AI Paper Atlas

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

TPAMI 2024 · Large language models › Retrieval and knowledge · approach: Knowledge graph reasoning with LLMs · 450 citations (OpenAlex)

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 changedWhereMeasuredEffect
Use ComplEx with DURA regularization instead of plain ComplExComplEx · FB15k-237 (transductive)MRR: 0.346 -> 0.370Improved
Replace TransE with rotation-based RotatETransE · WN18RR (transductive)MRR: 0.231 -> 0.476Improved
Replace RGCN with attention/composition GNN CompGCNRGCN · WN18RR (transductive)MRR: 0.427 -> 0.469Improved
Replace DistMult with ComplEx (complex embeddings)DistMult · FB15k-237 (transductive)MRR: 0.243 -> 0.346Improved
Replace ConvE with HousE (Householder-based geometric embedding)ConvE · WN18RR (transductive)MRR: 0.430 -> 0.511Improved
Use path-based GNN RED-GNN rather than embedding model HAKEHAKE · WN18RR (transductive)MRR: 0.497 -> 0.533Improved
Replace static TTransE with RNN/GNN-free ChronoR rotation modelTTransE · ICEWS14 (interpolation)MRR (%): 25.50 -> 62.53Improved
Use DyERNIE (Riemannian dynamic embeddings) instead of TComplExTComplEx · ICEWS14 (interpolation)MRR (%): 61.90 -> 66.90Improved
Add multi-modal fusion in MoSE versus image-based IKRLIKRL · FB15k-237-IMGHits@10 (%): 45.8 -> 56.5Improved
Use IMF interactive multi-modal fusion rather than MKGformerMKGformer · FB15k-237-IMGHits@10 (%): 49.9 -> 59.3Improved
Use pre-trained multi-modal transformer VisualBERT instead of dedicated KG modelMoSE · FB15k-237-IMGHits@10 (%): 56.5 -> 43.9Hurt
Use RGHAT instead of HousE on FB15k-237 but not on WN18RRHousE · FB15k-237 vs WN18RR (transductive)MRR: 0.361 -> 0.522 (FB15k-237); 0.511 -> 0.483 (WN18RR)Mixed

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy