Research map / Optimization / Nonconvex and smooth optimization
Stochastic Composite Mirror Descent: Optimal Bounds with High Probabilities
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 square-summable step sizes (theta=0.51 or 0.75) rather than theta=0.25 for multi-pass SGD (ablation) | SGD with hinge loss, linear kernel · LIBSVM binary classification (ADULT, IJCNN, PHISHING, SPLICE etc.) | test error vs passes | Helped |
| step size exponent 0.51 versus 0.75 (ablation) | multi-pass SGD · ADULT, IJCNN, PHISHING, SPLICE | test error | No clear effect |
Closest papers by meaning
- On the Generalization Error of Stochastic Mirror Descent for Quadratically-Bounded Losses: an Improved Analysis (NeurIPS 2023 · similarity 0.92)
- The Statistical Complexity of Early-Stopped Mirror Descent (NeurIPS 2020 · similarity 0.91)
- Stochastic Mirror Descent in Variationally Coherent Optimization Problems (NeurIPS 2017 · similarity 0.91)
- Rapid Overfitting of Multi-Pass SGD in Stochastic Convex Optimization (ICML 2025 · similarity 0.89)
- The Sample Complexity of Gradient Descent in Stochastic Convex Optimization (NeurIPS 2024 · similarity 0.89)
- High-probability complexity bounds for stochastic non-convex minimax optimization (NeurIPS 2024 · 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.