Knowledge Compilation for KLM-Style Defeasible Reasoning
摘要
In artificial intelligence, knowledge compilation involves transforming knowledge into a form that facilitates efficient reasoning. We introduce a method for knowledge compilation of ranked interpretations within the KLM framework for defeasible reasoning. Leveraging binary decision diagrams (BDDs), a well-established tool in knowledge compilation, we demonstrate that ranked interpretations can be effectively represented using a BDD with multiple terminal nodes, termed a ranked binary decision diagram (RBDD). Additionally, we show that an existing procedure for checking entailment against a BDD can be adapted with minimal modification to check defeasible entailment against an RBDD.