de
en
Schliessen
Detailsuche
Bibliotheken
Projekt
Impressum
Datenschutz
Schliessen
Publizieren
Besondere Sammlungen
Digitalisierungsservice
Hilfe
Impressum
Datenschutz
zum Inhalt
Detailsuche
Schnellsuche:
OK
Ergebnisliste
Titel
Titel
Inhalt
Inhalt
Seite
Seite
Im Werk suchen
Learning continuous representations for Knowledge Graphs / by Caglar Demir ; [Reviewers: Prof. Dr. Axel-Cyrille Ngonga Ngomo, Prof. Dr. Sören Auer, Jun.-Prof. Dr. Sebastian Peitz]. Paderborn, 2023
Inhalt
Contents
I Introduction
Introduction
Motivation
Research Questions and Challenges
Thesis Overview and Summary of Scientific Publications
Background
Machine Learning
Reinforcement Learning
Knowledge Graph and Link Prediction
Related Works for Knowledge Graph Embeddings
Description Logics, Knowledge Base and Concept Learning
Hypercomplex Numbers
Hadamard and Kronecker Products
II Contributions
A Physical Embedding Model for Knowledge Graphs
Methodology
Experiments & Results
A Shallow Neural Model for Relation Prediction
Methodology
Experiments & Results
Convolutional Complex Knowledge Graph Embeddings
Methodology
Experiments & Results
Convolutional Hypercomplex Embeddings for Link Prediction
Methodology
Experiments & Results
Kronecker Decomposition for Knowledge Graph Embeddings
Methodology
Experiments & Results
Polyak Parameter Ensemble for Knowledge Graph Embeddings
Methodology
Experiments & Results
Learning Permutation-Invariant Embeddings for Description Logic Concepts
Methodology
Experiments & Results
III Software Frameworks and Use cases
DICE Embeddings Framework
Large-scale Knowledge Graph Embeddings
DICE Embeddings: Hardware-agnostic Framework for Large-scale Knowledge Graph Embeddings
Downstream Applications
Ontolearn Framework
OWL Class Expressions in Python
Use Case: Deep Reinforcement Learning for Class Expression Learning
Methodology
Experiments & Results
Use Case: Rapid Explainability For Skill Description Learning
IV Conclusion and Future Work
Conclusion and Discussion
Future Works
References
Die detaillierte Suchanfrage erfordert aktiviertes Javascript.