Kazakhstan Tests Language-Vision AI Model in Pilot Program
The Institute of Smart Systems and Artificial Intelligence (NU ISSAI) at Nazarbayev University has introduced a groundbreaking AI model named Oylan. This multimodal model is capable of processing images and text in Kazakh, English, and Russian, providing a versatile tool for various industries.
In order to train Oylan, researchers at NU ISSAI compiled the largest AI dataset in Kazakhstan, consisting of over 10 million images and 50 million question-answer pairs. This extensive dataset is customized to meet the linguistic, cultural, and economic requirements of the country, allowing Oylan to be utilized in education, healthcare, industry, and finance sectors.
Askat Kuzdeuov, the lead data scientist, highlighted that the dataset covers a wide range of domains such as image captions, visual questions and answers, optical character recognition, document analysis, and problem-solving in fields like math, geometry, and physics. This comprehensive dataset enables Oylan to tackle various challenges effectively.
Despite facing limited computational resources, the research team at NU ISSAI employed efficient training techniques to enhance Oylan’s performance. The model is now open for public testing on the ISSAI Playground, which offers a user-friendly interface with voice interaction for a more immersive experience. Additionally, developers can access Oylan’s API to seamlessly integrate it into different applications.
With Oylan’s advanced capabilities and versatility, it is expected to make a significant impact in the field of artificial intelligence and smart systems. This innovative AI model represents a major milestone in Kazakhstan’s technological landscape and has the potential to drive advancements in multiple sectors.
