A neural network, speech-based approach to literacy

  • Authors:

  • John Fulcher;Russell Gluck;Marion Worthy;Kim Draisma;Wilma Vialle

  • Affiliations:

  • University of Wollongong, Australia;University of Wollongong, Australia;University of Wollongong, Australia;University of Wollongong, Australia;University of Wollongong, Australia

  • Venue:
  • ISICT '03 Proceedings of the 1st international symposium on Information and communication technologies
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

An automatic word recognition system is described which assists orally proficient literacy inefficient people to become literate within a minimum possible timeframe. The system enables people with a strong oral tradition to impart their stories directly in text form, without the assistance of a learning facilitator, which is the current best practice for such people. This project stems from work with indigenous communities, but has far-reaching repercussions beyond this community sector, with the potential for benefit to mainstream literacy education (Gluck et.al., 1999). It should be emphasized that our system differs from virtually all other speech-to-text recognition systems in that users are not required to interact via written text (e.g. Kohonen, 1988); at the heart of the system is a neural network-based pattern recognizer which translates speech patterns into visual rather than textual cues. Results obtained to date using this system are reported.