Back-Translation-Style Data Augmentation for End-to-End ASR

TL;DR

This paper introduces a novel data augmentation technique for end-to-end speech recognition that leverages unpaired text data through a neural text-to-encoder model, improving performance and reducing unknown words.

Contribution

The paper proposes a back-translation-style augmentation method using hidden states predicted from unpaired text, enhancing E2E-ASR training efficiency and accuracy.

Findings

Improved ASR performance on LibriSpeech dataset

Reduced number of unknown words

Effective use of unpaired text data for augmentation

Abstract

In this paper we propose a novel data augmentation method for attention-based end-to-end automatic speech recognition (E2E-ASR), utilizing a large amount of text which is not paired with speech signals. Inspired by the back-translation technique proposed in the field of machine translation, we build a neural text-to-encoder model which predicts a sequence of hidden states extracted by a pre-trained E2E-ASR encoder from a sequence of characters. By using hidden states as a target instead of acoustic features, it is possible to achieve faster attention learning and reduce computational cost, thanks to sub-sampling in E2E-ASR encoder, also the use of the hidden states can avoid to model speaker dependencies unlike acoustic features. After training, the text-to-encoder model generates the hidden states from a large amount of unpaired text, then E2E-ASR decoder is retrained using the generated hidden states as additional training data. Experimental evaluation using LibriSpeech dataset demonstrates that our proposed method achieves improvement of ASR performance and reduces the number of unknown words without the need for paired data.