Using Prior Knowledge: Problems and Solutions

Abstract

Encoding knowledge is time consuming and expensive. A possible solution to reduce the cost of developing a new knowledge base (KB) is to reuse existing knowledge. Previous work addressing this problem has focused on standards for representing, exchanging, and accessing knowledge (Genesereth and Fikes 1992), (Chaudhri et al. 1998), and on creating large repositories of knowledge (Lenat and Guha 1990). Results on the level of reuse achievable have been reported (Cohen et al. 1999). In this paper, we focus on the process of reuse and report a case study on constructing a KB by reusing existing knowledge. The reuse process involved the following steps: translation, comprehension, slicing, reformulation, and merging. We discuss technical problems encountered at each of these steps and explain how we solved them.