@inproceedings{zhao-etal-2025-looking,
    title = "Looking Beyond Text: Reducing Language Bias in Large Vision-Language Models via Multimodal Dual-Attention and Soft-Image Guidance",
    author = "Zhao, Haozhe  and
      Si, Shuzheng  and
      Chen, Liang  and
      Zhang, Yichi  and
      Sun, Maosong  and
      Chang, Baobao  and
      Zhang, Minjia",
    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/name-variant-enfa-fane/2025.emnlp-main.995/",
    doi = "10.18653/v1/2025.emnlp-main.995",
    pages = "19677--19701",
    ISBN = "979-8-89176-332-6",
    abstract = "Large vision-language models (LVLMs) have achieved impressive results in vision-language tasks. However, Therefore, we propose LACING, designed to address such bias with Mu$\underline{\textbf{L}}$timodal Du$\underline{\textbf{A}}$l-attention Me$\underline{\textbf{C}}$han$\underline{\textbf{I}}$sm (MDA) a$\underline{\textbf{N}}$d Soft-Image $\underline{\textbf{G}}$uidance (SIG). Specifically, MDA adopts a $\textbf{parallel dual-attention mechanism}$ that constructs separate attention for visual and text inputs to enhance integration of visual inputs across model. SIG uses a $\textbf{learnable soft visual prompt}$ during training and inference to replace visual inputs, designed to compel LVLMs to prioritize text inputs during inference. Experiments across different model architectures and scales demonstrate that LACING effectively debiases LVLMs from their language bias, enhancing visual comprehension and reducing hallucinations without additional resources."
}