融入对话上文整体信息的层次匹配回应选择(Learning Overall Dialogue Information for Dialogue Response Selection)

Bowen Si (司博文), Fang Kong (孔芳)


Abstract
对话是一个顺序交互的过程,回应选择旨在根据已有对话上文选择合适的回应,是自然语言处理领域的研究热点。已有研究取得了一定的成功,但仍然存在两个突出的问题。一是现有的编码器在挖掘对话文本语义信息上尚存在不足;二是只考虑每一回合对话与备选回应之间的关系,忽视了对话上文的整体语义信息。针对问题一,本文借助多头自注意力机制有效捕捉对话文本的语义信息;针对问题二,整合对话上文的整体语义信息,分别从单词、句子以及整体对话上文三个层次与备选回应进行匹配,充分保证匹配信息的完整。在Ubuntu Corpus V1和Douban Conversation Corpus数据集上的对比实验表明了本文给出方法的有效性。
Anthology ID:
2020.ccl-1.26
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:
266–276
Language:
Chinese
URL:
https://aclanthology.org/2020.ccl-1.26
DOI:
Bibkey:
Cite (ACL):
Bowen Si and Fang Kong. 2020. 融入对话上文整体信息的层次匹配回应选择(Learning Overall Dialogue Information for Dialogue Response Selection). In Proceedings of the 19th Chinese National Conference on Computational Linguistics, pages 266–276, Haikou, China. Chinese Information Processing Society of China.
Cite (Informal):
融入对话上文整体信息的层次匹配回应选择(Learning Overall Dialogue Information for Dialogue Response Selection) (Si & Kong, CCL 2020)
Copy Citation:
PDF:
https://aclanthology.org/2020.ccl-1.26.pdf