Search

Search Results

• DEEP LEARNING METHODS FOR NATURAL LANGUAGE TEXT PROCESSING

  V.V. Kureichik, S.I. Rodzin, V.V. Bova

  2022-05-26

  Abstract ▼

  The analysis of approaches based on deep learning (DL) to natural language processing
  (NLP) tasks is presented. The study covers various NLP tasks implemented using artificial neural
  networks (ANNs), convolutional neural networks (CNNs) and recurrent neural networks (RNNs).
  These architectures allow solving a wide range of natural language processing tasks that previously
  could not be effectively solved: sentence modeling, semantic role labeling, named entity
  recognition, answers to questions, text categorization, machine translation. Along with the advantages
  of using CNN to solve NLP problems, there are problems associated with a large number
  of variable network parameters and the choice of its architecture. We propose an evolutionary
  algorithm for optimizing the architecture of convolutional neural networks. The algorithm initializes
  a random population of a small number of agents (no more than 5) and uses the fitness function
  to get estimates of each agent in the population. Then a tournament selection is carried out
  between all agents and a crossover operator is applied between the selected agents. The algorithm
  has such an advantage as the small size of the network population, it uses several types of CNN
  layers: convolutional layer, maximum pooling layer (subdiscretization), medium pooling layer and
  fully connected layer. The algorithm was tested on a local computer with an ASUS Cerberus Ge-
  Force ® GTX 1050 Ti OC Edition 4 GB GDDR5, 8 GB of RAM and an Intel(R) Core(TM) i5-4670
  processor. The experimental results showed that the proposed neuroevolutionary approach is able
  to quickly find an optimized CNN architecture for a given data set with an acceptable accuracy
  value. It took about 1 hour to complete the algorithm execution. The popular TensorFlow framework
  was used to create and train CNN. To evaluate the algorithm, public datasets were used:
  MNIST and MNIST-RB. The kits contained black-and-white images of handwritten letters and
  numbers with 50,000 training samples and 10,000 test samples.

• CLASSIFIER OF IMAGES OF AGRICULTURAL CROPS SEEDS USING A CONVOLUTION NEURAL NETWORK

  V. A. Derkachev, V. V. Bakhchevnikov, A. N. Bakumenko

  2020-11-22

  Abstract ▼

  This article discusses the creation of a convolutional neural network architecture that classifies
  images of crops (in particular wheat) for subsequent use in an optical seed separator (photo
  separator). Interest in the design of neural networks for classifying images has recently increased
  significantly, which is associated both with the development of the theory of deep neural networks
  and the increased computing power of desktop computers, as well as the transfer of computing to
  graphic processors. The aim of the article is to develop the architecture of a neural network that
  allows the separation of the input flow of wheat seeds into two classes: “good” seeds and “bad”
  (with defects in shape and color) seeds. The architecture of the resulting neural network is convolutional,
  because, unlike a fully connected one, this class of neural networks is within certain limits
  immune to changes in the scale and angle of rotation of objects in the input data. In the work,
  for the formation of training, validation and test samples, seed images obtained using a household
  camera were used, which negatively affected the results of training and testing the neural network
  regarding the possible result of application in a real photo separator. The architecture of the developed
  neural network is preliminarily optimized for use on FPGAs, however, in the considered
  case, the transition from the values of weighting factors from the data type from a floating point to
  an integer type has not been made, which can lead to a decrease in the accuracy of the neural
  network, while significantly reducing the amount of resources FPGA. Application of the proposed
  architecture allows one to obtain a fairly accurate estimate of classified wheat seeds from verification
  and test data sets.

• USING FAST PROTOTYPING FACILITIES FOR IMPLEMENTATION OF A CONVOLUTION NEURAL NETWORK ON A FPGA

  V. V. Bakhchevnikov , V. A. Derkachev , A. N. Bakumenko

  2020-10-11

  Abstract ▼

  Research in the field of artificial intelligence is carried out with increasing interest every
  year. The fields of application of artificial intelligence are quite extensive: automation, analysis of
  a large amount of data, smart home technology, machine vision, etc. Artificial intelligence technologies
  are based on the use of artificial neural networks, which are based on the principles of
  the animal nervous system. In this case, the actual issue is the implementation of artificial neural
  networks on various software and hardware platforms: programmable logic integrated circuits of
  the FPGA type (Field Programmable Gate Array), on special purpose integrated circuits (Application-
  Specific Integrated Circuit, ASIC), GPU, CPU etc. FPGA performs best in low-power mobile
  systems. ASIC demonstrates the highest performance at a fairly high development cost.
  The problem of rapid prototyping of projects based on the use of artificial neural networks for
  FPGAs using conventional methods (using HDL languages, HDL encoders, graphic programming)
  is that either such a project is complex and time-consuming to debug (HDL languages), or
  the resulting code is not optimal (HDL encoders), or the duration of the project development and
  the complexity of reconfiguring the neural network (graphical programming) are high. Therefore,
  in the framework of this work, an effective method for designing fully connected and convolutional
  neural networks for their implementation on FPGAs using the Xilinx System Generator for DSP
  and Matlab / Simulink package is considered. Artificial neural networks generated in this way are
  easily reconfigurable and allow solving the following problems: image recognition, optimal filtering
  (for example, for problems of subsurface radar).

• DESIGNING MLP AND CNN NEURAL NETWORK MODULES ON FPGA FOR IMAGE CLASSIFICATION TASKS

  E. V. Melnik , D.Е. Blokh , А.I. Bezmeltsev , V.S. Panishchev , S.N. Poltoratsky

  214-229

  2025-11-10

  Abstract ▼

  Relevance. The development of machine learning methods and neural network architectures, as well as their spread into various industrial sectors, determine the relevance of solving problems related to their hardware implementation. The use of programmable logic integrated circuits in this area will increase data processing speed and the adaptability of the implemented algorithms. However, designing neural network architectures on programmable logic integrated circuits is associated with a number of methodological and technical difficulties, including the optimization of parallel computing, hardware resource management, and ensuring operation under conditions of limited computing resources. The purpose of this work is to analyze and compare two neural network architectures, the multilayer perceptron (MLP) and the convolutional neural network (CNN), in the context of their hardware implementation on programmable logic integrated circuits (PLICs). Particular attention is paid to the trade-off between classification accuracy and the efficient use of limited FPGA hardware resources. Research methods.
  To achieve the goal, two modules were developed and simulated on a Virtex 7 FPGA, a perceptron and a convolutional module. The MNIST dataset, reduced to 20×20 pixels, was used. The implementation included quantizing parameters to a fixed 16:16 format, optimizing hyperparameters, using tabular computations for nonlinear functions, and evaluating FPGA resource usage. Results and discussions.
  MLP achieved 93% accuracy using 11% of logic elements, while CNN achieved 98% accuracy but required significantly more resources. The use of internal buffers to store intermediate data in CNN resulted in exceeding the allowable resources. The forced transition to external memory increased delays and the number of I/O ports. Conclusions. The study showed that the choice of architecture depends on priorities: CNN provides better accuracy but is less resource-efficient. For embedded systems with memory and power consumption constraints, a simplified MLP implementation is preferable. The main problems remain the lack of internal memory and the high resource intensity of operations, which requires further research in the field of hardware optimization and adaptive computation control

• ANALYSIS OF ARTIFICIAL INTELLIGENCE METHODS APPLIED TO SOLVING PSYCHIATRY PROBLEMS

  E.S. Podoplelova

  2022-05-26

  Abstract ▼

  The use of artificial intelligence methods in the field of medicine has become widespread,
  helping to diagnose, analyze and make recommendations for treatment. Psychiatry is a branch of
  medicine that studies mental disorders, methods for their diagnosis and treatment. Her range of
  tasks includes not only diagnosis and treatment, but also observation, monitoring and subsequent
  rehabilitation of patients. This subject area has significant problems, such as objectivity, inconsistency
  in the diagnosis, the complexity of the classification of diseases, and the unpredictability
  of the course of the disease. With a number of these problems, the use of machine learning methods
  and artificial intelligence algorithms helps to cope. This paper is devoted to a review of research
  on artificial intelligence methods used to solve problems in the field of psychiatry.
  The relevance of the topic is due to the high need for improvements in this subject area. Specific
  issues are presented in this article. Among them, the main directions were identified: data deidentification,
  classification of symptom severity, accuracy of condition prediction. To solve them,
  the authors used such methods as latent semantic analysis for natural language processing, classification
  methods, convolutional neural networks for prediction, and cognitive modeling. Separately,
  the effectiveness of hybrid systems, including the implementation of several machine learning
  methods at once, is noted. The aim of the study was to highlight the main directions of development
  of research in the scientific community, which demonstrate the successful integration of artificial intelligence into psychiatry, as well as to compare them with each other according to the
  obtained estimates of the accuracy of the models. Which, in turn, implies the analysis and analysis
  of specific algorithms, their performance for specific tasks

• RESEARCH OF APPLICABILITY LIMITATIONS FOR ELBRUS MICROPROCESSORS FOR SOLVING TASKS OF TECHNICAL VISION

  К. А. Suminov, N. А. Bocharov

  2022-04-21

  Abstract ▼

  One of the key areas in the artificial intelligence is technical vision. For resource-intensive
  tasks of technical vision high-performance, computing systems are created with use of specialized
  accelerators. The use of such accelerators is necessary due to the inability of general-purpose
  microprocessors (GPM) to solve such problems in a given time due to a high computational load.
  However, the microprocessors of Elbrus series are successfully used to solve technical vision
  problems in both server and on-board modes, and the appearance of the sixth-generation Elbrus
  microprocessors should further improve performance on such tasks. Due to the high cost, greater
  complexity and limitations in the use of systems with specialized accelerators, the question arises
  of determining the conditions under which, it is sufficient to use CPU’s to solve the tasks of technical
  vision, for example, with the microprocessors of the Elbrus series without special accelerators.
  One of the most resource-intensive tasks in the field of technical vision are detection and
  classification of objects. For the detection of objects one of the popular methods is the Viola-Jones
  method. Convolutional neural networks are usually used to solve the classification problem.Mathematical models of computations have been developed for VGG16 and VGG19 neural networks
  in relation to the actual microprocessors of the Elbrus series. Using the developed models,
  the theoretical sufficiency of the performance of Elbrus microprocessors for technical vision tasks
  is substantiated. Also, based on these methods, programs for modeling detection and classifications
  objects in the image and video stream have been developed. The programs are written in
  C++ using the OpenCV library, OPO Elbrus, the GNS Platform library and the ImageNet competition
  database. Using the implemented programs, comparative testing was carried out on a number
  of high-performance computing systems with Elbrus and Intel CPU’s and NVidia video card.
  Based on the results obtained, it is shown that the Elbrus-8S is sufficient to solve the problem of
  searching for objects in the image for input resolutions up to 1920 x 1080, where the processing
  speed of the video stream is more than 20 frames per second.

• ALGORITHM FOR PRE-PROCESSING VIDEO IMAGES TO INCREASE THE ACCURACY OF SMALL OBJECT DETECTION

  V.V. Kovalev, N.E. Sergeev

  2021-12-24

  Abstract ▼

  Recognition of certain patterns in video images captured by a camera is carried out using
  training methods based on convolutional neural networks. The larger the number of images with
  multiple features and the more diverse the training sample of video images, the better the convolutional
  neural networks extract features from the sequence of video images that were not included in
  the training sample. This is a consequence of increasing the accuracy of detecting visual images on
  video images containing features of target images. However, there are limitations in improving the
  detection performance when the size of the image to be detected is much smaller than the background
  area, or when the image is described with little information. To solve problems of this kind, the authors
  of the article have developed an algorithm for the spatio-temporal integration of information
  about the movement of dynamic images. The algorithm processes a fixed number of video images at
  certain points in time and extracts new independent signs of motion of dynamic images based on
  space-time processing of video images. Further, it combines new local motion features with the original
  video image features. This allows you to add a motion feature of dynamic images while preserving
  the original image features that describe static images. Areas of the video image that characterize
  the motion feature are displayed in a «color» cluster. The use of pre-processing is aimed at improving the accuracy of pattern detection, provided there are dynamic visual images on a static background.
  If the camera is in scan mode, a static background can be provided with a video stabilizer.
  Experimentally, estimates of integral criteria for the accuracy of detection neural network algorithms
  have been obtained, showing an increase in the accuracy of detecting visual images using
  the algorithm for spatial-temporal integration of motion information.

• IMPLEMENTATION OF CONVENTIONAL NEURAL NETWORKS ON EMBEDDED DEVICES WITH A LIMITED COMPUTING RESOURCE

  V.V. Kovalev, N.E. Sergeev

  2022-01-31

  Abstract ▼

  Large amounts of video data captured by sensor sensors in various spectral ranges, the significant
  size of convolutional neural network architectures create problems with the implementation of
  neural network algorithms on peripheral devices due to significant limitations of computing resources
  on embedded computing devices. The article discusses the use of algorithms for automatic search and
  pattern recognition based on machine learning methods, implemented on embedded devices with a
  computing resource Graphics Processing Unit. Detection convolutional neural networks «You Only
  Look Once V3» and «You Only Look Once V3-Tiny» are used as a search and pattern recognition algorithm,
  which are implemented on embedded computing devices of the NVIDIA Jetson line, located in
  different price ranges and with different computing resources ... Also, in the work, the estimates ofalgorithms on embedded devices are experimentally calculated for such indicators as power consumption,
  forward passage time of a convolutional neural network, and detection accuracy.
  On the basis of solutions implemented, both at the hardware level and in software, presented by
  NVIDIA, it becomes possible to use deep neural network algorithms based on the convolution
  operation in real time. Computational optimization methods offered by NVIDIA are considered.
  Experimental studies of the influence of computations with reduced accuracy on the speed and
  accuracy of object detection in images of the investigated architectures of convolutional neural
  networks, which were previously trained on a sample of images consisting of the PASCAL VOC
  2007 and PASCAL VOC 2012 datasets, have been carried out.

• APPLICATION OF CONVOLUTIONAL NEURAL NETWORKS FOR TECHNICAL OBJECT RECOGNITION IN THE INTERESTS OF RADIO MONITORING

  D. V. Shumkov, I.V. Titkov, P.А. Gulevich

  2025-04-27

  Abstract ▼

  The article examines the possibility of using convolutional neural networks for technical object
  recognition in the context of radio monitoring. The focus is on the development and optimization of algorithms
  for processing radar signals using deep neural networks. Studies have shown that the use of CNN
  can significantly improve the classification accuracy of radio signals compared to traditional processingmethods. The developed approach is based on the extraction of hierarchical features from spectral images
  of radio signals and their subsequent classification using a trained neural network. The paper presents the
  results of experimental studies conducted on a dataset of more than 10,000 samples of radio signals of
  various types. It is shown that the proposed technique ensures recognition accuracy of up to 94% when
  working with noisy signals and the probability of a false alarm is no more than 0.05. Special attention is
  paid to the choice of neural network architecture for the specifics of the radio monitoring task. The options
  for converting radio signals into a spectral image for real-time processing were also considered in
  detail. Data preprocessing methods have been developed, including amplitude normalization, frequency
  correction, and interference elimination. The results of the study can be used in radio broadcast control
  systems and to ensure electromagnetic compatibility of electronic devices. The results obtained demonstrate
  the prospects of using CNN in the tasks of technical recognition of radio monitoring objects and
  open new opportunities for the development of intelligent radar information processing methods. Promising
  areas of further research include the development of adaptive neural network training methods in a
  changing radio environment and the creation of hybrid systems combining traditional signal processing
  methods with modern neural network algorithms

• APPLICATION OF COMPUTER VISION TECHNOLOGIES IN VISUAL INFORMATION PROCESSING SYSTEMS

  О.B. Lebedev , R.I. Cherkasov

  254-276

  2025-11-10

  Abstract ▼

  This paper considers the application of artificial intelligence technologies, in particular computer vision, in visual information processing systems. A comprehensive analysis of neural network approaches to solving computer vision problems is carried out, including systematization of key types of problems: image classification, object detection and semantic segmentation. The architectural principles of convolutional neural networks are studied in detail with an emphasis on the mechanisms of spatial feature extraction through convolutional layers, optimization of data representation through pooling operations and feature transformation in fully connected layers. Particular attention is paid to the evolution of object detection methods, where the problem of model selection is considered as an extension of classification due to the integration of spatial coordinate regression, and an assessment of the effectiveness of detectors is carried out based on the IoU, Precision, Recall and F1-score metrics, demonstrating a fundamental trade-off between localization accuracy and processing speed. The YOLOv7 algorithm is presented as an optimal solution for real-time systems. Its architecture is based on splitting the input image into a grid of S×S cells with direct prediction of the bounding box parameters (center coordinates, width, height) and class probabilities for each cell, as well as the use of specialized layers (SPP, PANet) for multi-scale feature aggregation. The structure of the neural network confirms the effectiveness of the approach used, which ensures high performance without critically reducing accuracy in strategically important applications of video surveillance, autonomous systems, and augmented reality. A comparative study of one-stage and two-stage detectors was conducted with an assessment of their performance by key metrics. Particular attention is paid to the practical aspects of using computer vision technologies in real visual information processing systems.

• A NEW REPRODUCIBILITY METRIC FOR COMPARING TIME SERIES CLASSIFIERS

  М. О. Dobrokhvalov , А.Y. Filatov , Е.А. Chegodaeva

  2026-02-27

  Abstract ▼

  Experimental reproducibility constitutes a critical cornerstone of modern machine learning research, yet random initialization seed selection substantially influences final model performance, creating challenges for principled comparison of different architectures and methods. Random seed effects on convolutional time series classifiers were quantified, and a principled comparison criterion was established. Two 1D architectures, FCN and ResNet, were trained on seven public datasets containing different data. 55 independent runs for each combination of model and dataset were performed nder controlled pseudorandomness in Python, NumPy, and PyTorch. Deterministic backends were enabled, and identical hyperparameters were used across runs. Normality of seed-wise accuracy distributions was assessed with the Shapiro–Wilk and Anderson–Darling tests. Accuracy variability attributable to seed choice reached up to 12 percentage points in some settings, with magnitude dependent on dataset and architecture. The distributions were found to be non-normal in most cases, indicating that confidence intervals predicated on normality are unreliable. To enable fair comparison across runs, a reproducibility meta-metric, RM, was introduced that subtracts a dispersion penalty from the mean and depends on the number of runs and a tunable coefficient λ. RM was shown to lie between the empirical minimum and the mean, to approach the lower bound for small sample sizes, and to converge toward the mean as the number of runs increases. Portability of the approach was examined on an additional architecture, DenseNet, confirming expected behavior. Practical value is provided by RM metric rankings reflect both performance and stability. In this way, reproducibility and the credibility of empirical conclusions are strengthened

• METHODOLOGICAL SUPPORT FOR ASSESSING THE AVAILABILITY OF GOODS IN DISTRIBUTED STORAGE BASED ON COMPUTER VISION METHODS

  А.R. Nedvigin , R.М. Sinetsky

  2026-02-27

  Abstract ▼

  This paper presents a formalization of the problem of automated monitoring of product availability on retail shelves and compliance with the prescribed planogram, leveraging computer vision and machine learning techniques. The aim of this research is to develop algorithmic solutions for the automatic assessment of product availability in distributed retail environments using computer vision methods, thereby addressing the challenge of maintaining optimal and necessary product assortments through continuous shelf monitoring and supporting data-driven managerial decision-making. A technological pipeline for visual data processing is proposed, comprising the stages of image normalization, segmentation, object localization, and classification, implemented with convolutional neural networks—specifically YOLO and U-Net architectures. An integrated product availability metric is introduced, which jointly accounts for physical, visual, and informational dimensions of availability. An optimization problem aimed at improving overall availability is formulated, and an adaptive neural network fine-tuning mechanism is implemented to enhance the accuracy of image recognition and segmentation, as well as the quality of analytical recommendations. Furthermore, an availability-improvement algorithm is proposed for a decision support system, based on the construction of an optimized merchandiser routing plan that prioritizes products and minimizes time expenditures. This routing problem is reduced to a generalized Traveling Salesman Problem (TSP) with priority-based weights. Methods for evaluating and enhancing product availability are proposed and described in detail. Based on the developed approaches and algorithms, a software system for monitoring and improving product availability has been implemented. Experimental results confirm the effectiveness of the proposed solutions: the average recognition accuracy reached 95.8%, and the integrated availability score achieved A = 0.93. The practical significance of this work lies in establishing an algorithmic foundation for intelligent shelf-monitoring systems that enable more efficient management of retail operations and inventory processes

• MODERN METHODS OF HYPERSPECTRAL IMAGE PROCESSING: SYSTEM ANALYSIS, ALGORITHMS AND PROSPECTS FOR APPLICATION IN CONSTRUCTION DIAGNOSTICS

  М. А. Filonova , S. N. Shirobokova

  188-208

  2026-07-07

  Abstract ▼

  The relevance of this study is determined by the growing interest in hyperspectral imaging as a tool for non-destructive testing of building materials and structures, as well as by the insufficient systematization of modern methods for processing such data. The aim of the work is to provide a systematic analysis of HSI processing algorithms, identify their advantages and limitations, and determine the prospects for their application in construction diagnostics. The study examines the specific features of hyperspectral data, including high dimensionality, noise, calibration errors, atmospheric distortions, and the shortage of labeled datasets. The evolution of approaches is shown: from classical machine learning methods and manual feature engineering to deep neural networks. Dimensionality reduction methods, kNN classifiers, Bayesian models, logistic regression, Random Forest, SVM, and MLP are analyzed, along with methods for incorporating spectral-spatial context. Special attention is paid to modern deep learning architectures: 1D, 2D, and 3D CNNs, RNNs, LSTM/GRU models, hybrid CNN–RNN models, transformers, and CNN–Transformer schemes. Transfer learning, semi-supervised learning, self-supervised learning, few-shot learning, meta-learning, and domain adaptation are considered separately as ways to overcome the limited availability of labeled data. The approaches are compared in terms of data requirements, computational complexity, robustness to noise, and their ability to account for spectral and spatial dependencies. It is shown that the most promising models for construction diagnostics are hybrid models that combine local convolutional features, the global context of attention mechanisms, and the possibility of fine-tuning on small specialized datasets. The paper summarizes the current state of the field and forms a basis for selecting methods for defect detection, moisture assessment, corrosion analysis, and evaluation of degradation in building structures. The conclusions formulated in the study can be used when designing experimental protocols and selecting architectures for further applied research in the field of building monitoring