Research map / Optimization / Implicit bias and generalization
Learning with SGD and Random Features
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 |
|---|---|---|---|
| Increase random Fourier features beyond a threshold of order √n | SGD with random features · SUSY and HIGGS test sets | classification error | No clear effect |
| Increase the mini-batch size and increase the step-size accordingly | SGD with random features · SUSY and HIGGS test sets, one pass | classification error | Helped |
| Use a mini-batch larger than order √n while keeping training to one pass | SGD with random features · SUSY and HIGGS test sets | classification error | Hurt |
| Use decreasing step-sizes instead of the analyzed constant-step regimes | SGD with random features · Nonparametric least-squares learning | excess risk | No clear effect |
Closest papers by meaning
- Optimal Learning for Multi-pass Stochastic Gradient Methods (NeurIPS 2016 · similarity 0.88)
- On Large-Batch Training for Deep Learning: Generalization Gap and Sharp Minima (ICLR 2017 · similarity 0.88)
- The Power of Random Features and the Limits of Distribution-Free Gradient Descent (ICML 2025 · similarity 0.87)
- Adaptive Sampling Probabilities for Non-Smooth Optimization (ICML 2017 · similarity 0.87)
- Can Gaussian Sketching Converge Faster on a Preconditioned Landscape? (ICML 2024 · similarity 0.86)
- Statistical Guarantees for High-Dimensional Stochastic Gradient Descent (NeurIPS 2025 · 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.