The GigaChat neural network model successfully passed the entrance exams to the Gnessin Russian Academy of Music for the "Musicology" educational program. This was reported by the press service of the RAM.
The GigaChat neural network model from Sber passed the entrance exams to one of the leading musical universities in the country — the Gnessin Russian Academy of Music. The experiment was part of a partnership project between Sber, the Gnessin RAM, and the Zvuk music service.
The exam for the "Musicology" program consisted of open and closed questions regarding the history of Russian, foreign, and pop-jazz music. In addition, the GigaChat neural network model successfully coped with complex tasks that are usually offered to applicants at the colloquium in the discipline "History of Music."
The neural network demonstrated the ability to understand the music of various composers.
GigaChat successfully completed the task and received a high score.
The answers were evaluated by members of the RAM examination committee: teachers of the historical-theoretical-compositional faculty, the vice-rector, and heads of educational departments.
GigaChat was additionally trained on scientific and educational materials, archival records, theoretical works, and reference books. Consultations and support during training were provided by teachers, researchers, and students of the academy. The project confirms the possibility of using AI in the field of art.
According to Alexander Ryzhinsky, Rector of the Academy, in the context of rapid technological progress, the use of artificial intelligence in creativity is becoming inevitable. According to him, the main task is to use the strengths of modern technologies to combine them with musical science and performing arts.
Earlier, Sber's artificial intelligence transformed the annual "Night at the Museum" event in Moscow. The GigaChat and Kandinsky neural networks created a new mascot — a wise Owlet. They also developed an original design for electronic tickets.
Read more on the topic:
GigaChat Learned to Understand Images
Total Control: How Smart Devices Spy on Users
Three Main Mistakes That Prevent You from Maximizing the Potential of Neural Networks
