Results - Details

Search command : Author="Πλεξουσάκης"  And Author="Δημήτρης"

Current Record: 9 of 88

[Add to Basket]
Identifier	000443666
Title	Knowledge graph embedding methods for entity alignment
Alternative Title	Μέθοδοι διανυμαστικής αναπαράστασης γράφων γνώσης για αντιστοίχιση οντοτήτων
Author	Φανουράκης, Νικόλαος Ε.
Thesis advisor	Χριστοφίδης, Βασίλης
Reviewer	Κομοδάκης, Νίκος Πλεξουσάκης, Δημήτρης Kotzinos, Dimitris
Abstract	In recent years, we have witnessed the proliferation of knowledge graphs (KG) in various domains, aiming to support applications like question answering, recommendations, etc. A frequent task when integrating knowledge from different KGs is to find which subgraphs refer to the same real-world entity. Recently, neural networks and deep learning techniques have been proposed for entity alignment tasks, that learn a vector-space representation (i.e., embedding) of entities which preserves their similarity in the original KGs. A wide variety of supervised, unsupervised, and semi-supervised embedding methods have been proposed in the literature that exploit the structural information (relation based) of entities in the KGs, or their attribute values (attribute based). Still, a quantitative assessment of their strengths and weaknesses on real-world KGs using different performance metrics and KG characteristics is missing. In this thesis, we conduct a meta-level analysis of popular embedding methods for entity alignment based on a statistically sound methodology. Our analysis reveals interesting statistically significant correlations of different embedding methods with various meta-features extracted by KGs. Finally, we are able to rank in a statistically significant way their performance across all real-world KGs of our testbed, while also to demonstrate their trade-offs in terms of effectiveness and efficiency. To the best of our knowledge, there are no experimental studies that provide such a type of meta-level analysis.
Language	English
Issue date	2021-11-26
Collection	School/Department--School of Sciences and Engineering--Department of Computer Science--Post-graduate theses
	Type of Work--Post-graduate theses
Permanent Link	https://elocus.lib.uoc.gr//dlib/5/5/5/metadata-dlib-1636970838-901868-14428.tkl
Views	391

Digital Documents
	Download document View document Views : 4