Benchmarking Dialectal Arabic-Turkish Machine Translation

Proceedings of Machine Translation Summit XIX, Vol. 1: Research Track, pages 261–271 September 4–8, 2023, Macau SAR, China. ©2023 The authors. This article is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0). 261 Benchmarking Dialectal Arabic-Turkish Machine Translation Hasan Alkheder hasan.alkhder2@ogr.sakarya.edu.tr Computer Engineering, Sakarya University, Serdivan, T¨urkiye Houda Bouamor hbouamor@cmu.edu Information Systems, Carnegie Mellon University Qatar, Doha, Qatar Nizar Habash nizar.habash@nyu.edu Computer Science, New York University Abu Dhabi, Abu Dhabi, UAE Ahmet Zengin azengin@sakarya.edu.tr Computer Engineering, Sakarya University, Serdivan, T¨urkiye Abstract Due to the significant influx of Syrian refugees in Turkey in recent years, the Syrian Arabic dialect has become increasingly prevalent in certain regions of Turkey. Developing a machine translation system between Turkish and Syrian Arabic would be crucial in facilitating com- munication between the Turkish and Syrian communities in these regions, which can have a positive impact on various domains such as politics, trade, and humanitarian aid. Such a sys- tem would also contribute positively to the growing Arab-focused tourism industry in Turkey. In this paper, we present the first research effort exploring translation between Syrian Arabic and Turkish. We use a set of 2,000 parallel sentences from the MADAR corpus containing 25 different city dialects from different cities across the Arab world, in addition to Modern Stan- dard Arabic (MSA), English, and French. Additionally, we explore the translation performance into Turkish from other Arabic dialects and compare the results to the performance achieved when translating from Syrian Arabic. We build our MADAR-Turk data set by manually trans- lating the set of 2,000 sentences from the Damascus dialect of Syria to Turkish with the help of two native Arabic speakers from Syria who are also highly fluent in Turkish. We evaluate the quality of the translations and report the results achieved. We make this first-of-a-kind data set publicly available to support research in machine translation between these important but less studied language pairs.1 1 Introduction The rapid advancements in machine translation technology have significantly helped to break down language barriers and facilitate cross-cultural communication using distant languages, including Turkish and Arabic. Given that Syria and Iraq border Turkey to the south, there is cultural overlap and close ties between those nations and Turkey, making Arabic and Turkish two of the most widely spoken languages in the Middle East. Despite this, there has been no significant research or machine translation effort that specifically addresses the translation of 1The MADAR-Turk data set is available from http://resources.camel-lab.com/. 262 dialectal Arabic and Turkish. The presence of more than 4 million Syrian and Iraqi refugees in Turkey, as well as the massive spread of Turkish drama (dubbed TV series) in the Arab world (Kraidy and Al-Ghazzi, 2013), not to mention the growing tourism industry in Turkey catering to Arab tourists, reinforce the urgent need for developing machine translation capa- bilities between Turkish and Arabic and its dialects to promote communication and cultural exchange between the Arab countries and Turkey. Efforts to develop neural machine translation between several language pairs …

Content truncated to 500 words