Summary of Imperceptible Adversarial Examples in the Physical World, by Weilin Xu et al.
Imperceptible Adversarial Examples in the Physical World
by Weilin Xu, Sebastian Szyller, Cory Cornelius, Luis Murillo Rojas, Marius Arvinte, Alvaro Velasquez, Jason Martin, Nageen Himayat
First submitted to arxiv on: 25 Nov 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 presents a novel approach to generating imperceptible adversarial examples in the physical world, which are typically difficult to produce due to the non-differentiable nature of visual sensing systems. The authors employ a straight-through estimator (STE) to overcome this challenge, allowing for fast generation of _bounded adversarial examples that can force zero classification accuracy or cause significant drops in object detection performance. The proposed method is demonstrated using printout photos and the CARLA simulator, showcasing its effectiveness in generating imperceptible adversarial patches. This work highlights the need to re-evaluate the threat of adversarial examples in the physical world. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper makes it possible for bad guys to create fake images that are almost indistinguishable from real ones. These fake images can be used to trick computers and make them make mistakes. The researchers developed a new way to create these fake images, called “adversarial examples,” using a special technique called the straight-through estimator (STE). They tested this method using pictures printed on paper and in a computer simulation, and it worked really well. |
Keywords
» Artificial intelligence » Classification » Object detection
