Research map / Adversarial robustness and security / Attacks on text and graphs
Generate (Non-Software) Bugs to Fool Classifiers
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 |
|---|---|---|---|
| optimize perturbation location and rotation with PEPG instead of using the patch-based method | GAN-generated moth-like image patches · ImageNet, first 50 NIPS '17 tasks | tasks with successful adversarial generation: five tasks | Helped |
| optimize perturbation location and rotation with PEPG instead of using the patch-based method | GAN-generated moth-like image patches · Road sign classifier, 128×128 | time to generate 100 adversarial examples: about an hour -> about 6 minutes | Helped |
| optimize perturbation location and rotation with PEPG instead of using the patch-based method | GAN-generated moth-like image patches · Road sign classifier, 128×128 | iterations to generate 100 adversarial examples: 5,340 -> 753 | Helped |
| increase perturbation size (ablation) | patch-based and PEPG-based image attacks · ImageNet, first 50 NIPS '17 tasks | success rate | Helped |
| increase perturbation size (ablation) | patch-based and PEPG-based image attacks · Road sign classifier | generation success: 32 × 32; 20,000 iterations | Helped |
| use PEPG-based location optimization rather than patch-based location robustness (ablation) | GAN-generated image patches · ImageNet perturbation relocation | success rate after random relocation | Hurt |
| generate bird-chirp-like audio perturbations and add them to speech clips | speech command classifier · Speech Commands Dataset, “yes” and “no” clips targeted as “stop” | targeted misclassification | Helped |
Closest papers by meaning
- Generative Adversarial Perturbations (CVPR 2018 · similarity 0.90)
- Generating Natural Adversarial Examples (ICLR 2018 · similarity 0.90)
- NAG: Network for Adversary Generation (CVPR 2018 · similarity 0.89)
- One Man's Trash Is Another Man's Treasure: Resisting Adversarial Examples by Adversarial Examples (CVPR 2020 · similarity 0.88)
- PixelDefend: Leveraging Generative Models to Understand and Defend against Adversarial Examples (ICLR 2018 · similarity 0.88)
- Wide Two-Layer Networks can Learn from Adversarial Perturbations (NeurIPS 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.