Multimodal sentiment analysis (MSA) is an important research topic for understanding human behavior by fusing information across modalities. Attention mechanisms in multimodal fusion highlight key information and improve recognition performance. However, conventional attention-based fusion methods effectively capture salient sentiment information while potentially overlooking subtle cues. Moreover, noise may arise both before and after multimodal fusion, which can readily disrupt downstream tasks. In this article, we leverage neighborhood information to analyze individual samples, extract contextual cues within neighborhoods, and detect residual noise after multimodal fusion. Accordingly, we propose N-AFD, a method that jointly performs alignment, fusion, and denoising. First, we present a cross-modal alignment method that constructs high-purity neighborhoods in the text representation space and aligns other modalities to text. Then, we employ a cross-modal fusion method based on neighborhood attention to capture subtle signals while preserving dominant semantics, and further extend it to a multispace and multiscale variant to enhance representational capacity. Next, we adopt an outlier-aware denoising method that masks isolated samples to reduce their interference with subsequent learning processes. Finally, extensive experiments on two benchmark datasets validate the effectiveness of the proposed N-AFD for the MSA task and demonstrate its competitive performance against other baseline methods.

In Internet of Vehicles (IoV), unmanned aerial vehicles (UAVs) assisted mobile edge computing (MEC) can improve the system performance and communication range of intelligent transportation systems (ITSs). However, the resource allocation and computation offloading in UAVs-assisted IoV systems still face huge challenges due to the growing number of vehicle terminals (VTs), potential privacy leakage, and inefficient problem-solving. Existing solutions cannot adapt to such dynamic multi-UAV scenarios and meet the real-time requirements of VTs. To address these challenges, we propose RACOMU, a novel resource allocation and collaborative offloading framework for multi-UAV-assisted IoV. First, we introduce the convex optimization theory to decouple the original problem and then obtain the near-optimal allocation of transmission power and computing resources by solving the Karush-Kuhn–Tucker (KKT) condition. Next, we design a new collaborative offloading strategy with federated deep reinforcement learning (FDRL), where the offloading requests from VTs are processed in a distributed manner to approach the global optimum while preserving data privacy. Extensive experiments verify the effectiveness of the proposed RACOMU. Compared to benchmark methods, RACOMU achieves better performance in terms of task processing latency, decision-making time, and load balancing degree under various scenarios.

By integrating subgraph learning with federated learning, federated subgraph learning realizes collaborative learning of subgraph information among distributed Unmanned Agents (UAs) while protecting data privacy, offering a promising solution for graph modeling in Internet of Unmanned Agents (IUA). However, due to the various manners of collecting data on different UAs, graph data exhibits the features of Non-Independent and Identically Distributed (Non-IID), while the structures and features of local graph data on UAs are quite diverse. These factors lead to convergence difficulties and insufficient generalization ability of federated subgraph learning during the training process. To address these important challenges, we propose PFedSL, a novel knowledge-sharing Personalized Federated Subgraph Learning framework for IUA. First, a new personalized model aggregation is performed based on the confidence score of UAs and their similarity to reduce the interference of Non-IID data on model performance. Next, a parameter selective activation is introduced for model updating to handle the heterogeneity issue of subgraph structural features. Finally, an original personalized single-view contrastive learning is designed to optimize node embedding, thereby enhancing local representation consistency. Using real-world benchmark graph datasets, extensive experiments demonstrate the superiority of the proposed PFedSL. The results show that PFedSL achieves higher node classification accuracy than state-of-the-art methods in different scenarios. Meanwhile, the effectiveness of the core components in PFedSL is validated via ablation studies.

As a key technique for future networks, the performance of emerging multi-edge caching is often limited by inefficient collaboration among edge nodes and improper resource configuration. Meanwhile, achieving optimal cache hit rates poses substantive challenges without effectively capturing the potential relations between discrete user features and diverse content libraries. These challenges become further sophisticated when caching schemes are exposed to adversarial attacks that seriously impair cache performance. To address these challenges, we introduce RoCoCache, a resilient collaborative caching framework that uniquely integrates robust federated deep learning with proactive caching strategies, enhancing performance under adversarial conditions. First, we design a novel partitioning mechanism for multi-dimensional cache space, enabling precise content recommendations in user classification intervals. Next, we develop a new Discrete-Categorical Variational Auto-Encoder (DC-VAE) to accurately predict content popularity by overcoming posterior collapse. Finally, we create an original training mode and proactive cache replacement strategy based on robust federated deep learning. Notably, the residual-based detection for adversarial model updates and similarity-based federated aggregation are integrated to avoid the model destruction caused by adversarial updates, which enables the proactive cache replacement adapting to optimized cache resources and thus enhances cache performance. Using the real-world testbed and datasets, extensive experiments verify that the RoCoCache achieves higher cache hit rates and efficiency than state-of-the-art methods while ensuring better robustness. Moreover, we validate the effectiveness of the components designed in RoCoCache for improving cache performance via ablation studies.

The digital twin acquires the environment states through device data synchronization, providing robust support for improving collaboration efficiency in edge-end collaborative networks. Grouping is an effective way to reduce competition and simplify scheduling, with multiple devices sharing spectrum resources within the group through power control to synchronize data. However, unreasonable grouping and power control will lead to uneven load and frequent collisions. We propose a data synchronization optimization algorithm for the digital twin based on grouped load balance and mask-assisted power control to handle this problem. First, we allocate periodic devices to groups that maximize the least common multiple of the period in the group to reduce collisions. Event-triggered devices are allocated to groups based on balancing each group’s cumulative average packet generation rate. Second, we apply multi-agent reinforcement learning for power control of grant-free devices. Masks are used to shield prohibited actions to reduce the impact on grant-based devices, and agents can utilize auxiliary experiences to learn the impact of prohibited actions without interaction. Finally, we design some experiments, and the results show that our schemes can reduce collisions while balancing the load and perform better than other schemes after convergence.

Data sharing helps the digital twin builders obtain the necessary data to construct a virtual replica of the physical entity. The consortium blockchain is one of the effective ways to solve the lack of trust among these participants, but the traditional PBFT cannot distinguish abnormal nodes and has poor scalability. We propose an efficient PBFT algorithm based on comprehensive reputation and node grouping for data sharing in digital twins. First, we design a comprehensive reputation evaluation mechanism to distinguish normal and abnormal nodes according to their behavior in the processes of data sharing and consensus. Secondly, a node grouping algorithm is proposed, in which nodes with high reputation form a committee, while other nodes are randomly assigned to different groups. Finally, we design consensus processes separately for the group and committee. Each transaction is recorded in the consortium blockchain when the group and committee have reached a consensus on it in sequence. Experimental results show that the CRNG-PBFT algorithm can effectively identify abnormal nodes, reduce transaction delay, and improve throughput.

With flexible mobility and broad communication coverage, Unmanned Aerial Vehicles (UAVs) have become an important extension of Multi-access Edge Computing (MEC) systems, exhibiting great potential for improving the performance of Federated Graph Learning (FGL). However, due to the limited computing and storage resources of UAVs, they may not well handle the redundant data and complex models, causing the inference inefficiency of FGL in UAV-assisted MEC systems. To address this critical challenge, we propose a novel LightWeight FGL framework, named LW-FGL, to accelerate the inference speed of classification models in UAV-assisted MEC systems. Specifically, we first design an adaptive Information Bottleneck (IB) principle, which enables UAVs to obtain well-compressed worthy subgraphs by filtering out the information that is irrelevant to downstream classification tasks. Next, we develop improved tiny Graph Neural Networks (GNNs), which are used as the inference models on UAVs, thus reducing the computational complexity and redundancy. Using real-world graph datasets, extensive experiments are conducted to validate the effectiveness of the proposed LW-FGL. The results show that the LW-FGL achieves higher classification accuracy and faster inference speed than state-of-the-art methods.

Sentiment analysis plays an indispensable part in human-computer interaction. Multimodal sentiment analysis can overcome the shortcomings of unimodal sentiment analysis by fusing multimodal data. However, how to extracte improved feature representations and how to execute effective modality fusion are two crucial problems in multimodal sentiment analysis. Traditional work uses simple sub-models for feature extraction, and they ignore features of different scales and fuse different modalities of data equally, making it easier to incorporate extraneous information and affect analysis accuracy. In this paper, we propose a Multimodal Sentiment Analysis model based on Multi-scale feature extraction and Multi-task learning (M$^{3}$SA). First, we propose a multi-scale feature extraction method that models the outputs of different hidden layers with the method of channel attention. Second, a multimodal fusion strategy based on the key modality is proposed, which utilizes the attention mechanism to raise the proportion of the key modality and mines the relationship between the key modality and other modalities. Finally, we use the multi-task learning approach to train the proposed model, ensuring that the model can learn better feature representations. Experimental results on two publicly available multimodal sentiment analysis datasets demonstrate that the proposed method is effective and that the proposed model outperforms baselines.

More vehicles are connecting to the Internet of Things (IoT), transforming Vehicle Ad hoc Networks (VANETs) into the Internet of Vehicles (IoV), providing a more environmentally friendly and safer driving experience. Vehicular announcement networks show promise in vehicular communication applications. However, two major issues arise when establishing such a system. First, user privacy cannot be guaranteed when messages are forwarded anonymously, thus the reliability of these messages is in question. Second, users often lack interest in responding to announcements. To address these problems, we introduce a Blockchain-based incentive announcement system called PIAS. This system enables anonymous message commitment in a semi-trusted environment and encourages witnesses to respond to requests for traffic information. Additionally, PIAS uses blockchain accounts as identities to participate in the system with incentives, ensuring privacy in anonymous announcements. PIAS successfully protects the privacy of participants and motivates witnesses to respond to requests. Furthermore, our assessment of security and compatibility shows that PIAS can maintain privacy and incentivization while being compatible with both the Bitcoin and Ethereum blockchains. Further evaluation has confirmed the system's efficiency in terms of performance.

Abstract As the scale of the system expands, processor failures are inevitable. Fault diagnosis has great significance in analyzing the reliability of multiprocessing systems. Probabilistic fault diagnosis is a method that attempts to diagnose nodes correctly with high probability. In this paper, we extend the threshold $t \leq 2$ to threshold $t=3$ for regular networks based on probabilistic diagnosis algorithm and determine the status of a cluster of nodes by analyzing the local performance. Moreover, we evaluate the global performance based on the Poisson distribution and the Binomial distribution and show that the achievement in terms of correctness demonstrates a good performance. Finally, we employ the probabilistic diagnosis scheme to explore some well-known networks, including complete cubic networks, dual cubes and hierarchical hypercubes as well.

Multimodal sentiment analysis aims to extract and integrate information from different modalities to accurately identify the sentiment expressed in multimodal data. How to effectively capture the relevant information within a specific modality and how to fully exploit the complementary information among multiple modalities are two major challenges in multimodal sentiment analysis. Traditional approaches fail to obtain the global contextual information of long time-series data when extracting unimodal temporal features, and they usually fuse the features from multiple modalities with the same method and ignore the correlation between different modalities when modeling inter-modal interactions. In this paper, we first propose an Attentional Temporal Convolutional Network (ATCN) to extract unimodal temporal features for enhancing the feature representation ability, then introduce a Multi-layer Feature Fusion (MFF) model to improve the effectiveness of multimodal fusion, which fuses the different-level features by different methods according to the correlation coefficient between the features, and cross-modal multi-head attention is used to fully explore the potential relationship between the low-level features. The experimental results on SIMS and CMU-MOSI datasets show that the proposed model achieves superior performance on sentiment analysis tasks compared to state-of-the-art baselines.

In this article, we proposedualtraceabledistributedattributebasedencryption withsubsetkeywordsearch system (DT-DABE-SKS, abbreviated as$\mathcal {DT}$) to simultaneously realize data source trace (secure provenance) and user trace (traitor trace) and flexible subset keyword search from polynomial interpolation. Leveraging non-interactive zero-knowledge proof technology,$\mathcal {DT}$preserves privacy for both data providers and users in normal circumstances, but a trusted authority can disclose their real identities if necessary, such as the providers deceitfully uploading false data or users maliciously leaking secret attribute key. Next, we introduce the new conception of updatable and transferable message-lock encryption (UT-MLE) for block-level dynamic encrypted file update, where the owner does not have to download the whole ciphertext, decrypt, re-encrypt and upload for minor document modifications. In addition, the owner is permitted to transfer file ownership to other system customers with efficient computation in an authenticated manner. A nontrivial integration of$\mathcal {DT}$and UT-MLE lead to the distributed ABSE with ownership transfer system ($\mathcal {DTOT}$) to enjoy the above merits. We formally define$\mathcal {DT}$, UT-MLE, and their security model. Then, the instantiations of$\mathcal {DT}$and UT-MLE, and the formal security proof are presented. Comprehensive comparison and experimental analysis based on real dataset affirm their feasibility.

Many existing searchable encryption schemes are inflexible in retrieval patterns. The data usage authorization is almost permanent valid as long as the user is not revoked. This “all-or-nothing” authorization mode is not compatible with the “pay-as-you-use” commercial billing model. In this article, we propose a new notion called time controlled expressive predicate query with accountable anonymity. It realizes time controlled data query, where a time server issues time token to authorize search privilege in designated time period. The data users can anonymously query on encrypted data and the anonymity is accountable in a way that the trusted authority is able to deanonymize data users if they misbehave in the system. The underlying techniques are anonymous credential, Pederson commitment and non-interactive zero-knowledge proof. We firstly design an efficient expressive predicate query (EPQ) scheme, which is proved secure to protect the privacy of expressive search predicate. Based on EPQ, we present a concrete system instantiation, which realizes key-escrow free and time token nontransferability. The formal definition and security models are given out. The system is formally proved indistinguishable against chosen keyword-set attacks, unforgeable of time tokens and accountable of anonymous users. The comparison and experiment results demonstrate its scalability and efficiency.

By sharing resources between edge servers and end users, edge-end cooperation is one important way to support various applications in Internet of Things, which have critical resource requirements on computing, storage, or bandwidth. How to price these resources and how to evaluate the service quality of edge servers are two key issues to support trusted resource allocation for blockchain-enabled Internet of Things. In this article, we provide a trusted resource allocation mechanism based on smart contracts, in which a group-buying pricing mechanism (GBPM) and a reputation evaluation mechanism (REM) are proposed to effectively address the problems existing in resources pricing and service quality evaluation of edge servers. In the trusted resource allocation mechanism, end users can choose a purchase mode from four pricing schemes in terms of actual demands on delay and price, and smart contracts can match end users with high-reputation edge servers automatically. Moreover, end users can submit reputation evaluations to smart contracts based on the behaviors of edge servers. Simulation results show the GBPM can provide differentiated prices and optimize the utility of end users accordingly, while the REM is more sensitive to edge servers with irregular behaviors and quickly reduces their reputations so that the success rate of transactions is improved.

By migrating tasks from the end devices to the edge or cloud, cooperative computing in the Internet of Things can support time-sensitive, high-dimensional, and complex applications while utilizing existing resources, such as the network bandwidth, computing resources, and storage capacity. How to design the multiresource allocation system efficiently is a significant research problem. In this article, we design a multiresource allocation system for cooperative computing in the Internet of Things based on deep reinforcement learning by redefining latency calculation models for communication, computation, and caching with the consideration of practical interference factors, such as the Gaussian noise and data loss. The proposed system uses actor–critic as the base model for rapidly approximating the optimal policy by updating parameters of the actor and critic in respective gradient directions. The balance control parameter is introduced to fit the desired learning rate and actual learning rate. At the same time, we use the method of double experience pool to limit the exploration direction of the optimal policy, which reduces the time complexity and space complexity of the problem solution and improves the adaptability and reliability of the scheme. Experiments have demonstrated that multiresource allocation algorithm based on deep reinforcement learning (DRL-MRA) performs well in terms of the average service latency under resource-constrained conditions, and the improvement is significant with the increase of network size.

In the cloud computing environment with massive information services and decision-making resources, the accuracy and reliability of information are more important than previous single closed systems. Therefore, ensuring the reliability of information and the stable operation of the system are the core problems in the research fields such as the Internet Plus and the Internet of Things. The connectivity and diagnosability are two important measures for the fault tolerance of multiprocessor systems. The g -good-neighbor conditional connectivity ( Rg -connectivity) is the minimum number of nodes that make the graph disconnected, and each node has at least g neighbors in every remaining component. The g -good-neighbor conditional diagnosability ( g -GNCD) is the maximum number of faulty processors that has been correctly identified in a system, and any fault-free processor has no less than g fault-free neighbors. Exchanged X -cubes are a class of irregular networks, obtained by deleting links from hypercubes and some variant networks of hypercubes ( X -cubes). They not only combine the advantages of X -cubes but also reduce the interconnection complexity. Exchanged X -cubes classify its nodes into two different classes clusters with a unique connecting rule. In this paper, we propose the generalized exchanged X -cubes framework so that architecture can be constructed by different connecting rules. Furthermore, we study the Rg -connectivity and g -GNCD of generalized exchanged X -cubes under the PMC and MM ∗ models. As applications, the Rg -connectivity and g -GNCD of generalized exchanged hypercubes, dual-cube-like networks, generalized exchanged crossed cubes, and locally generalized exchanged twisted cubes are determined, respectively.

In this paper we have considered the distributed scheduling problem for channel access in wireless mesh networks. The problem is to assign time-slots for each node in the network to access the control channels so that it is guaranteed that each node can broadcast the control packet to any one-hop neighbor in one scheduling cycle. The objective is to minimize the total number of different time-slots in the scheduling cycle. In the single-channel ad hoc networks, the known best result for this problem is proved to be (K2+ 1) in arbitrary graphs and 25K in unit disk graphs with K as the maximum node degree. The original contributions of this paper are that it has taken the large interference range problem into consideration for the first time and proposed two algorithms for the scheduling problem, namely, the one neighbor per cycle (ONPC) algorithm and the all neighbors per cycle (ANPC) algorithm. We prove that the number of time-slots by the second algorithm is upper-bounded by 2K in some special case. The fully distributed versions of these algorithms are given in this paper. Simulation results also show that the performance of ANPC is rather better than ONPC.

In sensor networks, energy efficiency is crucial to achieving satisfactory network life. Using the strategy of data aggregation and the technology of smart radio with adjustable transmission power, energy can be saved significantly. In this work we model the real-time requirement in sensor networks as two constraints with the data aggregation tree: node degree bounded and tree height bounded. We state with energy model as the FIRST ORDER RADIO MODEL [4], the maximum node degree of the MST for any graph in a plane is six, and it can be transformed into a MST with maximum node degree as five. Then, we propose three heuristic algorithms to build a MST with hop and degree constraints, namely Node-First Heuristic (NFH), Tree-First Heuristic (TFH), and Hop-Bounded Heuristic (HBH). Simulation results reveal that they are all suitable to solve the real-time data aggregation problem and the performance of NFH is the best.