STIMULUS SAMPLING PATTERN RECOGNITION

Abstract

A generalized computer model was constructed to simulate the behavior of organisms according to various versions of stimulus sampling theory. Results already reported dealt with the learning behavior of rats in a T-maze. With the addition of new routines to specify the simulated environment of the simulated organism, a similar basic program has learned to discriminate among visual patterns consisting of carelessly hand-drawn al phabetic characters, each of which is presented in a number of variations. Input to the program is a binary representation of the presence or absence of parts of the figure in a 20 x 20 matrix superimposed on the figure to be recognized. Each cell of the matrix is treated as a separate stimulus having various numbers of elements as sociated with all the responses that can be made. A random sample of elements from stimuli present in the visual field is used to determine the response any trial. When a correct response is made, all elements sampled on that trial become associated with that response. Changes in learning rates occur as the probability of selecting available stimulus elements is changed.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Mar 01, 1963
Accession Number
AD0406910

Entities

People

  • Frank N. Marzocco

Organizations

  • System Development Corporation

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Character Recognition
  • Computer Programming
  • Computer Programs
  • Computer Simulations
  • Computers
  • Environment
  • Government Procurement
  • Learning
  • Magnetic Tape
  • Mathematical Models
  • Models
  • Pattern Recognition
  • Probability
  • Punched Cards
  • Recognition
  • Sampling
  • Simulations

Fields of Study

  • Psychology

Readers

  • Computer Programming and Software Development.
  • Regression Analysis.
  • Vision Science/Vision Psychology/Cognitive Neuroscience.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference