Research map / Architectures and efficiency / CNN design
Residual Attention Network for Image Classification
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 |
|---|---|---|---|
| Use attention residual learning instead of direct soft-mask multiplication (ablation) | Attention-56 · CIFAR-10 test | Top-1 error (%): 5.89 -> 5.52 | Helped |
| Use attention residual learning instead of direct soft-mask multiplication (ablation) | Attention-92 · CIFAR-10 test | Top-1 error (%): 5.35 -> 4.99 | Helped |
| Use attention residual learning instead of direct soft-mask multiplication (ablation) | Attention-128 · CIFAR-10 test | Top-1 error (%): 5.57 -> 4.44 | Helped |
| Use attention residual learning instead of direct soft-mask multiplication (ablation) | Attention-164 · CIFAR-10 test | Top-1 error (%): 7.18 -> 4.31 | Helped |
| Use an encoder-decoder mask rather than local convolutions (ablation) | Attention-56 · CIFAR-10 test | Top-1 error (%): 6.48 -> 5.52 | Helped |
| Use unconstrained mixed attention rather than channel-only attention (ablation) | Attention-56 · CIFAR-10 test | Top-1 error (%): 6.24 -> 5.52 | Helped |
| Use unconstrained mixed attention rather than spatial-only attention (ablation) | Attention-56 · CIFAR-10 test | Top-1 error (%): 6.33 -> 5.52 | Helped |
| Add attention residual learning (ablation) | Residual Attention Network · CIFAR-10 with stacked attention modules | Top-1 error (%) | Helped |
| Use stacked Attention Modules | Residual Attention Network · CIFAR-10 and CIFAR-100 | Top-1 error (%) | Helped |
| Use Attention-92 rather than ResNet-164 | Image classification model · CIFAR-10 with 10% label noise | Error (%): 5.93 -> 5.15 | Helped |
| Use Attention-92 rather than ResNet-164 | Image classification model · CIFAR-10 with 30% label noise | Error (%): 6.61 -> 5.79 | Helped |
| Use Attention-92 rather than ResNet-164 | Image classification model · CIFAR-10 with 50% label noise | Error (%): 8.35 -> 7.27 | Helped |
Closest papers by meaning
- Deep Residual Learning for Image Recognition (CVPR 2016 · similarity 0.90)
- Learn to Pay Attention (ICLR 2018 · similarity 0.90)
- DIANet: Dense-and-Implicit Attention Network (AAAI 2020 · similarity 0.89)
- A^2-Nets: Double Attention Networks (NeurIPS 2018 · similarity 0.88)
- SimAM: A Simple, Parameter-Free Attention Module for Convolutional Neural Networks (ICML 2021 · similarity 0.88)
- ResMatch: Residual Attention Learning for Feature Matching (AAAI 2024 · 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.