Comparing Human Concept Acquisition to Models in a Cognitive Architecture

Abstract

A study funded by UK DERA at the University of Nottingham in the mid-1990s examined performance on a concept formation task, where subjects had to classify schematic aeroplanes as being either USA' or Australian'. Subjects displayed poor performance, but more intriguingly exhibited a wide range of variability. A simple model of the same task, constructed in ACT-R, also displayed great variability from run to run. The present project aimed to investigate the reasons for the variability in the model, and if possible also in human subjects; and also to understand better the nature of concept representation in this class of model. This report, (a) explains the model, (b) contents critically on aspects of the original studies, (c) analyses components of the variability, (d) offers an account for the variability in terms of random walk processes within the ACT-R learning mechanism, and (e) outlines a graphical depiction of the representation and gradual acquisition of the concept within the model.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 2002
Accession Number
ADA406924

Entities

People

  • Anna L. Cox
  • Richard M. Young

Organizations

  • University of Hertfordshire

Tags

Communities of Interest

  • Air Platforms

DTIC Thesaurus Topics

  • Acquisition
  • Air Force
  • Air Force Research Laboratories
  • Aircrafts
  • Airplanes
  • Coding
  • Cognitive Science
  • Concept Formation
  • Contracts
  • Experimental Design
  • Notation
  • Probability
  • Psychology
  • Random Walk
  • Statistical Analysis
  • Training
  • Universities

Readers

  • Computational Modeling and Simulation
  • European Security and Defence Policy (ESDP).
  • Theoretical Analysis.