UBC-NLP/PEARL

💎 PEARL Benchmark

Dataset ID: UBC-NLP/PEARL

Description: The PEARL Benchmark is a meticulously curated subset of 6,867 high-quality Question/Answer pairs derived from the larger PEARL dataset. It is specifically designed for evaluating Vision-Language Models' (VLMs) understanding of Arab cultural content. The benchmark covers ten significant cultural domains (e.g., architecture, clothing, cuisine) and thirteen diverse question types, tailored for robust model assessment of perception, knowledge recall, and reasoning about Arab culture.

Key Features:

• Size (Benchmark): 6,867 Question/Answer pairs (5,310 closed-form, 1,557 open-ended)
• Cultural Domains: 10 (e.g., architecture, food, clothing, festivals)
• Question Types: 13 (e.g., causal reasoning, comparative analysis, hypothesis formation)
• Languages: Arabic (questions and answers)
• Modality: Image-Text pairs

Paper: Pearl: A Multimodal Culturally-Aware Arabic Instruction Dataset
ArXiv Link: http://arxiv.org/abs/2505.21979
GITHUB: https://github.com/UBC-NLP/pearl\

Citation:

@article{Alwajih2025pearl,
  title={Pearl: A Multimodal Culturally-Aware {A}rabic Instruction Dataset},
  author={Fakhraddin Alwajih and Samar M. Magdy and Abdellah El Mekki and Omer Nacar and Youssef Nafea and Safaa Taher Abdelfadil and Abdulfattah Mohammed Yahya and Hamzah Luqman and Nada Almarwani and Samah Aloufi and Baraah Qawasmeh and Houdaifa Atou and Serry Sibaee and Hamzah A. Alsayadi and Walid Al-Dhabyani and Maged S. Al-shaibani and Aya El aatar and Nour Qandos and Rahaf Alhamouri and Samar Ahmad and Razan Khassib and Lina Hamad and Mohammed Anwar AL-Ghrawi and Fatimah Alshamari and Cheikh Malainine and Doaa Qawasmeh and Aminetou Yacoub and Tfeil moilid and Ruwa AbuHweidi and Ahmed Aboeitta and Vatimetou Mohamed Lemin and Reem Abdel-Salam and Ahlam Bashiti and Adel Ammar and Aisha Alansari and Ahmed Ashraf and Nora Alturayeif and Sara Shatnawi and Alcides Alcoba Inciarte and AbdelRahim A. Elmadany and Mohamedou cheikh tourad and Ismail Berrada and Mustafa Jarrar and Shady Shehata and Muhammad Abdul-Mageed},
  journal={arXiv preprint arXiv:2505.21979},
  year={2025}
}