| Paper | Code | Attack: PGD20 | Attack: AutoAttack | Attack: DeepFool | Robust Accuracy | ModelName | ReleaseDate |
|---|---|---|---|---|---|---|---|
| An Orthogonal Classifier for Improving the Adversarial Robustness of Neural Networks | ✓ Link | 78.680 | 44.150 | 51.310 | Xu et al. | 2021-05-19 | |
| Towards Deep Learning Models Resistant to Adversarial Attacks | ✓ Link | 48.440 | AdvTraining [madry2018] | 2017-06-19 | |||
| Theoretically Principled Trade-off between Robustness and Accuracy | ✓ Link | 45.900 | TRADES [zhang2019b] | 2019-01-24 | |||
| Ensemble everything everywhere: Multi-scale aggregation for adversarial robustness | ✓ Link | 78.13 | 3-ensemble of multi-resolution self-ensembles | 2024-08-08 | |||
| Enhancing Robust Representation in Adversarial Training: Alignment and Exclusion Criteria | ✓ Link | 59.70 | TRADES-ANCRA/ResNet18 | 2023-10-05 | |||
| Attention Masks Help Adversarial Attacks to Bypass Safety Detectors | ✓ Link | 1% | XU-Net | 2024-11-07 |