| Name | Matric Number |
|---|---|
| TAFA TAOFIK OLASUNKANMI | PEC233041 |
| YAP QI YUAN | MCS231025 |
| CAI HUA ZHU |
https://drive.google.com/drive/folders/1oPSHeq-5X-IiXxpqdSA3K-tv_uvJIK_A?usp=drive_link
This systematic literature review examines the current overview of machine translation (MT) performance for low-resourced and under-resourced languages, addressing the significant challenges and limitations these languages face due to limited digital resources and parallel corpora. Despite advancements in neural machine translation (NMT), the effectiveness of these MT systems on these languages remains constrained, necessitating a comprehensive synthesis of existing research to identify effective strategies and methodologies. The review follows PRISMA guidelines, systematically searching multiple academic databases for studies published between 2019 and 2023. A total of 69 relevant articles were studied to assess the MT performance, the challenges encountered, and the effectiveness of used or proposed solutions. Findings from the study reveal that while NMT has emerged as a dominant approach, its performance is often stalled by the scarcity of training data and Structural variability. The review highlights the importance of leveraging monolingual data and adapting model architectures to improve translation accuracy. Additionally, it identifies trends in research area, including the integration of multilingual sentiment analysis and speech-to-text translation systems. This study offers a thorough examination of the present state of MT for low-resourced and under-resourced languages. Furthermore, it emphasizes the necessity for future research to investigate underrepresented languages and construct datasets that encompass a wider range of languages. By fostering collaboration and innovation in this field, the study aims to contribute to more effective and equitable translation solutions, ultimately enhancing communication across diverse linguistic communities.
Keywords: Machine translation, low-resourced languages, under-resourced languages
Please create an Issue for any improvements, suggestions or errors in the content.