Summary of Any Target Can Be Offense: Adversarial Example Generation Via Generalized Latent Infection, by Youheng Sun et al.
Any Target Can be Offense: Adversarial Example Generation via Generalized Latent Infection
by Youheng Sun, Shengming Yuan, Xuanhan Wang, Lianli Gao, Jingkuan Song
First submitted to arxiv on: 17 Jul 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 A novel targeted adversarial attack method, Generalized Adversarial attacker (GAKer), is proposed to effectively mislead deep neural networks (DNNs) into recognizing any target object. GAKer overcomes the limitations of existing methods by crafting latently infected representations during adversarial example generation. This class-agnostic and model-agnostic approach reveals DNN vulnerabilities in a wider range of classes, achieving an approximately 14.13% higher attack success rate for unknown classes and 4.23% higher success rate for known classes compared to other generative methods. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Deep neural networks (DNNs) can be tricked into recognizing any object as long as they’re trained well! The new GAKer method helps figure out how good or bad a DNN is by making tiny changes to images, like adding tiny marks. This makes the DNN think an image is something it’s not. The cool thing about GAKer is that it works for both objects we know and ones we don’t yet. |
