Research map / Adversarial robustness and security / Black-box and transfer attacks
Improving the Transferability of Adversarial Samples by Path-Augmented Method
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 |
|---|---|---|---|
| use PAM instead of Admix | transfer-based attacks · ImageNet adversarially trained models (Inc-v3 ens3, Inc-v3 ens4, IncRes-v2 adv) | attack success rate: 44.8 vs 36.9; 43.4 vs 34.8; 22.4 vs 19.0 (Inc-v3 source) | Helped |
| add multiple augmentation paths to SIM (ablation) | SIM · ImageNet, IncRes-v2 target | attack success rate: 67.7 -> 79.7 | Helped |
| add Semantics Predictor to SIM (ablation) | SIM · ImageNet, IncRes-v2 target | attack success rate: 67.7 -> 68.3 | Helped |
| add Semantics Predictor to SIM with multiple paths (ablation) | SIM+paths · ImageNet, IncRes-v2 target | attack success rate: 79.7 -> 81.2 | Helped |
| add Semantics Predictor to SIM (ablation) | SIM · ImageNet, Res-v2 target | attack success rate: 63.2 -> 64.3 | Helped |
| add multiple augmentation paths to SIM (ablation) | SIM · ImageNet, Res-v2 target | attack success rate: 63.2 -> 76.4 | Helped |
| add Semantics Predictor to SIM with multiple paths (ablation) | SIM+paths · ImageNet, Res-v2 target | attack success rate: 76.4 -> 77.5 | Helped |
| add Semantics Predictor to SIM (ablation) | SIM · ImageNet, Inc-v3 ens4 target | attack success rate: 31.4 -> 32.5 | Helped |
| add multiple augmentation paths to SIM (ablation) | SIM · ImageNet, Inc-v3 ens4 target | attack success rate: 31.4 -> 40.7 | Helped |
| add Semantics Predictor to SIM with multiple paths (ablation) | SIM+paths · ImageNet, Inc-v3 ens4 target | attack success rate: 40.7 -> 43.4 | Helped |
| add Semantics Predictor to SIM (ablation) | SIM · ImageNet, IncRes-v2 adv target | attack success rate: 17.5 -> 18.3 | Helped |
| add multiple augmentation paths to SIM (ablation) | SIM · ImageNet, IncRes-v2 adv target | attack success rate: 17.5 -> 21.2 | Helped |
Closest papers by meaning
- Boosting Adversarial Transferability through Augmentation in Hypothesis Space (CVPR 2025 · similarity 0.89)
- Improving the Transferability of Adversarial Samples With Adversarial Transformations (CVPR 2021 · similarity 0.88)
- Improving Adversarial Transferability with Local Perturbation Augmentation (CVPR 2026 · similarity 0.87)
- Enhancing Adversarial Transferability with Adversarial Weight Tuning (AAAI 2025 · similarity 0.87)
- Prompting Adversarial Transferability via Path Flatness Attack (AAAI 2026 · similarity 0.87)
- PubDef: Defending Against Transfer Attacks From Public Models (ICLR 2024 · similarity 0.86)
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.