Research map / Architectures and efficiency / CNN design
Less Is More Important: An Attention Module Guided by Probability Density Function for Convolutional Neural Networks
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 |
|---|---|---|---|
| Add PdfAM-Gau attention | ResNet-18 on ImageNet · ImageNet validation | Top-1 accuracy: 70.33 % -> 71.32 % | Helped |
| Add PdfAM-Gau attention | ResNet-34 on ImageNet · ImageNet validation | Top-1 accuracy: 73.75 % -> 74.47 % | Helped |
| Add PdfAM-Gau attention | ResNet-50 on ImageNet · ImageNet validation | Top-1 accuracy: 76.34 % -> 77.41 % | Helped |
| Add PdfAM-Gau attention | ResNet-101 on ImageNet · ImageNet validation | Top-1 accuracy: 77.82 % -> 78.68 % | Helped |
| Add PdfAM-Gau attention | ResNeXt-50 on ImageNet · ImageNet validation | Top-1 accuracy: 77.47 % -> 78.13 % | Helped |
| Add PdfAM-Gau attention | MobileNetV2 on ImageNet · ImageNet validation | Top-1 accuracy: 71.90 % -> 72.39 % | Helped |
| Add PdfAM-Gau or PdfAM-Gmm | ISTA-RB on Brain MRI reconstruction · Brain MRI, 5%, 10%, and 15% sampling | PSNR: 30.74 -> 31.24; 34.91 -> 35.20; 37.12 -> 37.39 | Helped |
| Add PdfAM-Gau | HC-PGD on Brain MRI reconstruction · Brain MRI, 5%, 10%, and 15% sampling | PSNR: 30.52 -> 31.91; 34.75 -> 35.17; 36.77 -> 37.15 | Mixed |
| Add PdfAM-Gau | MADUN on Brain MRI reconstruction · Brain MRI, 5%, 10%, and 15% sampling | PSNR: 32.82 -> 32.99; 36.17 -> 36.28; 38.00 -> 38.09 | Helped |
| Increase GMM components from K=2 to K=4 (ablation) | PdfAM-Gmm on ResNet-20 · CIFAR-10 | Top-1 accuracy: 92.68 -> 92.84 | Helped |
| Increase EM iterations from T=3 to T=5 (ablation) | PdfAM-Gmm on ResNet-20 · CIFAR-10 | Top-1 accuracy: 92.68 -> 92.75 | Helped |
| Add learnable PDF-rescale parameters (ablation) | PdfAM-Gmm on ResNet-20 · CIFAR-10 | Top-1 accuracy: 92.49 -> 92.68 | Helped |
Closest papers by meaning
- Bridging the Divide: Reconsidering Softmax and Linear Attention (NeurIPS 2024 · similarity 0.86)
- PolaFormer: Polarity-aware Linear Attention for Vision Transformers (ICLR 2025 · similarity 0.84)
- Skip-Attention: Improving Vision Transformers by Paying Less Attention (ICLR 2024 · similarity 0.84)
- Breaking the Low-Rank Dilemma of Linear Attention (CVPR 2025 · similarity 0.84)
- EAN: An Efficient Attention Module Guided by Normalization for Deep Neural Networks (AAAI 2024 · similarity 0.83)
- Faster Neighborhood Attention: Reducing the O(n^2) Cost of Self Attention at the Threadblock Level (NeurIPS 2024 · similarity 0.83)
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.