Research map / Graph learning / Graph algorithms and clustering
On Graph Reconstruction via Empirical Risk Minimization: Fast Learning Rates and Scalability
Guillaume Papa, Aurélien Bellet, Stephan Clémençon
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 |
|---|---|---|---|
| raise sampled pairs from B=0.01n to 0.1n (ablation) | ERM metric-learning graph reconstruction · Synthetic graph n=1,000,000, q=100, p=0.05 | reconstruction error: 0.2272 -> 0.1543 | Improved |
| raise B from 0.1n to n (ablation) | ERM metric-learning graph reconstruction · Synthetic graph n=1,000,000, q=100, p=0.05 | reconstruction error: 0.1543 -> 0.1276 | Improved |
| raise B from n to 5n (ablation) | ERM metric-learning graph reconstruction · Synthetic graph n=1,000,000, q=100, p=0.05 | reconstruction error: 0.1276 -> 0.1185 | Improved |
| raise B from 5n to 10n (ablation) | ERM metric-learning graph reconstruction · Synthetic graph n=1,000,000, q=100, p=0.05 | reconstruction error: 0.1185 -> 0.1159 | No clear effect |
| raise B from 0.01n to 0.1n (ablation) | ERM metric-learning graph reconstruction · Synthetic graph n=1,000,000, q=100, p=0.05 | training time (seconds): 21 -> 398 | Hurt |
| raise B from n to 5n (ablation) | ERM metric-learning graph reconstruction · Synthetic graph n=1,000,000, q=100, p=0.05 | training time (seconds): 5,705 -> 20,815 | Hurt |
| raise B from 5n to 10n (ablation) | ERM metric-learning graph reconstruction · Synthetic graph n=1,000,000, q=100, p=0.05 | training time (seconds): 20,815 -> 42,574 | Hurt |
| learn reconstruction rule from B=100 pairs instead of Euclidean distance with optimal threshold | ERM metric learning · Illustrative graph n=50, q=2 | reconstruction error: 0.1311 -> 0.009 | Improved |
Closest papers by meaning
- Transfer Learning on Edge Connecting Probability Estimation Under Graphon Model (NeurIPS 2025 · similarity 0.87)
- Harnessing Multiple Correlated Networks for Exact Community Recovery (NeurIPS 2024 · similarity 0.86)
- Fast Wasserstein rates for estimating probability distributions of probabilistic graphical models (NeurIPS 2026 · similarity 0.86)
- Learning Graph Representation via Graph Entropy Maximization (ICML 2024 · similarity 0.86)
- Temporal Generalization Estimation in Evolving Graphs (ICLR 2024 · similarity 0.86)
- Distribution Learning Meets Graph Structure Sampling (NeurIPS 2025 · 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.
