Connectionism for fuzzy learning in rule-based expert systems
A novel approach to rule refinement based upon connectionism is presented. This approach is capable of performing rule deletion, rule addition, changing rule quality, and modification of rule strengths. The fundamental algorithm is referred to as the Consistent-Shift algorithm. Its basis for identifying incorrect connections is that incorrect connections will often undergo larger inconsistent weight shift than correct ones during training with correct samples. By properly adjusting the detection threshold, incorrect connections would be uncovered, which can then be deleted or modified. Deletion of incorrect connections and addition of correct connections then translate into various forms of rule refinement just mentioned.
KeywordsExpert System Neural Network
Unable to display preview. Download preview PDF.
- 1.Buchanan, B.G. and Shortliffe, E.H., Rule-Based Expert Systems, Addison-Wesley, Massachusetts, 1984.Google Scholar
- 3.Fu, L.M., Knowledge base refinement by backpropagation, to appear in Data and Knowledge Engineering, 7, 1992.Google Scholar
- 4.Holldobler, S., A connectionist unification algorithm, Technical report, ICSI-TR-90-012, International Computer Science Institute, Berkeley, CA., 1990.Google Scholar
- 5.Rumelhart, D.E., Hinton, G.E. and Williams, R.J., Learning internal representation by error propagation, In Parallel Distributed Processing: Explorations in the Microstructures of Cognition, Vol. 1, MIT press: Cambridge, 1986.Google Scholar
- 6.Towell, G.G., Shavlik, J.W., and Noordewier, M.O., “Refinement of approximate domain theories by knowledge-based neural networks”, in Proceeding of AAAI-90, BostonGoogle Scholar