Research map / Adversarial robustness and security / Certified and empirical robustness
On the Certified Robustness for Ensemble Models and Beyond
Zhuolin Yang, Linyi Li, Xiaojun Xu, Bhavya Kailkhura, Tao Xie, Bo Li
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 |
|---|---|---|---|
| Add DRT to WE with SmoothAdv base models | WE · MNIST, L2 radius 2.50 | certified accuracy: 25.0 -> 34.8 | Improved |
| Add DRT to WE with SmoothAdv base models | WE · CIFAR-10, L2 radius 1.50 | certified accuracy: 23.4 -> 30.3 | Improved |
| Add DRT to WE with SmoothAdv base models | WE · CIFAR-10, L2 radius 2.00 | certified accuracy: 17.5 -> 20.3 | Improved |
| Add DRT to MME with SmoothAdv base models | MME · ImageNet, L2 radius 3.00 | certified accuracy: 21.6 -> 23.4 | Improved |
| Add the Gradient Diversity loss (ablation) | DRT ensemble · MNIST and CIFAR-10 | certified accuracy | Improved |
| Add the Confidence Margin loss (ablation) | DRT ensemble · MNIST and CIFAR-10 | certified accuracy | Improved |
| Add the Gradient Diversity loss (ablation) | DRT ensemble · MNIST and CIFAR-10, larger radii | certified accuracy | Improved |
| Combine Gradient Diversity and Confidence Margin losses (ablation) | DRT ensemble · MNIST and CIFAR-10, all radii | certified accuracy | Improved |
Closest papers by meaning
- On the Perils of Cascading Robust Classifiers (ICLR 2023 · similarity 0.91)
- Training Robust Ensembles Requires Rethinking Lipschitz Continuity (ICLR 2025 · similarity 0.90)
- Improving Ensemble Robustness by Collaboratively Promoting and Demoting Adversarial Robustness (AAAI 2021 · similarity 0.90)
- Improving Adversarial Robustness via Promoting Ensemble Diversity (ICML 2019 · similarity 0.89)
- Learning Better Certified Models from Empirically-Robust Teachers (NeurIPS 2026 · similarity 0.89)
- Probably Approximately Global Robustness Certification (ICML 2025 · similarity 0.88)
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.
