1. Paper title
Generating Informative Conversational Response using Recurrent Knowledge-Interaction and Knowledge-Copy
2. link
https://www.aclweb.org/anthology/2020.acl-main.6.pdf
3. 摘要
Knowledge-driven conversation approaches have achieved remarkable research attention recently. However, generating an informative response with multiple relevant knowledge without losing fluency and coherence is still one of the main challenges. To address this issue, this paper proposes a method that uses recurrent knowledge interaction among response decoding steps to incorporate appropriate knowledge. Furthermore, we introduce a knowledge copy mechanism using a knowledge-aware pointer network to copy words from external knowledge according to knowledge attention distribution. Our joint neural conversation model which integrates recurrent Knowledge-Interaction and knowledge Copy (KIC) performs well on generating informative responses. Experiments demonstrate that our model with fewer parameters yields significant improvements over competitive baselines on two datasets Wizardof-Wikipedia(average Bleu +87%; abs.:0.034) and DuConv(average Bleu +20%; abs.:0.047) with different knowledge formats (textual & structured) and different languages (English & Chinese).
4. 要解决什么问题
在多重相关知识的场景中生成流利和连贯的应答。
5. 作者的主要贡献
在“response decoding”阶段使用“recurrent knowledge interaction”来“incorporate appropriate knowledge”。
[?] 指针网络
引入“knowledge copy”机制：使用“knowledge-aware”指针网络基于“knowledge attention distribution”从外部知识获取words。
6. 得到了什么结果
需要更少的参数。
比baseline有重大的提升：
Wizardof-Wikipedia(average Bleu +87%; abs.:0.034)
DuConv(average Bleu +20%; abs.:0.047)