mlsanityData to 7 Oct 2026

Adversarially Optimized Mixup for Robust Classification

arXiv 2103.11589 ↗First posted 22 Mar 2021cs.LG cs.CVcited by 7

Not on the map. Why?

Abstract (arXiv)

Mixup is a procedure for data augmentation that trains networks to make smoothly interpolated predictions between datapoints. Adversarial training is a strong form of data augmentation that optimizes for worst-case predictions in a compact space around each data-point, resulting in neural networks that make much more robust predictions. In this paper, we bring these ideas together by adversarially probing the space between datapoints, using projected gradient descent (PGD). The fundamental approach in this work is to leverage backpropagation through the mixup interpolation during training to optimize for places where the network makes unsmooth and incongruous predictions. Additionally, we also explore several modifications and nuances, like optimization of the mixup ratio and geometrical label assignment, and discuss their impact on enhancing network robustness. Through these ideas, we have been able to train networks that robustly generalize better; experiments on CIFAR-10 and CIFAR-100 demonstrate consistent improvements in accuracy against strong adversaries, including the recent strong ensemble attack AutoAttack. Our source code would be released for reproducibility.

Key facts. Cited by 7 papers in the map (data through 7 Oct 2026). Builds on 24 of its 27 references that are in the map. First posted on arXiv 22 Mar 2021.

Builds on

24 of 27 references.

References · 24

  1. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
  2. 2014Explaining and Harnessing Adversarial Examplescited by 7,287
  3. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,326
  4. 2017mixup: Beyond Empirical Risk Minimizationcited by 5,155
  5. 2013Intriguing properties of neural networkscited by 5,114
  6. 2016Wide Residual Networkscited by 3,057
  7. 2016Identity Mappings in Deep Residual Networkscited by 1,960
  8. 2019Theoretically Principled Trade-off between Robustness and Accuracycited by 1,510
  9. 2020Reliable evaluation of adversarial robustness with an ensemble of diverse parameter-free attackscited by 1,442
  10. 2019Natural Adversarial Examplescited by 1,401
  11. 2018Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examplescited by 1,259
  12. 2019Adversarial Examples Are Not Bugs, They Are Featurescited by 950
12 more
  1. 2020Fast is better than free: Revisiting adversarial trainingcited by 723
  2. 2019Square Attack: a query-efficient black-box adversarial attack via random searchcited by 682
  3. 2020On Adaptive Attacks to Adversarial Example Defensescited by 547
  4. 2020Overfitting in adversarially robust deep learningcited by 504
  5. 2019On Evaluating Adversarial Robustnesscited by 442
  6. 2019Adversarial Examples Improve Image Recognitioncited by 283
  7. 2018MixUp as Locally Linear Out-Of-Manifold Regularizationcited by 158
  8. 2019When NAS Meets Robustness: In Search of Robust Architectures against Adversarial Attackscited by 87
  9. 2020Adversarial Vertex Mixup: Toward Better Adversarially Robust Generalizationcited by 73
  10. 2019Mixup Inference: Better Exploiting Mixup to Defend Adversarial Attackscited by 54
  11. 2019Interpolated Adversarial Training: Achieving Robust Neural Networks without Sacrificing Too Much Accuracycited by 49
  12. 2019MixUp as Directional Adversarial Trainingcited by 14

Built on by

Cited by year

Citing papers per year: 2021: 1; 2022: 3; 2023: 0; 2024: 2; 2025: 1.

7 papers.

  1. 2022Sparsity Winning Twice: Better Robust Generalization from More Efficient Trainingcited by 41
  2. 2022A Survey of Mix-based Data Augmentation: Taxonomy, Methods, Applications, and Explainabilitycited by 30
  3. 2024A Survey on Mixup Augmentations and Beyondcited by 18
  4. 2022TGDM: Target Guided Dynamic Mixup for Cross-Domain Few-Shot Learningcited by 16
  5. 2021MixACM: Mixup-Based Robustness Transfer via Distillation of Activated Channel Mapscited by 9
  6. 2024CIMGEN: Controlled Image Manipulation by Finetuning Pretrained Generative Models on Limited Datacited by 0
  7. 2025Enlightenment Period Improving DNN Performancecited by 0