Research map / Optimization / Nonconvex and smooth optimization
Stochastic Parallel Block Coordinate Descent for Large-Scale Saddle Point Problems
Zhanxing Zhu, Amos Storkey
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 SP-BCD with K=2 instead of PDMM2 | RPCA · RPCA | Time (s): 750 -> 492 | Improved |
| Use SP-BCD with K=2 instead of PDMM2 | RPCA · RPCA | Frobenius norm of residual (10^-4): 4.55 -> 6.17 | Hurt |
| Use SP-BCD with K=2 instead of GSADMM | RPCA · RPCA | Time (s): 448 -> 492 | Mixed |
| Use SP-BCD instead of ADMM | Lasso · Lasso (1000,5000,500) | Time (s): 0.69 -> 0.70 | No clear effect |
| Use SP-BCD instead of FISTA | Lasso · Lasso (5000,20000,2000) | Time (s): 19.00 -> 13.32 | Improved |
| Use SP-BCD instead of ADMM | Lasso · Lasso (5000,20000,2000) | Time (s): 19.83 -> 13.32 | Improved |
| Use SP-BCD instead of SPDC | Lasso · Lasso | Objective convergence | Improved |
| Use matrix-structure-scaled primal and dual steps instead of PDCP's constant stepsize | SP-BCD · Group Lasso (MEMset Donor) | Objective convergence | Improved |
| Use fewer selected blocks per iteration (ablation) | SP-BCD · Group Lasso (MEMset Donor) | Convergence speed | Improved |
| Use SP-BCD with K=2 instead of K=1 (ablation) | SP-BCD · RPCA | Time (s): 784 -> 492 | Improved |
| Use SP-BCD with K=2 instead of K=3 (ablation) | SP-BCD · RPCA | Time (s): 553 -> 492 | Improved |
Closest papers by meaning
- High Probability and Risk-Averse Guarantees for a Stochastic Accelerated Primal-Dual Method (JMLR 2024 · similarity 0.87)
- Alternating Randomized Block Coordinate Descent (ICML 2018 · similarity 0.87)
- Stochastic Variance Reduction Methods for Saddle-Point Problems (NeurIPS 2016 · similarity 0.87)
- Projection-Free Methods for Solving Nonconvex-Concave Saddle Point Problems (NeurIPS 2023 · similarity 0.87)
- On Linear Convergence in Smooth Convex-Concave Bilinearly-Coupled Saddle-Point Optimization: Lower Bounds and Optimal Algorithms (ICML 2025 · similarity 0.86)
- Dealing With Unbounded Gradients in Stochastic Saddle-point Optimization (ICML 2024 · similarity 0.86)
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.
