Existing Research Work Research Articles

IPCADP-Equalizer: An Improved Multibalance Privacy Preservation Scheme against Backdoor Attacks in Federated Learning

Although there are some protection mechanisms in federated learning, its training process is still vulnerable to some powerful attacks, such as invisible backdoor attacks. Existing research work focuses more on how to prevent attacks in distributed training scenarios and improve the security of the FL training process, but it lacks consideration of utility and robustness, especially when the learning model of FL suffers from stealth backdoor attacks. This paper proposes an improved FL defense scheme IPCADP based on user-level differential privacy and variational autoencoders technology. The scheme can control and protect the privacy attribute of the image and can also eliminate the triggers that exist in the poisoned image. The experimental results show that compared with some existing defense schemes, IPCADP can defend against invisible backdoor attacks and improve the classification accuracy of the main task, while mitigating the impact of attacks on model robustness and stability. To a certain extent, the balance and unity of security, utility, and robustness are realized.

A Hybrid Metaheuristic Framework for Materialized View Selection in Data Warehouse Environments

The huge response time taken by the analytical queries is one of the most challenging problems in a data warehouse. The main reason for this high response time is the enormous amount of data being queried and the complex nature of the queries. This problem can be addressed with Materialized View Selection (MVS), where the optimal views with low response time are selected. To this end, the possible views for the analytical queries are randomly defined in the search space. Since most of the existing solutions are formulated based on semi-optimal solutions, this work introduces a hybrid metaheuristic-based framework to address the issue reliably. The major gaps in the existing research works are the resource wastage issue and the space constraint issue. These gaps are aimed to be addressed by the proposed hybrid framework. The proposed work integrates a Genetic Algorithm (GA) with Aquila Optimizer (AO) to search for the most optimal views from the search space. The presented solution follows highly efficient exploratory and exploitation behaviors to optimize the considered problem. Based on an iterative procedure, the most optimal views are selected by the hybrid framework, and these views are evaluated for minimum cost and time. Finally, performance evaluations are carried out to signify the practicality of the proposed approach.

INFRASTRUCTURAL FEASIBILITY INVESTIGATION FOR IMPLEMENTING AUTONOMOUS VEHICLES

Autonomous vehicles (AVs) have been in high demand over the last few years, and this demand is expected to continue growing. With this growth, infrastructure systems should be updated accordingly. Michigan is one of the AV-friendly states where the infrastructural systems are prepared to be ready for AVs. This research investigates infrastructure system enhancement for Michigan. Future infrastructure would not benefit significantly from AVs on their own, so it's still crucial to implement the proper policy interventions. Smart transportation addressing infrastructural management, sustainability, and development will include AVs as an additional component. Rather than analyzing AVs in isolation, it is necessary to incorporate the new technology context in which implementation occurs. The development of an engagement plan should include key personnel from original equipment manufacturers (OEMs) and AV technology firms. The engagement plan should aid in the identification of deployed and planned AV technologies and the investigation of technical infrastructure modification requirements. As a result of a detailed analysis of the existing research work and implementations on the effects of automation on infrastructure, this paper presents potential challenges to constructing the infrastructure required for automation. The research investigates the state of Michigan, its current applications in infrastructure investigation. The paper reviews the strategic framework in Michigan. As a result of all investigations carried out, the paper gives a feasibility investigation for Michigan’s infrastructure, considering existing challenges. The feasibility investigation mainly gives ideas to researchers and investors about infrastructural adaptation and readiness for autonomous vehicles in Michigan.

Energy Efficient Data Aggregation Approach for Sensor Network using Trust Mechanism

A wireless sensor network consists of numerous sensor nodes which are capable of entering and departing from the network according to their wish. The implementation of sensor networks is done at far places and they are small-sized. Thus, energy consumption becomes the main issue of WSN. Due to the limited energy of sensor nodes, the data that is collected from the intended environment is sent directly to the main station. The sink node receives the data that several sensor nodes send. The decision-making process is implemented by identifying and removing similarities between the data from various sensor nodes. In addition, the sink makes the deployment of obtained data locally as well as transmits these data to the networks which are executed far away. The existing research work employs CTNR, an energy-efficient protocol that is capable of enhancing the duration of WSN. The CTNR protocol consists of two-level hierarchies for mitigating the energy utilized in WSN. This protocol assists in choosing the CHs (cluster heads) on the basis of distance and energy. This work will focus on enhancing the CTNR routing algorithm so that the life span of the network can be prolonged.

A robust and efficient vector-based key management scheme for IoT networks

Information and communication security is a critical concern in the rapid growth of Internet of Things (IoT) networks that need to exchange sensitive data. Therefore, key management is essential to address such networks’ security problems. In this context, several existing research work focuses on key management solutions, which consider the resource-limitation of IoT devices. However, the weaknesses of these solutions are (1) the lack of protection for sensitive parameters during transmission, (2) security and performance metrics’ misbalancing, and (3) the vulnerability to node-compromising attacks. This paper presents a new key management scheme based on pre-distributed vectors to overcome these limits to secure key establishment, refresh, and revocation. Moreover, the network area is divided into subareas, which makes our solution lightweight, flexible, scalable, and resilient to multiple attacks while minimizing communications, computation, and storage overheads on IoT devices. We assess the proposed scheme by using (1) the BAN (Burrows Abadi Needham) formal verification logic, (2) the informal security to demonstrate its resilience against some known attacks, and (3) the performance analysis to prove its correctness. The obtained results show that the proposed scheme is more efficient than other schemes in terms of storage by saving more than 81.84%, communications by 100% during the group-wise key establishment, computation overheads by reducing the number of multiplication operations to 50%, and energy consumption by saving up to 99.99% during the group-wise key establishment phase. Moreover, the results show that our scheme is more resilient against node capture attacks by up to 96.43% during the initialization phase and by more than 9.89% after, making it suitable for use in resource-limited IoT networks.

Explore deep reinforcement learning for efficient task processing based on federated optimization in big data

In recent years, along with the extensive application of consumer electronics, the task execution with cloud computing for big data has become one of the research focuses. Nevertheless, the traditional theories and algorithms are still employed by existing research work to explore the feasible solutions, which takes a beating from low generalization performance, system load imbalance, more response delay, etc. To solve the matter, a task execution method called DROP (Deep Reinforcement network aided Optimization method aiming at task Processing) has been put forward, which is capable of completing task request allocation through virtual network embedding. The prominence of this method is explained by its effect in reducing load balancing degree, minimizing bandwidth resource overhead, and preserving electric energy as well as meeting customer demands. It makes use of Deep Deterministic Policy Gradient (DDPG) instead of depending on tons of iterations for better path selection schemes in previous methods, through continuous environment interaction and trial-and-error evaluation to get better strategy selection for virtual link embedding. To realize the virtual node embedding in the federated optimization based system architecture, the intentional deep feature learning network is applied. Compared with the cutting edge approaches, the performance benefits of DROP can be verified by the experimental results in terms of bringing down the extra cost on resources and energy of the substrate network during the task execution for big data.

Criteria system construction for sustainable supplier selection: A product-category-oriented intelligent model

In recent years, awareness of environmental protection and social responsibility has increased, making sustainable supply chain development and the selection of sustainable suppliers more important. The construction of a criteria system is the premise of sustainable supplier selection (SSS), which determines the final decision. Recognizing the limitations of existing research works, including subjectively determined evaluation criteria based only on the literature and personal experience, and neglecting the characteristics of different product categories and the combination effect of different criteria, this research designs a product-category-oriented intelligent SSS criteria system construction model by applying the random forest algorithm and the recursive feature elimination cross-validation method. The novelties and contributions of this paper include, first, a systematic SSS criteria system construction model from a new perspective of product category is designed to help buyers construct a comprehensive, reliable, and efficient SSS criteria system. Second, the appropriate integration of human experience and machine learning fully exploits the advantages of both subjective judgement and objective analysis. Third, the ideas and methods of group decision making are introduced for the first time into the SSS criteria system construction process, and the relative importance of different decision-makers are taken into account. Finally, the proposed model can construct the optimal SSS criteria system while considering the combination effect of the whole criteria system. The model was validated in a Chinese forklift company, helping the company to construct a criteria system containing nine criteria with 86.68% accuracy rate. The results of comparative and sensitivity analyses show that the accuracy of the proposed model can achieve an 8.11% improvement compared with other existing methods.

A dynamic holographic modelling method of digital twin scenes for bridge construction

ABSTRACT Holographic projection technology can provide a more intuitive and efficient visualization effect for a digital twin bridge construction scene. However, pre-rendering methods in the existing research work are usually used to implement holographic visualization, which is static display. The above-mentioned methods for static display have many shortcomings, such as poor adaptability, low rendering efficiency and lack of real-time. A dynamic holographic modelling approach is proposed for the augmented visualization of digital twin scenes for bridge construction. Firstly, a dynamic segmentation algorithm with adaptive screen size was designed to high-efficiently generate holographic scenes. Secondly, a motion blur control method was designed to improve the rendering efficiency of holographic scenes according to human visual characteristics. Finally, a prototype system was developed, and the corresponding experimental analysis was completed. The experimental results show that the method proposed in this article can support adaptive screen size image segmentation and real-time generation of holographic scenes for bridge construction. The amount of scene data can be reduced to more than 30%, which significantly improves rendering efficiency and reduces glare.

The random dynamic performances of EMUs with polygonal damages

Polygonalization is a typical non-uniform wear phenomenon that affects the performance and comfort of the vehicle/track system. Therefore, the vehicle system performance could be strictly troubled by polygonal damage close to the wheel circumference at the wheel/rail contact. A 3D vehicle/track an EMU type train consisting of 8 coupled vehicles model has been established. The coupled model car body and bogie frame are considered rigid and integrated with the flexible wheelset, rail, and slab track to investigate the elastic deformation caused by the impact load induced by the wheel polygonalization. The flexible track can be incorporated into the multi-body vehicle model using the movable reference frame formulation. Uncertainty parameters are presented, such as passenger weight, wind force, coefficient friction, and suspension variation. The proposed research is validated based on the existing research work and shows excellent agreement. The simulation result deliberated the vertical wheel-rail force, wheelset acceleration, rail acceleration, rail displacement, and slab track acceleration. The results of this study show that as the wheel polygonal damage amplitude increase, the value of wheel/rail normal forces and the dynamic vibrational response of the wheelset axle increase. Additionally, polygonal wear-induced impact stresses may cause the wheelset axle to resonate. When wheel polygonal damage amplitude and uncertainty parameters are corporate into the system dynamics, the maximum magnitude of the wheelset, rail, and slab track acceleration increases. Besides, the train failure degree is comparable to the polygonised wheel harmonic order, defect parameters, and amplitude of out-of-roundness.

ПРЕДПОСЫЛКИ ФОРМИРОВАНИЯ АРХИТЕКТУРЫ ВРЕМЕННЫХ ЖИЛИЩ ДЛЯ ЗОН ЧРЕЗВЫЧАЙНЫХ СИТУАЦИЙ РОССИЙСКОЙ ФЕДЕРАЦИИ

The article examines the role of temporary shelters in recovery programs for emergency zones of the Russian Federation. The frequency of natural disasters and emergencies is analyzed, as well as a number of problems associated with the need to resettle the affected population in the existing housing stock of settlements located in territories adjacent to the emergency zone or resettlement of the affected to other regions.
 The features of the use of temporary accommodation centers, both stationary and mobile, the possible duration of their use are considered, the provision of the regions of the Russian Federation with them is analyzed. A review of existing research works and developed solutions for mobile housing for temporary stay is carried out on the example of domestic and foreign works.
 The relevance of further research in the issue of determining common problems, patterns and technical solutions in the formation of temporary dwellings for emergency zones of the Russian Federation in the context of further development of principles is substantiated. Goals, boundaries, novelty, tasks of further research are highlighted. For the first time, the question of the possibility of using transforming mobile buildings from rigid elements as temporary housing for emergency zones in the Russian Federation is raised as part of a general study on the formation of principles.

Understanding the class separation process of convolutional neural networks

Extensive research shows that deep neural networks (DNNs) capture better feature representations than the previous handcrafted feature engineering, which leads to a significant performance improvement. However, it raises the question of how does the neural network spontaneously learn the intermediate representation, during the training process, to correctly separate the test data set into different categories? In this paper, based on the existing research work, we continue to take a small step towards understanding the dynamics of convolutional neural networks (CNN). Specifically, we model the test data set as a relationship graph, and each intermediate layer representation learned by CNN will transform the relationship graph correspondingly. Through these continuous transformations, we investigate the class separation process of CNN from two perspectives: 1) show the evolution of the degree of correlation between samples through the histogram of similarity between vertices; 2) by visualizing and quantifying the degree of separation (i.e., modularity) of the important relationship subgraph, we can show the contribution of each convolution layer of CNN to class separation. In the preliminary experiment, it can be found that: 1) each convolution layer of CNN gradually reduced the similarity of most edges between the samples to a very low level, that is, the dense relationship graph gradually evolved into a sparse graph; 2) the remaining important relationship subgraph consist of edges with high similarity shows evident modularity; 3) both visualization and quantification results show that the modularity of intra-class subgraph increases when the layers go deeper. Moreover, the degradation and plateau in the modularity curve reveal the existence of redundant layers. In this paper, we reveal some more in-depth dynamics of CNN, and introduced another modularity, which is widely used as the theoretical guidance tool of layer pruning. It can save more model parameters without losing the accuracy of classification.

The Retrieval and Effect of Core Parameters for Near-Field Inter-Body Coupling Communication

The potential of the Internet of Body (IoB) to support healthcare systems in the future lies in its ability to enable proactive wellness screening through the early detection and prevention of diseases. One promising technology for facilitating IoB applications is near-field inter-body coupling communication (NF-IBCC), which features lower power consumption and higher data security when compared to conventional radio frequency (RF) communication. However, designing efficient transceivers requires a profound understanding of the channel characteristics of NF-IBCC, which remain unclear due to significant differences in the magnitude and passband characteristics of existing research. In response to this problem, this paper clarifies the physical mechanisms of the differences in the magnitude and passband characteristics of NF-IBCC channel characteristics in existing research work through the core parameters that determine the gain of the NF-IBCC system. The core parameters of NF-IBCC are extracted through the combination of transfer functions, finite element simulations, and physical experiments. The core parameters include the inter-body coupling capacitance (CH), the load impedance (ZL), and the capacitance (Cair), coupled by two floating transceiver grounds. The results illustrate that CH, and particularly Cair, primarily determine the gain magnitude. Moreover, ZL mainly determines the passband characteristics of the NF-IBCC system gain. Based on these findings, we propose a simplified equivalent circuit model containing only core parameters, which can accurately capture the gain characteristics of the NF-IBCC system and help to concisely describe the channel characteristics of the system. This work lays a theoretical foundation for developing efficient and reliable NF-IBCC systems that can support IoB for early disease detection and prevention in healthcare applications. The potential benefits of IoB and NF-IBCC technology can, thus, be fully realized by developing optimized transceiver designs based on a comprehensive understanding of the channel characteristics.

Data hiding steganography model based on hyper chaos 2D compressive sensing inhabited with manchester encoder/decoder using circular queue exploiting modification direction

Reversible data hiding (RDH) based on Steganography is considered as one of the future related aspects in the field of security for the information hiding paradigm. Existing research work has been carried out based on secure data transmission as well as reducing the dataloss from one user to other users. But due to encryption data expansion over non-linear transformation, complexity in attacking caused due to keyspace, ineffective image compression, poor embedding ratio, poor quality, overflow/underflow problems, data loss etc., leads to inefficient data transmission causing a security risk. This paper proposes a novel method named Triple Secured Data Hiding Steganography Model which provides solutions to the above challenges. This work is initiated with Hyper Chaos 2D Compressive Sensing that performs image compression and encryption simultaneously. It provides control over low dimension chaos system bearing secure risks with suffering from data encrypted expansion while adopt non-linear transformation. In addition to reduce the error rate and providing signal synchronization as well as system reliability over the transmission channel, Manchester Encoder/Decoder is initiated. To cope up with data embedding and extraction our work has proposed Circular Queue Exploiting Modification Direction(CQEMD). Thus, overall proposed model enhances effective secure data transmission under RDH by inhabiting a triple secured system.

Performance Analysis of Downlink Multisatellite Joint Service System Toward SAGOI-Net

With the rapid development of communication, the future Internet of Things will greatly expand its coverage to form the space, air, ground, and ocean integrated network (SAGOI-Net) and provide globally ubiquitous applications and services. In SAGOI-Net, the existing research work focused on system design and neglected the theoretical analysis of system performance. Furthermore, the Nakagami-m fading model can fit the experimental data better than the Rayleigh and Rice fading models. Also, the Nakagami-m fading model can adapt to the SAGOI-Net environment by changing the value of the shaping parameter m. However, there is no work to analyze the performance of SAGOI-Net using Nakatomi-m fading. Therefore, in this paper, the synchronous downlink system is theoretically analyzed using the Nakagami-m fading model. However, Then, the statistical characteristics of multiple access interference (MAI) and MAI-plus noise are theoretically derived. Furthermore, the accurate expression of the Bit-Error Rate (BER) for the fading model is also derived. After that, we establish the relationship between the number of users, BER, and the information transmission rate to optimize energy efficiency. Finally, simulation results verify that the theoretical derivation results are reliable and effective.

Landweber-type method with uniformly convex constraints under conditional stability assumptions

The Landweber-type iteration method is one of the most prominent regularization methods for solving ill-posed problems when the data is corrupted by noise. By extending the previous results proposed by de Hoop et al. (2012), the current work systematically investigates the convergence rate of Landweber-type iteration in Banach spaces for nonlinear ill-posed problems. Two types of stopping rules including the discrepancy principle and the heuristic Hanke–Raus rule are provided which are novel in view of existing research work.

Abusive Content Detection in Arabic Tweets Using Multi-Task Learning and Transformer-Based Models

Different social media platforms have become increasingly popular in the Arab world in recent years. The increasing use of social media, however, has also led to the emergence of a new challenge in the form of abusive content, including hate speech, offensive language, and abusive language. Existing research work focuses on automatic abusive content detection as a binary classification problem. In addition, the existing research work on the automatic detection task surrounding abusive Arabic content fails to tackle the dialect-specific phenomenon. Consequently, this has led to two important issues in the automatic abusive Arabic content detection task. In this study, we used a multi-aspect annotation schema to tackle the automatic abusive content detection problem in Arabic countries, based on the multi-class classification task and the dialectal Arabic (DA)-specific phenomenon. More precisely, the multi-aspect annotation schema includes five attributes: directness, hostility, target, group, and annotator. We specifically developed a framework to automatically detecting abusive content on Twitter using natural language processing (NLP) techniques. The developed framework used different models of machine learning (ML), deep learning (DL), and pretrained Arabic language models (LMs) using the multi-aspect annotation dataset. In addition, to investigate the impact of the other approaches, such as multi-task learning (MTL), we developed four MTL models built on top of a pretrained DA language model (called MARBERT) and trained on the multi-aspect annotation dataset. Our MTL models and pretrained Arabic LMs enhanced the performance compared to the existing DL model mentioned in the literature.

New fixed-time synchronization results for discontinuous dynamic networks: A novel two-stage analysis approach

This paper concerns the issue of fixed-time synchronization (FXTS) for complex networks composed of dynamical nodes with discontinuous dynamics, abbreviated as discontinuous dynamic networks, under the concept of Filippov solution. In terms of [Formula: see text]-norm and [Formula: see text]-norm, two types of simple control schemes without the linear feedback part are, respectively, designed to compel the considered discontinuous dynamic network to fixed-timely synchronize with a desired trajectory. By utilizing a novel two-stage analysis approach, some new FXTS criteria are derived. Compared with the existing research works, the established theoretical results not only are more relaxed, but also improve the estimation accuracy of synchronous settling time. Finally, the validity of the analytical results is verified via numerical simulations.

A review on the fresh properties, mechanical and durability performance of graphene-based cement composites

In real-world applications, composite materials based on cement that have excellent mechanical and durability properties are always preferred. The research establishment as a whole agrees that concrete, the most widely used building material in the world, must be produced at the nanoscale, where its chemical and physiochemical properties can be greatly enhanced. The microstructure of hydration gel has been reported in the current literature to be greatly improved by the high-surface volume ratio of nanomaterials, including carbon nanotubes (CNTs), nano silica, and nano alumina, which when used as additives in cementitious materials as strengthening additives. Recent advancements in materials science and nanotechnology have resulted in another promising nanomaterial, graphene, which may be employed as a nano-sized addition for concrete mixtures. In terms of mechanical and durability properties, cement matrix mixed with nanomaterials based on graphene performs significantly better due to a significant reduction of micro-cracks owing to its high surface area and tremendous ultimate tensile strength of graphene (130 GPa) and elastic modulus (1.1TPa). In this paper, recent research on graphene-based concrete was critically reviewed. By contrasting the fresh, mechanical, and durability qualities, the performance of cement composite is examined. Additionally, a thorough discussion of the strategies for improving the dispersion characteristics of graphene and its byproducts (edge-oxidized graphene, graphene oxide, and minimized graphene oxide) was conducted.The optimum dosages of graphene/graphene derivatives in concrete with respect to compressive, tensile, flexural strengths, and permeability were articulated with respect to the existing research works. The key obstacles and potential of graphene-based reinforced cement composites will be reviewed in order to give positive insights and suggestions for future investigations.

Iterative heterogeneous graph learning for knowledge graph-based recommendation

Incorporating knowledge graphs into recommendation systems has attracted wide attention in various fields recently. A Knowledge graph contains abundant information with multi-type relations among multi-type nodes. The heterogeneous structure reveals not only the connectivity but also the complementarity between the nodes within a KG, which helps to capture the signal of potential interest of the user. However, existing research works have limited abilities in dealing with the heterogeneous nature of knowledge graphs, resulting in suboptimal recommendation results. In this paper, we propose a new recommendation method based on iterative heterogeneous graph learning on knowledge graphs (HGKR). By treating a knowledge graph as a heterogeneous graph, HGKR achieves more fine-grained modeling of knowledge graphs for recommendation. Specifically, we incorporate the graph neural networks into the message passing and aggregating of entities within a knowledge graph both at the graph and the semantic level. Furthermore, we designed a knowledge–perceiving item filter based on an attention mechanism to capture the user’s potential interest in their historical preferences for the enhancement of recommendation. Extensive experiments conducted on two datasets in the context of two recommendations reveal the excellence of our proposed method, which outperforms other benchmark models.

Performance analysis of video data transmission for telemedicine applications with 5G enabled Internet of Things

Fifth Generation (5G) communication access technology has been implemented to provide highly reliable and efficient video data streaming in telemedicine applications. The Internet of Things (IoT) advancement enhances the 5G network for smart healthcare services and applications. The existing research work focused only on Lagrangian Encoder (LE) based video compression technique with H.265 Protocol for video data transmission in 5G networks for telemedicine applications. This paper proposes a novel KNN classifier-based H.265 protocol with a single buffer model incorporated with multiple sensors for telemedicine applications. The proposed multiple sensors are placed at the transmitter and receiver base stations to exchange data efficiently and accurately between transmitter and receiver devices. The data transmission performance is measured using collision error, propagation error, sensing error, and visual security with encryption for the proposed and existing methods. The performance of the proposed model is compared with the existing LE-based single buffer and identifies the proposed KNN classifier-based single buffer with a multi-sensor technique that performs better.