Experiments in Minimally Distinct Word-Pair Discrimination Using the Multi-Layer Perception

Abstract

The multi-layer perceptron is investigated as a new approach to the automatic discrimination of spoken minimally distinct word-pairs. The choice of the parameters for the multi-layer perceptron is discussed and experimental results are reported. A comparison is made with hidden Markov modelling applied to the same data. The results, for this particular task, show that the discrimination accuracy obtained using the multi-layer perceptron is superior to that attained using hidden Markov modelling. Keywords: Great Britain.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Apr 01, 1988
Accession Number
ADA199526

Entities

People

  • R. K. Moore
  • S. M. Peeling

Organizations

  • Royal Signals and Radar Establishment

Tags

Communities of Interest

  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Accuracy
  • Algorithms
  • Automated Speech Recognition
  • Discrimination
  • Distribution Functions
  • Errors
  • Foreign Languages
  • Frequency Bands
  • Learning
  • Markov Models
  • Momentum
  • Pattern Recognition
  • Probability
  • Probability Distributions
  • Recognition
  • Test Sets
  • Training

Readers

  • Neural Network Machine Learning.
  • Speech Processing/Speech Recognition.