Research map / Architectures and efficiency / CNN design
Convolutional Neural Networks With Alternately Updated Clique
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 |
|---|---|---|---|
| Use Stage-II features rather than Stage-I features for the block representation and transition (ablation) | CliqueNet · CIFAR-10 | error rate (%): 6.64 -> 5.76 | Helped |
| Use Stage-II transition features while retaining Stage-I block features (ablation) | CliqueNet · CIFAR-10 | error rate (%): 6.64 -> 6.1 | Helped |
| Add attentional transition (ablation) | CliqueNet (k=36, T=12) · CIFAR-10 | error rate (%): 5.93 -> 5.8 | Helped |
| Add attentional transition (ablation) | CliqueNet (k=36, T=12) · CIFAR-100 | error rate (%): 27.32 -> 26.41 | Helped |
| Add compression (ablation) | CliqueNet (k=36, T=12) · CIFAR-10 | error rate (%): 5.93 -> 5.69 | Helped |
| Add both attentional transition and compression (ablation) | CliqueNet (k=36, T=12) · CIFAR-10 | error rate (%): 5.93 -> 5.61 | Helped |
| Add both attentional transition and compression (ablation) | CliqueNet (k=36, T=12) · CIFAR-100 | error rate (%): 27.32 -> 25.55 | Helped |
| Add attentional transition and compression (ablation) | CliqueNet (k=80, T=15) · SVHN | error rate (%): 1.56 -> 1.53 | Helped |
| Add attentional transition (ablation) | CliqueNet-S2 · ImageNet ILSVRC 2012 validation, single crop | top-1 error rate (%): 25.85 -> 24.82 | Helped |
| Add attentional transition (ablation) | CliqueNet-S3 · ImageNet ILSVRC 2012 validation, single crop | top-1 error rate (%): 24.98 -> 24.01 | Helped |
Closest papers by meaning
- ClcNet: Improving the Efficiency of Convolutional Neural Network Using Channel Local Convolutions (CVPR 2018 · similarity 0.87)
- Refining Architectures of Deep Convolutional Neural Networks (CVPR 2016 · similarity 0.87)
- Densely Connected Convolutional Networks (CVPR 2017 · similarity 0.87)
- CliqueCNN: Deep Unsupervised Exemplar Learning (NeurIPS 2016 · similarity 0.86)
- More Is Less: A More Complicated Network With Less Inference Complexity (CVPR 2017 · similarity 0.86)
- Learning Implicitly Recurrent CNNs Through Parameter Sharing (ICLR 2019 · 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.