The recurrent IML-network

Authors:
Joern Fischer
Affiliations:
Fraunhofer Gesellschaft, Autonomous Intelligent Systems, Augustin, Germany
Venue:
IWANN'03 Proceedings of the Artificial and natural neural networks 7th international conference on Computational methods in neural modeling - Volume 1
Year:
2003

Citing 5
Cited 0

Competitive learning algorithms for vector quantization

Neural Networks
A new competitive learning approach based on an equidistortion principle for designing optimal vector quantizers

Neural Networks
Color Image Segmentation using Competitive Learning

IEEE Transactions on Pattern Analysis and Machine Intelligence
Unsupervised fuzzy clustering

Pattern Recognition Letters
A self-organizing network for hyperellipsoidal clustering (HEC)

IEEE Transactions on Neural Networks

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this paper we propose a multivaluated recurrent neural network for vector quantization where the synaptic potential is given by a weigted sum of values of a function that evaluates the consensus between the states of the process units. Each process unit presents the state wilh the largest activation potential, that is, it depends on the state of the nearest process units (more strongly connected according to the synaptic weights). Like Hopfield network, it uses a computational energy function that always decreases (or remains constant.) as the system evolves according to its dynamical rule based on an energy function that, is equivalent to the distortion function of the vector quantization problem. It does not use tuning parameters and so it attains computational efficiency.