Exploring Back Translation with Typo Noise for Enhanced Inquiry Understanding in Task-Oriented Dialogue

Jihyun Lee, Junseok Kim, Gary Geunbae Lee


Abstract
This paper presents our approach to the DSTC11 Track 5 selection task, which focuses on retrieving appropriate natural language knowledge sources for task-oriented dialogue. We propose typologically diverse back-translation method with typo noise, which could generate various structured user inquries. Through our noised back translation, we augmented inquiries by combining three different typologies of language sources with five different typo noise injections. Our experiments demonstrate that typological variety and typo noise aids the model in generalizing to diverse user inquiries in dialogue. In the competition, where 14 teams participated, our approach achieved the 5th rank for exact matching metric.
Anthology ID:
2023.dstc-1.21
Volume:
Proceedings of The Eleventh Dialog System Technology Challenge
Month:
September
Year:
2023
Address:
Prague, Czech Republic
Editors:
Yun-Nung Chen, Paul Crook, Michel Galley, Sarik Ghazarian, Chulaka Gunasekara, Raghav Gupta, Behnam Hedayatnia, Satwik Kottur, Seungwhan Moon, Chen Zhang
Venues:
DSTC | WS
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
185–192
Language:
URL:
https://aclanthology.org/2023.dstc-1.21
DOI:
Bibkey:
Cite (ACL):
Jihyun Lee, Junseok Kim, and Gary Geunbae Lee. 2023. Exploring Back Translation with Typo Noise for Enhanced Inquiry Understanding in Task-Oriented Dialogue. In Proceedings of The Eleventh Dialog System Technology Challenge, pages 185–192, Prague, Czech Republic. Association for Computational Linguistics.
Cite (Informal):
Exploring Back Translation with Typo Noise for Enhanced Inquiry Understanding in Task-Oriented Dialogue (Lee et al., DSTC-WS 2023)
Copy Citation:
PDF:
https://aclanthology.org/2023.dstc-1.21.pdf