An extended version of this paper has been published at the the 34th AAAI Conference on Artificial Intelligence (AAAI) with the title “Learning Triple Embeddings from Knowledge Graphs”. Graph embedding techniques allow to learn high-quality low-dimensional graph representations useful in various tasks, from node classification to clustering. Knowledge graphs are particular types of graphs characterized by several distinct types of nodes and edges. Existing knowledge graph embedding approaches have only focused on learning embeddings of nodes and predicates. However, the basic piece of information stored in knowledge graphs are triples and thus, an interesting problem is that of learning embeddings of triples as a whole. In this paper we report on Triple2Vec, a new technique to directly compute triple embeddings in knowledge graphs. Triple2Vec leverages the idea of line graph and extends it to the context of knowledge graphs. Embeddings are then generated by adopting the SkipGram model, where sentences are replaced with walks on a wighted version of the line graph.

From Node Embeddings to Triple Embeddings

Fionda Valeria;
2022

Abstract

An extended version of this paper has been published at the the 34th AAAI Conference on Artificial Intelligence (AAAI) with the title “Learning Triple Embeddings from Knowledge Graphs”. Graph embedding techniques allow to learn high-quality low-dimensional graph representations useful in various tasks, from node classification to clustering. Knowledge graphs are particular types of graphs characterized by several distinct types of nodes and edges. Existing knowledge graph embedding approaches have only focused on learning embeddings of nodes and predicates. However, the basic piece of information stored in knowledge graphs are triples and thus, an interesting problem is that of learning embeddings of triples as a whole. In this paper we report on Triple2Vec, a new technique to directly compute triple embeddings in knowledge graphs. Triple2Vec leverages the idea of line graph and extends it to the context of knowledge graphs. Embeddings are then generated by adopting the SkipGram model, where sentences are replaced with walks on a wighted version of the line graph.
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/20.500.11770/336825
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 0
  • ???jsp.display-item.citation.isi??? ND
social impact