Research map / Graph learning / GNN architectures
Block Modeling-Guided Graph Convolutional Neural Networks
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 block-modeling-guided classified aggregation | BM-GCN · Texas | accuracy (%): 85.13±4.64 vs 84.59±4.37 | Helped |
| Use block-modeling-guided classified aggregation | BM-GCN · Squirrel | accuracy (%): 51.41±1.10 vs 44.07±1.95 | Helped |
| Use block-modeling-guided classified aggregation | BM-GCN · Chameleon | accuracy (%): 69.58±2.90 vs 69.78±1.97 | No clear effect |
| Use block-modeling-guided classified aggregation | BM-GCN · Cora | accuracy (%): 87.99±1.29 vs 87.69±1.37 | Helped |
| Use block-modeling-guided classified aggregation | BM-GCN · Citeseer | accuracy (%): 76.13±1.92 vs 75.95±2.18 | Helped |
| Use block-modeling-guided classified aggregation | BM-GCN · Pubmed | accuracy (%): 90.25±0.75 vs 89.08±0.67 | Helped |
| Use 3 graph convolution layers rather than 1 (ablation) | BM-GCN · Cora | accuracy (%): 81.27 -> 87.99 | Helped |
| Increase graph convolution layers from 3 to 5 (ablation) | BM-GCN · Cora | accuracy (%): 87.99 -> 60.10 | Hurt |
| Increase graph convolution layers from 3 to 6 (ablation) | BM-GCN · Cora | accuracy (%): 87.99 -> 40.16 | Hurt |
| Use 3 graph convolution layers rather than 1 (ablation) | BM-GCN · Chameleon | accuracy (%): 58.71 -> 69.58 | Helped |
| Increase graph convolution layers from 3 to 5 (ablation) | BM-GCN · Chameleon | accuracy (%): 69.58 -> 49.28 | Hurt |
| Increase graph convolution layers from 3 to 6 (ablation) | BM-GCN · Chameleon | accuracy (%): 69.58 -> 34.69 | Hurt |
Closest papers by meaning
- Powerful Graph Convolutional Networks with Adaptive Propagation Mechanism for Homophily and Heterophily (AAAI 2022 · similarity 0.92)
- Universal Graph Convolutional Networks (NeurIPS 2021 · similarity 0.91)
- Integrating Co-Training with Edge Discrimination to Enhance Graph Neural Networks Under Heterophily (AAAI 2025 · similarity 0.90)
- Refining Latent Homophilic Structures over Heterophilic Graphs for Robust Graph Convolution Networks (AAAI 2024 · similarity 0.88)
- Graph Neural Networks Are More Than Filters: Revisiting and Benchmarking from A Spectral Perspective (ICLR 2025 · similarity 0.87)
- Making Classic GNNs Strong Baselines Across Varying Homophily: A Smoothness–Generalization Perspective (NeurIPS 2025 · similarity 0.87)
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.