Alexandria: A Multi-Domain Dialectal Arabic Machine Translation Dataset for Culturally Inclusive and Linguistically Diverse LLMs
Abstract
Arabic is a highly diglossic language where most daily communication occurs in regional dialects rather than Modern Standard Arabic (MSA). Despite this, machine translation (MT) systems often generalize poorly to dialectal input, limiting their utility for millions of speakers. We introduce Alexandria, a large-scale, community-driven, human-translated dataset designed to bridge this gap. Alexandria covers 13 Arab countries and 11 high-impact domains, including health, education, and agriculture. Unlike previous resources, Alexandria provides unprecedented granularity by associating contributions with city-of-origin metadata, capturing authentic local varieties beyond coarse regional labels. The dataset consists of parallel English-Dialectal Arabic multi-turn conversational scenarios annotated with speaker-addressee gender configurations, enabling the study of gender-conditioned variation in dialectal use. Comprising 107K total turns, Alexandria serves as both a training resource and as a rigorous benchmark for evaluating MT and Large Language Models (LLMs). Our automatic and human evaluation benchmarks the current capabilities of Arabic-aware LLMs in translating across diverse Arabic dialects and sub-dialects while exposing significant persistent challenges. The Alexandria dataset, the creation prompts, the translation and revision guidelines, and the evaluation code are publicly available in the following repository: https://github.com/UBC-NLP/Alexandria
Cite
@article{arxiv.2601.13099,
title = {Alexandria: A Multi-Domain Dialectal Arabic Machine Translation Dataset for Culturally Inclusive and Linguistically Diverse LLMs},
author = {Abdellah El Mekki and Samar M. Magdy and Houdaifa Atou and Ruwa AbuHweidi and Baraah Qawasmeh and Omer Nacar and Thikra Al-hibiri and Razan Saadie and Hamzah Alsayadi and Nadia Ghezaiel Hammouda and Alshima Alkhazimi and Aya Hamod and Al-Yas Al-Ghafri and Wesam El-Sayed and Asila Al sharji and Mohamad Ballout and Anas Belfathi and Karim Ghaddar and Serry Sibaee and Alaa Aoun and Areej Asiri and Lina Abureesh and Ahlam Bashiti and Majdal Yousef and Abdulaziz Hafiz and Yehdih Mohamed and Emira Hamedtou and Brakehe Brahim and Rahaf Alhamouri and Youssef Nafea and Aya El Aatar and Walid Al-Dhabyani and Emhemed Hamed and Sara Shatnawi and Fakhraddin Alwajih and Khalid Elkhidir and Ashwag Alasmari and Abdurrahman Gerrio and Omar Alshahri and AbdelRahim A. Elmadany and Ismail Berrada and Amir Azad Adli Alkathiri and Fadi A Zaraket and Mustafa Jarrar and Yahya Mohamed El Hadj and Hassan Alhuzali and Muhammad Abdul-Mageed},
journal= {arXiv preprint arXiv:2601.13099},
year = {2026}
}
Comments
Accepted to ACL 2026 Main; Project resources will be available here: https://github.com/UBC-NLP/Alexandria