Research map / Large language models / Retrieval and knowledge
A Survey of Knowledge Graph Reasoning on Graph Types: Static, Dynamic, and Multi-Modal
Ke Liang, Lingyuan Meng, Meng Liu, Yue Liu, Wenxuan Tu, Siwei Wang, Sihang Zhou, Xinwang Liu, Fuchun Sun, Kunlun He
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 ComplEx with DURA regularization instead of plain ComplEx | ComplEx · FB15k-237 (transductive) | MRR: 0.346 -> 0.370 | Improved |
| Replace TransE with rotation-based RotatE | TransE · WN18RR (transductive) | MRR: 0.231 -> 0.476 | Improved |
| Replace RGCN with attention/composition GNN CompGCN | RGCN · WN18RR (transductive) | MRR: 0.427 -> 0.469 | Improved |
| Replace DistMult with ComplEx (complex embeddings) | DistMult · FB15k-237 (transductive) | MRR: 0.243 -> 0.346 | Improved |
| Replace ConvE with HousE (Householder-based geometric embedding) | ConvE · WN18RR (transductive) | MRR: 0.430 -> 0.511 | Improved |
| Use path-based GNN RED-GNN rather than embedding model HAKE | HAKE · WN18RR (transductive) | MRR: 0.497 -> 0.533 | Improved |
| Replace static TTransE with RNN/GNN-free ChronoR rotation model | TTransE · ICEWS14 (interpolation) | MRR (%): 25.50 -> 62.53 | Improved |
| Use DyERNIE (Riemannian dynamic embeddings) instead of TComplEx | TComplEx · ICEWS14 (interpolation) | MRR (%): 61.90 -> 66.90 | Improved |
| Add multi-modal fusion in MoSE versus image-based IKRL | IKRL · FB15k-237-IMG | Hits@10 (%): 45.8 -> 56.5 | Improved |
| Use IMF interactive multi-modal fusion rather than MKGformer | MKGformer · FB15k-237-IMG | Hits@10 (%): 49.9 -> 59.3 | Improved |
| Use pre-trained multi-modal transformer VisualBERT instead of dedicated KG model | MoSE · FB15k-237-IMG | Hits@10 (%): 56.5 -> 43.9 | Hurt |
| Use RGHAT instead of HousE on FB15k-237 but not on WN18RR | HousE · FB15k-237 vs WN18RR (transductive) | MRR: 0.361 -> 0.522 (FB15k-237); 0.511 -> 0.483 (WN18RR) | Mixed |
Closest papers by meaning
- Graph-constrained Reasoning: Faithful Reasoning on Knowledge Graphs with Large Language Models (ICML 2025 · similarity 0.90)
- Reasoning on Graphs: Faithful and Interpretable Large Language Model Reasoning (ICLR 2024 · similarity 0.90)
- Debate on Graph: A Flexible and Reliable Reasoning Framework for Large Language Models (AAAI 2025 · similarity 0.89)
- Diagnosing and Addressing Pitfalls in KG-RAG Datasets: Toward More Reliable Benchmarking (NeurIPS 2025 · similarity 0.89)
- Towards Graph Foundation Models: Training on Knowledge Graphs Enables Transferability to General Graphs (NeurIPS 2025 · similarity 0.89)
- Plan-on-Graph: Self-Correcting Adaptive Planning of Large Language Model on Knowledge Graphs (NeurIPS 2024 · similarity 0.89)
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.
