Benchmarking LF-MMI, CTC and RNN-T Criteria for Streaming ASR

In this work, to measure the accuracy and efficiency for a latency-controlled streaming automatic speech recognition (ASR) application, we perform comprehensive evaluations on three popular training criteria: LF-MMI, CTC and RNN-T. In transcribing social media videos of 7 languages with training dat...

Full description

Saved in:
Bibliographic Details
Published inarXiv.org
Main Authors Zhang, Xiaohui, Zhang, Frank, Liu, Chunxi, Schubert, Kjell, Chan, Julian, Prakash, Pradyot, Liu, Jun, Ching-Feng Yeh, Peng, Fuchun, Saraf, Yatharth, Zweig, Geoffrey
Format Paper
LanguageEnglish
Published Ithaca Cornell University Library, arXiv.org 09.11.2020
Subjects
Automatic control
Automatic speech recognition
Coders
Criteria
Digital media
Experimentation
Languages
Model accuracy
Modelling
Training
Online AccessGet full text

Cover

More Information
Summary:In this work, to measure the accuracy and efficiency for a latency-controlled streaming automatic speech recognition (ASR) application, we perform comprehensive evaluations on three popular training criteria: LF-MMI, CTC and RNN-T. In transcribing social media videos of 7 languages with training data 3K-14K hours, we conduct large-scale controlled experimentation across each criterion using identical datasets and encoder model architecture. We find that RNN-T has consistent wins in ASR accuracy, while CTC models excel at inference efficiency. Moreover, we selectively examine various modeling strategies for different training criteria, including modeling units, encoder architectures, pre-training, etc. Given such large-scale real-world streaming ASR application, to our best knowledge, we present the first comprehensive benchmark on these three widely used training criteria across a great many languages.
ISSN:2331-8422