Yandex is set to introduce YandexGPT, a groundbreaking language model capable of writing, processing texts, and conducting conversations in the Kazakh language.
Scheduled for launch in 2024, YandexGPT represents the first public generative neural network specifically trained to grasp and articulate in Kazakh, marking a significant milestone in language AI development for the region.
YandexGPT will be integrated into a variety of Yandex services, including a virtual assistant within the Yandex application, Search, and the Yandex Browser, exclusively for users in Kazakhstan.
This initiative underscores Yandex’s commitment to enhancing AI accessibility and functionality for Kazakh-speaking users, fostering greater inclusivity within the digital space.
The development of YandexGPT leveraged the foundational YandexGPT model, enriched with hundreds of thousands of Kazakh language queries and responses.
This extensive dataset encompasses a wide range of topics and conversational contexts, laying the groundwork for a highly capable and contextually aware AI model.
AI trainers from Yandex Kazakhstan play a crucial role in refining the neural network’s capabilities. They meticulously assess the AI’s output for accuracy, completeness, factual correctness, and adherence to grammatical and stylistic norms of the Kazakh language.
Through this iterative training process, involving the generation of standard queries and responses, the neural network’s response accuracy and contextual understanding are continually enhanced.
Looking ahead, Yandex plans to introduce a commercial version of YandexGPT in Kazakh for businesses, to be hosted on the Yandex Cloud platform.
This move is poised to open new avenues for businesses to leverage advanced AI tools for a variety of applications, from customer service to content creation, all in the Kazakh language.
This initiative by Yandex not only represents a leap forward in AI language processing capabilities but also highlights the company’s dedication to fostering technological advancements that cater to the linguistic diversity of its user base.
Source: TASS
