Research map / Graph learning / Knowledge and temporal graphs
Learning Representations of Bi-level Knowledge Graphs for Reasoning beyond Link Prediction
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 |
|---|---|---|---|
| represent relationships between base-level triplets and train embeddings on both graph levels | BiVE · FBH, FBHE, and DBHE triplet prediction | MRR: 0.104{ $\pm$0.000} vs 0.748{ $\pm$0.007}; 0.267{ $\pm$0.000} vs 0.555{ $\pm$0.007}; 0.191{ $\pm$0.001} vs 0.644{ $\pm$0. | Helped |
| condition base-level missing-entity prediction on another triplet through a higher-level relation | BiVE · FBH, FBHE, and DBHE conditional link prediction | MRR: 0.440{ $\pm$0.001} vs 0.762{ $\pm$0.007}; 0.528{ $\pm$0.001} vs 0.698{ $\pm$0.004}; 0.517{ $\pm$0.006} vs 0.801{ $\pm$0. | Helped |
| add augmented-triplet loss (ablation) | BiVE-QuatE · FBH triplet prediction | MR: 19.2 -> 18.7 | Helped |
| add augmented-triplet loss (ablation) | BiVE-QuatE · FBHE triplet prediction | MR: 28.1 -> 33.1 | Hurt |
| add augmented-triplet loss (ablation) | BiVE-QuatE · DBHE triplet prediction | MR: 65.4 -> 56.6 | Helped |
| add augmented-triplet loss (ablation) | BiVE-QuatE · FBH conditional link prediction | MR: 8.3 -> 7.0 | Helped |
| add augmented-triplet loss (ablation) | BiVE-QuatE · FBHE conditional link prediction | MR: 12.5 -> 11.0 | Helped |
| add augmented-triplet loss (ablation) | BiVE-QuatE · DBHE conditional link prediction | MR: 12.4 -> 12.5 | No clear effect |
| add higher-level-triplet loss (ablation) | BiVE-QuatE · FBH base-level link prediction | MR: 139.0 -> 138.4 | Helped |
| add augmented-triplet loss (ablation) | BiVE-QuatE · FBH, FBHE, and DBHE base-level link prediction | MR: 139.0 -> 124.7; 139.0 -> 125.2; 409.6 -> 404.9 | Helped |
Closest papers by meaning
- Learning Triple Embeddings from Knowledge Graphs (AAAI 2020 · similarity 0.91)
- Structure Is All You Need: Structural Representation Learning on Hyper-Relational Knowledge Graphs (ICML 2025 · similarity 0.88)
- Knowledge Graphs Can be Learned with Just Intersection Features (ICML 2024 · similarity 0.88)
- Toward a Unified Statistical Theory of Unsupervised Pretraining and Supervised Neural Knowledge Graph Learning (NeurIPS 2026 · similarity 0.88)
- Clustering then Propagation: Select Better Anchors for Knowledge Graph Embedding (NeurIPS 2024 · similarity 0.88)
- Towards Continual Knowledge Graph Embedding via Incremental Distillation (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.