Papers by Vladimir Golovko
International Journal of Computing, Aug 1, 2014
Applied Mathematics and Computation, 2007
The energy consumption in food and beverage industries in Iran was investigated. The energy consu... more The energy consumption in food and beverage industries in Iran was investigated. The energy consumption in this sector was modeled using artificial neural network (ANN), response surface methodology (RSM) and genetic algorithm (GA). First, the input data to the model were calculated according to the statistical source, balance-sheets and the method proposed in this paper. It can be seen that diesel and liquefied petroleum gas have respectively the highest and lowest shares of energy consumption compared with the other types of carriers. For each of the evaluated energy carriers (diesel, kerosene, fuel oil, natural gas, electricity, liquefied petroleum gas and gasoline), the best fitting model was selected after taking the average of runs of the developed models. At last, the developed models, representing the energy consumption of food and beverage industries by each energy carrier, were put into a finalized model using Simulink toolbox of Matlab software. The results indicated that consumption of natural gas is being increased in Iranian food and beverage industries, while in the case of fuel oil and liquefied petroleum gas a decreasing trend was estimated.
В данной работе рассмотрена задача распознавания графической метки ведущего робота в системе веду... more В данной работе рассмотрена задача распознавания графической метки ведущего робота в системе ведущийведомый роботы. Для решения задачи был разработан метод детектирования на основе RAM-based сетей позволяющий по расположению метки узнать положение и дальность ведущего робота по графическому паттерну. Обученная RAM-based сеть хранит характеристические особенности паттерна в разных секторах относительно ведомого робота. Сработавший дискриминатор сети будет указывать на сектор, в котором находится ведущий робот. Ключевые слова: RAM-based сети; стайные роботы, распознавание образов, выбор порога.
This work is devoted to the consideration of the most important factor providing semantic compati... more This work is devoted to the consideration of the most important factor providing semantic compatibility of intelligent computer systems and their componentsstandardization of intelligent computer systems, as well as standardization of methods and tools of their design.
Pattern Recognition and Image Analysis, 2021
Training methods for deep neural networks (DNNs) are analyzed. It is shown that maximizing the li... more Training methods for deep neural networks (DNNs) are analyzed. It is shown that maximizing the likelihood function of the distribution of the input data P(x) in the space of synaptic connections of a restricted Boltzmann machine (RBM) is equivalent to minimizing the cross-entropy (CE) of the network error function and minimizing the total mean squared error (MSE) of the network in the same space using linear neurons. The application of DNNs for the detection and recognition of productmarking is considered.
Communications in Computer and Information Science, 2020
International Journal of Computing, Aug 1, 2014
A goal of EEG signals analysis is not only human psychologically and functionality states definit... more A goal of EEG signals analysis is not only human psychologically and functionality states definition but also pathological activity detection. In this paper we present an approach for epileptiform activity detection by artificial neural network technique for EEG signal segmentation and for the highest Lyapunov's exponent computing. The EEG segmentation by the neural network approach makes it possible to detect an abnormal activity in signals. We examine our system for segmentation and anomaly detection on the EEG signals where the anomaly is an epileptiform activity.
International Journal of Computing, Aug 1, 2014
In this article the classification task in the domain of intrusion detection is considered. Often... more In this article the classification task in the domain of intrusion detection is considered. Often a chosen algorithm is not good enough for practical use. So the question arises how is it possible to improve the performance? In this case we can employ so-called Committee Machines that increase accuracy and reliability of the base classification model. These advantages are the result of dividing complex computational problems among several experts. The knowledge of each expert influences on the general conclusion of Committee Machine.
International Journal of Computing, 2014
Most current Intrusion Detection Systems (IDS) examine all data features to detect intrusion. Als... more Most current Intrusion Detection Systems (IDS) examine all data features to detect intrusion. Also existing intrusion detection approaches have some limitations, namely impossibility to process large number of audit data for real-time operation, low detection and recognition accuracy. To overcome these limitations, we apply modular neural network models to detect and recognize attacks in computer networks. It is based on combination of principal component analysis (PCA) neural networks and multilayer perceptrons (MLP). PCA networks are employed for important data extraction and to reduce high dimensional data vectors. We present two PCA neural networks for feature extraction: linear PCA (LPCA) and nonlinear PCA (NPCA). MLP is employed to detect and recognize attacks using feature-extracted data instead of original data. The proposed approaches are tested using KDD-99 dataset. The experimental results demonstrate that the designed models are promising in terms of accuracy and computa...
DZIOMIN V.V., KABYSH A.S., DUNETS I.P., DUNETS A.P., GOLOVKO V.V. Using RAM-based networks for vi... more DZIOMIN V.V., KABYSH A.S., DUNETS I.P., DUNETS A.P., GOLOVKO V.V. Using RAM-based networks for visual mark detection
Современные проблемы математики и вычислительной техники
Rubanov V. S., Golovko V. A., Sadykhov R. Kh., Lazakovich N. V., Kalinin A. I., Golenkov V. V., S... more Rubanov V. S., Golovko V. A., Sadykhov R. Kh., Lazakovich N. V., Kalinin A. I., Golenkov V. V., Starovoitov V. V., Dudkin A. A., Savchuk V. F., Raketskii V. M., Derechennik S. S., Makhnist L. P., Parfomuk S. I., Muravyev G. L., Savitsky Yu. V., Kostyuk D. A. Modern problems of mathematics and computer engineering
A comparative analysis of various methods and architectures used to solve the problem of object d... more A comparative analysis of various methods and architectures used to solve the problem of object detection is carried out. This allows so-called oneway neural networks architectures to provide high quality solutions to the problem. A neural network algorithm for labeling images in text documents is developed on the basis of image preprocessing that simplifies the localization of individual parts of a document and the subsequent recognition of localized blocks using a deep convolutional neural network. The resulting algorithm provides a high quality of localization and an acceptable level of subsequent classification.
In this paper a novel method for detection of network attacks and malicious code is described. Th... more In this paper a novel method for detection of network attacks and malicious code is described. The method is based on main principles of Artificial Immune Systems where immune detectors have an Artificial Neural Network’s structure. The main goal of proposed approach is to detect unknown, previous unseen cyber attacks (malicious code, intrusion detection, etc.). The mechanism of evolution of the neural network immune detectors allows increasing the detection rate. The proposed Intelligent Cyber Defense System can increase the reliability of intrusion detection in computer systems and, as a result, it may reduce financial losses of companies from cyber attacks.
International Journal of Computing, 2014
This paper describes a multi-agent influence learning approach and reinforcement learning adaptat... more This paper describes a multi-agent influence learning approach and reinforcement learning adaptation to it. This learning technique is used for distributed, adaptive and self-organizing control in multi-agent system. This technique is quite simple and uses agent’s influences to estimate learning error between them. The best influences are rewarded via reinforcement learning which is a well-proven learning technique. It is shown that this learning rule supports positive-reward interactions between agents and does not require any additional information than standard reinforcement learning algorithm. This technique produces optimal behavior of multi-agent system with fast convergence patterns.
International Journal of Computing, 2020
In this work, we draw attention to prediction of football (soccer) match winner. We propose the d... more In this work, we draw attention to prediction of football (soccer) match winner. We propose the deep multilayer neural network based on elastic net regularization that predicts the winner of the English Premier League football matches. Our main interest is to predict the match result (win, loss or draw). In our experimental study, we prove that using open access limited data such as team shots, shots on target, yellow and red cards, etc. the system has a good prediction accuracy and profitability. The proposed approach should be considered as a basis of Oracle engine for predicting the match outcomes.
Applied Intelligence, 2017
In this work we contribute to development of a "Human-like Visual-Attention-based Artificial Visi... more In this work we contribute to development of a "Human-like Visual-Attention-based Artificial Vision" system for boosting firefighters' awareness about the hostile environment in which they are supposed to move along. Taking advantage from artificial visual-attention, the investigated system's conduct may be adapted to firefighter's way of gazing by acquiring some kind of human-like artificial visual neatness supporting firefighters in interventional conditions' evaluation or in their appraisal of the rescue
Soft Computing, 2017
Fitting the skills of the natural vision is an appealing perspective for artificial vision system... more Fitting the skills of the natural vision is an appealing perspective for artificial vision systems, especially in robotics applications dealing with visual perception of the complex surrounding environment where robots and humans mutually evolve and/or cooperate, or in a more general way, those prospecting human-robot interaction. Focusing the visual attention dilemma through human eye-fixation paradigm, in this work we propose a model for artificial visual attention combining a statistical foundation of visual saliency and a genetic tuning of the related parameters for robots' visual perception. The computational issue of our model relies on the one hand on center-surround statistical features' calculations with a nonlinear fusion of different resulting maps, and on the other hand on an evolutionary tuning of human's gazing way resulting in emergence of a kind of artificial eye-fixation-based visual attention. Statistical foundation and bottom-up nature of the proposed model provide as well the advantage to make it usable without needing prior Communicated by V. Loia.
Communications in Computer and Information Science, 2017
At present the deep neural network is the hottest topic in the domain of machine learning and can... more At present the deep neural network is the hottest topic in the domain of machine learning and can accomplish a deep hierarchical representation of the input data. Due to deep architecture the large convolutional neural networks can reach very small test error rates below 0.4% using the MNIST database. In this work we have shown, that high accuracy can be achieved using reduced shallow convolutional neural network without adding distortions for digits. The main contribution of this paper is to point out how using simplified convolutional neural network is to obtain test error rate 0.71% on the MNIST handwritten digit benchmark. It permits to reduce computational resources in order to model convolutional neural network.
This paper presents an evolution of an ontology-based approach to designing batch manufacturing e... more This paper presents an evolution of an ontology-based approach to designing batch manufacturing enterprises. According to Industry 4.0 approach, instead of isolated view of a manufacturing process inside a single enterprise this new approach encompasses related business entities as well-raw material suppliers (e.g. dairy farms) and large-scale consumers (e.g. stores or retail chains). Special attention is paid to logistics processes: a short description of fundamental logistics processes of cottage cheese production is provided, as well as subject domain structure of logistics and an example of formal specification of emergency logistics situation. It is shown that multiagent industrial control system with agents interacting via shared memory is compliant with design principles of Industry 4.0 approach. Standards formalization topic is touched upon as well. PFC, a graphical procedural model specification language, formalization is discussed. PFC is specified in ISA-88.02 standard. G...
At present the deep neural network is the hottest topic in the domain of machine learning and can... more At present the deep neural network is the hottest topic in the domain of machine learning and can accomplish a deep hierarchical representation of the input data. Due to deep architecture the large convolutional neural networks can reach very small test error rates below 0.4% using the MNIST database. In this work we have shown, that high accuracy can be achieved using reduced shallow convolutional neural network without adding distortions for digits. The main contribution of this paper is to point out how using simplified convolutional neural network is to obtain test error rate 0.71% on the MNIST handwritten digit benchmark. It permits to reduce computational resources in order to model convolutional neural network.
Uploads
Papers by Vladimir Golovko