Research map / Optimization / Implicit bias and generalization
How to Escape Sharp Minima with Random Perturbations
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 Sharpness-Aware Perturbation instead of Randomly Smoothed Perturbation | flat-minimum finding algorithms · Training loss | iterations to find a flat minimum: d\epsilon^{-1} \min\{ 1, \epsilon d^3\} times larger than the decrease | Helped |
| use stochastic gradients for the outer perturbation instead of full-batch gradients (ablation) | Sharpness-Aware Perturbation · Training loss | trace of Hessian decrease: d times larger gradient | Helped |
| use batch size 1 instead of batch size 64 or 128 (ablation) | Sharpness-Aware Perturbation · CIFAR10 | escape from sharp minima | Helped |
| use Sharpness-Aware Perturbation instead of Randomly Smoothed Perturbation | ResNet-18 · CIFAR10 | escape from sharp minima | Helped |
Closest papers by meaning
- Zeroth-Order Optimization Finds Flat Minima (NeurIPS 2025 · similarity 0.89)
- Sharpness-Aware Minimization Efficiently Selects Flatter Minima Late In Training (ICLR 2025 · similarity 0.88)
- A Universal Class of Sharpness-Aware Minimization Algorithms (ICML 2024 · similarity 0.87)
- Improving Sharpness-Aware Minimization by Lookahead (ICML 2024 · similarity 0.87)
- Tilted Sharpness-Aware Minimization (ICML 2025 · similarity 0.87)
- Sharpness-Aware Minimization Enhances Feature Quality via Balanced Learning (ICLR 2024 · similarity 0.87)
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.