Streaming Speech-to-Confusion Network Speech Recognition
arXiv:2306.03778 · doi:10.21437/Interspeech.2023-486
Abstract
In interactive automatic speech recognition (ASR) systems, low-latency requirements limit the amount of search space that can be explored during decoding, particularly in end-to-end neural ASR. In this paper, we present a novel streaming ASR architecture that outputs a confusion network while maintaining limited latency, as needed for interactive applications. We show that 1-best results of our model are on par with a comparable RNN-T system, while the richer hypothesis set allows second-pass rescoring to achieve 10-20\% lower word error rate on the LibriSpeech task. We also show that our model outperforms a strong RNN-T baseline on a far-field voice assistant task.
arXiv commentSubmitted to Interspeech 2023
References in corpus (5)
- Attention-Based Models for Speech Recognition
- Sequence Transduction with Recurrent Neural Networks
- Finding consensus in speech recognition: word error minimization and other applications of confusion networks
- Towards Online End-to-end Transformer Automatic Speech Recognition
- Unified End-to-End Speech Recognition and Endpointing for Fast and Efficient Speech Systems