{ "Name": "ADH", "Volume": 552.0, "Unit": "sentences", "License": "MIT License", "Link": "https://github.com/saleml/arabic-dialect-hub", "HF_Link": "", "Year": 2024, "Domain": [ "other" ], "Form": "text", "Collection_Style": [ "LLM generated", "human annotation" ], "Description": "Cross-dialectal Arabic learning resource.", "Ethical_Risks": "Low", "Provider": [ "Mohamed Bin Zayed University of Artificial Intelligence" ], "Derived_From": [], "Paper_Title": "ArabicDialectHub: A Cross-Dialectal Arabic Learning Resource and Platform", "Paper_Link": "https://aclanthology.org/2026.abjadnlp-1.1.pdf", "Tokenized": false, "Host": "GitHub", "Access": "Free", "Cost": "", "Test_Split": false, "Tasks": [ "other" ], "Venue_Title": "", "Venue_Type": "", "Venue_Name": "", "Authors": [ "Salem Lahlou" ], "Affiliations": [ "Mohamed Bin Zayed University of Artificial Intelligence" ], "Abstract": "We present ArabicDialectHub, comprising two complementary contributions. First, we introduce a curated collection of 552 phrases across six Arabic varieties: Moroccan Darija, Lebanese, Syrian, Emirati, Saudi, and MSA. The collection was generated using large language models and validated by five native speakers, stratified by difficulty, and organized thematically. The open-source platform provides translation exploration, adaptive quizzing with algorithmic distractor generation, cloud synchronized progress tracking. Second, we present an open source interactive learning platform demonstrating the resource\u2019s practical utility through multiple learning modalities: a translation hub for phrase exploration, an adaptive quiz system with intelligent distractor generation, progress tracking with cloud synchronization, and cultural context cards highlighting regional sensitivities. Both the dataset and complete platform source code are released under MIT license. Platform: https://arabic-dialect-hub.netlify.app.", "Subsets": [], "Dialect": "mixed", "Language": "ar", "Script": "Arab", "Added_By": "qwen/qwen3.6-35b-a3b" }