Research map / Adversarial robustness and security / Certified and empirical robustness
Confidence-Aware Training of Smoothed Classifiers for Certified Robustness
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 |
|---|---|---|---|
| Use confidence-aware sample-wise Gaussian training with bottom-K and worst-case losses | smoothed classifiers · CIFAR-10, σ=1.00 | ACR: 0.773 → 0.815 | Helped |
| Use confidence-aware sample-wise Gaussian training with bottom-K and worst-case losses | smoothed classifiers · CIFAR-10-C | mAcc: 67.7 → 70.1 | Helped |
| Use confidence-aware sample-wise Gaussian training with bottom-K and worst-case losses | smoothed classifiers · ImageNet validation, σ=1.00 | ACR: 1.047 → 1.071 | Helped |
| Add worst-case loss weighted by λ (ablation) | CAT-RS · CIFAR-10, σ=0.50 | certified accuracy versus clean accuracy | Mixed |
| Increase the number of Gaussian noise samples M (ablation) | CAT-RS · CIFAR-10, σ=0.50 | approximate certified test accuracy | Mixed |
| Add the bottom-K Gaussian loss (ablation) | CAT-RS · CIFAR-10, σ=0.50 | certified robustness | Helped |
| Add the worst-case Gaussian loss (ablation) | CAT-RS · CIFAR-10, σ=0.50 | certified robustness | Helped |
Closest papers by meaning
- Consistency Regularization for Certified Robustness of Smoothed Classifiers (NeurIPS 2020 · similarity 0.94)
- Certified Adversarial Robustness via Randomized Smoothing (ICML 2019 · similarity 0.94)
- Provably Robust Deep Learning via Adversarially Trained Smoothed Classifiers (NeurIPS 2019 · similarity 0.93)
- DRF: Improving Certified Robustness via Distributional Robustness Framework (AAAI 2024 · similarity 0.89)
- Treatment of Statistical Estimation Problems in Randomized Smoothing for Adversarial Robustness (NeurIPS 2024 · similarity 0.89)
- Provably Cost-Sensitive Adversarial Defense via Randomized Smoothing (ICML 2025 · similarity 0.89)
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.