Deep Convolutional Neural Networks for Large-scale Speech Tasks

@article{Sainath2015DeepCN,
  title={Deep Convolutional Neural Networks for Large-scale Speech Tasks},
  author={Tara N. Sainath and Brian Kingsbury and George Saon and Hagen Soltau and Abdel-rahman Mohamed and George E. Dahl and Bhuvana Ramabhadran},
  journal={Neural networks : the official journal of the International Neural Network Society},
  year={2015},
  volume={64},
  pages={
          39-48
        }
}
Convolutional Neural Networks (CNNs) are an alternative type of neural network that can be used to reduce spectral variations and model spectral correlations which exist in signals. Since speech signals exhibit both of these properties, we hypothesize that CNNs are a more effective model for speech compared to Deep Neural Networks (DNNs). In this paper, we explore applying CNNs to large vocabulary continuous speech recognition (LVCSR) tasks. First, we determine the appropriate architecture to… CONTINUE READING

Citations

Publications citing this paper.
SHOWING 1-10 OF 208 CITATIONS

Persian Signature Verification using Fully Convolutional Networks

VIEW 9 EXCERPTS
CITES BACKGROUND
HIGHLY INFLUENCED

Multiresolution convolutional neural network for robust speech recognition

VIEW 10 EXCERPTS
CITES BACKGROUND & METHODS
HIGHLY INFLUENCED

Recent advances in LVCSR : A benchmark comparison of performances

VIEW 4 EXCERPTS
CITES METHODS
HIGHLY INFLUENCED

Very Deep Convolutional Neural Networks for Noise Robust Speech Recognition

VIEW 7 EXCERPTS
CITES BACKGROUND & METHODS
HIGHLY INFLUENCED

Convolutional neural networks for small-footprint keyword spotting

VIEW 5 EXCERPTS
CITES METHODS, RESULTS & BACKGROUND

Phone recognition with hierarchical convolutional deep maxout networks

  • László Tóth
  • Computer Science
  • EURASIP J. Audio, Speech and Music Processing
  • 2015
VIEW 4 EXCERPTS
CITES METHODS & BACKGROUND
HIGHLY INFLUENCED

Building DNN acoustic models for large vocabulary speech recognition

VIEW 5 EXCERPTS
HIGHLY INFLUENCED

Adversarial Training Methods for Boltzmann Machines

VIEW 5 EXCERPTS
CITES BACKGROUND
HIGHLY INFLUENCED

FILTER CITATIONS BY YEAR

2014
2020

CITATION STATISTICS

  • 19 Highly Influenced Citations

  • Averaged 49 Citations per year from 2017 through 2019

  • 10% Increase in citations per year in 2019 over 2018