Research On Knowledge Representation And Inference Method Based On Extenics

The artificial intelligence is being in the period of vigorous development. The close combination between it and computer has initiated the people into the research of knowledge processing in which knowledge representation and knowledge reasoning are especially more important. There are many methods of knowledge expression and reasoning which play an important role in the intelligent systems. However, these intelligent systems have only simulated the abstract logical thinking activities of human, scarcely simulated creative thinking activities. In order to overcome the shortcomings of current intelligent systems, in this paper the extension theory is applied to the research of knowledge representation and reasoning.First, the background, the goal and the significance of this thesis are introduced, and the elementary knowledge of artificial intelligence and extension theory are simply illustrated in the paper. Second, the applications of extension methods in artificial intelligence, especially the close relation between it and knowledge representation and reasoning, are analyzed in detail. Then the formal definition of extension knowledge representation is proposed and compared with several original methods of knowledge representation.In the end, the divergent extension reasoning is proposed according to knowledge extension. Based on it the concept of "inherit" is introduced and further the extension inheritance reasoning is proposed. These two new reasoning methods are illustrated detailedly with the example of an animal recognition system in the paper.