Artificial Intelligence Impressions

The Study of Fashion Items Generation Based on Big Data and Deep Learning

Yang Song, Zhijian Wang, Jianming Zhang, Yan Wang — Fri, 14 Oct 2022 00:00:00 +0000

Since the Third Industrial Revolution, computer network information technology has made continuous progress, and the era of big data and Artificial Intelligence (AI) has arrived. The rapid progress of big data and AI will not only be able to promote the continuous improvement of computer network information technology, but also improve the level of economic development and make positive contributions. In this chapter, we propose a style-based fashion items generator for clothing design with big data and deep learning. Designing a clothing item is complicated, time-consuming, and challenging for clothes designers and clothing industry; however, recent improvements in conditional image generation provide a feasible solution. With a desired fashion category, the proposed framework will generate a nonexistent fashion item which cannot be distinguished from the real ones. Generative adversarial networks (GANs) make it possible to perform our clothing design with semi-supervised conditions. Our method can generate clothing items conditioned on 15 fashion categories. Furthermore, due to the multimodality of the clothes images, we employ a style-based generator with disentangled networks and adopt a multistage discriminator to improve the results of image generation. The effectiveness of our approach is well demonstrated through quantitative experiments. Our method will spark inspirations for fashion designers in their work.

Big Multimodal Data-driven Medical Image Processing for the COVID-19 Diagnosis

Thu, 24 Nov 2022 00:00:00 +0000

Diagnosis using medical images is extremely labor-intensive and could even be subjective. With the exciting progress of artificial intelligence (AI) in the last decade, it has been increasingly realistic and promising to develop automatic diagnoses from medical images. In this chapter, we give comprehensive advances on recent data-driven AI technologies for the COVID-19 diagnosis based on medical images. Moreover, rather than using a massive volume of data to train a network, it is also valuable to take multi-modal multi-view data into a unified learning framework to improve the accuracy of diagnosis. To this end, two categories of representation learning methods are proposed to deal with this variety and variability of big medical image data. One is an average approximate hashing (AAH) method for searching large-scale multimedia databases, which projects data into different semantic spaces but shares a unified hash code. The other focuses on nonnegative matrix factorization-based clustering models for multi-view data. Experiments justify the effectiveness and efficiency of the proposed methods.

Parallel Optimization of Weld Design Based on Collective Decision-Making

Andrey Batranin, Raisa Krektuleva — Fri, 14 Oct 2022 00:00:00 +0000

Non-consumable electrode welding is widely used in manufacture, construction and other areas. Despite progress in welding technology and the emergence of new methods of joining materials, non-consumable electrode welding is confidently occupying its niche in the production area. In this work we consider the problem of optimal design in non-consumable electrode welding based on the solution of the inverse heat transfer problem by means of parallel computations. Heat transfer problem is solved by finite difference method for orthogonal geometry in transient mode. Authors used their own program, called the “Welding engineer virtual workplace” which is developed in Matlab. In this work we present the basis for solving the inverse heat transfer problem and an algorithm of parallel searching for the optimal technological parameters of the non-consumable electrode welding. The final result of design is based on a collective decision-making approach.

A Hybrid Method for Analysis of Multimodal Data for Establishing the Mechanism of Sorption of Oil and Petroleum Products by Natural Sorbent: Hair of Flyers of Horse Flowers

Olga Gennadievna Gorovykh, Baurzhan Alpysovich Alzhanov — Thu, 24 Nov 2022 00:00:00 +0000

The economic and economic value of cattail consists of its medicinal, technical (fuel, construction, roofing, garter, wicker, cooper), as well as food, fodder, perganose, indicator, decorative and other uses. However, far from all the possibilities of using this plant have been disclosed. An interesting application of cattail is the use of its inflorescences (ears) as a sorbent when carrying out work to eliminate emergency spills of oil and oil products on the water surface. The chapter presents the results of studies on the use of the pistillate part of cattail inflorescences as an oil sorbent. A rather unique set of properties of the hairs of fly mats inflorescences of cattail - long-term buoyancy (more than 100 days), high sorption capacity, low water absorption and the possibility of multiple regeneration requires explanation and establishment of the sorption mechanism. To establish the mechanism of sorption by the sorbent - cattail inflorescences, a hybrid method for analyzing the data obtained from individual studies was used. On the basis of the hybrid method, it has been established that the sorption mechanism consists in a sequence of such processes as: dissolution of the wax layer covering the surface of an individual hair of a cattail fly; the formation of ruptures and the opening of internal cavities; stratification of individual hairs into several ribbons. Such a complex combination of processes leads to a synergistic effect that ensures high performance of the sorbent and an increase in its active surface by 30 %. The operational properties of a sorbent based on cattail inflorescences are given: density, oil capacity in static and dynamic conditions for various hydrocarbons, regeneration by mechanical squeezing, buoyancy, higher heat of combustion determined on a calorimeter IKA- WERKE C 5003 contro. The complex of the revealed properties of cattail inflorescences allows them to be effectively used when carrying out work to eliminate emergency oil and oil products spills on the water surface. For this purpose, a method of using adhesive drum devices for collecting waste sorbent from the water surface is proposed.

Methods and Problems of Demercurization, New Effective Solutions for Decontamination of Mercury-Contaminated Objects

Thu, 24 Nov 2022 00:00:00 +0000

Statistical data are presented on the number of trips to eliminate emergency situations related to mercury contamination of premises for various purposes, vehicles and territories in the Republic of Belarus and the United States of America in recent years. The solutions and chemical materials proposed for use by various authors for demercurization works are considered. The problems arising from the use of these demercurizing compositions have been noted. It is shown that the currently used methods on the basis of chemical demercurizing solutions do not allow to carry out demercurization when the concentration of mercury vapor reaches below the maximum permissible concentration in a reasonable time. The importance of solving this problem has been shown. The necessity of development of new approaches to elimination of emergency situations connected with presence of mercury is specified. The method of decontamination of mercury-containing materials by means of ozonization is offered. The experimental unit and the instrument base used in carrying out the experiment are described. The regularities of change of ozone concentration by height of a room are established, at its receipt in the closed kill on its lower level. The obtained experimental data are compared with the theoretical rate of change of ozone concentration by height based on molecular diffusion of ozone. The explanation of the obtained results is offered. The average rate of spontaneous ozone decomposition without access of fresh air and presence of oxidized substances is established. The comparison of calculated theoretically and experimentally determined rate of ozone spontaneous decomposition in time is carried out. It is proposed to use the obtained data for development of methodical recommendations on application of ozone as demerchurizing agent. It is shown that the technology based on the use of ozone will allow to provide decontamination of surfaces and air in this volume efficiently and in the shortest possible time.

Hybrid Methods for Analysis of Structure and Phase Transitions in Alloys under Arс Heating

Raisa Krektuleva, Roman Cherepanov, Evgenii Fedin — Thu, 24 Nov 2022 00:00:00 +0000

The work solves the problem of comprehensive hybrid analysis of the state of the alloy material during machining by a moving electric arc heating source. The main stages of solving the problem include creating a numerical 3D model of thermophysical processes in the alloy material under the action of an electric arc, calculation of thermophysical characteristics of the alloy material and analysis of its state diagram, experimental validation of the calculated methodology. A special algorithm was developed to understand (data mining) the complex multi-cycle dynamics of thermal fields in a wide temperature range. This algorithm made it possible to record multiple cyclic states of the treated materials (heating - melting - evaporation - cooling - crystallization and formation of new phases) and to determine their geometry in the volume. At the last stage of the study, computer experiments were reproduced in reality. We obtained a good agreement between the calculated and in-situ experiments.

Research on the Rapid Prototyping Process Optimization Technology of Complex Heterogeneous Mould Oriented by Manufacturing Big Data

Yan Cao, Liang Huang, Yu Bai, Jiang Du — Thu, 24 Nov 2022 00:00:00 +0000

In order to meet the flexible market needs of complex heterogeneous moulds in recent years, and for slove the defects of small batch and multi-variety moulds in traditional CNC milling. This chapter proposes an additive mould manufacturing method based on Fused Deposition Modelling (FDM) rapid forming technology, which can further reduce the cost and cycle of such mould manufacturing while ensuring the quality of moulds. This method first combines with the cloud manufacturing design concept, and constructs the qualitative relationship between the process data and the forming quality and forming efficiency in the FDM overall forming process; Secondly, according to the forming mechanism model, construct a quantitative mapping model between process parameters on forming quality and efficiency; Finally, through the complex mould digital intelligent manufacturing system under the guidance of cloud manufacturing big data, to lay the foundation and contribution to the future cloud manufacturing technology of complex and heterogeneous products.

Using a Hybrid FNN Method for Image Classification of Satellite Remote Sensing Data

Thu, 24 Nov 2022 00:00:00 +0000

Satellite remote sensing images play important roles in many practical applications, including meteorology, natural resource identification, ecology, agriculture, emergency and disaster management, as well as mapping and surveying. With the rapid development of the aerospace industry, more and more onboard imaging systems are being constructed and launched, many with hyperspectral and high-resolution capabilities. To fully use the huge amounts of remotely sensed data provided by these systems requires appropriate algorithms, developing these is an ongoing challenge for researchers in academia and in industry. In this chapter the basic concepts of satellite remote sensing image analysis are discussed and the fuzzy neural network (FNN) approach described, both in the context of hyperspectral/multi-spectral images and in the context of images based on the red, green, blue colours of visible light. Experiments based on real-world examples of such images are carried out to illustrate the methods involved. The results are analyzed and show that the FNN approach can give good results for both multi-spectral and visible light images.

Big Data Analysis and Multi-Objective Optimization for Smart Grid

Lei Xu, Sergey Gorbachev, Dong Yue, Chunxia Dou — Thu, 24 Nov 2022 00:00:00 +0000

With the growth of photovoltaic power generation, the safe operation problems are becoming increasingly important. At the same time, the high uncertainty of load also brings great pressure to the power grid. It becomes necessary to improve the accuracy of distributed photovoltaic power generation and load prediction, and various emerging prediction methods are used based on smart grid, which provides support for accurate prediction of modern power system combined with big data analysis because of its high integration of information, and it has become an important research direction. This chapter will review the distributed photovoltaic power generation and load forecasting of smart grid from the perspective of technical methods. In addition, this chapter also sorts out the multi-objective optimization methods of smart grid, and introduces the optimization methods in the smart grid scenario with high proportion of distributed generation. At the same time, the distributed optimization methods with high efficiency and privacy reflected in power grid optimization is also analyzed in this article.

Thermal Performance Hybrid Optimization of Natural Gas Hydrate Wellbore

Ning Wang, Hui Wang, Victor Kuzin, Binhui Zheng, Taohong Xu — Thu, 24 Nov 2022 00:00:00 +0000

Studies show that there are massive gas hydrate reservoirs in offshore deepwater seabed. The main challenge of exploiting these resources is THE high-cost transportation of the extracted gas to the onshore facilities. Moreover, adjusting the temperature near the wellbore is another challenge to prevent regeneration of gas hydrate during extraction. Therefore, the hybrid analysis of the finite difference method and the enthalpy method is used to get a synergistic effect in order to realize the effective thermal property output. To analyze the transient heat transfer process, a 2D model is developed to simulate the wellbore area. Moreover, different materials are used for insulation and the corresponding thermal performances are calculated using the enthalpy method. The obtained results show that the soil surrounding the wellbore performs as an insulator. Adding an insulation layer outside the wellbore can effectively enhance the exit temperature. Moreover, the model indicates that the fluid holding time can be dramatically extended using the microphase change materials (MPCM) as an insulation layer.

Multi-Criteria Fuzzy AHP Method for Analyzing the Importance of E-Learning Platforms during Covid-19

Thu, 24 Nov 2022 00:00:00 +0000

The demand for e-learning has risen to new heights due to uncertainty during the Covid19 pandemic. Every organization try to find better ways of communication with their stakeholders and therefore most of the applications used widely during covid19 pandemic include virtual conferencing tools. However, the effectiveness of virtual conferencing tools, especially those used for teaching and learning purposes, needs to be assessed critically. As a standard practice, it is always useful to consider multiple inputs in decision making especially when there is a need to achieve optimal results. In this regard e-learning platform impacts have been analyzed in this paper with published research articles which are indexed in the Scopus database. The datasets include a lot of information including authors' nativity, journal information, date of publication, abstract of the research article etc. In line with this perspective, this paper considers the application of hybrid multi-criteria decision making for fuzzy logical analysis with analytic hierarchy process (AHP) and the possibility of analytic network process (ANP) to assist decision-making in assessing and choosing priority for high impact e-learning platform. This multi-criteria model can assist decision-making in assessing and choosing priority for high impact e-learning platforms accurately and effectively by using the Fuzzy AHP method. As per AHP and F-AHP results, its important to aware of the e-learning platforms impact using publication data perspectives, the common things in both analysis is criteria C5 («Abstract») with good impact result. And also other two criterion like C1 («Title of the article») and C7 («Location-Country») have significant impact on this research article publication as per findings of Fuzzy AHP.

Fuzzy Logic Traffic Light Control Model Using Arduino UNO Microcontroller

Sergey Gorbachev, Maxim Bobyr, Yang Yang, Fei Ding — Thu, 24 Nov 2022 00:00:00 +0000

The chapter is devoted to modeling the traffic light control process based on fuzzy logic with the possibility of adjusting the time intervals of traffic light signals depending on the traffic situation. The definition of input variables for a fuzzy logic control system of an intelligent traffic light is performed using a vision system. The proposed method of controlling a traffic control device is based on a fuzzy inference system and contains several stages: determination of clear input variables, fuzzification of the values of input variables, aggregation of data based on fuzzy rules, defuzzification of values and determination of the delay time of the permitting traffic light signal. According to the proposed fuzzy model, an experimental layout based on the Arduino Uno controller has been developed, simulating the operation of an intelligent traffic light control system. A specialized software model has been created, which has been patented ("A program for regulating traffic lights based on fuzzy logic"). The results of experimental studies have shown the high efficiency of the traffic light control model in the daily cycle. The model successfully copes with the estimation of the traffic density of cars and pedestrians, proportionally adjusting the operating time of traffic lights. The development and implementation of this model will ensure the safety and convenience of road traffic for all participants.

Leader-Follower Consensus of Fractional-Order Multi-Agent Systems Using Adaptive Fuzzy Control

Yilin Hao, Jinde Cao, Sergey Gorbachev, Heng Liu — Thu, 24 Nov 2022 00:00:00 +0000

This part studies the leader-follower consensus (LFC) issue of fractional-order multi-agent systems (FOMASs) with nonlinear or linear function. Firstly, a fractional Lyapunov method is proposed to prove the stability of the FOMAS. Secondly, based on a sliding mode control method, when the follower dynamics is linear, a controller is devised to realize this uniform tracking question. Thirdly, when the follower has a nonlinear dynamic function satisfying the Lipschitz condition, it is verified that the devised controller is still reasonable. Fourthly, suppose that the dynamics of the followers are unknown, an adaptive fuzzy variable structure controller is devised to guarantee the consistency of the FOMAS. Finally, noting that the proposed conditions for stability and consensus are easy to verify, simulation calculations are written to expound the adaptability of control scheme.

Preface

Sergey Gorbachev, Boris Gusev, Victor Kuzin, Shengli Xie, Dong Yue — Fri, 14 Oct 2022 00:00:00 +0000

This volume describes the solution of problems of Big Data intellectual analysis, control, design and optimization. Big Data - technologies that extract maximum benefit from big data, which are widespread in all spheres.

Starting with an official introduction to the basics of algorithm hybridization, this book combines many different aspects of current research on hybrid technologies, such as deep neural networks, fuzzy neural networks, multi-MISO ANFIS, fuzzy C-means, conditional disentangled networks, generative adversarial networks, finite difference method and enthalpy method.

The book also covers a wide range of applications and implementation problems, from pattern recognition and image generation to intelligent forecasting problems, automation of production in technical applications (3D-analysis, forecasting of distributed photovoltaic systems and loads) with due attention to modeling.

It covers a wide range of applications in the field of Big Data analysis, as well as Data Mining.

In addition to the traditional tasks of classification, clustering, forecasting, it also discusses original approaches to hybrid optimization and control in the tasks of multi-object optimization for smart grid, natural gas hydrate wellbore, parallel search for optimal technological parameters of the non-consumable electrode welding.

The articles are arranged in five thematic topics, I) Strongly coupled (functional) hybrid methods (articles 1-3); II) Loosely coupled (functional) hybrid methods (articles 4-5); III) Transformational hybrid methods (articles 6-7); IV) Integrated methods (articles 8-12) and V) Distributed hybrid intelligent methods (article 13).