Metraipolis › archive › dc_eaf13838b248
CEFR-oriented Russian Text Simplification Using Reinforcement Learning — arXiv:2610.27237
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.27237
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-21T07:08:50.985Z
- SHA-256 of the archived text
21e7dce155848d1e0a8fa6cdcf6c1260d5e280343bba045a254b3fd982c242e6- Stable URL of this page
https://metraipolis.com/documento/dc_eaf13838b248
How to cite
CEFR-oriented Russian Text Simplification Using Reinforcement Learning — arXiv:2610.27237 — arXiv (en), https://arxiv.org/abs/2610.27237 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_eaf13838b248 (Metraipolis archive dc_eaf13838b248, sha256 21e7dce155848d1e0a8fa6cdcf6c1260d5e280343bba045a254b3fd982c242e6)
BibTeX
@misc{dceaf13838b248,
title = {CEFR-oriented Russian Text Simplification Using Reinforcement Learning — arXiv:2610.27237},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.27237},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_eaf13838b248; sha256 21e7dce155848d1e0a8fa6cdcf6c1260d5e280343bba045a254b3fd982c242e6},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.27237. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
We present a study on controlled automatic simplification of Russian texts using preference-based optimization algorithms with a focus on adapting texts to the A1, A2 and B1 proficiency levels. We fine-tune YandexGPT using Rejection Sampling and Direct Preference Optimization (DPO) methods to improve control over the complexity of the generated text. We compare several training configurations that differ in dataset composition, text type and the strategy used to construct preference pairs for DPO. For automatic evaluation we train a regression model based on ruRoBERTa-large to predict continuous complexity scores corresponding to CEFR levels. We also consider semantic similarity, lexical coverage and quantitative readability metrics. Human evaluation assesses simplicity, fluency and adequacy. Our best model achieves 87.15\% target level accuracy, an MAE of 0.35 CEFR levels and a target interval MAE of 0.11. Manual assessment shows that 75% of the evaluated simplifications match the target morphological and syntactic level, and fluency and adequacy scores are especially high for A2 and B1. Although A1 remains the most challenging level, the model significantly reduces the proportion of vocabulary and grammatical constructions beyond the target level while generally preserving meaning. Our model is freely available on $\href{ this https URL }{HuggingFace}$.
Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2610.27237 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-21T07:05:58.104Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).