Distributed associative memory and the computation of membership functions

doi:10.1016/0020-0255(91)90076-7

Information Sciences

Volumes 57–58, September–December 1991, Pages 171-180

https://doi.org/10.1016/0020-0255(91)90076-7 Get rights and content

Abstract

The purpose of this paper is to introduce an extension and generalizations to sparse distributed memory introduced by Kanerva [7], which we shall call distributed associative memory (DAM), concerning both the manner in which information is stored and techniques for improving the reading mechanism. Its effectiveness as a learning tool is discussed using a statistical model. We then show how distributed associative memory can be used to compute membership functions for decision-making under uncertainty.

References (11)

J. Bourrely et al.
Mémoires associatives: évaluation et architectures
Comptes-.Rendus Acad. Science de Paris
(1989)
C.F. Foster
Content Addressable Parallel Processors
(1976)
E. Gelenbe
Random neural networks with negative and positive signals and product form solution
Neural Computation
(1989)
E. Gelenbe
An evaluation of sparse associative memory
Neural Computation
(1990)
D.O. Hebb
The Organization of Behavior
(1984)

There are more references available in the full text version of this article.

Cited by (10)

Survey of random neural network applications
2000, European Journal of Operational Research
This paper consists of a survey of various engineering applications based on the random neural network (RNN) model [Neural Computation 1(4) (1989) 502–511; 2(2) (1990) 239–247; Comptes-Rendus Acad. Sci. t. Série II 310(2) (1990) 177–180; Applied Math. Modelling 15 (1991) 534–541; Neural Computation 5(1) (1993) 154–164], and also a summary of the recent image processing techniques such as still image compression, image enlargement, and image fusion. The advantage of the RNN model is that it is closer to biophysical reality and mathematically more tractable than standard neural methods, especially when used as a recurrent structure.
Reinforcement learning with internal expectation for the random neural network
2000, European Journal of Operational Research
The reinforcement learning scheme proposed in Halici (1977) (Halici, U., 1997. Journal of Biosystems 40 (1/2), 83–91) for the random neural network (Gelenbe, E., 1989b. Neural Computation 1 (4), 502–510) is based on reward and performs well for stationary environments. However, when the environment is not stationary it suffers from getting stuck to the previously learned action and extinction is not possible. In this paper, the reinforcement learning scheme is extended by introducing a weight update rule which takes into consideration the internal expectation of reinforcement. With the proposed scheme, the system behaves as in learning with reward when the reward for the learned action is not below the internal expectation, otherwise it behaves as in learning with punishment so that other possibilities can be explored. Such a scheme has made extinction possible while resulting in a good convergence to the most rewarding action.
Associative list memory
1997, Neural Networks
This paper introduces an Associative List Memory (ALM) that has high recall fidelity with low memory and low processing requirements. This permits a simple implementation in software on a personal computer or space instrument microprocessor. Associative List Memory has a performance comparable with Sparse Distributed Memory (SDM) but differs from SDM in that convergence occurs during learning, rather than on recall, and in that the memory is in the form of a dynamic list rather than static randomly distributed locations. Associative List Memory is suitable for unsupervised finding of classes of phenomena in large databases. In particular, all of the class exemplars deduced can be easily accessed at any time to provide a summary of current database knowledge, being essentially the contents of the list. Examples are given where patterns of 1000 bits length with > 30% noise can be learned unsupervised to deduce the original pattern's noise free. A second pass through the data in recall mode can be used to assign to each input the appropriate original pattern, effectively removing all noise from the input data. At large input bit sizes the recall fidelity approaches closely to the maximum possible value. Associative List Memory compares well in recall fidelity with SDM and other associative memories. Its processing times on a personal computer are found to be practical for database applications. Implemented within a space instrument processor, ALM would greatly reduce downlink data transmission rates. © 1997 Elsevier Science Ltd.
G-networks and their applications to machine learning, energy packet networks and routing: Introduction to the special issue
2017, Probability in the Engineering and Informational Sciences
Random neural network model for supervised learning problems
2015, Neural Network World
A cascadable random neural network chip with reconfigurable topology
2010, Computer Journal

View all citing articles on Scopus

View full text

Distributed associative memory and the computation of membership functions

Abstract

Mémoires associatives: évaluation et architectures

Comptes-.Rendus Acad. Science de Paris

Content Addressable Parallel Processors

Random neural networks with negative and positive signals and product form solution

Neural Computation

An evaluation of sparse associative memory

Neural Computation

The Organization of Behavior