AI Paper Atlas

Research map / Graph learning / Knowledge and temporal graphs

UniHR: Hierarchical Representation Learning for Unified Knowledge Graph Link Prediction

Zhiqiang Liu, Yin Hua, Mingyang Chen, Yichi Zhang, Zhuo Chen, Lei Liang, Wen Zhang

AAAI 2026 · Graph learning › Knowledge and temporal graphs · approach: Knowledge graph embeddings · 6 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
add intra-fact message passing (ablation)UniHR · FBHE (NKG)MRR: 0.754 -> 0.793Improved
add inter-fact message passing (ablation)UniHR · wikidata12k (TKG)MRR: 0.319 -> 0.334Improved
add inter-fact message passing (ablation)UniHR · DB15K (HKG)MRR: 0.338 -> 0.348Improved
add fact-node initialization from main triple (ablation)UniHR · FBHE (NKG)MRR: 0.767 -> 0.793Improved
add relation projection Wr (ablation)UniHR · FBHE (NKG)MRR: 0.792 -> 0.793Improved
add edge-type weight and direction weight (ablation)UniHR · DB15K (HKG)MRR: 0.342/0.341 -> 0.348Improved
UniHR vs StarE customized GNNUniHR · WD50KMRR: 0.309 -> 0.348Improved
UniHR vs specialised HKG methods HAHE and HyperSATUniHR · WikiPeople subject/objectMRR: 0.498 -> 0.496No clear effect
UniHR vs NestE on nested triple predictionUniHR · FBHE triple predictionMRR: 0.851 -> 0.793Hurt
UniHR vs NestE on nested triple predictionUniHR · FBH triple predictionMRR: 0.922 -> 0.946Improved
UniHR vs specialised TKG methodsUniHR · wikidata12kH@10: 0.546 -> 0.527Hurt
joint training of HKG and TKG data (wikimix)UniHR · wikimixMR and othersImproved

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