SRS-Stories: Vocabulary-constrained multilingual story generation for language learning

TL;DR

This paper presents a method for generating personalized, vocabulary-constrained stories in multiple languages to aid language learning, combining large language models with lexical constraints and spaced repetition.

Contribution

It introduces a novel approach for multilingual story generation that integrates vocabulary constraints and optimizes learning through spaced repetition systems.

Findings

Generated stories are more grammatical and coherent.

Stories effectively teach new vocabulary in context.

The method outperforms standard constrained beam search.

Abstract

In this paper, we use large language models to generate personalized stories for language learners, using only the vocabulary they know. The generated texts are specifically written to teach the user new vocabulary by simply reading stories where it appears in context, while at the same time seamlessly reviewing recently learned vocabulary. The generated stories are enjoyable to read and the vocabulary reviewing/learning is optimized by a Spaced Repetition System. The experiments are conducted in three languages: English, Chinese and Polish, evaluating three story generation methods and three strategies for enforcing lexical constraints. The results show that the generated stories are more grammatical, coherent, and provide better examples of word usage than texts generated by the standard constrained beam search approach