We describe two applications of refinement operators that can generalise and specialise concepts expressed in the ALC description logic language. The first application addresses the problem of analysing the joint coherence of some given concepts w.r.t. a background ontology. To this end, we apply Thagard's computational theory of coherence, in combination with semantic similarity between concepts defined by means of the generalisation operator. The second application focuses on repairing an inconsistent collective ontology that may result from the vote on axioms of multiple experts based on principles from social choice theory and judgment aggregation. We use the refinement operators, together with a reference ontology, to weaken some axioms and to repair the collective ontology.
Confalonieri, R., Troquard, N., Galliani, P., Kutz, O., Peñaloza, R., Porello, D. (2018). Two applications of concept refinement. Intervento presentato a: 1st International Workshop on Comprehensibility and Explanation in AI and ML, Bari, Italy.
Two applications of concept refinement
Peñaloza R;
2018
Abstract
We describe two applications of refinement operators that can generalise and specialise concepts expressed in the ALC description logic language. The first application addresses the problem of analysing the joint coherence of some given concepts w.r.t. a background ontology. To this end, we apply Thagard's computational theory of coherence, in combination with semantic similarity between concepts defined by means of the generalisation operator. The second application focuses on repairing an inconsistent collective ontology that may result from the vote on axioms of multiple experts based on principles from social choice theory and judgment aggregation. We use the refinement operators, together with a reference ontology, to weaken some axioms and to repair the collective ontology.File | Dimensione | Formato | |
---|---|---|---|
CExAIIA_2017_paper_4.pdf
accesso aperto
Tipologia di allegato:
Publisher’s Version (Version of Record, VoR)
Dimensione
560.21 kB
Formato
Adobe PDF
|
560.21 kB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.