Through conducting equivalent model training within the paradigm of edge intelligence, the Digital Twin Edge Networks (DITEN) have been widely employed in the Industrial Internet-of-Things (IIoT) to facilitate the cost-effective execution without the operational disruption. However, due to the insufficient consideration of heterogeneity in computing and communication capabilities of distinct industrial terminals in the Digital Twin (DT) model training, the existing approaches of DT construction/update have unbalanced model training cost and loss in the whole life cycle of DT model, hindering the abilities of quick responding to complex and dynamic productions and ensuring the data consistency of virtual-real space. To address this issue, we define a global loss minimization problem with constraint, and propose an original approach of semi-asynchronous federated learning, named EDT-SaFL, as a promising solution. Considering the collaborative utilization of heterogeneous resources, and the contribution of local data quantity and quality to the global model update, the EDT-SaFL consists of three important operations,Terminal Selection for Model Training,Self-Adaptation of Local Training Iterations, andSemi-asynchronous Global Aggregation. With the analysis of convergence, complexity and communication overhead, the experiments have evidently demonstrated the superiority of EDT-SaFL on the datasets of CIFAR-10 and Industrial-Equipment.

GNSS/INS integrated navigation positioning methods are widely used in vehicle positioning services. However, high-accuracy and reliable vehicle localization remains a challenge under GNSS signal degradation or INS drift. We propose PUKF, a novel fusion algorithm combining Nonlinear Predictive Filtering (NPF) with Unscented Kalman Filtering (UKF), to address model error-induced degradation in GNSS/INS integrated navigation. PUKF introduces real-time model error estimation into the prediction phase of UKF to dynamically adapt to complex environments. Preliminary results from simulation and KITTI dataset-based experiments show that PUKF outperforms the traditional UKF in high-dynamic environments, especially when model errors are large, significantly enhancing positioning accuracy and system stability under abrupt dynamics and low-cost sensors. The proposed method has promising applications in realtime web-based vehicle services.

Gait-based WiFi user identification enables devicefree recognition by analyzing signal variations caused by human motion. However, existing methods face performance drops in new environments and cannot detect unauthorized users. This paper proposes GIWiD (Gait-based user Identification with WiFi Device-free sensing), which employs adversarial learning to extract domain-invariant features and enhance cross-domain generalization. A two-stage training strategy is adopted: the first stage performs identity recognition, and the second detects unauthorized users via reconstruction errors. GIWiD uses passive WiFi sensing to preserve privacy and ensure accuracy. Experiments on data from 13 volunteers across diverse indoor settings show that GIWiD outperforms existing methods in both crossdomain identification and unauthorized user detection.

High-quality data and effective feature representation are critical for tumor image detection. However, existing tumor detection models are limited in relying on large-scale annotated datasets, ineffective feature extraction, and ignoring local features. This study presents an enhanced feature representation approach with contrastive learning and the multiscale filter, termed CL-MFTD, to achieve accurate tumor detection. This model realizes the pre-training via the contrastive learning framework, which extracts and compares the target and predicted features generated from different augmented views to obtain the optimal pre-trained model. Then, we investigate a multiscale adaptive Wiener filter that dynamically adjusts noise suppression strength based on local variance estimation within wavelet subbands, which overcomes the limitations of conventional single-scale Wiener filters by preserving edge structures while suppressing modality-specific noise in medical images. Finally, based on the optimal pre-trained model, a detection model is reconstructed with Backbone, Neck, and Head modules to extract effective features from the denoised dataset. We design the experiment of model comparisons and ablation tests on four evaluation metrics to verify the proposed CL-MFTD model’s performance. The results demonstrate that the CL-MFTD model outperforms other baselines and achieves satisfactory results on small-scale medical datasets.

With the rapid advancement of 6G, the task offloading has emerged as a critical issue for enhancing computational efficiency in the Industrial Internet of Things (IIoT). However, industrial devices are often constrained by computing power, energy and mobility, challenging the delay-sensitive and compute-intensive tasks consisting of subtasks with complex dependencies, e.g., industrial image recognition. Given the increasing task complexity, developing efficient offloading strategies in dynamic multislot industrial scenarios with mobile devices remains a challenge. To address this issue, a task offloading scheme for industrial image recognition in digital twin edge networks (DITENs) is proposed. By leveraging the digital twin (DT) to accurately model the states of mobile industrial tasks and edge servers, the task offloading is formulated to optimize the weighted sum of task processing delay and energy consumption, that is proven to be NP-hard. Since the task of image recognition can be represented as a directed acyclic graph (DAG), the dependencies between subtasks are extracted using graph convolutional network (GCN) to generate optimal execution priorities for task offloading. Through proving the optimization problem as a Markov decision process (MDP), an improved multiagent deep deterministic policy gradient algorithm, named$\epsilon $-ATN-MADDPG, incorporating the$\epsilon $-greedy strategy and the self-attention mechanism to enable efficient decision-making in dynamic environments, is designed to offer a promising solution. Experimental results on the KolektorSDD dataset demonstrate that this solution outperforms compared methods.

The Internet of Everything (IoE) has recently become a hot topic. With the development of Internet of Things (IoT) technology, people can connect to networks in increasingly diverse ways. The surge in users, devices, and requests poses significant challenges to network capacity and backhaul links. Content caching technology has long been considered a promising approach to improving network performance. However, existing methods still have room for improvement in terms of content transmission efficiency and user access latency. To address these issues, this paper proposes an Influence-Driven Content Caching (IDCC) method. Specifically, based on a caching strategy of “caching content that is likely to have the greatest future influence on the most influential edge devices", this paper designs a comprehensive framework encompassing content selection, updating, and placement to optimize content caching efficiency, enhance network spectral efficiency, and improve user’s quality of experience (QoE). First, a content selection strategy based on the popularity dynamics prediction method is developed by utilizing graph neural networks and contrastive learning to model heterogeneous data. Second, a content update mechanism for cached content and key caching information is designed based on the popularity of content and Near-Field Communications (NFC) between users. Furthermore, interconnected network devices are represented as a graph, and the communication influence of key network nodes is predicted using autoencoders and graph neural networks to identify the optimal caching nodes for maximizing benefits. Finally, extensive experiments show that the proposed IDCC method offers significant advantages in reducing network latency and improving network utilization.

The sixth-generation (6 G) networks aim to achieve ubiquitous intelligent connectivity while ensuring extremely low latency, reducing energy consumption, and enhancing privacy protection. Mobile edge computing (MEC) offers an effective solution to reduce latency and energy consumption by leveraging resources near end devices for task offloading. However, MEC faces significant challenges in meeting the requirements of 6 G networks, including limited computational resources, high mobility, and strict data privacy demands. Efficiently allocating edge resources while preserving privacy has become a critical issue for realizing the objectives of 6 G networks. In this paper, we propose a privacy-preserving edge computing power network (EdgeCPN) model that jointly leverages the computing resources of edge computing nodes and protects sensitive computing power information through differential privacy methods. In addition, we propose a task matching problem that aims to minimize the privacy-budget-weighted energy consumption while ensuring privacy protection and meeting task requirements. We propose a dynamic graph-based multiagent reinforcement learning (MADRL) algorithm to find the optimal strategy for task matching and computing resource allocation with privacy protection. The results show that our proposed task matching model with energy and privacy tradeoffs can minimize the energy consumption in the matching process while ensuring privacy, and the algorithm can find the optimal strategy for task matching efficiently.

In 6G-enabled vehicle-to-edge networks, through deploying computing power resources closer to mobile vehicles for providing low latency and highly reliable services, Mobile Edge Computing (MEC) as an emerging paradigm has promoting mobile vehicles with limited capacities to come with diversified artificial intelligence (AI) applications. Nevertheless, the computing power of MEC server is still finite, seeking the optimal resource pricing and allocation strategies for MEC servers, and determining the optimal computation offloading for intelligent vehicle applications, still remain challenging issues that are necessary to be reasonably solved to improve the service experience. To address this issue, a scenario that multiple intelligent vehicle applications cooperatively initiate computation offloading requests in 6G-enabled multi-server multi-access vehicle-to-edge computing systems is considered in this paper, and with the defined reasonable utilities for MEC servers and intelligent vehicle applications, a solution of bidding-enabled dynamic resource pricing for computation offloading is proposed. Concretely, through considering the relationship of resource supply and demand, and the bidding among MEC servers, a dynamic resource pricing scheme is designed for MEC servers, meanwhile, with the complete consideration of dynamic resource pricing and time-varying wireless channel interference in multi-cell networks, a Q-learning based offloading decision algorithm is proposed for intelligent vehicle applications. Simulation experiments finally are conducted to demonstrate the efficiency in achieving the win-win situation with guaranteed utilities for both MEC servers and intelligent vehicle applications.

The spotlight on autonomous aerial vehicles (AAVs) is to enhance wireless communications while ignoring the potential risk of AAVs acting as adversaries. Due to their mobility and flexibility, AAV eavesdroppers pose an immeasurable threat to legitimate wireless transmissions. However, the existing fixed jamming scheme without cooperation cannot counter the flexible and dynamic AAV eavesdropping. In this article, a cooperative intelligent jamming scheme is proposed, authorizing ground jammers (GJs) to interfere with AAV eavesdroppers, generating specific jamming shields between AAV eavesdroppers and legitimate users. Toward this end, we formulate a secrecy capacity maximization problem and model the problem as a decentralized partially observable Markov decision process (Dec-POMDP). To address the challenge of the huge state space and action space with network dynamics, we leverage a deep reinforcement learning (DRL) algorithm with a dueling network and double-Q learning (i.e., dueling double deep Q-network) to train policy networks. Then, we propose a multi-agent mixing network framework (QMIX)-based collaborative jamming algorithm to enable GJs to independently make decisions without sharing local information. Additionally, we perform extensive simulations to validate the superiority of our proposed scheme and present useful insights into practical implementation by elucidating the relationship between the deployment settings of GJs and the instantaneous secrecy capacity.

Information diffusion prediction is a complex task due to the dynamic of information substitution present in large social platforms, such as Weibo and Twitter. This task can be divided into two levels: the macroscopic popularity prediction and the microscopic information diffusion prediction (who is next), which share the essence of modeling the dynamic spread of information. While many researchers have focused on the internal influence of individual cascades, they often overlook other influential factors that affect information diffusion, such as competition and cooperation among information, the attractiveness of information to users, and the potential impact of content anticipation on further diffusion. To address this issue, we propose a multiscale information diffusion prediction with minimal substitution (MIDPMS) neural network. This model simultaneously enables macroscale popularity prediction and microscale diffusion prediction. Specifically, information diffusion is modeled as a substitution system among different information. First, the life cycle of content, user preferences, and potential content anticipation are considered in this system. Second, a minimal-substitution-theory-based neural network is first proposed to model this substitution system to facilitate joint training of macroscopic and microscopic diffusion prediction. Finally, extensive experiments are conducted on Weibo and Twitter datasets to validate the performance of our proposed model on multiscale tasks. The results confirmed that the proposed model performed well on both multiscale tasks on Weibo and Twitter.

One of the key challenges of current online social platforms is predicting the size of information cascades, also known as popularity prediction or cascade prediction. Accurate popularity prediction can benefit various fields, including news distribution, market decisions, and rumor detection. However, existing popularity prediction approaches concentrate more on the historical sequences of single messages, overlooking the interactions between message diffusion and the dynamic nature of social networks, which limits the timeliness and accuracy of predictions. To address this, we propose an online popularity prediction service based on minimal substitution reinforcement learning calledMSRL. Specifically, we explore a substitution theory and design a minimal substitution reinforcement learning method that models diffusion as message substitution and considers mutual information diffusion. That helps the model gain a broader perspective, allowing it to fully exploit the cooperative, competitive, or dependent relationships between information diffusions. Furthermore, the reinforcement learning scheme enables the service to dynamically adjust its parameters to respond to the dynamic social network environment in real-time. Finally, extensive experiments on real-world datasets show that the MSRL outperforms state-of-the-art methods regarding accuracy and service agility.

In social media platforms such as Weibo, Twitter, and Facebook, a variety of information is diffused daily. Exploring and exploiting the diffusion patterns in this information play crucial roles in areas such as viral marketing, recommendation systems, and public opinion management. However, the diffusion of this information is not merely sequential propagation among users, as most researchers assume. When we observe the diffusion of information in the entire network from a macroscopic perspective, we discover that these phenomena of information diffusion exhibit a series of interconnected relationships, such as alternation or dependency. In traditional methods of information diffusion prediction (IDP), these aspects are often overlooked. To address this, we introduce a substitution theory of information diffusion, minimal substitution (MS), and we combine it with neural networks to design a network model known as MSNN. First, the incorporation of MS theory enables our model to effectively capture the complex interrelations among pieces of information. Second, we analyze the relationship of the multi-scale IDP task, develop a one-step MS-based microscopic IDP method and a dynamic MS-based macroscopic IDP method, and utilize these two methods for joint training to achieve multi-scale prediction. Finally, we validate the accuracy of the proposed MSNN model through training on two real-world datasets with different growth patterns.

With the global epidemic of the COVID-19, various rumors spread wantonly on social networks, which has seriously affected the stability and harmony of the entire society. To purify the network environment, some researchers have proposed to fight rumors from the perspectives of tracing the source of rumors, detecting the authenticity of information, and predicting explosive fake news. But their works are fragmented, and their performance are not significant. So we need strong antirumor methods to fight rumors. To this end, this article proposes a more comprehensive antirumor mechanism, which can realize rumors source location, rumor detection, and popularity prediction (RLDP). In particular, in the task of localization, we propose graph neural network-based method, which does not need to specify the underlying propagation mode and the number of rumor sources; in the task of detection, utilizing lightGBM, we construct a rumor detection model; in the task of popularity prediction, we construct a model based on contrastive learning while considering user engagements and information propagation, and the text of rumor. Finally, we verify the performance of the proposed RLDP by conducting extensive experiments.

How to localize the rumor source is a common interest of all sectors of the society. Many researchers have tried to use deep-learning-based graph models to detect rumor sources, but they have neglected how to train their deep-learning-based graph models in thenoisysocial network environmentefficiently. Especially for deep learning models, the performance relies on the data scale. However, even though its known that a substantial amount of rumor data distributed across multiple edge servers (e.g., cross-platform), due to conflicting business interests, its challenging to coordinate all parties to train a model driven by many samples while avoiding moving data. Federated learning, is an effective technique to bridge this gap. Therefore, this paper proposes aDistributedRumorSourceDetection viaBoostedFederatedLearning (DRSDBFL). Specifically, this paper proposes an effective rumor source detection method based on a deep-learning-based graph model with a denoising module. To the best of our knowledge, we are the first to attempt to the use of a denoising module to reduce the noisy effects of social networks. Then, we propose a novel boosted federated learning mechanism through boosting the high-quality edge worker to improve the training efficiency. Finally, the effectiveness of the proposed method is verified by extensive experiments.

Fabric metaverse employs intelligence fibers embedded with flexible sensors to unknowingly gather and transmit massive hypermodal data around humans to a deep neural network-based metaverse inference service (DMS) for continual and real-time analysis. Each DMS has one primary branch and multiple side branches that allow early termination of service with differential accuracy and energy consumption. However, the continual provisioning of compute-intensive DMS with varying requirements for service model, accuracy, delay, and reliability poses a challenge for edge servers characterized by restricted computing resources and intermittent green energy. In this paper, we focus on a continual individualized DMS provisioning problem in the fabric metaverse consisting of a side branch insertion subproblem and a server activation and service deployment subproblem, and formulate them as Integer linear Programming and Markov Decision Process, respectively. Then, we propose a green continual inference (GCI) system, where a pruner with provable approximation ratios trims superfluous branches of every model to the given number$K$to minimize total overflow accuracy between accuracy demands and reserved branches assigned to users. Based on this exit result, each DMS is further divided into several blocks with dependencies to exploit constrained resources of computing and energy in a fine-grained manner. Finally, a learning-based scheduler is merged into GCI to maximize request throughput while minimizing the activation number of edge servers on different demand scenarios, by adaptively activating suitable servers and deploying required blocks and their corresponding backups on selected servers. Theoretical analyses, simulations, and experiments demonstrate that the GCI is promising compared with baseline algorithms.

Content caching in mobile networks is a highly promising technology for reducing traffic load latency and energy consumption levels. Its fundamental goal is to satisfy the supply-and-demand relationships between content providers and content-requesting users. However, previous research primarily focused on the optimization goals of mobile network operators, and although these caching strategies yield improved latency and energy consumption levels, they fall short of satisfying the diverse content demands of users in real-world scenarios. Therefore, this paper proposes a diversity-driven proactive caching strategy that considers multiple stakeholders' requirements, in which the cache hit rate, cache gain for network operators, and content diversity are jointly optimized. Specifically, a novel improved Latent Dirichlet Allocation (LDA) is designed for radio access network caching, which enables diverse topic associations. For user device caching at the network edge, a Gradient-Guided Contrastive Learning (GGCL) is proposed to optimize the multiple objectives of cache systems with limited labeled data resources. Finally, extensive experiments conducted on the MovieLens dataset demonstrate that the proposed method significantly outperforms other methods in various aspects, including content diversity, the cache hit rate, and some network performance metrics, such as the traffic load and cache gain.

Given an input image, scene graph generation (SGG) aims to generate comprehensive visual relationships between objects in the form of graphs. Recently, more attention to the design of complex networks and complicated strategies has been paid to the long tail issue caused by the imbalanced class distribution. However, most existing methods adopt the concatenated features of two objects in real space as the final relation representation for a given triplet. We mainly argue that such a simple concatenation may neglect the importance of complex interactions between objects, which results in the diversity of visual relations. In addition, the representation learning in real space is also inadequate to express this property. To alleviate these issues, we seamlessly incorporate Hermitian inner product into existing models to facilitate the generation of scene graphs by learning Relation Embedding in Complex space (CoRE). More specifically, we first introduce the concept of complex-valued representations for entities and then formulate the relation triplets with Hermitian inner product in complex space. Finally, we investigate the effect of utilizing only real component or both of Hermitian inner product on inferring more reasonable interaction between objects for scene graphs. Comprehensive experiments on two widely used benchmark datasets, Visual Genome (VG) and Open Image, demonstrate our effectiveness, superiority, and generalization on various metrics for biased or unbiased inference.

Supervising illicit activities on blockchain networks, such as money laundering, fraud, extortion, Ponzi schemes, and funding for terrorist organizations, presents significant challenges. Emerging machine learning methods for detecting abnormal transactions face hurdles due to high labeling costs, limited labeled data, and data imbalance. To address this, this paper proposes aMinimalSubstitution-basedLabelPropagation(MSLP) model to provide more labeled data to balance the graph data and complement the sample for anomalous transaction detection service in the blockchain networks. As far as we know, MSLP is the first method that utilizes the minimal substitution theory from the social computing field to find more abnormal transactions with under-labeling budget constraints. This approach has the potential to obtain more high-quality labeled data with minimal computational cost by utilizing a small amount of labeled graph data. Then, a label evaluation mechanism is proposed to decide the number of samples to be adopted for each class, ensuring the performance of downstream graph neural networks. Finally, extensive experiments were conducted and the proposed model improved the F1 score of illegal transaction node detection by 2.6% to 8.2%.

Due to the significant Influence of social networks, information diffusion prediction, which aims to study the spread of messages among users, has become a crucial objective in various scenarios. Existing works have mostly attempted to integrate multiple source features such as social relationships and user preferences for prediction. Although it can improve the accuracy of predictions to some extent, there are still issues with complex models, low computational efficiency, and limited prediction performance. To address these challenges, this paper proposes a Gaussian-Noise-based Information Diffusion Prediction Model (GNIDP). As we know, we are the first to model the information diffusion in the social network as Gaussian noise diffusion, specifically, GNIDP utilizes Gaussian noise to simulate information diffusion patterns during information propagation, while capturing the evolving diffusion patterns over time across different time slots. Consequently, our model does not rely on social topology to achieve prediction, while improving the efficiency and accuracy of predictions. Experiments on two real social network datasets demonstrates that GNIDP achieves 13.54% relative gains over the best baseline on average. Therefore, the proposed model can be applied in scenarios where social topology is missing or difficult to obtain particularly in large-scale datasets requiring efficient diffusion prediction.

Due to the lack of centralized regulatory authorities, the cryptocurrency trading market has witnessed an increase in illicit activities. Recently, more and more researchers have started exploring the application of machine learning techniques to achieve anomaly detection in these transaction networks. However, in this transaction network, it’s hard to find the abnormal nodes because the labeling cost is high, without enough labeled and imbalanced data, which greatly limits the performance of the detection model. Therefore, this paper proposes a Minimal Substitution-based Label Propagation (MSLP) model to provide more labeled data to balance the graph data and complement the sample for downstream anomalous transaction detection in the blockchain networks. Specifically, MSLP is the first method that utilizes the minimal substitution theory from the social computing field to find more minority nodes from the unlabeled nodes. This approach has the potential to obtain more high-quality labeled data with minimal computational cost by utilizing a small amount of labeled graph data. Then, a label evaluation mechanism is proposed to decide the number of samples to be adopted for each class, ensuring the performance of downstream graph neural networks. Finally, extensive experiments were conducted and the proposed model improve the F1 score of illegal transaction node identification by 2.6% to 8.2%.

In Internet of Things (IoT)-based e-Health Systems (IoTEHS), medical devices form a large network that continuously sense and share the healthcare data with the nearby edge devices or cloud servers. The health data is subsequently made available to various IoTEHS stakeholders (such as doctors, nurses and patients) to track and monitor patients under observation. However, the entire IoTEHS stakeholders communicate with each other over a wireless unsecured public communication channel. This is a major security and privacy loophole wherein the attacker can exploit the vulnerability of the system and can launch various attacks on the ongoing communication. Motivated by the aforementioned challenges, a secure data dissemination scheme using AI and blockchain is proposed. In this scheme, the transaction collected through healthcare sensors installed around the patients premises act as data sets that is forwarded to the nearby edge devices. The collected data is first filtered using AI-based intrusion detection system located at the edge of the network. Second, a secure health monitoring network is designed using blockchain. Specifically, the filtered or normal transactions are transmitted to centralized cloud servers where the smart contact-enabled consensus mechanism is used to validate the transactions. Once the transaction gets validated, it is stored on distributed InterPlanetary File System (IPFS) of cloud and returned transaction hash is stored on the blockchain ledger located at edge devices making data exchange faster. The detailed experimental investigation demonstrates that the proposed schemes are efficient (in terms of computing and processing time) as well as its resistance to a variety of security attacks.

With the rise of the Industrial Internet of Things (IIoT), more and more industrial devices can be connected via the network. Data collection, processing, analysis, task execution, and other devices that can product network traffic volume are gradually being deployed to IIoT. However, under the limited spectrum resources and low-cost and low-energy production requirements of enterprises, how to ensure the interconnection and intercommunication of industrial networks while realizing the effective use of network communication resources is currently a hot topic. Among them, network traffic prediction is considered to be a very important task. The time variability and interpretability, especially the time-varying features of traffic sequences, greatly challenge this task. To address those, this article proposes a method calledFlow2graphto predict network traffic in IIoT. Specifically, some key segments, i.e., shapelets are extracted from the network traffic sequence according to time-varying traffic; then uses the relationship between the traffic sequence and shapelets to convert the flow into a shapelets conversion graph; Subsequently, the graph isomorphism network are used to learn the specificity of the flow sequence from different devices, thereby to predict its traffic value for a period of time in the future; finally, we conduct extensive experiments on real data to verify the effectiveness of the proposed method.

In the Cognitive Internet of Vehicles (CIoV), vehicles, road side units (RSU) and other key nodes have been equipped with more and more software to support intelligent transportation system (ITS), vehicle automatic control and intelligent road information services. Additionally, technological innovation forces the software in the CIoV to update and upgrade in time. However, escalation is critical to the safety, stability, and maintenance cost of transportation systems. It can be assumed that when the intelligent services supporting CIoV can realize self-perception and escalation, the cognitive ability and coordination ability of the entire CIoV will be greatly improved. To address this, we first propose a deep learning-based method for Software Escalation Prediction (SEP) in CIoV. Specifically, the pretraining mechanism of transformers in the field of natural language processing is combined with software upgrade-related events to dynamically model software sequence activities. To capture the event association in the software activities, we use graph modeling software’s state log and utilize a graph neural network (GNN) to learn the complex life activity rule of software. Finally, the above characteristics are deeply integrated. The proposed method has a 6%–8% improvement over the RoBERTa methods.

No abstract available.

Nowadays, artificial intelligence (AI) provides tremendous prospects for driving future healthcare while empowering patients and service providers. The extensive use of digital healthcare produces a massive amount of multimedia healthcare data continuously (e.g., MRI, X-Ray, ultrasound images, etc.). Hence, it needs special data analytics techniques to provide a smart diagnosis to the patients. Recent advancements in artificial intelligence and machine learning techniques, particularly Deep learning (DL) methods, have demonstrated tremendous medical diagnosis progress and achievements. Diabetic Retinopathy (DR), cataract, macular degeneration, and glaucoma are the most common eye problems due to diabetes. Numerous models have been proposed using deep learning models to diagnose diabetic retinopathy, but no model is perfect for detecting DR diseases. This article presents a deep learning model to analyze diabetic retinopathy images to classify DR patients’ severity levels. The model applies a custom-weighted loss function in the model’s training and achieves 92.49% accuracy and a 0.945 Cohen Kappa score on test data. The model’s weighted average precision was 93%, recall 92%, and f1 score 93%. The model is compared with several state-of-the-art pre-trained models. We observe that the proposed model performs better in accuracy results and Cohen Kappa score.

Aspect-based sentiment classification aims to predict the sentiment polarity of an aspect term in a sentence. It has been verified that syntactic dependent trees, especially integrating with graph convolutional networks (GCN) can provide crucial syntactic features for sentiment classification. However, it can not consider the dependency label information between the aspect and context in the sentence, which contains rich semantic information. In this paper, we propose a bidirectional edge-enhanced graph convolutional networks (BE-GCN), which combines the syntactic structure and the dependency label information effectively. Specifically, we design an edge-enhanced module to dynamically update the dependency edge according to dependency relation and contextual information. Then the dependency edge can update the word representation reversely, thereby selectively outputting sentiment features according to the given aspect. Comparison experiments demonstrate the effectiveness of our model using dependency label information and syntactic structure.

Summary Analyzing the correlation between two funds can help investors control investment risks and optimize investment portfolios, which has a strong guiding significance for fund investment in reality. Constructing an intelligent investment system with fund correlation analysis capabilities can help investors automatically make profits from financial markets. In previous research, many researchers have built intelligent investment systems using Bayesian networks, support vector machines (SVM), and LSTM models. However, the strong historical dependence between fund data and the high‐dimensional and high‐noise characteristics of fund data prevent traditional methods from obtaining excellent performance in fund analysis. This paper designs a deep learning‐based fund intelligent trading system‐DLIFT which has functions such as investment push, income prediction, and risk control. The systems data analysis module is implemented using the Improved RNN model. This model employed encoder‐decoder architecture. The encoder is responsible for analyzing the fund's feature, and the decoder is responsible for analyzing the dependency relationship between the historical correlation and the current correlation. LSTM and an attention mechanism are simultaneously applied to the encoder and decoder, which enabled the discovery of the implicit dependence of time series data. This article places the designed system on a historical dataset containing multiple public funds for verification. In specific experiments, the experimental results of the comparative experiments show the superiority of our model. At the same time, the results of the ablation experiment results show that LSTM and attention mechanism play critical role in the proposed system.

The development of the smart medical industry and equipment has made great progress due to the fusion of the IoT, cloud computing, and big data. In smart IoT healthcare, patients can collect vital parameters from various medical sensors attached to them to detect diseases and make initial diagnoses by themselves. With the powerful storage and computing functions of cloud computing, medical sensor devices deployed in a cloud environment can effectively solve the problems that the devices are highly dispersed, heterogeneous, and have limited processing capabilities. As a result, this method can effectively provide customized and scalable smart medical services for patients. However, because these medical resources share computing resources on the cloud platform, changes in equipment workloads will lead to service performance competition among tenants. Therefore, determining how to achieve performance isolation between tenants and guarantee the service-level agreements (SLAs) of the tenants has become the most concerning issue for cloud service providers. In this article, we propose a performance isolation algorithm for multitenant IoT clouds, which can effectively provide performance isolation between tenants. Experiments show that the message processing delay of tenants working within the allocated quota can be reduced by 82%.

The unified management of IoT devices with interoperability can be inspired by cloud computing. In addition, sinking the 5G core network to the edge brings chances for the deployment of end-to-end ultralow-latency services. However, the resource efficiency brought by heterogeneous computing devices in 5G spectrum multiplexing environments has encountered challenges. To discuss this issue from a comprehensive perspective, this article first proposes an ultralow-latency service deployment architecture in 5G heterogeneous networks, and three cognitive engines are the key components for efficient service communication across the terminal/edge/cloud computing structure. Then we give an analysis of application task model in the proposed architecture, and following the service response time models are established. In addition, it is efficient to deploy multiuser tasks with constraint resources when the differentiated user requirements are met. Finally, we conducted some experiments and the result statistics are up to our expectations. The first one is the system performance under two microcloud covered cells, and the second one is the performance comparison of the proposed solution with three single scenes of terminal computing, edge computing and cloud computing.

At present, computed tomography (CT) is widely used to assist disease diagnosis. Especially, computer aided diagnosis (CAD) based on artificial intelligence (AI) recently exhibits its importance in intelligent healthcare. However, it is a great challenge to establish an adequate labeled dataset for CT analysis assistance, due to the privacy and security issues. Therefore, this paper proposes a convolutional autoencoder deep learning framework to support unsupervised image features learning for lung nodule through unlabeled data, which only needs a small amount of labeled data for efficient feature learning. Through comprehensive experiments, it shows that the proposed scheme is superior to other approaches, which effectively solves the intrinsic labor-intensive problem during artificial image labeling. Moreover, it verifies that the proposed convolutional autoencoder approach can be extended for similarity measurement of lung nodules images. Especially, the features extracted through unsupervised learning are also applicable in other related scenarios.

Future industrial cyber-physical system (CPS) devices are expected to request a large amount of delay-sensitive services that need to be processed at the edge of a network. Due to limited resources, service placement at the edge of the cloud has attracted significant attention. Although there are many methods of design schemes, the service placement problem in industrial CPS has not been well studied. Furthermore, none of existing schemes can optimize service placement, workload scheduling, and resource allocation under uncertain service demands. To address these issues, we first formulate a joint optimization problem of service placement, workload scheduling, and resource allocation in order to minimize service response delay. We then propose an improved deep Q-network (DQN)-based service placement algorithm. The proposed algorithm can achieve an optimal resource allocation by means of convex optimization where the service placement and workload scheduling decisions are assisted by means of DQN technology. The experimental results verify that the proposed algorithm, compared with existing algorithms, can reduce the average service response time by 8-10%.

Currently, advanced communications and networks greatly enhance user experiences and have a major impact on all aspects of people's lifestyles in terms of work, society, and the economy. However improving competitiveness and sustainable vehicle network services, such as higher user experience, considerable resource utilization and effective personalized services, is a great challenge. Addressing these issues, this paper proposes a virtual network resource management based on user behavior to further optimize the existing vehicle communications. In particular, ensemble learning is implemented in the proposed scheme to predict the user's voice call duration and traffic usage for supporting user-centric mobile services optimization. Sufficient experiments show that the proposed scheme can significantly improve the quality of services and experiences and that it provides a novel idea for optimizing vehicle networks.

In the intelligent transportation system, the session data usually represents the users' demand. However, the traditional approaches only focus on the sequence information or the last item clicked by the user, which cannot fully represent user preferences. To address this issue, this paper proposes an Multi-aspect Aware Session-based Recommendation (MASR) model for intelligent transportation services, which comprehensively considers the user's personalized behavior from multiple aspects. In addition, it developed a concise and efficient transformer-style self-attention to analyze the sequence information of the current session, for accurately grasping the user's intention. Finally, the experimental results show that MASR is available to improve user satisfaction with more accurate and rapid recommendations, and reduce the number of user operations to decrease the safety risk during the transportation service.

Autism spectrum disorder (ASD) is a serious hazard to the physical and mental health of children, which limits the social activities of patients throughout their lives and places a heavy burden on families and society. The developments of communication techniques and artificial intelligence (AI) have provided new potential methods for the treatment of autism. The existing treatment systems based on AI for children with ASD focus on detecting health status and developing social skills. However, the contradiction between the terminal interaction capability and availability cannot meet the needs for real application scenarios. At the same time, the lack of diverse data cannot provide individualized care for autistic children. To explore this robot-based approach, a novel AI-based first-view-robot architecture is proposed in this article. By providing care from the first-person perspective, the proposed wearable robot overcomes the difficulty of the absence of cognitive ability in the third-view of traditional robotics and improves the social interaction ability of children with ASD. The first-view-robot architecture meets the requirements of dynamic, individualized, and highly immersed interaction services for autistic children. First, the multi-modal and multi-scene data collection processes of standard, static, and dynamic datasets are introduced in detail. Then, to comprehensively evaluate the learning ability of children with ASD through mental states and external performances, a learning assessment model with emotion correction is proposed. Besides, a wearable robot-assisted environment perception and expression enhancement mechanism for children with ASD is realized by reinforcement learning, which can be adapted to interactive environments with optimal action policies. An interactive testbed for children with ASD treatments is demonstrated and experimental cases for test subjects are presented. Last, three open issues are discussed from data processing, robot designing, and service responding perspectives.

With the spread of COVID-19 worldwide, people¡¯s production and life have been significantly affected. Artificial intelligence and big data technologies have been vigorously developed in recent years. It is very significant to use data science and technology to help humans in a timely and accurate manner to prevent and control the development of the epidemic, maintain social stability and assess the impact of the epidemic. This paper explores how data science can play a role from the perspectives of epidemiology, social networking, and economics. In particular, for the existing epidemic model SIR, we present a parameter learning method using particle swarm optimization (PSO) and the least squares method, and use it to predict the trend of the epidemic. Aiming at the social network data, we provide a specific method to realize sentiment analysis during the epidemic and propose an explainable fake news detection technique based on a variety of data mining methods.

The interactive robot is expected to support emotion analysis and utilize the deep learning and machine learning to provide users with continuous emotional care. However, it is a great challenge to securely acquire sufficient data for emotion analysis such that the privacy of emotional data is adequately protected. To address the security issue, this paper proposes a security policy based on identity authentication and access control to ensure the security certificate through an interactive robot or edge devices while the access control of private data stored in the edge cloud is adequately protected. Specifically, this paper adopts a polynomial-based access control policy and designs a secure and effective access control scheme. At the same time, this paper puts forward an identity authentication mechanism in view of edge cloud systems, which can reduce the computational overhead and authentication delay in a collaborative authentication of multiple edge clouds. The effectiveness of the proposed access control policy and identity authentication mechanism is verified by an actual testbed platform.

As the occurrence of cloud computing, the exponential growth of various application services results in the urgent demand for green computing and resource sustainability on the premise of guaranteeing service performance. Especially, Cloud Radio Access Network (CRAN) has been recognized as a promising approach to provide smart computing and sustainable resource usage for fulfilling the increasing traffic demand. Moreover, the secure network environment is also a vital element for achieving reliable application services. In this paper, we propose a cluster-based secure Cloud Radio Access Network (CSC-RAN), which optimizes the trade-off between performance and resource utilization to satisfy the requirements of the sustainable green network. Based on the powerful ability of cloud computing, the abundant network resource can be dynamically allocated according to the varying traffic. A traffic-aware RRHs cluster formation (TRCF) algorithm is proposed for realizing efficient resource utilization while improving the Quality of Service(QoS). Furthermore, a spectrum allocation genetic algorithm (SAGA) is introduced to solve the optimal spectrum allocation problem for the formatted cluster, which is proved to be a mixed-integer programming problem. Finally, the effectiveness of the TRCF and SAGA algorithms are verified through a series of numerical simulations.

Human wearable helmet is a useful tool for monitoring the status of miners in the mining industry. However, there is little research regarding human emotion recognition in an extreme environment. In this paper, an emotional state evoked paradigm is designed to identify the brain area where the emotion feature is most evident. Next, the correct electrode position is determined for the collection of the negative emotion by the electroencephalograph (EEG) based on the international 10-20 system of electrode placement. And then, a fusion algorithm of the anxiety level is proposed to evaluate the person's mental state using the θ, α, and β rhythms of an EEG. Experiments demonstrate that the position Fp2 is the best electrode position for obtaining the anxiety level parameter. The most visible EEG changes appear within the first two seconds following stimulation. The amplitudes of the θ rhythm increase most significantly in the negative emotional state.

The advances in information technology have facilitated great progress in healthcare technologies in various domains. However, these new technologies have also made healthcare data not only much larger in size but also substantially more difficult to handle and process. Moreover, because the data are created from a variety of devices within a short time span, these data are stored in different formats and created quickly, which can to a large extent be regarded as a big data problem. This paper discusses how to develop intelligent patientcentric healthcare applications and services from the perspectives of mobile computing and big data analytics technologies. This healthcare system consists of a data collection layer with a unified standard, a data management layer for distributed storage and parallel computing, and a data-oriented service layer. Furthermore, various healthcare applications are discussed to show that mobile computing and big data technologies enhance the performance of the system toward improving a humans well-being.

In the era of Internet of Things (IoT) and big data, data has increased dramatically. Computers have been used in various fields. Algorithmic trading is beginning to develop rapidly in the trading market, more and more algorithms begin to be used in the transaction market. As a form of machine learning, neural network can fully reveal the complex trading market. Based on the characteristics of commodity futures market, this paper chooses back propagation neural network to establish price forecasting model. And then, according to the rules of futures market, a self-evolving commodity futures trading strategy is proposed. We also use the data of the Shanghai Futures Exchange and the Dalian Futures Exchange to back-testing the strategy. Finally, we compare the proposed strategies and traditional strategies, and illustrate the evolution of our strategy. Experiments show that our strategies are superior to other compared strategies in the proposed evaluation indicator. Our strategy has a good performance both in yield and risk. It also proves the feasibility of the model and the strategy. The research of this paper is significant to the research of the futures market, and it also provides a new idea for the application of machine learning in algorithmic trading.

Cognitive Internet of Things (IoT) is the use of cognitive computing technologies, which is derived from cognitive science and artificial intelligence, in combination with data generated by connected devices and the actions those devices can perform. Cognitive IoT provides high performance of communicating, computing, controlling, and even high degree of machine intelligence. Cognitive IoT redefines the relationship between human and their pervasive digital environment. They may play the role of assistant or coach for the user. Specifically, the IoT generated big data, when used to power predictive analytics algorithms or to develop a corps for a cognitive computing solution, can provide insights that would never be discovered in time to be useful if the departmental silos do not collaboration in data sensing and analysis. It is the integration of this data that enables cognitive computing applications for IoT of the next decade. Therefore, the services of a cognitive IoT could be constructive, prescriptive, or instructive in nature.

It is entering an era of big data, which facilitated great improvement in various sectors. Particularly, assisted by wireless communications and mobile computing, mobile devices have emerged with a great potential to renovate the healthcare industry. Although the advanced techniques will make it possible to understand what is happening in our body more deeply, it is extremely difficult to handle and process the big health data anytime and anywhere. Therefore, data analytics and mobile computing are significant for the healthcare systems to meet many technical challenges and problems that need to be addressed to realize this potential. Furthermore, the advanced healthcare systems have to be upgraded with new capabilities such as machine learning, data analytics, and cognitive power for providing human with more intelligent and professional healthcare services. To explore recent advances and disseminate state‐of‐the‐art techniques related to data analytics and mobile computing on designing, building, and deploying novel technologies, to enable intelligent healthcare services and applications, this paper presents the detailed design for developing intelligent healthcare systems assisted by data analytics and mobile computing. Moreover, some representative intelligent healthcare applications are discussed to show that data analytics and mobile computing are available to enhance the performance of the healthcare services.

This paper concerns the design of classifiers for medical decision making, and proposes a novel probabilistic neural network classifier with the assistance of a genetic algorithm. Unlike the conventional probabilistic neural networks that use all patterns in data sets as hidden nodes, the proposed neural network adopts some centers through clustering algorithms and the output of data set as the hidden nodes. Comparing with the conventional probability neural network classifiers, the proposed approach significantly decreases the time consuming. Furthermore, a genetic algorithm is utilized to optimize feature selection from the data set. Experimental results are presented on several benchmarks illustrating the relationship between selected features and disease.

Wireless sensor networks deployment and operation are most likely to take place under hazardous conditions. One extreme scenario is the deployment of a wireless sensor network in a mountainous and forested region in which a fire has ignited, for the purpose of localizing and/or tracking, in real time, its spread. Some methods developed in this paper are expected to provide superior performances under these conditions. Existing fires in the deployment region will also affect the nodes’ coverage because the nodes that will fall in the middle of the fire will likely cease to operate. In our work, we develop a coverage hole bypassing algorithm for storing and maintaining holes-information in the network based on which a boundary node can provide the communication path which is more efficient to try to bypass a hole. Simulation results are presented to illustrate the proposed method and evaluate its run-time and message count per node. We show that these routing can be used in many applications.

In order to detect cerebral microbleed more efficiently, we developed a novel computer-aided detection method based on susceptibility-weighted imaging. We enrolled five CADASIL patients and five healthy controls. We used a 20x20 neighboring window to generate samples on each slice of the volumetric brain images. The sparse autoencoder (SAE) was used to unsupervised feature learning. Then, a deep neural network was established using the learned features. The results over 10x10-fold cross validation showed our method yielded a sensitivity of 93.20±1.37%, a specificity of 93.25±1.38%, and an accuracy of 93.22±1.37%. Our result is better than Roy's method, which was proposed in 2015.

Wireless sensor networks deployment and operation are most likely to take place under hazardous conditions. One extreme scenario is the deployment of a wireless sensor network in a mountainous and forested region in which a fire has ignited, for the purpose of localizing and/or tracking, in real time, its spread. Some methods developed in this paper are expected to provide superior performances under these conditions. Existing fires in the deployment region will also affect the nodes' coverage because the nodes that will fall in the middle of the fire will likely cease to operate. In our future work, we show that these routing can be used in many applications.

In recent years, an extensive integration of cyber, physical and social spaces has been occurring. Cyber-Physical-Social Systems (CPSSs) have become the basic paradigm of evolution in the information industry, through which traditional computer science will evolve into cyber-physical-social computational science. Intelligent recommender systems, which are an important fundamental research topic in the CPSS field and one of the key techniques for the implementation of personalized and intelligent computing, have great significance in CPSS development. This paper proposes a group-centric recommender system in the CPSS domain, which consists of activity-oriented group discovery, the revision of rating data for improved accuracy, and group preference modeling that supports sufficient context mining from multiple sources. Through experiments, it is verified that the proposed recommender system is efficient, objective and accurate, thereby providing a strong foundation for personalized computing in the CPSS paradigm.

The intelligent mobile robot platform based on Android features multiple wireless communication functions, which, via Internet and WLAN, may control the movement of the robot. The platform integrates cloud speech recognition and offline speech recognition technologies to control the movement of the robot and have simple man-machine conversation. In addition, with a camera, this platform may realize the remote real-time transmission of video. Since the platform is characterized with sound hardware compatibility and expansibility, we may conduct the research of the robot and rapidly develop an intelligent mobile robot which applies to a specific application scenario on the platform.

Multipath routing in wireless multimedia sensor network makes it possible to transfer data simultaneously so as to reduce delay and congestion and it is worth researching. However, the current multipath routing strategy may cause problem that the node energy near sink becomes obviously higher than other nodes which makes the network invalid and dead. It also has serious impact on the performance of wireless multimedia sensor network (WMSN). In this paper, we propose a pair-wise directional geographical routing (PWDGR) strategy to solve the energy bottleneck problem. First, the source node can send the data to the pair-wise node around the sink node in accordance with certain algorithm and then it will send the data to the sink node. These pair-wise nodes are equally selected in 360° scope around sink according to a certain algorithm. Therefore, it can effectively relieve the serious energy burden around Sink and also make a balance between energy consumption and end-to-end delay. Theoretical analysis and a lot of simulation experiments on PWDGR have been done and the results indicate that PWDGR is superior to the proposed strategies of the similar strategies both in the view of the theory and the results of those simulation experiments. With respect to the strategies of the same kind, PWDGR is able to prolong 70% network life. The delay time is also measured and it is only increased by 8.1% compared with the similar strategies.

With the development of e-commerce, a growing number of people prefer to purchase medicine online for the sake of convenience. However, it is a serious issue to purchase medicine blindly without necessary medication guidance. In this paper, we propose a novel cloud-based medicine recommendation, which can recommend users with top-N related medicines according to symptoms. Firstly, we cluster the drugs into several groups according to the functional description information, and design a basic personalized medicine recommendation based on user collaborative filtering. Then, considering the shortcomings of collaborative filtering algorithm, such as computing expensive, cold start, and data sparsity, we propose a cloud-based approach for enriching end-user Quality of Experience (QoE) of medicine recommendation, by modeling and representing the relationship of the user, symptom and medicine via tensor decomposition. Finally, the proposed approach is evaluated with experimental study based on a real dataset crawled from Internet.