Research map / Graph learning / Knowledge and temporal graphs
Embedding of Hierarchically Typed Knowledge Bases
Richong Zhang, Fanshuang Kong, Chenyue Wang, Yongyi Mao
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| add type-constraint costs (entity-type and relation-type) to TransE | TransE · FB15K | Hit@10 / mean rank (DIM=300): 48.82 -> 50.17 Hit@10; 193.09 -> 178.98 rank | Improved |
| add type-constraint costs to TransH | TransH · FB15K | Hit@10 (DIM=300): 49.20 -> 50.85 | Improved |
| add type costs to TransE (reciprocal triples removed) | TransE · FB15K* | Hit@10 / mean rank (DIM=300): 34.97 -> 38.73 Hit@10; 363.29 -> 315.48 rank | Improved |
| add type costs to TransH | TransH · FB15K* | Hit@10 (DIM=300): 35.50 -> 39.19 | Improved |
| add type costs to multi-fold mTransH | mTransH · JF17K | Hit@10 / mean rank (DIM=300): 47.41 -> 50.99 Hit@10; 223.28 -> 150.42 rank | Improved |
| use proposed typed model instead of TKRL-WHE | TransE-T · FB15K | Hit@10 (DIM=300): 49.32 -> 50.17 | Improved |
| use proposed typed model instead of TKRL at low dimension | TransH-T · FB15K | Hit@10 (DIM=100): 51.97 (TKRL-WHE) vs 48.73 | Hurt |
| increase embedding dimension for TKRL-RHE | TKRL-RHE · FB15K | Hit@10: 51.84 -> 11.40 (DIM 100 -> 200) | Hurt |
Closest papers by meaning
- Data-Dependent Learning of Symmetric/Antisymmetric Relations for Knowledge Base Completion (AAAI 2018 · similarity 0.88)
- Entity-Agnostic Representation Learning for Parameter-Efficient Knowledge Graph Embedding (AAAI 2023 · similarity 0.88)
- LENA: Locality-Expanded Neural Embedding for Knowledge Base Completion (AAAI 2019 · similarity 0.86)
- DeepType: Multilingual Entity Linking by Neural Type System Evolution (AAAI 2018 · similarity 0.86)
- Improved Knowledge Graph Embedding Using Background Taxonomic Information (AAAI 2019 · similarity 0.86)
- On Completing Sparse Knowledge Base with Transitive Relation Embedding (AAAI 2019 · similarity 0.86)
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.
