We introduce a data augmentation technique based on byte pair encoding and a BERTlike 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

https://www.aclweb.org/anthology/2021.eacl-main.159.pdf