Research map / Adversarial robustness and security / Attacks on text and graphs
Defense Against Adversarial Attacks Using Feature Scattering-based Adversarial Training
Haichao Zhang, Jianyu Wang
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 |
|---|---|---|---|
| feature-scattering adversarial training vs Madry PGD training | adversarial training · CIFAR10 PGD20 | accuracy: 44.9 -> 70.5 | Improved |
| feature scattering vs Bilateral | adversarial training · CIFAR10 PGD20 | accuracy: 57.5 -> 70.5 | Improved |
| feature scattering vs Madry | adversarial training · CIFAR100 PGD20 | accuracy: 22.6 -> 47.2 | Improved |
| feature scattering vs Madry | adversarial training · CIFAR10 clean | accuracy: 85.7 -> 90.0 | Improved |
| feature scattering vs Bilateral | adversarial training · CIFAR10 clean | accuracy: 91.2 -> 90.0 | Hurt |
| generate perturbations by feature scattering instead of supervised loss (ablation) | adversarial training · CIFAR10 PGD20 | accuracy: 56.0 -> 70.5 | Improved |
| feature scattering instead of random perturbation (ablation) | adversarial training · CIFAR10 PGD20 | accuracy: 29.9 -> 70.5 | Improved |
| OT matching instead of uniform matching (ablation) | feature scattering · CIFAR10 PGD20 | accuracy: 57.1 -> 70.5 | Improved |
| OT matching instead of identity matching (ablation) | feature scattering · CIFAR10 PGD20 | accuracy: 57.5 -> 70.5 | Improved |
| IPOT solver instead of Sinkhorn (ablation) | feature scattering · CIFAR10 PGD20 | accuracy: 70.5 -> 69.9 | No clear effect |
Closest papers by meaning
- Breaking Barriers in Physical-World Adversarial Examples: Improving Robustness and Transferability via Robust Feature (AAAI 2025 · similarity 0.87)
- Stylized Adversarial Defense (TPAMI 2022 · similarity 0.87)
- Adversarial Training Can Provably Improve Robustness: Theoretical Analysis of Feature Learning Process Under Structured Data (ICLR 2025 · similarity 0.87)
- Adversarial Robustness without Adversarial Training: A Teacher-Guided Curriculum Learning Approach (NeurIPS 2021 · similarity 0.87)
- DAT: Improving Adversarial Robustness via Generative Amplitude Mix-up in Frequency Domain (NeurIPS 2024 · similarity 0.87)
- Ensemble Adversarial Training: Attacks and Defenses (ICLR 2018 · 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.
