A Comparative Analysis of Machine Translation Approaches for Low-Resource Languages and Future Enhancements
D. I. De Silva, D. G. P. Hansadi · 2024
This paper provides a comparative analysis of machine translation and its evolution, with a specific focus on the challenges and advancements in translating three low-resource languages. It begins by exploring the emergence of machine translation and compares it with human translation, highlighting their respective strengths and weaknesses. The paper further discusses the evolution of various machine translation approaches, including rule-based, statistical, neural, and hybrid methods, offering a detailed contrast of these models. A comparative analysis of existing research is conducted, followed by a review of findings and an evaluation of translation quality through quantitative and qualitative metrics. The paper concludes by proposing future directions for machine translation, specifically addressing the translation of low-resource languages. Suggested solutions include the integration of data augmentation techniques, transfer learning, and hybrid approaches to enhance translation quality, mitigate resource limitations, and improve the overall effectiveness of machine translation systems for low-resource language pairs.