Year: 2026
Type: article
Abstract: Machine translation (MT) has become an imperative facilitator of the increase in availability of culturally enriched learning resources in multilingual learning settings. Folk narratives as carriers of indigenous knowledge, moral codes, and linguistic innovation are still overrepresented in the informal sphere of learning because, despite the language barrier and the lack of quality translations, the number of them is limited. This paper will suggest a special machine translation system to folk stories in education, both linguistically and culturally. Based on the progress in neural machine translation, the model combines domain-adaptive training, cultural embeddings and human-machine training to assist low-resource and high-context languages. A systematized multilingual folk narrative collection of folk narratives is created by conducting a systematic collection of the folk narratives, marking them with the cultural markers, and recognizing the figurative expressions, metaphors, and oral storytelling patterns. It is evaluated by standard metrics of the MT such as BLEU, METEOR, and TER as well as the suggested cultural fidelity and education usability scores to measure narrative coherence, pedagogical relevance, and the understanding of learners. The findings show that the performance of folklore-adapted models is much better than generic NMT systems, especially when it comes to maintaining culturally significant expressions and plot structure.
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
SDG: Quality education
Open Access status: diamond
OpenAlex ID: W7130506337