Research map / Graph learning / Knowledge and temporal graphs
SSP: Semantic Space Projection for Knowledge Graph Embedding with Text Descriptions
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 |
|---|---|---|---|
| Project triple loss into a semantic hyperplane using entity descriptions | SSP (Joint) · FB15K knowledge graph completion (entity prediction, Filter HITS@10) | HITS@10: 77.1 | Helped |
| Project triple loss into a semantic hyperplane using entity descriptions | SSP (Joint) · WN18 knowledge graph completion (entity prediction, Filter HITS@10) | HITS@10: 93.2 | Helped |
| Jointly learn topic semantics with triple embeddings instead of fixing pretrained semantic vectors (ablation) | SSP · FB15K entity classification | MAP: 93.2 -> 94.4 | Helped |
| Jointly learn topic semantics with triple embeddings instead of fixing pretrained semantic vectors (ablation) | SSP · FB20K entity classification (zero-shot) | MAP: 67.4 | Helped |
| Use joint rather than standard semantic-vector training (ablation) | SSP · WN18 knowledge graph completion (entity prediction, Filter HITS@10) | HITS@10: 91.4 -> 93.2 | Helped |
| Use joint rather than standard semantic-vector training (ablation) | SSP · FB15K knowledge graph completion (relation prediction, Filter Mean Rank) | Mean Rank: 1.22 -> 1.47 | Hurt |
| Use text-aware semantic projection rather than TransE without textual descriptions | SSP (Joint) · FB15K knowledge graph completion (entity prediction, Filter HITS@10) | HITS@10: 66.1 -> 77.1 | Helped |
| Use strong semantic interaction rather than first-order text/triple constraints | SSP (Joint) · FB15K knowledge graph completion (entity prediction, Filter Mean Rank) | Mean Rank: 39 -> 85 | Mixed |
| Use joint text and triple learning instead of text-only NMF | SSP (Joint) · FB20K entity classification (zero-shot) | MAP: 59.6 -> 67.4 | Helped |
| Use semantic projection to discriminate hard negative triples | SSP (Std.) · Link prediction hard examples | correct discrimination rate: 82.0% | Helped |
Closest papers by meaning
- Interpreting Knowledge Graph Relation Representation from Word Embeddings (ICLR 2021 · similarity 0.90)
- Knowledge Graph Completion with Relation-Aware Anchor Enhancement (AAAI 2025 · similarity 0.88)
- KGDM: A Diffusion Model to Capture Multiple Relation Semantics for Knowledge Graph Embedding (AAAI 2024 · similarity 0.88)
- Select and Augment: Enhanced Dense Retrieval Knowledge Graph Augmentation (Abstract Reprint) (AAAI 2024 · similarity 0.87)
- Towards Global-Topology Relation Graph for Inductive Knowledge Graph Completion (AAAI 2025 · similarity 0.87)
- KG-FIT: Knowledge Graph Fine-Tuning Upon Open-World Knowledge (NeurIPS 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.