Doha: In the rapidly evolving landscape of artificial intelligence, researchers at Hamad Bin Khalifa University’s Qatar Computing Research Institute (QCRI) are using advanced AI techniques to promote Arabic language.
The research projects address the challenges related to lack of content and equally important, extracting that content, analysing and transforming it. The researchers are working on advancing large language models (LLMs) to enhance AI’s understanding of the Arabic language.
AI models like chatGPT have not yet been able to master Arabic and its dialects very well. Reports suggest that a key challenge in creating AI models focused on Arabic is the limited availability of high-quality Arabic data online, which makes up just about 0.5 percent of the overall content.
To overcome this, one of the most important projects that researchers are working on at QCRI is the Arabic large language model.
Among this, the Fanar, a series of Arabic centric LLMs, is a great example. This relies on highly curated native Arabic data. “The most important factor that affects large language models is data. At Fanar we strive to ensure that Arabic remains authentic. Fanar supports modern standard Arabic as well as popular dialects like Egyptian, Levantine, and Gulf dialects,” said Dr. Mohamed Eltabakh, Principal Scientist, QCRI, in a video message published by Qatar Foundation.
