This article discusses the formation and connection of layers in the task of classifying images of traffic signs, as well as calculating weights on the corresponding layers of the neural network. The authors describe the biological structure of brain neurons, as well as their comparison with artificial neural networks. A conceptual model of an artificial neuron and a neural network with a description of structural elements is presented. The matrix structure of the weights of the neural network is given. The process of converting an RGB image of a road sign into an input layer of a neural network is described. A corresponding description is provided for each hidden layer. In addition, a description of the convolution layers and the maximum pool is given, as well as an explanation of the need to use this type of layers in a convolutional neural network. The authors also described an algorithm for the formation of a convolutional neural network for the classification of road signs. Examples of the operation of this neural network are given.
Keywords: convolutional neural networks, classification, deep learning, big data, mathematical modeling, computer science
This paper describes the issue of choosing a distributed registry platform when designing information systems in the financial sector of the economy. The relevance of these studies is due to the ever-increasing growth in demand for information systems of the financial sector of the economy formed using distributed registry technology. The growing interest in this technology is associated with the need to ensure reliable storage of information, the change of which will be monitored by the participants of this transaction. The purpose of this work is to determine the most suitable platform using the hierarchy analysis method. In the course of the work, the main platforms of the distributed registry were identified, as well as the key criteria for these frameworks were determined, taking into account the requirements of business process participants. These criteria were evaluated. For each alternative evaluation matrix, the indicators of the maximum eigenvalue vector were determined according to separate criteria, and the consistency of the judgment was proved, including the determination of the consistency index, the local priority index and the consistency ratio. A synthetic analysis of the criteria under consideration was carried out. Based on the data obtained during the synthetic analysis, the most promising platform was selected. Conclusions on the evaluated systems are formed.
Keywords: distributed registry, hierarchy analysis method, system analysis, information systems, computer science
This article discusses the problems of constructing convolutional neural networks for determining road objects. The general relevance and formulation of the problem of determining road objects is presented. The rationale for the use of artificial neural networks for determining road objects has been formed. The Retinanet network architecture is used as the main architecture of an artificial neural network for determining road objects. The general concept of this architecture and the main subnets are visualized. Error functions for the main subnets of the Retina net network are described. The design description of algorithms for constructing data annotation for training an artificial neural network, as well as algorithms for constructing the neural network architecture of classification, regression and feature pyramid is given. The dynamics of changes in the general error function when determining road objects is determined. The result of training an artificial neural network is presented.
Keywords: convolutional neural networks, classification, regression, convolutional neural networks, deep learning, big data, mathematical modeling, computer science, RetinaNet architecture
In this article discusses the problems of determining traffic signs for driving a motor vehicle using an artificial neural network apparatus. The relevance of research at this point in time is described, as well as the advantages of using neural networks in determining traffic signs. The input data for determining traffic signs for convolutional neural networks are presented. The architecture of the convolutional neural classification network is formed, in particular, the sequence of layers of the image classification network is considered. A mathematical description of the modeling of the error function and the stochastic gradient descent method is given. A mathematical model of the learning process of an artificial neural network, as well as activation functions: linear functions and sigmoids is presented. An algorithm for forming an artificial neural network model is proposed. The learning process of this function is visualized on the graph. The result of the training is presented.
Keywords: artificial neural networks, classification, convolutional neural networks, deep learning, big data, mathematical modeling, informatics
This article discusses the problems of determining the road surface for automatic control of a vehicle using an artificial neural network. The current state of the industry is described, as well as the relevance of these studies. Describes the input data for determining the road surface. The idea of the applicability of the image segmentation method for determining the road surface is substantiated. The structure of an artificial neural network based on the U-NET architecture is being formed. In particular, the structure of the sequence of layers is described. Particular attention is paid to the mathematical modeling of the convolution process and the maximum pool. A mathematical model of the learning process of an artificial neural network, as well as activation functions: linear functions and sigmoids, is given. An algorithm for forming an artificial neural network model is proposed. The learning process of this function is visualized on the graph. The result of the training is presented.
Keywords: artificial neural networks, UNET, data analysis,, machine learning, deep lerning, convolutional neural networks, convolution, maximum pool, image segmentation, modeling
This article deals with the problem of determining the cost of renting real estate. The idea of minimizing the absolute error function using artificial neural networks is substantiated. Particular attention is paid to the process of determining the input data of the neural network. In particular, the problems of determining such parameters as the improvement of the region and premises. The article clarifies the features of determining the weight coefficients to determine the technical equipment of the room using a genetic algorithm. A model of neural network architecture is proposed. The model of change of weight coefficients is described. As a result, the model was tested on test data, and the model of data correction taking into account price dynamics was described.
Keywords: neural network, data mining, data analysis, real estate rental, regression, genetic algorithm, Informatics, machine learning, cost estimation, modeling, extrapolation
This article discusses the problems of optimal placement of information resources on the nodes of a computer network. The main methods used in solving this problem are presented. In particular, the method of random allocation of resources, optimization of allocation of resources using the branch and bound method, and optimization of allocation of resources using a genetic algorithm are considered. For these methods, the structure of the input and output data has been determined, in addition, the internal structure of resource allocation has been demonstrated for the presented methods. A key aspect of the consideration in this article is the formulation of the problem and the modeling of its solution using the algorithms presented. As a result, testing of the developed module on the input data and analysis of the prospects for using the module are presented.
Keywords: computer network, information resource, random allocation of resources, simplex method, branch and bound method, genetic algorithm, design, modeling
This article presents a mathematical model of the distributed registry as a Queuing network. The main components of this network, as well as their formal representation are considered. The model of the peer-to-peer network is visualized, the vector of the network state is defined, and the restrictions of the state space are defined. After that, the laws of distribution of individual flows and service time were presented. In addition, the design elements of the infinitesimal matrix were determined. Based on the data obtained, a simulation model of this process was produced. For simulation, the Anylogic package was used. The results of simulation were analyzed and the most optimal parameters were selected.
Keywords: Queuing network, information security, distributed registries, computer science and engineering, mathematical modeling information system, corda
The article describes the general model of the information system for managing information flows when renting real estate using the technology of distributed registries on the corda platform. The description of participants of the information system is presented. The key concepts of the use of distributed register technology for the implementation of the information system for real estate management are considered. The methods of data analysis of the information system for real estate management are presented. A method for approximating data using the Lagrange polynomial is considered. A method for estimating the value of real estate using artificial neural network technology is described.
Keywords: rent, real estate, data analysis, information security, neural network, distributed registries, Informatics, information system, corda