Forgetting-Based Inconsistency Measure

Philippe Besnard (1, 2)

1 IRIT-LILaC - Logique, Interaction, Langue et Calcul
2 CNRS - Centre National de la Recherche Scientifique

We propose to apply a variant of forgetting, a simple method to restore consistency, in order to get a new inconsistency measure from the following intuitive idea: How much effort is needed to restore consistency of a knowledge base is presumably indicative of how inconsistent the knowledge base is. We discuss properties of the inconsistency measure obtained, in particular in the face of well-known postulates for inconsistency measures. We also mention in what sense this new measure does not fall into the dichotomy of inconsistency measures proposed in the literature: alphabet-based approaches vs formula-based approaches.

Mots clés: Knowledge base; Propositional logic

Domaines: Apprentissage [cs.LG]; Logique en informatique [cs.LO]

Philippe Besnard. Forgetting-Based Inconsistency Measure. 10th International Conference on Scalable Uncertainty Management (SUM 2016), Sep 2016, Nice, France. pp.331-337, ⟨10.1007/978-3-319-45856-4_23⟩. ⟨hal-02365628⟩