Research map / Graph learning / Graph algorithms and clustering
Reconstructing Undirected Graphs from Eigenspaces
Yohann De Castro, Thibault Espinasse, Paul Rochet
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 bagging over half-samples to the backward algorithm (ablation) | Backward algorithm · Kite graph N=5, n=500 | exact support recovery: 23% -> 61% | Improved |
| add bagging over half-samples to the backward algorithm (ablation) | Backward algorithm · Kite graph N=5, n=500 | mean support error: 1.95 -> 0.68 | Improved |
| add bagging over half-samples to the backward algorithm (ablation) | Backward algorithm · Graph G2 N=15, n=10000 | exact support recovery: 26% -> 69% | Improved |
| add bagging over half-samples to the backward algorithm (ablation) | Backward algorithm · Graph G2 N=15, n=10000 | mean support error: 25 -> 1 | Improved |
| add bagging over half-samples to the backward algorithm (ablation) | Backward algorithm · Graph G2 N=15, n=10000 | CPU time (s): 2.5 -> 256 | Hurt |
| use bagging backward instead of l2 thresholding | l2 thresholding · Graph G2 N=15, n=10000 | exact support recovery: 22% -> 69% | Improved |
| use oracle-calibrated l2 thresholding instead of bagging backward | Bagging backward algorithm · Kite graph N=5, n=500 | exact support recovery: 61% -> 75% | Mixed |
| change link function from exp(t) to (1-t)^-2 (more spread spectrum) | Bagging backward algorithm · Graph G2 N=15, n=10000 | exact support recovery: 69% -> 97% | Improved |
| decrease sample size from 10000 to 1000 (ablation) | Bagging backward algorithm · Graph G2, K=(1-W)^-2 | exact support recovery: 97% -> 13% | Hurt |
| use bagging backward instead of GGMselect (default parameters) | GGMselect · France lightning | computation time and robustness: 0.3 s -> 400 s | Mixed |
Closest papers by meaning
- Exact Community Recovery under Side Information: Optimality of Spectral Algorithms (ICLR 2025 · similarity 0.85)
- Weisfeiler-Leman Is Incomplete on Simple Spectrum Graphs, so Canonicalize Them (NeurIPS 2026 · similarity 0.85)
- Structure-Aware Spectral Sparsification via Uniform Edge Sampling (NeurIPS 2025 · similarity 0.85)
- Nonlinear Laplacians: Tunable principal component analysis under directional prior information (NeurIPS 2025 · similarity 0.85)
- Fast exact recovery of noisy matrix from few entries: the infinity norm approach (NeurIPS 2025 · similarity 0.84)
- On the Robustness of Spectral Algorithms for Semirandom Stochastic Block Models (NeurIPS 2024 · similarity 0.84)
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.
