Machine Translation of Dialects into Standard Arabic: A Systematic Review and Comparative Analysis

Abstract

This paper provides an in-depth examination of the linguistic, technological, and methodological challenges of translating Moroccan Darija into Standard Arabic. It underscores the inherent complexity of Darija, which is shaped by significant regional variations, a diverse multilingual vocabulary, and the absence of a standardized written form. Through a critical and structured review of recent research, the paper presents a state-of-the-art overview of existing translation approaches, including rule-based, statistical, and neural methods. A comparative table summarizes the most important contributions, detailing the datasets used, the techniques applied, the results achieved, and their respective strengths and limitations. Furthermore, the analysis explores the potential of large language models (LLMs) as a promising direction, while emphasizing the need for highquality parallel corpus and context-aware fine-tuning to improve translation accuracy and adaptability across different scenarios. The study also provides practical recommendations aimed at guiding future research and fostering the development of robust translation tools tailored to the unique characteristics of Darija.