Research map / Graph learning / Graph algorithms and clustering
Isolation Graph Kernel
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 IGK instead of WWL | graph classification · ENZYMES | accuracy (%): 74.2 -> 77.0 | Helped |
| use IGK instead of WWL | graph classification · AIDS | accuracy (%): 99.5 -> 98.5 | Hurt |
| use IGK instead of WWL | graph classification · DHFR | accuracy (%): 78.9 -> 76.6 | Hurt |
| use IGK instead of WWL | graph classification · TWITTER | accuracy (%) | Helped |
| add dataset-wide Isolation Kernel node embedding (ablation) | IGK · ENZYMES | accuracy (%): 46.7 -> 77.0 | Helped |
| replace Gaussian layer-1 feature map with Isolation Kernel (ablation) | IGK · ENZYMES | accuracy (%): 47.8 -> 77.0 | Helped |
| use Isolation Kernel rather than Gaussian kernel in layer 3 (ablation) | IGK mean embedding · ENZYMES | accuracy (%): 70.8 -> 77.0 | Helped |
| add Isolation Kernel as the first embedding layer | WWL · ENZYMES | accuracy (%): 74.2% to 75.5% | Helped |
| use IGK instead of WWL | graph classification · 13 benchmark datasets | average rank: 2.76 -> 1.53 | Helped |
| use IGK instead of Wasserstein-distance graph kernel | large-scale graph classification · TWITTER | runtime: >24hr -> 1,431 seconds | Helped |
Closest papers by meaning
- Wasserstein Weisfeiler-Lehman Graph Kernels (NeurIPS 2019 · similarity 0.93)
- Graph Filtration Kernels (AAAI 2022 · similarity 0.89)
- Matching Node Embeddings for Graph Similarity (AAAI 2017 · similarity 0.88)
- Rethinking Kernel Methods for Node Representation Learning on Graphs (NeurIPS 2019 · similarity 0.88)
- DHAKR: Learning Deep Hierarchical Attention-Based Kernelized Representations for Graph Classification (AAAI 2025 · similarity 0.85)
- Exploring Consistency in Graph Representations: from Graph Kernels to Graph Neural Networks (NeurIPS 2024 · similarity 0.85)
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.