Data Augmentation for Voice-Assistant NLU using BERT-based Interchangeable Rephrase

Akhila Yerukola, Mason Bretan, Hongxia Jin


Abstract
We introduce a data augmentation technique based on byte pair encoding and a BERT-like self-attention model to boost performance on spoken language understanding tasks. We compare and evaluate this method with a range of augmentation techniques encompassing generative models such as VAEs and performance-boosting techniques such as synonym replacement and back-translation. We show our method performs strongly on domain and intent classification tasks for a voice assistant and in a user-study focused on utterance naturalness and semantic similarity.
Anthology ID:
2021.eacl-main.159
Volume:
Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume
Month:
April
Year:
2021
Address:
Online
Editors:
Paola Merlo, Jorg Tiedemann, Reut Tsarfaty
Venue:
EACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
1852–1860
Language:
URL:
https://aclanthology.org/2021.eacl-main.159
DOI:
10.18653/v1/2021.eacl-main.159
Bibkey:
Cite (ACL):
Akhila Yerukola, Mason Bretan, and Hongxia Jin. 2021. Data Augmentation for Voice-Assistant NLU using BERT-based Interchangeable Rephrase. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume, pages 1852–1860, Online. Association for Computational Linguistics.
Cite (Informal):
Data Augmentation for Voice-Assistant NLU using BERT-based Interchangeable Rephrase (Yerukola et al., EACL 2021)
Copy Citation:
PDF:
https://aclanthology.org/2021.eacl-main.159.pdf