Research map / Learning theory and algorithms / Combinatorial and search optimization
Descending, lifting or smoothing: Secrets of robust cost optimization
Christopher Zach, Guillaume Bourmaud
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 |
|---|---|---|---|
| add graduated optimization (GOM+) on top of IRLS instead of plain IRLS | IRLS · Weak membrane energy, Lena | final objective: 231.8133±1.9040 -> 45.0463±3.66e-13 | Improved |
| use convexified Newton approximation instead of Gauss-Newton in lifting | lifted methods · Weak membrane energy, Lena | final objective: 45.0811±0.0861 -> 45.0496±0.0446 | Improved |
| use the relative stopping criterion (GOM+) instead of uniform iteration allocation (GOM) (ablation) | graduated optimization · Weak membrane energy, Lena | convergence speed | Improved |
| use sigmoid parametrization instead of w(u)=u^2 for lifting variables (ablation) | lifted Gauss-Newton and Newton · Robust mean, D=3, 100 runs | final objective | Improved |
| use lifted Newton instead of lifted Gauss-Newton | lifting · Robust mean, D=3 | final objective | Improved |
| initialize lifting variables to their optimal values (l-G-N*) instead of large values (ablation) | lifted Gauss-Newton · Linearized bundle adjustment | objective | Hurt |
| use GOM+ instead of lifted formulations | lifting-based methods · Linearized bundle adjustment (10 instances) | objective | Improved |
| use lifted Newton instead of Gauss-Newton for non-linear metric bundle adjustment | lifting · Metric bundle adjustment | objective | Hurt |
Closest papers by meaning
- Stochastic Gradient Methods for Distributionally Robust Optimization with f-divergences (NeurIPS 2016 · similarity 0.87)
- Second-order Optimization under Heavy-Tailed Noise: Hessian Clipping and Sample Complexity Limits (NeurIPS 2025 · similarity 0.87)
- Distributionally Robust Optimization with Bias and Variance Reduction (ICLR 2024 · similarity 0.86)
- Outlier-Robust Distributionally Robust Optimization via Unbalanced Optimal Transport (NeurIPS 2024 · similarity 0.86)
- Statistical Properties of Robust Satisficing (ICML 2024 · similarity 0.86)
- Drago: Primal-Dual Coupled Variance Reduction for Faster Distributionally Robust Optimization (NeurIPS 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.
