Vadym SlyusarCentral Scientific Research Insitute of Armaments and Military Equipment of Armed Forces of Ukraine · Group of chief research scientists for scientific research management
Vadym Slyusar
Doctor of Science, Professor
About
379
Publications
114,644
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
518
Citations
Introduction
Vadym Slyusar has more as 35 years of research experience in the areas of radar systems, smart antennas for wireless communications and digital beamforming. He earned Ph.D. in 1992, Doctor of Sciences in 2000, Professor in 2005, Honored Scientist and Technician of Ukraine in 2008 and has 68 patents, 805 publications in these areas ( 63 publications was included in SCOPUS (Author ID 7004240035) and 275 publications included in ORCID 0000-0002-2912-3149). He is authority in digital signals processing. Since 2009 - a Member of Editorial Board of Journal "Radioelectronics and Communications Systems".
Scientific advisor of 15 PhDs (candidates of technics) and 2 Doctors of Science.
Google Scholar Citations: wSegaWsAAAAJ
www.slyusar.kiev.ua
Additional affiliations
November 2019 - December 2022
NATO Science and Technology Organization
Position
- Tecnical Member
Description
- AVT-334 RTG Membership , Report co-writer and reviewer
September 2018 - December 2020
NATO Science and Technology Organization
Position
- Tecnical Member
Description
- AVT-290 RTG Membership , Report co-writer and reviewer
Education
August 1981 - July 1985
Orenburg High Military School of Air Defense
Field of study
- Radar technics, digital signal processing, digital antenna arrays
Publications
Publications (379)
In 2007 60 years from the moment of the publication of the first theoretical work which have proved fundamental limits of achievable parameter!, so-called Electrically Small Antennas (ESA) are executed. On the eve of a anniversary of the given event -it is high time to sum up the 60-years way gone by theorists and developers of aerials in understan...
The purpose of this work is to consider the UAV-communication channel multiplexing in the case of nonorthogonal frequency-division multiplexing (N-OFDM) . Keywords: UAV, orthogonal frequency-division multiplexing (OFDM), nonorthogonal frequency-division multiplexing (N-OFDM), fast Fourier transform (FFT). The majority of known approaches to the imp...
The perspective direction in the development of dismounted soldiers systems is use exoskeletons. Need be note the possible application of exoskeleton elements in a role of antenna for transfer and reception of radio signals. In case of metallic exoskeleton, the full exoskeleton construction or his separate segments can carry out functions of the an...
A new model of the NDPP are considered.
In this abstract the tensor-matrix theory of Artificial Intelligency is considered and evolved.
The article proposes a new type of broadband patch antennas based on Micro QR codes with integrated metamaterial cell as a marker. In this research, several design options were considered that used the variation of the metamaterial cell parameters. For their analysis, the numerical modeling methods of the Ansys EM Suite program
were used due to the...
A new approach to the reverse dictionary task is put forward that is based on segmentation of a text into subwords. We use the physical texts and split the words therein into subwords that include morphemes, separate characters, and their combinations. In contrast to the wordwise segmentation, the subword vocabulary is finite that excludes the prob...
This is the video record of the presentation "An Approach to the Reverse Dictionary Task Based on Automatic Smart Subword Segmentation of Ukrainian Physical Texts" that was presented on the/ Modern Machine Learning Technologies and Data Science Workshop (MoMLeT&DS 2024).
A new approach to the reverse dictionary task is put forward that is based on...
The article addresses hardware-software and other key aspects of the artificial intelligence development strategy for mobile technologies. The proposed components of the strategy include a series of approaches to address issues related to the development and deployment of large language models on mobile devices, as well as suggestions for improving...
Classifying metals based on microstructural characteristics is crucial in materials science and metallurgy, as it unveils mechanical properties, processing history, and
potential applications of the material. Traditionally, this complex process required significant expertise, relying on a deep understanding of the relevant field. However, recent ad...
The method of joint processing of pulses and OFDM (N-OFDM) signals is proposed. The corresponding analytical relations for the lower Cramer-Rao boundary on the dispersion of OFDM (N-OFDM) signals amplitude ratings in the presence of sources of pulsed radiation are obtained. Using mathematical modeling properties and limitations of the demodulation...
The article highlights the importance of text segmentation in the field of natural language processing (NLP), especially in light of the development of large language models such as GPT-4. It discusses the use of specialized segmentation neural networks for various tasks, such as processing passport data and other documents, and points out the poss...
Informational reference on the scientific activity of the doctor of technical sciences, professor, Honored Worker of Science and Technology of Ukraine Vadym Slyusar
Based on the Gemini Ultra test results, can be say that GPT-4v is definitely better than Gemini Ultra!!!
The hype around the absolute benefits of Gemini Ultra is just a purely business PR campaign that mainly misleads users and tries to pass off wishful thinking.
The multimodal capabilities of the Gemini Ultra v 1.0 are actually very limited and do...
The full text can be read on the link: https://rdcu.be/dx60e.
In the article, the substantiated proposals for the use of YOLO family neural networks to detect the underwater undetonated munitions are proposed. At the same time, the YOLO3, YOLO4 and YOLO5 neural networks previously trained on the MS COCO dataset are used. The retraining of YOLO3 an...
In this presentation, the idea of creating a personal assistant to a soldier was proposed.
The technical aspects of implementing such an idea based on a mixture of experts (MoE) LLM/SLM are discussed in combination with a description of typical use cases and functionality of a soldier's personal assistant.
The research article addresses the problem of improving the efficiency of image transmission from unmanned aerial vehicles by using state-of-the-art image processing techniques along with advanced machine learning algorithms. The research introduces a new methodology that involves the use of split autoencoders to facilitate the compression of image...
Seasonal changes in the environment and landscape require a high level of flexibility and adaptability from Object Detection systems. Neural networks that do not take these changes into account may have reduced accuracy and effectiveness in different seasonal conditions. Therefore, including seasonal variations in training and testing data sets is...
Seasonal changes in the environment and landscape require a high level of flexibility and adaptability from Object Detection systems. Neural networks that do not take these changes into account may have reduced accuracy and effectiveness in different seasonal conditions. Therefore, including seasonal variations in training and testing data sets is...
This study provides an image enhancement method based on a segmented
autoencoder for marine unmanned platforms. It optimally compresses and reconstructs data using an encoder on the vehicle and a decoder on the base, providing efficient real-time visualization in challenging marine environments
The presentation highlights the application of advanced artificial intelligence technologies, particularly large language models (LLM), in the military sector. It thoroughly examines the potential of LLM for analyzing large volumes of textual data, automating intelligence activities, preparing reports, and ensuring prompt communicative interaction...
This paper is devoted to the analysis of the specific focuses, directions, and
peculiarities of the Strategy of Artificial Intelligence (AI) Development in Ukraine. The main paper’s components are an analysis of the current state of the justification, development, and governmental approval of the National Strategy of AI in Ukraine; key elements and...
This paper is devoted to the analysis of the specific focuses, directions, and peculiarities of the Strategy of Artificial Intelligence (AI) Development in Ukraine. The main paper's components are an analysis of the current state of the justification, development, and governmental approval of the National Strategy of AI in Ukraine; key elements and...
This presentation is devoted to the analysis of the specific focuses, directions, and peculiarities of the Strategy of Artificial Intelligence (AI) Development in Ukraine. The main paper’s components are an analysis of the current state of the justification, development, and governmental approval of the National Strategy of AI in Ukraine; key eleme...
Traditional Data Farming (DF) consists of a toolbox of established analysis techniques that are available for an analyst-led study of a particular military operation in support of a single decision-maker. Multi-Dimensional Data Farming (MDDF) is a new and automated analytical process that provides accelerated support to military decision-making at...
This paper proposes an optimized approach for image transmission from Unmanned Aerial Vehicles using split autoencoders and super-resolution techniques. The autoencoder, partitioned between the UAV and receiver, compresses and reconstructs images, saving energy and bandwidth. Super-resolution methods
are applied at the receiver to counter quality l...
This paper is devoted to the analysis of the specific focuses, directions, and peculiarities of the Strategy of Artificial Intelligence Development in Ukraine.
This is video record of the presentation "Implementation Peculiarities of the Strategy of Artificial Intelligence Development in Ukraine", September 2023. Conference: ХI International Scientific Conference «Information-Management Systems and Technologies» 21th – 23th September, 2023. Odessa, Ukraine
Various options for the 3D design of a metamaterial cell based on two-ribbon split rings have been proposed and studied. When synthesizing the geometry of the outer tape, a quasi-fractal approach was used, which made it possible to form periodic sequences of depressions (slots) in its end surfaces. Their alternation from both ends of the tape was c...
The purpose of this presentation is to use for text segmentation neural networks, which by their architecture are used for image processing, and, in particular, image segmentation.
What is it for? Now there is a trend towards multimodal neural networks, when the same neural network complex or hyperneural network will perform various functions, with...
In the article the authors present improved Dijkstra's method for the military communication system by using the incidence matrix of nodes and communication lines instead of the incidence matrix of nodes. The improved method was adapted to the model of the military communication system, which is formalized by a block matrix of the incidence of node...
The presentation proposes a strategy for creating a System of Systems Standards (S3) for quantum technologies, particularly in the field of quantum communications. It highlights the main advantages of quantum communications, with a focus on their resilience in the presence of electromagnetic interference, secure communication capabilities via quant...
The proposed vector-matrix form based on wedge-shaped matrices assembled with symmetric elements allows to compactly present the responses of a two-channel I/Q-demodulator and simplifies the process of its synthesis. Suppression of frequency response side lobes achieved by this method makes it suitable for demodulation of N-OFDM signals, which will...
The report discusses image segmentation techniques of water meter dials, in particular deep learning models of neural networks. Evaluating the effectiveness of these methods takes into account their accuracy, robustness, and computational efficiency. The main emphasis in the conducted research is on the consideration of the simplest variants of neu...
The article proposes a new approach to solving text classification tasks using pre-trained convolutional neural networks for image processing. A comparison of the training results of different neural network architectures was performed for the dataset of text reviews about the Tesla electric car. The obtained results allowed us to conclude that amo...
The monograph examines the prerequisites and scientific foundations for creation of the Strategy for Artificial Intelligence Development in Ukraine as well as means and ways of its effective implementation. For specialists, postgraduate, and graduate students in the field of artificial intelligence, information technologies, philosophy, state forma...
The article proposes a new type of metamaterial cells (metacells) based on double Moebius strip. In this research, several design options have been considered which applied the variation of the metamaterial cells design and parameters. For their analysis, the Numerical modeling methods in Ansys EM Suite program were used due to the complexity of de...
The purpose of the presentation is to demonstration the electromagnetic parameters of 3D SRR structures based on two-ribbon split rings, in which the constituent elements are modeled on the basis of quasi-fractals.
The report discusses options for solving the image segmentation problem of displaying digital indicators of analogue water or gas meters using neural networks. The results of a comparative analysis of the application of various implementation options for neural networks based on PSP, U-Net, and U-Net2 are presented. The Water Meters Dataset, which...
The presentation proposes a new approach to solving text classification tasks using pre-trained convolutional neural networks for image processing. A comparison of the training results of different neural network architectures was performed for the dataset of text reviews about the Tesla electric car. The obtained results allowed us to conclude tha...
The results of studying the accuracy of various architectures of U-Net neural networks with an attention mechanism in solving the problem of image segmentation of analogue counter dials are considered. It is concluded that there is no effect of increasing accuracy due to the integration of one- and two-channel attention schemes, while the achieved...
The results of studying the accuracy of various architectures of U-Net neural networks with an attention mechanism in solving the problem of image segmentation of analogue counter dials are considered. It is concluded that there is no effect of increasing accuracy due to the integration of one-and two-channel attention schemes, while the achieved s...
The report discusses options for solving the image segmentation problem of displaying digital indicators of analogue water or gas meters using neural networks. The results of a comparative analysis of the application of various implementation options for neural networks based on PSP, U-Net, and U-Net2 are presented. The Water Meters Dataset, which...
This presentation proposes a technique for forming the response of two-channel demodulators. At the same time, a string vector is used, the elements of which are the weighting coefficients of the demodulator, which are used to form the quadrature components, as well as a wedge-shaped matrix with an inversely symmetrical sequence of elements. The am...
Various PSP and U-Net architectures for forest segmentation in remote sensing pictures have been proposed and investigated. The main improvements of the proposed architectures are based on using the BathNormalization layers, replacing MaxPool2D layers with AveragePooling2D, changing Conv2DTranspose to UpSampling2D blocks, etc. For the training of n...
This document proposes a technique forforming the response of two-channel demodulators. At the same time, a string vector is used, the elements of which are the weighting coefficients of the demodulator, which are used to form the quadrature components, as well as a wedge-shaped matrix with an inversely symmetrical sequence of elements. The amplitu...
The paper proposes variants of 3D models of split ring resonators (SRR) based on ring structures as unit cells of doubly negative (DNG) metamaterials. Evaluation of the efficiency of the synthesized 3D SRR variants is based on determining their relative bandwidth, for which the real part of permittivity and permeability are negative. To analyze the...
Combat models based on Lanchester equations describe a homogeneous blue force against a red force.
However, in reality, a conflict involves multiple factions operating in multiple domains. For example, in the Vietnam War, US force fought alongside the South Vietnamese against the North Vietnamese. With this viewpoint, we provide a general model of...
The Ukrainian Strategy of Artificial Intelligence Development is designed with the aim of
achieving sustainable development and adopting artificial intelligence. It will effectively improve competitiveness, ensure a high rank of Ukraine in global civilizational processes, and raise living standards. As long as the scientific potential of the countr...
This article is dedicated on 100 years since the birth of V. A. Varyukhin, the founder of the theory of multichannel analysis, and creator of the scientific school on digital antenna arrays.
This presentation includes a few approaches to freedom of will of artificial intelligence in the context of autonomous weapons systems
The article proposes ways to solve the problem of structural synthesis of neural networks architectures; based on non-traditional approaches to their analytical formalization and application of new ones data processing operations. The example of the CIFAR10 dataset proves the possibility of improving the accuracy of the solution image classificatio...
The article proposes ways to solve the problem of structural synthesis of neural networks architectures; based on non-traditional approaches to their analytical formalization and application of new ones data processing operations. The example of the CIFAR10 dataset proves the possibility of improving the accuracy of the solution image classificatio...
The family of new antenna designs based on the double Moebius strip principle is proposed in this article. Three possible options of the power supply port location had been reviewed in this respect: the cut external strip with the signal port located directly in the cut area is used; the continuous rings without discontinuities are applied, with th...
5G communication systems ensure high data rate, low latency, network reliability, energy efficiency and high throughput that require new and very efficient antenna designs. In this paper, we proposed a simple and very effective antenna with a centre frequency 28GHz designed on an RF4 substrate of 1. a 6mm thickness. The performance characteristics...
![Fig. 3. YOLOv5 network architecture [15]: CSP (Cross Stage Partial...](profile/Vadym-Slyusar/publication/359084024/figure/fig1/AS:1146512373497857@1650361259373/YOLOv5-network-architecture-15-CSP-Cross-Stage-Partial-Network-multi-stage-partial_Q320.jpg)
This paper considers a model of object detection on aerial photographs and video using a neural network in
unmanned aerial systems. The development of artificial intelligence and computer vision systems for unmanned
systems (drones, robots) requires the improvement of models for detecting and recognizing objects in images and video
streams. The res...
This paper considers a model of object detection on aerial photographs and video using a neural network in unmanned aerial systems. The development of artificial intelligence and computer vision systems for unmanned systems (drones, robots) requires the improvement of models for detecting and recognizing objects in images and video streams. The res...
![Fig. 3. YOLOv5 network architecture [15]: CSP (Cross Stage Partial...](profile/Vasil-Melkin/publication/359985038/figure/fig1/AS:1145202597208064@1650048984191/YOLOv5-network-architecture-15-CSP-Cross-Stage-Partial-Network-multi-stage-partial_Q320.jpg)
This paper considers a model of object detection on aerial photographs and video using a neural network in unmanned aerial systems. The development of artificial intelligence and computer vision systems for unmanned systems (drones, robots) requires the improvement of models for detecting and recognizing objects in images and video streams. The res...
В работе предложена совокупность новых тензорноматричных операций для обработки данных в нейронных сетях.
In this presentation, the demonstration of technology for underwater object detection using YOLO4 neural network is considered.
This paper considers a model of the neural network for semantically segmenting the images of monitored objects on aerial photographs. Unmanned aerial vehicles monitor objects by analyzing (processing) aerial photographs and video streams. The results of aerial photography are processed by the operator in a manual mode; however, there are objective...
"Lions" in the Neural Networks "Zoo" was presented in this papaer in the context of CiFAR10 dataset.
The proposed architecture based on the Xception had 97.2 % BalancedRecall value.
Multimodal quasi-fractal neural networks were proposed in this paper.
"Lions" in the Neural Networks "Zoo" was presented in this presentation in the context of CiFAR10 dataset. The proposed architecture based on the Xception had 97.2 % BalancedRecall value.
Multimodal quasi-fractal neural networks were proposed in this presentation.
The ornamental design for a double-negative metamaterial unit cell, as shown and described.
In this article, the theory of tensor matrices U-net and GANs was proposed (in Russian, but all mathematical formulas will be understandable for any language)
The paper proposes a new type of broadband patch antennas based on MicroQR codes. To synthesize them, the word “antenna” was used. In this research, several design options were considered that used the variation of the antenna's feeder parameters. For their analysis, the Numerical modeling methods were used due to the complexity of describing the i...
The presentation proposes a new type of broadband patch antennas based on MicroQR codes. To synthesize them, the word “antenna” was used. In this research, several design options were considered that used the variation of the antenna's feeder parameters. For their analysis, the Numerical modeling methods were used due to the complexity of describin...
Although blockchain technology (BT) was originally developed for digital currency, different potential uses cases are discussed in the military communities. Blockchain can be defined as a digital information system, which stores data in an encrypted, distributed format on the combat field in different domains. In this regard, BT can be considered a...
Situational awareness is a decisive aspect of survivability and effective decision making for combat vehicle crews and soldiers in the battlefield, and Augmented Reality (AR) technologies have the potential to take it to the next level. The introduction of optical, IR and acoustic sensors for monitoring the battlefield, and in particular the integr...
This presentation considers a general view of the use of XR for training and education in the military area.
This article explores the concept of battlefield virtualization based on a combination of androids and virtual reality.
Detection and recognition of objects in images is the main problem to be solved by computer vision systems. As part of solving
this problem, the model of object recognition in aerial photographs taken from unmanned aerial vehicles has been improved. A study
of object recognition in aerial photographs using deep convolutional neural networks has bee...
In this paper, the pandemic approach to performance assessment of combined armor protection systems was proposed (in Russian language)
This paper investigates the properties of patch genetics antennas.
IVth International scientific-practical conference dedicated to the 50th anniversary of the Department of Information Systems and Technologies (October 21-22, 2021)
In this paper, the tensor-matrix model of LeNet5 is proposed.
In this paper, Neural Networks as Smart Home equipment was considered.
This paper proposes an approach to technologies foresight based on neural networks.
This paper considers the example of Neural Networks developments on Java Language
This presentation is devoted to the tensor-matrix theory of the traditional approach to multichannel signal parameters measurements based analytical model of signals and the tensor-matrix theory of Neural Networks for Mechanical Measurements.
This presents the basic concepts of matrix operations that can be used for ultrasonic, sonar, radar system...
The utilization of AR in Military Operations for planning, engagement and training has two different aspects: AR is the advanced means of Situation Awareness, and AR is a communication bridge between Humans and Artificial Intelligence (AI), which use for Decision Making. These aspects are considered in this presentation.
This presentation proposes a new type of printed antennas, called chaotic due to the peculiar geometry of their outline, as possible applicants for antennas that meet the above requirements. The synthesis of the topology of chaotic antennas was carried out on the basis of the application of the ant colony optimization (ACO) algorithm for solving th...
The paper proposes a new type of multiband broadband printed antennas with a chaotic outline. To synthesize them, the topologies for solving the traveling salesman problem based on the ant colony optimization algorithm are used. Removal of topological restrictions on the trajectory of traversing the matrix of cities made it possible to switch from...
This is the answer to the reviewer of MDPI journal regarding article
https://www.researchgate.net/publication/352405011_The_Concept_of_Networked_Distributed_Engine_Control_System_of_Future_Air_Vehicles
Questions
Questions (58)
Based on the my personal Gemini Ultra test results, I can say that GPT-4v is definitely better than Gemini Ultra!!!
Presentation Comparison of Gemini Ultra and GPT-4v
The hype around the absolute benefits of Gemini Ultra is just a purely business PR campaign that mainly misleads users and tries to pass off wishful thinking. The multimodal capabilities of the Gemini Ultra v 1.0 are actually very limited and do not meet the requirements. At the same time, ideally, it is necessary to use these different LLMs, supplementing the gaps of one with the advantages of the other.
Please share your experience regarding this.
