We study the effectiveness of Knowledge Graph Embeddings (KGE) for knowledge graph (KG) completion with rule mining. More specifically, we mine rules from KGs before and after they have been completed by a KGE to compare possible differences in the rules extracted. We apply this method to classical KGEs approaches, in particular, TransE, DistMult and ComplEx. Our experiments indicate that there can be huge differences between the extracted rules, depending on the KGE approach for KG completion. In particular, after the TransE completion, several spurious rules were extracted.
@article{arxiv.2206.00983,
title = {On the Effectiveness of Knowledge Graph Embeddings: a Rule Mining Approach},
author = {Johanna Jøsang and Ricardo Guimarães and Ana Ozaki},
journal= {arXiv preprint arXiv:2206.00983},
year = {2022}
}
Comments
The paper has 24 pages, 3 figures, presented at the first International Workshop on Knowledge Representation for Hybrid Intelligence (KR4HI'22); added references, clarifications, and acknowledgements; fixed minor typos