Learning to diagnose by doing

  • Authors:
  • Jayant Kalagnanam;Eswaran Subrahmanian

  • Affiliations:
  • Department of Engineering and Public Policy, Carnegie Mellon University, Pittsburgh, PA;Engineering Design Research Center, Carnegie Mellon University, Pittsburgh, PA

  • Venue:
  • IJCAI'89 Proceedings of the 11th international joint conference on Artificial intelligence - Volume 1
  • Year:
  • 1989

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper is a study on the process of evolution of a novice to an expert in a diagnostic context. In this paper, we have chosen an abstract example of a diagnostic problem. The results in this article are based on a longitudinal study of a single subject. The empirical base is a protocol of the subject as he solved this problem until he mastered the most sophisticated strategy. Based on an analysis of the protocol, we have identified four different strategies that were used by the subject to solve the given set of problems. These strategies vary in their efficiency of diagnosis and in their modes of reasoning. We also identify the different operators that were used by the subject to transform one strategy into a more efficient one. The learning process has been implemented as a computer simulation. Finally, we discuss the hypotheses that are suggested by this experiment and the implications of our observations.