Retrieval process of an associative memory with a general input-output function

Hidetoshi Nishimori, Ioan Opris

Research output: Contribution to journalArticle

26 Citations (Scopus)

Abstract

The retrieval process of an associative memory with a general input-output function is studied by means of a signal-to-noise ratio analysis. We derive a set of recursion relations for macroscopic variables to describe the time development of the network. By taking the equilibrium limit of the recursion relations, we find that a certain type of nonmonotonic input-output relation of a single neuron yields an enhanced memory capacity of the network compared with the conventional monotonic relation. This behavior is in agreement with the prediction of Morita et al. who used numerical simulations as well as geometrical arguments to reach their conclusion. Our method reveals the relation between the type of the input-output function and the memory capacity in a generic associative memory.

Original languageEnglish (US)
Pages (from-to)1061-1067
Number of pages7
JournalNeural Networks
Volume6
Issue number8
DOIs
StatePublished - Jan 1 1993
Externally publishedYes

Fingerprint

Data storage equipment
Signal-To-Noise Ratio
Neurons
Signal to noise ratio
Computer simulation

Keywords

  • Amari-Maginu theory
  • Associative memory
  • Gaussian noise
  • Input-output function
  • Macrovariables
  • Monte Carlo simulation
  • Retrieval dynamics
  • Signal-to-noise ratio analysis

ASJC Scopus subject areas

  • Cognitive Neuroscience
  • Artificial Intelligence

Cite this

Retrieval process of an associative memory with a general input-output function. / Nishimori, Hidetoshi; Opris, Ioan.

In: Neural Networks, Vol. 6, No. 8, 01.01.1993, p. 1061-1067.

Research output: Contribution to journalArticle

@article{39b3ce7f98c9478caa3612806505e753,
title = "Retrieval process of an associative memory with a general input-output function",
abstract = "The retrieval process of an associative memory with a general input-output function is studied by means of a signal-to-noise ratio analysis. We derive a set of recursion relations for macroscopic variables to describe the time development of the network. By taking the equilibrium limit of the recursion relations, we find that a certain type of nonmonotonic input-output relation of a single neuron yields an enhanced memory capacity of the network compared with the conventional monotonic relation. This behavior is in agreement with the prediction of Morita et al. who used numerical simulations as well as geometrical arguments to reach their conclusion. Our method reveals the relation between the type of the input-output function and the memory capacity in a generic associative memory.",
keywords = "Amari-Maginu theory, Associative memory, Gaussian noise, Input-output function, Macrovariables, Monte Carlo simulation, Retrieval dynamics, Signal-to-noise ratio analysis",
author = "Hidetoshi Nishimori and Ioan Opris",
year = "1993",
month = "1",
day = "1",
doi = "10.1016/S0893-6080(09)80017-8",
language = "English (US)",
volume = "6",
pages = "1061--1067",
journal = "Neural Networks",
issn = "0893-6080",
publisher = "Elsevier Limited",
number = "8",

}

TY - JOUR

T1 - Retrieval process of an associative memory with a general input-output function

AU - Nishimori, Hidetoshi

AU - Opris, Ioan

PY - 1993/1/1

Y1 - 1993/1/1

N2 - The retrieval process of an associative memory with a general input-output function is studied by means of a signal-to-noise ratio analysis. We derive a set of recursion relations for macroscopic variables to describe the time development of the network. By taking the equilibrium limit of the recursion relations, we find that a certain type of nonmonotonic input-output relation of a single neuron yields an enhanced memory capacity of the network compared with the conventional monotonic relation. This behavior is in agreement with the prediction of Morita et al. who used numerical simulations as well as geometrical arguments to reach their conclusion. Our method reveals the relation between the type of the input-output function and the memory capacity in a generic associative memory.

AB - The retrieval process of an associative memory with a general input-output function is studied by means of a signal-to-noise ratio analysis. We derive a set of recursion relations for macroscopic variables to describe the time development of the network. By taking the equilibrium limit of the recursion relations, we find that a certain type of nonmonotonic input-output relation of a single neuron yields an enhanced memory capacity of the network compared with the conventional monotonic relation. This behavior is in agreement with the prediction of Morita et al. who used numerical simulations as well as geometrical arguments to reach their conclusion. Our method reveals the relation between the type of the input-output function and the memory capacity in a generic associative memory.

KW - Amari-Maginu theory

KW - Associative memory

KW - Gaussian noise

KW - Input-output function

KW - Macrovariables

KW - Monte Carlo simulation

KW - Retrieval dynamics

KW - Signal-to-noise ratio analysis

UR - http://www.scopus.com/inward/record.url?scp=0027139024&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0027139024&partnerID=8YFLogxK

U2 - 10.1016/S0893-6080(09)80017-8

DO - 10.1016/S0893-6080(09)80017-8

M3 - Article

VL - 6

SP - 1061

EP - 1067

JO - Neural Networks

JF - Neural Networks

SN - 0893-6080

IS - 8

ER -