Proteus: Visuospatial analogy in problem-solving

  • Authors:
  • Jim Davies;Ashok K. Goel;Patrick W. Yaner

  • Affiliations:
  • Institute of Cognitive Science, Carleton University, Ottawa, Ont., Canada K1S 5B6;College of Computing Georgia Institute of Technology Atlanta, GA 30332-0760, USA;College of Computing Georgia Institute of Technology Atlanta, GA 30332-0760, USA

  • Venue:
  • Knowledge-Based Systems
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

We hypothesize that visuospatial knowledge alone is sufficient for analogical transfer of problem-solving procedures. We developed a theory of visuospatial analogical problem-solving. Proteus, our implemented program, provides a content account, with vocabulary and data structures, and a process account, with methods and algorithms. Proteus addresses all major subtasks of analogy, and identifies a new subtask: dynamic mapping generation between the intermediate knowledge states in the source and the target, when a step in the transferred procedure creates a new object. Examining the limitations of use of visuospatial knowledge alone, we help identify the functional roles of non-visuospatial knowledge.