A high-level network of neural classifiers

College

College of Computer Studies

Department/Unit

Information Technology

Document Type

Archival Material/Manuscript

Publication Date

2001

Abstract

The high-level Neural Network model described in this paper is a multi-layered feedforward network where each hidden and output unit is also a Neural Network. Each of the units which compose the Neural Network, termed classifier unit, is an incremental network that adjusts its architecture depending on the complexity of the input-output association task that is assigned to it. The various ways by which such a high-level Neural Network can learn are presented. These are discussed in the context of hybrid systems which incorporate the advantages of Expert Systems and Neural Networks.

html

Disciplines

Computer Sciences

Note

Undated; Publication/creation date supplied

Keywords

Neural networks (Computer science); Hybrid systems

Upload File

wf_no

This document is currently not available here.

Share

COinS