Converse, Focus and Guess - Towards Multi-Document Driven Dialogue

Authors

  • Han Liu Beijing University of Posts and Telecommunications, Beijing, China
  • Caixia Yuan Beijing University of Posts and Telecommunications, Beijing, China
  • Xiaojie Wang Beijing University of Posts and Telecommunications, Beijing, China
  • Yushu Yang Meituan, Beijing, China
  • Huixing Jiang Meituan, Beijing, China
  • Zhongyuan Wang Meituan, Beijing, China

DOI:

https://doi.org/10.1609/aaai.v35i15.17579

Keywords:

Conversational AI/Dialog Systems

Abstract

We propose a novel task, Multi-Document Driven Dialogue (MD3), in which an agent can guess the target document that the user is interested in by leading a dialogue. To benchmark progress, we introduce a new dataset of GuessMovie, which contains 16,881 documents, each describing a movie, and associated 13,434 dialogues. Further, we propose the MD3 model. Keeping guessing the target document in mind, it converses with the user conditioned on both document engagement and user feedback. In order to incorporate large-scale external documents into the dialogue, it pretrains a document representation which is sensitive to attributes it talks about an object. Then it tracks dialogue state by detecting evolvement of document belief and attribute belief, and finally optimizes dialogue policy in principle of entropy decreasing and reward increasing, which is expected to successfully guess the user's target in a minimum number of turns. Experiments show that our method significantly outperforms several strong baseline methods and is very close to human's performance.

Downloads

Published

2021-05-18

How to Cite

Liu, H., Yuan, C., Wang, X., Yang, Y., Jiang, H., & Wang, Z. (2021). Converse, Focus and Guess - Towards Multi-Document Driven Dialogue. Proceedings of the AAAI Conference on Artificial Intelligence, 35(15), 13380-13387. https://doi.org/10.1609/aaai.v35i15.17579

Issue

Section

AAAI Technical Track on Speech and Natural Language Processing II