arXiv · cs.AI · Chenyu Yi, Jiawen Zong, Sami Hormi, Bingquan Shen, Xudong Jiang, Alex Chichung Kot · 5 小时前模型评分41
用文字而非像素：针对大型视觉语言模型的自然文本不可学习样本
The Pen is Mightier than the Pixel: Natural Textual Unlearnable Example for Large Vision Language Models
导读
研究者提出自然文本不可学习样本（NTUE，Natural Textual Unlearnable Examples），这是首次针对大型视觉语言模型（LVLMs，Large Vision Language Models）探索不可学习样本，通过微调图像标签在感兴趣身份与另一目标间建立捷径来误导模型，且不引入事实错误、不影响可读性。
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
- “We propose a novel text-based perturbation method, Natural Textual Unlearnable Examples (NTUE), which can effectively mislead LVLMs by subtly manipulating the image label, establishing a shortcut between an identity of interest and a separate target, without introducing factual misinformation or affecting readability.”
- “Our findings show that NTUE achieves a high success rate with a poison rate as low as 5%, underscoring its potential as a practical and robust solution for protecting multimodal data in the era of LVLMs.”
来源：arXiv · cs.AI · arxiv.org