Research map / Graph learning / Graph self-supervised learning
Data Representation and Learning With Graph Diffusion-Embedding Networks
Check 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 equilibrium graph diffusion with random walk with restart in place of GCN propagation | GDEN · Citeseer | accuracy: 71.0% -> 72.8% | Helped |
| Use normalized Laplacian diffusion in place of GCN propagation | GDEN · Cora | accuracy: 83.6% -> 85.1% | Helped |
| Use Laplacian diffusion in place of GCN propagation | GDEN · Pubmed | accuracy: 78.3% -> 78.9% | Helped |
| Use normalized Laplacian diffusion instead of GCN propagation | GDEN · SVHN, 1000 labels | accuracy: 81.07±0.52% -> 82.88±0.54% | Helped |
| Use normalized Laplacian diffusion instead of GCN propagation | GDEN · CIFAR10, 4000 labels | accuracy: 71.55±0.29% -> 75.46±0.21% | Helped |
| Use adaptive diffusion over multiple graphs instead of averaging graphs for GCN | M-GDEN · MSRC-v1, 10% labels | accuracy: 80.36 -> 84.70 | Helped |
| Use adaptive diffusion over multiple graphs instead of averaging graphs for GCN | M-GDEN · Caltech101-7, 30% labels | accuracy: 94.68 -> 96.74 | Helped |
| Replace normalized Laplacian diffusion with RWR diffusion (ablation) | GDEN · CIFAR10 | accuracy: 69.08±0.53% -> 68.60±0.52%; 72.56±0.29% -> 72.12±0.42%; 75.46±0.21% -> 75.19±0.35% | Mixed |
| Replace normalized Laplacian diffusion with RWR diffusion (ablation) | GDEN · SVHN | accuracy: 82.88±0.54% -> 82.72±0.53%; 84.80±0.39% -> 84.76±0.33%; 86.24±0.36% -> 86.36±0.41% | Mixed |
| Vary the number of hidden units (ablation) | Two-layer GDEN · CIFAR10 and SVHN | accuracy | No clear effect |
| Vary the number of hidden layers (ablation) | GDEN · CIFAR10 and SVHN | accuracy | No clear effect |
| Use multiple graphs with adaptive graph weights instead of a single graph (ablation) | M-GDEN · MSRC-v1 and Caltech101-7 | accuracy: 80.36 -> 84.70; 91.05 -> 92.69 | Helped |
Closest papers by meaning
- Diffusion-Convolutional Neural Networks (NeurIPS 2016 · similarity 0.92)
- Diffusion Improves Graph Learning (NeurIPS 2019 · similarity 0.90)
- Diffusion-Guided Graph Data Augmentation (NeurIPS 2025 · similarity 0.89)
- Learning to Approximate Adaptive Kernel Convolution on Graphs (AAAI 2024 · similarity 0.88)
- Dissecting the Diffusion Process in Linear Graph Convolutional Networks (NeurIPS 2021 · similarity 0.88)
- A Generalized Neural Diffusion Framework on Graphs (AAAI 2024 · similarity 0.88)
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.