DocumentCode
3325812
Title
Statistical neurodynamics of various versions of correlation associative memory
Author
Amari, Shun-Ichi
Author_Institution
Fac. of Eng., Tokyo Univ., Japan
fYear
1988
fDate
24-27 July 1988
Firstpage
633
Abstract
The statistical neurodynamical method is used to study characteristics of various versions of correlation-type associative memory models: They are a cross-correlation model, a cascaded associative memory model, a bilateral associative memory model (BAM) and its generalization k-AM, a sequence-recalling model, and an autoassociative model. They have different dynamical noise reduction properties. It is shown that BAM and an autocorrelation CAM are not so good as others.<>
Keywords
content-addressable storage; neural nets; neurophysiology; statistics; autoassociative model; bilateral associative memory; cascaded associative memory; content addressable storage; correlation associative memory; cross-correlation model; k-AM; neural nets; neurophysiology; noise reduction; sequence-recalling model; statistical neurodynamical method; Associative memories; Nervous system; Neural networks; Statistics;
fLanguage
English
Publisher
ieee
Conference_Titel
Neural Networks, 1988., IEEE International Conference on
Conference_Location
San Diego, CA, USA
Type
conf
DOI
10.1109/ICNN.1988.23900
Filename
23900
Link To Document