/Awesome-Adversarial-Training

Adversarial Training paper list

Awesome-Adversarial-Training

Here is a paper list about Adversarial Training

Research Groups

MIT madry Lab
Bo Li
Jun Zhu
Quanshi Zhang

Survey

Recent Advances in Adversarial Training for Adversarial Robustness[paper](IJCAI2021)

Classic papers

(FGSM)Explaining and Harnessing Adversarial Examples[paper](ICLR2015)
(PGD)Towards deep learning models resistant to adversarial attacks[paper](ICLR2018)
(TRADES)Theoretically Principled Trade-off between Robustness and Accuracy[paper](ICML2019)
(AutoAttack)Reliable evaluation of adversarial robustness with an ensemble of diverse parameter-free attacks[papercode](ICML2020)
(C&W)Towards Evaluating the Robustness of Neural Networks[paper]
adversarial examples are not bugs they are features[paper](NIPS2019)
Adversarially Robust Generalization Requires More Data[paper](NIPS2019)
On Adaptive Attacks to Adversarial Example Defenses[paper](NIPS2020)

Data Augmentation

Overfitting in adversarially robust deep learning[paper](ICML2020)
ROBUST LOCAL FEATURES FOR IMPROVING THE GENERALIZATION OF ADVERSARIAL TRAINING[paper](ICLR2020) Fixing Data Augmentation to Improve Adversarial Robustness[paper](NIPS2021)

Unsupervised/Semi-supervised

Are labels required for improving adversarial robustness?(NIPS2019)
Unlabeled data improves adversarial robustness(NIPS2019)
Adversarially robust generalization just requires more unlabeled data(arXiv2019)
Using self-supervised learning can improve model robustness and uncertainty(NIPS2019)

Adaptive epsilon/Curriculum Learning

Attacks Which Do Not Kill Training Make Adversarial Learning Stronger(ICML2020)
Curriculum adversarial training(IJCAI2018)
On the convergence and robustness of adversarial training.(ICML2019)
Instance adaptive adversarial training: Improved accuracy tradeoffs in neural nets(arXiv2019)

Efficient Adversarial Training

Ensemble Adversarial Training

Adversarial Regularization

Transformer and other network architectures

Adversarial Robustness Comparison of Vision Transformer and MLP-Mixer to CNNs[paper](arXiv2021)
ON THE ADVERSARIAL ROBUSTNESS OF VISION TRANSFORMERS[paper](arXiv2021)

Robust Architecture

Others

ADVERSARIAL ROBUSTNESS THROUGH THE LENS OF CAUSALITY[paper](ICLR2022)

Except for Adversarial Training

Adversarial Robustness without Adversarial Training: A Teacher-Guided Curriculum Learning Approach[paper](NIPS2021)

Robust Transfer Learning(forked from jindongwang)