@inproceedings{wu-etal-2025-uva,
    title = "{U}v{A}-{MT}{'}s Participation in the {WMT}25 General Translation Shared Task",
    author = "Wu, Di  and
      Meng, Yan  and
      Nachesa, Maya Konstantinovna  and
      Aycock, Seth  and
      Monz, Christof",
    editor = "Haddow, Barry  and
      Kocmi, Tom  and
      Koehn, Philipp  and
      Monz, Christof",
    booktitle = "Proceedings of the Tenth Conference on Machine Translation",
    month = nov,
    year = "2025",
    address = "Suzhou, China",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-emnlp/2025.wmt-1.45/",
    pages = "688--694",
    ISBN = "979-8-89176-341-8",
    abstract = "This paper presents UvA-MT{'}s submission to the WMT 2025 shared task on general machine translation, competing in the unconstrained track across all 16 translation directions. Unusually, this year we use only WMT25{'}s blind test set (source sentences only) to generate synthetic data for LLM training, and translations are produced using pure beam search for submission. Overall, our approach can be seen as a special variant of data distillation, motivated by two key considerations: (1) perfect domain alignment, where the training and test domains are distributionally identical; and (2) the strong teacher model, GPT-4o-mini, offers high-quality outputs as both a reliable reference and a fallback in case of mere memorization.Interestingly, the outputs of the resulting model, trained on Gemma3-12B using Best-of-N (BoN) outputs from GPT-4o-mini, outperform both original BoN outputs from GPT-4o-mini and Gemma3-12B in some high-resource languages across various metrics. We attribute this to a successful model ensemble, where the student model (Gemma3-12B) retains the strengths of the teacher (GPT-4o-mini) while implicitly avoiding their flaws."
}