Research map / Optimization / Implicit bias and generalization
Global Optimality Beyond Two Layers: Training Deep ReLU Networks via Convex Programs
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 |
|---|---|---|---|
| Train with the proposed convex program instead of SGD | Three-layer ReLU network · acute-inflamation | training objective: 0.0029 -> 0.0013 | Helped |
| Train with the proposed convex program instead of SGD | Three-layer ReLU network · acute-nephritis | training objective: 0.0039 -> 0.0021 | Helped |
| Train with the proposed convex program instead of SGD | Three-layer ReLU network · balloons | training objective: 0.7901 -> 0.6695 | Helped |
| Train with the proposed convex program instead of SGD | Three-layer ReLU network · breast-tissue | training objective: 0.5219 -> 0.3979 | Helped |
| Train with the proposed convex program instead of SGD | Three-layer ReLU network · fertiliy | training objective: 0.125 -> 0.1224 | Helped |
| Train with the proposed convex program instead of SGD | Three-layer ReLU network · pittsburg-bridges-span | training objective: 0.1723 -> 0.1668 | Helped |
| Train with the proposed convex program instead of SGD | Three-layer ReLU network · acute-inflamation | test error: 0.0224 -> 0.0217 | Helped |
| Train with the proposed convex program instead of SGD | Three-layer ReLU network · acute-nephritis | test error: 0.0198 -> 0.0192 | Helped |
| Train with the proposed convex program instead of SGD | Three-layer ReLU network · balloons | test error: 0.2693 -> 0.1496 | Helped |
| Train with the proposed convex program instead of SGD | Three-layer ReLU network · breast-tissue | test error: 1.4082 -> 1.0377 | Helped |
| Train with the proposed convex program instead of SGD | Three-layer ReLU network · fertiliy | test error: 0.3551 -> 0.5050 | Hurt |
| Train with the proposed convex program instead of SGD | Three-layer ReLU network · pittsburg-bridges-span | test error: 1.4373 -> 1.3112 | Helped |
Closest papers by meaning
- Convex Relaxations of ReLU Neural Networks Approximate Global Optima in Polynomial Time (ICML 2024 · similarity 0.92)
- Global Minimizers of $\ell^p$-Regularized Objectives Yield the Sparsest ReLU Neural Networks (NeurIPS 2025 · similarity 0.90)
- Convex Formulations for Training Two-Layer ReLU Neural Networks (ICLR 2025 · similarity 0.90)
- Activation-Descent Regularization for Input Optimization of ReLU Networks (ICML 2024 · similarity 0.90)
- Compelling ReLU Networks to Exhibit Exponentially Many Linear Regions at Initialization and During Training (ICML 2025 · similarity 0.90)
- Optimal Rates for Generalization of Gradient Descent for Deep ReLU Classification (NeurIPS 2025 · similarity 0.89)
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.