Research map / Optimization / Implicit bias and generalization
Gradient Descent with Large Step Sizes: Chaos and Fractal Convergence Region
Shuang Liang, Guido Montufar
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 |
|---|---|---|---|
| Add l2 regularization to scalar factorization loss (ablation) | Gradient descent with near-critical step size · Scalar factorization convergence region | boundary regularity: smooth boundary -> fractal boundary | Hurt |
| Use near-critical step sizes instead of small step sizes | Gradient descent · Depth-3 ReLU network on 2-class CIFAR-10 subset (MSE) | predictability of final norm and sharpness | Hurt |
| Use step sizes in the edge-of-stability regime | Gradient descent · Depth-3 ReLU network on CIFAR-10 subset | final sharpness: sharpness near 2/eta | No clear effect |
| Add Polyak momentum | Gradient descent · Depth-3 ReLU network on CIFAR-10 subset | final sharpness | Mixed |
| Add weight decay (ablation) | Gradient descent · Depth-3 ReLU network on CIFAR-10 subset | fractal convergence structure | No clear effect |
| Use cross-entropy loss instead of MSE (ablation) | Gradient descent · Depth-3 ReLU network on CIFAR-10 subset | two-regime behavior | No clear effect |
| Initialize in orthogonal subspace W (incl. identity init) | Gradient descent matrix factorization · General matrix factorization | chaos and fractal boundary | No clear effect |
Closest papers by meaning
- Towards Resolving the Implicit Bias of Gradient Descent for Matrix Factorization: Greedy Low-Rank Learning (ICLR 2021 · similarity 0.90)
- Preconditioning Matters: Fast Global Convergence of Non-convex Matrix Factorization via Scaled Gradient Descent (NeurIPS 2023 · similarity 0.90)
- On the Crucial Role of Initialization for Matrix Factorization (ICLR 2025 · similarity 0.89)
- Provable Acceleration of Nesterov's Accelerated Gradient for Asymmetric Matrix Factorization and Linear Neural Networks (NeurIPS 2024 · similarity 0.89)
- Deep Weight Factorization: Sparse Learning Through the Lens of Artificial Symmetries (ICLR 2025 · similarity 0.87)
- In-depth Analysis of Low-rank Matrix Factorisation in a Federated Setting (AAAI 2025 · 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.
