Summary of Patch Is Enough: Naturalistic Adversarial Patch Against Vision-language Pre-training Models, by Dehong Kong et al.
Patch is Enough: Naturalistic Adversarial Patch against Vision-Language Pre-training Models
by Dehong Kong, Siyuan Liang, Xiaopeng Zhu, Yuansheng Zhong, Wenqi Ren
First submitted to arxiv on: 7 Oct 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper proposes a novel strategy to improve adversarial attacks on Visual Language Pre-training (VLP) models by exclusively using image patches, preserving the integrity of original text. The approach leverages prior knowledge from diffusion models and utilizes cross-attention mechanism to optimize patch placement. In a white-box setting for image-to-text scenarios, the proposed method achieves a 100% attack success rate, outperforming existing techniques. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The researchers developed a new way to make VLP models vulnerable to attacks. They focused on images rather than text and used a special technique called cross-attention to find the best places to add fake information. This approach was very successful in making the models fail, with a 100% success rate. It also worked well when switching from image-to-text to text-to-image tasks. |
Keywords
» Artificial intelligence » Cross attention » Diffusion
