热点
"对抗性扰动" 相关文章
De-AntiFake: Rethinking the Protective Perturbations Against Voice Cloning Attacks
cs.AI updates on arXiv.org 2025-07-04T04:08:45.000000Z
Microsoft Researchers Propose MedFuzz: A New AI Method for Evaluating the Robustness of Medical Question-Answering LLMs to Adversarial Perturbations
MarkTechPost@AI 2024-09-14T05:05:32.000000Z