𝔖 Bobbio Scriptorium
✦   LIBER   ✦

Associative memory neural networks with concatenated vectors and nonzero diagonal terms

✍ Scribed by Emanuel Marom


Publisher
Elsevier Science
Year
1990
Tongue
English
Weight
498 KB
Volume
3
Category
Article
ISSN
0893-6080

No coin nor oath required. For personal study only.

✦ Synopsis


The efJects oJ vector concatenation as well as that of eliminating the zero diagonal restriction of associative memory neural network matrices is anah, zed. Extensive computer simulations seem to indicate that the t~e of concatenated vectors increases the storage capaci O' of the association matrices, beyond the accepted practical limit Of 1). 14N, N being the length of the stored vector,s.


πŸ“œ SIMILAR VOLUMES


Global stability analysis of bidirection
✍ Jiye Zhang; Yiren Yang πŸ“‚ Article πŸ“… 2001 πŸ› John Wiley and Sons 🌐 English βš– 108 KB πŸ‘ 1 views

## Abstract In this paper, without assuming the boundedness, monotonicity and differentiability of the activation functions, we present new conditions ensuring existence, uniqueness, and global asymptotical stability of the equilibrium point of bidirectional associative memory neural networks with