Research map / Graph learning / GNN architectures
VQ-GNN: A Universal Framework to Scale up Graph Neural Networks using Vector Quantization
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 VQ-GNN instead of sampling-based scalable training | GCN on PPI · PPI | F1-score: $.8852\pm.0066$ -> $.9549\pm.0058$ | Helped |
| Use VQ-GNN instead of the strongest reported scalable baseline | GCN on Reddit · Reddit | accuracy: $.9264\pm.0034$ -> $.9399\pm.0021$ | Helped |
| Use VQ-GNN instead of NS-SAGE | SAGE-Mean on Reddit · Reddit | accuracy: $.9615\pm.0089$ -> $.9449\pm.0024$ | Hurt |
| Use VQ-GNN instead of full-graph training | GAT on ogbl-collab · ogbl-collab | Hits@50: $.4048\pm.0125$ -> $.4102\pm.0099$ | Helped |
| Use VQ-GNN instead of full-graph training | SAGE-Mean on ogbl-collab · ogbl-collab | Hits@50: $.4810\pm.0081$ -> $.4673\pm.0164$ | Hurt |
| Use VQ-GNN instead of sampling-based inference | SAGE-Mean on ogbn-arxiv · ogbn-arxiv | inference time: 1.61 s -> 0.40 s | Helped |
| Use VQ-GNN rather than GraphSAINT-RW at the same message count | GCN on ogbn-arxiv · ogbn-arxiv; 1.5M messages per batch | peak memory: 661.6 MB -> 485.5 MB | Helped |
| Use VQ-GNN rather than GraphSAINT-RW at the same node count | GCN on ogbn-arxiv · ogbn-arxiv; 85K nodes per batch | peak memory: 501.5 MB -> 758.0 MB | Hurt |
| Use VQ-GNN instead of full-graph training | GCN on ogbn-arxiv · ogbn-arxiv | accuracy: $.7029\pm.0036$ -> $.7055\pm.0033$ | No clear effect |
Closest papers by meaning
- Generalizing CNNs to graphs with learnable neighborhood quantization (NeurIPS 2024 · similarity 0.93)
- Efficient Learning of Linear Graph Neural Networks via Node Subsampling (NeurIPS 2023 · similarity 0.91)
- VQGraph: Rethinking Graph Representation Space for Bridging GNNs and MLPs (ICLR 2024 · similarity 0.90)
- MuseGNN: Forming Scalable, Convergent GNN Layers that Minimize a Sampling-Based Energy (ICLR 2025 · similarity 0.90)
- A Manifold Perspective on the Statistical Generalization of Graph Neural Networks (ICML 2025 · similarity 0.90)
- Can Classic GNNs Be Strong Baselines for Graph-level Tasks? Simple Architectures Meet Excellence (ICML 2025 · similarity 0.89)
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.