Lite-BD: A Lightweight Black-box Backdoor Defense via Reviving Multi-Stage Image Transformations

Documenti analoghi

• Exploiting the Vulnerability of Large Language Models via Defense-Aware Architectural Backdoor
  di: Miah, Abdullah Arafat, et al.
  Pubblicazione: (2024)
• BadSNN: Backdoor Attacks on Spiking Neural Networks via Adversarial Spiking Neuron
  di: Miah, Abdullah Arafat, et al.
  Pubblicazione: (2026)
• BadImplant: Injection-based Multi-Targeted Graph Backdoor Attack
  di: Khan, Md Nabi Newaz, et al.
  Pubblicazione: (2026)
• NoiseAttack: An Evasive Sample-Specific Multi-Targeted Backdoor Attack Through White Gaussian Noise
  di: Miah, Abdullah Arafat, et al.
  Pubblicazione: (2024)
• Probe before You Talk: Towards Black-box Defense against Backdoor Unalignment for Large Language Models
  di: Yi, Biao, et al.
  Pubblicazione: (2025)