AI Paper Atlas

Research map / Adversarial robustness and security / Certified and empirical robustness

Amplitude-Phase Recombination: Rethinking Robustness of Convolutional Neural Networks in Frequency Domain

Guangyao Chen, Peixi Peng, Li Ma, Jia Li, Lin Du, Yonghong Tian

ICCV 2021 · Adversarial robustness and security › Certified and empirical robustness · approach: Robust overfitting and regularization · 136 citations (OpenAlex)

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 changedWhereMeasuredEffect
add APR-S and APR-P together (APR-SP) to APR-S (ablation)APR-S · CIFAR-100-Cmean corruption error: 36.8 -> 33.9Improved
add APR-S to APR-P (APR-SP) (ablation)APR-P · CIFAR-10-Cmean corruption error: 19.7 -> 10.0Improved
replace Mixup with APR-P (amplitude swap, phase label)Standard training · CIFAR-100-Cmean corruption error: 52.6 -> 46.6Improved
replace AugMix with APR-SStandard training · CIFAR-100-Cmean corruption error: 38.3 -> 36.8Improved
add APR-SP on top of ImageNet augmentationStandard ResNet-50 · ImageNet-CmCE: 80.6 -> 65.0Improved
add APR-SP to DeepAugmentDeepAugment · ImageNet-CmCE: 63.1 -> 57.5Improved
add APR-SP to DeepAugmentDeepAugment · ImageNet cleantest error: 26.3 -> 26.4No clear effect
add APR-SP to standard trainingStandard ResNet-50 · ImageNet cleantest error: 23.9 -> 24.4Hurt
add APR-P instead of CutoutCE · CIFAR-10 OOD detectionmean AUROC: 91.0 -> 93.1Improved
add APR-P to APR-S (APR-SP) (ablation)APR-S · CIFAR-10 OOD detectionmean AUROC: 91.3 -> 94.7Improved
add APR-SP vs Cutout under FSGM adversarial trainingFSGM · CIFAR-10 AutoAttackrobust accuracy: 41.6 -> 45.7Improved
add Cutout to FSGMFSGM · CIFAR-10 AutoAttackrobust accuracy: 43.2 -> 41.6Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy