Research map / Graph learning / Graph self-supervised learning
Multi-resolution Spectral Coherence for Graph Generation with Score-based Diffusion
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 |
|---|---|---|---|
| add multi-resolution wavelet spectral coherence to the score model (Wave-GD vs GDSS) | GDSS · Ego-small | average MMD: 0.017 -> 0.009 | Helped |
| add multi-resolution wavelet spectral coherence (Wave-GD vs GDSS) | GDSS · Community-small | average MMD: 0.046 -> 0.022 | Helped |
| add multi-resolution wavelet spectral coherence (Wave-GD vs GDSS) | GDSS · Grid | average MMD: 0.062 -> 0.056 | No clear effect |
| add multi-resolution wavelet spectral coherence (Wave-GD vs GDSS) | GDSS · QM9 | average of validity, novelty, uniqueness: 93.48 -> 95.20 | Helped |
| add multi-resolution wavelet spectral coherence (Wave-GD vs GDSS) | GDSS · QM9 | FCD: 2.90 -> 3.85 | Hurt |
| add multi-resolution wavelet spectral coherence (Wave-GD vs GDSS) | GDSS · QM9 | generation time (s): 1.14e2 -> 1.21e2 | No clear effect |
| evaluate with 1024 generated samples instead of test-set size (Wave-GD vs GDSS) | GDSS · Ego-small | average MMD: 0.016 -> 0.011 | Helped |
| evaluate with 1024 generated samples (Wave-GD vs GraphRNN) | GraphRNN · Community-small | average MMD: 0.017 -> 0.031 | Hurt |
| increase GMH depth D (oversmoothing study), Wave-GD vs GDSS (ablation) | GDSS · Community-small | average MMD | Helped |
| vary number of scales J from 3 to 7 (ablation) | Wave-GD · Community-small | average MMD: 0.028, 0.031, 0.024, 0.022, 0.027 | No clear effect |
| vary number of scales J from 4 to 7 (ablation) | Wave-GD · QM9 | average of validity, uniqueness, novelty: 89.87 -> 94.85 | Mixed |
| vary number of scales J from 8 to 12 (ablation) | Wave-GD · Ego-small | average MMD: 0.015, 0.011, 0.009, 0.009, 0.016 | Mixed |
Closest papers by meaning
- WaveNet: Tackling Non-stationary Graph Signals via Graph Spectral Wavelets (AAAI 2024 · similarity 0.90)
- Generating Graphs via Spectral Diffusion (ICLR 2025 · similarity 0.89)
- A General Graph Spectral Wavelet Convolution via Chebyshev Order Decomposition (ICML 2025 · similarity 0.89)
- Graph Generation with Diffusion Mixture (ICML 2024 · similarity 0.87)
- When Edge Independence Fails: Joint Graph Diffusion with Latent Sociability Priors (NeurIPS 2026 · similarity 0.87)
- Advancing Graph Generation through Beta Diffusion (ICLR 2025 · similarity 0.86)
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.