Research map / Learning theory and algorithms / Sample complexity
Sharpness, Restart and Acceleration
Vincent Roulet, Alexandre d'Aspremont
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 |
|---|---|---|---|
| restart schemes (Adap, Crit, Mono) instead of plain accelerated gradient | Accelerated gradient with restart · Least squares on UCI Sonar | f(x)-f* | Improved |
| restart schemes instead of plain accelerated gradient | Accelerated gradient with restart · Logistic loss on UCI Sonar | f(x)-f* | No clear effect |
| scheduled restart instead of restart heuristics | Accelerated gradient with restart · Dual SVM (hinge loss, regularized) on Sonar | f(x)-f* | Improved |
| restart schemes instead of plain accelerated gradient | Accelerated gradient with restart · LASSO | f(x)-f* | Improved |
| adaptive restart (Adap) instead of critical-parameter restart (Crit) | Accelerated gradient with restart · Dual SVM and LASSO | convergence consistency | Improved |
Closest papers by meaning
- Step-Size Stability in Stochastic Optimization: A Theoretical Perspective (ICML 2026 · similarity 0.87)
- Adaptive and Optimal Second-order Optimistic Methods for Minimax Optimization (NeurIPS 2024 · similarity 0.86)
- Stability and Sharper Risk Bounds with Convergence Rate $\tilde{O}(1/n^2)$ (NeurIPS 2025 · similarity 0.86)
- Fundamental Benefit of Alternating Updates in Minimax Optimization (ICML 2024 · similarity 0.85)
- Optimal Acceleration for Minimax and Fixed-Point Problems is Not Unique (ICML 2024 · similarity 0.85)
- Directional Smoothness and Gradient Methods: Convergence and Adaptivity (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.
