A Patch-based Cross-view Regularized Framework for Backdoor Defense in Multimodal Large Language Models

Documenti analoghi

• Multimodal Backdoor Attack on VLMs for Autonomous Driving via Graffiti and Cross-Lingual Triggers
  di: Wang, Jiancheng, et al.
  Pubblicazione: (2026)
• PAD: Patch-Agnostic Defense against Adversarial Patch Attacks
  di: Jing, Lihua, et al.
  Pubblicazione: (2024)
• Explore Internal and External Similarity for Single Image Deraining with Graph Neural Networks
  di: Wang, Cong, et al.
  Pubblicazione: (2024)
• A Dual-Purpose Framework for Backdoor Defense and Backdoor Amplification in Diffusion Models
  di: Truong, Vu Tuan, et al.
  Pubblicazione: (2025)
• BOSC: A Backdoor-based Framework for Open Set Synthetic Image Attribution
  di: Wang, Jun, et al.
  Pubblicazione: (2024)