Research map / Adversarial robustness and security / Black-box and transfer attacks
Adversarial Attack on Attackers: Post-Process to Mitigate Black-Box Score-Based Query Attacks
Sizhe Chen, Zhehao Huang, Qinghua Tao, Yingwen Wu, Cihang Xie, Xiaolin Huang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 AAA post-processing defense | WideResNet-28 · CIFAR-10 Square 2500 queries | adversarial accuracy %: 00.09 -> 80.59 | Improved |
| AAA instead of adversarial training | WideResNet-28 · CIFAR-10 Square 2500 queries | adversarial accuracy %: 67.44 -> 80.59 | Improved |
| add AAA | WideResNet-28 · CIFAR-10 clean | accuracy %: 94.78 -> 94.84 | No clear effect |
| add AAA | WideResNet-28 · CIFAR-10 clean | ECE %: 3.52 -> 2.46 | Improved |
| add adversarial training | WideResNet-28 · CIFAR-10 clean | accuracy %: 94.78 -> 87.02 | Hurt |
| add random noise defense RND | WideResNet-28 · CIFAR-10 clean | ECE %: 3.52 -> 6.32 | Hurt |
| combine AAA with adversarial training (ablation) | AT model · CIFAR-10 Square 2500 queries l2=2.5 | adversarial accuracy %: 25.19 -> 73.72 | Improved |
| combine AAA with adversarial training (ablation) | AT model · CIFAR-10 clean | ECE %: 11.00 -> 10.56 | No clear effect |
| increase AAA optimization iterations from 0 to 20 (ablation) | AAA-linear · CIFAR-10 Square 100 queries | adversarial accuracy %: 39.38 -> 79.29 | Improved |
| increase iterations from 60 to 100 (ablation) | AAA-linear · CIFAR-10 Square 100 queries | adversarial accuracy %: 81.37 -> 81.36 | No clear effect |
| adaptive opposite-direction attacker against linear design | AAA-linear · CIFAR-10 Square 100 queries | adversarial accuracy %: 81.36 -> 57.31 | Hurt |
| use sine target loss instead of linear (ablation) | AAA · CIFAR-10 adaptive Square 100 queries | adversarial accuracy %: 57.31 -> 76.41 | Improved |
Closest papers by meaning
- Square Attack: a query-efficient black-box adversarial attack via random search (ECCV 2020 · similarity 0.91)
- BRUSLEATTACK: A QUERY-EFFICIENT SCORE- BASED BLACK-BOX SPARSE ADVERSARIAL ATTACK (ICLR 2024 · similarity 0.89)
- Improving Query Efficiency of Black-box Adversarial Attack (ECCV 2020 · similarity 0.88)
- Understanding the Robustness of Randomized Feature Defense Against Query-Based Adversarial Attacks (ICLR 2024 · similarity 0.87)
- Mind the Gap: Detecting Black-box Adversarial Attacks in the Making through Query Update Analysis (CVPR 2025 · similarity 0.86)
- Closing the Gap: Achieving Better Accuracy-Robustness Tradeoffs against Query-Based Attacks (AAAI 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.
