SELF-LEARNING AUTOMATIC SYSTEMS (SELECTED ARTICLES),

Abstract

Algorithms for teaching a machine to recognize patterns without encouragement are considered in this paper. Objects are presented to the machine which belong to different patterns and all that is indicated is the number of classes into which these objects are to be divided; the machine is provided with no information as to which pattern each object presented belongs. What is required is that after the 'self-instruction' process the division of the objects by the machine into classes coincide with the true and factually existing breakdown. A receptor space X is considered, such that a point in this space corresponds to each object presented to the machine. For the sake of simplicity, it is further assumed that there are two classes: A and B. Just as in the case of teaching a machine to recognize patterns with encouragement, the purpose of the automation in this article is to draw a surface which will divide these two sets. A graphic method of representation is employed. (Author)

Document Details

Document Type
Technical Report
Publication Date
Sep 22, 1967
Accession Number
AD0670116

Entities

Organizations

  • National Air and Space Intelligence Center

Tags

DTIC Thesaurus Topics

  • Adaptive Control Systems
  • Adaptive Systems
  • Algorithms
  • Automatic
  • Automation
  • Control Systems
  • Education
  • Instructions
  • Learning
  • Mathematics

Readers

  • Computational Linguistics
  • Mathematical Modeling and Probability Theory.
  • STEM Education

Technology Areas

  • Space
  • Space - Space Objects