Research map / Adversarial robustness and security / Certified and empirical robustness
Provable Certificates for Adversarial Examples: Fitting a Ball in the Union of Polytopes
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 GeoCert instead of Fast-Lip | 70NetBin · Binary MNIST, ℓ∞ | distance: 0.136 -> 0.191 | Helped |
| Use GeoCert instead of Fast-Lip | 40NetBin · Binary MNIST, ℓ∞ | distance: 0.132 -> 0.187 | Helped |
| Use GeoCert instead of Fast-Lip | 70NetBin · Binary MNIST, ℓ2 | distance: 1.289 -> 1.555 | Helped |
| Use GeoCert instead of Fast-Lip | 40NetBin · Binary MNIST, ℓ2 | distance: 1.319 -> 1.607 | Helped |
| Use GeoCert instead of MIP | Binary MNIST network [784, 20, 20, 20, 2] · Binary MNIST, 300s/example, five examples | lower-bound distance: GeoCert: 2.251, 1.356, 1.620, 2.499, 2.402; MIP: 2.0, 1.0, 1.0, 2.0, 2.0 | Helped |
| Use GeoCert instead of Fast-Lip | Binary MNIST network [784, 20, 20, 20, 2] · Binary MNIST, 300s/example, five examples | lower-bound distance: GeoCert: 2.251, 1.356, 1.620, 2.499, 2.402; Fast-Lip: 1.782, 1.319, 1.501, 1.975, 1.871 | Helped |
| Use GeoCert instead of MIP for exact verification | 70NetBin and 40NetBin · Binary MNIST, ℓ2 and ℓ∞ | time: 1.300 vs 0.947; 4.031 vs 0.689; 4.789 vs 4.030; 21.852 vs 5.831 | Mixed |
Closest papers by meaning
- Hierarchical Verification for Adversarial Robustness (ICML 2020 · similarity 0.90)
- Fast Geometric Projections for Local Robustness Certification (ICLR 2021 · similarity 0.89)
- Globally-Robust Neural Networks (ICML 2021 · similarity 0.88)
- Second-Order Provable Defenses against Adversarial Attacks (ICML 2020 · similarity 0.87)
- Compositional Curvature Bounds for Deep Neural Networks (ICML 2024 · similarity 0.85)
- LEVIS: Large Exact Verifiable Input Spaces for Neural Networks (ICML 2025 · similarity 0.85)
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.