AI Chat AI Image Generator AI Video Text to Speech

Homophone-based Label Smoothing in End-to-End Automatic Speech Recognition

04/07/2020

∙

by Yi Zheng, et al.

∙

∙

A new label smoothing method that makes use of prior knowledge of a language at human level, homophone, is proposed in this paper for automatic speech recognition (ASR). Compared with its forerunners, the proposed method uses pronunciation knowledge of homophones in a more complex way. End-to-end ASR models that learn acoustic model and language model jointly and modelling units of characters are necessary conditions for this method. Experiments with hybrid CTC sequence-to-sequence model show that the new method can reduce character error rate (CER) by 0.4

Yi Zheng
22 publications
Xianjie Yang
1 publication
Xuyong Dang
1 publication

page 1

page 2

page 3

page 4

research

∙ 03/24/2022

Automatic Speech recognition for Speech Assessment of Preschool Children

The acoustic and linguistic features of preschool speech are investigate...

0 Amirhossein Abaskohi, et al. ∙

research

∙ 06/15/2020

Exploration of End-to-End ASR for OpenSTT – Russian Open Speech-to-Text Dataset

This paper presents an exploration of end-to-end automatic speech recogn...

0 Andrei Andrusenko, et al. ∙

research

∙ 10/18/2019

End-to-End Speech Recognition: A review for the French Language

Recently, end-to-end ASR based either on sequence-to-sequence networks o...

0 Florian Boyer, et al. ∙

research

∙ 06/09/2023

Improving Frame-level Classifier for Word Timings with Non-peaky CTC in End-to-End Automatic Speech Recognition

End-to-end (E2E) systems have shown comparable performance to hybrid sys...

0 Xianzhao Chen, et al. ∙

research

∙ 02/16/2021

End-to-End Automatic Speech Recognition with Deep Mutual Learning

This paper is the first study to apply deep mutual learning (DML) to end...

0 Ryo Masumura, et al. ∙

research

∙ 02/02/2023

Improving Rare Words Recognition through Homophone Extension and Unified Writing for Low-resource Cantonese Speech Recognition

Homophone characters are common in tonal syllable-based languages, such ...

0 Holam Chung, et al. ∙

research

∙ 10/27/2022

SAN: a robust end-to-end ASR model architecture

In this paper, we propose a novel Siamese Adversarial Network (SAN) arch...

0 Zeping Min, et al. ∙