Summary of Adversarial Attacks on Parts Of Speech: An Empirical Study in Text-to-image Generation, by G M Shahariar et al.
Adversarial Attacks on Parts of Speech: An Empirical Study in Text-to-Image Generation
by G M Shahariar, Jia Chen, Jiachen Li, Yue Dong
First submitted to arxiv on: 21 Sep 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Machine Learning (cs.LG)
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 The abstract discusses the vulnerability of text-to-image (T2I) models to adversarial attacks, particularly with noun perturbations in text prompts. The study investigates how different parts-of-speech (POS) tags within text prompts affect the images generated by T2I models. To achieve this, the researchers created a high-quality dataset for realistic POS tag token swapping and performed gradient-based attacks to find adversarial suffixes that mislead T2I models into generating images with altered tokens. The results show that the attack success rate varies significantly among different POS tag categories, with nouns, proper nouns, and adjectives being the easiest to attack. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper explores how text-to-image (T2I) models can be tricked into generating wrong images by adding special words to the text prompts. The researchers found that certain types of words, like names and common nouns, are easier to trick than others. They also discovered why this happens and made their program publicly available. |
Keywords
» Artificial intelligence » Token
