Link Prediction Playground
MVP wireframe for benchmarking graph embedding models (node classification, link prediction) against a small RDF-derived dataset.
Concept, MVP scope: convert a small RDF graph to a PyG dataset, run two or three standard embedding baselines, and report link-prediction metrics side by side. The goal is a fast sandbox for trying an idea, not a benchmarking suite.
Temporal knowledge graph forecasting — the harder, more research-flavored version of this problem — is a deliberate future extension, not part of the MVP.