Research map / Architectures and efficiency / Quantization
Rotated Binary Neural Network
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 RBNN instead of IR-Net | binary ResNet-18 · CIFAR-10 | accuracy: 91.5% -> 92.2% | Helped |
| use RBNN instead of IR-Net | binary ResNet-20 · CIFAR-10 | accuracy: 85.4% -> 86.5% | Helped |
| use RBNN instead of IR-Net | binary ResNet-20 Bi-Real · CIFAR-10 | accuracy: 86.5% -> 87.8% | Helped |
| use RBNN instead of IR-Net | binary VGG-small · CIFAR-10 | accuracy: 90.4% -> 91.3% | Helped |
| use RBNN instead of IR-Net | binary ResNet-18 · ImageNet | top-1 accuracy: 58.1% -> 59.9% | Helped |
| use RBNN instead of IR-Net | binary ResNet-18 · ImageNet | top-5 accuracy: 80.0% -> 81.9% | Helped |
| use RBNN instead of IR-Net | binary ResNet-34 · ImageNet | top-1 accuracy: 62.9% -> 63.1% | Helped |
| use RBNN instead of IR-Net | binary ResNet-34 · ImageNet | top-5 accuracy: 84.1% -> 84.4% | Helped |
| add weight rotation to XNOR-Net binarization (ablation) | B (XNOR-Net binarization) · CIFAR-10, ResNet-20 Bi-Real | accuracy: 83.7% -> 86.4% | Helped |
| add training-aware approximation to XNOR-Net binarization (ablation) | B (XNOR-Net binarization) · CIFAR-10, ResNet-20 Bi-Real | accuracy: 83.7% -> 86.6% | Helped |
| add both training-aware approximation and weight rotation (ablation) | B (XNOR-Net binarization) · CIFAR-10, ResNet-20 Bi-Real | accuracy: 83.7% -> 87.1% | Helped |
| add the adjustable rotated weight scheme (ablation) | B + T + R · CIFAR-10, ResNet-20 Bi-Real | accuracy: 87.1% -> 87.8% | Helped |
Closest papers by meaning
- SA-BNN: State-Aware Binary Neural Network (AAAI 2021 · similarity 0.86)
- BiPer: Binary Neural Networks using a Periodic Function (CVPR 2024 · similarity 0.86)
- Binarized Neural Networks (NeurIPS 2016 · similarity 0.86)
- Fast and Accurate Binary Neural Networks Based on Depth-Width Reshaping (AAAI 2023 · similarity 0.85)
- Training Binary Neural Networks via Gaussian Variational Inference and Low-Rank Semidefinite Programming (NeurIPS 2024 · similarity 0.85)
- A&B BNN: Add&Bit-Operation-Only Hardware-Friendly Binary Neural Network (CVPR 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.