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

TL;DR

This paper presents a novel data augmentation method using BERT-based rephrasing to improve spoken language understanding in voice assistants, outperforming other techniques in classification accuracy and naturalness.

Contribution

The authors introduce a BERT-based rephrasing augmentation technique that enhances voice assistant NLU performance beyond existing methods.

Findings

Our method improves domain and intent classification accuracy.

It outperforms VAEs, synonym replacement, and back-translation.

User study shows increased utterance naturalness and semantic consistency.

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.