From bd0a56724307604083ae3ffd76d352514f7bfe92 Mon Sep 17 00:00:00 2001
From: masader-bot <zaidalyafeai@gmail.com>
Date: Mon, 8 Jun 2026 09:13:39 +0000
Subject: [PATCH] Creating datasets/arabicdialecthub.json

---
 datasets/arabicdialecthub.json | 48 ++++++++++++++++++++++++++++++++++
 1 file changed, 48 insertions(+)
 create mode 100644 datasets/arabicdialecthub.json

diff --git a/datasets/arabicdialecthub.json b/datasets/arabicdialecthub.json
new file mode 100644
index 00000000..c980e216
--- /dev/null
+++ b/datasets/arabicdialecthub.json
@@ -0,0 +1,48 @@
+{
+    "Name": "ArabicDialectHub",
+    "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": "ArabicDialectHub: cross-dialectal 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://arxiv.org/pdf/2601.22987v1.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, a cross-dialectal learning resource comprising a curated collection of 552 phrases across six Arabic varieties: Moroccan Darija, Lebanese, Syrian, Emirati, Saudi, and MSA. Generated using LLMs and validated by five native speakers, the collection is stratified by difficulty and organized thematically. The open-source platform provides translation exploration, adaptive quizzing with algorithmic distractor generation, cloud-synchronized progress tracking, and cultural context cards. Our contributions address gaps in pedagogical structuring and dialectal learning focus, providing a fully functional interactive system released under MIT license.",
+    "Subsets": [],
+    "Dialect": "mixed",
+    "Language": "ar",
+    "Script": "Arab",
+    "Added_By": "qwen/qwen3.6-35b-a3b"
+}
\ No newline at end of file