Characteristics of associative chaotic neural networks with weighted pattern storage-a pattern is stored stronger than others

Author

Adachi, Masakazu ; Aihara, Kazuyuki

Author_Institution

Dept. of Electron. Eng., Tokyo Denki Univ., Japan

Volume

3

fYear

1999

fDate

1999

Firstpage

1028

Abstract

Associative chaotic neural networks with weighted pattern storage are studied. Values of the synaptic weights of conventional associative neural networks are determined by an auto-associative matrix. On the other hand, in this paper, we use a weighted auto-associative matrix in order to store a pattern that is stronger than the other stored patterns. Retrieval characteristics and dynamical properties of associative chaotic neural networks with this weighted auto-associative matrix are numerically analysed. As a result, the network retrieves the strongly stored pattern more frequently than other stored patterns, even in the case where the dynamics of the network is chaotic

Keywords

associative processing; chaos; content-addressable storage; information retrieval; neural nets; associative chaotic neural networks; chaotic dynamics; dynamical properties; numerically analysis; pattern retrieval characteristics; strongly stored pattern; synaptic weights; weighted auto-associative matrix; weighted pattern storage; Biological neural networks; Brain modeling; Chaos; Convergence; Educational institutions; Neural networks; Neurofeedback; Neurons; Paper technology; Physics;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Information Processing, 1999. Proceedings. ICONIP '99. 6th International Conference on

Conference_Location

Perth, WA

Print_ISBN

0-7803-5871-6

Type

conf

DOI

10.1109/ICONIP.1999.844677

Filename

844677