Research map / Adversarial robustness and security / Certified and empirical robustness
Defense Through Diverse Directions
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 the mean penalty (ablation) | Bayesian CNN · MNIST | PGD adversarial accuracy (%): 0.9 -> 55.2 | Helped |
| Add the mean penalty (ablation) | Bayesian CNN · MNIST | Black-Box adversarial accuracy (%): 56.4 -> 60.1 | No clear effect |
| Add the non-sparse penalty to mean penalty (ablation) | Bayesian CNN · MNIST | PGD adversarial accuracy (%): 55.2 -> 94.0 | Helped |
| Add the non-sparse penalty to mean penalty (ablation) | Bayesian CNN · MNIST | Black-Box adversarial accuracy (%): 60.1 -> 54.3 | Hurt |
| Add the variance penalty to mean and non-sparse penalties (ablation) | Bayesian CNN · MNIST | PGD adversarial accuracy (%): 94.0 -> 95.8 | Helped |
| Add offline adversarial training to mean and non-sparse penalties (ablation) | Bayesian CNN · MNIST | Black-Box adversarial accuracy (%): 54.3 -> 89.9 | Helped |
| Add the mean penalty (ablation) | Bayesian VGG16 · CIFAR-10 | Black-Box adversarial accuracy (%): 24.5 -> 64.1 | Helped |
| Add the non-sparse penalty to mean penalty (ablation) | Bayesian VGG16 · CIFAR-10 | PGD adversarial accuracy (%): 6.1 -> 47.7 | Helped |
| Add the non-sparse penalty to mean penalty (ablation) | Bayesian VGG16 · CIFAR-10 | Standard accuracy (%): 90.1 -> 88.4 | Hurt |
| Add offline adversarial training to mean and non-sparse penalties (ablation) | Bayesian VGG16 · CIFAR-10 | PGD adversarial accuracy (%): 47.7 -> 72.3 | Helped |
Closest papers by meaning
- Bayesian Learning with Information Gain Provably Bounds Risk for a Robust Adversarial Defense (ICML 2022 · similarity 0.92)
- Adv-BNN: Improved Adversarial Defense through Robust Bayesian Neural Network (ICLR 2019 · similarity 0.92)
- Bayesian Adversarial Learning (NeurIPS 2018 · similarity 0.91)
- Certifying Some Distributional Robustness with Principled Adversarial Training (ICLR 2018 · similarity 0.89)
- Adversarial Robustness via Deformable Convolution with Stochasticity (ICML 2025 · similarity 0.88)
- Adversarial Training Can Provably Improve Robustness: Theoretical Analysis of Feature Learning Process Under Structured Data (ICLR 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.