Skip to content

Latest commit

 

History

History
37 lines (27 loc) · 1.18 KB

README.md

File metadata and controls

37 lines (27 loc) · 1.18 KB
Raw

KG2Vec

🏎 KG2Vec: Expeditious Generation of Knowledge Graph Embeddings

Usage

sh kg2vec_<scoring>.sh <dataset_id> <training_data> <dimensions> <test_data> <verbalization_type> <neg_sampling> <training_epochs>

LSTM-based scoring function

sh kg2vec_lstm.sh aksw-bib aksw-bib.train+valid.nt 10 aksw-bib.test.nt output random 100

Analogy-based scoring function

sh kg2vec_analogy.sh aksw-bib aksw-bib.train+valid.nt 10 aksw-bib.test.nt output

Use cases

  • An add-on for the Genesis Linked Data browser uses a low-dimensional KG2Vec model trained on DBpedia for retrieving similar resources.

Cite

  • Presented at the 5th European Conference on Data Analysis (ECDA 2018) as "A Simple and Fast Approach to Knowledge Graph Embedding".
  • Working paper: https://arxiv.org/abs/1803.07828
@proceedings{soru-kg2vec-2018,
    author = "Tommaso Soru and Stefano Ruberto and Diego Moussallem and Edgard Marx and Diego Esteves and Axel-Cyrille {Ngonga Ngomo}",
    title = "Expeditious Generation of Knowledge Graph Embeddings",
    year = "2018",
}