Summary
2011 International Symposium on Nonlinear Theory and Its Applications
2011
Session Number:B1L-D
Session:
Number:B1L-D1
Complex-valued Associative Memory with Strong Thresholds
Michimasa Kitahara, Masaki Kobayashi,
pp.362-365
Publication Date:2011/9/4
Online ISSN:2188-5079
DOI:10.34385/proc.45.B1L-D1
PDF download (78.2KB)
Summary:
A Complex-valued Associative Memory (CAM) stores not only training patterns but also their rotated patterns. They reduce noise robustness. In this paper, we introduce strong thresholds to the neurons. Thresholds relieve storing rotated patterns and improve noise robustness. If thresholds are simply used, their effect is absorbed in quantification of complex-valued neurons. So we use strong thresholds to improve noise robustness.