Research map / Learning theory and algorithms / Matrix and tensor methods
An Alternating Proximal Splitting Method with Global Convergence for Nonconvex Structured Sparsity 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 |
|---|---|---|---|
| replace convex l1 with nonconvex l0 regularizer (ablation) | graph-guided fused logistic regression (AFBS) · 20newsgroup rec. vs rest | accuracy (%): 86.34(±0.017) -> 90.07(±0.009) | Helped |
| replace convex l1 with nonconvex l0 regularizer (ablation) | graph-guided fused logistic regression (AFBS) · 20newsgroup sci. vs rest | accuracy (%): 79.53(±0.016) -> 85.58(±0.005) | Helped |
| replace convex l1 with capped-l1 regularizer (ablation) | graph-guided fused logistic regression (AFBS) · 20newsgroup com. vs rest | accuracy (%): 82.32(±0.013) -> 84.83(±0.014) | Helped |
| replace capped-l1 with l0 regularizer (ablation) | graph-guided fused logistic regression (AFBS) · 20newsgroup com. vs rest | accuracy (%): 84.83(±0.014) -> 84.93(±0.013) | No clear effect |
| use nonconvex capped-l1 / l0 instead of l1 in overlapping group lasso (ablation) | overlapping group lasso (AFBS) · simulated sparse recovery, n=500, K=60 | variable and group selection error | Helped |
| use AFBS instead of CCCP with SLEP inner solver for capped-l1 group lasso | capped-l1 overlapping group lasso · simulated, n=10000, K=300 | running time and final objective | Helped |
| use AFBS/AFBS_ACC instead of Picard, SLEP, ADMM on convex group lasso | overlapping group lasso (l1) · simulated, n=4000, K=120 | running time | Helped |
| use AFBS instead of Picard/SLEP | convex overlapping group lasso · simulated, n=4000, K=120 | iterations to converge | Hurt |
| add Nesterov-style acceleration to AFBS (ablation) | AFBS · simulated overlapping group lasso | convergence speed | Helped |
Closest papers by meaning
- On Quadratic Convergence of DC Proximal Newton Algorithm in Nonconvex Sparse Learning (NeurIPS 2017 · similarity 0.90)
- Efficient Learning with a Family of Nonconvex Regularizers by Redistributing Nonconvexity (ICML 2016 · similarity 0.89)
- SequentialAttention++ for Block Sparsification: Differentiable Pruning Meets Combinatorial Optimization (NeurIPS 2024 · similarity 0.88)
- Differentiable Sparsity via $D$-Gating: Simple and Versatile Structured Penalization (NeurIPS 2025 · similarity 0.87)
- Revisiting Frank-Wolfe for Structured Nonconvex Optimization (NeurIPS 2025 · similarity 0.86)
- Dynamic Sparse Training with Structured Sparsity (ICLR 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.