mlsanityData to 7 Oct 2026

How Does Frequency Bias Affect the Robustness of Neural Image Classifiers against Common Corruption and Adversarial Perturbations?

arXiv 2205.04533 ↗First posted 9 May 2022cs.LG cs.CVcited by 11

Not on the map. Why?

Abstract (arXiv)

Model robustness is vital for the reliable deployment of machine learning models in real-world applications. Recent studies have shown that data augmentation can result in model over-relying on features in the low-frequency domain, sacrificing performance against low-frequency corruptions, highlighting a connection between frequency and robustness. Here, we take one step further to more directly study the frequency bias of a model through the lens of its Jacobians and its implication to model robustness. To achieve this, we propose Jacobian frequency regularization for models' Jacobians to have a larger ratio of low-frequency components. Through experiments on four image datasets, we show that biasing classifiers towards low (high)-frequency components can bring performance gain against high (low)-frequency corruption and adversarial perturbation, albeit with a tradeoff in performance for low (high)-frequency corruption. Our approach elucidates a more direct connection between the frequency bias and robustness of deep learning models.

Key facts. Cited by 11 papers in the map (data through 7 Oct 2026). Builds on 37 of its 50 references that are in the map. First posted on arXiv 9 May 2022.

Builds on

37 of 50 references.

References · 37

  1. 2014Explaining and Harnessing Adversarial Examplescited by 7,287
  2. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,326
  3. 2013Intriguing properties of neural networkscited by 5,114
  4. 2016Towards Evaluating the Robustness of Neural Networkscited by 2,976
  5. 2019Benchmarking Neural Network Robustness to Common Corruptions and Perturbationscited by 2,857
  6. 2016Identity Mappings in Deep Residual Networkscited by 1,960
  7. 2020The Many Faces of Robustness: A Critical Analysis of Out-of-Distribution Generalizationcited by 1,842
  8. 2020Reliable evaluation of adversarial robustness with an ensemble of diverse parameter-free attackscited by 1,442
  9. 2018ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustnesscited by 1,430
  10. 2019Natural Adversarial Examplescited by 1,401
  11. 2018Sanity Checks for Saliency Mapscited by 1,149
  12. 2017SmoothGrad: removing noise by adding noisecited by 1,097
25 more
  1. 2019Adversarial Examples Are Not Bugs, They Are Featurescited by 950
  2. 2019AugMix: A Simple Data Processing Method to Improve Robustness and Uncertaintycited by 897
  3. 2018Robustness May Be at Odds with Accuracycited by 873
  4. 2018AutoAugment: Learning Augmentation Policies from Datacited by 728
  5. 2019Minimally distorted Adversarial Examples with a Fast Adaptive Boundary Attackcited by 333
  6. 2019A Fourier Perspective on Model Robustness in Computer Visioncited by 288
  7. 2017Improving the Adversarial Robustness and Interpretability of Deep Neural Networks by Regularizing their Input Gradientscited by 236
  8. 2020Understanding and Improving Fast Adversarial Trainingcited by 213
  9. 2019Adversarial Robustness through Local Linearizationcited by 146
  10. 2016Technical Report on the CleverHans v2.1.0 Adversarial Examples Librarycited by 123
  11. 2019Improving Robustness Without Sacrificing Accuracy with Patch Gaussian Augmentationcited by 105
  12. 2019Defense Against Adversarial Attacks Using Feature Scattering-based Adversarial Trainingcited by 95
  13. 2018Improving DNN Robustness to Adversarial Attacks using Jacobian Regularizationcited by 88
  14. 2019On the Connection Between Adversarial Robustness and Saliency Map Interpretabilitycited by 80
  15. 2020On the Loss Landscape of Adversarial Training: Identifying Challenges and How to Overcome Themcited by 59
  16. 2019On the Effectiveness of Low Frequency Perturbationscited by 56
  17. 2019Jacobian Adversarially Regularized Networks for Robustnesscited by 40
  18. 2019What it Thinks is Important is Important: Robustness Transfers through Input Gradientscited by 34
  19. 2021Improving robustness against common corruptions with frequency biased modelscited by 34
  20. 2018On the Structural Sensitivity of Deep Convolutional Networks to the Directions of Fourier Basis Functionscited by 31
  21. 2020Hold me tight! Influence of discriminative features on deep network boundariescited by 31
  22. 2021Impact of Aliasing on Generalization in Deep Convolutional Networkscited by 25
  23. 2020Compounding the Performance Improvements of Assembled Techniques in a Convolutional Neural Networkcited by 24
  24. 2020An Effective Anti-Aliasing Approach for Residual Networkscited by 23
  25. 2020From a Fourier-Domain Perspective on Adversarial Examples to a Wiener Filter Defense for Semantic Segmentationcited by 4

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 1; 2023: 5; 2024: 1; 2025: 3; 2026: 1.

11 papers.

  1. 2023A Survey on the Robustness of Computer Vision Models against Common Corruptionscited by 20
  2. 2023HybridAugment++: Unified Frequency Spectra Perturbations for Model Robustnesscited by 12
  3. 2024Towards Combating Frequency Simplicity-biased Learning for Domain Generalizationcited by 6
  4. 2025Robust Asymmetric Heterogeneous Federated Learning with Corrupted Clientscited by 5
  5. 2023DFM-X: Augmentation by Leveraging Prior Knowledge of Shortcut Learningcited by 4
  6. 2023Fourier Analysis on Robustness of Graph Convolutional Neural Networks for Skeleton-based Action Recognitioncited by 2
  7. 2025FastDINOv2: Frequency Based Curriculum Learning Improves Robustness and Training Speedcited by 2
  8. 2022Aliasing is a Driver of Adversarial Attackscited by 1
  9. 2025Adversarial Examples in Environment Perception for Automated Driving (Review)cited by 1
  10. 2023Exploiting Frequency Spectrum of Adversarial Images for General Robustnesscited by 0
  11. 2026Expanding the Role of Diffusion Models for Robust Classifier Trainingcited by 0