COLLECTED TECHNICAL PAPERS. VOLUME I

Abstract

THE PERCEPTRON: A MODEL FOR BRAIN FUNCTION, by H. D. Block: A perceptron consists of a retina of sensory units (for example photocells); these are connected (for example by wires) to associator units. When a stimulus is presented to the retina (for example as a pattern of illumination) impulses are conducted from the activated sensory units to the associator units. If the total signal arriving at an associator unit exceeds a certain threshold, then the associator becomes active and sends an impulse to units to which it is connected. The magnitude of the impulse carried by certain connections depends on the past activity of their termini according to certain preassigned reinforcement rules. ANALYSIS OF A 4-LAYER SERIES-COUPLED PERCEPTRON, by H. D. Block, B. W. Knight, Jr., and F. Rosenblatt: The perceptron of this paper is self-organizing in the sense that during the training period the experimenter does not tell the machine the category of each stimulus. The only contact between the experimenter and the machine is the presentation of the stimuli. PRELIMINARY EXPERIMENTS ON PERCEPTRON APPLICATION TO BUBBLE CHAMBER EVENT RECOGNITION, by Carl Kesler: Computer simulations are described which relate to the general problem of building a perceptron to recognize the various types of events in bubble chamber photographs.

Document Details

Document Type
Technical Report
Publication Date
Jan 01, 1961
Accession Number
AD0253960

Entities

People

  • Carl Kesler
  • H.d. Block

Organizations

  • Cornell University

Tags

DTIC Thesaurus Topics

  • Bubble Chambers
  • Chambers
  • Computer Simulations
  • Computers
  • Illumination
  • Photoelectric Cells (Semiconductor)
  • Photographs
  • Photography
  • Recognition
  • Simulations
  • Simulators
  • Training

Readers

  • Neural Network Machine Learning.
  • Statistical inference.
  • Vision Science/Vision Psychology/Cognitive Neuroscience.