Kazakhstan Unveils Multilingual AI Model Oylan
The Institute of Smart Systems and Artificial Intelligence (NU ISSAI) at Nazarbayev University has launched Oylan, a cutting-edge multimodal AI model designed to process images and text in Kazakh, English, and Russian, according to a press release from the university.
Oylan was trained using the largest AI dataset in Kazakhstan, which includes over 10 million images and 50 million question-answer pairs. This dataset was specifically tailored to meet Kazakhstan’s linguistic, cultural, and economic needs, making the AI model applicable in sectors like education, healthcare, industry, and finance.
Askat Kuzdeuov, the lead data scientist, highlighted that the dataset covers various domains, including image captions, visual question answering, optical character recognition, document analysis, and problem solving in subjects like math, geometry, and physics.
Despite limited computational resources, the team optimized Oylan's performance using advanced training techniques. The model is now available for public testing on the ISSAI Playground, offering voice interaction for ease of use. Developers can also access Oylan's API to incorporate it into different applications.
Photo credit: oylan.nu.edu.kz
