Bridging the Resource Gap for Malay-to-Arabic Translation: Evaluating Machine Translation of News Headlines

Abstract

Machine translation (MT) technology has become essential for cross-lingual communication, surpassing traditional human-centric translation methods. Many translation studies and MT developments have focused primarily on English, leading to significant improvements in MT systems for English. However, the quality may not be as reliable when translating between less widely spoken languages, such as Malay and Arabic, due to the scarcity of resources and research on improving MT for these language pairs. While MT systems provide good translations for widely spoken languages like English, there is a need for more research and development to improve the quality of translation for less common language pairs like Malay–Arabic. This study aims to address this gap by focusing on translating Malay news headlines into Arabic, contributing to improving MT systems for these language pairs, and providing a resource for translation students and professionals, where Arabic translation materials can be scarce, especially in Malaysian institutions. This study used a mixed-methods approach, integrating quantitative analysis of 20 news headlines scored on accuracy, style, and clarity by evaluators with 5-27 years of expertise in Arabic language and translation. A qualitative thematic analysis was conducted by the researcher to achieve the aim of this study. Results showed significant variations in MT system performance. While some systems preserved linguistic features and accuracy, cultural nuances were often lost, with common errors in idioms and structure. This study evaluates user feedback on Google Translate (GT) and Bing Microsoft Translator (BMT). Due to unequal participant distribution, potential bias exists. The findings highlight the need for advanced MT tools for Malay-Arabic translation and enhance MT technology, promoting cross-cultural understanding in the news industry. Future studies should aim for a more balanced sample size for better comparability.