Research map / Large language models / Retrieval and knowledge
Query2box: Reasoning over Knowledge Graphs in Vector Space Using Box Embeddings
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 queries with boxes rather than point vectors | Query2Box versus strongest GQE baseline · FB15k | H@3 (Avg): 0.386 -> 0.484 | Helped |
| represent queries with boxes rather than point vectors | Query2Box versus strongest GQE baseline · FB15k-237 | H@3 (Avg): 0.23 -> 0.268 | Helped |
| represent queries with boxes rather than point vectors | Query2Box versus strongest GQE baseline · NELL995 | H@3 (Avg): 0.248 -> 0.306 | Helped |
| use attention-weighted intersection centers instead of averaging (ablation) | Query2Box · FB15k | H@3 (Avg): 0.468 -> 0.484 | Helped |
| use attention-weighted intersection centers instead of averaging (ablation) | Query2Box · FB15k-237 | H@3 (Avg): 0.249 -> 0.268 | Helped |
| use attention-weighted intersection centers instead of averaging (ablation) | Query2Box · NELL995 | H@3 (Avg): 0.283 -> 0.306 | Helped |
| use attention-weighted intersection centers instead of a DeepSets intersection (ablation) | Query2Box · FB15k | H@3 (Avg): 0.467 -> 0.484 | Helped |
| use attention-weighted intersection centers instead of a DeepSets intersection (ablation) | Query2Box · FB15k-237 | H@3 (Avg): 0.259 -> 0.268 | Helped |
| use attention-weighted intersection centers instead of a DeepSets intersection (ablation) | Query2Box · NELL995 | H@3 (Avg): 0.293 -> 0.306 | Helped |
| train on complex logical queries in addition to 1p queries (ablation) | Query2Box · FB15k | H@3 (Avg): 0.385 -> 0.484 | Helped |
| train on complex logical queries in addition to 1p queries (ablation) | Query2Box · FB15k-237 | H@3 (Avg): 0.219 -> 0.268 | Helped |
| train on complex logical queries in addition to 1p queries (ablation) | Query2Box · NELL995 | H@3 (Avg): 0.274 -> 0.306 | Helped |
Closest papers by meaning
- ConE: Cone Embeddings for Multi-Hop Reasoning over Knowledge Graphs (NeurIPS 2021 · similarity 0.91)
- Faithful Embeddings for Knowledge Base Queries (NeurIPS 2020 · similarity 0.90)
- Query Embedding on Hyper-Relational Knowledge Graphs (ICLR 2022 · similarity 0.90)
- Rethinking Complex Queries on Knowledge Graphs with Neural Link Predictors (ICLR 2024 · similarity 0.88)
- Context-aware Inductive Knowledge Graph Completion with Latent Type Constraints and Subgraph Reasoning (AAAI 2025 · similarity 0.87)
- A Foundation Model for Zero-shot Logical Query Reasoning (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.