Adversarial Attacks on Parts of Speech: An Empirical Study in Text-to-Image Generation
This work addresses vulnerabilities in text-to-image generation for AI security, but it is incremental as it builds on known adversarial attack methods.
The study investigated how adversarial attacks on different parts of speech (POS) tags affect text-to-image models, finding that nouns, proper nouns, and adjectives are the easiest to attack with varying success rates.
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.