Research map / Learning theory and algorithms / Sample complexity
A consistently adaptive trust-region method
Fadi Hamad, Oliver Hinder
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 |
|---|---|---|---|
| CAT vs Newton trust-region on linear dynamical systems | CAT · Linear dynamical systems, 60 instances | geo-mean #iter: 480.1 -> 308.1 | Improved |
| CAT vs Newton trust-region on matrix completion | CAT · Matrix completion (Ausgrid) | geo-mean #iter: 1000 -> 216.4 | Improved |
| CAT vs Newton trust-region on CUTEst | CAT · CUTEst, 67 instances | geo-mean #iter: 44.5 -> 41.5 | Improved |
| CAT vs Newton trust-region on CUTEst | CAT · CUTEst, 67 instances | #failures: 4 -> 3 | Improved |
| CAT vs ARC g-rule on CUTEst iterations | CAT · CUTEst, 67 instances | geo-mean #iter: 38.1 -> 41.5 | Hurt |
| CAT vs ARC g-rule on CUTEst gradient evaluations | CAT · CUTEst, 67 instances | geo-mean #gradient evals: 26.6 -> 44.4 | Hurt |
| CAT vs ARC g-rule on CUTEst failures | CAT · CUTEst, 67 instances | #failures: 1 -> 3 | Hurt |
| set theta=0 (remove theta/2 ||d|| ||grad|| term from rho-hat) vs default theta=0.1 (ablation) | CAT · CUTEst test set | fraction of problems solved vs iterations | No clear effect |
| set theta=0 instead of theta=0.1 on hard worst-case example, r1=1.5 (ablation) | CAT · Hard example from [45], eps=10^-3 | iterations to terminate | Hurt |
Closest papers by meaning
- Trust Region Methods for Nonconvex Stochastic Optimization beyond Lipschitz Smoothness (AAAI 2024 · similarity 0.88)
- Regional Expected Improvement for Efficient Trust Region Selection in High-Dimensional Bayesian Optimization (AAAI 2025 · similarity 0.84)
- Improving Convergence Guarantees of Random Subspace Second-order Algorithm for Nonconvex Optimization (ICLR 2025 · similarity 0.84)
- Trust Region Constrained Measure Transport in Path Space for Stochastic Optimal Control and Inference (NeurIPS 2025 · similarity 0.84)
- Generalized Smooth Variational Inequalities: Methods with Adaptive Stepsizes (ICML 2024 · similarity 0.84)
- Affine-Invariant Global Non-Asymptotic Convergence Analysis of BFGS under Self-Concordance (NeurIPS 2025 · similarity 0.83)
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.
