@inproceedings{wang-etal-2025-vlminferslow,
    title = "{VLMI}nfer{S}low: Evaluating the Efficiency Robustness of Large Vision-Language Models as a Service",
    author = "Wang, Xiasi  and
      Yao, Tianliang  and
      Chen, Simin  and
      Wang, Runqi  and
      Ye, Lei  and
      Gao, Kuofeng  and
      Huang, Yi  and
      Yao, Yuan",
    editor = "Che, Wanxiang  and
      Nabende, Joyce  and
      Shutova, Ekaterina  and
      Pilehvar, Mohammad Taher",
    booktitle = "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2025",
    address = "Vienna, Austria",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingestion-acl-25/2025.acl-long.781/",
    pages = "16035--16050",
    ISBN = "979-8-89176-251-0",
    abstract = "Vision-Language Models (VLMs) have demonstrated great potential in real-world applications. While existing research primarily focuses on improving their accuracy, the efficiency remains underexplored. Given the real-time demands of many applications and the high inference overhead of VLMs, efficiency robustness is a critical issue. However, previous studies evaluate efficiency robustness under unrealistic assumptions, requiring access to the model architecture and parameters{---}an impractical scenario in ML-as-a-service settings, where VLMs are deployed via inference APIs. To address this gap, we propose VLMInferSlow, a novel approach for evaluating VLM efficiency robustness in a realistic black-box setting. VLMInferSlow incorporates fine-grained efficiency modeling tailored to VLM inference and leverages zero-order optimization to search for adversarial examples. Experimental results show that VLMInferSlow generates adversarial images with imperceptible perturbations, increasing the computational cost by up to 128.47{\%}. We hope this research raises the community{'}s awareness about the efficiency robustness of VLMs."
}