Thresholding in Self-Organizing Machines

Abstract

The threshold unit is a simplified analog of the living neuron cell. Self-organizing machines were constructed, using the threshold unit as a Boolean function generator. These machines may be represented by a mathematical model. The model may be analyzed by vector and matrix methods. Analysis of a simple Perceptron by these methods indicates that the ability of the machine to identify noisy patterns may depend on the existence of one threshold unit which assigns all noisy patterns to the same class as the corresponding correct patterns. A method based on matrix representation of Karnaugh maps may be used as a design tool for analysis of threshold units and functions.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 1962
Accession Number
AD0286921

Entities

People

  • Robert B. Stuart

Organizations

  • Air Force Institute of Technology

Tags

Communities of Interest

  • Materials and Manufacturing Processes
  • Weapons Technologies

DTIC Thesaurus Topics

  • Air Force
  • Air Force Facilities
  • Algorithms
  • Artificial Intelligence
  • Circuits
  • Electrical Circuits
  • Engineering
  • Equations
  • Identification
  • Mathematical Models
  • Models
  • Neurons
  • Pattern Recognition
  • Probability
  • Recognition
  • Switching
  • Switching Circuits
  • United States

Fields of Study

  • Biology

Readers

  • Computer Engineering
  • Neural Network Machine Learning.
  • Theoretical Analysis.