From July 20-23, 2026, researchers from our Knowledge Representation and Methods Group attended the 23rd International Conference on Principles of Knowledge Representation and Reasoning (KR) in Lisbon, Portugal. They submitted several papers and were actively involved in program planning as program chairs or members of the program committee.

Program Committee and Chairs

• Prof. Franz Baader, Principal Investigator at ScaDS.AI Dresden/Leipzig, was one of the two program chairs for the conference.
• In addition, Prof. Markus Krötzsch was heading the “KR in the Wild” section.
• Dr. Dörthe Arndt and Dr. Sergei Obiedkov helped him shape the track’s content.
• Numerous ScaDS.AI researchers were also contributing to the program as area chairs. These include Prof. Sebastian Rudolph, Prof. Gerhard Brewka, Prof. Carsten Lutz, and Prof. Ringo Baumann.

Best paper award

Marvin Grosser received the Ray Reiter Best Paper Prize for the paper Fitting Horn DL Ontologies to ABox and Query Examples: A Tale of Simulation Quantifiers and Finite Models. The award was introduced in 2004 in honour of the contributions made by Ray Reiter and sponsored by the Artificial Intelligence Journal.

Carsten Lutz: The Logical Expressive Power of Graph Neural Networks

The invited talk The Logical Expressive Power of Graph Neural Networks by Prof. Carsten Lutz reviewed approaches for leveraging logics developed in symbolic AI. This should assess the expressive power of subsymbolic (neural) learning methods like graph neural networks. He provided an overview of graph neural networks (GNNs) and the latest advancements in characterizing their precise expressive capabilities.

Two submissions of Sergei Obiedkov

Dr. Sergei Obiedkov had two submissions accepted at the KR Conference. One was a regular paper in the main track titled “Computing Extensions of Abstract Argumentation Frameworks by Enumerating Closed Sets.” In the paper, he presents a new approach for computing complete, stable, and preferred extensions of abstract argumentation frameworks. His approach solves the problem directly by leveraging the fact that these extensions are contained within certain closure systems. The second paper, PAC Learning of Concept Inclusions for Ontology-Mediated Query Answering is an extended abstract. It presents a probably approximately correct algorithm for learning the terminological part of a description logic knowledge base via subsumption queries.

Workshop contribution by Christian Alrabbaa

At the Joint Workshop on “Explainable Logic-Based Knowledge Representation (XLoKR) and Explanations with Constraints and Satisfiability (ExCoS)” Christian Alrabbaa presents a tool develop jointly with members of PI Dachselt’s group. It provides visual decision support for the interactive repair of ontologies. This work is an important contribution towards making symbolic AI techniques more trust-worthy and explainable.

Workshop contribution by Oscar Barreca

At the workshop on “Logical Approaches to Handling Inconsistent Data”, ScaDS.AI PhD student Oscar Barreca presents a poster. It points out that errors in knowledge-bases need not manifest themselves as inconsistency. Thus, not only approaches for inconsistency-tolerant reasoning, but also ones for error-tolerant reasoning must be considered. This again is a step towards making symbolic AI techniques more trust-worthy by developing ways for dealing with modeling errors.

KR 2026

The International Conference on Principles of Knowledge Representation and Reasoning (KR) is the leading forum for the in-depth and timely presentation of advances in the theory and practice of representing and computationally managing knowledge. The conference is dedicated to formal modeling and the application of knowledge in intelligent systems capable of logical reasoning and rational decision-making. Research in this field focuses on the expressiveness, computational efficiency, and explainability of various representation forms.