Research map / Graph learning / GNN architectures
Full-Spectrum Graph Neural Networks: Expressive and Scalable
Xiaohan Wang, Deyu Bo, Longlong Li, Kelin Xia
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 off-diagonal eigenspace interactions (ablation) | FSpecGNN(Cheb) · Texas | accuracy: 51.0 -> 55.3 | Improved |
| add off-diagonal eigenspace interactions (ablation) | FSpecGNN(Cheb) · Squirrel | accuracy: 34.6 -> 39.6 | Improved |
| add off-diagonal eigenspace interactions (ablation) | FSpecGNN(Cheb) · Roman Empire | accuracy: 53.8 -> 54.4 | Improved |
| add in-filter f (ablation) | FSpecGNN(Cheb) · Texas | accuracy: 49.5 -> 55.3 | Improved |
| add in-filter f (ablation) | FSpecGNN(Cheb) · Squirrel | accuracy: 29.9 -> 39.6 | Improved |
| add in-filter f (ablation) | FSpecGNN(Cheb) · Roman Empire | accuracy: 52.0 -> 54.4 | Improved |
| add in-filter f and off-diagonal interactions together (ablation) | FSpecGNN(Cheb) · Texas | accuracy: 39.7 -> 55.3 | Improved |
| add in-filter f and off-diagonal interactions together (ablation) | FSpecGNN(Cheb) · Squirrel | accuracy: 24.5 -> 39.6 | Improved |
| add in-filter f and off-diagonal interactions together (ablation) | FSpecGNN(Cheb) · Roman Empire | accuracy: 47.1 -> 54.4 | Improved |
| add full-spectrum eigenspace interaction to Chebyshev filter (ChebNet to FSpecGNN(Cheb)) | ChebNet · Squirrel | accuracy: 22.58 -> 39.57 | Improved |
| add full-spectrum eigenspace interaction to Chebyshev filter (ChebNet to FSpecGNN(Cheb)) | ChebNet · Texas | accuracy: 36.76 -> 55.32 | Improved |
| add full-spectrum interaction to Bernstein filter (BernNet to FSpecGNN(Bern)) | BernNet · Questions | ROC-AUC: 65.41 -> 77.11 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- BernNet — BernNet: Learning Arbitrary Graph Spectral Filters via Bernstein Approximation (NeurIPS 2021) · better on 4 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Graph Neural Networks Are More Than Filters: Revisiting and Benchmarking from A Spectral Perspective (ICLR 2025 · similarity 0.93)
- SLOG: An Inductive Spectral Graph Neural Network Beyond Polynomial Filter (ICML 2024 · similarity 0.92)
- Unifying Homophily and Heterophily for Spectral Graph Neural Networks via Triple Filter Ensembles (NeurIPS 2024 · similarity 0.92)
- Position: Spectral GNNs Rely Less on Graph Fourier Basis than Conceived (ICML 2025 · similarity 0.92)
- How Universal Polynomial Bases Enhance Spectral Graph Neural Networks: Heterophily, Over-smoothing, and Over-squashing (ICML 2024 · similarity 0.92)
- On the Expressive Power of Spectral Invariant Graph Neural Networks (ICML 2024 · similarity 0.91)
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.
