A Look at AI in Russia

A Look at AI in Russia

Feature articles |
The technology of the modern digital economy is largely bound up with the opportunities offered by artificial intelligence (AI). This report gives an overview of how Russian AI research is progressing.
By eeNews Europe


These opportunities have a downside, though, which is being pondered by people like Elon Musk. This fall, he pointed to a statement about AI made by President Vladimir Putin during the All-Russia Open Lesson: “Whoever becomes the leader in this area will own the world.” Musk’s reaction on Twitter was no less unequivocal: “China, Russia, soon all countries w strong computer science. Competition for AI superiority at national level most likely cause of WW3.”

Such an impression made the businessman turn to the UN in order to push for a global ban on AI-based weapons. This overview is an attempt to provide an answer to Musk (and those who share his point of view), with an emphasis on the Russian side of efforts to improve AI technology.


How to accelerate AI: Technology and relevant projects

Experts usually understand AI to mean the branch of knowledge and technology that allows computers, based on certain tools and accumulated knowledge, to answer questions and draw their own conclusions. The computer in this case does not simply accumulate data, but generates new knowledge that was not previously entered into it by humans. Precisely this AI capability gave rise to such currently popular fields as neural networks, machine learning and pattern recognition.

According to SAP analysts, almost 1,500 AI research projects in Russia have received financial support from the state and private sector over the past 10 years, with more than half of the projects paid for by the state or implemented as part of federal targeted programs. For example, the Global Competitiveness Improvement Project for Leading Russian Universities, 5-100, has brought together the strongest Russian universities so that they could work in advanced research fields with the support of the Ministry of Education and Science. The creation of artificial intelligence is one such field.

In 2017, several major international conferences were held in Moscow, such as BICA 2017 and Neuroinformatics-2017. They brought together scientists from over 30 countries to discuss ways AI can help humankind in a variety of areas, such as medicine: for example, robots that react to the slightest signals from users can help people with disabilities adapt to the environment, or predicting a variety of consequences where computer simulators allow users to simulate certain – including emergency – situations in the most unpredictable combinations, providing the appropriate response and improving during the process of interaction.

A team led by Professor Alexei Samsonovich is developing AI technology called Virtual Actor, which is able to understand the context of what is happening and which also has emotional intelligence. Such virtual actors can effectively adapt to human psychology, behavior and emotions. A unique social experiment on play interactions between people and computer bots has already been conducted, showing that it has become very difficult to distinguish between the bots and the humans.

Another university, Moscow Institute of Physics and Technology (MIPT), which is a participant in the 5-100 program, is contributing to the development of artificial intelligence with its iPavlov project, which is being carried out as part of the Neuronet roadmap of the National Technological Initiative. It is based on the “conversational” technology of self-learning AI and is an intermediate step in creating a platform on which AI can effectively address bank client requests in order to automate this work to a significant degree.

MIPT specialists pay great attention to AI medical applications. They have presented a neural network architecture to generate new molecular imprints with required properties, into which a neuron indicating a decrease in the number of malignant cells after treatment is introduced. Such a neural network was used to screen 72 million compounds and to select candidate molecules that have potential antitumor properties.

A joint project of the IT, Mechanics and Optics University (ITMO – St. Petersburg) and the Far Eastern Federal University (Vladivostok) focuses on the future of AI. Its goal is to build, in conjunction with the world’s leading educational centers, a trans-Eurasian network with the Master’s program, Big Data and Artificial Intelligence for Digital Liberal Arts. This program will make it possible to train the necessary number of professionals in the relevant fields according to the latest standards.

Studies are underway regarding the use of neural networks for managing nuclear power plant systems. Scientists believe that this will improve performance and minimize the number of automatic power controller errors due to the use of multi-purpose computer simulation of a water-water power reactor (VVER).

How to stop AI: Regulation and the singularity barrier

In addition to the technological component of AI research, it is also necessary to develop the legal side. In the first half of 2018, the relevant committee of the State Duma will consider a draft convention on robotics and artificial intelligence. How will the Russian authorities determine the degree of accountability borne by AI machines? This issue concerns not only lawyers, but researchers as well.

One cannot help noticing that artificial intelligence is rapidly making its way into everyday life, addressing an increasing number of tasks in industry, economy and the social sphere,” Anton Pyrkin, dean of the Department of Control Systems and Robotics at ITMO University, said. “So, the issues of legal relations between humans and robots or computer software that imitates people cannot be left unattended – especially when money will become the subject of discussions or disputes. In my opinion, the responsibility for the AI programs’ actions should be borne by developers or copyright holders, which put a particular tool capable of causing damage into circulation. An attempt to impart an AI program with personal legal liability is, I believe, an attempt to shirk responsibility by its developers and possible beneficiaries.”

Experts point out that humankind is faced with a number of technical and ethical-moral issues as it develops AI. After all, the cognitive functions of AI are not fundamentally limited, as compared with humans, and are sure to significantly exceed human capabilities in the future. In this case, robots should be made full-fledged subjects of law.

The calculation of the point of technological singularity – a big leap in scientific and technological development – is one of the most common criteria for assessing the time it will take to create a full-fledged AI similar to human intelligence (known as a “strong version” of AI). This is estimated to occur between 2030 and 2050. Nevertheless, no one can say that AI in its “strong version” will definitely be created by this time.

“We are now at a lab stage, but already by the 2020s we will be able to take the next step involving a fully functional market, and a multitude of actual start-up companies,” Valentin Klimov, deputy director of the Institute of Intelligent Cybernetic Systems of the National Research Nuclear University MEPhI, said.

However, some players on the AI market believe that the technological gap between the West and Russia is too vast, and the only option Russia has is building into existing Western platforms. This policy of “least resistance” suits far from everyone.  “Sooner or later, things have to start from scratch, and now is precisely this time: software tools have already reached a certain point of stagnation, where they cannot include a full-fledged emotional component into AI,” Professor Alexei Samsonovich, research director at the Institute of Intelligent Cyber Systems, said. “All of that, I believe, should be discarded. AI should be built on the basis of cognitive research, rather than purely mathematical logic. So, Russia has every chance to succeed: we must strive to create an ‘understanding’ intelligent agent, which will be fully compatible with humans socially. Only after that it can have applied functions added to it, even if the latter are not obvious at first.”

Thus, Russian experts in the sphere of AI technology are trying to demonstrate their serious but cautious attitude to the field which elicited such an extreme response from Elon Musk. In this sense, they are benefitting from both current conditions for implementing their projects, and the overall amount of time before the singularity is reached.


About the authors:

Varvara Novikova and Alexander Titkov are in charge of promoting Project 5-100 – https://5top100.com, a Russian initiative set to maximize the competitive position of a group of leading Russian universities in the global research and education market.

Linked Articles
eeNews Embedded