Improving Noise Tolerance in Hopfield Networks

  • Kim, Young-Tae (Department of Computer Science and Statistics, Iowa State University) ;
  • Park, Jeong-Hyun (Department of Computer Science and Statistics, Kwandong University)
  • Published : 1997.10.30

Abstract

Adding a noise tolerance factor to the Relaxation learning algorithm in Hop-field network improves noise tolerance without effecting storage capacity. The new algorithm is called the Pseudo-Relaxation algorithm, and the convergence of the algorithm has been proved. It is also shown that the noise tolerance factor does not effect learning speed.

Keywords