Blending adversarial training and representation-conditional purification via aggregation improves adversarial robustness

Documenti analoghi

• Deepfake detectors are DUMB: A benchmark to assess adversarial training robustness under transferability constraints
  di: Serrano, Adrian, et al.
  Pubblicazione: (2026)
• D2R: dual regularization loss with collaborative adversarial generation for model robustness
  di: Liu, Zhenyu, et al.
  Pubblicazione: (2025)
• Trainwreck: A damaging adversarial attack on image classifiers
  di: Zahálka, Jan
  Pubblicazione: (2023)
• Robust width: A lightweight and certifiable adversarial defense
  di: Peck, Jonathan, et al.
  Pubblicazione: (2024)
• Verification of Neural Networks against Convolutional Perturbations via Parameterised Kernels
  di: Brückner, Benedikt, et al.
  Pubblicazione: (2024)