Accelerate Acoustic Likelihood Computations on GPU for Speech Recognition

DOI

10.2991/iccsee.2013.256How to use a DOI?

Keywords

Speech Recognition, GMM,GPU

Abstract

Acoustic likelihood computations is one of the most computationally intensive part in the large vocabulary continuous speech recognition (LVCSR). In this paper, we give an introduction of accelerating acoustic likelihood computations for graphical processing units(GPUs). According to the optimal method, we achieve 11× speedup on the acoustic probability evaluation.

Copyright

© 2013, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Yong Liu
AU  - Zhen Zhang
AU  - Yujing Si
AU  - Qingwei Zhao
AU  - Yonghong Yan
PY  - 2013/03
DA  - 2013/03
TI  - Accelerate Acoustic Likelihood Computations on GPU for Speech Recognition
BT  - Proceedings of the 2nd International Conference on Computer Science and Electronics Engineering (ICCSEE 2013)
PB  - Atlantis Press
SP  - 1019
EP  - 1022
SN  - 1951-6851
UR  - https://doi.org/10.2991/iccsee.2013.256
DO  - 10.2991/iccsee.2013.256
ID  - Liu2013/03
ER  -