On inconsistency-minimizing, paraconsistent reasoning in knowledge bases

Aus International Center for Computational Logic
Wechseln zu:Navigation, Suche

On inconsistency-minimizing, paraconsistent reasoning in knowledge bases

Vortrag von Walter Forkel
Many different solutions have been proposed to enable working with inconsistent description logic knowledge bases. In this work we introduce a novel approach that combines the ideas of a paraconsistent four-valued description logic with the assumption that models with fewer inconsistencies are to be preferred over non-minimal models. We argue that our formalism is well suited for reasoning in the presence of inconsistencies. Furthermore we show that it extends usual $\ALC$ in a natural way, providing the same results when used on consistent knowledge bases. Finally our approach has the advantage that it can be embedded in usual circumscribed $\ALC$ and thus does not require an extra reasoner and also keeps the same complexity of reasoning.