Veri cation and Validation of KBS With Neural Network Components

Artiicial Neural Networks(ANN) play an important role in developing robust Knowledge Based Systems(KBS). The ANN based components used in these systems learn to give appropriate predictions through training with correct input-output data patterns. Unlike traditional KBS that depends on a rule database and a production engine, the ANN based system mimics the decisions of an expert without speciically formulating the if-then type of rules. In fact, the ANNs demonstrate their superiority when such if-then type of rules are hard to generate by human expert. Veriication of traditional knowledge based system is based on the proof of consistency and completeness of the rule knowledge base and correctness of the production engine. These techniques , however, can not be directly applied to ANN based components. In this position paper, we propose a veriication and validation procedure for KBS with ANN based components. The essence of this procedure is to obtain an accurate system speciication through incremental modii-cation of the speciications using an ANN rule extraction algorithm. First, the ANN based components are speciied using available domain knowledge and implemented based on this speciica-tion. Next, past data sets are used to train the ANN components. An rule extraction algorithm is then applied to the trained ANN. Extracted rules are then analyzed and incrementally incorporated into the system speciication. Finally, the modiied speciications are veriied for cor-rectness and the product tested against the correct speciications.