| Autor: |
Andrecut, Mircea |
| Předmět: |
|
| Zdroj: |
International Journal of Modern Physics B: Condensed Matter Physics; Statistical Physics; Applied Physics; 3/30/2000, Vol. 14 Issue 8, p869, 8p |
| Abstrakt: |
A simple random walk learning algorithm for associative memories is described. The Hebbian memory matrix optimized by the random walk algorithm leads to a perfect learning in associative memories. Also, in the special case of a binary memory matrix, the random walk learning algorithm leads to an increase of the critical storage density from α[sub c] = 0.102 to α[sub c] = 0.25. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
