A SIMPLE MODEL OF A PATTERN RECOGNITION SYSTEM.

Abstract

This model is to be used to demonstrate the functional structure of a learning machine or pattern recognition system that is similar in basic principle to the learning machine recently designed and constructed at SRI. In addition, the model shows the important role that mask design plays. A specific method for selecting ''efficient'' masks for this model, by computer simulation, is also presented. The model is a highly simplified and therefore not very close approximation to the learning machine, MINOS II. It can be expected to discriminate only among the three letters of the alphabet with which it has been trained. The paper model displays many of the fundamental features and logical-system design principles encountered in the full-scale machine, which is capable of performing tasks more interesting than recognizing three free-hand letters of the alphabet. The model exhibits the ability to distinguish between different classes (discrimination), and to ignore variations within a class (generalization), which are the basic properties of a pattern recognition system. (Author)

Document Details

Document Type
Technical Report
Publication Date
Apr 01, 1964
Accession Number
AD0600079

Entities

People

  • D. J. Hall

Organizations

  • SRI International

Tags

DTIC Thesaurus Topics

  • Alphabets
  • Computer Simulations
  • Computers
  • Discrimination
  • Learning
  • Learning Machines
  • Pattern Recognition
  • Recognition
  • Simulations
  • Simulators

Fields of Study

  • Computer science

Readers

  • Computational Linguistics
  • Theoretical Analysis.
  • Vision Science/Vision Psychology/Cognitive Neuroscience.

Technology Areas

  • AI & ML
  • AI & ML - Neural Networks