Machine learningTraining techniques
对抗训练
对抗训练是一种针对深度神经网络的鲁棒优化过程，模型训练时不仅使用干净数据，还使用在训练过程中精心构造的最坏情况扰动输入。Madry et al. (2018) 将其形式化为一个最小-最大鞍点问题，该方法在每次梯度更新之前，使用投影梯度下降（PGD）在有界 Lp 扰动集内生成强对抗样本，从而迫使网络学习在此类扰动下保持稳定的决策边界。
阅读完整方法
仅限会员
登录使用免费账户登录即可阅读本节。
Method map
The neighbourhood of related methods — select a node to explore.
来源
- Madry, A., Makelov, A., Schmidt, L., Tsipras, D., & Vladu, A. (2018). Towards deep learning models resistant to adversarial attacks. International Conference on Learning Representations (ICLR). link ↗
如何引用本页
ScholarGate. (2026, June 2). Adversarial Training (Robust Optimization for DL). ScholarGate. https://scholargate.app/zh/deep-learning/adversarial-training
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
Compare side by side →