BOOLEAN MATRICES AND THE STABILITY OF NEURAL NETS

Abstract

THE FUNCTIONAL STATES OF A NEURAL (COMBINATORIAL) NET OF PRESCRIBED INTERCONNECTION GEOMETRY AND OVER-ALL NET FUNCTION ARE CHARACTERIZED AND ENUMERATED. The problem i formalized for nets having an arbitrary number of inputs and outputs by introducing the notion of a Boolean matrix. Various theorems forming a basis for the treatment of other neural net problems are proved.A discu sion of the concept of staility leads to a definition of a measure of net stability closely related to the set of functional s tes over which the input-output function i invariant. (Author)

Document Details

Document Type
Technical Report
Publication Date
Jan 01, 1962
Accession Number
AD0275323

Entities

People

  • R.h. Urbano

Organizations

  • Air Force Cambridge Research Laboratories

Tags

DTIC Thesaurus Topics

  • Geometry

Fields of Study

  • Mathematics

Readers

  • Graph Algorithms and Convex Optimization.
  • Neural Network Machine Learning.
  • Snow Cover Descriptors for Reptiles and Their Illustrations.