Russian text simplification model matches language learning levels closely
CEFR-oriented Russian Text Simplification Using Reinforcement Learning
Computation and Language
Summary
Simplifying texts so they match language learners’ abilities helps them understand and learn better. This work improves a Russian language AI model to rewrite texts at easier levels defined by a common language proficiency scale. The authors trained the model using human preferences to better control text complexity, making sure the rewritten text stays clear and meaningful while using simpler words and grammar. Their best model successfully targets specific learner levels most of the time, especially for intermediate learners, and the authors made their tool freely available online.
What this means in practice
- •For language learning platforms: Automatically simplify Russian texts to match specific learner proficiency levels improving personalized learning materials.
- •For content localization teams: Adapt Russian content complexity on demand to better fit target audience language levels while preserving meaning.
Authors
Kristina Mustakova, Alexey Sorokin
Abstract
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 $\href{https://huggingface.co/kmustakova/RuSimplifier_DPO}{HuggingFace}$.