Research map / Learning theory and algorithms / Clustering
Spectral Embedding of Regularized Block Models
Nathan De Lara, Thomas Bonald
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 |
|---|---|---|---|
| Add complete-graph regularization (α=1) instead of using α=0 (ablation) | Spectral embedding + K-Means · SBM, K=K_truth | homogeneity: 0.19 -> 0.36 | Improved |
| Add complete-graph regularization (α=1) instead of using α=0 (ablation) | Spectral embedding + K-Means · 20newsgroup, K=K_truth | homogeneity: 0.40 -> 0.46 | Improved |
| Add complete-graph regularization (α=0.1) instead of using α=0 (ablation) | Spectral embedding + K-Means · Wikipedia for Schools, K=K_truth | homogeneity: 0.23 -> 0.26 | Improved |
| Increase regularization from α=1 to α=10 (ablation) | Spectral embedding + K-Means · 20newsgroup, K=K_truth | homogeneity: 0.46 -> 0.37 | Hurt |
| Increase regularization from α=0.1 to α=10 (ablation) | Spectral embedding + K-Means · Wikipedia for Schools, K=K_truth | homogeneity: 0.26 -> 0.19 | Hurt |
| Add 1% isolated nodes with self loops without regularization (ablation) | Spectral embedding + K-Means · Wikipedia for Schools | homogeneity: 0.23 -> 0.00 | Hurt |
| Add 1% isolated nodes with self loops while using α=1 (ablation) | Spectral embedding + K-Means · Wikipedia for Schools | homogeneity: 0.23 -> 0.24 | No clear effect |
| Add 5% isolated nodes with self loops without regularization (ablation) | Spectral embedding + K-Means · Wikipedia for Schools | homogeneity: 0.23 -> 0.00 | Hurt |
| Add 5% isolated nodes with self loops while using α=1 (ablation) | Spectral embedding + K-Means · Wikipedia for Schools | homogeneity: 0.23 -> 0.23 | No clear effect |
| Add 10% isolated nodes with self loops without regularization (ablation) | Spectral embedding + K-Means · Wikipedia for Schools | homogeneity: 0.23 -> 0.00 | Hurt |
| Add 10% isolated nodes with self loops while using α=1 (ablation) | Spectral embedding + K-Means · Wikipedia for Schools | homogeneity: 0.23 -> 0.24 | No clear effect |
| Regularize the biadjacency matrix instead of the full adjacency matrix (ablation) | Spectral embedding + K-Means · 20newsgroup, K=K_truth/2, α=1 | homogeneity: 0.38 -> 0.41 | Improved |
Closest papers by meaning
- Convergence Guarantees for the DeepWalk Embedding on Block Models (ICML 2024 · similarity 0.88)
- Structure-Aware Spectral Sparsification via Uniform Edge Sampling (NeurIPS 2025 · similarity 0.87)
- Optimal Graph Clustering without Edge Density Signals (NeurIPS 2025 · similarity 0.87)
- On the Robustness of Spectral Algorithms for Semirandom Stochastic Block Models (NeurIPS 2024 · similarity 0.86)
- Pass-Efficient Algorithms for Graph Spectral Clustering (Student Abstract) (AAAI 2024 · similarity 0.86)
- Community Detection Guarantees using Embeddings Learned by Node2Vec (NeurIPS 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.
