RDF2vec at Scale

On larger knowledge graphs, RDF2vec models can be very expensive to train. In this chapter, we look at two techniques that make RDF2vec easier to use with large knowledge graphs. First, we look at a knowledge graph embedding server called KGvec2go, which serves pre-trained embedding vectors for well-known knowledge graphs such as DBpedia as a service. Second, we look at how we can train partial RDF2vec models only for instances of interest with RDF2vec Light.