Research map / Graph learning / GNN architectures
Differentiable Graph Module (DGM) for Graph Convolutional Networks
Anees Kazi, Luca Cosmo, Seyed‐Ahmad Ahmadi, Nassir Navab, Michael M. Bronstein
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 a DGM layer that learns the graph (ablation) | GCN · Cora | accuracy: 80.90 -> 84.70 | Improved |
| add a DGM layer that learns the graph (ablation) | GCN · CiteSeer | accuracy: 71.90 -> 74.70 | Improved |
| add a DGM layer that learns the graph (ablation) | GCN · PubMed | accuracy: 86.40 -> 87.60 | Improved |
| stack a third DGM layer (ablation) | dDGM · Cora | accuracy: 84.70 -> 82.40 | Hurt |
| use k=5 instead of k=20 neighbors (ablation) | dDGM · Cora | accuracy: 70.60 -> 84.60 | Improved |
| use k=5 instead of k=1 neighbors (ablation) | dDGM · CiteSeer | accuracy: 67.00 -> 74.80 | Improved |
| use Euclidean instead of hyperbolic latent space (ablation) | dDGM · PubMed | accuracy: 86.60 -> 87.60 | Improved |
| use Euclidean instead of hyperbolic latent space (ablation) | dDGM · Cora | accuracy: 84.40 -> 84.60 | No clear effect |
| use input graph with GCN embedding function instead of MLP (ablation) | dDGM · Cora | accuracy: 63.30 -> 84.70 | Improved |
| use GCN diffusion instead of EdgeConv (ablation) | dDGM · CiteSeer | accuracy: 52.20 -> 74.80 | Improved |
| use GCN diffusion instead of GAT (ablation) | dDGM · Cora | accuracy: 81.30 -> 84.60 | Improved |
| add DGM before EdgeConv diffusion (ablation) | EdgeConv · Cora | accuracy: 67.00 -> 71.60 | Improved |
Closest papers by meaning
- Bayesian Graph Convolutional Neural Networks for Semi-Supervised Classification (AAAI 2019 · similarity 0.87)
- Convolutional Kernel Networks for Graph-Structured Data (ICML 2020 · similarity 0.87)
- Latent Graph Inference using Product Manifolds (ICLR 2023 · similarity 0.87)
- From Latent Graph to Latent Topology Inference: Differentiable Cell Complex Module (ICLR 2024 · similarity 0.86)
- DFA-GNN: Forward Learning of Graph Neural Networks by Direct Feedback Alignment (NeurIPS 2024 · similarity 0.86)
- Generalizing CNNs to graphs with learnable neighborhood quantization (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.
