On the Existence of Consistent Adversarial Attacks in High-Dimensional Linear Classification
Not on the map. Why?
Abstract (arXiv)
What fundamentally distinguishes an adversarial attack from a misclassification due to limited model expressivity or finite data? In this work, we investigate this question in the setting of high-dimensional binary classification, where statistical effects due to limited data availability play a central role. We introduce a new error metric that precisely capture this distinction, quantifying model vulnerability to consistent adversarial attacks -- perturbations that preserve the ground-truth labels. Our main technical contribution is an exact and rigorous asymptotic characterization of these metrics in both well-specified models and latent space models, revealing different vulnerability patterns compared to standard robust error measures. The theoretical results demonstrate that as models become more overparameterized, their vulnerability to label-preserving perturbations grows, offering theoretical insight into the mechanisms underlying model sensitivity to adversarial attacks.
Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 56 of its 91 references that are in the map. First posted on arXiv 14 Jun 2025.
Builds on
56 of 91 references.
References · 56
- 2014Explaining and Harnessing Adversarial Examplescited by 7,364
- 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,415
- 2013Intriguing properties of neural networkscited by 5,155
- 2017Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithmscited by 4,349
- 2021Convex optimizationcited by 2,515
- 2019Benign Overfitting in Linear Regressioncited by 549
- 2019Surprises in High-Dimensional Ridgeless Least Squares Interpolationcited by 513
- 2020Overfitting in adversarially robust deep learningcited by 506
- 2019The generalization error of random features regression: Precise asymptotics and double descent curvecited by 413
- 2023Better Diffusion Models Further Improve Adversarial Trainingcited by 208
- 2020Spectrum Dependent Learning Curves in Kernel Regression and Wide Neural Networkscited by 200
- 2020Understanding and Mitigating the Tradeoff Between Robustness and Accuracycited by 149
38 more
- 2018Rademacher Complexity for Adversarially Robust Generalizationcited by 132
- 2020Generalisation error in learning with random features and the hidden manifold modelcited by 127
- 2021Learning curves of generic features maps for realistic datasets with a teacher-student modelcited by 126
- 2017Optimal Errors and Phase Transitions in High-Dimensional Generalized Linear Modelscited by 116
- 2020The Gaussian equivalence of generative models for learning with shallow neural networkscited by 113
- 2022More Than a Toy: Random Matrix Models Predict How Real-World Neural Representations Generalizecited by 91
- 2020Precise Tradeoffs in Adversarial Training for Linear Regressioncited by 77
- 2019Adversarially Robust Generalization Just Requires More Unlabeled Datacited by 76
- 2022Universality of empirical risk minimizationcited by 75
- 2020A Random Matrix Analysis of Random Fourier Features: Beyond the Gaussian Kernel, a Precise Phase Transition, and the Corresponding Double Descentcited by 70
- 2020The role of regularization in classification of high-dimensional noisy Gaussian mixturecited by 63
- 2021Learning Gaussian Mixtures with Generalised Linear Models: Precise Asymptotics in High-dimensionscited by 54
- 2020Generalization error in high-dimensional perceptrons: Approaching Bayes error with convex optimizationcited by 51
- 2022Dimension free ridge regressioncited by 51
- 2020Precise Statistical Analysis of Classification Accuracies for Adversarial Trainingcited by 50
- 2020Adversarial Learning Guarantees for Linear Hypotheses and Neural Networkscited by 47
- 2022The curse of overparametrization in adversarial training: Precise analysis of robust generalization for random features regressioncited by 41
- 2023Deterministic equivalent and error universality of deep random features learningcited by 35
- 2020Sharp Statistical Guarantees for Adversarially Robust Gaussian Classificationcited by 28
- 2023Universality laws for Gaussian mixtures in generalized linear modelscited by 28
- 2022Gaussian Universality of Perceptrons with Random Labelscited by 26
- 2024Dimension-free deterministic equivalents and scaling laws for random feature regressioncited by 23
- 2022Overparameterized Linear Regression under Adversarial Attackscited by 21
- 2022Theoretical characterization of uncertainty in high-dimensional linear classificationcited by 18
- 2023Regularization properties of adversarially-trained linear regressioncited by 18
- 2022Error Scaling Laws for Kernel Classification under Source and Capacity Conditionscited by 17
- 2022On double-descent in uncertainty quantification in overparametrized modelscited by 16
- 2021Interpolation can hurt robust generalization even when there is no noisecited by 15
- 2023High-dimensional robust regression under heavy-tailed data: Asymptotics and Universalitycited by 15
- 2022Why adversarial training can hurt robust accuracycited by 14
- 2025Asymptotics of Non-Convex Generalized Linear Models in High-Dimensions: A proof of the replica formulacited by 14
- 2024Asymptotics of Learning with Deep Structured (Random) Featurescited by 13
- 2024Ensemble everything everywhere: Multi-scale aggregation for adversarial robustnesscited by 12
- 2023Classification of Heavy-tailed Features in High Dimensions: a Superstatistical Approachcited by 11
- 2023Reliable learning in challenging environmentscited by 10
- 2024A High Dimensional Statistical Model for Adversarial Training: Geometry and Trade-Offscited by 7
- 2024Efficient Optimization Algorithms for Linear Adversarial Trainingcited by 6
- 2025Computational Thresholds in Multi-Modal Learning via the Spiked Matrix-Tensor Modelcited by 6
50 most cited of 56 All 56 (JSON)
Built on by
Citing papers per year: 2025: 1; 2026: 1.
2 papers.