Content |
The main articles are:
2024
How neuromorphic technologies are developing in Russia and what are their prospects
Neuromorphic technologies, the architecture of which is as close as possible to the structure of biological ones, are neural networks actively developing in the world, gradually becoming one of the key areas in the field of artificial intelligence. Russia There are few such projects, but research and development is underway, and certain successes have already been achieved. TAdviser Market participants told about what domestic scientists and companies have achieved in the neuromorphic field in June 2024.
Neuromorphic technologies rely on the principles of the human brain. This allows you to create systems that can self-learn and adapt to new conditions. The term neuromorphism itself implies the creation of intelligent, autonomous and systems that do not require synchronization, explained the founder of the AI service BotB2B Victor Popov.
Russia already has the first domestic neuromorphic chip developed by the Novosibirsk company Motive NT in cooperation with Kaspersky Lab. The processor was named "Altai" (AltAI). Andrei Lavrentiev, head of the technology development department at Kaspersky Lab, told TAdviser that calculations and demonstration projects show that the power consumption of Altai chips when working with pulse neural networks is at least a thousand times less than that of classic neural networks on graphics maps.
For several years now, Rosatom has had a neuromorphic computing competence center based on Tsifrum, which deals with the applied use of neuromophysical technologies in the nuclear industry and scientific research in this area.
Scientists from the Moscow Institute of Physics and Technology (MIPT) and the K. A. Valiev Institute of Physics and Technology have carried out a lot of analytical work to find effective technologies in the context of microelectronics that have a stimulating effect on the human brain. Experts analyzed various developments for neuromorphic systems and described further prospects for the technology.
St. Petersburg State Technical University "LETI" develops neuromorphic integrated circuit approaches that are aimed at designing circuits and topologies of the main nodes and functional blocks of architectures of multibit neuromorphic systems for their implementation on a chip and subsequent conformal integration into superminiature devices of intelligent electronics. The Russian Academy of Sciences (RAS) and Moscow State University (MSU ) are also conducting research on the development of pulsed neural networks and their hardware implementation.
Experts interviewed by TAdviser agree that neuromorphic chips can significantly increase energy efficiency and speed up the work of new generation artificial intelligence systems. They will be useful in areas such as the Internet of Things, robotics, unmanned vehicles, cyberphysical security systems, facial recognition and intelligent processing of big data (for example, in data centers).
Kaspersky Lab sees prospects for using neuromorphic technologies in vision systems for video surveillance. Many city video surveillance systems function on the basis of neural network technologies: cameras recognize car numbers, people's faces, dangerous situations. Neuromorphic chips will make the operation of such systems more efficient in a number of parameters, according to the company.
The HSE believes that neuromorphic technologies can become a driver of Russia's innovative development: they can not only increase the competitiveness of domestic products in the world market, but also create new jobs in high-tech sectors of the economy. In addition, the development of these technologies can help strengthen the country's scientific and technical potential and attract investments in the field of high technologies, says Olga Martynova, a leading researcher at the Center for Neuroeconomics and Cognitive Research, researcher at the Institute of Cognitive Neurosciences at the Higher School of Economics.
According to Anna Dudnik, co-founder of НейроИнститута.РФ, the development of neuromorphic technologies in Russia is constrained by the lack of a sufficient number of specialists, as well as the need to attract investment capital. However, the prospects for neuromorphic technologies in Russia are great. They promise significant improvements in medicine, education and security, offering solutions for tasks that previously seemed impossible, Dudnik is sure.
Professor of the Department of Micro- and Nanoelectronics of St. Petersburg State Technical University "LETI" Natalya Andreeva notes that a feature of the work of neuromorphic processor modules is the lack of information transfer between RAM and long-term memory, therefore, areas of application will be determined by such priority areas of scientific and technological development as "Security of information acquisition, storage, transmission and processing" and "Preventive and personalized medicine."
According to the head of the laboratory of the Institute for Advanced Materials and Technologies of the National Research University of MIET Pyotr Lazarenko (grantee of the Russian Scientific Foundation), the subsequent development of electronic neuromorphic computing systems can be carried out by using elements that can change their physical properties depending on external influences obtained at the training stage and save changes in time. By mid-2024, work is underway to develop electronic neuromoric computing systems based on RRAM memristors, spintronics and PCRAM, which provide the possibility of a multi-level change in their state and its non-volatile maintenance, he said.
According to Denis Larionov, head of artificial intelligence at Tsifrum, neuromorphic hardware and software complexes as a class of AI systems are applicable in all industries. For example, neuromorphic systems are well suited for controlling rapidly developing processes - keeping plasma in a tokamak, controlling the movement of UAVs. Neuromorphic processors coupled with event cameras (DVS) can be used for visual vibration diagnostics of equipment, measuring vibrations without using physical sensors, he said.
Sergey Shchanikov, representative of the National Center for Physics and Mathematics (NCPM; among its founders is Rosatom), I agree that neuromorphic technologies have great prospects in Russia: you can work out the issues of their further application in the field of robotics, for example, to create self-learning robotic complexes, in technical vision systems for various purposes, to create human-machine interfaces and neurohibride systems. The field of application of neuromorphic systems is very wide and is consistent with all strategic directions of the country's development, approved by the President of the Russian Federation and the government, Shchanikov believes. He stressed that research on neuromorphic systems in Russia by 2024 is at a high level comparable to the international one.
 | This is the area in which Russian scientists have a good fundamental groundwork and deep knowledge both about the principles of functioning of living biological systems, obtained as a result of neuroscientific projects, and about micro- and nanoelectronics technologies aimed at creating new types of electronic component base based on brain-like principles of information storage and processing, - added the representative of the NCPM. |  |
The first Russian neuristor has been developed, which will become the basis of a neuromorphic computer
On June 21, 2024, it became known that Russian scientists developed the first domestic neuristor, which will become the basis for the creation of a neuromorphic computer. Read more here.
Putin instructed to create neuromorphic processors
In February 2024, Russian President Vladimir Putin approved changes to the National Strategy for the Development of Artificial Intelligence. The document, among other things, involves the creation of neuromorphic and tensor processors in Russia. Read more here.
2023
The world's first neuromorphic supercomputer has been developed
On December 13, 2023, Australian researchers at Western Sydney University reported the development of the world's first supercomputer capable of simulating pulsed neural networks at the scale of the human brain. A neuromorphic machine called DeepSouth is expected to be able to process huge amounts of data at high speed, while consuming much less power than conventional supercomputers. Read more here.
A neuromorphic processor was created at the Kurchatov Institute. It is built on artificial biomaterials
At the end of November 2023, the Kurchatov Institute national research center announced the creation of a neuromorphic processor, which is built on artificial biological materials. Read more here.
Russian company developed the first domestic neuromorphic processor
In November 2023, the Motif Neuromorphic Technologies (Motif NT) company announced the first neuromorphic processor in Russia. It was named "Altai" and is intended for vision systems, intelligent and robotic systems and data centers. The neuroprocessor will be used for the tasks of recognizing and classifying video and audio images, integrating and processing information from various sensors and sensors. Read more here.
2022: Start of sales of Akida neuromorphic processor
January 18, 2022 BrainChip announced the start of sales of its neural network processor. Akida According to the developers, this product is the first commercial among neuromorphic - AI chips that can provide the benefits of ultra-low power consumption and performance compared to traditional approaches. More. here
2021: The Ministry of Industry and Trade has taken up the study of processors that mimic the work of the brain
In July 2021, the Ministry of Industry and Trade of the Russian Federation announced a tender for a marketing study of "domestic and global industries of hardware and software complexes for the purposes of artificial intelligence." The agency is ready to pay up to 70 million rubles to the winner of the competition. Read more here.
| The site content is translated by machine translation software powered by PROMT. The machine-translated articles are not always perfect and may contain errors in vocabulary, syntax or grammar. Read original article If you find inaccuracies or errors in the results of machine translation, please write to editor@tadviser.ru. We will make every effort to correct them as soon as possible. |