Russian neural network has passed accreditation for a doctor

Sbera's neural network model GigaChat MAX has successfully passed the test and solved situational tasks corresponding to the theoretical stages of the primary accreditation in the specialty of "Medicine", the company's representatives reported.

The exam was organized on the basis of the accreditation center of the I.M. Sechenov First Moscow State Medical University of the Russian Ministry of Health and fully met the requirements for doctors.

"Every medical student must undergo such accreditation to qualify to work in the Russian healthcare system. Our neural network also passed the testing corresponding to the primary accreditation for the first time in Russia. This is a very important step towards the practical use of AI in healthcare. We will develop GigaChat MAX in certain specialties, including therapy, cardiology, neurology, gastroenterology, and others," said Sergey Zhdanov, director of Sberbank's Health Industry Center.

"I consider this a landmark event on the path of AI development in our country. By assisting the doctor in decision-making, GigaChat MAX will save his time and reduce the routine workload. An agreement has already been reached with Sber on further AI training in the specialties of clinical pharmacology, urology and pulmonology", - said the first vice-rector of Sechenov University Andrey Svistunov.

Sber recalled that in February 2024, the GigaChat neural network model was one of the first in the world to pass the final exam for sixth-year students required to obtain the qualification of "medical doctor". In the second, accreditation exam, the neural network, like medical professionals, is tested for more in-depth knowledge. For this purpose, GigaChat MAX has been additionally trained by hundreds of physicians on a considerable amount of medical training materials. Successfully passing the accreditation confirms that the model now has deeper medical knowledge, Sber emphasized.