Research map / Architectures and efficiency / Pruning and efficient tuning
Cascaded Projection: End-To-End Network Compression and Acceleration
Check 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 |
|---|---|---|---|
| Add optimized projection matrices instead of using random projections (ablation) | CaP compression with kernel relaxation · CIFAR100 | Accuracy: 76.27 -> 77.47 | Helped |
| Add kernel relaxation after projection compression (ablation) | CaP compression · CIFAR100 | Accuracy: 76.65 -> 77.47 | Helped |
| Optimize the projections rather than using random projections (ablation) | CaP with kernel relaxation · CIFAR100 | Accuracy: 76.27 -> 77.47 | Helped |
| Combine projection optimization and kernel relaxation rather than training the compressed model from scratch (ablation) | ResNet18 compression · CIFAR100 | Accuracy: 77.22 -> 77.47 | Helped |
| Use linear combinations of feature channels instead of pruning entire channels | CaP · ImageNet2012 | Top-5 accuracy: 82.0 -> 88.23 | Helped |
| Cascade the projection into the next layer to avoid feature reprojection | CaP · ImageNet2012 | GPU speedup: 2.5* -> 3.44 | Helped |
| Use simultaneous projection optimization and deep supervision across layers | CaP · CIFAR10 | Accuracy | Helped |
Closest papers by meaning
- Compressing Neural Networks: Towards Determining the Optimal Layer-wise Decomposition (NeurIPS 2021 · similarity 0.89)
- Towards Compact CNNs via Collaborative Compression (CVPR 2021 · similarity 0.87)
- "Lossless" Compression of Deep Neural Networks: A High-dimensional Neural Tangent Kernel Approach (NeurIPS 2022 · similarity 0.87)
- On Compressing Deep Models by Low Rank and Sparse Decomposition (CVPR 2017 · similarity 0.87)
- Distribution-Aware Tensor Decomposition for Compression of Convolutional Neural Networks (NeurIPS 2025 · similarity 0.87)
- Compression-aware Training of Deep Networks (NeurIPS 2017 · similarity 0.87)
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.