The GigaChat neural network model is developing very quickly and specializing in new areas of knowledge. This was stated by Sergei Zhdanov, director of the Sberbank Health Industry Center.
It was noted that GigaChat passed the exam of a higher medical institution in the field of “General Medicine” training, which is necessary to gain the qualification of “medical doctor”. Thus, the model became one of the first in the world to achieve this.
Sberbank announced that the bank’s artificial intelligence successfully passed the test at the level of a medical school graduate and received a final grade of 4 by answering the questions on the ticket. The exam was conducted by a commission consisting of professors of therapy, surgery and obstetrics. and gynecology from the eponymous National Center for Medical Research Institute of Medical Education. VA Almazova.
The standard ticket for the oral exam includes three situational tasks in therapy, surgery, obstetrics and gynecology and 3-5 questions for them. GigaChat also passed the 100-question test with a score of 82%, with a 70% passing threshold.
“I would like to acknowledge this role and thank the employees of the Almazov Center, who provided control over the training and validation of the model. Zhdanov noted that the first stage – passing the doctor’s examination – has been completed.
According to him, in the future, a neural network can become the basis for creating assistants to a doctor and a patient, and at the same time provide new conditions for care and information for each person about their health.
“The use of large language models and their successors will be one of the key technologies in developing human-centered healthcare,” he added.
According to Almazov, the center’s general director, and Evgeniy Shlyakhto, president of the Russian Society of Cardiology, the project to train the large language model GigaChat in medical knowledge has become a major challenge for the center, which includes hundreds of teachers and researchers.
“Citizens and students actively participated in the work. “We are satisfied with the current results and the training of the model will continue,” he said.
Shlyakhto emphasized that together with Sber, a number of GigaChat-based application solutions for medical institutions, patients and doctors are already planned. Their development will begin this year.