Towards Understanding the Impact of Graph Structure on Knowledge Graph Embeddings

Knowledge graphs (KGs) are an established paradigm for integrating heterogeneous data and representing knowledge. As such, there are many different methodologies for producing KGs, which span notions of expressivity, and are tailored for different use-cases and domains. Now, as neurosymbolic methods rise in prominence, it is important to understand how the development of KGs according to these methodologies impact downstream tasks, such as link prediction using KG embeddings (KGE). In this paper, we modify FB15k-237 in several ways (e.g., by increasingly including semantic metadata). This significantly changes the graph structure (e.g., centrality). We assess how these changes impact the link prediction task, using six KGE models.