Summary
International Symposium on Nonlinear Theory and its Applications
2009
Session Number:B2L-D
Session:
Number:B2L-D1
Chaotic Rotor Associative Memory
Michimasa Kitahara, Masaki Kobayashi, Motonobu Hattori,
pp.-
Publication Date:2009/10/18
Online ISSN:2188-5079
Summary:
Complex-valued Associative Memory (CAM) can deal with multi-valued patterns. Rotor Associative Memory (RAM) is an advanced model of the CAM. In this paper, we propose Chaotic Rotor Associative Memory (CRAM) which uses a chaotic neuron model. It is known that the CAM stores not only given learning patterns but also the rotated patterns. All of them appear in the recall process of the Chaotic CAM (CCAM). Our proposed model, the CRAM, can remove the rotated patterns. In addition, we found that most of the superimposed patterns vanished by computer simulations.