@inproceedings{an-etal-2026-toward,
    title = "Toward Native Multimodal Modeling: A Roadmap",
    author = "An, Siyu  and
      Lu, Junru  and
      Dong, Junnan  and
      Wang, Qiufeng  and
      Li, Yinghui  and
      Fei, Weizhi  and
      Yu, Zichao  and
      Yuan, Zheng  and
      Liu, Biao  and
      Wang, Haopeng  and
      Liang, Renzhao  and
      Yang, Yixuan  and
      Shen, Yunhang  and
      Ke, Bo  and
      Chen, Keyu  and
      Luo, Linhao  and
      Zou, Difan  and
      Huang, Xiao  and
      Yin, di  and
      Qiao, Ruizhi  and
      Sun, Xing",
    editor = "Lee, Hung-Yi  and
      Dev, Sunipa  and
      Li, Jessy  and
      Elliott, Desmond",
    booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
    month = oct,
    year = "2026",
    address = "Budapest, Hungary",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2026.findings-emnlp.220/",
    pages = "4594--4644",
    ISBN = "979-8-89176-462-0"
}