Special Issue on Challenges for Reasoning under Uncertainty, Inconsistency, Vagueness, and Preferences

Gabriele Kern-Isberner, Thomas Lukasiewicz
2017 Künstliche Intelligenz  
Tabia investigate different approaches for handling inconsistent knowledge bases in the description logic DL-Lite when the ABox is prioritized and inconsistent with the TBox. Such inconsistency problems often occur when ABoxes are provided by multiple conflicting sources of different reliability levels. The authors propose different inference strategies for selecting one consistent ABox, called preferred repair, along with polynomial algorithms for computing the preferred repairs in the
more » ... t cases. An inconsistency measure maps a knowledge base to a non-negative real number, where larger values indicate the presence of more significant inconsistencies in the knowledge base. To assess the quality of a particular inconsistency measure, a wide range of rationality postulates has been proposed in the literature. In his paper "On the compliance of rationality postulates for inconsistency measures: a more or less complete picture", Matthias Thimm surveys 15 recent inconsistency measures and compares them relative to their compliance with eight rationality postulates, providing new insights into the adequacy of measures and the significance of postulates. When reasoning qualitatively from a conditional knowledge base, two established approaches are p-entailment and System Z, using all and just one ranking model(s), respectively, as semantics of a conditional knowledge base. Between these two extremes, the approach of c-representations generates a subset of all ranking models with certain constraints. In "A practical comparison of qualitative inferences with preferred ranking models", Christoph Beierle, Christian Eichhorn, and Steven Kutsch follow this idea of using preferred ranking models as the semantics of a conditional knowledge base. Decision theory is often based on quantitative utility theory. However, in "Analyzing a bipolar decision structure through qualitative decision theory", Florence Dupin Managing uncertainty, inconsistency, vagueness, and preferences has been extensively explored in artificial intelligence (AI). During the recent years, especially with the emerging of smart services and devices, technologies for managing uncertainty, inconsistency, vagueness, and preferences to tackle the problems of dynamic, real-world scenarios have started to play a key role also in other areas, such as information systems and the (social and/or semantic) Web. These application areas have sparked another wave of strong interest into formalisms and logics for dealing with uncertainty, inconsistency, vagueness, and preferences. Important examples are fuzzy and probabilistic approaches for description logics, or rule systems for handling vagueness and uncertainty in the Semantic Web, or formalisms for handling user preferences in the context of ontological knowledge in the social semantic web. While scalability of these approaches is an important issue to be addressed, also the need for combining various of these approaches with each other and/or with more classical ways of reasoning have become obvious (hybrid reasoning under uncertainty). This special issue presents several state-ofthe-art formalisms and methodologies for managing uncertainty, inconsistency, vagueness, and preferences. In "Polynomial algorithms for computing a single preferred assertional-based repair"
doi:10.1007/s13218-016-0479-z fatcat:rdduceqyzzatljzzbmlo2vhjty