Ukrainian-to-English folktale corpus: Parallel corpus creation and augmentation for machine translation in low-resource languages
This work provides a resource for machine translation in low-resource languages, specifically for Ukrainian folktales, though it is incremental as it focuses on corpus creation rather than novel methods.
The authors created a Ukrainian-to-English parallel corpus of folktales to address the scarcity of translated texts, using a domain-specific approach for building and augmenting it, resulting in a word- and sentence-aligned dataset tailored for machine translation training.
Folktales are linguistically very rich and culturally significant in understanding the source language. Historically, only human translation has been used for translating folklore. Therefore, the number of translated texts is very sparse, which limits access to knowledge about cultural traditions and customs. We have created a new Ukrainian-To-English parallel corpus of familiar Ukrainian folktales based on available English translations and suggested several new ones. We offer a combined domain-specific approach to building and augmenting this corpus, considering the nature of the domain and differences in the purpose of human versus machine translation. Our corpus is word and sentence-aligned, allowing for the best curation of meaning, specifically tailored for use as training data for machine translation models.