Translation in progress Some entries still include text in the original language.
Failure-First·Indexed by this site · Originally published
Exploring Adversarial Vulnerabilities in Robotic Vision-Language-Action Models
[Daily Paper] Exploring the Adversarial Vulnerabilities of Vision-Language-Action Models in Robotics
AI summary
一个空间感知的对抗攻击框架显示，视觉-语言-动作（VLA）机器人系统存在显著安全漏洞，可导致任务完全失败。该框架表明，观测空间中的对抗补丁能够完全破坏机器人的轨迹执行。
Read originalfailurefirst.org