基于对话约束的回复生成研究(Research on Response Generation via Dialogue Constraints)

Mengyu Guan (管梦雨), Zhongqing Wang (王中卿), Shoushan Li (李寿山), Guodong Zhou (周国栋)


Abstract
现有的对话系统中存在着生成“好的”、“我不知道”等无意义的安全回复问题。日常对话中,对话者通常围绕特定的主题进行讨论且每句话都有明显的情感和意图。因此该文提出了基于对话约束的回复生成模型,即在Seq2Seq模型的基础上,结合对对话的主题、情感、意图的识别。该方法对生成回复的主题、情感和意图进行约束,从而生成具有合理的情感和意图且与对话主题相关的回复。实验证明,该文提出的方法能有效地提高生成回复的质量。
Anthology ID:
2020.ccl-1.22
Volume:
Proceedings of the 19th Chinese National Conference on Computational Linguistics
Month:
October
Year:
2020
Address:
Haikou, China
Editors:
Maosong Sun (孙茂松), Sujian Li (李素建), Yue Zhang (张岳), Yang Liu (刘洋)
Venue:
CCL
SIG:
Publisher:
Chinese Information Processing Society of China
Note:
Pages:
225–235
Language:
Chinese
URL:
https://aclanthology.org/2020.ccl-1.22
DOI:
Bibkey:
Cite (ACL):
Mengyu Guan, Zhongqing Wang, Shoushan Li, and Guodong Zhou. 2020. 基于对话约束的回复生成研究(Research on Response Generation via Dialogue Constraints). In Proceedings of the 19th Chinese National Conference on Computational Linguistics, pages 225–235, Haikou, China. Chinese Information Processing Society of China.
Cite (Informal):
基于对话约束的回复生成研究(Research on Response Generation via Dialogue Constraints) (Guan et al., CCL 2020)
Copy Citation:
PDF:
https://aclanthology.org/2020.ccl-1.22.pdf