Research map / Architectures and efficiency / CNN design
CNN filter pruning
Prunes or factorizes convolutional filters using frequency-domain, gating, and redundancy criteria.
43 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2025) · an approach in CNN design, keyword cluster “cnn · filter · trained”
State History Benchmarks Tried Preprints Export
Where it stands
Quiet None of the tracked benchmarks has a report from 2025–26.
2 papers in 2023–24 and 1 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 1.5% to 1.0%.
No longer reported by papers here: ImageNet (Top-1 accuracy, last in 2022).
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2016 · Learning Structured Sparsity in Deep Neural Networks (NeurIPS 2016 · 458 citations)
- 2017 · The Reversible Residual Network: Backpropagation Without Storing Activations (NeurIPS 2017 · 351 citations)
- 2018 · Rethinking the Smaller-Norm-Less-Informative Assumption in Channel Pruning of Convolution Layers (ICLR 2018 · 149 citations)
- 2019 · Filter Pruning via Geometric Median for Deep Convolutional Neural Networks Acceleration (CVPR 2019 · 1,263 citations)
- 2020 · SCOP: Scientific Control for Reliable Neural Network Pruning (NeurIPS 2020 · 96 citations)
- 2021 · Convolutional Neural Network Pruning With Structural Redundancy Reduction (CVPR 2021 · 203 citations)
- 2022 · Convolutional Neural Network Compression through Generalized Kronecker Product Decomposition (AAAI 2022 · 21 citations)
- 2024 · REPrune: Channel Pruning via Kernel Representative Selection (AAAI 2024 · 6 citations)
- 2025 · Frequency-Biased Synergistic Design for Image Compression and Compensation (CVPR 2025 · 4 citations)
Earliest papers here
- Learning Structured Sparsity in Deep Neural Networks (NeurIPS 2016 · 458 citations)
- Learning Sparse High Dimensional Filters: Image Filtering, Dense CRFs and Bilateral Neural Networks (CVPR 2016 · 205 citations)
- Doubly Convolutional Neural Networks (NeurIPS 2016 · 32 citations)
Benchmark records
The best number each year that beat every earlier one, among papers of this approach reporting their own method. Numbers are extracted by AI from each paper's text; check the paper before citing.
ImageNet · Top-1 accuracy (higher is better)
4 papers, 2018–2022 · not reported since 2022
| Year | Best so far | Set by |
|---|---|---|
| 2018 | 56.82 | Frequency-Domain Dynamic Pruning for Convolutional Neural Networks (NeurIPS 2018 · 99 citations) |
| 2019 | 77.47 | Approximated Oracle Filter Pruning for Destructive CNN Width Optimization (ICML 2019 · 45 citations) |
What has been tried
42 papers here report 442 measured changes (a component added, removed or swapped, compared with a baseline): helped 63%, no clear effect 10%, hurt 24%, mixed 4%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Replace Ward linkage with single linkage | CIFAR-10, ResNet-56 · Top-1 accuracy | 93.40% -> 92.74% | Hurt | REPrune: Channel Pruning via Kernel Representative Selection (AAAI 2024 · 6 citations) |
| replace default quantization with Lo-Hi sensitive quantization | JPEG · BD-rate over LPIPS | -11.50% | Helped | Frequency-Biased Synergistic Design for Image Compression and Compensation (CVPR 2025 · 4 citations) |
| add dyadic decomposition front-end | NorthEastUS · MSE | 1.35 -> 1.05 | Helped | SoundCount: Sound Counting from Raw Audio with Dyadic Decomposition Neural Network (AAAI 2024 · 3 citations) |
Recent preprints
Loading recent arXiv preprints…
Use it
Download BibTeX of 23 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Filter design and trained-filter analysis (128 papers)
- Compact CNN architectures (114 papers)
- Theory of convolutional inductive bias (64 papers)
- Biologically inspired CNN explanation (46 papers)
Elsewhere in the topic
- Efficient CNN scaling and dynamic inference (104 papers)
- Dense matching and estimation CNNs (93 papers)
- Invertible and wide-network analysis (65 papers)
- Large and dynamic kernel design (62 papers)
- Group symmetry equivariant networks (55 papers)
- Steerable rotation-equivariant CNNs (40 papers)
- Higher-order pooling (20 papers)
- Group convolution theory (19 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).