Research map / Learning theory and algorithms / Sample complexity
Recovery Guarantees for One-hidden-layer Neural Networks
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 |
|---|---|---|---|
| initialize W and v with tensor methods, then optimize W while fixing v | Squared-ReLU 1NN gradient descent · Synthetic Gaussian regression | convergence speed | Helped |
| use random Gaussian initialization for W and v, then optimize both (ablation) | Squared-ReLU 1NN gradient descent · Synthetic Gaussian regression | convergence | Hurt |
| initialize W randomly while fixing v to its ground-truth value (ablation) | Squared-ReLU 1NN gradient descent · Synthetic Gaussian regression | convergence rate | Helped |
| remove resampling between gradient-descent iterations (ablation) | Tensor-initialized squared-ReLU 1NN · Synthetic Gaussian regression | algorithm performance | No clear effect |
| increase the sample count | Tensor initialization · Synthetic Gaussian regression | initialization error | Helped |
| use tensor initialization followed by gradient descent instead of random initialization | One-hidden-layer neural network · Synthetic Gaussian regression | recovery rate | Helped |
Closest papers by meaning
- On the Complexity of Learning Neural Networks (NeurIPS 2017 · similarity 0.85)
- A Recovery Guarantee for Sparse Neural Networks (ICLR 2026 · similarity 0.85)
- Optimization and Generalization of Shallow Neural Networks with Quadratic Activation Functions (NeurIPS 2020 · similarity 0.85)
- Efficiently Learning One Hidden Layer ReLU Networks From Queries (NeurIPS 2021 · similarity 0.84)
- Loss Landscape of Shallow ReLU-like Neural Networks: Stationary Points, Saddle Escape, and Network Embedding (ICLR 2025 · similarity 0.83)
- Wide Neural Networks Trained with Weight Decay Provably Exhibit Neural Collapse (ICLR 2025 · similarity 0.82)
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.