Research map / Graph learning / Graph algorithms and clustering
On clustering network-valued 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 log-traces of adjacency-matrix powers instead of graphlet kernels | Network clustering without node correspondence · Simulated networks | clustering error: 0 vs 0.5 (GK3), 0.36 (GK4), 0.26 (GK5) | Helped |
| Use log-traces of adjacency-matrix powers instead of global graph statistics | Network clustering without node correspondence · Simulated networks | clustering error: 0 vs 0.37 | Helped |
| Use log-traces of adjacency-matrix powers instead of top adjacency eigenvalues | Network clustering without node correspondence · Simulated networks | clustering error: 0 vs 0.18 | Helped |
| Use log-traces of adjacency-matrix powers instead of graphlet kernels | Network clustering without node correspondence · Real networks | clustering error: 0.1 vs 0.6 | Helped |
| Use log-traces of adjacency-matrix powers instead of global graph statistics | Network clustering without node correspondence · Real networks | clustering error: 0.1 vs 0.16 | Helped |
| Use log-traces of adjacency-matrix powers instead of top adjacency eigenvalues | Network clustering without node correspondence · Real networks | clustering error: 0.1 vs 0.32 | Helped |
| Use log-traces of adjacency-matrix powers instead of graphlet kernels | Network clustering without node correspondence · Real networks | time (s): 2.7 vs 45, 50, and 60 | Helped |
| Use log-traces of adjacency-matrix powers instead of global graph statistics | Network clustering without node correspondence · Real networks | time (s): 2.7 vs 765 | Helped |
| Tune graphlet-kernel clustering to use the top two eigenvectors and four clusters (ablation) | GK3, GK4, and GK5 · Simulated networks | clustering error: 0.5 -> 0.08 (GK3); 0.36 -> 0.025 (GK4); 0.26 -> 0.03 (GK5) | Helped |
| Increase the number of log moments from J = 5 to J = 8 (ablation) | NCLM · Simulated networks | clustering error | No clear effect |
| Use graphon estimates and spectral clustering instead of vectorized adjacency matrices | NCGE · Simulated networks with node correspondence | clustering accuracy | Helped |
| Use neighborhood smoothing rather than USVT for graphon estimation | NCGE · Smooth graphon simulation with node correspondence | clustering accuracy | Helped |
Closest papers by meaning
- Community Detection in Attributed Graphs: An Embedding Approach (AAAI 2018 · similarity 0.89)
- Incorporating Network Embedding into Markov Random Field for Better Community Detection (AAAI 2019 · similarity 0.88)
- Community Detection Guarantees using Embeddings Learned by Node2Vec (NeurIPS 2024 · similarity 0.86)
- Learning Persistent Community Structures in Dynamic Networks via Topological Data Analysis (AAAI 2024 · similarity 0.86)
- Local Graph Clustering with Noisy Labels (ICLR 2024 · similarity 0.85)
- Network Tight Community Detection (ICML 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.