Research map / Learning theory and algorithms / Matrix and tensor methods
Speeding Up Latent Variable Gaussian Graphical Model Estimation via Nonconvex Optimization
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 matrix-factorized low-rank component and alternating gradient descent with hard thresholding instead of convex nuclear-norm optimization | LVGGM estimation · Synthetic LVGGM, d=5000, r=10, n=2 × 10^5 | Time (s): 21090.7900 -> 445.6730 | Helped |
| Use matrix-factorized low-rank component and alternating gradient descent with hard thresholding instead of convex nuclear-norm optimization | LVGGM estimation · Synthetic LVGGM, d=5000, r=10, n=2 × 10^5 | Precision Frobenius error: 1.5012±0.0240 -> 1.3449±0.0084 | Helped |
| Use matrix-factorized low-rank component and alternating gradient descent with hard thresholding instead of convex nuclear-norm optimization | Sparse plus low-rank Gaussian precision estimation · Synthetic sparse plus low-rank, d=5000, r=10, n=2 × 10^5 | Time (s): 20393.3650 -> 287.8630 | Helped |
| Use matrix-factorized low-rank component and alternating gradient descent with hard thresholding instead of convex nuclear-norm optimization | Sparse plus low-rank Gaussian precision estimation · Synthetic sparse plus low-rank, d=5000, r=10, n=2 × 10^5 | Precision Frobenius error: 1.5324±0.0085 -> 1.2068±0.0032 | Helped |
| Use AltGD rather than the fastest reported baseline | LVGGM estimation on luminal breast cancer data · Breast cancer gene expression, luminal subtype | Time (s): 7.6700 -> 0.1500 | Helped |
| Include latent-variable structure rather than fitting only a sparse precision matrix | Regulatory network recovery · Breast cancer luminal subtype TF network | Correctly recovered edges | Helped |
Closest papers by meaning
- A Completely Tuning-Free and Robust Approach to Sparse Precision Matrix Estimation (ICML 2022 · similarity 0.90)
- A Fast and Scalable Joint Estimator for Integrating Additional Knowledge in Learning Multiple Related Sparse Gaussian Graphical Models (ICML 2018 · similarity 0.90)
- Graphical Nonconvex Optimization via an Adaptive Convex Relaxation (ICML 2018 · similarity 0.89)
- Conditional Matrix Flows for Gaussian Graphical Models (NeurIPS 2023 · similarity 0.89)
- Learning Local Neighborhoods of Non-Gaussian Graphical Models (AAAI 2025 · similarity 0.87)
- Learning Latent Variable Models via Jarzynski-adjusted Langevin Algorithm (NeurIPS 2025 · 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.