Rule-Based Inference in Large Knowledge Bases

The process of inference can be conceived as the continual redescription of a given structure In the system knowledge base until a desired structure Is formed. Inference rules then represent single transformations to be used In this redescription process. If the rules are themselves Integrated into the knowledge base, the rule application mechanism can take advantage of knowledge base organization to maintain an efficient inference process, even in systems that must deal with large amounts of knowledge. This paper presents a methodology for implementing rule application, and gives two examples of its use in knowledge-based systems,