Correct reaction neural network

Koji Nakajima, Yoshihiro Hayakawa

Research output: Contribution to journalArticlepeer-review

7 Citations (Scopus)

Abstract

We describe a new method to analyze and to design Hopfield-like neural networks. Our method is based on the "force" derived from the equations of motion for neural networks instead of the energy functions of the system. The properties of local minima in the energy function of Hopfield networks can be investigated by using the new method. A novel design technique to eliminate these local minima in the networks has been developed.

Original languageEnglish
Pages (from-to)217-222
Number of pages6
JournalNeural Networks
Volume6
Issue number2
DOIs
Publication statusPublished - 1993

Keywords

  • Basin of attractor
  • Energy function
  • Information strage
  • Local minimum
  • Optimization circuit

ASJC Scopus subject areas

  • Cognitive Neuroscience
  • Artificial Intelligence

Fingerprint

Dive into the research topics of 'Correct reaction neural network'. Together they form a unique fingerprint.

Cite this