Research map / Graph learning / Graph self-supervised learning
Convolutional Kernel Networks for Graph-Structured Data
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 path aggregation instead of walk aggregation (ablation) | GCKN · PTC | accuracy: 65.9\pm 2.0 -> 67.3\pm 5.0 | Helped |
| Add the subtree layer to the path model (ablation) | GCKN · PTC | accuracy: 67.3\pm 5.0 -> 70.8\pm 4.6 | Helped |
| Add the subtree layer to the path model (ablation) | GCKN · NCI1 | accuracy: 81.4\pm 1.6 -> 83.9\pm 1.6 | Helped |
| Add the subtree layer to the path model (ablation) | GCKN · PROTEINS (continuous) | accuracy: 76.3\pm 0.5 -> 77.5\pm 0.3 | Mixed |
| Add a third layer with short paths (ablation) | GCKN · MUTAG | accuracy: 95.0\pm 5.2 -> 97.2\pm 2.8 | Helped |
| Train subtree filters end-to-end rather than unsupervised (ablation) | GCKN-subtree · PTC | accuracy: 70.8\pm 4.6 -> 68.4\pm 7.4 | Hurt |
| Train subtree filters end-to-end rather than unsupervised (ablation) | GCKN-subtree · BZR | accuracy: 85.8\pm 0.9 -> 86.4\pm 0.5 | Helped |
| Increase the number of Nyström filters (ablation) | Unsupervised GCKN-subtree · NCI1, PROTEINS, IMDB-MULTI | accuracy | Helped |
Closest papers by meaning
- Learning Convolutional Neural Networks for Graphs (ICML 2016 · similarity 0.93)
- Unsupervised Neighborhood Propagation Kernel Layers for Semi-supervised Node Classification (AAAI 2024 · similarity 0.89)
- DHAKR: Learning Deep Hierarchical Attention-Based Kernelized Representations for Graph Classification (AAAI 2025 · similarity 0.89)
- Graph Classification via Reference Distribution Learning: Theory and Practice (NeurIPS 2024 · similarity 0.89)
- On the Stability of Graph Convolutional Neural Networks: A Probabilistic Perspective (NeurIPS 2025 · similarity 0.89)
- Learning to Approximate Adaptive Kernel Convolution on Graphs (AAAI 2024 · 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.