Researchers debate adversarial examples: attackability depends on perturbation budget
Researchers note that adversarial examples are hard to find on CIFAR-10 under a 1/255 Linf perturbation budget, yet no adversarially robust model exists on the benchmark even by 2026.
2026-11-23 ~ 2026-11-23 · 2 related posts
- It's 2026 and still no adversarially robust model exists on CIFAR-10 — scaling01 · 2026-11-23
- Adversarial examples aren't ubiquitous: budget matters, argues researcher — maksym_andr · 2026-11-23