Abstract graphs and abstract paths for knowledge graph completion

• Verfasst von: Nastase, Vivi [VerfasserIn]
• Verfasst von: Kotnis, Bhushan [VerfasserIn]
• Titel: Abstract graphs and abstract paths for knowledge graph completion
• Verf.angabe: Vivi Nastase, Bhushan Kotnis
• E-Jahr: 2019
• Jahr: June 2019
• Umfang: 11 S.
• Fussnoten: Gesehen am 16.07.2019
• Titel Quelle: Enthalten in: Joint Conference on Lexical and Computational Semantics (8. : 2019 : Minneapolis, Minn.)Lexical and Computational Semantics (*SEM) - proceedings of the eighth conference
• Ort Quelle: Stroudsburg, PA : Association for Computational Linguistics (ACL), 2019
• Jahr Quelle: 2019
• Band/Heft Quelle: (2019), Seite 147-157
• ISBN Quelle: 978-1-948087-93-3
• Datenträger: Online-Ressource
• Sprache: eng
• Bibliogr. Hinweis: Forschungsdaten Nastase, Vivi: Abstract graphs, abstract paths, grounded paths for Freebase and NELL
• K10plus-PPN: 1669204804

Abstract

Knowledge graphs, which provide numerous facts in a machine-friendly format, are incomplete. Information that we induce from such graphs - e.g. entity embeddings, relation representations or patterns - will be affected by the imbalance in the information captured in the graph - by biasing representations, or causing us to miss potential patterns. To partially compensate for this situation we describe a method for representing knowledge graphs that capture an intensional representation of the original extensional information. This representation is very compact, and it abstracts away from individual links, allowing us to find better path candidates, as shown by the results of link prediction using this information.