Research map / Architectures and efficiency / CNN design
Learning One Convolutional Layer with Overlapping Patches
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 |
|---|---|---|---|
| Use Convotron instead of randomly initialized SGD | One-layer convolutional networks · 1D and 2D patch-and-stride experiments | failure probability: almost 0.5 | Helped |
| Use Convotron instead of SGD | One-layer convolutional networks · Identity-covariance Gaussian inputs | failure probability | Helped |
| Use Convotron instead of SGD | One-layer convolutional networks · Non-identity covariance inputs | failure probability | Helped |
| Update using only one disjoint patch rather than all patch directions (ablation) | Convotron · Identity-covariance inputs with one non-overlapping patch | weight recovery | Helped |
| Use stochastic Convotron updates rather than gradient descent | One-layer convolutional networks with overlapping patches · General symmetric input distributions satisfying the stated patch conditions | weight recovery | Helped |
Closest papers by meaning
- Efficient Learning of CNNs using Patch Based Features (ICML 2022 · similarity 0.90)
- Tight Sample Complexity of Learning One-hidden-layer Convolutional Neural Networks (NeurIPS 2019 · similarity 0.87)
- Approximation and Learning with Deep Convolutional Models: a Kernel Perspective (ICLR 2022 · similarity 0.87)
- Deep Layer Aggregation (CVPR 2018 · similarity 0.86)
- RePr: Improved Training of Convolutional Filters (CVPR 2019 · similarity 0.86)
- PICNN: A Pathway towards Interpretable Convolutional Neural Networks (AAAI 2024 · 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.