Adversarial Attacks on Parts of Speech: An Empirical Study in Text-to-Image Generation
On the map: LLM safety and security → Vision-language and text-to-image safety
Abstract (arXiv)
Recent studies show that text-to-image (T2I) models are vulnerable to adversarial attacks, especially with noun perturbations in text prompts. In this study, we investigate the impact of adversarial attacks on different POS tags within text prompts on the images generated by T2I models. We create a high-quality dataset for realistic POS tag token swapping and perform gradient-based attacks to find adversarial suffixes that mislead T2I models into generating images with altered tokens. Our empirical results show that the attack success rate (ASR) varies significantly among different POS tag categories, with nouns, proper nouns, and adjectives being the easiest to attack. We explore the mechanism behind the steering effect of adversarial suffixes, finding that the number of critical tokens and content fusion vary among POS tags, while features like suffix transferability are consistent across categories. We have made our implementation publicly available at - https://github.com/shahariar-shibli/Adversarial-Attack-on-POS-Tags.
Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Placed in Vision-language and text-to-image safety, one of 12 directions in LLM safety and security. Builds on 21 of its 32 references that are in the map. First posted on arXiv 21 Sep 2024.
Builds on
21 of 32 references.
Closest ancestors · 2
Same direction, newest first
- 2023SneakyPrompt: Jailbreaking Text-to-image Generative Models20 May
- 2023Black Box Adversarial Prompting for Foundation Models8 Feb
Other references · 19
- 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
- 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
- 2014Microsoft COCO: Common Objects in Contextcited by 17,498
- 2021Diffusion Models Beat GANs on Image Synthesiscited by 8,185
- 2022Hierarchical Text-Conditional Image Generation with CLIP Latentscited by 6,539
- 2022Photorealistic Text-to-Image Diffusion Models with Deep Language Understandingcited by 6,322
- 2022BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generationcited by 4,310
- 2023SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesiscited by 4,290
- 2021GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Modelscited by 3,595
- 2022Reproducible scaling laws for contrastive language-image learningcited by 1,163
- 2022eDiff-I: Text-to-Image Diffusion Models with an Ensemble of Expert Denoiserscited by 921
- 2022What the DAAM: Interpreting Stable Diffusion Using Cross Attentioncited by 233
7 more
- 2023Adversarial Example Does Good: Preventing Painting Imitation from Diffusion Models via Adversarial Examplescited by 163
- 2023Raising the Cost of Malicious AI-Powered Image Editingcited by 160
- 2023A Pilot Study of Query-Free Adversarial Attack against Stable Diffusioncited by 75
- 2022Analyzing Encoded Concepts in Transformer Language Modelscited by 37
- 2022Adversarial Attacks on Image Generation With Made-Up Wordscited by 29
- 2023Discovering Failure Modes of Text-guided Diffusion Models via Adversarial Searchcited by 25
- 2023Evaluating the Robustness of Text-to-image Diffusion Models against Real-world Attackscited by 24
Built on by
Citing papers per year: 2024: 1; 2025: 1.
2 papers.