Research map / Adversarial robustness and security / Black-box and transfer attacks
Weight-covariance alignment for adversarially robust neural networks
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 |
|---|---|---|---|
| Add WCA | WRN-34-10 · CIFAR-10, PGD (ε=8) | accuracy: 0.12 -> 0.73 | Helped |
| Add WCA | Tuned ResNet-18 · CIFAR-10, PGD (ε=8) | accuracy: 0.01 -> 0.71 | Helped |
| Use a tuned ResNet-18 architecture with WCA instead of the submitted RN18-WCA | WCA-Net · CIFAR-10, PGD-100 (ε=8) | accuracy: 0.61 -> 0.72 | Helped |
| Use a tuned ResNet-18 architecture with WCA instead of the submitted RN18-WCA | WCA-Net · CIFAR-10, Square (ε=8) | accuracy: 0.44 -> 0.49 | Helped |
| Add WCA | ResNet-18 · ImageNette, PGD (ε=8) | accuracy: 0 -> 0.49 | Helped |
| Add WCA | ResNet-18 · tiny-ImageNet, PGD (ε=8) | accuracy: 0 -> 0.30 | Helped |
| Use the MaxEnt regularizer instead of the WCA loss to train anisotropic noise (ablation) | ResNet-18 with anisotropic noise · CIFAR-10 (ε=8/255), FGSM | accuracy: 72.4 -> 58.1 { (-14.3)} | Hurt |
| Use the MaxEnt regularizer instead of the WCA loss to train anisotropic noise (ablation) | ResNet-18 with anisotropic noise · CIFAR-10 (ε=8/255), PGD | accuracy: 71.7 -> 54.0 { (-17.7)} | Hurt |
Closest papers by meaning
- How Sampling Impacts the Robustness of Stochastic Neural Networks (NeurIPS 2022 · similarity 0.90)
- Enhancing Adversarial Robustness in SNNs with Sparse Gradients (ICML 2024 · similarity 0.90)
- First Line of Defense: A Robust First Layer Mitigates Adversarial Attacks (AAAI 2025 · similarity 0.89)
- Robust NAS under adversarial training: benchmark, theory, and beyond (ICLR 2024 · similarity 0.88)
- Adversarial Training Can Provably Improve Robustness: Theoretical Analysis of Feature Learning Process Under Structured Data (ICLR 2025 · similarity 0.87)
- Adversarial Robustness via Deformable Convolution with Stochasticity (ICML 2025 · similarity 0.87)
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.