Scaffolding in teaching knowledge representation

  • Authors:

  • Daniel Bryce

  • Affiliations:

  • Utah State University, Old Main Hill, Logan UT

  • Venue:
  • Journal of Computing Sciences in Colleges
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

Knowledge representation is one of the most difficult tasks for students in introductory Artificial Intelligence courses. We study the impact of varying levels and types of scaffolding in teaching students how to model automated planning problems in a first-order logic-like language called the planning domain definition language (PDDL). We select two well-known problems: the sliding tile puzzle and the towers of Hanoi for our evaluation. The results suggest that scaffolding does increase proficiency, and that the type of scaffolding gives rise to different obstacles in learning. We discuss these learning obstacles and implications that they have upon the modeling process.