Research map / Architectures and efficiency / CNN design
Circulant Binary Convolutional Networks: Enhancing the Performance of 1-Bit DCNNs With Circulant Back Propagation
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 |
|---|---|---|---|
| Use circulant binary convolution instead of XNOR convolution | LeNet · MNIST | error rate (%): 3.76 -> 1.91 | Helped |
| Use circulant binary convolution instead of XNOR convolution | LeNet · MNIST-rot | error rate (%): 17.26 -> 5.76 | Helped |
| Use circulant binary convolution instead of XNOR convolution | ResNet18 · CIFAR10-rot | error rate (%): 40.75 -> 19.68 | Helped |
| Use circulant binary convolution instead of XNOR convolution | ResNet18 · ImageNet | Top-1 accuracy (%): 51.2 -> 61.4 | Helped |
| Replace XNOR BConv with CBConv (ablation) | CIFAR10 ResNet18 with unchanged shortcuts · CIFAR10 | accuracy (%): 76.3 -> 84.79 | Helped |
| Add doubled shortcuts (ablation) | CBCN (K=4), CIFAR10 ResNet18 · CIFAR10 | accuracy (%): 84.79 -> 89.10 | Helped |
| Add center loss (ablation) | Doubled-shortcut CBCN (K=4), CIFAR10 ResNet18 · CIFAR10 | accuracy (%): 89.10 -> 89.6 | Helped |
| Use Gaussian rather than piecewise-polynomial gradient approximation (ablation) | Doubled-shortcut CBCN (K=4), CIFAR10 ResNet18 · CIFAR10 | accuracy (%): 89.10 -> 90.22 | Helped |
| Increase CiF orientations from K=2 to K=4 to K=8 (ablation) | Doubled-shortcut CBCN, CIFAR10 ResNet18 · CIFAR10 | accuracy (%): 85.79 -> 89.10 -> 90.80 | Helped |
Closest papers by meaning
- CircConv: A Structured Convolution with Low Complexity (AAAI 2019 · similarity 0.88)
- Modulated Convolutional Networks (CVPR 2018 · similarity 0.87)
- Training binary neural networks with real-to-binary convolutions (ICLR 2020 · similarity 0.86)
- BD-Net: Has Depth-Wise Convolution Ever Been Applied in Binary Neural Networks? (AAAI 2026 · similarity 0.86)
- Projection Convolutional Neural Networks for 1-bit CNNs via Discrete Back Propagation (AAAI 2019 · similarity 0.86)
- Bi-ViT: Pushing the Limit of Vision Transformer Quantization (AAAI 2024 · similarity 0.84)
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.