Abstract
Despite the fact that adversarial training has become the de facto method for improving the robustness of deep neural networks, it is well-known that vanilla adversarial training suffers from daunting robust overfitting, resulting in unsatisfactory robust generalization. A number of approaches have been proposed to address these drawbacks such as extra regularization, adversarial weights perturbation, and training with more data over the last few years. However, the robust generalization improvement is yet far from satisfactory. In this paper, we approach this challenge with a brand new perspective – refining historical optimization trajectories. We propose a new method named Weighted Optimization Trajectories (WOT) that leverages the optimization trajectories of adversarial training in time. We have conducted extensive experiments to demonstrate the effectiveness of WOT under various state-of-the-art adversarial attacks. Our results show that WOT integrates seamlessly with the existing adversarial training methods and consistently overcomes the robust overfitting issue, resulting in better adversarial robustness. For example, WOT boosts the robust accuracy of AT-PGD under AA-\(L_{\infty }\) attack by 1.53%–6.11% and meanwhile increases the clean accuracy by 0.55%–5.47% across SVHN, CIFAR-10, CIFAR-100, and Tiny-ImageNet datasets.
T. Huang, S, Liu and T. Chen—These authors contributed equally to this research.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
- 2.
Appendix can be found in https://arxiv.org/pdf/2306.14275v2.pdf.
References
Alayrac, J.B., Uesato, J., Huang, P.S., Fawzi, A., Stanforth, R., Kohli, P.: Are labels required for improving adversarial robustness?. In: Advances in Neural Information Processing Systems 32 (2019)
Andriushchenko, M., Croce, F., Flammarion, N., Hein, M.: Square attack: a query-efficient black-box adversarial attack via random search. In: Vedaldi, A., Bischof, H., Brox, T., Frahm, J.-M. (eds.) ECCV 2020. LNCS, vol. 12368, pp. 484–501. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-58592-1_29
Athalye, A., Carlini, N., Wagner, D.: Obfuscated gradients give a false sense of security: Circumventing defenses to adversarial examples. arXiv preprint arXiv:1802.00420 (2018)
Balaji, Y., Goldstein, T., Hoffman, J.: Instance adaptive adversarial training: Improved accuracy tradeoffs in neural nets. arXiv preprint arXiv:1910.08051 (2019)
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)
Carmon, Y., Raghunathan, A., Schmidt, L., Duchi, J.C., Liang, P.S.: Unlabeled data improves adversarial robustness. In: Advances in Neural Information Processing Systems 32 (2019)
Chen, T., Zhang, Z., Liu, S., Chang, S., Wang, Z.: Robust overfitting may be mitigated by properly learned smoothening. In: International Conference on Learning Representations (2020)
Chen, T., et al.: Sparsity winning twice: Better robust generaliztion from more efficient training. arXiv preprint arXiv:2202.09844 (2022)
Cohen, J.M., Rosenfeld, E., Kolter, J.Z.: Certified adversarial robustness via randomized smoothing. arXiv preprint arXiv:1902.02918 (2019)
Croce, F., Hein, M.: Minimally distorted adversarial examples with a fast adaptive boundary attack. In: International Conference on Machine Learning, pp. 2196–2205. PMLR (2020)
Croce, F., Hein, M.: Reliable evaluation of adversarial robustness with an ensemble of diverse parameter-free attacks. In: International Conference on Machine Learning, pp. 2206–2216. PMLR (2020)
Deng, J., Dong, W., Socher, R., Li, L.J., Li, K., Fei-Fei, L.: Imagenet: a large-scale hierarchical image database. In: 2009 IEEE Conference on Computer Vision and Pattern Recognition, pp. 248–255. IEEE (2009)
Dong, Y., Pang, T., Su, H., Zhu, J.: Evading defenses to transferable adversarial examples by translation-invariant attacks. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 4312–4321 (2019)
Dong, Y., et al.: Exploring memorization in adversarial training. arXiv preprint arXiv:2106.01606 (2021)
Dziugaite, G.K., Roy, D.M.: Computing nonvacuous generalization bounds for deep (stochastic) neural networks with many more parameters than training data. arXiv preprint arXiv:1703.11008 (2017)
Elisseeff, A., Evgeniou, T., Pontil, M., Kaelbing, L.P.: Stability of randomized learning algorithms. J. Mach. Learn. Res. 6(1) (2005)
Engstrom, L., Ilyas, A., Athalye, A.: Evaluating and understanding the robustness of adversarial logit pairing. arXiv preprint arXiv:1807.10272 (2018)
Foret, P., Kleiner, A., Mobahi, H., Neyshabur, B.: Sharpness-aware minimization for efficiently improving generalization. arXiv preprint arXiv:2010.01412 (2020)
Girshick, R., Donahue, J., Darrell, T., Malik, J.: Rich feature hierarchies for accurate object detection and semantic segmentation. In: Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition (2014). https://doi.org/10.1109/CVPR.2014.81
Goodfellow, I.J., Shlens, J., Szegedy, C.: Explaining and Harnessing Adversarial Examples (Dec 2014). arxiv.org/abs/1412.6572
Hardt, M., Recht, B., Singer, Y.: Train faster, generalize better: Stability of stochastic gradient descent. In: International Conference on Machine Learning, pp. 1225–1234. PMLR (2016)
He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 770–778 (2016)
Hinton, G., et al.: Deep neural networks for acoustic modeling in speech recognition. IEEE Signal Process. Mag. (2012). https://doi.org/10.1109/MSP.2012.2205597
Huang, T., Menkovski, V., Pei, Y., Pechenizkiy, M.: Bridging the performance gap between fgsm and pgd adversarial training. arXiv preprint arXiv:2011.05157 (2020)
Huang, T., Menkovski, V., Pei, Y., Pechenizkiy, M.: Calibrated adversarial training. In: Asian Conference on Machine Learning, pp. 626–641. PMLR (2021)
Huang, T., Menkovski, V., Pei, Y., Wang, Y., Pechenizkiy, M.: Direction-aggregated attack for transferable adversarial examples. ACM J. Emerging Technol. Comput. Syst. (JETC) 18(3), 1–22 (2022)
Jiang, Y., Neyshabur, B., Mobahi, H., Krishnan, D., Bengio, S.: Fantastic generalization measures and where to find them. arXiv preprint arXiv:1912.02178 (2019)
Krizhevsky, A., Nair, V., Hinton, G.: Cifar-10 (canadian institute for advanced research). https://www.cs.toronto.edu/~kriz/cifar.html (2010)
Madry, A., Makelov, A., Schmidt, L., Tsipras, D., Vladu, A.: Towards deep learning models resistant to adversarial attacks. arXiv preprint arXiv:1706.06083 (2017)
Madry, A., Makelov, A., Schmidt, L., Tsipras, D., Vladu, A.: Towards Deep Learning Models Resistant to Adversarial Attacks, arxiv.org/abs/1706.06083 (Jun 2017)
Moosavi-Dezfooli, S.M., Fawzi, A., Frossard, P.: DeepFool: a simple and accurate method to fool deep neural networks. In: Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, vol. 2016-Decem, pp. 2574–2582. IEEE Computer Society (Dec 2016). https://doi.org/10.1109/CVPR.2016.282
Moosavi-Dezfooli, S.M., Fawzi, A., Uesato, J., Frossard, P.: Robustness via curvature regularization, and vice versa, arXiv: 1811.09716 (Nov 2018)
Netzer, Y., Wang, T., Coates, A., Bissacco, A., Wu, B., Ng, A.Y.: Reading digits in natural images with unsupervised feature learning (2011)
Pang, T., Lin, M., Yang, X., Zhu, J., Yan, S.: Robustness and accuracy could be reconcilable by (proper) definition. arXiv preprint arXiv:2202.10103 (2022)
Qin, C., et al.: Adversarial Robustness through Local Linearization, arXiv: 1907.02610 (Jul 2019) ’
Rebuffi, S.A., Gowal, S., Calian, D.A., Stimberg, F., Wiles, O., Mann, T.: Fixing data augmentation to improve adversarial robustness. arXiv preprint arXiv:2103.01946 (2021)
Rice, L., Wong, E., Kolter, Z.: Overfitting in adversarially robust deep learning. In: International Conference on Machine Learning, pp. 8093–8104. PMLR (2020)
Ross, A., Doshi-Velez, F.: Improving the adversarial robustness and interpretability of deep neural networks by regularizing their input gradients. In: Proceedings of the AAAI Conference on Artificial Intelligence, vol. 32 (2018)
Schmidt, L., Santurkar, S., Tsipras, D., Talwar, K., Madry, A.: Adversarially robust generalization requires more data. In: Advances in Neural Information Processing Systems, pp. 5014–5026 (2018)
Sehwag, V., et al.: Robust learning meets generative models: Can proxy distributions improve adversarial robustness? arXiv preprint arXiv:2104.09425 (2021)
Simonyan, K., Zisserman, A.: Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556 (2014)
Singla, V., Singla, S., Feizi, S., Jacobs, D.: Low curvature activations reduce overfitting in adversarial training. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 16423–16433 (2021)
Stutz, D., Hein, M., Schiele, B.: Relating adversarially robust generalization to flat minima. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 7807–7817 (2021)
Szegedy, C., et al.: Intriguing properties of neural networks, arXiv: 1312.6199 (dec 2013)
Uesato, J., O’donoghue, B., Kohli, P., Oord, A.: Adversarial risk and the dangers of evaluating against weak attacks. In: International Conference on Machine Learning, pp. 5025–5034. PMLR (2018)
Wang, Y., Zou, D., Yi, J., Bailey, J., Ma, X., Gu, Q.: Improving adversarial robustness requires revisiting misclassified examples. In: International Conference on Learning Representations (2020)
Wu, D., Wang, Y., Xia, S.T.: Revisiting loss landscape for adversarial robustness. arXiv preprint arXiv:2004.05884 (2020)
Wu, D., Xia, S.T., Wang, Y.: Adversarial weight perturbation helps robust generalization. In: Advances in Neural Information Processing Systems 33 (2020)
Xie, C., et al.: Improving transferability of adversarial examples with input diversity. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 2730–2739 (2019)
Yu, C., et al.: Robust weight perturbation for adversarial training (2021)
Zagoruyko, S., Komodakis, N.: Wide residual networks. arXiv preprint arXiv:1605.07146 (2016)
Zhang, C., et al.: Musings on deep learning: Properties of sgd. Tech. rep, Center for Brains, Minds and Machines (CBMM) (2017)
Zhang, H., Yu, Y., Jiao, J., Xing, E., El Ghaoui, L., Jordan, M.: Theoretically principled trade-off between robustness and accuracy. In: International Conference on Machine Learning, pp. 7472–7482. PMLR (2019)
Zhang, J., et a;.: Attacks which do not kill training make adversarial learning stronger. In: International Conference on Machine Learning, pp. 11278–11287. PMLR (2020)
Zhang, J., Zhu, J., Niu, G., Han, B., Sugiyama, M., Kankanhalli, M.: Geometry-aware instance-reweighted adversarial training. arXiv preprint arXiv:2010.01736 (2020)
Zhou, Y., Liang, Y., Zhang, H.: Generalization error bounds with probabilistic guarantee for sgd in nonconvex optimization. arXiv preprint arXiv:1802.06903 (2018)
Acknowledgement
This work is partially supported by NWO EDIC and EDF KOIOS projects. Part of this work used the Dutch national e-infrastructure with the support of the SURF Cooperative using grant no. NWO2021.060, EINF-5587 and EINF-5587/L1. We would like to express our deepest gratitude to the anonymous reviewers whose insightful comments and suggestions significantly improved the quality of this paper.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Ethics declarations
Ethical statement
Our proposed method, Weighted Optimization Trajectories (WOT), enhances the robustness of deep neural networks against adversarial attacks. While the primary goal is to improve AI system security, it is crucial to consider the ethical implications of our work:
− Personal Data Protection: Researchers and practitioners must ensure proper data handling, privacy, and compliance with data protection laws when working with personal data
− Privacy Preservation: Improved robustness could inadvertently increase the capacity to infer personal information from data. Privacy-preserving techniques, like differential privacy, should be employed to mitigate these risks.
Rights and permissions
Copyright information
© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Huang, T. et al. (2023). Enhancing Adversarial Training via Reweighting Optimization Trajectory. In: Koutra, D., Plant, C., Gomez Rodriguez, M., Baralis, E., Bonchi, F. (eds) Machine Learning and Knowledge Discovery in Databases: Research Track. ECML PKDD 2023. Lecture Notes in Computer Science(), vol 14169. Springer, Cham. https://doi.org/10.1007/978-3-031-43412-9_7
Download citation
DOI: https://doi.org/10.1007/978-3-031-43412-9_7
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-43411-2
Online ISBN: 978-3-031-43412-9
eBook Packages: Computer ScienceComputer Science (R0)