An agreement signed at the KazanForum between the Academy of Sciences of the Republic of Tatarstan, Innopolis University, and the AIRI Institute provides for the training of artificial intelligence to work with the languages of small ethnic groups of Russia, primarily Tatar. This was reported by the forum's press service, citing Ayrat Khasyanov, Vice-President of the Academy of Sciences of the Republic of Tatarstan. The AI will operate in the field of Natural Language Processing (NLP), a field that studies the problems of computer analysis (understanding) and synthesis (generation) of texts in natural languages.

"Today, we signed an agreement with the AIRI Institute, which is the leading structure in Russia engaged in artificial intelligence. NLP developments can be used for all Turkic languages. We plan to move towards translating the results for application in solving problems of other languages of Russia. The immediate next steps we plan to take are to hold a joint conference at Kazan Digital Week dedicated to AI and to conduct a hackathon on natural language processing technologies" Khasyanov noted.

Under the agreement, specialists from the AIRI Institute will be involved in the development of language models, including those in the Tatar language. Bauman Moscow State Technical University (MSTU) is the hardware partner. The university is working on the "Teragraf" supercomputers, on which the translation of the large digital linguistic heritage of Tatarstan is taking place. The head of the "Teragraf" project at Bauman MSTU announced the creation of a technological platform to implement promising language models. This implies the use of knowledge graphs, which signifies a certain departure from an exclusively neural network approach to the natural representation of knowledge in the form of related facts. This task is considered a new computational idea requiring the redesign of hardware.

ORIENT