Research map / Optimization / Implicit bias and generalization
Phase diagram of Stochastic Gradient Descent in high-dimensional two-layer neural networks
Rodrigo Veiga, Ludovic STEPHAN, Bruno Loureiro, Florent Krzakala, Lenka Zdeborova
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 |
|---|---|---|---|
| scale learning rate down with input dimension (gamma ~ d^-delta, delta>0) at fixed width | one-pass SGD, two-layer net · noisy teacher-student, kappa=0 | asymptotic population risk: slope -0.494 for delta=1/2 | Improved |
| grow hidden width with d so that kappa+delta>0 | one-pass SGD, two-layer net · teacher-student simulation d=100 | population risk | Improved |
| increase learning rate with d so kappa+delta<0 (delta<0) | one-pass SGD, two-layer net · teacher-student simulation | population risk | Hurt |
Closest papers by meaning
- High-dimensional limit theorems for SGD: Effective dynamics and critical scaling (NeurIPS 2022 · similarity 0.92)
- Emergence and scaling laws in SGD learning of shallow neural networks (NeurIPS 2025 · similarity 0.91)
- High-dimensional Limit of SGD for Diagonal Linear Networks (NeurIPS 2026 · similarity 0.90)
- The Global Convergence Time of Stochastic Gradient Descent in Non-Convex Landscapes: Sharp Estimates via Large Deviations (ICML 2025 · similarity 0.89)
- Learning quadratic neural networks in high dimensions: SGD dynamics and scaling laws (NeurIPS 2025 · similarity 0.89)
- Statistical Guarantees for High-Dimensional Stochastic Gradient Descent (NeurIPS 2025 · 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.
