Research map / Image generation / Image restoration
A Tour of Convolutional Networks Guided by Linear Interpreters
Pablo Navarrete Michelini, Hanwen Liu, Yunhua Lu, Xingqun Jiang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 a larger or better ImageNet classifier family member (e.g. VGG-11 to VGG-19) | VGG classifiers · ImageNet-1k val | average residual contribution to top-1 score (%): 82.2% -> 85.5% | Improved |
| use DenseNet-161 instead of AlexNet | classification CNNs · ImageNet-1k val | average residual contribution to top-1 score (%): 78.5% -> 95.0% | Improved |
| use ResNet-152 instead of ResNet-34 | ResNet classifiers · ImageNet-1k val | average residual contribution to top-1 score (%): 83.7% -> 81.1% | No clear effect |
| add batch normalization to VGG training | VGG-19 · ImageNet-1k val | layer-wise bias contributions: 85.5% -> 85.8% | Mixed |
| replace storing activations (DFV) with LinearScope extra batch | linear interpreter · speed | speedup: 2x to 10^4x speedup over DFV | Improved |
| apply FGSM adversarial noise to inputs | VGG classifiers · ImageNet-1k val | layer-wise contributions and pixel votes | Hurt |
Closest papers by meaning
- Linear CNNs Discover the Statistical Structure of the Dataset Using Only the Most Dominant Frequencies (ICML 2023 · similarity 0.92)
- Non-Linear Convolution Filters for CNN-Based Learning (ICCV 2017 · similarity 0.89)
- LSNet: See Large, Focus Small (CVPR 2025 · similarity 0.88)
- KernelWarehouse: Rethinking the Design of Dynamic Convolution (ICML 2024 · similarity 0.88)
- PICNN: A Pathway towards Interpretable Convolutional Neural Networks (AAAI 2024 · similarity 0.88)
- AsCAN: Asymmetric Convolution-Attention Networks for Efficient Recognition and Generation (NeurIPS 2024 · similarity 0.87)
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.
