Research map / Graph learning / Graph self-supervised learning
Dual-Geometry Graph Network: Unifying Local and Global Priors for Few-Shot Learning
Zheng Han, Xiaobin Zhu, Chun Yang, Jingyan Qin, Xu-Cheng Yin
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 the local curvature pathway to a baseline GNN (ablation) | baseline GNN · CUB-200-2011 5-way 1-shot | accuracy: 83.09 -> 87.92 | Improved |
| add the global resistive pathway to a baseline GNN (ablation) | baseline GNN · TieredImageNet 5-way 1-shot | accuracy: 76.25 -> 80.37 | Improved |
| add the global pathway to the local-only GNN (ablation) | local-only GNN · CUB-200-2011 5-way 1-shot | accuracy: 87.92 -> 90.71 | Improved |
| add the local pathway to the global-only GNN (ablation) | global-only GNN · TieredImageNet 5-way 1-shot | accuracy: 80.37 -> 84.77 | Improved |
| add curvature regularization Lorc (ablation) | DGGN · MiniImageNet 5-way 1-shot | accuracy: 73.39 -> 74.08 | Improved |
| add connectivity regularization Lres (ablation) | DGGN · MiniImageNet 5-way 1-shot | accuracy: 73.39 -> 74.78 | Improved |
| add both regularization losses (ablation) | DGGN · TieredImageNet 5-way 1-shot | accuracy: 82.13 -> 84.77 | Improved |
| use cascaded Local to Global instead of parallel pathways (ablation) | DGGN · MiniImageNet 5-way 1-shot | accuracy: 71.90 -> 75.27 | Improved |
| use Local to Global order instead of Global to Local (ablation) | DGGN · MiniImageNet 5-way 1-shot | accuracy: 73.42 -> 75.27 | Improved |
| add residual connections (ablation) | DGGN · TieredImageNet 5-way 5-shot | accuracy: 88.91 -> 89.80 | Improved |
| rebuild k-NN graph each layer instead of fixed edges (ablation) | DGGN · TieredImageNet 5-way 1-shot | accuracy: 82.69 -> 84.77 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DPGN — DPGN: Distribution Propagation Graph Network for Few-Shot Learning (CVPR 2020) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Adaptive Riemannian Graph Neural Networks (AAAI 2026 · similarity 0.87)
- Graph Geometry-Preserving Autoencoders (ICML 2024 · similarity 0.86)
- DAPrompt: Dual Alignment Prompt of Structure and Semantics for Few-shot Graph Learning (AAAI 2026 · similarity 0.86)
- Geometry Awakening: Cross-Geometry Learning Exhibits Superiority over Individual Structures (NeurIPS 2024 · similarity 0.86)
- Distilling Graph Geometry: Knowledge Gap from GNNs to MLPs (NeurIPS 2026 · similarity 0.86)
- Effective Structural Encodings via Local Curvature Profiles (ICLR 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.
