Research map / Adversarial robustness and security / Black-box and transfer attacks
AdaEA++: Enhanced Adaptive Model Ensemble Adversarial Attack for Boosting Adversarial Transferability
Bin Chen, Jia-Li Yin, Bo‐Hao Chen, Ximeng Liu, Yuanlong Yu
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 adaptive gradient modulation (AGM) (ablation) | naive ensemble attack · CIFAR-10 | average ASR: 27.29 -> 39.79 (+12.5) | Improved |
| add disparity-reduced filter (DRF) (ablation) | naive ensemble attack · CIFAR-10 | average ASR: 27.29 -> 42.42 (+15.13) | Improved |
| add DRF (ablation) | naive ensemble attack · CIFAR-10 (ViT targets) | average ASR: 23.94 -> 47.57 | Improved |
| combine AGM and DRF (ablation) | naive ensemble attack · CIFAR-10 | average ASR: 27.29 -> 44.78 (+22.4) | Improved |
| increase weighting scale beta in AGM (ablation) | AdaEA · CIFAR-10 (Res-18, ViT-T) | average ASR | Mixed |
| increase binarization threshold eta in DRF (ablation) | AdaEA · CIFAR-10 (Res-18, ViT-T) | average ASR | Mixed |
| replace Ens with AdaEA on I-FGSM | ensemble attack · CIFAR-10 | average ASR: 26.56 -> 45.09 | Improved |
| replace Ens with AdaEA on I-FGSM | ensemble attack · ImageNet | average ASR: 46.38 -> 49.36 | Improved |
| use SVRE instead of Ens with MI-FGSM | ensemble attack · CIFAR-10 | average ASR: 37.07 -> 23.04 | Hurt |
| use SVRE instead of Ens with DI2-FGSM | ensemble attack · CIFAR-10 | average ASR: 72.44 -> 33.49 | Hurt |
| replace Ens with AdaEA on DI2-FGSM | ensemble attack · CIFAR-10 | average ASR: 72.44 -> 80.32 | Improved |
| use four surrogate models instead of two (ablation) | AdaEA · CIFAR-10 | average ASR | Improved |
Later papers that compared with this method
1 later accepted paper reports AdaEA as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Boosting Adversarial Transferability via Ensemble Non-Attention (AAAI 2026 · better on 2 of 2 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Rethinking Model Ensemble in Transfer-based Adversarial Attacks (ICLR 2024 · similarity 0.92)
- Understanding Model Ensemble in Transferable Adversarial Attack (ICML 2025 · similarity 0.92)
- Enhancing Adversarial Transferability with Adversarial Weight Tuning (AAAI 2025 · similarity 0.89)
- Ensemble Diversity Facilitates Adversarial Transferability (CVPR 2024 · similarity 0.89)
- Enhancing Adversarial Transferability with Checkpoints of a Single Model's Training (CVPR 2025 · similarity 0.88)
- Boosting Adversarial Transferability through Augmentation in Hypothesis Space (CVPR 2025 · similarity 0.88)
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.
