Research map / Optimization / Nonconvex and smooth optimization
Parametric Dual Maximization for Non-Convex Learning Problems
Yuxun Zhou, Zhaoyi Kang, Costas Spanos
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 |
|---|---|---|---|
| increase approximation degree m from 0.1 to 0.2 length(p) (ablation) | PDM · 2moons S3VM | test error rate (%): 1.9 -> 0.2 | Improved |
| increase approximation degree m from 0.1 to 0.2 length(p) (ablation) | PDM · coil S3VM | test error rate (%): 5.3 -> 1.1 | Improved |
| increase approximation degree m from 0.1 to 0.2 length(p) (ablation) | PDM · robot S3VM | test error rate (%): 9.5 -> 3.3 | Improved |
| increase approximation degree m from 0.1 to 0.2 length(p) (ablation) | PDM · Wave LSVM | test error rate (%): 19.9 -> 13.1 | Improved |
| increase approximation degree m from 0.1 to 0.2 length(p) (ablation) | PDM · 2moons S3VM | normalized objective value: 1.03 -> 1.00 | Improved |
| increase approximation degree m from 0.1 to 0.2 length(p) (ablation) | PDM · Wave LSVM | normalized objective value: 1.54 -> 1.00 | Improved |
| increase approximation degree m from 0.1 to 0.2 length(p) (ablation) | PDM · Wave LSVM | time: 2.5h -> 4.8h | Hurt |
| increase m beyond about 300 (range 80 to 650) (ablation) | PDM · robot S3VM | objective and testing accuracy | No clear effect |
| use PDM instead of local optimizers (GD, CCCP, AO, LCS) | S3VM/LSVM training · 2moons S3VM | test error rate (%): 51.4 -> 1.9 | Improved |
| use PDM instead of local optimizers | LSVM training · Wave LSVM | test error rate (%): 36.7 -> 13.1 | Improved |
| use PDM instead of local optimizers | LSVM training · Wave LSVM | normalized objective value: 10.2 -> 1.54 | Improved |
| use PDM instead of exact branch and bound | S3VM · robot S3VM | feasibility | Improved |
Closest papers by meaning
- Near-Optimal Solutions of Constrained Learning Problems (ICLR 2024 · similarity 0.89)
- Uniform Convergence of Gradients for Non-Convex Learning and Optimization (NeurIPS 2018 · similarity 0.87)
- Coordinate Descent Methods for DC Minimization: Optimality Conditions and Global Convergence (AAAI 2023 · similarity 0.87)
- Semi-infinite Nonconvex Constrained Min-Max Optimization (NeurIPS 2025 · similarity 0.87)
- A Generalization Result for Convergence in Learning-to-Optimize (ICML 2025 · similarity 0.86)
- From Linear to Linearizable Optimization: A Novel Framework with Applications to Stationary and Non-stationary DR-submodular Optimization (NeurIPS 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.
