Integrating Ontologies and Vector Space Embeddings Using Conceptual Spaces (Invited Paper)
Résumé
Ontologies and vector space embeddings are among the most popular frameworks for encoding conceptual knowledge. Ontologies excel at capturing the logical dependencies between concepts in a precise and clearly defined way. Vector space embeddings excel at modelling similarity and analogy. Given these complementary strengths, there is a clear need for frameworks that can combine the best of both worlds. In this paper, we present an overview of our recent work in this area. We first discuss the theory of conceptual spaces, which was proposed in the 1990s by Gärdenfors as an intermediate representation layer in between embeddings and symbolic knowledge bases. We particularly focus on a number of recent strategies for learning conceptual space representations from data. Next, building on the idea of conceptual spaces, we discuss approaches where relational knowledge is modelled in terms of geometric constraints. Such approaches aim at a tight integration of symbolic and geometric representations, which unfortunately comes with a number of limitations. For this reason, we finally also discuss methods in which similarity, and other forms of conceptual relatedness, are derived from vector space embeddings and subsequently used to support flexible forms of reasoning with ontologies, thus enabling a looser integration between embeddings and symbolic knowledge.
Mots clés
Conceptual Spaces
Ontologies
Vector Space Embeddings
Learning and Reasoning
Computing methodologies → Knowledge representation and reasoning
2012 ACM Subject Classification Computing methodologies → Knowledge representation and reasoning phrases Conceptual Spaces, Ontologies, Vector Space Embeddings, Learning and Reasoning Digital Object Identifier 10.4230/OASIcs.AIB.2022.3 Camille Bourgaux, Ana Ozaki, and Rafael Peñaloza Article No. 3 pp. 3:1-3:30 OpenAccess Series in Informatics Schloss Dagstuhl -Leibniz-Zentrum für Informatik, Dagstuhl Publishing, Germany
2012 ACM Subject Classification Computing methodologies → Knowledge representation and reasoning phrases Conceptual Spaces, Ontologies, Vector Space Embeddings, Learning and Reasoning Digital Object Identifier 10.4230/OASIcs.AIB.2022.3 Camille Bourgaux, Ana Ozaki, and Rafael Peñaloza
Article No. 3
pp. 3:1-3:30 OpenAccess Series in Informatics Schloss Dagstuhl -Leibniz-Zentrum für Informatik, Dagstuhl Publishing, Germany
Domaines
Intelligence artificielle [cs.AI]Origine | Fichiers produits par l'(les) auteur(s) |
---|