@inproceedings{liu-etal-2025-vla,
    title = "{VLA}-Mark: A cross modal watermark for large vision-language alignment models",
    author = "Liu, Shuliang  and
      Qi, Zheng  and
      Xu, Jesse Jiaxi  and
      Yan, Yibo  and
      Zhang, Junyan  and
      Geng, He  and
      Liu, Aiwei  and
      Jiang, Peijie  and
      Liu, Jia  and
      Tam, Yik-Cheung  and
      Hu, Xuming",
    editor = "Christodoulopoulos, Christos  and
      Chakraborty, Tanmoy  and
      Rose, Carolyn  and
      Peng, Violet",
    booktitle = "Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing",
    month = nov,
    year = "2025",
    address = "Suzhou, China",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-emnlp/2025.emnlp-main.1342/",
    pages = "26420--26438",
    ISBN = "979-8-89176-332-6",
    abstract = "Vision-language models demand watermarking solutions that protect intellectual property without compromising multimodal coherence. Existing text watermarking methods disrupt visual-textual alignment through biased token selection and static strategies, leaving semantic-critical concepts vulnerable. We propose VLA-Mark, a vision-aligned framework that embeds detectable watermarks while preserving semantic fidelity through cross-modal coordination. Our approach integrates multiscale visual-textual alignment metrics, combining localized patch affinity, global semantic coherence, and contextual attention patterns, to guide watermark injection without model retraining. An entropy-sensitive mechanism dynamically balances watermark strength and semantic preservation, prioritizing visual grounding during low-uncertainty generation phases. Experiments show 7.4{\%} lower PPL and 26.6{\%} higher BLEU than conventional methods, with near-perfect detection (98.8{\%} AUC). The framework demonstrates 96.1{\%} attack resilience against attacks such as paraphrasing and synonym substitution, while maintaining text-visual consistency, establishing new standards for quality-preserving multimodal watermarking."
}