TY - GEN
T1 - From machine reading comprehension to dialogue state tracking
T2 - 2nd Workshop on NLP for Conversational AI, NLP4ConvAI 2020 at the 58th Annual Meeting of the Association for Computational Linguistics, ACL 2020
AU - Gao, Shuyang
AU - Agarwal, Sanchit
AU - Chung, Tagyoung
AU - Jin, Di
AU - Hakkani-Tur, Dilek
N1 - Publisher Copyright:
© 2020 Association for Computational Linguistics.
PY - 2020
Y1 - 2020
N2 - Dialogue state tracking (DST) is at the heart of task-oriented dialogue systems. However, the scarcity of labeled data is an obstacle to building accurate and robust state tracking systems that work across a variety of domains. Existing approaches generally require some dialogue data with state information and their ability to generalize to unknown domains is limited. In this paper, we propose using machine reading comprehension (RC) in state tracking from two perspectives: model architectures and datasets. We divide the slot types in dialogue state into categorical or extractive to borrow the advantages from both multiple-choice and span-based reading comprehension models. Our method achieves near the current state-of-the-art in joint goal accuracy on MultiWOZ 2.1 given full training data. More importantly, by leveraging machine reading comprehension datasets, our method outperforms the existing approaches by many a large margin in few-shot scenarios when the availability of in-domain data is limited. Lastly, even without any state tracking data, i.e., zero-shot scenario, our proposed approach achieves greater than 90% average slot accuracy in 12 out of 30 slots in MultiWOZ 2.1.
AB - Dialogue state tracking (DST) is at the heart of task-oriented dialogue systems. However, the scarcity of labeled data is an obstacle to building accurate and robust state tracking systems that work across a variety of domains. Existing approaches generally require some dialogue data with state information and their ability to generalize to unknown domains is limited. In this paper, we propose using machine reading comprehension (RC) in state tracking from two perspectives: model architectures and datasets. We divide the slot types in dialogue state into categorical or extractive to borrow the advantages from both multiple-choice and span-based reading comprehension models. Our method achieves near the current state-of-the-art in joint goal accuracy on MultiWOZ 2.1 given full training data. More importantly, by leveraging machine reading comprehension datasets, our method outperforms the existing approaches by many a large margin in few-shot scenarios when the availability of in-domain data is limited. Lastly, even without any state tracking data, i.e., zero-shot scenario, our proposed approach achieves greater than 90% average slot accuracy in 12 out of 30 slots in MultiWOZ 2.1.
UR - http://www.scopus.com/inward/record.url?scp=85098380041&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85098380041&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:85098380041
T3 - Proceedings of the Annual Meeting of the Association for Computational Linguistics
SP - 79
EP - 89
BT - ACL 2020 - NLP for Conversational AI, Proceedings of the 2nd Workshop
PB - Association for Computational Linguistics (ACL)
Y2 - 9 July 2020
ER -