Autor*innen mit den meisten Dokumenten
Auflistung nach:
Neueste Veröffentlichungen
- ZeitschriftenartikelThe AAA ABox Abduction Solver(KI - Künstliche Intelligenz: Vol. 34, No. 4, 2020) Pukancová, Júlia; Homola, MartinAAA is a sound and complete ABox abduction solver based on the Reiter’s MHS algorithm and the Pellet reasoner. It supports DL expressivity up to $$\mathcal {SROIQ}$$ SROIQ (i.e., OWL 2). It supports multiple observations, and allows to specify abducibles.
- ZeitschriftenartikelRethinking Computer Science Through AI(KI - Künstliche Intelligenz: Vol. 34, No. 4, 2020) Kersting, Kristian
- ZeitschriftenartikelNews(KI - Künstliche Intelligenz: Vol. 34, No. 4, 2020)
- ZeitschriftenartikelError-Tolerance and Error Management in Lightweight Description Logics(KI - Künstliche Intelligenz: Vol. 34, No. 4, 2020) Peñaloza, RafaelThe construction and maintenance of ontologies is an error-prone task. As such, it is not uncommon to detect unwanted or erroneous consequences in large-scale ontologies which are already deployed in production. While waiting for a corrected version, these ontologies should still be available for use in a “safe” manner, which avoids the known errors. At the same time, the knowledge engineer in charge of producing the new version requires support to explore only the potentially problematic axioms, and reduce the number of exploration steps. In this paper, we explore the problem of deriving meaningful consequences from ontologies which contain known errors. Our work extends the ideas from inconsistency-tolerant reasoning to allow for arbitrary entailments as errors, and allows for any part of the ontology (be it the terminological elements or the facts) to be the causes of the error. Our study shows that, with a few exceptions, tasks related to this kind of reasoning are intractable in general, even for very inexpressive description logics.
- ZeitschriftenartikelSpecial Issue on Ontologies and Data Management: Part II(KI - Künstliche Intelligenz: Vol. 34, No. 4, 2020) Schneider, Thomas; Šimkus, Mantas
- ZeitschriftenartikelDefeasible Description Logics(KI - Künstliche Intelligenz: Vol. 34, No. 4, 2020) Varzinczak, IvanThe present paper is a summary of a habilitation ( Habilitation à Diriger des Recherches , in French), which has been perused and evaluated by a committee composed by the following members: Franz Baader, Stéphane Demri, Hans van Ditmarsch, Sébastien Konieczny, Pierre Marquis, Marie-Laure Mugnier, Odile Papini and Leon van der Torre. It was defended on 26 November 2019 at Université d’Artois in Lens, France.
- ZeitschriftenartikelA Lightweight Defeasible Description Logic in Depth(KI - Künstliche Intelligenz: Vol. 34, No. 4, 2020) Pensel, MaximilianIn this thesis we study KLM-style rational reasoning in defeasible Description Logics. We illustrate that many recent approaches to derive consequences under Rational Closure (and its stronger variants, lexicographic and relevant closure) suffer the fatal drawback of neglecting defeasible information in quantified concepts. We propose novel model-theoretic semantics that are able to derive the missing entailments in two differently strong flavours. Our solution introduces a preference relation to distinguish sets of models in terms of their typicality (amount of defeasible information derivable for quantified concepts). The semantics defined through the most typical (most preferred) sets of models are proven superior to previous approaches in that their entailments properly extend previously derivable consequences, in particular, allowing to derive defeasible consequences for quantified concepts. The dissertation concludes with an algorithmic characterisation of this uniform maximisation of typicality, which accompanies our investigation of the computational complexity for deriving consequences under these new semantics.
- ZeitschriftenartikelInterview with Diego Calvanese(KI - Künstliche Intelligenz: Vol. 34, No. 4, 2020) Calvanese, Diego; Šimkus, Mantas
- ZeitschriftenartikelRewriting Approaches for Ontology-Mediated Query Answering(KI - Künstliche Intelligenz: Vol. 34, No. 4, 2020) Ahmetaj, ShqiponjaA most promising approach to answering queries in ontology-based data access (OBDA) is through query rewriting. In this paper we present novel rewriting approaches for several extensions of OBDA. The goal is to understand their relative expressiveness and to pave the way for efficient query answering algorithms.
- ZeitschriftenartikelTowards Explanatory Interactive Image Captioning Using Top-Down and Bottom-Up Features, Beam Search and Re-ranking(KI - Künstliche Intelligenz: Vol. 34, No. 4, 2020) Biswas, Rajarshi; Barz, Michael; Sonntag, DanielImage captioning is a challenging multimodal task. Significant improvements could be obtained by deep learning. Yet, captions generated by humans are still considered better, which makes it an interesting application for interactive machine learning and explainable artificial intelligence methods. In this work, we aim at improving the performance and explainability of the state-of-the-art method Show, Attend and Tell by augmenting their attention mechanism using additional bottom-up features. We compute visual attention on the joint embedding space formed by the union of high-level features and the low-level features obtained from the object specific salient regions of the input image. We embed the content of bounding boxes from a pre-trained Mask R-CNN model. This delivers state-of-the-art performance, while it provides explanatory features. Further, we discuss how interactive model improvement can be realized through re-ranking caption candidates using beam search decoders and explanatory features. We show that interactive re-ranking of beam search candidates has the potential to outperform the state-of-the-art in image captioning.
- «
- 1 (current)
- 2
- 3
- »