Embedding-centrality: Generic centrality computation using neural networks

Rami Puzis, Zion Sofer, Dvir Cohen, Matan Hugi

Research output: Contribution to journalConference articlepeer-review

1 Scopus citations


Deriving vector representations of vertices in graphs, a.k.a. vertex embedding, is an active field of research. Vertex embedding enables the application of relational data mining techniques to network data. Unintended use of vertex embedding unveils a novel generic method for centrality computation using neural networks. The new centrality measure, termed Embedding Centrality, proposed in this paper is defined as the dot product of a vertex and the center of mass of the graph. Simulation results confirm the validity of Embedding Centrality which correlates well with other commonly used centrality measures. Embedding Centrality can be tailored to specific applications by devising the appropriate context for vertex embedding and can facilitate further understanding of supervised and unsupervised learning methods on graph data.

Original languageEnglish
Pages (from-to)87-97
Number of pages11
JournalSpringer Proceedings in Complexity
Issue number219279
StatePublished - 1 Jan 2018
Event9th International Conference on Complex Networks, CompleNet 2018 - Boston, United States
Duration: 5 Mar 20188 Mar 2018

ASJC Scopus subject areas

  • Applied Mathematics
  • Modeling and Simulation
  • Computer Science Applications


Dive into the research topics of 'Embedding-centrality: Generic centrality computation using neural networks'. Together they form a unique fingerprint.

Cite this