Improving Proper Noun Recognition in End-to-End ASR By Customization of the MWER Loss Criterion

TL;DR

This paper enhances end-to-end speech recognition systems' ability to accurately recognize proper nouns by customizing the MWER loss function, achieving significant improvements without additional data or external models.

Contribution

The authors introduce two novel MWER-based loss criteria that specifically improve proper noun recognition in E2E ASR systems without requiring extra data or external models.

Findings

Achieved 2-7% relative improvement on benchmarks.

No additional data or external models needed.

Enhanced proper noun recognition accuracy.

Abstract

Proper nouns present a challenge for end-to-end (E2E) automatic speech recognition (ASR) systems in that a particular name may appear only rarely during training, and may have a pronunciation similar to that of a more common word. Unlike conventional ASR models, E2E systems lack an explicit pronounciation model that can be specifically trained with proper noun pronounciations and a language model that can be trained on a large text-only corpus. Past work has addressed this issue by incorporating additional training data or additional models. In this paper, we instead build on recent advances in minimum word error rate (MWER) training to develop two new loss criteria that specifically emphasize proper noun recognition. Unlike past work on this problem, this method requires no new data during training or external models during inference. We see improvements ranging from 2% to 7% relative on several relevant benchmarks.