Knowledge representation and reasoning (KRR), on the one hand, and machine learning (ML), on the other hand, have largely been developed as independent research trends in artificial intelligence (AI). Human reasoning, however, is often based on an intricate combination of processes that are related to learning (e.g. induction or extrapolation) and processes that are closer to deductive reasoning. Similarly, we can expect that progress in AI will increasingly need to rely on hybrid approaches that combine the explainability and teachability of KRR methods with the robustness and data-driven nature of ML methods. The ambitious aim of truly integrating reasoning and learning beyond one-way linkage raises many new questions, which this workshop hopes to explore. Beyond a study of the underlying principles, this workshop will also focus on applications, with a particular emphasis on the use of spatial and temporal knowledge in everyday tasks. The workshop will serve as a forum for researchers from different fields (including Automated Theorem Proving, Cognitive Computing, Cognitive Robotics, Commonsense Reasoning, Constraint Solving, Logic, Mathematics, Machine Learning, Natural Language Processing, Theoretical Computer Science, Qualitative Reasoning) to discuss open problems, methodology, and recent advancements in the field. It also provides a forum for early career researchers to present their current work and to build up networks. We invite technical contributions mixing or linking advanced ideas from knowledge representation and reasoning with ideas from machine learning or data mining.
This workshop is the first under this name and with this scope. However it is to a limited extent a follow-up of previous ECAI and IJCAI workshops (already co-organized by 3 of the co-authors of the present proposal):
- the successful series of WL4AI workshops (Weighted Logics for Artificial Intelligence: ECAI-2012, IJCAI-2013, IJCAI-2015)
- the IJCAI-2017 workshop on Logical Foundations for Uncertainty and Learning (LFU)
Still WL4AI did not consider learning, and LFU prioritized the provision of semantical foundations for learning. L&R looks broadly at the intersection of logical formalisms and learning, by unifying the themes of WL4AI and LFU, and additionally encourages submissions touching on defeasible reasoning and nonmonotonic frameworks among other issues.
Vaishak Belle, University of Edinburgh, UK
Lluís Godo, IIIA - CSIC, Barcelona, Spain
Henri Prade, IRIT, Toulouse, France
Jochen Renz, The Australian National University, Australia
Steven Schockaert, University of Cardiff, UK
Ute Schmid, University of Bamberg, Germany
Diedrich Wolter, University of Bamberg, Germany