Neural Machine Translation for the Arabic-English Language Pair
Ahmed Amine Aliane, Nasreddine Semmar, Hassina Aliane · 2024
Machine Translation is an early Artificial Intelligence field of research that has a rich history and a substantial body of literature. If statistical methods dominated over two decades, the success of neural approaches opened up new promises, particularly for languages with rich morphology and limited resources like Arabic. We investigate through this research Arabic-English Machine Translation within the new paradigm of Neural Machine Translation. We describe and discuss our implementation of both traditional models, LSTM and F-Conv, and the recent transformer model. If we report fairly honorable results for all models, the transformer model unsurprisingly achieves an outstanding BLEU-score.