arXiv · cs.AI · Yukiya Horiba, Koshiro Aoki, Shunsuke Yasuki, Bum Jun Kim, Taiki Miyanishi · 8 小时前模型评分34
检测与抑制：针对 VLA 模型对抗补丁的机制性防御
Detect and Suppress: A Mechanistic Defense against Adversarial Patches in VLA Models
导读
研究者用稀疏自编码器（SAE，Sparse Autoencoder）分析视觉-语言-动作（VLA，Vision-Language-Action）模型的内部表示，找到一个激活与对抗补丁强相关的特征，仅在线性探针检测到攻击时于推理阶段抑制该特征，无需微调即可提升鲁棒性。在 LIBERO-10 上的测试显示，间歇性攻击下按需干预能提高成功率，而持续施加同一干预会明显损害策略性能。
分类置信度
中 · 待人工复核
- 只有单一媒体来源，需要人工复核。
来源级别：媒体与个人
事件类型：研究发布
证据摘录
- “In this work, we mechanistically analyze VLA representations using a sparse autoencoder (SAE) and identify a feature whose activation strongly correlates with the presence of an adversarial patch.”
- “we suppress the identified feature at inference time only when a linear probe detects an attack”
- “This intervention improves robustness without the cost of fine-tuning the VLA.”
- “Conditional intervention improves success rate under intermittent attacks, whereas continuously applying the same intervention substantially degrades policy performance.”
来源：arXiv · cs.AI · arxiv.org