Research map / Graph learning / GNN architectures
Deep Graph Reprogramming
Yongcheng Jing, Chongbin Yuan, Li Ju, Yiding Yang, Xinchao Wang, Dacheng Tao
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 task-adaptive ReAgg to MetaFP | MetaFP reusing a Citeseer-pretrained GNN · Pubmed | accuracy: 0.7790 -> 0.8030 | Improved |
| Add task-adaptive ReAgg to MetaFP | MetaFP reusing a Citeseer-pretrained GNN · Computers | accuracy: 0.9085 -> 0.9229 | Improved |
| Add task-adaptive ReAgg to MetaFP | MetaFP reusing a Citeseer-pretrained GNN · Photo | accuracy: 0.8909 -> 0.9213 | Improved |
| Add task-adaptive ReAgg to MetaFP | MetaFP reusing a Citeseer-pretrained GNN · Cora | accuracy: 0.8335 -> 0.8312 | Hurt |
| Add MetaFP feature padding rather than vanilla reuse | Frozen GNN reused across graph classification and regression tasks · QM7b | MAE: 24.18 -> 2.3093 | Improved |
| Add MetaFP feature padding rather than vanilla reuse | Frozen GNN reused across graph classification and regression tasks · PROTEINS | accuracy: 0.3304 -> 0.6071 | Improved |
| Add MetaGP meta-graph padding rather than vanilla reuse | GNN pretrained on ogbg-molbace · ogbn-molbbbp | ROC-AUC: 0.5136 -> 0.6691 | Improved |
| Add MetaGP meta-graph padding rather than vanilla reuse | GNN pretrained on ogbg-molbace · ogbn-molesol | RMSE: 6.950 -> 2.050 | Improved |
| Add MetaGP meta-graph padding rather than vanilla reuse | DGCNN pretrained on ModelNet40 · ShapeNet | accuracy: 0.1545 -> 0.6090 | Improved |
| Use MetaFP padding rather than vanilla reuse | GNN pretrained on Citeseer · Cora | accuracy: 0.1571 -> 0.8335 | Improved |
| Use random padding placement rather than fixed placement (ablation) | MetaFP · Computers → Photo | accuracy: 0.9183; 0.9161; 0.9212; 0.9165 | No clear effect |
| Change feature-padding position among front, center, end, and random (ablation) | MetaFP · Photo → Pubmed | accuracy: 0.8420; 0.8430; 0.8420; 0.8440 | No clear effect |
Closest papers by meaning
- Towards Unsupervised Open-Set Graph Domain Adaptation via Dual Reprogramming (NeurIPS 2025 · similarity 0.88)
- Locality-Aware Graph Rewiring in GNNs (ICLR 2024 · similarity 0.87)
- Model Reprogramming: Resource-Efficient Cross-Domain Machine Learning (AAAI 2024 · similarity 0.87)
- GraphTOP: Graph Topology-Oriented Prompting for Graph Neural Networks (NeurIPS 2025 · similarity 0.86)
- Towards Pre-trained Graph Condensation via Optimal Transport (NeurIPS 2025 · similarity 0.86)
- DFA-GNN: Forward Learning of Graph Neural Networks by Direct Feedback Alignment (NeurIPS 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.
