MOSCOW, January 25. / TASS /. The library PuzzleLib, developed by the Skolkovo Foundation’s resident company, which makes it possible to assemble a neural network for various tasks, is included in the Unified Register of Russian programs for electronic computers and databases, which gives development priority over other software (software) for all government organizations. This was announced on Friday by the Fund’s press service.
“The development of the company” Neuronet Ashmanov “, a resident of the Information Technology Cluster of the Skolkovo Foundation, is included in the Unified Register of Russian programs for electronic computers and databases. reads the message.
The PuzzleLib library is software that allows you to build a neural network for different tasks, run its training and place it on the network for future use. Only foreign companies currently have analogs, for example, Google TensorFlow, Facebook PyTorch, and others. Among the advantages of the Russian development are high speed, automatic optimization of calculations, a large set of neural network modules, of which there are more than 60 types and a number of others.
“I consider the emergence of domestic technology to work with neural networks, also optimized for the Russian computing equipment stack, a very important milestone in the development of IT technologies in the country. This fact expands the circle of customers of this software (software) and makes it possible to create neural network technologies of special application”, – said Sergey Voinov, head of IoT, Electronics, Digital Health, IT cluster of the Skolkovo Foundation, Sergey Voinov, whose words are quoted in the message.
Stanislav Ashmanov, CEO of Neuronets Ashmanov, explained that the inclusion of PuzzleLib in the registry will give the library priority over other software for all government organizations. “The transition to the domestic library is to be made by all state organizations (first of all, organizations of the military-industrial complex),” the press service quotes him.
From a technical point of view, as representatives of the company noted, it will be easy to rebuild to a new library, since customers are provided with detailed advice and training materials on its operation.
About neural networks
Artificial computer neural networks are built on the same principles as the networks of nerve cells in a living organism. The processes of memorization and pattern recognition are based on the transmission of signals between neurons. One of the main advantages of neural networks over traditional algorithms is the opportunity to learn. In the process of learning, such a network is able to identify complex dependencies between input data and output data, as well as perform a generalization.
Therefore, areas of application of neural networks are those where human intelligence is ineffective, and traditional calculations are too laborious, for example, in telecommunications, speech recognition, medical diagnostics, and so on.