For deep networks, the whole equals the sum of the parts.

Journal: The Behavioral and brain sciences
Published Date: Dec 6, 2023

Abstract

Deep convolutional networks exceed humans in sensitivity to local image properties, but unlike biological vision systems, do not discover and encode abstract relations that capture important properties of objects and events in the world. Coupling network architectures with additional machinery for encoding abstract relations will make deep networks better models of human abilities and more versatile and capable artificial devices.

Authors

  • Philip J Kellman
    Department of Psychology, University of California, Los Angeles, Los Angeles, California, United States of America.
  • Nicholas Baker
    Department of Psychology, University of California, Los Angeles, Los Angeles, California, United States of America.
  • Patrick Garrigan
    Department of Psychology, St. Joseph's University, Philadelphia, PA, USA patrick.garrigan@sju.edu; https://sjupsych.org/faculty_pg.php.
  • Austin Phillips
    Department of Psychology, University of California, Los Angeles, Los Angeles, CA, USA asphillips@ucla.edu; https://kellmanlab.psych.ucla.edu/.
  • Hongjing Lu
    Department of Statistics, UCLA, Los Angeles, CA 90095, USA.

Keywords

External Resources

Popular Topics
Recent Journals