Research map / Optimization / Nonconvex and smooth optimization
Gradient-Normalized Smoothness for Optimization with Approximate Hessians
Andrei Semenov, Martin Jaggi, Nikita Doikov
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 |
|---|---|---|---|
| use approximate Weighted Gauss-Newton Hessian instead of exact Hessian (ablation) | Gradient-regularized Newton · LogSumExp (libsvm datasets) | convergence speed | No clear effect |
| use second-order method instead of Gradient Method | Gradient-regularized Newton with inexact Hessian · LogSumExp | convergence speed | Improved |
| use function-value adaptive search instead of gradient-norm search (ablation) | Gradient-regularized Newton · LogSumExp | step-size gamma_k | Improved |
| add Gauss-Newton preconditioner to Gradient Method (ablation) | Gradient Method · Nonlinear Equations (linear operator, p=2..5) | convergence speed | Improved |
| use Fisher-type rank-one Hessian approximation instead of exact Newton (ablation) | Gradient-regularized Newton · Nonlinear Equations (MNIST, synthetic) | wall-clock time | Improved |
| use inexact Hessian instead of exact Hessian (ablation) | Gradient-regularized Newton · Rosenbrock Nonlinear Equations | robustness to starting point | Improved |
| use Gradient Method with our adaptive search at large p (ablation) | Gradient Method · Nonlinear Equations (linear operator) | convergence speed | Hurt |
Closest papers by meaning
- Efficiently Escaping Saddle Points under Generalized Smoothness via Self-Bounding Regularity (NeurIPS 2025 · similarity 0.89)
- Skip the Hessian, Keep the Rates: Globalized Semismooth Newton with Lazy Hessian Updates (NeurIPS 2026 · similarity 0.89)
- Directional Smoothness and Gradient Methods: Convergence and Adaptivity (NeurIPS 2024 · similarity 0.89)
- Methods for Convex $(L_0,L_1)$-Smooth Optimization: Clipping, Acceleration, and Adaptivity (ICLR 2025 · similarity 0.88)
- A Regularized Newton Method for Nonconvex Optimization with Global and Local Complexity Guarantees (NeurIPS 2025 · similarity 0.88)
- Optimizing $(L_0, L_1)$-Smooth Functions by Gradient Methods (ICLR 2025 · similarity 0.88)
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.
