Research map / Graph learning / GNN architectures
Efficient and Stable Graph Scattering Transforms via Pruning
Vassilis N. Ioannidis, Siheng Chen, Georgios B. Giannakis
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 |
|---|---|---|---|
| prune scattering tree by spectrum criterion (pDS, 61 of 781 features) | graph scattering transform with diffusion wavelets · Authorship attribution (N=244) | classification accuracy: 781 -> 61 coefficients (12.8%) | No clear effect |
| prune scattering tree (pMCS, 30 features) | graph scattering transform with monic cubic wavelets · Authorship attribution (N=244) | classification accuracy: 781 -> 30 coefficients (3.8%) | No clear effect |
| prune scattering tree (pTHS, 80 features) | graph scattering transform with tight Hann wavelets · Authorship attribution (N=244) | classification accuracy: 781 -> 80 coefficients (10.2%) | No clear effect |
| prune scattering tree | graph scattering transform · Facebook source localization (N=234) | classification accuracy vs SNR | No clear effect |
| increase pruning threshold tau (ablation) | pGST · Facebook source localization | classification accuracy | Hurt |
| increase number of layers L and filters J (ablation) | pGST · Facebook source localization | classification accuracy | Improved |
| prune the scattering tree (pGST vs GST) | gradient boosting on scattering features · Graph classification, Proteins | accuracy (%): 76.23 -> 78.57 | Improved |
| prune the scattering tree (pGST vs GST) | gradient boosting on scattering features · Graph classification, D&D | accuracy (%): 79.28 -> 81.27 | Improved |
| prune the scattering tree (pGST vs GST) | gradient boosting on scattering features · Graph classification, Enzymes | accuracy (%): 59.84 -> 60.25 | Improved |
| prune the scattering tree (pGST vs GST) | gradient boosting on scattering features · Graph classification, Collab | accuracy (%): 77.32 -> 78.40 | Improved |
| prune the scattering tree (pGST vs GST) | scattering features with MLP · Cora node classification | accuracy (%): 81.9 -> 81.9 | No clear effect |
| prune the scattering tree (pGST vs GST) | scattering features · ModelNet40 point clouds | classification accuracy | Improved |
Closest papers by meaning
- Pruned Graph Scattering Transforms (ICLR 2020 · similarity 0.97)
- Scattering GCN: Overcoming Oversmoothness in Graph Convolutional Networks (NeurIPS 2020 · similarity 0.91)
- Limiting Over-Smoothing and Over-Squashing of Graph Message Passing by Deep Scattering Transforms (JMLR 2026 · similarity 0.88)
- Unsupervised Parameter-free Simplicial Representation Learning with Scattering Transforms (ICML 2024 · similarity 0.88)
- GDeR: Safeguarding Efficiency, Balancing, and Robustness via Prototypical Graph Pruning (NeurIPS 2024 · similarity 0.86)
- Graph Context Transformation Learning for Progressive Correspondence Pruning (AAAI 2024 · 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.
