Published May 9, 2023 | Version v2
Dataset Open

SemOpenAlex Embeddings

  • 1. Karlsruhe Institute of Technology (KIT)

Description

Here, we provide knowledge graph embeddings for SemOpenAlex.

After conducting experiments using different knowledge graph embedding algorithms, including TransE, DistMult, ComplEx, GraphSAGE, and GAT, we found that DistMult achieved the highest MRR scores. Therefore, we provide the embeddings based on this technique.

More information about the training on high-performance computing (using up to 716GB of CPU RAM) can be found online at https://github.com/metaphacts/semopenalex/tree/main/embeddings-generation

Files

Files (190.2 GB)

Name Size Download all
md5:d88b55dd774e17027518d776ad95632c
190.2 GB Download
md5:478697cbefdc67b0d5f845ff15cd73f2
6.6 kB Download