Weronika Węgier, M.Sc. and Piotr Gramacki, M.Sc., are among twelve people from our University whose research has been supported with grants under the next edition of the Preludium programme held by the National Science Centre. The received a total of PLN 2 million.
Preludium is a globally unique competition, which allows researchers at an early stage of their careers to gain experience in conducting research independently, before obtaining their doctoral degree. In this formula, novice scientists do not have to compete with more experienced scientists, and an important element of the project is the involvement of a research supervisor who supports the manager in its implementation.
In the 23rd edition, a total of PLN 61.2 million in grants was awarded to 362 projects. This group includes 10 researchers from Wrocław University of Science and Technology. They will receive almost 2 million PLN.
Weronika Węgier, M.Sc. (Faculty of Information and Communication Technology)
„Application of explainable artificial intelligence algorithms in order to improve imbalanced data methods.” Amount of funding: 90,036 PLN.
The growing popularity of artificial intelligence brings about the need to understand it better. The newly proposed methods aim to explain the mechanisms behind decision models by determining what influenced the system's response. They help inter alia to improve algorithms, especially in cases where the defined reasoning is based on incorrect premises, such as artifacts in photos.
In her project, Weronika Węgier proposes using explainable artificial intelligence (XAI) methods in the problem of unbalanced data, i.e. data in which one of the classes has a much smaller number than the others. An example of such an application could be diagnosing rare diseases or detecting network attacks, in which significant instances (a sick patient or an attack) are less common than the instances less important for detection (a healthy patient, standard network traffic).
Here, XAI methods are intended to indicate which features of samples, especially less common classes, negatively affect their correct recognition. This information will allow the quality of models to be improved, e.g. through their guided pruning or processing the data used for training.
Piotr Gramacki, M.Sc. (Faculty of Information and Communication Technology)
„Geospatially oriented language models” Amount of funding: 210,000 PLN.
Spatial data are a valuable source of data because they link semantic information with specific geographic locations. Large Language Models (LLM), currently used on a large scale, lack mechanisms for efficient processing and use of this data. This type of data is time-variable, which is why we cannot rely on the ability of these models to remember spatial information. Furthermore, due to the uneven representation of different regions of the world in the training sets, currently existing models perform poorly in tasks involving less developed areas of the world.
Therefore, in his project, Piotr Gramacki intends to develop language models that integrate spatial data, ensuring correct operation regardless of geographical location. Our scientist plans to create a diverse set of text data enriched with spatial context. On its basis, he will develop language models capable of understanding and generating coherent text taking into account the spatial context.
As a result, the models will be able to correctly answer questions about locations, recognize spatial dependencies and use available sources of spatial data to provide information consistent with geographical facts. In the future, they can help build tools that will be used in city planning, tourism, crisis management or environmental monitoring.
Patryk Fałat, M.Sc. (Faculty of Chemistry)
„POLY-LLAMA: Functional POLYmeric surfaces for Light-triggered LAnthanide-doped-Materials-Aided disinfection." Amount of funding: 209,999 PLN.
The overuse of chemical disinfectants and antibiotics in the treatment of infectious diseases contributes to the emergence of extremely resistant strains of bacteria, viruses and fungi. As a result, chemical disinfection of surfaces in medical facilities and public transport is ineffective. An alternative is disinfection methods based on physical factors, e.g. ultraviolet radiation. However, UV lamps used in medical offices and operating rooms, due to their negative impact on health, can only be turned on periodically.
The aim of Patryk Fałat's project is therefore to develop polymer functional surfaces containing materials doped with lanthanide ions, which are capable of converting visible (VIS) or infrared (NIR) radiation into low-intensity UV radiation.
This solution will allow continuous, localized neutralization of pathogens in a way that is safe for human health. The synthesized nanomaterials and microcrystals containing lanthanide ions will emit UV radiation thanks to the process of upconversion of photon energy. They will be placed in polymer matrices characterized by high permeability to UV radiation and good mechanical properties, e.g. abrasion resistance.
In the final stage of the project, the effectiveness of photoinduced disinfection of the produced surfaces will be tested in experiments using bacteria in the form of biofilm, viruses and fungi. Functional polymer surfaces for photoinduced disinfection supported by materials doped with lanthanide ions are to ensure effective and safe control of pathogens, which is of great importance in the context of the growing problem of microbial resistance to conventional disinfection methods.
Dominika Benkowska-Biernacka, M.Sc. (Faculty of Chemistry)
„Gold Nanorods for Lipidic Mesophases Imaging and Modification (GLIM)” Amount of funding: 209,991 PLN.
An important aspect of research into soft matter is to develop new concepts leading to a better understanding of biological structures with liquid crystalline properties. One of the basic examples of mesophases of biological importance is a microstructure composed of concentrically ordered lipid bilayers, known as myelin figures (MFs). Pathological changes in this lipid-rich biomembrane lead to less effective conduction of the nerve impulse and are associated with the occurrence of incurable diseases of the central nervous system with causes not fully understood (e.g. multiple sclerosis).
Therefore, much attention is paid to research on the mechanisms underlying disorders in lipid organization, which can contribute to progress in the diagnosis and treatment of demyelinating diseases, often using model structures for this purpose. Interdisciplinary research on lipid mesophases may provide new perspectives for in vitro and in vivo imaging. In this context, the combination of gold nanorods (GNRs) with MFs may lead to the development of a new approach to detecting subtle changes in lipid mesophases.
During the project, Dominika Benkowska-Biernacka intends to conduct experiments aimed at selecting optimal GNRs for potential applications in myelin sheath studies. A key aspect of the research will be the preparation of hybrid materials composed of lipid bilayers and GNRs, which will be characterized using advanced microscopic techniques, such as two-photon microscopy with temperature and polarization control.
The first result of the project will be a better understanding of the interactions between GNRs and lipid bilayers and a description of the effect of nanoparticles on MFs in order to design new materials with specific morphology and thermal stability. The second result will be the development of a novel approach to imaging lipid mesophases of biological importance.
Sindu Daniarta, M.Sc. (Faculty of Mechanical and Power Engineering)
„Experimental and predictive study of wet-to-dry volumetric expansion process in organic Rankine cycle systems” Amount of funding: 209,840 PLN.
The concept of an organic Rankine cycle with a two-phase expansion system has a significant theoretical potential, offering the possibility of increasing the efficiency of energy conversion in various applications. Despite promising prospects, this area remains relatively poorly explored, opening interesting opportunities for further research.
Therefore, the aim of Sindu Daniarty's project is to fill this gap by conducting a comprehensive experimental analysis of the two-phase expansion process, with particular emphasis on understanding and predicting the isentropic efficiency of the two-phase expansion process under various conditions. By providing experimental data and developing predictive models, the research aims to significantly expand the area of knowledge on two-phase expansion systems.
The results will not only allow the verification of theoretical models, but also contribute to the development of more efficient, reliable and sustainable energy systems. The project is carried out under the supervision of Prof. Piotr Kolasiński.
Michał Biały, M.Sc. (Faculty of Mechanical Engineering)
„Thermoplastic processing-induced changes in serrated flow behaviour and plastic deformation mechanisms of bulk metallic glasses” Amount of funding: 209,960 PLN.
Bulk metallic glasses (BMGs) are advanced metal alloys with exceptional mechanical properties, which are increasingly used in the automotive, space and biomedical industries. Unlike traditional, crystalline metal alloys, metallic glasses have an amorphous internal structure, which is characterized by a lack of long-range atomic order and is the source of unique properties of this class of materials. Due to the amorphous structure, heated BMGs show a continuous decrease in viscosity over a wide temperature range. Despite being metals, they can be formed like polymers or typical oxide glasses.
The project aims to increase the practical use of BMGs in industry and to observe the growing importance of thermoplastic processing as a method of forming very complex shapes. In the project, the our researcher will therefore focus on investigating how thermoplastic processing affects the behaviour of BMGs during deformation at lower temperatures. This information is crucial for the safe use of elements manufactured with the above-mentioned method. The research will include a detailed analysis of the mechanical behaviour of materials before and after thermoplastic processing. The project will use advanced nano- and microindentation techniques to determine the parameters characterizing the basic units of plastic deformation in these materials, the so-called shear transformation zones (STZ). It is expected that the project will provide more information on the effect of thermoplastic processing on the mechanisms of plastic deformation of BMG and determine the optimal processing parameters to obtain the desired mechanical properties. The results will be the basis for designing BMGs with improved plasticity, for better meeting the requirements of modern industry.
Ewa Mańskowska, M.Sc. (Faculty of Electronics, Photonics and Microsystems)
Analysis of innovative p-type semitransparent thin film nanostructures based on mixed oxides of copper and titanium for gas sensors application” Amount of funding: 207,400 PLN.
The project focuses on the analysis of sensing properties of thin layers of mixed copper and titanium oxides, produced by magnetron sputtering. Gas sensors based on metal oxides such as tin, tungsten, titanium or copper, allow the detection of explosive and harmful gases thanks to changes in the layer resistance resulting from the adsorption/desorption of gases on the oxide surface. Mixtures of oxides having different types of conductivity are promising gas-sensitive materials, which often exhibit better properties than single oxides. The heterostructures of copper and titanium oxides are still poorly understood.
Therefore, the aim of Ewa Mańkowska's project is to produce and analyse the properties of layers with different proportions of oxides (e.g. 30% copper oxide and 70% titanium oxide) and to investigate the possibility of using them as gas sensors. The planned research includes measurements of gas detection (e.g. hydrogen, ethanol, nitrogen oxides) at different concentrations in the air. The analysis will also include the structure and morphology of thin films, which will be modified by appropriately designing the material composition and thermal treatment.
A key step will be to determine the optimal operating conditions of the sensors, such as operating temperature, and an in-depth analysis of the sensor response in correlation with surface morphology, chemical composition and electrical properties.
The results will recommendations of methods for producing layers with the best sensing properties. In addition, the analysis of additional properties, such as transparency, may lead to the development of multifunctional layers that combine the ability to detect gases with other functional features.
Mateusz Moj, M.Sc. (Faculty of Civil Engineering)
Evaluation of selected functional properties of concrete industrial flooring surface layers based on nondestructive testing using artificial intelligence (SURFACE AI) Amount of funding: 189,453 PLN.
The aim of the project is to develop a comprehensive system for assessing concrete floors, which would be based on non-destructive methods supported by artificial intelligence techniques. The key objective is to replace traditional destructive tests with methods of comparable accuracy, which at the same time minimize interference with the tested structure. The project includes the use of acoustic, sclerometric methods and surface morphology analysis, the accuracy of which will be optimized by using machine learning algorithms.
The planned research will answer such important questions as: Is it possible to completely eliminate destructive methods from the process of assessing the functional properties of concrete floors? What technological errors have the greatest impact on the functionality of floors? How does the microstructure of the surface layer change when modifying the Dry Shake Topping technology? What is the relationship between surface morphology and abrasion resistance?
The answers will bring benefits to both the scientific community and the industrial sector, where minimizing damage and precise quality assessment are important. The innovativeness of the project is emphasized by the use of artificial intelligence, including algorithms supporting the assessment of strength and surface analysis, which can revolutionize the method of identifying selected properties of concrete floors.
The project has the potential to become a breakthrough solution in the field of diagnostics and assessment of building materials.
Aleksandra Piasecka, MSc. (Faculty of Fundamental Problems of Technology)
„QF-BEC: Quantum Fluctuations and dynamics of Bose Einstein Condensate of photons in semiconductor lasers” Amount of funding: 139,080 PLN.
Laser action is similar to the Bose-Einstein condensation phenomenon in many respects, because it is characterized by a macroscopic number of particles and oscillations of particles in phase, and thus also by coherence. Recently, our scientists at WUST discovered that BE condensation of photons can occur in standard semiconductor lasers. This discovery has naturally led to the question: what is the difference between lasing and condensation and how to distinguish one from the other?
Aleksandra Piasecka's project aims to help answer these questions by systematically studying the particle number fluctuations and dynamic response of conventional semiconductor lasers operating in the BE photon condensation mode. The intensity fluctuations will be studied by means of experimental measurements of second-order photon quantum correlations. In the photon condensation mode, the fluctuations are expected to be extremely large, while in the case of lasing, the intensity fluctuations are strongly suppressed, which is a known property of coherent light emission. The experimental observation of large fluctuations would prove that the photon condensate is a thermodynamic source of coherent light, in which the macroscopically occupied emitting state is the result of a statistical distribution of bosons, and not of stimulated emission.
Aleksandra Piasecka also hopes to prove that correlations will be a tool to distinguish conventional lasers from photon condensates.
In addition, the project will investigate the dynamic properties of photon condensation in a semiconductor laser. This will be achieved by characterizing the dynamic frequency response of the photon condensate to perturbations in the carrier density induced by low-amplitude modulation of the control current. Owing to this dynamic response, it is possible to investigate the nonlinearities of the system and their effect on the condensation of light.
Adrian Kowaliński, M.Sc. (Faculty of Chemistry)
„Design, synthesis and catalytic activity studies of heterometallic rare-earth complexes in the synthesis of polymer anticorrosion coatings” Amount of funding: 138,958 PLN.
Our researcher's innovative project involves the development of new, precise methods for the synthesis of heterometallic clusters of rare earth metals and selected d-block elements with specific physicochemical properties and reactivity. The key element of the experimental studies will be the use of homometallic precursors of zinc and magnesium.
This will allow for the creation of structures with a strictly controlled arrangement of metal ions in the crystal lattice. Such an approach will facilitate the study of the relationship between the structure of these compounds and their catalytic activity and other properties.
The resulting complexes will be used as catalysts in the alcoholysis of cyclic anhydrides, which together with the resulting mono- and diesters and glycerol will be used as reagents for the synthesis of alkyd resins, i.e. biodegradable polymers used in the production of anti-corrosion protective coatings for metal surfaces. The use of initiators differing only in the type of metal or the arrangement of ligands will enable the production of polymers with varied corrosion resistance. They will enable the production of materials adapted to specific environmental conditions and technical requirements.
It has also been proven that the presence of rare earth ions in alkyd resins can significantly improve their corrosion resistance and even enhance the self-healing properties of such coatings.
Kamila Łupińska, M.Sc. (Faculty of Chemistry)
„Random number generation based on dual lasing emission in scattering media” Amount of funding: 69,906 PLN.
The development of information technology has contributed to the increase in the number of cybercrimes, which means that the protection of personal data requires increasingly advanced methods, such as digital signatures. A key element of their operation are Random Number Generators (RNG), and their most important feature – unpredictability. They are used in financial institutions, cryptography, cybersecurity, gambling and when creating random passwords.
The aim of Kamila Łupińska's project is to generate random numbers based on double laser emission in scattering media, using inter alia the AIE (Aggregation Induced Emission) phenomenon, which enhances light emission and supports multiple scattering for feedback. The research will compare static disorder obtained using thin polymer layers doped with an organic dye with dynamic disorder generated in a solution of the same compound.
The hypothesis assumes that dynamic disorder can facilitate the generation of random numbers, because the quasimodes occurring in random lasering (RL) change with each laser action. The project results will help understand how the aggregation phenomenon affects the randomness of RL and contribute to the development of alternative methods for creating RNGs.
Kamila Jankowska, M.Sc. (Faculty of Electrical Engineering)
„Application of deep learning methods in detection, classification and compensation of the measuring sensors faults in permanent magnet synchronous motor drive system” Amount of funding: 60,634 PLN.
Permanent magnet synchronous motors are an integral part of modern industrial applications. The most important areas of their application include electromobility, space technologies, aviation and broadly understood manufacturing technologies. Improving the level of safety and reliability is becoming crucial for the optimization of industrial processes. To meet these requirements, fault-tolerant control systems are used, which include detection, classification and compensation of the largest possible number of failures. For this purpose, complex algorithms for machine diagnostics are being developed, based largely on measuring sensors.
The main goal of Kamila Jankowska's project is to develop new algorithms for detection, classification and compensation of measuring sensor failures, based on deep learning methods. The operation of the developed algorithms will be analysed for motors of different rated power in simulation and experimental studies, in order to obtain a possibly universal solution.
There are also plans to conduct research on the use of deep neural networks in the estimation of state variables, in order to finally develop an innovative, fully automated fault-tolerant control system.
The project assumes the use of deep learning methods to eliminate the shortcomings of solutions described in the literature so far. This approach will allow solutions to be obtained faster than in the current methods of damage detection and will allow the classification of a greater number of failure types.
Gallery
