mlsanityData to 14 Oct 2026

Exploring Geometry of Blind Spots in Vision Models

arXiv 2310.19889 ↗First posted 30 Oct 2023cs.CV cs.AI cs.LGcited by 2

Not on the map. Why?

Abstract (arXiv)

Despite the remarkable success of deep neural networks in a myriad of settings, several works have demonstrated their overwhelming sensitivity to near-imperceptible perturbations, known as adversarial attacks. On the other hand, prior works have also observed that deep networks can be under-sensitive, wherein large-magnitude perturbations in input space do not induce appreciable changes to network activations. In this work, we study in detail the phenomenon of under-sensitivity in vision models such as CNNs and Transformers, and present techniques to study the geometry and extent of "equi-confidence" level sets of such networks. We propose a Level Set Traversal algorithm that iteratively explores regions of high confidence with respect to the input space using orthogonal components of the local gradients. Given a source image, we use this algorithm to identify inputs that lie in the same equi-confidence level set as the source image despite being perceptually similar to arbitrary images from other classes. We further observe that the source image is linearly connected by a high-confidence path to these inputs, uncovering a star-like structure for level sets of deep networks. Furthermore, we attempt to identify and estimate the extent of these connected higher-dimensional regions over which the model maintains a high degree of confidence. The code for this project is publicly available at https://github.com/SriramB-98/blindspots-neurips-sub

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 34 of its 38 references that are in the map. First posted on arXiv 30 Oct 2023.

Builds on

34 of 38 references.

References · 34

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  3. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,305
  4. 2018The Unreasonable Effectiveness of Deep Features as a Perceptual Metriccited by 11,177
  5. 2014Explaining and Harnessing Adversarial Examplescited by 7,364
  6. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,415
  7. 2013Intriguing properties of neural networkscited by 5,155
  8. 2020Training data-efficient image transformers & distillation through attentioncited by 4,586
  9. 2015Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classificationcited by 3,611
  10. 2016Towards Evaluating the Robustness of Neural Networkscited by 2,998
  11. 2016Adversarial examples in the physical worldcited by 1,769
  12. 2018Glow: Generative Flow with Invertible 1x1 Convolutionscited by 1,545
22 more
  1. 2019Theoretically Principled Trade-off between Robustness and Accuracycited by 1,517
  2. 2018Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examplescited by 1,268
  3. 2016Adversarial Machine Learning at Scalecited by 936
  4. 2017Ensemble Adversarial Training: Attacks and Defensescited by 840
  5. 2020RobustBench: a standardized adversarial robustness benchmarkcited by 661
  6. 2019On Evaluating Adversarial Robustnesscited by 447
  7. 2019Minimally distorted Adversarial Examples with a Fast Adaptive Boundary Attackcited by 334
  8. 2020Do Adversarially Robust ImageNet Models Transfer Better?cited by 316
  9. 2019Towards Stable and Efficient Training of Verifiably Robust Neural Networkscited by 205
  10. 2018i-RevNet: Deep Invertible Networkscited by 137
  11. 2020Perceptual Adversarial Robustness: Defense Against Unseen Threat Modelscited by 137
  12. 2019Wasserstein Adversarial Examples via Projected Sinkhorn Iterationscited by 89
  13. 2023Revisiting Adversarial Training for ImageNet: Architectures, Training and Generalization across Threat Modelscited by 79
  14. 2018Excessive Invariance Causes Adversarial Vulnerabilitycited by 66
  15. 2020Guided Adversarial Attack for Evaluating and Enhancing Adversarial Defensescited by 62
  16. 2015Adversarial Manipulation of Deep Representationscited by 56
  17. 2020Fundamental Tradeoffs between Invariance and Sensitivity to Adversarial Perturbationscited by 55
  18. 2019An Alternative Surrogate Loss for PGD-based Adversarial Testingcited by 45
  19. 2020Second-Order Provable Defenses against Adversarial Attackscited by 40
  20. 2021Improved, Deterministic Smoothing for L_1 Certified Robustnesscited by 39
  21. 2021Adversarial Robustness against Multiple and Single lp-Threat Models via Quick Fine-Tuning of Robust Classifierscited by 23
  22. 2020Span Recovery for Deep Neural Networks with Applications to Input Obfuscationcited by 4

Built on by

Cited by year

Citing papers per year: 2024: 2.

2 papers.

  1. 2024Input Space Mode Connectivity in Deep Neural Networkscited by 4
  2. 2024A singular Riemannian Geometry Approach to Deep Neural Networks III. Piecewise Differentiable Layers and Random Walks on n-dimensional Classescited by 2