Constant-time neural decoders for some BCH codes

Yuen Hsien Tseng, Ja Ling Wu

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

High-order neural networks (HONN) are shown to decode some BCH codes in constant-time with very low hardware complexity. HONN is a direct extension of the linear perceptron: it uses a polynomial consisting of a set of product terms as its discriminant function. Because a product term is isomorphic to a parity function and a two-layer perceptron for the parity function has been shown by Rumelhart, Hinton, and Williams (1986), HONN has a simple realization if it is considered as having a set of parity networks in the first-half layer, followed by a linear perceptron in the second-half layer. The main problem in using high-order neural networks for a specific application is to decide a proper set of product terms. We apply genetic algorithms to this structure-adaptation problem.

Original languageEnglish
Title of host publicationProceedings - 1994 IEEE International Symposium on Information Theory, ISIT 1994
PublisherIEEE
Number of pages1
ISBN (Print)0780320158, 9780780320154
DOIs
Publication statusPublished - 1994 Dec 1
Event1994 IEEE International Symposium on Information Theory, ISIT 1994 - Trondheim, Norway
Duration: 1994 Jun 271994 Jul 1

Publication series

NameIEEE International Symposium on Information Theory - Proceedings
ISSN (Print)2157-8095

Other

Other1994 IEEE International Symposium on Information Theory, ISIT 1994
CountryNorway
CityTrondheim
Period94/6/2794/7/1

    Fingerprint

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Information Systems
  • Modelling and Simulation
  • Applied Mathematics

Cite this

Tseng, Y. H., & Wu, J. L. (1994). Constant-time neural decoders for some BCH codes. In Proceedings - 1994 IEEE International Symposium on Information Theory, ISIT 1994 [394675] (IEEE International Symposium on Information Theory - Proceedings). IEEE. https://doi.org/10.1109/ISIT.1994.394675