Research map / Optimization / Nonconvex and smooth optimization
Stochastic PCA with $\ell_2$ and $\ell_1$ Regularization
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 |
|---|---|---|---|
| Add l2 regularization (strongly convex objective, step 1/(lambda t)) to MSG | MSG convex relaxation of PCA · Synthetic and MNIST | suboptimality vs number of samples | Helped |
| Add l1 (trace) regularization that shifts spectrum by mu*eta_t I | MSG · Synthetic and MNIST | rank of iterates | Helped |
| Add l1 regularization to MSG | MSG · Synthetic and MNIST | overall CPU runtime to reach suboptimality | Helped |
| Add elastic-net (l1+l2) regularization to MSG | MSG · Synthetic and MNIST | overall CPU runtime and iteration complexity | Helped |
| Use Oja's algorithm instead of regularized MSG variants | Oja baseline · Synthetic and MNIST | suboptimality vs iterations and runtime | Hurt |
| Choose regularization parameters within admissible range (lambda < gap, mu small) | l1 and l2 regularized PCA · Theory | optimum preserved | No clear effect |
Closest papers by meaning
- Efficient Convex Relaxations for Streaming PCA (NeurIPS 2019 · similarity 0.95)
- Local Linear Convergence of Gradient Methods for Subspace Optimization via Strict Complementarity (NeurIPS 2022 · similarity 0.93)
- Convergence of Stochastic Gradient Descent for PCA (ICML 2016 · similarity 0.92)
- Rethinking PCA Through Duality (NeurIPS 2025 · similarity 0.89)
- Efficient Sparse PCA via Block-Diagonalization (ICLR 2025 · similarity 0.89)
- Fast and Provable Algorithms for Sparse PCA with Improved Sample Complexity (ICML 2025 · similarity 0.87)
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.