@inproceedings{349ac33dc3f24124b2f95d76cc6fb7e8,
title = "QADI: Arabic Dialect Identification in the Wild",
abstract = "Proper dialect identification is important for a variety of Arabic NLP applications. In this paper, we present a method for rapidly constructing a tweet dataset containing a wide range of country-level Arabic dialects-covering 18 different countries in the Middle East and North Africa region. Our method relies on applying multiple filters to identify users who belong to different countries based on their account descriptions and to eliminate tweets that either write mainly in Modern Standard Arabic or mostly use vulgar language. The resultant dataset contains 540k tweets from 2,525 users who are evenly distributed across 18 Arab countries. Using intrinsic evaluation, we show that the labels of a set of randomly selected tweets are 91.5% accurate. For extrinsic evaluation, we are able to build effective countrylevel dialect identification on tweets with a macro-averaged F1-score of 60.6% across 18 classes.",
author = "Ahmed Abdelali and Hamdy Mubarak and Younes Samih and Sabit Hassan and Kareem Darwish",
note = "Publisher Copyright: {\textcopyright} WANLP 2021 - 6th Arabic Natural Language Processing Workshop; 6th Arabic Natural Language Processing Workshop, WANLP 2021 ; Conference date: 19-04-2021",
year = "2021",
language = "English",
series = "WANLP 2021 - 6th Arabic Natural Language Processing Workshop, Proceedings of the Workshop",
publisher = "Association for Computational Linguistics (ACL)",
pages = "1--10",
editor = "Nizar Habash and Houda Bouamor and Hazem Hajj and Walid Magdy and Wajdi Zaghouani and Fethi Bougares and Nadi Tomeh and Farha, {Ibrahim Abu} and Samia Touileb",
booktitle = "WANLP 2021 - 6th Arabic Natural Language Processing Workshop, Proceedings of the Workshop",
address = "United States",
}