Firoj Alam

🌟 A Journey from Curiosity to Community Impact: Evaluating LLMs with Everyday and Native Knowledge 🌟 Back at the end of 2023, when we started to work on building Arabic LLM, #fanar, one question kept coming up: "How do we meaningfully evaluate LLMs on the kinds of knowledge…

Presenting at, #ACL2025 #ACL2025NLP , if you’re around and want to learn more about our work, please stop by…

🚨 It’s happening! The Welcome Address is LIVE at #ACL2025NLP 🇦🇹 🎙️ Local welcome + key program stats 📈 Selection insights from the chairs 🎯 Preview of what’s ahead Follow along as we officially kick off the conference! #NLProc #ACL2025NLP

📣 Presenting at ACL 2025, Vienna! 🗓️ Monday, July 28 | 🕕 18:00–19:30 | 📍 Hall 4/5 If you're attending ACL 2025, come visit our poster session and learn more about our latest work on culturally-grounded and low-resource LLMs: 🔹 NativQA: Multilingual Culturally-Aligned…

The Call for Papers for the second edition of the Bangla Language Processing (BLP) Workshop is out (blp-workshop.github.io/cfp). This year's BLP workshop will be co-located with the IJCNLP-AACL conference from December 20-24, 2025, in Mumbai, India. We invite researchers to submit…

🎉 We’re excited to announce IWSDS 2026 will take place Feb 26 – Mar 1, 2026 in Trento🇮🇹! 🏔️✨ 📢 Call for Papers is now open! 🔗 sites.google.com/unitn.it/iwsds… #IWSDS26 #ConversationalAI @UniTrento @UniTrento_DISI

🌍✨ MAHED 2025 Shared Task: Pioneering Hope & Hate Detection in Arabic NLP Amid global conflicts, innovation offers hope. Join us for MAHED 2025 , the first shared task dedicated to Hope and Hate Detection in Arabic content, co-located with EMNLP 2025 at ArabicNLP 2025. This…

we planned to launch our open-weight model next week. we are delaying it; we need time to run additional safety tests and review high-risk areas. we are not yet sure how long it will take us. while we trust the community will build great things with this model, once weights are…

🌍 Are you working on culturally and regionally aligned Question-Answering (QA) datasets in native languages? Take a look at the NativQA Framework — an open-source solution designed to make your workflow seamless. ✅ Easy to set up and run ✅ Supports native and low-resource…

20 more days to close the registration. However you should start early if you want to participate and submit the system description paper

If you're participating in ArabicNLP 2025 shared tasks or experimenting with large language models (LLMs), we’d like to share a valuable resource that can streamline your workflow and enhance your results: 📷 LLMeBench + Fanar API A free and easy-to-use integration that…

MAHED 2025 Multimodal Detection of Hope and Hate Emotions in Arabic Content. More: marsadlab.github.io/mahed2025/ Data/Code: github.com/marsadlab/MAHE… Registration: Task 1: codabench.org/competitions/9… Task 2: codabench.org/competitions/9… Task 3: codabench.org/competitions/9… Consider participating.…

🌐 Introducing NativQA Framework: A powerful tool for building culturally & regionally aligned QA datasets in any language! NativQA empowers researchers to create native, local, and everyday knowledge resources—addressing fairness & bias in LLMs. Already used to generate 300K+…

🎯 Participating in a shared task? Benchmarking LLMs or testing prompts? Try LLMeBench + Fanar API ✅ Free ✅ Easy to integrate ✅ Great for multilingual benchmarks! 🔗 llmebench.qcri.org 🔗 api.fanar.qa/docs #LLMs #AI #SharedTask #NLP #PromptEngineering

🚨 Calling all #AI & #NLProc enthusiasts! 🚨 Join the MAHED 2025 Shared Task at #ArabicNLP2025 (co-located with #EMNLP2025, Suzhou, China, Nov 5-9)! Test your models on real Arabic data in 3 exciting tracks: 1️⃣ Hate & Hope Speech (text) 2️⃣ Emotion, Offensive & Hate Detection…

#NLProc Welcome to MAHED 2025 shared task at ArabicNLP 2025! MAHED 2025 - Multimodal Detection of Hope and Hate Emotions in Arabic Content as a shared task for ArabicNLP 2025. This task aims to address the critical issue of detecting hope speech, hate speech and emotional…

If you are attending ICWSM, you can check this paper: ThatiAR: Subjectivity Detection in Arabic News Sentences arxiv.org/pdf/2406.05559

As AI continues to evolve rapidly, Fanar presents a pioneering model that combines technological advancement with a firm commitment to cultural identity and Islamic values. #Fanar #ArtificialIntelligence #ArabicAI #MCITQatar

Attending Interspeech 2025? Make sure to check out some great work from ALT/QCRI and our amazing collaborators! We’re diving into everything from speech and text foundation models to culturally aware spoken queries and Arabic TTS. 1.From Words to Waves: Analyzing Concept…