Degree Type

Dissertation

Date of Award

1991

Degree Name

Doctor of Philosophy

Department

Computer Science

First Advisor

Suresh C. Kothari

Abstract

The conventional computer systems can solve complex mathematical problems very fast, yet it can't efficiently process high-level intelligent functions of human brain such as pattern recognition, categorization, and associative memory;A neural network is proposed as a computational structure for modeling high-level intelligent functions of human brain. Recently, neural networks have attracted considerable attentions as a novel computational system because of the following expected benefits which are often considered as generic characteristics of human brain: (1) massive parallelism, (2) learning as a means of efficient knowledge acquisition, and (3) robustness arising from distributed information processing;Neural networks are being studied from a different point of view in many disciplines such as psychology, mathematics, statistics, physics, engineering, computer science, neuroscience, biology, and linguistics. Depending on disciplines, neural networks have diverse nomenclature as artificial neural networks, connectionism, PDPs, adaptive systems, adaptive networks, and neurocomputers;We study the neural networks from the computer scientist's point of view. The objectives of this research work are: (1) providing a global picture of the current state of the art by surveying a score of neural networks chronologically and functionally, (2) providing a theoretical justification for well-known empirical results about the information capacity of Hopfield neural network, and (3) providing an experimental logical database system using Hopfield neural network as an inference engine.

DOI

https://doi.org/10.31274/rtd-180813-9227

Publisher

Digital Repository @ Iowa State University, http://lib.dr.iastate.edu/

Copyright Owner

Kesig Lee

Language

en

Proquest ID

AAI9126214

File Format

application/pdf

File Size

86 pages

Share

COinS