Research map / Architectures and efficiency / CNN design
Fully Learnable Group Convolution for Acceleration of Deep Neural Networks
Check 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 standard group convolution with FLGC at G=8 | ResNet50 · LFW | verification accuracy (%): 98.30 -> 98.73 | Helped |
| Replace standard group convolution with FLGC at G=4 | ResNet50 · LFW | verification accuracy (%): 98.78 -> 98.82 | Helped |
| Replace standard group convolution with FLGC at G=2 | ResNet50 · LFW | verification accuracy (%): 98.81 -> 98.82 | Helped |
| Use FLGC instead of standard group convolution at G=2 | MobileNetV2 · CIFAR-10 | classification error (%): 6.04 -> 5.89 | Helped |
| Use FLGC instead of standard group convolution at G=4 | MobileNetV2 · CIFAR-10 | classification error (%): 6.64 -> 5.84 | Helped |
| Use FLGC instead of standard group convolution at G=8 | MobileNetV2 · CIFAR-10 | classification error (%): 7.51 -> 6.91 | Helped |
| Replace standard group convolution with FLGC | CondenseNet · ImageNet | Top-1 classification error (%): 29.0 -> 25.3 | Helped |
| Replace standard group convolution with FLGC | CondenseNet · ImageNet | Top-5 classification error (%): 9.9 -> 7.9 | Helped |
| Replace CondenseNet's learnable group convolution with FLGC | CondenseNet · ImageNet | Top-1 classification error (%): 26.2 -> 25.3 | Helped |
| Replace CondenseNet's learnable group convolution with FLGC | CondenseNet · ImageNet | Top-5 classification error (%): 8.3 -> 7.9 | Helped |
| Replace 1x1 convolutions with FLGC | ResNet50 · CIFAR-10 | classification error (%): 8.2 -> 7.95 | Helped |
| Replace 1x1 convolutions with FLGC | ResNet50 · CIFAR-10 | classification error (%): 6.94 -> 6.77 | Helped |
Closest papers by meaning
- Exploiting Redundancy: Separable Group Convolutional Networks on Lie Groups (ICML 2022 · similarity 0.88)
- Group Equivariant Convolutional Networks (ICML 2016 · similarity 0.87)
- Group Downsampling with Equivariant Anti-aliasing (ICLR 2025 · similarity 0.85)
- Continuous Rotation Group Equivariant Network Inspired by Neural Population Coding (AAAI 2024 · similarity 0.84)
- Variational Partial Group Convolutions for Input-Aware Partial Equivariance of Rotations and Color-Shifts (ICML 2024 · similarity 0.84)
- On the VC dimension of deep group convolutional neural networks (NeurIPS 2025 · similarity 0.83)
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.