Research map / Learning theory and algorithms / Clustering
Spectral Clustering with Graph Neural Networks for Graph Pooling
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 MinCutPool instead of DiffPool for node clustering | GNN clustering · Cora and Citeseer | NMI: Cora: 0.315 ± 0.005 -> 0.404 ± 0.018; Citeseer: 0.139 ± 0.016 -> 0.287 ± 0.047 | Helped |
| Use MinCutPool instead of spectral clustering | GNN clustering · Cora, Citeseer, and Pubmed | NMI: Cora: 0.025 ± 0.014 -> 0.404 ± 0.018; Citeseer: 0.014 ± 0.003 -> 0.287 ± 0.047; Pubmed: 0.182 ± 0.000 -> 0.200 ± 0.020 | Mixed |
| Use MinCutPool instead of DiffPool | Graph classification GNN · Bench-easy | accuracy: 98.6 ± 0.4 -> 99.0 ± 0.0 | Helped |
| Use MinCutPool instead of NDP | Graph classification GNN · Bench-hard | accuracy: 72.6 ± 0.9 -> 73.8 ± 1.9 | Helped |
| Use MinCutPool instead of the strongest listed baseline | Graph classification GNN · Mutagenicity | accuracy: 81.7 ± 1.1 -> 79.9 ± 2.1 | Hurt |
| Use MinCutPool instead of NDP | Graph classification GNN · Proteins | accuracy: 73.3 ± 3.7 -> 76.5 ± 2.6 | Helped |
| Use MinCutPool instead of DiffPool | Graph classification GNN · DD | accuracy: 79.3 ± 2.4 -> 80.8 ± 2.3 | Helped |
| Use MinCutPool instead of No-pool | Graph classification GNN · COLLAB | accuracy: 82.1 ± 1.8 -> 83.4 ± 1.7 | Helped |
| Use MinCutPool instead of DiffPool | Graph classification GNN · Reddit-Binary | accuracy: 86.8 ± 2.1 -> 91.4 ± 1.5 | Helped |
| Use MinCutPool instead of the flat no-pooling baseline | QM9 graph regression GNN · QM9 mu | MSE: 0.559 -> 0.538 | Helped |
| Use MinCutPool instead of the flat no-pooling baseline | QM9 graph regression GNN · QM9 alpha, homo, lumo, gap, r2, zpve, u0_atom, cv | MSE: alpha: 0.078 vs 0.065; homo: 0.526 vs 0.435; lumo: 0.540 vs 0.515; gap: 0.584 vs 0.552; r2: 0.261 vs 0.204; zpve: 0.328 | Hurt |
Closest papers by meaning
- Total Variation Graph Neural Networks (ICML 2023 · similarity 0.91)
- SpectralNet: Spectral Clustering using Deep Neural Networks (ICLR 2018 · similarity 0.90)
- Graph Neural Networks Are More Than Filters: Revisiting and Benchmarking from A Spectral Perspective (ICLR 2025 · similarity 0.88)
- Pass-Efficient Algorithms for Graph Spectral Clustering (Student Abstract) (AAAI 2024 · similarity 0.87)
- Geometry-Aware Edge Pooling for Graph Neural Networks (NeurIPS 2025 · similarity 0.87)
- Revisiting Self-Supervised Heterogeneous Graph Learning from Spectral Clustering Perspective (NeurIPS 2024 · 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.