Investigating Preferential Reasoning in Formal Concept Analysis
Aus International Center for Computational Logic
Investigating Preferential Reasoning in Formal Concept Analysis
Vortrag von Lucas Carr
- Veranstaltungsort: APB-2026
- Beginn: 11. Dezember 2025 um 11:00
- Ende: 11. Dezember 2025 um 12:00
- Forschungsgruppe: Computational Logic
- Event series: Research Seminar Logic and AI
- iCal
Formal Concept Analysis (FCA) provides a framework for studying Boolean data through concepts and their corresponding structures. The implication logic underlying FCA is relatively simple - analogous to propositional Horn, and is monotonic. As a result, FCA is largely unable to reason through exceptions that may exist in given data, impeding the discovery of partially holding, but useful, rules. We explore how the style of non-monotonic reasoning developed by Kraus, Lehmann, and Magidor (1990) (frequently initialised to the KLM framework), and more generally preferential semantics, may be introduced to FCA, and what becomes of the (defeasible) implication logic, as well as how the notions of formal concepts may be reframed to accommodate the aforementioned preferential semantics.
