Research map / Adversarial robustness and security / Certified and empirical robustness
Towards Robust Detection of Adversarial Examples
Tianyu Pang, Chao Du, Yinpeng Dong, Jun Zhu
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 |
|---|---|---|---|
| Use RCE training instead of CE with K-density detection | Resnet-32 · MNIST; FGSM | AUC score (10^-2): 98.8 -> 99.4 | Improved |
| Use RCE training instead of CE with K-density detection | Resnet-32 · CIFAR-10; FGSM | AUC score (10^-2): 99.7 -> 98.0 | Hurt |
| Use RCE training instead of CE with K-density detection | Resnet-32 · MNIST; BIM | AUC score (10^-2): 90.0 -> 91.8 | Improved |
| Use RCE training instead of CE with K-density detection | Resnet-32 · CIFAR-10; BIM | AUC score (10^-2): 100.0 -> 100.0 | No clear effect |
| Use RCE training instead of CE with K-density detection | Resnet-32 · MNIST; ILCM | AUC score (10^-2): 96.2 -> 98.6 | Improved |
| Use RCE training instead of CE with K-density detection | Resnet-32 · CIFAR-10; ILCM | AUC score (10^-2): 84.2 -> 93.9 | Improved |
| Use RCE training instead of CE with K-density detection | Resnet-32 · MNIST; JSMA | AUC score (10^-2): 97.7 -> 99.0 | Improved |
| Use RCE training instead of CE with K-density detection | Resnet-32 · CIFAR-10; JSMA | AUC score (10^-2): 85.8 -> 95.4 | Improved |
| Use RCE training instead of CE with K-density detection | Resnet-32 · MNIST; C&W | AUC score (10^-2): 99.4 -> 99.8 | Improved |
| Use RCE training instead of CE with K-density detection | Resnet-32 · CIFAR-10; C&W | AUC score (10^-2): 95.3 -> 98.2 | Improved |
| Use RCE training instead of CE with K-density detection | Resnet-32 · MNIST; C&W-hc | AUC score (10^-2): 91.1 -> 99.6 | Improved |
| Use RCE training instead of CE with K-density detection | Resnet-32 · CIFAR-10; C&W-hc | AUC score (10^-2): 75.4 -> 91.8 | Improved |
Closest papers by meaning
- Improving Adversarial Robustness Requires Revisiting Misclassified Examples (ICLR 2020 · similarity 0.89)
- GAT: Generative Adversarial Training for Adversarial Example Detection and Robust Classification (ICLR 2020 · similarity 0.88)
- Are Adversarial Examples Created Equal? A Learnable Weighted Minimax Risk for Robustness under Non-uniform Attacks (AAAI 2021 · similarity 0.88)
- Certified Adversarial Robustness with Additive Noise (NeurIPS 2019 · similarity 0.87)
- Rethinking Softmax Cross-Entropy Loss for Adversarial Robustness (ICLR 2020 · similarity 0.87)
- Identifying and Understanding Cross-Class Features in Adversarial Training (ICML 2025 · 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.
