KRUSTWorks Knowledge Refinement

The KRUSTWorks toolkit allows a knowledge engineer to construct a refinement tool, customised for a particular rule-based system. The refinement tool applied Machine Learning techniques to identify possible faults, and suggest potential repairs, to improve the problem-solving of the target knowledge-based system (KBS). A refinement framework defined a set of generic KBS concepts and refinement steps, and an extensible toolkit provided refinement components to achieve these steps. KRUSTWorks provides refinement facilities that apply to KBSs developed using a number of different KBS tools (e.g. NASA's Clips, IntelliCorp's Kappa, Logica's Formulogic). The research was highly relevant to IntelliCorp, was successfully applied to 2 aerospace KBSs from ESA and NASA. Compared to AstraZeneca's manual knowledge engineering of a rule-based tablet formulation system, KRUSTWorks removed faults from buggy rules and successfully re-engineered the knowledge after a substantial change in formulation policy.    [More] Partners AstraZeneca, IntelliCorp, LogicaCMG Funding EPSRC GR/L38387 Research Team Susan Craw (PI) Robin Boswell Nirmalie Wiratunga Gareth Palmer Related Projects Complexity Modelling for CBR Systems

Publications

Wiratunga, N., & Craw, S. (2001). Informed selection of filter examples for knowledge refinement. 21st BCS SGES Expert Systems 01 Conference, Cambridge, England

Boswell, R., & Craw, S. (2000). Experiences with a generic refinement toolkit. 12th International Conference on Knowledge Engineering and Knowledge Management (EKAW2000), Juan Les Pins, France

Craw, S., & Boswell, R. (2000). Debugging knowledge-based applications with a generic toolkit. 12th International Conference on Tools with AI (ICTAI-2000), Vancouver, Canada

Wiratunga, N., & Craw, S. (2000). Informed selection of training examples for knowledge refinement. 2th International Conference on Knowledge Engineering and Knowledge Management (EKAW2000), Juan Les Pins, France

Boswell, R., & Craw, S. (1999). Knowledge modelling for a generic refinement framework. In Knowledge Based Systems, Vol 12, page(s) 317-325

Craw, S., & Boswell, R. (1999). Representing problem-solving for knowledge refinement. Sixteenth National Conference on Artificial Intelligence (AAAI'99), Orlando, FL

Craw, S., Boswell, R., & Rowe, R. (1999). Automated knowledge refinement for rule-based formulation expert system. In Pharmaceutical Science and Technology Today, Vol 2(9), page(s) 383-385

Wiratunga, N., & Craw, S. (1999). Incorporating backtracking search with knowledge refinement. Validation and Verification of Knowledge Based Systems, Proceedings of the 5th European Symposium on the Validation and Verification of Knowledge Based Systems (EUROVAV'99), Oslo, Norway

Wiratunga, N., & Craw, S. (1999). Sequencing training examples for iterative knowledge refinement. 19th BCS SGES Expert Systems Conference, Cambridge, UK

Boswell, R., Craw, S., & Rowe, R. (1997). Knowledge Refinement for a Design System. 10th European Workshop Knowledge Acquisition, Modeling and Management (EKAW), Sant Feliu de Guixols, Spain

Craw, S., Boswell, R., & Rowe, R. (1997). Knowledge Refinement to Debug and Maintain a Tablet Formulation System. 9th IEEE International Conference on Tools with Artificial Intelligence, Newport Beach, CA

Craw, S. (1996). Refinement Complements Verification and Validation. In International Journal of Human-Computer Studies, Vol 44(2), page(s) 245-256

Palmer, G., & Craw, S. (1996). The Role of Test Cases in Automated Knowledge Refinement. 16th BCS SGES Expert Systems Conference, Cambridge, UK

Craw, S., & Sleeman, D. (1995). Refinement in Response to Validation. In Expert Systems with Applications, Vol 18(3), page(s) 343-349

Craw, S. (1993). Refinement Assists Quality Assurance. IJCAI Workshop on Validation of Knowledge-based Systems, Chambery, France

Craw, S., & Sleeman, D. (1991). The Flexibility of Speculative Refinement. 8th International Machine Learning Workshop, Evanston, IL

Craw, S., & Sleeman, D. (1990). Automating the Refinement of Knowledge-Based Systems. European Conference on Artificial Intelligence (ECAI'90), Stockholm, Sweden