Research map / Optimization / Nonconvex and smooth optimization
Certificate-Guided Pruning for Stochastic Lipschitz Optimization
Ibne Farabi Shihab, Sanjeda Akter, Anuj Sharma
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 GP refinement within the active set when local smoothness ratio rho < 0.5 (ablation) | CGP · Branin (T=200) | simple regret (x1e-2): 2.0 -> 1.4 | Improved |
| add GP refinement (hybrid) (ablation) | CGP · Rosenbrock (T=200) | simple regret (x1e-2): 3.8 -> 3.4 | Improved |
| add GP refinement (hybrid) on rough function (ablation) | CGP · Needle-2D (T=200) | simple regret (x1e-2): 1.2 -> 1.1 | No clear effect |
| replace GP-UCB with CGP | GP-UCB · Needle-2D (non-differentiable) | simple regret (x1e-2): 2.1 -> 1.2 | Improved |
| replace TuRBO with CGP-TR trust regions plus local certificates | TuRBO · Rover-60 (T=500) | regret: 12.4 -> 11.2 | Improved |
| use CGP-TR instead of vanilla CGP for high dimension | CGP · d>20 benchmarks | tractability | Improved |
| initialize CGP-Adaptive with L underestimated by 100x (ablation) | CGP-Adaptive · Hartmann-6 | simple regret (x1e-2): 2.9 -> 3.2 | Hurt |
| learn L online via doubling instead of LIPO | LIPO · Hartmann-6 | simple regret (x1e-2): 6.2 -> 3.2 | Improved |
| remove pruning certificates (ablation) | CGP · ablation (Table 8, not shown in text) | regret: +78% | Hurt |
| remove coverage penalty (ablation) | CGP · ablation (Table 8, not shown in text) | regret: +44% | Hurt |
| remove replication (ablation) | CGP · ablation (Table 8, not shown in text) | regret: +33% | Hurt |
| use certified eps_t stopping rule | CGP · stopping rule (Table 9) | samples saved: 33% saved, no regret penalty | Improved |
Closest papers by meaning
- Optimizing $(L_0, L_1)$-Smooth Functions by Gradient Methods (ICLR 2025 · similarity 0.86)
- Adaptive Proximal Gradient Methods Are Universal Without Approximation (ICML 2024 · similarity 0.86)
- Decentralized Gradient-Free Methods for Stochastic Non-smooth Non-convex Optimization (AAAI 2024 · similarity 0.86)
- Finding Simple Proofs for First-Order Optimization (NeurIPS 2026 · similarity 0.86)
- Non-Asymptotic and Non-Lipschitzian Bounds on Optimal Values in Stochastic Optimization Under Heavy Tails (ICML 2025 · similarity 0.85)
- Improved Sample Complexity for Private Nonsmooth Nonconvex Optimization (ICML 2025 · 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.
