@inproceedings{atuhurra-etal-2026-vlures,
    title = "{VLUR}es: Benchmarking Long-Text Grounding and Cross-Lingual Robustness in Vision Language Models",
    author = "Atuhurra, Jesse  and
      Ali, Iqra  and
      Iwakura, Tomoya  and
      Kamigaito, Hidetaka  and
      Hiraoka, Tatsuya",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-nlpsi/2026.findings-acl.1367/",
    doi = "10.18653/v1/2026.findings-acl.1367",
    pages = "27426--27481",
    ISBN = "979-8-89176-395-1",
    abstract = "We introduce \textit{\textbf{VLURes}}, a multilingual benchmark for evaluating Vision-Language Models (VLMs) under \textit{long-text grounding}: selecting and reasoning over the image-relevant subset of article-length text that contains distractors and ungrounded claims. \textit{VLURes} contains \textbf{4,000} web-curated \textit{image + long-text} pairs across \textbf{English (En), Japanese (Ja), Swahili (Sw), and Urdu (Ur)} and \textbf{10} topical categories, and defines \textbf{eight} tasks spanning image-only perception (OR, SU, RU, SS, IC) and image+text grounding (ITM, \textit{Unrelatedness}, VQA). To construct web-realistic pairs, we apply language-adapted CLIP alignment to select representative images and filter weakly grounded pages. Across \textbf{10} proprietary and open VLMs evaluated under zero-shot and one-shot prompting, with and without rationales, the best model (GPT-4o) reaches \textbf{90.8{\%}} overall accuracy but remains \textbf{6.7} points below human performance (\textbf{97.5{\%}}) on Object Recognition, and cross-lingual sensitivity persists, while open models are substantially weaker and often lack reliable multilingual VL support. \textit{VLURes} provides a practical testbed for long-text grounding and multilingual robustness in web-realistic agent settings."
}