Research map / Adversarial robustness and security / Attacks on text and graphs
Modeling Adversarial Noise for Adversarial Training
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add instance-dependent label-transition modeling to AT | ResNet-18 · CIFAR-10, adaptive FWA-40 | adversarial accuracy (%): 15.44 -> 29.53 | Helped |
| Add instance-dependent label-transition modeling to AT | ResNet-18 · CIFAR-10, adaptive CW₂ | adversarial accuracy (%): 0.00 -> 43.17 | Helped |
| Add instance-dependent label-transition modeling to AT | ResNet-18 · CIFAR-10, adaptive DDN | adversarial accuracy (%): 0.09 -> 10.63 | Helped |
| Add instance-dependent label-transition modeling to AT | ResNet-18 · CIFAR-10, natural (None) | accuracy (%): 83.39 -> 82.72 | Hurt |
| Add instance-dependent label-transition modeling to AT | ResNet-18 · Tiny-ImageNet, adaptive PGD-40 | adversarial accuracy (%): 17.35 -> 18.15 | Helped |
| Add instance-dependent label-transition modeling to AT | ResNet-18 · Tiny-ImageNet, adaptive CW₂ | adversarial accuracy (%): 0.00 -> 16.27 | Helped |
| Add instance-dependent label-transition modeling to TRADES | ResNet-18 · CIFAR-10, adaptive PGD-40 | adversarial accuracy (%): 46.29 -> 48.65 | Helped |
| Add instance-dependent label-transition modeling to MART | ResNet-18 · CIFAR-10, adaptive FWA-40 | adversarial accuracy (%): 25.56 -> 31.23 | Helped |
| Add instance-dependent label-transition modeling to TRADES | ResNet-18 · CIFAR-10, adaptive CW₂ | adversarial accuracy (%): 0.00 -> 1.46 | Helped |
| Add instance-dependent label-transition modeling to MART | ResNet-18 · CIFAR-10, adaptive CW₂ | adversarial accuracy (%): 0.02 -> 1.53 | Helped |
| Jointly adversarially train the target model and transition network instead of fixing the target model (ablation) | MAN · adaptive attacks (setting described for ablation) | adversarial accuracy | Helped |
| Use a two-parallel-ResNet-18 AT baseline with a similar parameter count instead of MAN (ablation) | CIFAR-10 adaptive-attack comparison · CIFAR-10, adaptive attacks | adversarial accuracy | Hurt |
Closest papers by meaning
- Label Noise in Adversarial Training: A Novel Perspective to Study Robust Overfitting (NeurIPS 2022 · similarity 0.89)
- A law of adversarial risk, interpolation, and label noise (ICLR 2023 · similarity 0.88)
- Soften to Defend: Towards Adversarial Robustness via Self-Guided Label Refinement (CVPR 2024 · similarity 0.88)
- On the Role of Label Noise in the Feature Learning Process (ICML 2025 · similarity 0.86)
- First Line of Defense: A Robust First Layer Mitigates Adversarial Attacks (AAAI 2025 · similarity 0.85)
- Understanding and Mitigating the Label Noise in Pre-training on Downstream Tasks (ICLR 2024 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.