Research map / Graph learning / GNN architectures
Hybrid Graph Neural Networks for Few-Shot Learning
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use HGNN feature adaptation instead of ProtoNet alone (ablation) | ProtoNet with ResNet-12 · MiniImageNet 5-shot | accuracy (%): 80.49 -> 83.00 | Helped |
| Add PGNN (ablation) | ProtoNet · MiniImageNet 5-shot | accuracy (%): 80.49 -> 82.62 | Helped |
| Add IGNN (ablation) | ProtoNet · MiniImageNet 5-shot | accuracy (%): 80.49 -> 82.20 | Helped |
| Combine PGNN and IGNN (ablation) | ProtoNet · MiniImageNet 5-shot | accuracy (%): 82.62 -> 83.00 | Helped |
| Add KL consistency loss (ablation) | HGNN · MiniImageNet 1-shot | accuracy (%): 66.70 -> 67.02 | Helped |
| Add KL consistency loss (ablation) | HGNN · MiniImageNet 5-shot | accuracy (%): 82.61 -> 83.00 | Helped |
| Replace KL consistency with L1 consistency (ablation) | HGNN · MiniImageNet 1-shot and 5-shot | accuracy (%): 66.82; 82.71 | Mixed |
| Replace KL consistency with MSE consistency (ablation) | HGNN · MiniImageNet 1-shot and 5-shot | accuracy (%): 66.75; 82.65 | Mixed |
| Change the adjacency operator (ablation) | HGNN · MiniImageNet 5-shot | accuracy (%): 83.00; 82.61; 82.93; 82.64; 83.01 | No clear effect |
| Increase IGNN depth from 1 layer to 2 or 3 layers (ablation) | IGNN · MiniImageNet 5-shot | accuracy (%): 82.20 -> 81.57; 78.08 | Hurt |
| Increase PGNN depth from 1 layer to 2 or 3 layers (ablation) | PGNN · MiniImageNet 5-shot | accuracy (%): 82.62 -> 81.71; 81.38 | Hurt |
| Use HGNN rather than the strongest listed baseline | Few-shot classifiers · MiniImageNet, TieredImageNet, and CUB-200-2011, 1-shot and 5-shot | accuracy (%) | Helped |
Closest papers by meaning
- Graph Few-Shot Learning via Knowledge Transfer (AAAI 2020 · similarity 0.91)
- FEW-SHOT LEARNING ON GRAPHS VIA SUPER-CLASSES BASED ON GRAPH SPECTRAL MEASURES (ICLR 2020 · similarity 0.90)
- Unlocking the Potential of Black-box Pre-trained GNNs for Graph Few-shot Learning (AAAI 2025 · similarity 0.90)
- LLMs as Zero-shot Graph Learners: Alignment of GNN Representations with LLM Token Embeddings (NeurIPS 2024 · similarity 0.88)
- Fast Graph Sharpness-Aware Minimization for Enhancing and Accelerating Few-Shot Node Classification (NeurIPS 2024 · similarity 0.87)
- HGPrompt: Bridging Homogeneous and Heterogeneous Graphs for Few-Shot Prompt Learning (AAAI 2024 · 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.