Developing RNN-T Models Surpassing High-Performance Hybrid Models with Customization Capability

TL;DR

This paper presents advancements in RNN-T models for speech recognition, achieving superior accuracy and lower latency than hybrid models, with effective domain customization using text-to-speech data.

Contribution

Introduces improved RNN-T training techniques and demonstrates effective domain adaptation methods leveraging text-to-speech data.

Findings

RNN-T models outperform hybrid models in accuracy and latency.

Enhanced training reduces GPU memory usage.

Domain adaptation via TTS data is highly effective.

Abstract

Because of its streaming nature, recurrent neural network transducer (RNN-T) is a very promising end-to-end (E2E) model that may replace the popular hybrid model for automatic speech recognition. In this paper, we describe our recent development of RNN-T models with reduced GPU memory consumption during training, better initialization strategy, and advanced encoder modeling with future lookahead. When trained with Microsoft's 65 thousand hours of anonymized training data, the developed RNN-T model surpasses a very well trained hybrid model with both better recognition accuracy and lower latency. We further study how to customize RNN-T models to a new domain, which is important for deploying E2E models to practical scenarios. By comparing several methods leveraging text-only data in the new domain, we found that updating RNN-T's prediction and joint networks using text-to-speech generated from domain-specific text is the most effective.