Designing Scripts to Guide Users in Modifying Knowledge-Based Systems

Proceedings of the AAAI Conference on Artificial Intelligence, 16

Volume

Proceedings of the AAAI Conference on Artificial Intelligence, 16

Knowledge Representation

Knowledge Acquisition (KA) Scripts capture typical modification sequences that users follow when they modify knowledge bases. KA tools can use these scripts to guide users in making these modifications, ensuring that they follow all the ramifications of the change until it is completed. This paper describes our approach to design, develop, and organize a library of KA Scripts. We report the results of three different analysis to develop this library, including a detail study of actual modification scenarios in two knowledge bases. In addition to identifying a good number of KA Scripts, we found a set of useful attributes to describe and organize the KA Scripts. These attributes allow us to analyze the size of the library and generate new KA Scripts in a systematic way. We have implemented a portion of this library and conducted two different studies to evaluate it. The result of this evaluation showed a 28 to 36 percent time savings in modifying knowledge bases and that the library included relevant and useful KA Scripts to assist users in realistic settings.