Abstract

Role-limiting approaches have been successful for acquiring knowledge from domain experts. However most systems using this approach ask the user a pre-determined sequence of questions and are therefore not flexible enough for a wide range of situations. They also typically do not support acquiring problem-solving knowledge, but only instance and type information. We extend the role-limiting approach with a knowledge acquisition tool that dynamically generates questions for the user based on a background theory of the domain and on the user’s previous answers. The tool uses KA scripts to give an overall structure to a session. We show how to augment a background theory based on problem-solving methods to support this style of interaction. We present results from tool ablation experiments with domain experts in two different domains based on a problem solving method for plan evaluation. When the tool was used, tasks were completed in substantially less time than when the ablated version was used. In addition participants were able to complete substantially more tasks on average, twice as many in one of the studies.

Date

October 14, 2025

Authors

Jim Blythe, Y Gil

Journal

ECAI 2000 Workshop on Applications of Ontologies and Problem-Solving Methods