Page 50 - 《软件学报》2026年第4期

翟浩杰 等: 基于自适应策略优化的鲁棒泛化权衡学习 1491

[19] Andriushchenko M, Croce F, Flammarion N, Hein M. Square attack: A query-efficient black-box adversarial attack via random search.
In: Proc. of the 16th European Conf. on Computer Vision. Glasgow: Springer, 2020. 484–501. [doi: 10.1007/978-3-030-58592-1_29]
[20] Wu DX, Xia ST, Wang YS. Adversarial weight perturbation helps robust generalization. In: Proc. of the 34th Conf. on Neural
Information Processing Systems. Vancouver: Curran Associates Inc., 2020. 2958–2969. [doi: 10.5555/3495724.3495973]
[21] Qian YG, Liang XY, Kang M, Wang B, Gu ZQ, Wang X, Wu CM. GAAT: Group adaptive adversarial training to improve the trade-off
between robustness and accuracy. Int’l Journal of Pattern Recognition and Artificial Intelligence, 2022, 36(13): 2251015. [doi: 10.1142/
S0218001422510156]
[22] Wang Y, Zhang DM, Zhang HY. Weighted adaptive perturbations adversarial training for improving robustness. In: Proc. of the 19th
Pacific Rim Int’l Conf. on Artificial Intelligence. Shanghai: Springer, 2022. 402–415. [doi: 10.1007/978-3-031-20865-2_30]
[23] Yu CJ, Zhou DW, Shen L, Yu J, Han B, Gong MM, Wang NN, Liu TL. Strength-adaptive adversarial training. arXiv:2210.01288, 2022.
[24] Xiao JC, Qin ZY, Fan YB, Wu BY, Wang J, Luo ZQ. Adaptive smoothness-weighted adversarial training for multiple perturbations with
its stability analysis. arXiv:2210.00557, 2022.
[25] Wang R, Li YX, Liu SJ. Robust mode connectivity-oriented adversarial defense: Enhancing neural network robustness against diversified
l p attacks. arXiv:2303.10225, 2023.
[26] Huang ZC, Fan YB, Liu C, Zhang WZ, Zhang Y, Salzmann M, Süsstrunk S, Wang J. Fast adversarial training with adaptive step size.
IEEE Trans. on Image Processing, 2023, 32: 6102–6114. [doi: 10.1109/TIP.2023.3326398]
[27] Ma LH, Liang L. Improving adversarial robustness of deep neural networks via adaptive margin evolution. Neurocomputing, 2023, 551:
126524. [doi: 10.1016/j.neucom.2023.126524]
[28] Guo RY, Chen QY, Liu H, Wang WQ. Adversarial robustness enhancement for deep learning-based soft sensors: An adversarial training
strategy using historical gradients and domain adaptation. Sensors, 2024, 24(12): 3909. [doi: 10.3390/s24123909]
[29] Hardy I, Yetukuri J, Liu Y. Adaptive adversarial training does not increase recourse costs. In: Proc. of the 2023 AAAI/ACM Conf. on AI,
Ethics, and Society. Montréal: ACM, 2023. 432–442. [doi: 10.1145/3600211.3604704]
[30] Su JW, Vargas DV, Sakurai K. One pixel attack for fooling deep neural networks. IEEE Trans. on Evolutionary Computation, 2019,
23(5): 828–841. [doi: 10.1109/TEVC.2019.2890858]
[31] Bai ZX, Wang HJ. Adversarial example generation method based on improved genetic algorithm. Computer Engineering, 2023, 49(5):
139–149 (in Chinese with English abstract). [doi: 10.19678/j.issn.1000-3428.0065260]
[32] Williams PN, Li K. Black-box sparse adversarial attack via multi-objective optimisation. In: Proc. of the 2023 IEEE/CVF Conf. on
Computer Vision and Pattern Recognition. Vancouver: IEEE, 2023. 12291–12301. [doi: 10.1109/CVPR52729.2023.01183]
[33] Jia XJ, Zhang Y, Wu BY, Ma K, Wang J, Cao XC. LAS-AT: Adversarial training with learnable attack strategy. In: Proc. of the 2022
IEEE/CVF Conf. on Computer Vision and Pattern Recognition. New Orleans: IEEE, 2022. 13388–13398. [doi: 10.1109/CVPR52688.
2022.01304]
[34] Zhang JF, Xu XL, Han B, Niu G, Cui LZ, Sugiyama M, Kankanhalli M. Attacks which do not kill training make adversarial learning
stronger. In: Proc. of the 37th Int’l Conf. on Machine Learning. New York: JMLR.org, 2020. 11278–11287. [doi: 10.5555/3524938.
3525984]
[35] Zhang JF, Zhu JN, Niu G, Han B, Sugiyama M, Kankanhalli M. Geometry-aware instance-reweighted adversarial training. arXiv:
2010.01736, 2020.
[36] Wang QZ, Liu F, Han B, Liu TL, Gong C, Niu G, Zhou MY, Sugiyama M. Probabilistic margins for instance reweighting in adversarial
training. In: Proc. of the 35th Conf. on Neural Information Proc. Systems. 2021. 23258–23269. [doi: 10.48550/arXiv.2106.07904]
[37] Zhu KJ, Hu XX, Wang JD, Xie X, Yang G. Improving generalization of adversarial training via robust critical fine-tuning. In: Proc. of the
2023 IEEE/CVF Int’l Conf. on Computer Vision. Paris: IEEE, 2023. 4401–4411. [doi: 10.1109/ICCV51070.2023.00408]
[38] van der Maaten L, Hinton G. Visualizing data using t-SNE. Journal of Machine Learning Research, 2008, 9(86): 2579–2605.

附中文参考文献
[31] 白祉旭, 王衡军. 基于改进遗传算法的对抗样本生成方法. 计算机工程, 2023, 49(5): 139–149. [doi: 10.19678/j.issn.1000-3428.
0065260]

作者简介
翟浩杰, 硕士生, 主要研究领域为深度学习.
王冉, 博士, 副教授, 主要研究领域为机器学习, 深度学习, 不确定性建模.
邬文慧, 博士, 副教授, 主要研究领域为机器学习, 图像增强, 网络数据聚类.
贾育衡, 博士, 副教授, CCF 高级会员, 主要研究领域为机器学习, 数据表示, 半监督学习.