Cross-Lingual Transfer for Machine Translation in Turkic Languages

arXiv:2607.29355 · cs.CL, cs.AI · Submitted 2026-07-31 · Read on arXiv

Omer Burak Cinar, Mehmet Mert Dalkilic, Cagri Toraman

cs.CL, cs.AI

Submitted: 2026-07-31

Code: https://github.com/TurkicTransfer/Cross-LingualTransferforMachineTranslationinTurkicLanguages

License: http://creativecommons.org/licenses/by-nc-sa/4.0/

The gist: Cross-lingual transfer is central to low-resource machine translation, but its behavior within closely related language families remains insufficiently characterized.

Terminology

Abstract

Cross-lingual transfer is central to low-resource machine translation, but its behavior within closely related language families remains insufficiently characterized. We study transfer among five Turkic languages; Turkish, Azerbaijani, Uzbek, Kazakh, and Kyrgyz; using pairwise transfer matrices. In this setting, each model is fine-tuned with one transfer source and evaluated on a different transfer target while the translation target remains the same. Across mT5 experiments, we find that transfer is strongest between closely related Turkic pairs, especially Turkish-Azerbaijani and Kazakh-Kyrgyz. We also show that transfer direction matters, and that the same transfer source-transfer target pair can behave differently when the translation target changes. Latinization improves BLEU and chrF in several script-mismatched settings, but its effect is not uniform across metrics. Additional analyses show that transfer sources are mostly stable across different datasets and model settings.

Sources

Related papers