Research map / Adversarial robustness and security / Certified and empirical robustness
Semialgebraic Representation of Monotone Deep Equilibrium Models and Applications to Certification
Tong Chen, Jean B. Lasserre, Victor Magron, Edouard Pauwels
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 Robustness Model instead of Lipschitz Model | certification · MNIST Linf eps=0.05 | certified ratio: 0% -> 24% | Improved |
| Use Robustness Model instead of Lipschitz Model | certification · MNIST Linf eps=0.01 | certified ratio: 24% -> 99% | Improved |
| Use Ellipsoid Model instead of Lipschitz Model | certification · MNIST Linf eps=0.01 | certified ratio: 24% -> 92% | Improved |
| Use Ellipsoid Model instead of Lipschitz Model | certification · MNIST L2 eps=0.1 | certified ratio: 91% -> 99% | Improved |
| Use Robustness Model instead of Ellipsoid Model | certification · MNIST Linf eps=0.05 | certified ratio: 0% -> 24% | Improved |
| Use Robustness Model instead of Ellipsoid Model | certification · MNIST Linf eps=0.01 | certified ratio: 92% -> 99% | Improved |
| Use Robustness Model instead of Ellipsoid Model | running time per example · MNIST | seconds: 500s -> 1350s | Hurt |
| Use Lipschitz Model's directly computed constant instead of bound from [33] | Lipschitz upper bound · q=2 | bound: 4.80 -> 4.67 | Improved |
| Use Lipschitz Model's directly computed constant instead of bound from [33] | Lipschitz upper bound · q=inf | bound: 824.14 -> 108.84 | Improved |
| Use Lipschitz Model's directly computed constant instead of bound from [33] | certification · MNIST Linf eps=0.01 | certified ratio: 0% -> 24% | Improved |
| Use Lipschitz Model instead of bound from [33] | computation time · Lipschitz bound q=inf | seconds: 1898.65 | Hurt |
| Increase perturbation eps from 0.01 to 0.1 | Robustness Model · MNIST Linf | certified ratio: 99% -> 0% | Hurt |
Closest papers by meaning
- Certified Robustness for Deep Equilibrium Models via Interval Bound Propagation (ICLR 2022 · similarity 0.88)
- Certified Robustness for Deep Equilibrium Models via Serialized Random Smoothing (NeurIPS 2024 · similarity 0.86)
- Positive Concave Deep Equilibrium Models (ICML 2024 · similarity 0.85)
- Solving hidden monotone variational inequalities with surrogate losses (ICLR 2025 · similarity 0.84)
- Lyapunov-Stable Deep Equilibrium Models (AAAI 2024 · similarity 0.83)
- Smooth Min-Max Monotonic Networks (ICML 2024 · similarity 0.83)
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.
