Current initiative topics of the department (as of December 2025):
0121U108261 “High-performance computer systems and networks: theory, methods, and means of hardware and software implementation.” Head of research: A. M. Volokyta.
0119U102212 “Methods and means of increasing the efficiency of parallel computing in systems-on-a-chip.” Head of research: V. I. Zhabin.
0124U003323 “Nonlinear and multi-criteria mathematical models for Data Science and Embedded Systems technologies.” Head of research: O. A. Pysarchuk.
2026 “Development of models and methods for the functioning of fault-tolerant cyber-physical distributed systems with dynamic structure.” Head of research: M. A. Novotarskyi.
“Knowledge At the Tip of Your fingers: Clinical Knowledge for Humanity” (KATY), Grant agreement ID: 101017453, a project implemented within the European Union’s HORIZON 2020 program. European Union. European Commission. (2021 – 2024, Project Lead: Doctor of Technical Sciences, Prof. Yu. I. Yakymenko; Principal Investigator from the Department of Computer Engineering: Prof. S. H. Stirenko).
- Artificial Intelligence (AI) offers vast potential for the future of personalized medicine. By promising individual treatment for patients, AI can help win the battle against serious diseases such as cancer. However, the implementation of AI-supported personalized medicine also poses challenges. Chief among these is the transformation of AI-based suggestions into practical decision-making processes and treatment strategies. The EU-funded KATY project will develop an AI-based personalized medicine system that will significantly assist medical professionals and researchers in utilizing and interpreting AI data in their daily work. This next-generation technology will bridge the gap between AI data and medical applications, thus becoming a powerful tool for diagnosing, treating, and overcoming serious diseases.
- Method for adapting standard convolutional neural networks (CNNs) for medical applications.
- An analysis of electroencephalography (EEG) data for various activities related to HCI was performed, which may be useful for supporting the daily lives of such individuals. Recently, several approaches based on artificial intelligence methods, such as neural networks (NN)—for example, Fully Connected NN (FCN), Convolutional NN (CNN), and Recurrent NN (RNN)—have been successfully used for EEG data analysis. Some new attention-based NN architectures (wA) are very promising in various applications. This work is dedicated to investigating various hybrid combinations, such as FCN-CNN, CNN-RNN, CNN-wA, RNN-wA, CNN-RNN-wA, etc., regarding EEG data analysis. These hybrid models were trained on the “Grasp-and-Lift” (GAL) dataset, where users use their hand to manipulate a smartphone.
- Method for adapting combined hybrid neural networks based on convolutional, graph, recurrent components, and attention mechanisms for medical applications.
“Artificial Intelligence Platform for Remote Automated Detection and Diagnosis of Human Diseases” (2020 – 2021, Project Lead: Prof. S. H. Stirenko).
- Advanced technology and developed software for the analysis of images obtained via CT, MRI, and radiography (X-ray) based on an artificial intelligence system using deep neural networks, which can automatically help identify signs of 14 lung diseases (including coronavirus pneumonia).
- A study was conducted on a theoretically and practically significant problem that is highly relevant today, namely the improvement of video image quality using neural networks. Several attempts to enhance image quality by modifying the block components of the considered method were examined, and the feasibility of such changes was evaluated to improve image and video quality metrics under practical application conditions.
- A study was conducted on image compression methods to determine the core features that directly influence the application of the researched methods in experiments. Analysis results were obtained with statistically significant accuracy regarding the performance of an image compression method based on a neural network using channel-level data quantization applied to the Kodak dataset. A method for determining the effectiveness of accuracy enhancement—achieved by increasing the number of features in the encoder and decoder—was proposed and defined.
“High-performance computer systems and networks: theory, methods, and means of hardware and software implementation” (2022 – 2023, Project Lead: Prof. H. M. Lutskyi).
- Methods and tools for network organization in data centers.
- Methods and tools for increasing the efficiency of IoT systems.
- Methods and tools for increasing the operational efficiency of information systems based on modern computer systems and networks.
- Improvement of computer network technologies using Software-Defined Networking (SDN).
- Artificial Intelligence (AI).
- A method of network organization in data centers that allows achieving specified characteristics by combining tree-based topologies with a graph based on de Bruijn code transformations.
- Methods and ways to accelerate the recovery of the replication factor in distributed data storage.
- Methodology for developing a scalable information and control system based on a dynamic functional model of the complex hierarchical structure of an educational institution.
- Improved architectural and functional concept of IoT monitoring systems using intelligent analysis methods based on neural networks.
- A new method for classifying accelerometer linear acceleration indicators using convolutional neural networks (CNN).
“Methods and means of increasing the efficiency of parallel computing in systems-on-a-chip,” state registration № 0119U102212 (2019 – 2023, Project Lead: Prof. V. I. Zhabin).
- Algorithms for accelerating the execution of data-dependent basic binary operations in parallel streaming systems. A dynamic method for error correction during the execution of parallel tasks in the event of system computing module failures has been proposed.
- Methods of dynamic reconfiguration for data-flow controlled systems in the event of hardware failure, which allow for a reduction in reconfiguration and system recovery time when an operational module fails.
- Methods for increasing fault tolerance of distributed systems controlled by descriptor flow.
- Methods for increasing the efficiency of implementing data-dependent operations.
- Method for increasing the efficiency of fine-grained computing in DATAFLOW systems and a method for increasing the fault tolerance of a data-flow controlled computing system (computing device – utility model application).
- Method for synthesizing fault-tolerant topologies using a Latin square with redundant encoding of vertex numbers. Methods for filling the square were considered, several topologies were synthesized, and a characterization analysis was performed. The possibilities of utilizing redundancy were analyzed. (Honcharenko).
- Method for increasing the efficiency of fine-grained computing in DATAFLOW systems.
- Method for increasing the fault tolerance of a data-flow controlled computing system.
- Software and hardware tools for increasing the efficiency of production activities using Internet of Things (IoT) technology at an enterprise. The implementation of GPS technology is proposed for the control and monitoring of personnel quality.
- Methods and tools for increasing the efficiency of performing bit-level parallel computing in systems-on-a-chip.
“Methods and tools for mapping streaming algorithms onto configurable computers,” state registration № 0117U005087 (2017 – 2022, Project Lead: Prof. A. M. Serhienko).
- Software and hardware tools, algorithms, and programs for the genetic synthesis of pipelined devices for configurable computers.
- Method and tool for genetic programming of pipelined devices.
- Methods for the synthesis of advanced digital filters.
- Method and tools for detecting keypoints in an image.
- Method and tool for hardware modeling of acoustic processes.
- Architecture and functional models of stack processors along with programming tools.
- Improved devices for computing elementary functions, file compression/decompression, and stereo radio signal generation.
- Algorithm and program for the genetic synthesis of pipelined devices for configurable computers and a method for hardware-software design of XML document parsers in such computers.
“Organization of computations in scalable distributed computing systems and networks,” state registration № 0113U002314 (2013 – 2022, Project Lead: Prof. V. P. Symonenko).
- Developed a system for data collection and processing in distributed computing systems, including GRID and Cloud. A software system for static scheduling in FPGA-based computing systems.
- A methodology is proposed for increasing the degree of information security of a computing node based on the audit mode. Information security methods for computing systems to combat spam have been investigated.
- Developed an adaptive software system for protecting the file system from external interference. A comprehensive system. A monitoring system for computing node workload for GRID and Cloud scheduling systems.
- Developed a method for fake news recognition using Natural Language Processing technologies and the Levenshtein algorithm. The results of testing and a comparative analysis of the proposed method are presented. The high rate of recognized fake news has been experimentally verified.
- Dynamic irrigation management system for agricultural crops; video surveillance system for residential premises; monitoring system for human vital signs; production scheduling system; information and advisory system for media resource search.
Development archive
