The rapid proliferation of diverse IoT scenarios necessitates intelligent attack detection as a critical safeguard for IoT security. However, the massive volumes of heterogeneous data generated by IoT devices, coupled with increasingly sophisticated covert attack vectors, pose significant challenges to ensuring robust security. A pivotal challenge lies in effectively extracting discriminative features from IoT device data to accurately classify normal and malicious activities. To address this, we propose WaveMamba, a novel IoT attack detection model that integrates wavelet transforms into a Mamba-based architecture for enhanced feature extraction. The wavelet transform provides superior time-frequency localization capabilities, decomposing raw signals into multi-scale approximation and detail coefficients. Meanwhile, the Mamba module, built upon continuous state-space models (SSMs), excels at capturing long-range dependencies in sequential data. Our framework strategically embeds wavelet analysis within the Mamba architecture to achieve synergistic fusion of time-frequency representations and multi-scale modeling advantages. Experimental results demonstrate that WaveMamba significantly outperforms existing methods. This advancement holds substantial promise for elevating the precision of intelligent attack detection in the AI era.

The Traveling Salesman Problem (TSP), a well-known NP-hard problem, is widely applied across manufacturing, biology, transportation, and other fields. In recent years, deep reinforcement learning (DRL), particularly DRL based on Transformer architectures, has emerged as a popular approach to solving TSP. However, the quadratic computational and space complexities of Transformer models pose challenges for handling larger-scale TSP instances. This paper introduces Sym-Favformer, an end-to-end DRL method based on symmetry regularization in the Transformer. In this approach, the encoder incorporates a Fast Attention Via positive Orthogonal Random features (FAVOR+) mechanism with linear complexity, while the decoder adopts a distance-based dynamic selection mechanism, and employs the Lion optimizer to reduce training time and memory consumption. To enhance performance and generalization, Sym-Favformer applies a random transformation data augmentation technique within the encoder and introduces enhanced context embedding in the decoder. Subsequently, the model is trained using a symmetry-regularized REINFORCE method. Experimental results show that Sym- Favformer significantly outperforms existing end-to-end DRL methods on both randomly generated benchmark instances and TSPLIB datasets, demonstrating strong generalization to larger-scale TSP instances.

Network management and security are significant for the development of the Internet of Things (IoT), The traditional traffic identification and detection methods face growing challenges with the widespread deployment of IoT devices, the extensive use of encryption technologies, and the continuous increase in malicious network traffic. These challenges further complicate traffic monitoring and malicious behavior detection in network management. Consequently, research is increasingly focusing on deep learning techniques to tackle these issues. Although new traffic classification models based on deep learning methods such as CNNs and Transformers have been introduced, they still encounter issues like insufficient feature extraction and poor performance with similar traffic types. To address these challenges, this paper introduces a malicious traffic classification method based on 1D ConvNeXt and multi-scale patches. This novel approach applies 1D ConvNeXt to malicious traffic classification, integrating 1D depth wise sep-arable convolution and multi-scale patches to capture spatio-temporal traffic features. Our method achieves 99.56% and 99.33% accuracy on the USTC-TFC2016 and CICIoT2022 datasets, respectively, significantly reducing sample confusion among highly similar traffic types. The results show that our approach outperforms the other baseline models.

Link delay is a key factor to evaluate and ensure the stringent network service quality required by the Industrial Internet of Things (IIoT). Because link delay is seriously affected by traffic, obtaining link delay features associated with network traffic is important. This article presents a traffic-associated link delay learning solution for the IIoT. In our solution, the network of the IIoT is divided into many local networks and a software-defined network (SDN). Our solution uses low-loaded methods to collect traffic-delay samples, and uses a traffic-interval-based mechanism to solve the traffic-associated delay statistics problem. We present a link traffic-delay model learning method for local networks of the IIoT. This method uses path traffic-delay samples, independent from specific network paradigms. Our solution uses a particular deep neural network structure to explore the information implied in path traffic-delay samples. We also propose a link traffic-delay model learning method for the SDN, which selects source links by a feature-similarity-based method and generates link traffic-delay models based on transfer learning. Our solution evaluates the accuracy of link traffic-delay models, and further improves the models with low accuracy.

Blockchain is renowned for its immutability, a feature that ensures recorded data cannot be modified or deleted. However, malicious entities can exploit this immutability to permanently embed objectionable data. Moreover, the immutability of blockchain can conflict with the “right to be forgotten” provision in the privacy protection laws of the GDPR. Therefore, a seminal redactable blockchain solution is proposed to address the above problem. Recently, one of the primary focuses in redactable blockchain solutions is the design of global editing permission control. The design leverages the consensus voting mechanism, with its core objective being to prevent excessive centralization of editing power, thereby preserving the decentralized nature of blockchain technology. Meanwhile, such schemes exhibit the following phenomena: (i) the members of the editorial decision-making group are fixed and unchanging, and (ii) blockchain nodes display lazy voting behaviors during the editorial voting process. Considering these factors, we introduce a Publicly Redactable Blockchain scheme with an off-chain reputation-based Consensus Protocol (PRBCP). In this scheme, any node in the blockchain network has the opportunity to become an editorial node and perform editing operations. We design a reputation-based off-chain editorial voting consensus protocol leveraging the threshold signature scheme, which enables dynamic updates to the editorial decision-making group membership and enhances nodes’ participation in editorial voting. In addition, we conduct rigorous security proofs and experimental efficiency analyses for our scheme. The results demonstrate that the PRBCP is both secure and efficient. Finally, we instantiate our scheme as a redactable medical blockchain (RMB) system for storing electronic medical records (EMRs).

In the Industrial Internet of Things, it is necessary to reserve enough bandwidth resources according to the maximum traffic peak. However, bandwidth reservation based on the maximum traffic peak leads to low resource utilization. In this paper, we propose a data transfer optimization solution, based on the cooperation of different entities in the local area, which strives to deliver data acquired by sensors to the cloud in a reliable manner and improve bandwidth utilization to save limited network resources. In our solution, the data transfers from the sensors in a local network are controlled by a local controller and some edge gateways with acceptable cost such that no congestion occurs in the path to the cloud and the bandwidth requirement of each flow can be met. To obtain a tradeoff between resource utilization and transfer delay, we study the problem of minimizing the maximum rate peak of periodic real-time traffic from distributed sensors and propose an algorithm to solve this problem with a desirable lower boundary of the performance. In addition, we design an application-level forwarding method that significantly improves resource utilization and a method of implementing reliable sampling instant adjustment. The experimental results show that our solution significantly improves resource utilization without producing network congestion.

In recent years, visual tracking methods based on convolutional neural networks and transformers have achieved remarkable performance and have been successfully applied in fields such as autonomous driving. However, the numerous security issues exposed by deep learning models have gradually affected the reliable application of visual tracking methods in real-world scenarios. Therefore, how to reveal the security vulnerabilities of existing visual trackers through effective adversarial attacks has become a critical problem that needs to be addressed. To this end, we propose an adaptive meta-gradient adversarial attack (AMGA) method for visual tracking. This method integrates multimodel ensemble and meta-learning strategies, combining momentum mechanisms and Gaussian smoothing, which can significantly enhance the transferability and attack effectiveness of adversarial examples. AMGA randomly selects models from a large model repository, constructs diverse tracking scenarios, and iteratively performs both white- and black-box adversarial attacks in each scenario, optimizing the gradient directions of each model. This paradigm minimizes the gap between white- and black-box adversarial attacks, thus achieving excellent attack performance in black-box scenarios. Extensive experimental results on large-scale datasets, such as OTB2015, LaSOT, and GOT-10 k demonstrate that AMGA significantly improves the attack performance, transferability, and deception of adversarial examples.

In addressing the multi-path problem of cloud-edge data transmission, this paper introduces a multi-path orchestration method based on deep reinforcement learning (DRL) within the context of software-defined networks (SDN), termed Deep Q-Routing (DQR). This method establishes optimal data transmission paths between the cloud and multiple edges, based on available bandwidth and transmission delay, while avoiding path congestion to improve bandwidth utilization and reduce transmission delay. The optimal routing in SDN is defined through the design of state space sets, reward functions, and neural networks. Simulation experiments were conducted to evaluate the DQR method. Compared to existing routing methods, the proposed DQR method, which includes a reward function with weighted parameters, demonstrates superior overall network performance.

Encrypted traffic classification is a challenging task involving precisely categorizing various types of network traffic and applications, a challenge exacerbated by the continuous emergence of new applications. Traditional methods predominantly rely on feature extraction from datasets and the utilization of complex deep learning models for classification, necessitating larger datasets and more intricate computational models. The challenge lies in balancing leveraging unlabeled traffic to enhance classification datasets and controlling model complexity. In response to these challenges, this paper introduces ETKD, an encrypted traffic classification framework based on knowledge distillation. ETKD is designed to address the trade-off between classification accuracy and model complexity. The framework capitalizes on image recognition techniques to extract encrypted traffic features from extensive unlabeled traffic data. Subsequently, these features are fine-tuned using a teacher-student model, achieving a harmonious equilibrium between reducing model complexity and maintaining high classification accuracy. The experimental results presented in this paper provide compelling evidence of the framework's effectiveness. Compared to state-of-the-art models, the minimal model employed in our approach achieves notably higher accuracy and F1-score, underscoring the practicality and efficiency of the proposed ETKD method.

Delays are crucial factors in the service management of networks, especially software-defined networks. Unfortunately, it is very difficult to accurately model a traffic-delay mapping without any assumptions on an uncertain network. In this article, we present a machine learning-based solution to generate a mapping between link traffic and link delay in software-defined networks. The proposed solution only requires a small number of link delay samples from the production network. The small number of link delay samples is not sufficient for learning link traffic-delay mapping. To solve the above problem, we extend the link delay-related data via a sample transfer method and a distributed path delay data collection method without the assistance of the controller. We design a link traffic-delay mapping learning solution using the above three classes of data. This solution uses a traffic segment-based statistical mechanism to deduce the mean link delay effectively from the collected path delay information and implements effective sample transfer via a distance-based approximation. On the basis of specially designed deep learning structures and training procedures, the proposed learning solution effectively builds traffic-delay mapping models using the samples transferred from an experimental network and the samples of the production network.

With the widespread application of network traffic encryption, traffic identification has become increasingly critical. As the types of encryption protocols continue to grow, identifying encrypted traffic with limited training samples has become more challenging. In recent years, pre-training models have been extensively applied in natural language processing due to their ability to utilize a large amount of unlabeled data effectively. However, when applied to encrypted traffic identification, these methods lack sufficient information extraction from encrypted network traffic, resulting in the loss of some essential features and negatively impacting the recognition performance of such approaches. Therefore, we proposed an encrypted traffic classification model based on a Transformer named FlowBERT. In FlowBERT, the semantic features of the traffic can be learned from two dimensions: payload and packet length sequence in large-scale, unlabeled encrypted traffic scenarios. Length sequences are encoded to extract traffic sequence features efficiently, enabling the model to learn the contextual semantic relationships within the sequences. Simultaneously, the pre-training process is improved by balancing data samples, enhancing the performance of the pre-training model. We validated the performance of this method on both classic encrypted traffic classification datasets and the novel network protocol DoH dataset. We concluded that our approach demonstrates robustness and superior recognition performance compared to similar methods.

We study the intelligent offloading problem for a multiple unmanned aerial vehicle (multi-UAV)-assisted mobile-edge computing (MEC) system in an MEC scenario where a natural disaster has damaged the edge server. The study has two steps. First, the task offloading destination is determined by minimizing the total energy consumption of the multi-UAVs in the system. We propose the server selection game-theoretic (SSGT) algorithm and demonstrate its convergence through simulation experiments. Second, we propose an offloading incentive mechanism to price computing resources for a single unmanned aerial vehicle (UAV)-MEC server. Considering the UAV’s power consumption and mobile users’ willingness, we model the interaction between the UAV-MEC server and mobile users as a Stackelberg game. We prove the existence of a Nash equilibrium by theoretical analysis and experimental verification and design the multiround iterative game (MRIG) algorithm based on arithmetic descent to achieve the optimal solution, i.e., the utility tradeoff between the UAV-MEC server and mobile users. Finally, the simulation results show that our proposed scheme can increase the value of overall user satisfaction (SoU) more than other schemes, which proves that the incentive mechanism of resource pricing can supply computing power support for ground mobile users in a UAV-assisted MEC system more effectively.

Service intention refers to what service or which service the server provides. The former includes service classification, service type, or service behavior classification. The latter contains service content classification, such as shopping online or uploading and downloading, etc.. Port-based classification and payload-based classification are two widely used service classification schemes, both of which have many limitations, such as only focusing on server-side scenarios or just designing for non-encrypted requests. In this paper, we propose an encryption-independent approach from a network-side perspective by analyzing the communication behavior of the IPs. Firstly, we identify similar service behavior clusters by employing service influence metrics. Then, we devise a semantic mining mechanism to infer whether they serve a fixed user group or provide interactive service. Finally, we use open-source benchmark datasets, synthetic datasets, and the real Netflow data collected from the China Education Research Network backbone (CERNET) to verify our proposal. Experimental results demonstrate that the accuracy and recall rate of the proposed approach is better than other similar state-of-the-art methods. Besides, our work can also distinguish malicious behavior clusters. Extensive experiments demonstrate that our work is efficient for network management and security monitoring.

With the potential of implementing computing-intensive applications, edge computing is combined with digital twinning (DT)-empowered Internet of vehicles (IoV) to enhance intelligent transportation capabilities. By updating digital twins of vehicles and offloading services to edge computing devices (ECDs), the insufficiency in vehicles’ computational resources can be complemented. However, owing to the computational intensity of DT-empowered IoV, ECD would overload under excessive service requests, which deteriorates the quality of service (QoS). To address this problem, in this article, a multiuser offloading system is analyzed, where the QoS is reflected through the response time of services. Then, a service offloading (SOL) method with deep reinforcement learning, is proposed for DT-empowered IoV in edge computing. To obtain optimized offloading decisions, SOL leverages deep Q-network (DQN), which combines the value function approximation of deep learning and reinforcement learning. Eventually, experiments with comparative methods indicate that SOL is effective and adaptable in diverse environments.

With the advent of the Internet of Things (IoT) and the increasing popularity of the intelligent transportation system, a large number of sensing devices are installed on the road for monitoring traffic dynamics in real time. These sensors can collect streaming traffic data distributed across different traffic sites, which constitute the main source of big traffic data. Analyzing and mining such big traffic data in massive IoT can help traffic administrations to make scientific and reasonable traffic scheduling decisions, so as to avoid prospective traffic congestions in the future. However, the above traffic decision making often requires frequent and massive data transmissions between distributed sensors and centralized cloud computing centers, which calls for lightweight data integrations and accurate data analyses based on large-scale traffic data. In view of this challenge, a big data-driven and nonparametric model aided by 6G is proposed in this article to extract similar traffic patterns over time for accurate and efficient short-term traffic flow prediction in massive IoT, which is mainly based on time-aware locality-sensitive hashing (LSH). We design a wide range of experiments based on a real-world big traffic data set to validate the feasibility of our proposal. Experimental reports demonstrate that the prediction accuracy and efficiency of our proposal are increased by 32.6% and 97.3%, respectively, compared with the other two competitive approaches.

A least squares support vector machine (LS-SVM) offers performance comparable to that of SVMs for classification and regression. The main limitation of LS-SVM is that it lacks sparsity compared with SVMs, making LS-SVM unsuitable for handling large-scale data due to computation and memory costs. To obtain sparse LS-SVM, several pruning methods based on an iterative strategy were recently proposed but did not consider the quantity constraint on the number of reserved support vectors, as widely used in real-life applications. In this article, a noniterative algorithm is proposed based on the selection of globally representative points (global-representation-based sparse least squares support vector machine, GRS-LSSVM) to improve the performance of sparse LS-SVM. For the first time, we present a model of sparse LS-SVM with a quantity constraint. In solving the optimal solution of the model, we find that using globally representative points to construct the reserved support vector set produces a better solution than other methods. We design an indicator based on point density and point dispersion to evaluate the global representation of points in feature space. Using the indicator, the top globally representative points are selected in one step from all points to construct the reserved support vector set of sparse LS-SVM. After obtaining the set, the decision hyperplane of sparse LS-SVM is directly computed using an algebraic formula. This algorithm only consumes O(N2) in computational complexity and O(N) in memory cost which makes it suitable for large-scale data sets. The experimental results show that the proposed algorithm has higher sparsity, greater stability, and lower computational complexity than the traditional iterative algorithms.

With the rapid development of the next-generation sequencing technology, a large amount of genomics information has been obtained. The scale of biological sequencing data is particularly large and complex. The tensor robust principal component analysis (TRPCA) method can effectively preserve the spatial structure of tensor data, so it has received extensive attention. However, the low-rank tensor obtained by TRPCA may be damaged to a certain extent. To solve this problem, this paper proposes a model for weighting low-rank data based on the method of TRPCA. This model has an additional constraint penalty term that can repair corrupted low-rank data and the effective information in it can be fully utilized. In addition, the norm is used to constrain the sparse tensor to make the sparse effect better. In the experimental part, TRPCA model clusters samples by low-rank tensor. The experimental results on cancer omics data show that our method is superior to other methods.