Research map / Graph learning / Graph algorithms and clustering
Taming graph kernels with random features
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 GRFs instead of brute-force kernel computation | Regularized Laplacian kernel computation · Speed test, N=800 | FLOPS: 512.64M -> 960K | Helped |
| Use GRFs instead of brute-force kernel computation | Regularized Laplacian kernel computation · Speed test, N=1000 | FLOPS: 1001M -> 1.4M | Helped |
| Use GRFs instead of brute-force kernel computation | Regularized Laplacian kernel computation · Speed test, N=3000 | FLOPS: 27B -> 10.2M | Helped |
| Apply JLT or anchor-point compression (ablation) | GRFs, d=1 · citeseer | clustering error: 0.020 -> 0.010 | Helped |
| Apply JLT or anchor-point compression (ablation) | GRFs, d=1 · Databases | clustering error: 0.170 -> 0.097 | Helped |
| Apply JLT or anchor-point compression (ablation) | GRFs, d=1 · polbooks and karate | clustering error: 0.28 -> 0.323; 0.11 -> 0.314 | Hurt |
| Apply JLT or anchor-point compression (ablation) | GRFs, d=2 · citeseer | clustering error: 0.008 -> 0.010 | Hurt |
| Apply JLT or anchor-point compression (ablation) | GRFs, d=2 · Databases, polbooks, and karate | clustering error: 0.140 -> 0.098; 0.12 -> 0.264; 0.032 -> 0.198 | Hurt |
| Use termination probability p_term=0.1 rather than longer average walks | GRF kernel estimation · Tested graphs | relative Frobenius norm error | No clear effect |
| Increase the number of random walks | GRF kernel estimation · Tested graphs | relative Frobenius norm error | Helped |
Closest papers by meaning
- General Graph Random Features (ICLR 2024 · similarity 0.95)
- Computationally-efficient Graph Modeling with Refined Graph Random Features (ICML 2026 · similarity 0.93)
- Quasi-Monte Carlo Graph Random Features (NeurIPS 2023 · similarity 0.92)
- RetGK: Graph Kernels based on Return Probabilities of Random Walks (NeurIPS 2018 · similarity 0.90)
- Manifold Random Features (NeurIPS 2026 · similarity 0.90)
- Graph Random Features for Scalable Gaussian Processes (ICLR 2026 · similarity 0.90)
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.