Active integrity constraints for general-purpose knowledge bases

• Published in: Annals of mathematics and artificial intelligence Vol. 83; no. 3-4; pp. 213 - 246
• Main Authors: Cruz-Filipe, Luís, Gaspar, Graça, Nunes, Isabel, Schneider-Kamp, Peter
• Format: Journal Article
• Language: English
• Published: Cham Springer International Publishing 01.08.2018; Springer; Springer Nature B.V
• Subjects: Algorithms; Analysis; Artificial Intelligence; Complex Systems; Computer Science; Context; Ethical aspects; Integrity; Iterative algorithms; Iterative methods; Knowledge bases (artificial intelligence); Mathematics; Ontology; Reasoning; Relational data bases; Validity; Ontologies; 68P15; Multi-context systems; 03B80; Repairs; Integrity constraints
• ISSN: 1012-2443; 1573-7470
• DOI: 10.1007/s10472-018-9577-y

In the database world, integrity constraints are essential to guarantee database integrity. The related problem of database repair deals with finding the best way to change a database so that it satisfies its integrity constraints. These two topics have been studied intensively since the 1980s. The formalism of active integrity constraints , proposed in 2004, aims at addressing them jointly, by providing a syntax whereby a particular subclass of integrity constraints can be specified together with preferred ways to repair inconsistency. In the last decade, several authors have proposed adaptations of the notion of integrity constraints to other reasoning frameworks than relational databases. In this article, we extend this line of work in two ways. First, we target multi-context systems, a general-purpose framework for combining heterogeneous reasoning systems, able to model most other reasoning frameworks, as we demonstrate. Second, we extend the notions of active integrity constraints and grounded repairs to this generalized setting. This way of including repair suggestions inside integrity constraints, subject to a validity check, enables us to define simple iterative algorithms to find all possible grounded repairs for an inconsistent multi-context system, avoiding the need to solve complex or undecidable problems.