A Computational Method for Enforcing Knowledge that Cannot be Subsumed

Abstract

We introduce a practical computational method that enforces some incoming piece of information [Formula: see text] in a clausal Boolean logic knowledge base [Formula: see text] in such a way that [Formula: see text] is not strictly subsumed by the resulting knowledge base. [Formula: see text] is not strictly subsumed by [Formula: see text] iff for every piece of information [Formula: see text] that is entailed by [Formula: see text] and that is such that [Formula: see text] entails [Formula: see text], we have that [Formula: see text] entails [Formula: see text], too. We claim that this paradigm is a useful form of reasoning for both human and artificial intelligence systems. Under a usual minimal change policy, it amounts to computing one cardinality-maximal satisfiable subset of [Formula: see text] that contains [Formula: see text] but that does not strictly subsume [Formula: see text]. Although this task is intractable in the worst case, we provide a practical method that appears experimentally efficient very often, even for large and complex knowledge bases.