Research map / Optimization / Nonconvex and smooth optimization
Frank-Wolfe-based Algorithms for Approximating Tyler's M-estimator
Lior Danon, Dan Garber
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 |
|---|---|---|---|
| Use GAFW instead of fixed-point iterations | Tyler's M-estimator solver · Tyler M-estimator, contaminated Gaussian and t-dist (p=50, n=2500) | spectral-norm distance and f(Q_t)-f(Q*) vs normalized runtime | Improved |
| Use GAFW instead of FPI under outlier contamination | Tyler's M-estimator solver · Gaussian with outliers | convergence speed | Improved |
| Use GAFW instead of FW (ablation) | Frank-Wolfe variants · Tyler M-estimator, contaminated Gaussian and t-dist (p=50, n=2500) | convergence speed | Improved |
| Use GAFW instead of AFW (ablation) | Frank-Wolfe variants · Tyler M-estimator, contaminated Gaussian and t-dist (p=50, n=2500) | convergence speed | Improved |
| Add away-steps (AFW versus FW) (ablation) | Frank-Wolfe · Tyler M-estimator, contaminated Gaussian and t-dist (p=50, n=2500) | convergence speed | Hurt |
| Replace FPI by FW variants for per-iteration cost | Tyler's estimator · n=p^2 | per-iteration runtime | Improved |
Closest papers by meaning
- Robust Gaussian Covariance Estimation in Nearly-Matrix Multiplication Time (NeurIPS 2020 · similarity 0.86)
- Loss Minimization and Parameter Estimation with Heavy Tails (JMLR 2016 · similarity 0.85)
- A Spectral Algorithm for List-Decodable Covariance Estimation in Relative Frobenius Norm (NeurIPS 2023 · similarity 0.84)
- TIC-TAC: A Framework For Improved Covariance Estimation In Deep Heteroscedastic Regression (ICML 2024 · similarity 0.84)
- Sub-Gaussian Confidence Intervals for Heavy-Tailed Data: Characterizing the Limits of Inference (NeurIPS 2026 · similarity 0.83)
- Robust Sparse Regression with Non-Isotropic Designs (NeurIPS 2024 · similarity 0.83)
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.
