Zobrazeno 1 - 10
of 33
pro vyhledávání: '"Glauer, Martin"'
Deep learning models are often unaware of the inherent constraints of the task they are applied to. However, many downstream tasks require logical consistency. For ontology classification tasks, such constraints include subsumption and disjointness r
Externí odkaz:
http://arxiv.org/abs/2405.02083
Integrating human knowledge into neural networks has the potential to improve their robustness and interpretability. We have developed a novel approach to integrate knowledge from ontologies into the structure of a Transformer network which we call o
Externí odkaz:
http://arxiv.org/abs/2301.08577
In ontology development, there is a gap between domain ontologies which mostly use the web ontology language, OWL, and foundational ontologies written in first-order logic, FOL. To bridge this gap, we present Gavel, a tool that supports the developme
Externí odkaz:
http://arxiv.org/abs/2210.03497
We describe a novel approach to explainable prediction of a continuous variable based on learning fuzzy weighted rules. Our model trains a set of weighted rules to maximise prediction accuracy and minimise an ontology-based 'semantic loss' function i
Externí odkaz:
http://arxiv.org/abs/2208.12523
Reference ontologies provide a shared vocabulary and knowledge resource for their domain. Manual construction enables them to maintain a high quality, allowing them to be widely accepted across their community. However, the manual development process
Externí odkaz:
http://arxiv.org/abs/2109.09202
Autor:
Booshehri, Meisam, Emele, Lukas, Flügel, Simon, Förster, Hannah, Frey, Johannes, Frey, Ulrich, Glauer, Martin, Hastings, Janna, Hofmann, Christian, Hoyer-Klick, Carsten, Hülk, Ludwig, Kleinau, Anna, Knosala, Kevin, Kotzur, Leander, Kuckertz, Patrick, Mossakowski, Till, Muschner, Christoph, Neuhaus, Fabian, Pehl, Michaja, Robinius, Martin, Sehn, Vera, Stappel, Mirjam
Publikováno v:
In Energy and AI September 2021 5
Publikováno v:
Semantic Web (1570-0844); 2024, Vol. 15 Issue 4, p937-958, 22p
We present an institution for UML state machines without hierarchical states. The interaction with UML class diagrams is handled via institutions for guards and actions, which provide dynamic components of states (such as valuations of attributes) bu
Externí odkaz:
http://arxiv.org/abs/1411.4495
Publikováno v:
In Energy Strategy Reviews November 2018 22:264-269
Publikováno v:
Semantic Web; 20240101, Issue: Preprints p1-16, 16p