Browsing by Author "Mager Hois, Jesus Manuel"

Filter results by typing the first few letters
Now showing 1 - 1 of 1
  • Thumbnail Image
    ItemOpen Access
    Low resource NLP for polysynthetic languages : morphological segmentation and machine translation
    (2024) Mager Hois, Jesus Manuel; Vu, Ngoc Thang (Prof. Dr.)
    This thesis explores the application of Natural Language Processing (NLP) techniques to morphologically rich indigenous languages of the Americas, focusing on low-resource scenarios. The work addresses the challenges of modeling morphological segmentation and machine translation for these languages, often lacking large annotated datasets and facing issues like code-switching and orthographic normalization. Contributions include the development of new datasets, the adaptation of neural network models for specific tasks, and the investigation of the impact of morphological segmentation on machine translation performance. Additionally, the thesis delves into the ethical implications of applying NLP technologies to these languages, considering the perspectives of native speakers and community leaders.