Research map / Large language models / Retrieval and knowledge
Benchmarking the Combinatorial Generalizability of Complex Query Answering on Knowledge Graphs
Zihao Wang, Hang Yin, Yangqiu Song
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 |
|---|---|---|---|
| Use DNF+IU instead of DM (ablation) | LogicE · FB15k-237 | MRR: 10.00 -> 10.46 | Improved |
| Use DNF+IU instead of DM (ablation) | LogicE · FB15k | MRR: 20.53 -> 21.89 | Improved |
| Use DNF+IU instead of DM (ablation) | LogicE · NELL | MRR: 11.13 -> 13.07 | Improved |
| Replace binary i/u with multi-input I/U in DNF+IU (ablation) | LogicE · FB15k-237 | MRR: 10.01 -> 10.46 | Improved |
| Replace binary i/u with multi-input I/U in DNF+IU (ablation) | LogicE · FB15k | MRR: 20.55 -> 21.89 | Improved |
| Replace binary i/u with multi-input I/U in DNF+IU (ablation) | LogicE · NELL | MRR: 11.14 -> 13.07 | Improved |
| Use DNF+IUD instead of DNF+IUd (ablation) | NewLook · FB15k-237 | MRR: 9.11 -> 9.13 | Improved |
| Use DNF+IUD instead of DNF+IUd (ablation) | NewLook · FB15k | MRR: 19.80 -> 19.87 | Improved |
| Use DNF+IUD instead of DNF+IUd (ablation) | NewLook · NELL | MRR: 9.88 -> 9.90 | Improved |
| Add deeper projection query types to training (1p,2p,3p rather than 1p) (ablation) | LogicE · FB15k-237, 2p–5p | MRR: 4.98 -> 9.01 (2p); 3.95 -> 7.98 (3p); 3.17 -> 7.22 (4p); 2.93 -> 7.15 (5p) | Improved |
| Add deeper projection query types to training (1p,2p,3p rather than 1p) (ablation) | LogicE · FB15k-237, 1p | MRR: 19.36 -> 17.81 | Hurt |
| Add 3i to the training query types (ablation) | LogicE · FB15k-237, 3i and 4i | MRR: 41.66 -> 42.67 (3i); 52.37 -> 52.70 (4i) | Improved |
Closest papers by meaning
- Is Complex Query Answering Really Complex? (ICML 2025 · similarity 0.93)
- A Foundation Model for Zero-shot Logical Query Reasoning (NeurIPS 2024 · similarity 0.89)
- Rethinking Complex Queries on Knowledge Graphs with Neural Link Predictors (ICLR 2024 · similarity 0.87)
- Consistent Query Answering over Existential Rules with Open and Closed Predicates (AAAI 2025 · similarity 0.87)
- Diagnosing and Addressing Pitfalls in KG-RAG Datasets: Toward More Reliable Benchmarking (NeurIPS 2025 · similarity 0.86)
- Bayesian Inference with Complex Knowledge Graph Evidence (AAAI 2024 · similarity 0.85)
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.
