Context dependent state tying for speech recognition using deep neural network acoustic models

Author

Bacchiani, Michiel ; Rybach, David

Author_Institution

Google Inc., New York, NY, USA

fYear

2014

fDate

4-9 May 2014

Firstpage

230

Lastpage

234

Abstract

This paper proposes an algorithm to design a tied-state inventory for a context dependent, neural network-based acoustic model for speech recognition. Rather than relying on a GMM/HMM system that operates on a different feature space and is of a different model family, the proposed algorithm optimizes state tying on the activation vectors of the neural network directly. Experiments show the viability of the proposed algorithm reducing the WER from 36.3% for a context independent system to 16.0% for a 15000 tied-state system.

Keywords

Gaussian processes; acoustics; hidden Markov models; mixture models; neural nets; speech recognition; vectors; GMM system; Gaussian mixture model; HMM system; activation vectors; context dependent state tying; context independent system; deep neural network acoustic models; feature space; hidden Markov models; speech recognition; tied-state inventory; Acoustics; Context; Entropy; Hidden Markov models; Neural networks; Training; Vectors; Acoustic Modeling; Context Modeling; Deep Neural Networks; State Tying;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech and Signal Processing (ICASSP), 2014 IEEE International Conference on

Conference_Location

Florence

Type

conf

DOI

10.1109/ICASSP.2014.6853592

Filename

6853592