Research map / Optimization / Nonconvex and smooth optimization
Efficient Learning with a Family of Nonconvex Regularizers by Redistributing Nonconvexity
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 the original nonconvex regularizer optimization with N2C convexification | Sparse group lasso · Synthetic sparse group lasso | CPU time (sec): 0.65±0.06 -> 0.48±0.05 | Helped |
| Use a nonconvex regularizer instead of the convex FISTA regularizer | Sparse group lasso · Synthetic sparse group lasso | RMSE: 53.8±1.7 -> 50.6±2.0 | Helped |
| Use a nonconvex regularizer instead of the convex FISTA regularizer | Sparse group lasso · Synthetic sparse group lasso | ABS: 10.6±0.3 -> 5.7±0.2 | Helped |
| Replace N2C with the approximate proximal-average approach GD-PAN | Sparse group lasso · Synthetic sparse group lasso | RMSE: 50.6±2.0 -> 52.3±2.0 | Hurt |
| Replace N2C with convex FISTA | Tree-structured group lasso · JAFFE anger vs not-anger | testing accuracy (%): 97.2±1.8 -> 99.6±0.9 | Hurt |
| Use N2C rather than convex FISTA | Tree-structured group lasso · JAFFE anger vs not-anger | sparsity (%): 9.2±0.2 -> 5.1±0.2 | Helped |
| Use N2C rather than FISTA | Tree-structured group lasso · JAFFE anger vs not-anger | CPU time (sec): 1.0±0.4 -> 1.9±0.3 | Hurt |
| Use N2C-FW rather than FaNCL | Nonconvex low-rank matrix completion · MovieLens 100K | CPU time (sec): 0.4±0.1 -> 0.2±0.1 | Helped |
| Use N2C-FW rather than FaNCL | Nonconvex low-rank matrix completion · MovieLens 1M | CPU time (sec): 16.6±0.6 -> 9.3±0.1 | Helped |
| Use N2C-FW rather than FaNCL | Nonconvex low-rank matrix completion · MovieLens 10M | CPU time (sec): 615.7±13.2 -> 313.0±6.6 | Helped |
| Use N2C-FW rather than FaNCL | Nonconvex low-rank matrix completion · MovieLens 100K | testing RMSE: 0.857±0.003 -> 0.855±0.004 | Helped |
| Use N2C-FW rather than FaNCL | Nonconvex low-rank matrix completion · MovieLens 1M | testing RMSE: 0.786±0.001 -> 0.785±0.001 | No clear effect |
Closest papers by meaning
- A Regularized Newton Method for Nonconvex Optimization with Global and Local Complexity Guarantees (NeurIPS 2025 · similarity 0.87)
- General framework for online-to-nonconvex conversion: Schedule-free SGD is also effective for nonconvex optimization (ICML 2025 · similarity 0.87)
- Weakly Convex Regularisers for Inverse Problems: Convergence of Critical Points and Primal-Dual Optimisation (ICML 2024 · similarity 0.87)
- Sketching for Convex and Nonconvex Regularized Least Squares with Sharp Guarantees (ICLR 2025 · similarity 0.86)
- Spectral Preconditioning for Gradient Methods on Graded Non-convex Functions (ICML 2024 · similarity 0.86)
- Cauchy-Schwarz Regularizers (ICLR 2025 · 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.