Contents

 
• List of Figures
• List of Tables
• Introduction
  • The Project
  • Biological Neural Networks
  • Definitions
  • Memorization and Generalization
  • Acquiring Knowledge by Learning
  • Approaches to Neural computing
    • The Connectionist Approach
    • The Weightless Logical Approach
  • Applications
• Multilayer Networks and Backpropagation
  • An Artificial Neuron
  • A Single Layer Network
  • Multilayer Neural Network
  • Other Methods to Train MLPs
  • Capabilities and Limitations of MLPs
• Logical Neural Networks
  • An Adaptive Node
  • A RAM Based Logical Neural Network
  • Selecting a Logical Network
  • An Example of a Logical Neural Network
  • Improvements on Logical Neural Networks
• Modularity
  • Modularity in the Nervous System
  • Modularity Derived from Psychology
  • Modularity in ANN
    • Motivation: 4-Bit-Parity Problem
  • Multiple and Modular Artificial Neural Networks
  • Publications on Aspects of Modular Neural Network
    • Modularity as a Powerful Concept
    • Reducing the Complexity of the Problem
  • Modular Neural Network Applications
• A New Modular Neural Network
  • The Evolution of the Model
  • The Network Architecture
  • Training the System
  • Calculation of the Output
  • Analysis of the New Model
    • A Fast Network
    • Learning Problems
    • On Generalization
• The Implementation
  • The Concept
  • The Software Platform
  • Preprocessing of the Data
  • The Neural Network Library
    • New Data Types
    • The Class CData
    • The Class CSupport
    • The Class CEncode
    • The Class CCell
    • The Class CLayer
    • The Class CBackPro
    • The Class CMLayer
    • The Class CMul2L
    • The Class CUserIF
  • The File Formats
    • The Data File
    • The Network Description File
    • The Error Output File
    • The Logfile
  • The Developed Programs
    • The Modular Neural Network: modnet
    • The Monolithic Neural Network: bpnet
    • The Logical Neural Network: logicnet
• Experimental Evaluation
  • Test Methodology
  • Binary Pattern Recognition
  • Problems with a Small Input Dimension
    • Prediction on Diabetes Data
    • Prediction on Heart Disease Data
    • Prediction on Credit Card Data
    • Classification of Radar Signals
    • Classification of Sonar Data
  • Problems with a Large Continuous
    Input Space
    • Discrimination Between Good and Faulty Wood
    • Gray Scale Picture Recognition
      • Manually Distorted Pictures
      • Noisy Pictures
  • Experiments on the Fault Tolerance of the System
• Conclusion
  • Further Work
• Conference Paper
• Example Files
  • A Typical Logfile
  • The Network Description File
• Binary Pictures
• References
• About this document ...