SEPAL: Mini YAGO3 Knowledge Graph for Downstream ML Tasks

The Mini YAGO3 knowledge graph and associated downstream tasks used in the SEPAL paper (https://arxiv.org/pdf/2507.00965). SEPAL is a scalable feature learning method designed for very large knowledge graphs to produce embeddings for regression and classification. The dataset was created by inria-soda and last updated on 2026-03-17.

Use Cases

• Training knowledge graph embedding models based on the Mini YAGO3 graph structure.
• Evaluating embedding quality for downstream regression tasks mentioned in the SEPAL paper.
• Evaluating embedding quality for downstream classification tasks mentioned in the SEPAL paper.
• Benchmarking scalable feature learning methods on large knowledge graphs.

Strengths

• Dataset is associated with a specific, peer-reviewed research paper (SEPAL).
• Last updated on 2026-03-17, indicating recent maintenance.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count, file formats, and dataset size are unknown, which may limit suitability assessment.

Provenance

Source

inria-soda

Collection Method

Likely derived from the YAGO3 knowledge graph for research purposes.

Freshness

Last updated 2026-03-17 14:56:25.