@inproceedings{quispe-callanaupa-etal-2026-optimization,
    title = "Optimization of Voice Translation Systems for Indigenous Languages: Retraining the {NLLB}-200 Model for the {Q}uechua{--}{S}panish Pair",
    author = "Quispe Calla{\~n}aupa, Mitzuko Davis  and
      Toledo Bernal, Max Erixon  and
      Torres Bautista, Ronil Nilo  and
      Pumaccahua Huallpa, Patrick Michael",
    editor = "Salesky, Elizabeth  and
      Anastasopoulos, Antonios  and
      Negri, Matteo  and
      Federico, Marcello",
    booktitle = "Proceedings of the 23rd International Conference on Spoken Language Translation ({IWSLT} 2026)",
    month = jul,
    year = "2026",
    address = "San Diego, USA (in-person and online)",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/bulk-corrections-2026-07-02/2026.iwslt-1.35/",
    doi = "10.18653/v1/2026.iwslt-1.35",
    pages = "316--317",
    ISBN = "979-8-89176-411-8",
    abstract = "This article describes the fine-tuning and incremental retraining process of the massive NLLB-200 model applied to the Quechua (Chanka and Collao variants) and Spanish language pair. Using a curated dataset of 22,891 parallel pairs, a robust cleaning strategy and optimized training for consumer hardware (NVIDIA RTX 3060) were implemented. The results demonstrate a progressive improvement in the BLEU metric, reaching a competitive state for translation tasks in low-resource scenarios, in line with the challenges posed by the IWSLT 2026 shared task."
}