Advertisement

Sparse Adversarial Attack via Perturbation Factorization

Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 12367)

Abstract

This work studies the sparse adversarial attack, which aims to generate adversarial perturbations onto partial positions of one benign image, such that the perturbed image is incorrectly predicted by one deep neural network (DNN) model. The sparse adversarial attack involves two challenges, i.e., where to perturb, and how to determine the perturbation magnitude. Many existing works determined the perturbed positions manually or heuristically, and then optimized the magnitude using a proper algorithm designed for the dense adversarial attack. In this work, we propose to factorize the perturbation at each pixel to the product of two variables, including the perturbation magnitude and one binary selection factor (i.e., 0 or 1). One pixel is perturbed if its selection factor is 1, otherwise not perturbed. Based on this factorization, we formulate the sparse attack problem as a mixed integer programming (MIP) to jointly optimize the binary selection factors and continuous perturbation magnitudes of all pixels, with a cardinality constraint on selection factors to explicitly control the degree of sparsity. Besides, the perturbation factorization provides the extra flexibility to incorporate other meaningful constraints on selection factors or magnitudes to achieve some desired performance, such as the group-wise sparsity or the enhanced visual imperceptibility. We develop an efficient algorithm by equivalently reformulating the MIP problem as a continuous optimization problem. Extensive experiments demonstrate the superiority of the proposed method over several state-of-the-art sparse attack methods. The implementation of the proposed method is available at https://github.com/wubaoyuan/Sparse-Adversarial-Attack.

Keywords

Perturbation factorization Sparse adversarial attack Mixed integer programming 

Notes

Acknowledgement

This work is supported by Tencent AI Lab. The participation of Yujiu Yang is supported by The Key Program of National Natural Science Foundation of China under Grant No. U1903213.

Supplementary material

504482_1_En_3_MOESM1_ESM.pdf (344 kb)
Supplementary material 1 (pdf 344 KB)

References

  1. 1.
    Achanta, R., Shaji, A., Smith, K., Lucchi, A., Fua, P., Süsstrunk, S.: Slic superpixels compared to state-of-the-art superpixel methods. IEEE Trans. Pattern Anal. Mach. Intell. 34(11), 2274–2282 (2012)CrossRefGoogle Scholar
  2. 2.
    Akhtar, N., Mian, A.: Threat of adversarial attacks on deep learning in computer vision: a survey. IEEE Access 6, 14410–14430 (2018)CrossRefGoogle Scholar
  3. 3.
    Bach, F., Jenatton, R., Mairal, J., Obozinski, G., et al.: Optimization with sparsity-inducing penalties. Found. Trends® Mach. Learn. 4(1), 1–106 (2012)Google Scholar
  4. 4.
    Bai, J., et al.: Targeted attack for deep hashing based retrieval. In: ECCV (2020)Google Scholar
  5. 5.
    Bibi, A., Wu, B., Ghanem, B.: Constrained k-means with general pairwise and cardinality constraints. arXiv preprint arXiv:1907.10410 (2019)
  6. 6.
    Boyd, S., Parikh, N., Chu, E., Peleato, B., Eckstein, J., et al.: Distributed optimization and statistical learning via the alternating direction method of multipliers. Found. Trends® Mach. Learn. 3(1), 1–122 (2011)Google Scholar
  7. 7.
    Carlini, N., Wagner, D.: Towards evaluating the robustness of neural networks. In: 2017 IEEE Symposium on Security and Privacy (SP), pp. 39–57. IEEE (2017)Google Scholar
  8. 8.
    Chen, H., Zhang, H., Chen, P.Y., Yi, J., Hsieh, C.J.: Show-and-fool: crafting adversarial examples for neural image captioning. arXiv preprint arXiv:1712.02051 (2017)
  9. 9.
    Chen, W., Zhang, Z., Hu, X., Wu, B.: Boosting decision-based black-box adversarial attacks with random sign flip. In: Proceedings of the European Conference on Computer Vision (2020)Google Scholar
  10. 10.
    Croce, F., Hein, M.: Sparse and imperceivable adversarial attacks. In: ICCV, pp. 4724–4732 (2019)Google Scholar
  11. 11.
    Deng, J., Dong, W., Socher, R., Li, L.J., Li, K., Fei-Fei, L.: ImageNet: a large-scale hierarchical image database. In: CVPR, pp. 248–255. IEEE (2009)Google Scholar
  12. 12.
    Dong, Y., et al.: Efficient decision-based black-box adversarial attacks on face recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 7714–7722 (2019)Google Scholar
  13. 13.
    Feng, Y., Chen, B., Dai, T., Xia, S.: Adversarial attack on deep product quantization network for image retrieval. In: AAAI (2020)Google Scholar
  14. 14.
    Feng, Y., Wu, B., Fan, Y., Li, Z., Xia, S.: Efficient black-box adversarial attack guided by the distribution of adversarial perturbations. arXiv preprint arXiv:2006.08538 (2020)
  15. 15.
    Goodfellow, I.J., Shlens, J., Szegedy, C.: Explaining and harnessing adversarial examples. In: ICLR (2015)Google Scholar
  16. 16.
    He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: CVPR, pp. 770–778 (2016)Google Scholar
  17. 17.
    Karmon, D., Zoran, D., Goldberg, Y.: LaVAN: localized and visible adversarial noise. In: ICML (2018)Google Scholar
  18. 18.
    Krizhevsky, A., Hinton, G.: Learning multiple layers of features from tiny images. Technical report, Citeseer (2009)Google Scholar
  19. 19.
    Legge, G.E., Foley, J.M.: Contrast masking in human vision. JOSA 70(12), 1458–1471 (1980)CrossRefGoogle Scholar
  20. 20.
    Li, T., Wu, B., Yang, Y., Fan, Y., Zhang, Y., Liu, W.: Compressing convolutional neural networks via factorized convolutional filters. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 3977–3986 (2019)Google Scholar
  21. 21.
    Li, Y., et al.: Toward adversarial robustness via semi-supervised robust training. arXiv preprint arXiv:2003.06974 (2020)
  22. 22.
    Li, Y., Yang, X., Wu, B., Lyu, S.: Hiding faces in plain sight: Disrupting AI face synthesis with adversarial perturbations. arXiv preprint arXiv:1906.09288 (2019)
  23. 23.
    Liu, A., et al.: Spatiotemporal attacks for embodied agents. In: European Conference on Computer Vision (2020)Google Scholar
  24. 24.
    Liu, A., et al.: Perceptual-sensitive GAN for generating adversarial patches. In: 33rd AAAI Conference on Artificial Intelligence (2019)Google Scholar
  25. 25.
    Liu, A., Wang, J., Liu, X., Cao, b., Zhang, C., Yu, H.: Bias-based universal adversarial patch attack for automatic check-out. In: European Conference on Computer Vision (2020)Google Scholar
  26. 26.
    Luo, B., Liu, Y., Wei, L., Xu, Q.: Towards imperceptible and robust adversarial example attacks against neural networks. In: AAAI (2018)Google Scholar
  27. 27.
    Modas, A., Moosavi-Dezfooli, S.M., Frossard, P.: Sparsefool: a few pixels make a big difference. In: CVPR, pp. 9087–9096 (2019)Google Scholar
  28. 28.
    Papernot, N., McDaniel, P., Jha, S., Fredrikson, M., Celik, Z.B., Swami, A.: The limitations of deep learning in adversarial settings. In: 2016 IEEE European Symposium on Security and Privacy (EuroS&P), pp. 372–387. IEEE (2016)Google Scholar
  29. 29.
    Sarikaya, R., Hinton, G.E., Deoras, A.: Application of deep belief networks for natural language understanding. IEEE/ACM Trans. Audio Speech Lang. Process. (TASLP) 22(4), 778–784 (2014)Google Scholar
  30. 30.
    Su, J., Vargas, D.V., Sakurai, K.: One pixel attack for fooling deep neural networks. IEEE Trans. Evol. Comput. 23, 828–841 (2019)CrossRefGoogle Scholar
  31. 31.
    Sun, Y., Liang, D., Wang, X., Tang, X.: DeepID3: face recognition with very deep neural networks. arXiv preprint arXiv:1502.00873 (2015)
  32. 32.
    Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., Wojna, Z.: Rethinking the inception architecture for computer vision. In: CVPR, pp. 2818–2826 (2016)Google Scholar
  33. 33.
    Szegedy, C., et al.: Intriguing properties of neural networks. In: ICLR (2014)Google Scholar
  34. 34.
    Wang, H., et al.: CosFace: large margin cosine loss for deep face recognition. In: CVPR, pp. 5265–5274 (2018)Google Scholar
  35. 35.
    Wu, B., et al.: Tencent ML-images: a large-scale multi-label image database for visual representation learning. IEEE Access 7, 172683–172693 (2019)CrossRefGoogle Scholar
  36. 36.
    Wu, B., Ghanem, B.: \(l_p\)-box ADMM: a versatile framework for integer programming. IEEE Trans. Pattern Anal. Mach. Intell. (TPAMI) 41, 1695–1708 (2018)CrossRefGoogle Scholar
  37. 37.
    Wu, B., Shen, L., Zhang, T., Ghanem, B.: Map inference via \(l_2\)-sphere linear program reformulation. Int. J. Comput. Vis. 128, 1–24 (2020)CrossRefGoogle Scholar
  38. 38.
    Xu, K., et al.: Structured adversarial attack: Towards general implementation and better interpretability. In: ICLR (2019)Google Scholar
  39. 39.
    Xu, Y., et al.: Exact adversarial attack to image captioning via structured output learning with latent variables. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 4135–4144 (2019)Google Scholar
  40. 40.
    Yuan, M., Lin, Y.: Model selection and estimation in regression with grouped variables. J. R. Stat. Soc.: Ser. B (Stat. Methodol.) 68(1), 49–67 (2006)MathSciNetCrossRefGoogle Scholar
  41. 41.
    Zhao, P., Liu, S., Wang, Y., Lin, X.: An ADMM-based universal framework for adversarial attacks on deep neural networks. In: 2018 ACMMM, pp. 1065–1073. ACM (2018)Google Scholar

Copyright information

© Springer Nature Switzerland AG 2020

Authors and Affiliations

  1. 1.Tencent AI LabShenzhenChina
  2. 2.Tsinghua Shenzhen International Graduate SchoolTsinghua UniversityBeijingChina

Personalised recommendations