This approach can be implemented in both emulators and real quantum processors, said Alexei Fedorov, a 28-year-old graduate of the Bauman Moscow State Technical University, who is considered a physics prodigy.

As reported this Friday by Fedorov, head of the scientific group of the Quantum Center in Moscow, for the first time it was possible to implement the proposed approach to solve the problem of classifying four classes of images using eight quantum bits to encode data and four auxiliary quantum bits.

The creation of the so-called quantum artificial intelligence is one of the main tasks of all the main participants in the world “quantum race”.

Under this term, scientists understand the widespread use of quantum technologies and effects to speed up the work of neural network algorithms and physical devices that emulate some of the properties of biological neural networks.

The group of researchers led by Fedorov, who is also a professor at the Moscow Institute of Physics and Technology, developed a hybrid approach for the first time in the world, allowing the use of quantum computers to speed up the work of individual layers of neural networks capable of recognizing different types of images with a high level of precision.

Neural networks are one of the most popular approaches in developing artificial intelligence systems, in which the initial data is processed using several overlapping layers of artificial similarities of neurons with different properties.

Their applications makes it possible to gradually bring out the most important key features of the analyzed images or other forms of information and use them to classify objects.

jrr/llp/jcm/gfa