Experimental Studies of Integrated Cognitive Systems

Abstract

In this paper, we examine the issues that arise in the experimental study of integrated cognitive systems. We review the reasons why such artifacts are difficult to evaluate, then consider some dependent measures that can be used to characterize their behavior. Next we discuss independent variables that can influence this behavior, in particular features of the domain and characteristics of the system, including its knowledge and experience. We then turn to domains and testbeds that support experiments with such systems, giving examples of some promising candidates. We conclude with a discussion of the scientific goals of experimentation, which involve understanding the mapping from domain and system characteristics onto behavior.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 2004
Accession Number
ADA515755

Entities

People

  • Elena Messina
  • Pat Langley

Organizations

  • Stanford University

Tags

Communities of Interest

  • Air Platforms
  • Autonomy

DTIC Thesaurus Topics

  • Aircrafts
  • Algorithms
  • Artificial Intelligence
  • Classification
  • Communities
  • Computational Science
  • Data Sets
  • Information Science
  • Intelligent Agents
  • Intelligent Systems
  • Language
  • Machine Learning
  • Search And Rescue
  • Standards
  • Statistical Tests
  • Test And Evaluation
  • Urban Areas

Readers

  • Systems Analysis and Design
  • Team-Based Human-Centered Cognitive Task Decision Making and Information Performance.