|
inferencing in Cyc The Cyc inference engine performs general logical deduction (including modus ponens, modus tollens, and universal and existential quantification), with AI's well-known named inference mechanisms (inheritance, automatic classification, etc.) as special cases. Cyc performs best-first search over proof-space using a set of proprietary heuristics, and uses microtheories to optimize inferencing by restricting search domains. Because the Cyc KB contains hundreds of thousands of assertions (aka "rules"), many approaches commonly taken by other inference engines (such as frame-based expert system shells, RETE match, Prolog, etc.) just don't scale up to KBs of this size. As a result, the Cyc team has been forced to develop other techniques. Cyc also includes several special-purpose inferencing modules for handling a few specific classes of inference. One such module handles reasoning concerning collection membership/disjointness. Others handle equality reasoning, temporal reasoning, and mathematical reasoning.
For more information, see the detailed description of CycL language features.
|