Research map / Architectures and efficiency / CNN design
You Look Twice: GaterNet for Dynamic Filter Selection in CNNs
Zhourong Chen, Yang Li, Samy Bengio, Si Si
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| Add GaterNet dynamic filter selection to ResNet-20 | ResNet-20 · CIFAR-10 test | error rate (%): 8.06 -> 6.88 | Improved |
| Add GaterNet dynamic filter selection to ResNet-20 | ResNet-20 · CIFAR-100 test | error rate (%): 32.39 -> 30.79 | Improved |
| Add GaterNet dynamic filter selection to ResNet-56 | ResNet-56 · CIFAR-10 test | error rate (%): 6.74 -> 5.72 | Improved |
| Add GaterNet dynamic filter selection to ResNet-56 | ResNet-56 · CIFAR-100 test | error rate (%): 28.87 -> 27.71 | Improved |
| Add GaterNet dynamic filter selection to ResNet-164 | ResNet-164 · CIFAR-10 test | error rate (%): 5.61 -> 4.80 | Improved |
| Add GaterNet dynamic filter selection to ResNet-164 | ResNet-164 · CIFAR-100 test | error rate (%): 25.39 -> 23.56 | Improved |
| Add GaterNet dynamic filter selection | DenseNet-BC and Shake-Shake · CIFAR-10 test | error rate (%): 4.48 -> 4.03; 3.61 -> 3.31; 3.52 -> 3.39; 3.05 -> 2.89; 2.82 -> 2.64 | Improved |
| Add GaterNet dynamic filter selection to ResNet-34 | ResNet-34 · ImageNet 2012 validation | top-1 error (%): 26.56 -> 26.04 | Improved |
| Add GaterNet dynamic filter selection to ResNet-34 | ResNet-34 · ImageNet 2012 validation | top-5 error (%): 8.48 -> 8.34 | Improved |
| Add GaterNet dynamic filter selection to ResNet-101 | ResNet-101 · ImageNet 2012 validation | top-1 error (%): 23.36 -> 21.51 | Improved |
| Add GaterNet dynamic filter selection to ResNet-101 | ResNet-101 · ImageNet 2012 validation | top-5 error (%): 6.56 -> 5.78 | Improved |
| Add GaterNet dynamic filter selection to ResNet-152 | ResNet-152 · ImageNet 2012 validation | top-1 error (%): 22.34 -> 21.19 | Improved |
Closest papers by meaning
- Decoupled Dynamic Filter Networks (CVPR 2021 · similarity 0.85)
- DS-Net++: Dynamic Weight Slicing for Efficient Inference in CNNs and Vision Transformers (TPAMI 2022 · similarity 0.85)
- Dynamic Slimmable Network (CVPR 2021 · similarity 0.85)
- Dynamic Deep Neural Networks: Optimizing Accuracy-Efficiency Trade-Offs by Selective Execution (AAAI 2018 · similarity 0.84)
- Convolutional Differentiable Logic Gate Networks (NeurIPS 2024 · similarity 0.83)
- Hypernetworks for Dynamic Feature Selection (NeurIPS 2026 · 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.
