Research map / Architectures and efficiency / RNNs and normalization
SCAN: A Scalable Neural Networks Framework Towards Compact and Efficient Models
Linfeng Zhang, Zhanhong Tan, Jiebo Song, Jingwei Chen, Chenglong Bao, Kaisheng Ma
Read the paperCheck your idea against 137,217 papers
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 SCAN classifiers and ensemble their predictions | VGG16(BN) · CIFAR100 | accuracy (%): 72.46 -> 76.80 | Improved |
| Add SCAN classifiers and ensemble their predictions | VGG19(BN) · CIFAR100 | accuracy (%): 72.25 -> 75.43 | Improved |
| Add SCAN classifiers and ensemble their predictions | ResNet18 · CIFAR100 | accuracy (%): 77.09 -> 80.46 | Improved |
| Add SCAN classifiers and ensemble their predictions | ResNet50 · CIFAR100 | accuracy (%): 77.68 -> 81.78 | Improved |
| Add SCAN classifiers and ensemble their predictions | ResNet101 · CIFAR100 | accuracy (%): 77.98 -> 82.06 | Improved |
| Add SCAN classifiers and ensemble their predictions | ResNet152 · CIFAR100 | accuracy (%): 79.21 -> 82.94 | Improved |
| Add SCAN classifiers and ensemble their predictions | WRN20-8 · CIFAR100 | accuracy (%): 74.61 -> 80.04 | Improved |
| Add SCAN classifiers and ensemble their predictions | WRN44-8 · CIFAR100 | accuracy (%): 76.22 -> 80.35 | Improved |
| Add SCAN shallow classifiers and distillation training | ResNet18 · ImageNet (ILSVRC2012) validation | accuracy (%): 68.02 -> 69.32 | Improved |
| Add SCAN shallow classifiers and distillation training | ResNet50 · ImageNet (ILSVRC2012) validation | accuracy (%): 74.47 -> 75.88 | Improved |
| Add SCAN shallow classifiers and distillation training | ResNet101 · ImageNet (ILSVRC2012) validation | accuracy (%): 75.24 -> 76.32 | Improved |
| Add attention modules to shallow classifiers (ablation) | ResNet50 with self distillation · CIFAR100 | accuracy (%): +5.46%, 4.13% ,and 5.16% | Improved |
Closest papers by meaning
- HydraNets: Specialized Dynamic Architectures for Efficient Inference (CVPR 2018 · similarity 0.86)
- Sampling weights of deep neural networks (NeurIPS 2023 · similarity 0.85)
- Adaptive Neural Networks for Efficient Inference (ICML 2017 · similarity 0.85)
- A Theoretical Framework for Inference Learning (NeurIPS 2022 · similarity 0.85)
- Diving into the shallows: a computational perspective on large-scale shallow learning (NeurIPS 2017 · similarity 0.85)
- Feedback-Prop: Convolutional Neural Network Inference Under Partial Evidence (CVPR 2018 · 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.
