Dmitry Shadrin, Head of Department at Skolkovo Institute of Science and Technology: Russia’s AI developments to be beneficial for BRICS Countries

Dmitry Shadrin graduated from the Moscow Institute of Physics and Technology (MIPT). He earned a degree from Skolkovo Institute of Science and Technology (Skoltech). Since 2023, Shadrin has been heading a group at Skoltech working on processing Earth remote sensing data while also teaching at Irkutsk National Research Technical University. In 2024, he became a laureate in the “Artificial Intelligence Technologies” category of the BRICS Young Innovators Prize.

Your work is closely connected to artificial intelligence (AI) technologies. How would you assess their development in Russia and BRICS countries?

I would assess it as very high. In recent times, there have been many publications on both fundamental and applied research in the field of AI, both in Russia and in BRICS countries, including China. These countries are focused not only on the development of core technologies – for example, new algorithms and architectures – but also on the applied use of AI, which is very important in demonstrating the value of this technology.

In which areas are AI and machine learning technologies most commonly applied? In which fields could they open up new opportunities?

The most obvious areas for the application of machine learning and artificial intelligence are image processing, extracting various characteristics from images, and image classification. We see this daily – from street surveillance cameras to satellite remote sensing data processing, which allows us, for example, to automatically generate surface characteristic maps across Russia. This greatly simplifies human effort.

AI algorithms are also applied in various predictive tasks. For example, it is possible to forecast equipment failures based on current operational parameters, which can optimise production processes and save resources.

In addition, fundamental models based on LLMs (large language models) are widely used, enabling non-specialist users to handle specific tasks via certain requests. This simply makes everyday life easier.

If we talk about areas where this could still be applied, I believe that almost all directions have already been covered. Now the focus is on hybridising different approaches to solve common problems in a way that requires minimal task formalisation from a person.

You have developed a unique technology for predicting wildfires. Could you tell us more about it?

This technology makes it possible to predict the likelihood of wildfires occurring and spreading in a given area. The solution has already been integrated into the operations of the Ministry of Emergency Situations of Russia.

Input data includes weather and meteorological conditions, as well as satellite data describing the Earth’s surface – what kind of forest is growing there, whether it is dry or not. Additionally, some parameters relate to population density – areas with higher population density have a greater risk of fires. Other data – topographical and static – also affect the probability of a fire starting and spreading.

At the core of this technology are AI algorithms, including convolutional and recurrent neural networks. The solution has already proven effective, and several fires in Russia have been predicted using it. The technology is now being improved, including the addition of new features that can influence forecast accuracy.

Wildfires and landscape fires often occur due to human negligence. How is the human factor accounted for in this development?

Accounting for the human factor is the most difficult part, as there are no formalised databases capable of automatically supplying information about people’s behaviour. At present, we factor it in indirectly – through population density and the distance of infrastructure from roads. In the future, it would be possible to formalise certain datasets in a way that makes them useful and informative for AI algorithms.

Is there potential for this technology to be deployed in BRICS partner countries?

Absolutely, we have already had discussions with colleagues. Brazil has vast forests, and China has large territories. So, this initiative is primarily aimed at BRICS countries, where environmental and fire monitoring is an important issue for economic development.

Our solution can be easily adapted for these countries – the entire algorithm, the architecture, and the approach can be transferred without changes. It might need to be adjusted to suit the specifics of a region, but this would only take a few months.

You received an award for third place in the BRICS Young Innovators Prize and were the only Russian on the winners’ podium. How would you describe the competition?

The competition was quite intense. It was a very interesting experience for me – I discovered a lot about the challenges BRICS countries face and how they are using AI to solve specific problems. There were projects of the highest level. For example, a scientist from Brazil working on sugarcane processing technologies with the help of different modelling methods. This is critically important for their economy.

You teach AI and computer vision disciplines at Irkutsk National Research Technical University. How interested are the students?

The students are genuinely interested in learning this technology because they have strong skills in solving applied problems and know a great deal about the specifics of certain industries – mining and processing. They can integrate AI into traditional technologies, such as gold ore flotation, helping to optimise the process and achieve a new level of efficiency.

Students see that AI technologies can be applied everywhere in their practical work. For example, during the flotation process, it is possible to assess the foam density in real time and automatically adjust parameters. All this happens in an automated system where human participation is not required, and artificial intelligence manages it more efficiently.

What kind of scientific and methodological framework exists in our country for teaching these disciplines? What experience from BRICS and BRICS+ countries would be of interest to Russia? What expertise could we share with other nations?

We have a very large scientific foundation in our country. There are AI centres where professionals in this field are trained. All the specialists working on applied projects in these centres can act as lecturers, sharing their research and broadening the knowledge base.

Additionally, we have strong computational resources. For example, at Skoltech, there is a GPU cluster capable of running large models. This enables students to learn how to work with such hardware and develop new architectures adapted for it. We are ready to share this experience with BRICS countries.

I believe we would be very interested in the applied AI developments in BRICS countries aimed at addressing specific industrial challenges, and in adopting these for solving certain tasks here in Russia.

How important is the research aspect in education? What does it give to school pupils and university students?

It would be great if schools included more research-based tasks and projects because school pupils often learn formulas up until Year 11 without seeing how they can be applied. In the senior years, it would be possible to combine physics and mathematics and show that they are not two different sciences.

Of course, diploma work offers the opportunity to feel like a researcher. For example, from my second year at university, I was based at the Space Research Institute, part of the Russian Academy of Sciences, where serious scientists work. This gave me a chance to interact with them, be inspired by their work and start developing myself in research.

Naturally, the research aspect is important. It proves that everything we study is not pointless and that we can draw our own conclusions based on this knowledge.

Watch the full interview here.

Photo: TV BRICS