Starting November 2023, the large international event RUSSIA EXPO is held in Moscow at the VDNH premises and presents the main scientific and technological accomplishments of the country. During his visit to the event, Andrey Ronzhin, Director of the St. Petersburg Federal Research Center of the Russian Academy of Sciences, spoke about current trends in cyber and digital security, robotics, artificial intelligence, as well as about key scientific projects of SPC RAS in the above areas.
All in all your institute does research and developments in computer science and intelligent systems. And what are the best relevant areas in recent years?
– On the one hand, and you are right – our Institute marks the computer science and deals with computer security, mathematical modeling, remote sensing of the Earth's surface, robotics: ground, air, water, and so on.
However, on the other hand, four years ago, the Institute underwent a large-scale reorganization, resulted in establishing the St. Petersburg Federal Research Center of the Russian Academy of Sciences based on our institute and five other large successful organizations.
After that, the emphasis of our research got changed. We have become stronger involved in interdisciplinary projects, particularly, in the areas of ecology and agriculture. In the course of this work, we receive not only basic but also applied scientific results aimed, for example, at improving the quality of agricultural products and improving environmental management.
Laboratory of autonomous robotic systems well fits the Center’s composition, and currently its developments are presented at the RUSSIA EXPO. What robots are being created by this team and what applications they are intended for?
– No doubts that I know this laboratory pretty well, actually for years I used to be its head. Initially, we were only engaged in ground robotics. Particularly the work dealt with a rather complex technical area: anthropomorphic robots (specifics is that they walk on two legs, like a human). And it was a truly exciting research area, involving both students and mature researchers; eventually this combination leads to good enough results?
Am I right, that we are speaking about working out the software “stuffing” as well as the hardware of these robots?
– Yes, you are absolutely right. Unfortunately, in the past, while being the institute of computer science, we concentrated mainly on software development, on control systems, and had no chance to create our own robots in hardware. Right now we develop both software and hardware at our own Engineering Center equipped with the necessary instrumentation and information and measurement systems; develop various electromechanical constructions and design full-fledged robotic systems.
You have mentioned anthropomorphic robots. Are such robots still being developed in the world and for what areas of life?
- The peculiarity of anthropomorphic robots is that they are quite energy-intensive. Say, that energy is not required in case of static standing on a wheeled platform, while for the case the anthropomorphic robot needs to spend energy in order to maintain a stable standing state. This is one of important reasons why this direction goes on with difficulties.
Nevertheless, anthropomorphic robots have many advantages: they are more passable and more human-like. The latter is important because, for example, if the robot approached you, the communication with it would be intuitively easier than with a wheeled platform. Contact between man and machine and man is established faster, and from this point of view, the development of anthropomorphic robots is a promising topic.
And we continue to foster it: now we are concentrated on sensors, that is, how to make robot gently grasp some object without crushing it, how to shake hands and so on. Once we get through the sensory stage, we can possibly return to anthropomorphic robots
According to your estimates, how many years will pass to introduce anthropomorphic robots in massive use? What will it depend on?
–– Hard to say. Good thing that over the past two years the attitude of society to science has changed significantly. Media are more and more active in covering science. We are not simply motivated, we feel the influx of young people in science. These are the burning eyes of schoolchildren, students who are ready to work. Moreover, many of them join us without any
And I believe that progress in science, and in anthropomorphic robotics particularly, will soon be very serious.
Should we speak generally about artificial intelligence, what spheres in Russia implement it most actively now and what might be some non-obvious areas of its use?
– There exist quite a number of such spheres. For instance, our Center is engaged in the artificial intelligence introduction to agriculture areas like: crop production, animal husbandry, fish farming.
While considering new directions, lately, in cooperation with institutions dealing with law, we have started such an interesting lead as Ligaltech, that is, automation of the legal process. And here we face a lot of pitfalls regarding the artificial intelligence use, yet we possess the competencies that can be applied in this area.
Our specialists have developed a solid foundation for automatic text processing so that the machine can process billions of documents. This is important because we address a lot of public regulatory legal acts, that are unfortunately often contradictory because of their enormous volume. Even lawyers often find these legal holes that are very difficult to close and recycle. This is where AI technology can help. The goal is to correct mistakes, and most importantly, to help society and people with their specific legal problem
Moreover, artificial intelligence can fight against malefactors: now there is a lot of fraudulent sites that allegedly offer legal services for money. After interacting such sites, people can be stripped off their money, and find themselves helpless in a situation. And by creating such legal decision support systems that recycle the entire existing legal framework, the user will be able to find a competent solution.
The current urgent question for many people is: to what extent AI is safe and secure, are there any mechanisms to control it?
– Indeed, a truly classic question. Here we do understand that in case we neglect dealing with artificial intelligence issues, someone else will do it. Perhaps it will be, relatively speaking, our enemy, attacker or competitor, and this opposer will do it better for the sake of own interests and against us. This is applicable to any society over any territory.
If we neglected studying AI technologies ourselves, sooner or later they will be used against us. I always say even to the students and schoolchildren: don't learn to be users, learn to be developers. Then you will know the system from the inside and it will be easier for you, knowing its laws, to resist it and create your own, more reliable, cheaper and more effective.
Time to speak about another interesting development of your Center: a wireless power transmission system. How promising it is?
– Actually already in April, the main developer of our technology will defend his PhD thesis – that is Dima Krestovnikov a smart guy, young promising. I hope that eventually he will get developed into a great scientist.
He independently assembled an initiative group and received additional financial support for the development. They started testing their system for transmitting energy through the air. And now they are successfully doing it underwater. The system has a great future, the guys are creating universal receiving and transmitting devices able to transfer electricity both in one direction and in the other. These devices are demanded and found applications in underwater and agricultural robotics.
So I hope that Dima's thesis will be successfully defended, and he will further proceed, establish his own laboratory, and receive new major projects. And in this research area of robotics and energy, we will raise the banner of our Center.
We wish Dmitry good luck. Now our talk will dwell upon cybersecurity. What attention is paid to its progress at SPC RAS? And what systems are considered to be under the protection?
– We have several units that deal with this topic. Firstly, cybersecurity should be distinguished from information security.
Information security is the protection of the information itself both from a malefactor and protection from unwanted and malicious information.
In this area, our scientists do research on social networks and groups, where analyzed the users, their characteristics, everything they post on the Internet, who and when they communicate with, and so on. This allows for identifying and categorizing the negative attempts of malefactors to influence other people. In addition, such systems are used to select candidates at employment. This also refers the video analytics direction.
So far, we have developed a comprehensive system that analyzes a person's profile on social networks and their video interviews. And according to these two criteria, the system tells whether a candidate can be a manager, whether he can be a manager or a salesperson. So, the program helps, on the one hand, to find a suitable job for the applicant, and on the other hand, the employer assesses whether the candidate skills match the requirements for filling the announced vacancy.
Если в первом случае мы работали с информацией, как правило, в виде текста и изображений, то кибербезопасность – это работа с ноликами и единичками, то есть с бинарным представлением информации. Здесь мы анализируем как на этапе проектирования, так уже и готовой к работе системы, насколько она уязвима от внешних атак, например, вирусов. И предлагаем решения по разработке контрмер по противодействию этим атакам и устранению этих уязвимостей.
If in the first case we worked with information, as a rule, in a form of text and images, then the cybersecurity assumes a work with zeroes and ones, that is, with a binary representation of information. Here we analyze both at the stages of design and the ready-to-use system, how vulnerable it is to external attacks, for example, viruses. And we offer solutions to develop counter-measures to oppose these attacks and eliminate these vulnerabilities.
In particular, we have another division dealing with this, the International Center for Digital Forensics, which works closely on legal issues. They cooperate with the St. Petersburg Academy of the Investigative Committee, with law universities, and deliver lectures. The lecture on digital hygiene is very popular.
This is indeed a very important topic. For example, even my colleagues used to write me almost every day and ask questions about as if my information concerning communications with a minister or with the FSB supervisors, who demanded to immediately carry out something allegedly previously entrusted to me, the director. And pretty often we faced the case at the time of this tendency emergence when the Center professionals arranged for big meetings, accomplished as if my tasks, and in result furnished me with some documents. And in response to my questions: “What for?” they used to answere: “You told me yourself, referring to the minister, that it was urgent.” It happens to us too.
Yes, now it is really quite an active type of fraud. To go further with the topic, if we speak about the development of systems for citizens, what developments, ideas, novelties exist in this direction?
– Here, as an example, you can cite systems that allow for analyzing a user's text message to social structures, annotate it and offer one of the solutions, like, how to correctly write an application for a subsidy due at the birth of a child.
I understand that after the child delivery, any young mother is in a happy state, and needs to get familiar with a lot of documents, to proper understand and write everything correctly, that indeed can be very important for a young mother. And such systems already exist, however, they should be developed and updated so that to provide information more accurately, more personified. It is important that they be able to give template solutions, as well as to consider each case separately, because in legal matters they all have their own peculiarities.
Another case is that right now we have our own very good logo recognition system in the direction of Ligaltech, that we plan to use in Rospatent and other structures.
Today, experts work in this area with minimal involvement of automated systems. This means that such a specialist must review millions of existing logos, compare them and understand whether there are any violations.
We propose our automated system that implements several types of classifiers: by text, by color, by semantic characteristics, by perception, and so on. It allows for selecting certain subsets of existing logos, close to the new sample according to these criteria, what significantly improves the quality and speed of processing.
The interview is available at the link: “Russia”