Research map / Architectures and efficiency / CNN design
Optimization Landscape and Expressivity of Deep CNNs
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 first-layer filter count from 10 to 30 (ablation) | MNIST CNN · MNIST train | train error: 8 -> 0 | Helped |
| Increase first-layer filter count from 30 to 100 (ablation) | MNIST CNN · MNIST test | test error: 174 -> 139 | Mixed |
| Use a first hidden layer wide enough to have at least as many neurons as training samples | wide CNN with a fully connected next layer · training data under the stated architecture assumptions | training loss | Helped |
Closest papers by meaning
- Large Scale Structure of Neural Network Loss Landscapes (NeurIPS 2019 · similarity 0.92)
- Loss Landscape Characterization of Neural Networks without Over-Parametrization (NeurIPS 2024 · similarity 0.88)
- Exploring The Loss Landscape Of Regularized Neural Networks Via Convex Duality (ICLR 2025 · similarity 0.87)
- No Wrong Turns: The Simple Geometry Of Neural Networks Optimization Paths (ICML 2024 · similarity 0.87)
- Flat Channels to Infinity in Neural Loss Landscapes (NeurIPS 2025 · similarity 0.87)
- Exploring Landscapes for Better Minima along Valleys (NeurIPS 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.