Memory search using complex dynamics in a recurrent neural network model

Authors:
Shigetoshi Nara;Peter Davis;Hiroo Totsuji
Affiliations:
Okayama University, Japan;ATR Optical and Radio Communications Research Laboratories, Japan;Okayama University, Japan
Venue:
Neural Networks
Year:
1993

Citing 3
Cited 7

Clustering, coding, switching, hierarchical ordering, and control in a network of chaotic elements

Physica D
Original Contribution: Dynamic link of memory-Chaotic memory map in nonequilibrium neural networks

Neural Networks
Learning Patterns and Pattern Sequences by Self-Organizing Nets of Threshold Elements

IEEE Transactions on Computers

Switched-Capacitor Multi-Internal-State Chaotic Neuron Circuit with Unipolar and Bipolar Output Functions

MICRONEURO '99 Proceedings of the 7th International Conference on Microelectronics for Neural, Fuzzy and Bio-Inspired Systems
A solution for two-dimensional mazes with use of chaotic dynamics in a recurrent neural network model

Neural Computation
2005 Special issue: A mixed analog/digital chaotic neuro-computer system for quadratic assignment problems

Neural Networks - 2005 Special issue: IJCNN 2005
Chaotic dynamics for multi-value content addressable memory

Neurocomputing
Tracking a Moving Target Using Chaotic Dynamics in a Recurrent Neural Network Model

Neural Information Processing
Large memory capacity in chaotic artificial neural networks: a view of the anti-integrable limit

IEEE Transactions on Neural Networks
Markovian perturbations of discrete iterations: Lyapunov functions, global minimization, and associative memory

Mathematical and Computer Modelling: An International Journal

Quantified Score

Hi-index	0.01

Visualization

Abstract

Complex dynamics is expected to enable us to avoid combinatorial explosion and diverging program complexity that has been one of the essential difficulties in solving complex problems (for instance, an ill-posed problem) by serial processing algorithms. In this paper, it is shown by numerical investigations that a complicated memory search task is executable using complex dynamics in a recurrent neural network model with asymmetric synaptic connection. In spite of the simplicity of the proposed search algorithm, the complexity of chaotic wandering orbit in the state space offiring pattern leads us to a certain number of successful cases of search tasks with better efficiencies than random search. It is shown that the dynamical structure of chaotic orbit has a strong influence on the efficiency of search performance. A learning rule is proposed in order to realize ''constrained chaos'' that has, in the state space, a wandering structure suited to a given search task.