Using access paths to guide inference with conceptual graphs

Abstract

Conceptual Graphs (CGs) are a natural and intuitive notation for expressing first-order logic statements. However, the task of performing inference with a large-scale CG knowledge base remains largely unexplored. Although basic inference operators are defined for CGs, few methods are available for guiding their application during automated reasoning. Given the expressive power of CGs, this can result in inference being intractable.

In this paper we show how a method used elsewhere for achieving tractability — namely the use of access paths — can be applied to conceptual graphs. Access paths add to CGs domain-specific information that guides inference by specifying preferred chains of subgoals for each inference goal (and hence, other chains will not be tried). This approach trades logical completeness for focussed inference, and allows incompleteness to be introduced in a controlled way (through the knowledge engineer's choice of which access paths to attach to CGs). The result of this work is an inference algorithm for CGs that significantly improves the efficiency of reasoning.