Research map / Adversarial robustness and security / Black-box and transfer attacks
Enhancing Cross-Task Black-Box Transferability of Adversarial Examples With Dispersion Reduction
Check 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 |
|---|---|---|---|
| Reduce intermediate feature-map dispersion instead of using task-specific attack losses | Cross-task black-box attacks on detection models · COCO/VOC/ImageNet validation | mAP: 12.8 / 7.4 vs TI-DIM: 16.7 / 11.2 | Helped |
| Reduce intermediate feature-map dispersion instead of using task-specific attack losses | Cross-task black-box attacks on segmentation models · COCO/VOC/ImageNet validation | mIoU: 20.0 / 16.9 vs DIM: 25.9 / 21.8 | Helped |
| Reduce intermediate feature-map dispersion | GCV APIs using VGG-16 as the source model · GCV Labels | accuracy: 41% -> 23% | Helped |
| Reduce intermediate feature-map dispersion | GCV APIs using VGG-16 as the source model · GCV Objects | mAP (IoU=0.5): 36.5 -> 32.9 | Helped |
| Reduce intermediate feature-map dispersion | GCV APIs using VGG-16 as the source model · GCV SafeSearch | accuracy: 57% -> 35% | Helped |
| Reduce intermediate feature-map dispersion | GCV APIs using VGG-16 as the source model · GCV Texts | AP (IoU=0.5): 29.9 -> 20.9 | Helped |
| Reduce intermediate feature-map dispersion | GCV APIs using VGG-16 as the source model · GCV Texts | C.R.W: 15.9% -> 4.1% | Helped |
| Target a middle convolutional layer rather than a top or bottom layer (ablation) | DR on VGG-16 · PASCAL VOC2012 validation, detection and segmentation | mAP / mIoU | Helped |
Closest papers by meaning
- Everywhere Attack: Attacking Locally and Globally to Boost Targeted Transferability (AAAI 2025 · similarity 0.90)
- Enhancing Adversarial Transferability with Checkpoints of a Single Model's Training (CVPR 2025 · similarity 0.90)
- Enhancing Adversarial Transferability with Adversarial Weight Tuning (AAAI 2025 · similarity 0.89)
- Transferable Adversarial Attacks for Object Detection Using Object-Aware Significant Feature Distortion (AAAI 2024 · similarity 0.89)
- Harnessing the Computation Redundancy in ViTs to Boost Adversarial Transferability (NeurIPS 2025 · similarity 0.89)
- Rethinking Model Ensemble in Transfer-based Adversarial Attacks (ICLR 2024 · similarity 0.89)
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.