our model can generate better personalized dialogues for di erent users. Quantitatively, our method achieves better performance than the state-of-theart dialogue systems in terms of BLEU score, perplexity, and human evaluation. |
| 语种 | 英语 |
| URL标识 | 查看原文 |
| 源URL | [http://ir.siat.ac.cn:8080/handle/172644/14088]  |
| 专题 | 深圳先进技术研究院_数字所 |
| 推荐引用方式 GB/T 7714 | Min Yang,Qiang Qu,Kai Lei,et al. Investigating Deep Reinforcement Learning Techniques in Personalized Dialogue Generation[C]. 见:. 美国圣地亚哥. 2018. |
入库方式： OAI收割
来源：深圳先进技术研究院
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。