Development of a system for creation, computer-aided alignment, and search in parallel corpora of low-resource languages

Abstract

This work presents the development of a web system for alignment, management, and search of parallel corpus, focusing on machine translation applications and linguistic studies. The system allows users to upload texts in different languages, perform automatic or manual alignment of corresponding sentences, and publicly share their corpora for research. The creation of high-quality parallel corpora is essential for Natural Language Processing (NLP) applications, yet low-resource languages face significant challenges due to data scarcity and linguistic diversity. By combining human expertise with computational efficiency, this system bridges the gap between scalability and accuracy in corpus creation. It is particularly suited for under-resourced languages, where fully automated methods struggle. The modular architecture of the system allows integration with existing NLP pipelines, fostering applications in machine translation, language preservation, and digital inclusion.