We consider finite temperature neural networks learned by the Edimburgh’s algorithm in the symmetric version. We prove that domains of attraction of stored patterns are independent from thermal noise.
Independence from Temperature of Domains of Attraction in Neural Networks
International Neural Network Conference ; Chapter : 149 ; 973-976
1990-01-01
4 pages
Article/Chapter (Book)
Electronic Resource
English
Las Vegas method of region-of-attraction enlargement in neural networks
British Library Conference Proceedings | 1994
|School Trip Attraction Modeling, Using Neural & Fuzzy-Neural Approaches
British Library Conference Proceedings | 2005
|Online Contents | 1996