Using Synthetic Audio to Improve The Recognition of Out-Of-Vocabulary Words in End-To-End ASR Systems

TL;DR

This paper introduces a method using synthetic TTS-generated audio to improve recognition of out-of-vocabulary words in end-to-end ASR systems, achieving significant WER reduction without harming overall performance.

Contribution

The study demonstrates that fine-tuning RNN-T models with synthetic OOV audio and elastic weight consolidation enhances OOV word recognition in ASR systems.

Findings

57% relative WER reduction on OOV words

No degradation on overall test set performance

Effective use of synthetic data for OOV recognition improvement

Abstract

Today, many state-of-the-art automatic speech recognition (ASR) systems apply all-neural models that map audio to word sequences trained end-to-end along one global optimisation criterion in a fully data driven fashion. These models allow high precision ASR for domains and words represented in the training material but have difficulties recognising words that are rarely or not at all represented during training, i.e. trending words and new named entities. In this paper, we use a text-to-speech (TTS) engine to provide synthetic audio for out-of-vocabulary (OOV) words. We aim to boost the recognition accuracy of a recurrent neural network transducer (RNN-T) on OOV words by using the extra audio-text pairs, while maintaining the performance on the non-OOV words. Different regularisation techniques are explored and the best performance is achieved by fine-tuning the RNN-T on both original training data and extra synthetic data with elastic weight consolidation (EWC) applied on the encoder. This yields a 57% relative word error rate (WER) reduction on utterances containing OOV words without any degradation on the whole test set.