Title :
Modified Hebbian auto-adaptive impulse neural circuits
Author :
Nintunze, N. ; Wu, Aimin
Author_Institution :
Dept. of Electr. & Comput. Eng., Washington State Univ., Pullman, WA, USA
Abstract :
Artificial neural networks learn by adapting interconnection weights. A generalised weight adaptation expression for associative learning has been implemented using synapse circuits based on floating gate devices. A reinforcement depending on the correlation of a synapse input and a neuronal output is used. The circuits also illustrate the influence of the conditioning stimuli amplitude on the conditioning rate.
Keywords :
learning systems; neural nets; Hebbian auto-adaptive impulse neural circuits; adapting interconnection weights; adaptive control; artificial intelligence; artificial neural nets; associative learning; conditioning rate; conditioning stimuli amplitude; floating gate devices; generalised weight adaptation expression; synapse circuits;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:19901002
