Abstract

A knowledge acquisition tool should provide a user with maximum guidance in extending and debugging a knowledge based system, by preventing inconsistencies and knowledge gaps that may arise inadvertently. At the same time, it should free the user from the need to know the details of how the system is organized and implemented. Most current acquisition tools guide the acquisition process having expectations about the form and content of knowledge that will be added. Because these expectations are built into the tool, the tools are not flexible. In this paper, we present the EXPECT framework, which achieves greater flexibility by analyzing the structure of the knowledge base to dynamically form expectations rather than using expectations that are built into the acquisition tool. Achieving this flexibility requires changes to the knowledge based system architecture to make it more declarative, and hence, easier to analyze, and to the knowledge acquisition tool itself. In the EXPECT architecture, domain facts and goals are represented declaratively, and the problem solver keeps records of their functionality within the task domain. When the user corrects the system's knowledge, EXPECT tracks any possible implications of this change in the overall system and cooperates with the user to correct any additional errors that may arise. The key of the flexibility of this knowledge acquisition tool is in that it adapts its guidance as the knowledge bases evolve in response to changes introduced by the user.

Date

1994

Authors

Yolanda Gil, WR Swartout

Journal

Proceedings of the 1994 Workshop of the ARPA-Rome Laboratory Knowledge-Based Planning and Scheduling Initiative