Attention based on-device streaming speech recognition with large speech corpus

TL;DR

This paper introduces a large-corpus, on-device streaming speech recognition system using monotonic chunk-wise attention, achieving high accuracy, model compression, and effective domain adaptation.

Contribution

It presents a novel on-device ASR system with large-scale training, model compression, and domain adaptation techniques, improving accuracy and efficiency.

Findings

Achieved around 90% word recognition rate on general domain.

Compressed models by over 3.4 times with minimal accuracy loss.

Improved domain-specific WER by 36% through fusion with n-gram models.

Abstract

In this paper, we present a new on-device automatic speech recognition (ASR) system based on monotonic chunk-wise attention (MoChA) models trained with large (> 10K hours) corpus. We attained around 90% of a word recognition rate for general domain mainly by using joint training of connectionist temporal classifier (CTC) and cross entropy (CE) losses, minimum word error rate (MWER) training, layer-wise pre-training and data augmentation methods. In addition, we compressed our models by more than 3.4 times smaller using an iterative hyper low-rank approximation (LRA) method while minimizing the degradation in recognition accuracy. The memory footprint was further reduced with 8-bit quantization to bring down the final model size to lower than 39 MB. For on-demand adaptation, we fused the MoChA models with statistical n-gram models, and we could achieve a relatively 36% improvement on average in word error rate (WER) for target domains including the general domain.