Title

A comparison of two neural network architectures for vector quantization

Document Type

Conference Proceeding

Publication Date

12-1-1991

Abstract

The authors investigate the performance of two neural network architectures for vector quantization (VQ). The two architectures are the multilayer feedforward network and the Hopfield analog neural network. It is found that for the feedforward network to have reasonably good performance, the number of hidden units must be unrealistically high: exponential in the number of dimensions and codewords. For the Hopfield analog model, on the other hand, the number of processors required is equal to the number of codewords and the resulting performance is very close to the optimum mean squared error.

Publication Source (Journal or Book title)

Proceedings. IJCNN-91-Seattle: International Joint Conference on Neural Networks

First Page

391

Last Page

396

This document is currently not available here.

COinS