Skoltech3D

In June 2024, the Skoltech AI team and its Russian collaborators presented a study on the AI trends of 2023, which was included in Stanford University’s AI Index Report. The paper introduces Skoltech3D, a large dataset that can be used to generate 3D models of real-world objects. Stanford experts noted a significant advancement in Skoltech3D compared with earlier datasets: For the first time, large-scale data with high-precision references was collected from multiple devices: cameras and lidars.

In 2024, the Skoltech AI team achieved a breakthrough by developing a method that can reliably determine whether text was written by a human or a machine, even with limited data. According to the researchers, this achievement resulted from combining their unique findings with proven AI models. The team presented its research at a language modeling conference hosted by the University of Pennsylvania and received the Outstanding Paper Award.

Scientists from Skoltech AI and AIRI Institute have proposed a new approach using optimal transport to solve a wide range of complex computational problems that frequently arise in machine learning and mathematical modeling. The new approach can train models three to 10 times faster. The researchers presented their work at one of the world’s largest AI conferences, NeurIPS-2024, in Vancouver.

In 2024, Skoltech AI, in cooperation with the Shirshov Institute of Oceanology of RAS, developed an innovative system for forecasting ice conditions throughout the Arctic, including along the Northern Sea Route, with a high resolution of about 3 kilometers. Leveraging advanced AI technology, the system integrates machine learning with physical models and data from satellites, stations, and buoys. This significantly increases the speed and accuracy of forecasts, achieving over 90% accuracy for ice consolidation forecasts. This solution is of interest to Russian and international companies operating in this strategic region. It is expected to save up to 100 billion rubles per year by reducing risks, accelerating transportation, and boosting cargo turnover on the Northern Sea Route.

Skoltech researchers and their colleagues from Sber proposed deep learning models that predict droughts using climate data. Long-term forecasts of this kind are in high demand among agricultural producers planning their operations, as well as insurance companies and banks evaluating climate risks and fine-tuning corporate credit ratings. The research was conducted as part of an industrial partnership between Skoltech AI and Sber.

Researchers from Skoltech AI developed a model that can predict new wildfires up to five days in advance. The model successfully completed its pilot operation and acceptance tests at the Russian Ministry of Emergency Situations (MChS) in 2024. This year, the model will be integrated in the Ministry’s Atlas of Hazards and Risks, and its predictions will be used in the ministry’s operations. The technology has received high international acclaim and could be adopted by other countries with extensive wildlands.

A team from Skoltech Bio, along with international colleagues, found out how the PARIS antiviral immunity system works. The researchers also demonstrated a new function of viral tRNA molecules. This discovery helps answer a question that has plagued scientists for over 50 years: Why do viruses encode their own tRNA molecules? The study reveals these molecules to be used by viruses to inhibit the host immune system. The presence of such tRNAs in human-infecting viruses underscores the universality of the discovery.

* The teams’ findings were examined by the CoLab.ws project and the editors of Chimica Techo Acta. The best papers were selected based on the journal’s impact factor.

Life and health sciences and agrotechnology

In December 2024, a team from Skoltech Neuro, Motorica, and the Far Eastern Federal University announced a breakthrough. During a study launched in 2021, the scientists created a prototype of a sensory feedback-enabled prosthetic arm. Previously, amputees using the prostheses could only identify objects as hard or soft and determine their size by touch. Now, they can feel movement in their prosthetic fingertips and distinguish between freely moved objects. One patient with an amputated foot reported feeling the heel of the prosthesis thanks to a smart insole.

Researchers from Skoltech Neuro created a “fatness map” of the human brain. Since mental and cognitive disorders are associated with disturbances in the brain lipidome, the map will enable a better understanding of the mechanisms behind such disorders.

A team from Skoltech Agro developed a next-generation platform for forecasting crop yields and monitoring crop health, dubbed EASY for Environment-Augmented Smart Yield. It provides early-season forecasts with an accuracy of up to 80%, far surpassing the industry standard of 50%-55% for crops such as sugar beets, corn, sunflower, and soybeans. The solution is already in high demand among agribusinesses.

Determining an animal’s genotype helps scientists better understand its genome. Experts in quantitative genetics at Skoltech Agro used genotype imputation and genome-wide association studies to study beef productivity. They found that 1,675 single-nucleotide polymorphisms (SNPs) in the cattle genome were associated with important economically relevant traits, such as carcass weight, meat marbling score, rib area, and back fat thickness. Current commercial SNP chips lack these characteristics. The Skoltech team has successfully patented its SNPs.

Researchers at Skoltech Omics Technologies developed and validated a method for the quantitative analysis of over 100 blood plasma proteins using a set of isotope-labeled reference peptides, including potential biomarkers of Alzheimer’s disease. The method will facilitate the early prediction of the disease using quantitative clinical proteomics. The study was conducted as part of the Ministry of Education and Science project titled “Development of Molecular Tools for the Analysis and Management of Neuropathologies.”

Researchers from Skoltech Engineering and their international team developed a groundbreaking method called Dimming InfeRred Estimate of CME Direction (DIRECD). By combining fundamental achievements in solar physics with operational space weather forecasting tasks, this method enables the rapid 3D estimation of coronal mass ejections and their propagation directions. DIRECD can significantly reduce the risk to satellites, navigation systems, and critical infrastructure, preventing the potentially massive costs for having to deal with the aftermath of a particularly severe solar storm.

Advanced materials and engineering

Researchers at Skoltech Petroleum found a way to produce hydrogen from natural gas right in the gas field by injecting steam and a catalyst into a well and adding oxygen to create a high-temperature environment in the reservoir. The catalyst- and vapor-assisted combustion produces a mixture of carbon monoxide and easily extractable hydrogen. This technology will accelerate the transition from fossil fuels to clean hydrogen power.

Researchers from Skoltech Petroleum created a digital twin of a polymer composite material with an embedded 2D sensor and successfully used it for structural integrity testing. The new technology can be used to manufacture — and monitor the state of — various large structures, such as aircraft wings, wind turbine blades, and bridge spans, which are made of polymer composites. The layer of carbon nanotubes that serves as the sensor, which the team inserted between the fiber-reinforced composite layers, is so thin that it does not affect the thickness, mechanical properties, and overall design of the part.

Researchers at Skoltech Petroleum developed and successfully tested a unique experimental setup for simulating gas dynamic processes in cryolithozone rock and predicting the probability of explosive methane emissions. The setup, which is the only one of its kind in Russia, can be used to conduct experiments with permafrost samples under a variety of simulated natural conditions. A specific algorithm can be employed to adjust the medium’s temperature and gas pressure, and the unique sample holder can compact the soil and measure the changes in its linear dimensions, porosity, and permeability during the experiment. The setup is fitted with acoustic sensors to determine the ratio of ice and unfrozen water in the rock pores. The solution will facilitate the study of a wide range of complex natural processes that affect frozen rock in the Arctic.

As part of the development of a 3D printer for the Russian segment of the International Space Station and the future Russian orbital station, researchers at Skoltech Materials created a prototype for a five-axis composite printing technology, where the nozzle moves in three dimensions and the platform rotates around two axes. The prototype of the new 3D printer, which will be used to make fiber-reinforced polymers, is expected to be ready by the end of 2025. The team is also developing an extruder to recycle plastic waste from printing, which will help optimize the development of 3D objects on board the ISS.

Researchers from Skoltech Materials and their colleagues from other research organizations enhanced the quality of a carbon-based electrode material by exposing it to air plasma. This helped improve the electrode’s properties, which are a key constraint for the performance of fuel cells — promising devices for efficient and environmentally friendly generation of electricity from combustible fuels. Fuel cells are used in industrial facilities and households, especially in hard-to-reach, unelectrified areas. A perfect solution for backup power sources, this mode of power generation is used in spacecraft, submarines, refrigerated warehouse loaders, cars, buses, trains, and boats.

From May 29 to June 1, 2024, Skoltech hosted the second Sino-Russian Symposium on Chemistry and Materials Science. The participants were greeted by Skoltech President Alexander Kuleshov, Distinguished Professor Artem R. Oganov, who heads the Material Discovery Laboratory, RAS Vice President Stepan Kalmykov, and Chinese Academy of Sciences Vice President Wu Zhaohui.

Fostering large-scale transition to electric vehicles and renewable energy

Researchers from Skoltech Energy presented a model that facilitates the design and operation of industrial vanadium redox flow batteries, which will play a significant role in the energy transformation. These huge batteries are already widely used to mitigate peak loads in the power grids of China, Germany, and the United States. This or similar technology is essential for supporting the widespread adoption of electrical vehicles and renewable energy. It will also enhance the efficiency and safety of nuclear power plants and provide backup power for critical infrastructure.

Energy efficiency and energy transition

Researchers at Skoltech Energy Transition developed new machine learning models to simulate the structure and properties of high-entropy functional materials. They shared the simulation results with their colleagues at Tomsk Polytechnic University, who successfully synthesized and analyzed the properties of new materials, which have potential applications in the chemical, mining, and electronics industries.

As part of the climate cooperation agreement signed in September 2024, Skoltech and Sibur will develop innovative decarbonization projects. Together, they will identify advanced solutions to improve energy and resource efficiency and reduce greenhouse gas emissions at Sibur’s production facilities, as well as in processes and projects indirectly related to its operations. One pilot area of collaboration will involve using microalgae to capture carbon dioxide from flue gases and convert it into useful products.

Researchers from the Plasmonics Laboratory at Skoltech Physics reviewed the current approaches to silicon integrated photonics. They described the photonic integrated circuit components and devices developed based on these approaches and compared them with alternative technological platforms. Silicon integrated photonics forms the core of optical modulators, photodetectors, optical filters and switches, multiplexers and demultiplexers, and optical transceivers. In many respects, silicon integrated photonics competes with alternative platforms based on indium phosphide, silicon nitride, silicon oxide, and other materials.

Telecommunications, photonics, and quantum materials

Researchers at Skoltech Photonics and their international and Russian colleagues demonstrated the spontaneous formation and synchronization of multiple quantum vortices in optically excited semiconductor microcavities. The researchers showed that polariton quantum vortices formed in the neighboring cells of optically generated lattices tend to have an opposite topological vortex charge. In other words, the vortices are “antiferromagnetically coupled.”

Researchers from Skoltech Photonics and the University of Wuppertal in Germany created a universal logic element, NOR, where N is a negation operator and OR is a logic sum operator. Based on polariton condensates, the gate functions at room temperature, has multiple inputs, and is hundreds of times faster than electronic gates. Also, it is fully optical, meaning it does not require an electric current. Importantly, such logic gates can be assembled into optical circuits and cascaded.

A joint team of physicists from the Nanomaterials Laboratory at Skoltech Photonics, MIPT, and ITMO University developed an optical component that helps manage the properties of a terahertz beam and split it into several channels. The new device can be used as a modulator and generator of terahertz vortex beams in medicine, microscopy, and 6G communications.

Researchers from the Biophotonics Laboratory at Skoltech Photonics synthesized tannic acid-coated magnetic beads capable of extracting membrane vesicles called exosomes from biological fluids with 60% efficiency. The chemical composition of the tiny exosomes, which measure 30 to 100 nanometers in diameter, can help detect cancer in its early stages and assess treatment effectiveness. Going forward, the team plans to improve the surface structure of the beads to further enhance their performance.

Researchers from Skoltech Photonics presented a new solution at Holoexpo 2024, a major conference on advanced photonics. The optical system, which was showcased in the Digital Holography and Visualization Techniques section, can be used in basic research, including optical printing technology, the development of analog simulators and optimizers, and some practical applications, such as optical tweezers, metrology, free-space optical communication lines, and lidars.

Researchers from the Krichever Center and their colleagues from other research institutes pioneered the use of parallel computing on graphics cards to simulate acoustic turbulence. Previously requiring a supercomputer, this type of simulation can now be run on a standard personal computer. This advancement will improve the accuracy of weather forecasting models and enable the application of turbulence theory in various fields of physics, including in astrophysics — to calculate the trajectories and propagation speeds of acoustic waves in the universe.

Advanced studies

The first international school titled “Integrable Systems and Algebraic Geometry” took place in Beijing from June 24 to July 5, 2024. It was dedicated to Igor Krichever, the late director of Skoltech’s Center for Advanced Studies, which now bears his name. Employees of Skoltech and the Beijing Institute of Mathematical Sciences and Applications (BIMSA)

Skoltech’s core facilities have deployed process, assembly, and measurement equipment, manufactured and characterized diverse passive and active components, from waveguides to diffraction gratings to thermo-optical modulators. Currently, the team is compiling a component library (PDK) for the first passive elements, including waveguides and multimode interference splitters of various configurations. A prototype is in development that will incorporate a radiation source, an integrated optical modulator, and a radiation receiver. Skoltech’s Research Facilities Center includes six core facilities fitted with state-of-the-art equipment.

Research Facilities Center