Research map / Graph learning / Graph self-supervised learning
MUG: Meta-path-aware Universal Heterogeneous Graph Pre-Training
Lianze Shan, Jitao Zhao, Dongxiao He, Yongqi Huang, Zhiyong Feng, Weixiong Zhang
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 |
|---|---|---|---|
| add MUG universal pretraining over HeCo/HGMAE/HERO (SVD-unified inputs) | MUG · cross-domain node classification (all 4 targets, ACM/Freebase source) | Macro-F1: AMiner 44.08 -> 76.35 (vs HERO, ACM-pretrained) | Improved |
| add MUG on source-matched target ACM | MUG · ACM (pretrained on Freebase) | Macro-F1: 84.78 -> 85.21 (vs HGMAE) | No clear effect |
| MUG in few-shot setting | MUG · DBLP and ACM few-shot (1/3/5-shot) | Macro-F1: DBLP 1-shot 61.46 -> 84.24 | Improved |
| MUG on Freebase target in few-shot | MUG · Freebase few-shot 1-shot | Macro-F1: 32.20 -> 33.24 (vs HERO) | No clear effect |
| remove contextual structural encoding (CSE) (ablation) | MUG · cross-domain node classification (trained on Freebase) | Macro-F1 | Improved |
| remove alignment loss (ablation) | MUG · cross-domain node classification (trained on Freebase) | Macro-F1 | Improved |
| remove scattering loss (ablation) | MUG · cross-domain node classification (trained on Freebase) | Macro-F1 | Improved |
Closest papers by meaning
- HeterSEED: Semantics–Structure Decoupling for Heterogeneous Graph Learning under Heterophily (NeurIPS 2026 · similarity 0.88)
- Harnessing Language Model for Cross-Heterogeneity Graph Knowledge Transfer (AAAI 2025 · similarity 0.87)
- UniGTE: Unified Graph–Text Encoding for Zero-Shot Generalization across Graph Tasks and Domains (NeurIPS 2025 · similarity 0.87)
- Multiplex Graph Representation Learning with Homophily and Consistency (AAAI 2025 · similarity 0.87)
- Cross-Domain Graph Data Scaling: A Showcase with Diffusion Models (NeurIPS 2025 · similarity 0.86)
- Heterogeneous Graph Neural Network on Semantic Tree (AAAI 2025 · similarity 0.86)
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.
