Research map / Architectures and efficiency / Pruning and efficient tuning
DessiLBI: Exploring Structural Sparsity of Deep Networks via Differential Inclusion Paths
Yanwei Fu, Chen Liu, Donghao Li, Xinwei Sun, Jinshan Zeng, Yuan Yao
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 |
|---|---|---|---|
| Use DessiLBI with momentum and weight decay instead of SGD Mom-Wd | AlexNet · ImageNet-2012 | Top-1 accuracy: 56.55 -> 57.09 | Improved |
| Use DessiLBI with momentum and weight decay instead of SGD Mom-Wd | AlexNet · ImageNet-2012 | Top-5 accuracy: 79.09 -> 79.86 | Improved |
| Use DessiLBI with momentum and weight decay instead of SGD Nesterov | ResNet-18 · ImageNet-2012 | Top-1 accuracy: 70.19 -> 70.55 | Improved |
| Use DessiLBI with momentum and weight decay instead of SGD Nesterov | ResNet-18 · ImageNet-2012 | Top-5 accuracy: 89.30 -> 89.56 | Improved |
| Add weight decay to DessiLBI with momentum (ablation) | AlexNet · ImageNet-2012 | Top-1 accuracy: 56.23 -> 57.09 | Improved |
| Add weight decay to DessiLBI with momentum (ablation) | AlexNet · ImageNet-2012 | Top-5 accuracy: 78.48 -> 79.86 | Improved |
| Add momentum and weight decay to naive DessiLBI (ablation) | AlexNet · ImageNet-2012 | Top-1 accuracy: 55.06 -> 57.09 | Improved |
| Add momentum and weight decay to naive DessiLBI (ablation) | AlexNet · ImageNet-2012 | Top-5 accuracy: 77.69 -> 79.86 | Improved |
| Add weight decay to DessiLBI with momentum (ablation) | ResNet-18 · ImageNet-2012 | Top-1 accuracy: 68.55 -> 70.55 | Improved |
| Add weight decay to DessiLBI with momentum (ablation) | ResNet-18 · ImageNet-2012 | Top-5 accuracy: 87.85 -> 89.56 | Improved |
| Add momentum and weight decay to naive DessiLBI (ablation) | ResNet-18 · ImageNet-2012 | Top-1 accuracy: 65.26 -> 70.55 | Improved |
| Add momentum and weight decay to naive DessiLBI (ablation) | ResNet-18 · ImageNet-2012 | Top-5 accuracy: 86.57 -> 89.56 | Improved |
Closest papers by meaning
- Exploring Structural Sparsity of Deep Networks Via Inverse Scale Spaces (TPAMI 2022 · similarity 0.95)
- Sparsity in Deep Learning: Pruning and growth for efficient inference and training in neural networks (JMLR 2021 · similarity 0.88)
- Data-Driven Sparse Structure Selection for Deep Neural Networks (ECCV 2018 · similarity 0.88)
- Random Sparse Lifts: Construction, Analysis and Convergence of finite sparse networks (ICLR 2024 · similarity 0.87)
- Sparse maximal update parameterization: A holistic approach to sparse training dynamics (NeurIPS 2024 · similarity 0.86)
- Position: A Theory of Deep Learning Must Include Compositional Sparsity (ICML 2025 · similarity 0.86)
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.
