@inproceedings{li-etal-2026-self,
    title = "Self-Supervised On-Policy Distillation for Omni-Modal Video Captioning",
    author = "Li, Zinuo  and
      Guo, Yongxin  and
      Li, Xin  and
      Tan, Hao  and
      Zhong, Zhao  and
      Bo, Liefeng  and
      Bennamoun, Mohammed  and
      Boussaid, Farid  and
      Ke, Qiuhong",
    editor = "Lee, Hung-Yi  and
      Dev, Sunipa  and
      Li, Jessy  and
      Elliott, Desmond",
    booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
    month = oct,
    year = "2026",
    address = "Budapest, Hungary",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2026.emnlp-main.651/",
    pages = "14974--14995",
    ISBN = "979-8-89176-459-0"
}