Research map / Adversarial robustness and security / Certified and empirical robustness
Randomization matters How to defend against strong adversarial attacks
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 |
|---|---|---|---|
| Replace a single adversarially trained classifier with the BAT mixture | CIFAR10 · CIFAR10, Adaptive-ℓ∞-PGD (ε∞ = 0.031) | accuracy: 0.42 -> 0.55 | Helped |
| Replace a single adversarially trained classifier with the BAT mixture | CIFAR10 · CIFAR10, Adaptive-ℓ2-C&W (ε2 = 0.4) | accuracy: 0.60 -> 0.60 | No clear effect |
| Replace a single adversarially trained classifier with the BAT mixture | CIFAR10 · CIFAR10, Adaptive-ℓ2-C&W (ε2 = 0.6) | accuracy: 0.47 -> 0.57 | Helped |
| Replace a single adversarially trained classifier with the BAT mixture | CIFAR10 · CIFAR10, Adaptive-ℓ2-C&W (ε2 = 0.8) | accuracy: 0.35 -> 0.53 | Helped |
| Replace a single adversarially trained classifier with the BAT mixture | CIFAR100 · CIFAR100, Adaptive-ℓ∞-PGD (ε∞ = 0.031) | accuracy: 0.26 -> 0.40 | Helped |
| Replace a single adversarially trained classifier with the BAT mixture | CIFAR100 · CIFAR100, Adaptive-ℓ2-C&W (ε2 = 0.4) | accuracy: 0.38 -> 0.45 | Helped |
| Replace a single adversarially trained classifier with the BAT mixture | CIFAR100 · CIFAR100, Adaptive-ℓ2-C&W (ε2 = 0.6) | accuracy: 0.29 -> 0.41 | Helped |
| Replace a single adversarially trained classifier with the BAT mixture | CIFAR100 · CIFAR100, Adaptive-ℓ2-C&W (ε2 = 0.8) | accuracy: 0.22 -> 0.38 | Helped |
Closest papers by meaning
- Mixed Nash Equilibria in the Adversarial Examples Game (ICML 2021 · similarity 0.92)
- A Game Theoretic Analysis of Additive Adversarial Attacks and Defenses (NeurIPS 2020 · similarity 0.91)
- Towards Optimal Randomized Strategies in Adversarial Example Game (AAAI 2023 · similarity 0.89)
- Adversarial Robustness via Deformable Convolution with Stochasticity (ICML 2025 · similarity 0.88)
- Understanding the Robustness of Randomized Feature Defense Against Query-Based Adversarial Attacks (ICLR 2024 · similarity 0.86)
- Adversarial Training Should Be Cast as a Non-Zero-Sum Game (ICLR 2024 · 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.