Research map / Architectures and efficiency / CNN design
Compressing Deep Convolutional Neural Networks by Stacking Low-dimensional Binary Convolution Filters
Read the paperCheck 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 |
|---|---|---|---|
| Increase compression by changing SLBF from (f1=1/2, f2=1/2) to (f1=1/2, f2=1/8) (ablation) | SLBF with VGG-16 · CIFAR-10 | accuracy (%): 91.30 -> 89.24 | Hurt |
| Increase compression by changing SLBF from (f1=1/2, f2=1/2) to (f1=1/2, f2=1/8) (ablation) | SLBF with ResNet-18 · CIFAR-100 | accuracy (%): 74.19 -> 67.89 | Hurt |
| Add scaling factors to the selected binary-filter blocks (ablation) | SLBF with VGG-16 · CIFAR-10 | accuracy | Helped |
| Use batch normalization with scaling and shifting (ablation) | SLBF · CIFAR-10 | accuracy | Helped |
| Use ADAM instead of SGD with momentum (ablation) | SLBF (f1=1/2, f2=1/2) · CIFAR-10 | accuracy | No clear effect |
| Initialize proxies from a pre-trained LegoNet rather than randomly (ablation) | SLBF (f1=1/2, f2=1/2) · CIFAR-10 | accuracy | No clear effect |
| Use larger batch sizes (ablation) | SLBF · CIFAR-10 | accuracy | No clear effect |
Closest papers by meaning
- Compressing Convolutional Neural Networks via Factorized Convolutional Filters (CVPR 2019 · similarity 0.90)
- FSNet: Compression of Deep Convolutional Neural Networks by Filter Summary (ICLR 2020 · similarity 0.89)
- On Compressing Deep Models by Low Rank and Sparse Decomposition (CVPR 2017 · similarity 0.89)
- CNNpack: Packing Convolutional Neural Networks in the Frequency Domain (NeurIPS 2016 · similarity 0.89)
- Compressing Neural Networks: Towards Determining the Optimal Layer-wise Decomposition (NeurIPS 2021 · similarity 0.88)
- Auto-Balanced Filter Pruning for Efficient Convolutional Neural Networks (AAAI 2018 · similarity 0.88)
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.