Research map / Adversarial robustness and security / Certified and empirical robustness
Input-Specific Robustness Certification for Randomized Smoothing
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 input-specific rather than input-agnostic sample sizes at a similar certification time | Randomized smoothing certification · ImageNet (σ=1.0) | ACR: 0.917 -> 0.958 | Helped |
| Use input-specific rather than input-agnostic sample sizes at a similar certification time | Randomized smoothing certification · ImageNet (σ=0.5) | ACR: 0.77 -> 0.794 | Helped |
| Use input-specific rather than input-agnostic sample sizes at a similar certification time | Randomized smoothing certification · CIFAR-10 (σ=1.0) | ACR: 0.763 -> 0.78 | Helped |
| Use input-specific sampling with a 0.10 absolute-decline bound instead of IAS at comparable runtime | Randomized smoothing certification · ImageNet (σ=1.0) | ACR: 0.967 -> 1.005 | Helped |
| Use relative-decline-constrained input-specific sampling instead of IAS at comparable runtime | Randomized smoothing certification · ImageNet (σ=1.0) | ACR: 0.959 -> 0.966 | Helped |
| Use relative-decline-constrained input-specific sampling instead of IAS at comparable runtime | Randomized smoothing certification · ImageNet (σ=1.0) | ACR: 0.927 -> 0.935 | Helped |
| Use absolute-decline-constrained input-specific sample sizes rather than a fixed sample size (ablation) | Randomized smoothing certification · ImageNet | ACR | Mixed |
| Use relative-decline constraints instead of absolute-decline constraints (ablation) | ISS · ImageNet | Radius decline across inputs | Helped |
| Vary p_A while holding the ISS decline type and other settings fixed (ablation) | ISS · ISS sample-size mapping | Required sample size | Mixed |
| Increase the desired maximum sample size (ablation) | ISS · ISS sample-size mapping | Required sample size | Helped |
| Increase the allowed absolute or relative decline (ablation) | ISS · ISS sample-size mapping | Required sample size | Helped |
| Set the initial estimation fraction k0/overline{k} too low (ablation) | ISS · ImageNet and CIFAR-10 | Runtime | Hurt |
Closest papers by meaning
- Average Certified Radius is a Poor Metric for Randomized Smoothing (ICML 2025 · similarity 0.87)
- Mitigating the Curse of Dimensionality for Certified Robustness via Dual Randomized Smoothing (ICLR 2024 · similarity 0.86)
- Towards Large Certified Radius in Randomized Smoothing Using Quasiconcave Optimization (AAAI 2024 · similarity 0.86)
- Effects of Exponential Gaussian Distribution on (Double Sampling) Randomized Smoothing (ICML 2024 · similarity 0.86)
- Certified Robustness for Deep Equilibrium Models via Serialized Random Smoothing (NeurIPS 2024 · similarity 0.86)
- Halt Fast! Early Stopping for Certified Robustness (NeurIPS 2026 · 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.