Multi-Domain Data Enhanced Network for Task-Oriented Dialogue

Yuanyuan Cai, Yutong Shi, Haitao Xiong

2022

Abstract

Intelligent dialogue system is a new approach to human-machine interaction, replacing highly repetitive and standardized customer service. It has been applied widely for various purposes in many domains, such as weather inquiry and hotel booking. However, the performance of dialogue systems is limited to domain-specific data which is insufficient to train a high-quality dialogue model. To address this issue, we purpose a multi-domain data enhanced neural model with an end-to-end framework. This model fuses integrated-domain features and individual-domain features to improve the performance on dialogue generation in each specific domain. The experimental results on the two datasets show that our dialogue model outperforms existing methods, which indicates it has high flexibility in domains with small-scale data.

Paper Citation

in Harvard Style

Cai Y., Shi Y. and Xiong H. (2022). Multi-Domain Data Enhanced Network for Task-Oriented Dialogue. In Proceedings of the 3rd International Symposium on Automation, Information and Computing - Volume 1: ISAIC; ISBN 978-989-758-622-4, SciTePress, pages 299-305. DOI: 10.5220/0011923700003612

in Bibtex Style

@conference{isaic22,
author={Yuanyuan Cai and Yutong Shi and Haitao Xiong},
title={Multi-Domain Data Enhanced Network for Task-Oriented Dialogue},
booktitle={Proceedings of the 3rd International Symposium on Automation, Information and Computing - Volume 1: ISAIC},
year={2022},
pages={299-305},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0011923700003612},
isbn={978-989-758-622-4},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 3rd International Symposium on Automation, Information and Computing - Volume 1: ISAIC
TI - Multi-Domain Data Enhanced Network for Task-Oriented Dialogue
SN - 978-989-758-622-4
AU - Cai Y.
AU - Shi Y.
AU - Xiong H.
PY - 2022
SP - 299
EP - 305
DO - 10.5220/0011923700003612
PB - SciTePress