A New Model for Knowledge Representation and Automatic Deduction for A.I applications

摘要

In spite of the variety of the given Artificial intelligence definitions, those definitions remain all rights on the fact that the purpose of A.I, is to translate the intelligent human behaviour to the computer. This do necessary, the modelling of the said intelligent human behaviour. However, the modelling task is not at all evident one. When the considered application is related to the human intuition (i.e acquiring and use knowledge), the modelling process move away from formal, and depends essentially on the expert perception of the human comportment to be modelled. Note that the expert perception mights be incomplete, and depends on research results given on each domain related to the human cognition. In spite of that, results given by some models conceived and applied in A.I domain, gave an encouraging results. Most persuasive example can be the neural networks. In this paper, we will present a simple model oriented to knowledge acquiring and use; we mean by use, the task commonly called "deduction". We based our model on some underlined human cognition behaviours. The main remarkable characteristics of that model can be the high level of abstraction of the model, the flexibility to adjunction, suppression and substitution, the graphical simple representation, and the tolerance to errors which can occur during knowledge acquisition.