Research map / Graph learning / Knowledge and temporal graphs
An Open-World Extension to Knowledge Graph Completion Models
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use OWE with 300-dimensional Wikipedia2Vec embeddings instead of 200-dimensional Glove embeddings | ComplEx-OWE · DBPedia50k tail prediction | MRR: 57.7 -> 60.3 | Helped |
| Use OWE with 300-dimensional Wikipedia2Vec embeddings instead of 200-dimensional Glove embeddings | ComplEx-OWE · FB15k-237-OWE tail prediction | MRR: 37.3 -> 40.1 | Helped |
| Use OWE with 300-dimensional Wikipedia2Vec embeddings instead of 200-dimensional Glove embeddings | ComplEx-OWE · FB20k tail prediction | MRR: 52.3 -> 53.1 | Helped |
| Apply OWE with ComplEx rather than TransE | OWE · FB15k-237-OWE tail prediction | Filtered MRR: 28.7 -> 35.2 | Helped |
| Apply OWE with ComplEx rather than DistMult | OWE · FB15k-237-OWE tail prediction | Filtered MRR: 34.4 -> 35.2 | Helped |
| Replace affine mapping with linear mapping (ablation) | ComplEx-OWE-300 · FB15k-237-OWE tail prediction | Filtered MRR: 35.2 -> 33.2 | Hurt |
| Replace affine mapping with a four-layer MLP (ablation) | ComplEx-OWE-300 · FB15k-237-OWE tail prediction | Filtered MRR: 35.2 -> 32.2 | Hurt |
| Remove entity descriptions while retaining entity names, using Wikipedia2Vec (ablation) | ComplEx-OWE-300 · FB15k-237-OWE | prediction accuracy: 2-3% | Hurt |
| Remove entity descriptions while retaining entity names, using Glove (ablation) | ComplEx-OWE · FB15k-237-OWE | prediction accuracy | Hurt |
| Remove names and descriptions for half of the training entities (ablation) | ComplEx-OWE-300 · FB15k-237-OWE | prediction accuracy: less than 1% | Hurt |
| Remove names and descriptions for 90% of training entities (ablation) | ComplEx-OWE-300 · FB15k-237-OWE | prediction accuracy: 90% | Hurt |
| Use ComplEx-OWE-300 rather than ConMask | Open-world link prediction · FB15k-237-OWE tail prediction | MRR: 29.9 -> 40.1 | Helped |
Closest papers by meaning
- Knowledge Graph Completion with Relation-Aware Anchor Enhancement (AAAI 2025 · similarity 0.91)
- Open-World Knowledge Graph Completion (AAAI 2018 · similarity 0.91)
- Select and Augment: Enhanced Dense Retrieval Knowledge Graph Augmentation (Abstract Reprint) (AAAI 2024 · similarity 0.89)
- Replacing Paths with Connection-Biased Attention for Knowledge Graph Completion (AAAI 2025 · similarity 0.89)
- KG-FIT: Knowledge Graph Fine-Tuning Upon Open-World Knowledge (NeurIPS 2024 · similarity 0.88)
- KGDM: A Diffusion Model to Capture Multiple Relation Semantics for Knowledge Graph Embedding (AAAI 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.