Research map / Architectures and efficiency / Pruning and efficient tuning
Dynamic Network Pruning with Interpretable Layerwise Channel Selection
Yulong Wang, Xiaolu Zhang, Xiaolin Hu, Bo Zhang, Hang Su
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 |
|---|---|---|---|
| dynamic discrete channel-selection pruning vs best static/dynamic baselines (VGG16-BN, ~50% MACs) | VGG16-BN · CIFAR10 | accuracy: 93.03% (FBS) -> 93.45% | Improved |
| increase number of actions from 1 (static) to 5 in decision unit (ablation) | ResNet56 · CIFAR10 | MACs ratio (%) (lower better): 62.68 -> 47.38 | Improved |
| increase number of actions beyond 5 up to 10 (ablation) | ResNet56 · CIFAR10 | accuracy: 92.57 -> 92.00 | Hurt |
| vary action number m (1 to 10) (ablation) | VGG16-BN · CIFAR10 | accuracy: 93.38 (m=1) vs 93.45 (m=5) | No clear effect |
| discrete decision features with tree classifiers vs Mahalanobis detector | adversarial detection (ResNet-34) · CIFAR10 CW | AUROC (%): 95.85 -> 99.27 | Improved |
| decision features with GBDT vs Mahalanobis detector | adversarial detection (ResNet-34) · CIFAR10 DeepFool | AUROC (%): 91.53 -> 91.44 | No clear effect |
| decision features with GBDT vs Mahalanobis detector | adversarial detection (ResNet-34) · CIFAR100 DeepFool | AUROC (%): 83.93 -> 81.03 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- ThiNet — ThiNet: A Filter Level Pruning Method for Deep Neural Network Compression (ICCV 2017) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Fire Together Wire Together: A Dynamic Pruning Approach With Self-Supervised Mask Prediction (CVPR 2022 · similarity 0.86)
- Rethinking the Value of Network Pruning (ICLR 2019 · similarity 0.86)
- Efficient Distortion-Minimized Layerwise Pruning (TPAMI 2025 · similarity 0.86)
- Pruning from Scratch (AAAI 2020 · similarity 0.86)
- BilevelPruning: Unified Dynamic and Static Channel Pruning for Convolutional Neural Networks (CVPR 2024 · similarity 0.86)
- RG-GAN: Dynamic Regenerative Pruning for Data-Efficient Generative Adversarial Networks (AAAI 2024 · 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.
