RSS
Логотип
Баннер в шапке 1
Баннер в шапке 2

Digital Polygons for AI Development

Product
Developers: Moscow Institute of Physics and Technology (MIPT)
Date of the premiere of the system: July 2023
Branches: Information Technology

2023: Creating Polygons

MIPT specialists have created digital polygons to accelerate the development of AI systems and reduce the cost of testing them. The press service of the university told about this in early July 2023.

They noted that MIPT scientists are one of the first in the Russian Federation to create simulators and doubles of physical and technical objects within the framework of multifysical modeling, while additionally including control and artificial intelligence algorithms. The virtual polygon simulates the external test environment of the prototypes of the target system, creating all possible abnormal and force majeure situations. When testing systems with artificial intelligence, this becomes especially important, said Roman Gorbachev, head of the laboratory of Wave Processes and Control Systems at MIPT.

Virtual Polygon Simulates External Test Environment for Target System Prototypes
File:Aquote1.png
If the fourth technological revolution allows you to transfer models from the virtual world to the physical one, then our technology complements the range of its capabilities. It allows you to transfer algorithms and AI capabilities from a virtual model object to a dual physical object, "explained Vyacheslav Kondratyev, director of the MIPT Research Center for Applied Artificial Intelligence Systems.
File:Aquote2.png

According to him, at the development stage, MIPT specialists first embed a machine learning system not into a physical object, but into its digital metamodel, and then the already trained model is embedded into a physical object.

This approach, as noted by the developers, allows you to comprehensively check the operation of machines and devices even before they are manufactured. This allows you to save time, resources and money for testing, as well as quickly identify possible causes of problems in the operation of systems. In the near future, the developers plan to expand the scale of the digital polygon, which will make it even more effective, MIPT reported in early July 2023.[1]

Notes