Research map / Optimization / Nonconvex and smooth optimization
A Block Coordinate Descent Method for Nonsmooth Composite Optimization under Orthogonality Constraints
Ganzhao Yuan
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 OBCD-R (random working-set block coordinate descent) instead of LADMM/RADMM/SPM | L0-regularized SPCA · 10 real/synthetic datasets (r=20, lambda 10-500) | objective value after 40s: w1a, lambda=100: 2199.819 -> 1999.667 | Improved |
| OBCD-R vs baselines on some datasets (e.g. lambda=10 and lambda=500 on Gisette, tied with others) | L0-regularized SPCA · Gisette-3000-1000 | objective value: 199.979 vs 199.980 (lambda=10) | No clear effect |
| use OBCD, a greedy monotone descent method, rather than methods with fluctuating objectives | L0-regularized SPCA · convergence curves (lambda=100) | objective trajectory | Improved |
Closest papers by meaning
- Block Acceleration Without Momentum: On Optimal Stepsizes of Block Gradient Descent for Least-Squares (ICML 2024 · similarity 0.89)
- Inertial Block Proximal Methods for Non-Convex Non-Smooth Optimization (ICML 2020 · similarity 0.87)
- Cyclic Block Coordinate Descent With Variance Reduction for Composite Nonconvex Optimization (ICML 2023 · similarity 0.87)
- On Faster Convergence of Cyclic Block Coordinate Descent-type Methods for Strongly Convex Minimization (JMLR 2018 · similarity 0.86)
- Enhanced Cyclic Coordinate Descent Methods for Elastic Net Penalized Linear Models (NeurIPS 2025 · similarity 0.85)
- Smoothing Proximal Gradient Methods for Nonsmooth Sparsity Constrained Optimization: Optimality Conditions and Global Convergence (ICML 2024 · 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.
