Advances in all-neural speech recognition

This paper advances the design of CTC-based all-neural (or end-to-end) speech recognizers. We propose a novel symbol inventory, and a novel iterated-CTC method in which a second system is used to transform a noisy initial output into a cleaner version. We present a number of stabilization and initia...

Full description

Saved in:
Bibliographic Details
Published in2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) pp. 4805 - 4809
Main Authors Zweig, Geoffrey, Chengzhu Yu, Droppo, Jasha, Stolcke, Andreas
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.03.2017
Subjects
Acoustics
CTC
Decoding
end-to-end training
Hidden Markov models
Neural networks
recurrent neural network
Speech recognition
Training
Online AccessGet full text

Cover

More Information
Summary:This paper advances the design of CTC-based all-neural (or end-to-end) speech recognizers. We propose a novel symbol inventory, and a novel iterated-CTC method in which a second system is used to transform a noisy initial output into a cleaner version. We present a number of stabilization and initialization methods we have found useful in training these networks. We evaluate our system on the commonly used NIST 2000 conversational telephony test set, and significantly exceed the previously published performance of similar systems, both with and without the use of an external language model and decoding technology.
ISSN:2379-190X
DOI:10.1109/ICASSP.2017.7953069