Capacity of oscillatory associative-memory networks with error-free retrieval

Takashi Nishikawa, Ying-Cheng Lai, Frank C. Hoppensteadt

Research output: Contribution to journalArticle

49 Citations (Scopus)

Abstract

Networks of coupled periodic oscillators (similar to the Kuramoto model) have been proposed as models of associative memory. However, error-free retrieval states of such oscillatory networks are typically unstable, resulting in a near zero capacity. This puts the networks at disadvantage as compared with the classical Hopfield network. Here we propose a simple remedy for this undesirable property and show rigorously that the error-free capacity of our oscillatory, associative-memory networks can be made as high as that of the Hopfield network. They can thus not only provide insights into the origin of biological memory, but can also be potentially useful for applications in information science and engineering.

Original languageEnglish
Article number108101
Number of pages4
JournalPhysical Review Letters
Volume92
Issue number10
DOIs
Publication statusPublished - 12 Mar 2004

Keywords

  • neural network
  • natural frequencies
  • synchronization
  • patterns
  • cortex
  • model

Cite this

Capacity of oscillatory associative-memory networks with error-free retrieval. / Nishikawa, Takashi; Lai, Ying-Cheng; Hoppensteadt, Frank C.

In: Physical Review Letters, Vol. 92, No. 10, 108101, 12.03.2004.

Research output: Contribution to journalArticle

Nishikawa, Takashi ; Lai, Ying-Cheng ; Hoppensteadt, Frank C. / Capacity of oscillatory associative-memory networks with error-free retrieval. In: Physical Review Letters. 2004 ; Vol. 92, No. 10.
@article{8c353c53ccfc4ac1983a53e003855704,
title = "Capacity of oscillatory associative-memory networks with error-free retrieval",
abstract = "Networks of coupled periodic oscillators (similar to the Kuramoto model) have been proposed as models of associative memory. However, error-free retrieval states of such oscillatory networks are typically unstable, resulting in a near zero capacity. This puts the networks at disadvantage as compared with the classical Hopfield network. Here we propose a simple remedy for this undesirable property and show rigorously that the error-free capacity of our oscillatory, associative-memory networks can be made as high as that of the Hopfield network. They can thus not only provide insights into the origin of biological memory, but can also be potentially useful for applications in information science and engineering.",
keywords = "neural network, natural frequencies, synchronization, patterns, cortex, model",
author = "Takashi Nishikawa and Ying-Cheng Lai and Hoppensteadt, {Frank C.}",
year = "2004",
month = "3",
day = "12",
doi = "10.1103/PhysRevLett.92.108101",
language = "English",
volume = "92",
journal = "Physical Review Letters",
issn = "0031-9007",
publisher = "American Physical Society",
number = "10",

}

TY - JOUR

T1 - Capacity of oscillatory associative-memory networks with error-free retrieval

AU - Nishikawa, Takashi

AU - Lai, Ying-Cheng

AU - Hoppensteadt, Frank C.

PY - 2004/3/12

Y1 - 2004/3/12

N2 - Networks of coupled periodic oscillators (similar to the Kuramoto model) have been proposed as models of associative memory. However, error-free retrieval states of such oscillatory networks are typically unstable, resulting in a near zero capacity. This puts the networks at disadvantage as compared with the classical Hopfield network. Here we propose a simple remedy for this undesirable property and show rigorously that the error-free capacity of our oscillatory, associative-memory networks can be made as high as that of the Hopfield network. They can thus not only provide insights into the origin of biological memory, but can also be potentially useful for applications in information science and engineering.

AB - Networks of coupled periodic oscillators (similar to the Kuramoto model) have been proposed as models of associative memory. However, error-free retrieval states of such oscillatory networks are typically unstable, resulting in a near zero capacity. This puts the networks at disadvantage as compared with the classical Hopfield network. Here we propose a simple remedy for this undesirable property and show rigorously that the error-free capacity of our oscillatory, associative-memory networks can be made as high as that of the Hopfield network. They can thus not only provide insights into the origin of biological memory, but can also be potentially useful for applications in information science and engineering.

KW - neural network

KW - natural frequencies

KW - synchronization

KW - patterns

KW - cortex

KW - model

U2 - 10.1103/PhysRevLett.92.108101

DO - 10.1103/PhysRevLett.92.108101

M3 - Article

VL - 92

JO - Physical Review Letters

JF - Physical Review Letters

SN - 0031-9007

IS - 10

M1 - 108101

ER -