Fine-tuning multilingual sentence transformers for low-resource skill-concept matching across Kazakh, Russian, and English
Sandugash Serikbayeva, Madina Sambetbayeva, Valiya Ramazanova, Aigerim Yerimbetova, Zhanar Lamasheva, Zhanna Sadirmekova, Ardak Batyrkhanov, Yersaiyn Mailybayev · 2026
Read Fine-tuning multilingual sentence transformers for low-resource skill-concept matching across Kazakh, Russian, and English by Sandugash Serikbayeva, Madina Sambetbayeva, Valiya Ramazanova, Aigerim Yerimbetova, Zhanar Lamasheva, Zhanna Sadirmekova, Ardak Batyrkhanov, Yersaiyn Mailybayev (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13600825.txt.
Description
Serikbayeva and colleagues fine-tune in twenty twenty-six a multilingual text encoder to match skill phrases across English, Russian, and Kazakh. Short skill labels are terse, mixed-language, and domain-specific, so general encoders stumble, above all in low-resource Kazakh. The team trained the model in stages on a three-language paraphrase set, with extra Russian pairs.
English skill matching held steady while Russian and Kazakh scores rose, Kazakh most on one measure. Grouping skills into a taxonomy also grew cleaner, with far fewer wrong clusters. The authors read the Kazakh gain as cross-language transfer from its richer neighbours. They advise setting match thresholds per language in real hiring tools.
Catalog record
| CREATOR | Sandugash Serikbayeva, Madina Sambetbayeva, Valiya Ramazanova, Aigerim Yerimbetova, Zhanar Lamasheva, Zhanna Sadirmekova, Ardak Batyrkhanov, Yersaiyn Mailybayev |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 82 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |