Cihan University-Erbil Conferences, The 3rd International Conference on Language and Education at Cihan University-Erbil

Font Size: 
A Synchronic Assessment of Google Translate, Microsoft Translator and Reverso in English<>Arabic Translation

Last modified: 2023-07-20


The study aimed to evaluate three multilingual MT systems that provide free English<> Arabic translation services. It assessed the effectiveness of the three systems in dealing with WHO, UN, Petra News Agency, and Arab League documents.  The study adopted (LDC, 2002) intelligibility and fidelity scales to highlight the limitations and strengths of the systems understudy. The study found that Google Translate outperformed the three systems in providing intelligible and fluent output without flaws with 230 segments, equaling 57% of the corpus size, followed by Microsoft Translator with 48% and Reverso with 45% translation free of errors. On the other hand, the study found that Google Translate outperformed the other systems in terms of adequacy and fluency scales, followed by Microsoft Translator and Reverso interchangeably. However, the three systems are still unable to substitute human translators. Therefore, the study recommends conducting a diachronic study to trace the three systems' development over five years scientifically.

Keywords— Error Analysis; Fidelity; Intelligibility; Machine Translation.


Published Date: 01/06/2023


Full Text: PDF