Research map / Architectures and efficiency / CNN design
Skew Orthogonal Convolutions
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 |
|---|---|---|---|
| Replace BCOP convolutions with SOC convolutions | LipConvnet-20 · CIFAR-10, l2 radius 36/255 | standard accuracy: 69.84% -> 76.43% | Helped |
| Replace BCOP convolutions with SOC convolutions | LipConvnet-20 · CIFAR-10, l2 radius 36/255 | robust accuracy: 52.10% -> 61.92% | Helped |
| Replace BCOP convolutions with SOC convolutions | LipConvnet-20 · CIFAR-10, l2 radius 36/255 | time per epoch: 170.009 -> 77.226 | Helped |
| Replace BCOP convolutions with SOC convolutions | LipConvnet-20 · CIFAR-100, l2 radius 36/255 | standard accuracy: 36.13% -> 43.07% | Helped |
| Replace BCOP convolutions with SOC convolutions | LipConvnet-20 · CIFAR-100, l2 radius 36/255 | robust accuracy: 22.50% -> 29.18% | Helped |
| Replace BCOP convolutions with SOC convolutions | LipConvnet-40 · CIFAR-10, l2 radius 36/255 | standard accuracy: 60.17% -> 71.63% | Helped |
| Replace BCOP convolutions with SOC convolutions | LipConvnet-40 · CIFAR-10, l2 radius 36/255 | robust accuracy: 38.87% -> 54.36% | Helped |
| Replace BCOP convolutions with SOC convolutions | LipConvnet-40 · CIFAR-100, l2 radius 36/255 | standard accuracy: 19.97% -> 41.83% | Helped |
| Replace BCOP convolutions with SOC convolutions | LipConvnet-40 · CIFAR-100, l2 radius 36/255 | robust accuracy: 8.66% -> 27.98% | Helped |
Closest papers by meaning
- An Adaptive Orthogonal Convolution Scheme for Efficient and Flexible CNN Architectures (ICML 2025 · similarity 0.89)
- A Rescaling-Invariant Lipschitz Bound Based on Path-Metrics for Modern ReLU Network Parameterizations (ICML 2025 · similarity 0.86)
- Approximation theory for 1-Lipschitz ResNets (NeurIPS 2025 · similarity 0.86)
- ECLipsE: Efficient Compositional Lipschitz Constant Estimation for Deep Neural Networks (NeurIPS 2024 · similarity 0.85)
- Enhancing Certified Robustness via Block Reflector Orthogonal Layers and Logit Annealing Loss (ICML 2025 · similarity 0.85)
- DP-SGD Without Clipping: The Lipschitz Neural Network Way (ICLR 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.