Abstract In practical scenarios, security events triggered by abnormal network traffic often result from the collective behavior of multiple data streams, embodying group security events with collective characteristics. Existing research methods, focusing on individual data streams, lack a macroscopic analysis and struggle with challenges of analyzing massive, imbalanced data sets. To address these challenges, this paper adopts a multi-instance learning approach, mapping multiple data streams into a bag with a coarse-grained approach, where each bag corresponds to a security event label and each data stream represents an instance. We propose a multi-instance network traffic conversion method, Bag2Image, which transforms temporal multi-instance network traffic data into image representations, preserving the spatio-temporal characteristics of instances within the bag through image channels and pixels. This strategy allows the network security event prediction task to be approached as an image classification problem, leveraging advanced image classification techniques for prediction. Our cross-experiments with six advanced multi-instance learning (MIL) algorithms and six different classification models demonstrate the superior performance of our method on both the UNSW-NB15 dataset and a private dataset. Specifically, our method achieved the highest F1 scores of 77.9% and 74.4% on these datasets, respectively, representing improvements of 4.1% and 13.5% over the second-best MIL algorithm. The recall rates also saw increases of 4.1% and 13.2%, respectively.

In the industrial IoT scenario, where massive data generation occurs, network traffic classification is crucial for operational security. The Internet of Unmanned Agents (IUA) is an emerging concept within the IoT framework. It focuses on the connectivity and interaction of various unmanned agents, such as drones, autonomous robots, and smart sensors. These unmanned agents collect and transmit large amounts of data in real-time, further contributing to the complexity of data in the IoT environment. The IUA aims to enable seamless cooperation and coordination among these agents, enhancing the overall efficiency and intelligence of industrial operations. The primary challenges in the IUA scenario lie in developing effective models and meeting real-time processing demands. Traditional methods struggle with large, high-dimensional data, while transformer-based models, although achieving good results, are difficult to deploy due to their size, training times, and complex tuning. In this article, we introduce a simple distributed architecture MVTC, which incorporates prior domain knowledge and delivers comparable results to transformer-based models but with shorter processing times and easier deployment. And it does not require large-scale unlabeled data for pretraining, which makes it highly suitable for real-world network traffic classification. The experiments demonstrate that the proposed method outperforms most existing approaches by up to 1.53% while using only 15.26% of the parameters.

This paper presents BM1684X, a cutting-edge AI processor from SOPHGO designed to meet the demanding requirements of broad AI applications. Firstly, we employ SIMD architecture with very large data width to design our TPU to reduce the area ratio of the instruction unit and greatly improves the computing power density. Secondly, the customization of special acceleration instructions within the EU enables the dynamic pipeline execution, leading to a reduction in the total number of instructions and execution time. This customization enhances the performance of TPU in processing RQ and DQ operations, crucial for AI computations. Thirdly, the CUBE array within the TPU implements the multiplication and addition operations of 64 pairs of INT8 operands in the channel dimension of the feature map. By utilizing an addition tree instead of a conventional adder, the implementation significantly reduces both area and power consumption, optimizing the efficiency of TPU. Additionally, the BM1684X processor incorporates a 64-input, 64-output, 8-bit crossbar within the lane, facilitating high-performance data gathering. This crossbar design enhances data gathering capabilities, enabling efficient data processing and manipulation within the TPU architecture. Furthermore, BM1684X offers three distinct memory access modes, showing the processor's versatility in addressing a wide range of AI processing needs and optimizing DRAM utilization for various tasks and workloads. Finally, we design a TPU-MLIR toolchain, highlighting its rich features such as unified processing of multiple frameworks, hierarchical design of model abstractions, correctness guarantees, and traceability of each transformation step. BM1684X excels in providing high-performance computing for a variety of AI models including large models, demonstrating its capabilities through comprehensive evaluations with industry-leading peers.

When workers are heterogeneous in computing and transmission capabilities, the global efficiency of federated learning suffers from the straggler issue, i.e., the slowest worker drags down the overall training process. We propose a novel and efficient federated learning framework named FedPAGE, where workers perform distributed pruning adaptively towards global efficiency, i.e., fast training and high accuracy. For fast training, we develop a pruning rate learning approach generating an adaptive pruning rate for each worker, making the overall update time approximate to the fastest worker’s update time, i.e., no stragglers. For high accuracy, we find that structural similarity between sub-models is essential to global model accuracy in the distributed pruning, and thus propose the CIG_X pruning scheme to ensure maximum similarity. Meanwhile, we adopt the sparse training and design model aggregating of different size sub-models to cope with distributed pruning. We prove the convergence of FedPAGE and demonstrate the effectiveness of FedPAGE on image classification and natural language inference tasks. Compared with the state-of-the-art, FedPAGE achieves higher accuracy with the same speedup ratio.

The satellite-terrestrial networks (STNs) network has the characteristics of open links, node movement, dynamic network topology, and diverse collaborative algorithms, which will lead to frequent passive handovers and continuous reauthentication of user equipment (UE). This causes a waste of valuable computing and storage resources in the STNs network and seriously affects the user’s network experience. To address this, we propose an access authentication protocol with user anonymity and traceability to reduce the communication delay and signaling cost of access authentication. In addition, we also propose a hierarchical group key distribution scheme to implement cross-domain handover authentication between different groups of UE, thereby effectively avoiding reauthentication. We take strict security analysis to prove that the authentication protocol we propose is secure. Compared with the three-related existing arts, our protocol can greatly reduce the communication interaction delay and improve the efficiency of initial access and handover authentication of UE.

Virtual reality (VR) streaming is becoming a popular mobile application that requires ultralow latency and active participation of devices. Meanwhile, blockchain is a promising paradigm to decentralize the traditional ledger of a single trusted entity. In this article, a deep deterministic policy gradient (DDPG)-based scheme is proposed to tackle the joint resource allocation and replica selection challenge of VR streaming in blockchain-enabled fog radio access networks (F-RANs). The proposed scheme balances the load on VR streaming and blockchain maintenance and fully exploits the edge caching and computing resources on fog-based access points (F-APs). Thus, less energy is consumed compared with other learning schemes based on extensive simulations.

The current Internet of Underwater Things (IoUT) for marine observations and emergency responses suffers from two critical issues: 1) energy efficient and 2) timely data collection. Autonomous underwater vehicles (AUVs), serving as tools for collecting and forwarding distributed data, can deal with the unbalanced power consumption in a traditional multihop underwater communication network. However, the low speed of the AUV has not been able to guarantee the timeliness of delay-sensitive data. In this article, we introduce a hybrid data collection scheme (HDCS), taking both real-time data collection and energy efficiency (EE) issues into consideration. All sensor nodes (SNs) are first clustered based on their locations in the network. We develop an analytic expression to describe the attenuation of Value of Information (VoI), involving the relationship between the importance degree and timeliness; initial VoI could be measured by historical data. The emergency can be recognized by the presented criterion, and the transmission mode of cluster heads (CHs) in the same layer is judged by CHs themselves according to VoI. The selected CHs shall transmit the urgent data via multihop routing to avoid over attenuation of VoI. The normal data are collected by AUVs visiting all remaining CHs, and the shortest trajectory is achieved by addressing a variation of the classic traveling salesman problem (TSP). Our simulation experiments show that this mechanism can effectively increase long-term VoI while significantly improving EE.

Nowadays, by integrating the smart devices carried by users with existing communication infrastructures to provide large-scale, fine-grained and complex sensing services, crowdsensing as a novel sensing paradigm has significantly enriched the applications of smart city and promoted the development of Internet of Things (IoT). However, privacy has become skyrocketing concern for crowdsensing and gravely affected the deployment of crowdsensing. In this article, we present a framework to make the tradeoff between minimizing data aggregation error and guaranteeing system stability by jointly considering the privacy of participants, the randomness of sensing task arrival and the cost of platform. We propose an online control mechanism by exploiting Lyapunov stochastic optimization technique. Additionally, considering that, in reality, it always takes different time for different tasks to make sensing decisions, we extend standard Lyapunov stochastic optimization technique to make separate decisions for different types of sensing tasks in consecutive time. Through rigorous theoretical analysis, we prove that our time-average data aggregation error is approximately optimal while still maintaining system stability. By carrying out extensive simulations, we demonstrate the superiority of our proposed mechanisms.

With the rise of the Internet of Things (IoT), the number of mobile devices with sensing and computing capabilities increases dramatically, paving the way toward an emerging paradigm, i.e., crowdsensing that facilitates the interactions between humans and the surrounding physical world. Despite its superiority, particular attention is paid to be able to submit sensing data to the platform wherever possible to avoid leaking the sensitive information of participants and to incentivize them to improve sensing quality. In this article, we propose an incentive mechanism for participants, aiming to protect them from privacy leakage, ensure the availability of sensing data, and maximize the utilities of both platforms and participants by means of distributing different sensing tasks to different participants. More specifically, we formulate the interactions between platforms and participants as a multileader-multifollower Stackelberg game and derive the Stackelberg equilibrium (SE) of the game. Due to the difficulty to obtain the optimal strategy, a reinforcement learning algorithm, i.e., Q-learning is adopted to obtain the optimal sensing contributions of participants. In order to accelerate learning speed and reduce overestimation, a deep learning algorithm combined with Q-learning in a dueling network architecture, i.e., double deep Q network with dueling architecture (DDDQN) is proposed to obtain the optimal payment strategies of platforms. To evaluate the performance of our proposed mechanism, extensive simulations are conducted to show the superiority of our proposed mechanism compared with state-of-the-art approaches.

Community detection for Location-based Social Networks (LBSNs) has been received great attention mainly in the field of large-scale Wireless Communication Networks. In this paper, we present a Differential Privacy Latent Trajectory cOmmunity Discovering (DP-LTOD) scheme, which obfuscates original trajectory sequences into differential privacy-guaranteed trajectory sequences for trajectory privacy-preserving, and discovers latent trajectory communities through clustering the uploaded trajectory sequences. Different with traditional trajectory privacy-preserving methods, we first partition original trajectory sequence into different segments. Then, the suitable locations and segments are selected to constitute obfuscated trajectory sequence. Specifically, we formulate the trajectory obfuscation problem to select an optimal trajectory sequence which has the smallest difference with original trajectory sequence. In order to prevent privacy leakage, we add Laplace noise and exponential noise to the outputs during the stages of location obfuscation matrix generation and trajectory sequence function generation, respectively. Through formal privacy analysis, we prove that DP-LTOD scheme can guarantee ϵ-differential private. Moreover, we develop a trajectory clustering algorithm to classify the trajectories into different kinds of clusters according to semantic distance and geographical distance. Extensive experiments on two real-world datasets illustrate that our DP-LTOD scheme can not only discover latent trajectory communities, but also protect user privacy from leaking.

The Internet of Things has such a profound impact that we have witnessed crowdsensing has emerged as the most popular sensing paradigm where participants sense and aggregate data to the platform by smart devices. However, the participants may not be willing to involve in data sensing and aggregation if they are not sufficiently compensated or their personalized private information are disclosed. In order to overcome the above issues, this article proposes a payment-privacy protection level (PPL) game, where each participant submits his sensing data with a specified PPL while the platform chooses a corresponding payment to the participant. Additionally, we derive the Nash equilibrium point of the game. Considering that the payment-PPL model is unknown in practice, we employ a reinforcement learning technique, i.e., Q-learning to obtain the payment-PPL strategy in a dynamic payment-PPL game. We further use the deep Q network (DQN), which combines a deep-learning technique with Q-learning to accelerate the learning speed. Through extensive simulations, we verify that our proposed algorithm using DQN achieves superior performance in terms of utilities of both platform and participants and data aggregation accuracy compared with the one using Q-learning.

An identifier network (IN), as one of the promising network architectures to solve the IP dual properties problems, has been applied in many areas, including Internet of Things (IoT) for smart cities scenario. The separation mechanisms of an access/core network and identifier/location can benefit IoT in terms of trust management, access control, and privacy protection. The core network in IN is independent of the access network by introducing two namespaces, which makes the core network difficult to be attacked but easy for trust management. However, access network, such as access wireless sensor network (WSN), is facing serious trust and security challenge. Therefore, this article addresses this problem by using network address shuffling. We present an optimization framework of defense cost for IoT security and formulate it as a stochastic cost optimization problem by considering the impacts of network address shuffling control, network autoimmunity control, and defense cost. To improve its generality, we adopt a Lyapunov optimization theory and transform the formulated optimization problem into a queue stability problem, and further decompose the queue stability problem into two subproblems to solve the initial optimization problem. Finally, a novel stochastic cost minimization mechanism (SCMM) consisting of two algorithms for the derived subproblems is proposed. Through logical theoretical analyses, it is proved that our proposed mechanism can achieve the optimized results while maintaining the security of each access WSN and guaranteeing the limited network resources. The extensive simulation results verify that the tradeoff between defense strategy and limited network resources can be well tackled by the balancing factor.

As a novel sensing paradigm, crowdsensing has gained great attention due to large-scale user participation, low cost and wide data source, replacing traditional sensor based sensing in intelligent transportation, environmental monitoring, urban public management, etc. In crowdsensing, however, user privacy leakage is a common but fatal problem, where the participants in crowdsensing might not provide their data if their sensing data expose their personal private information or even lead to malicious attacks. Additionally, it is still challenging for the platform to consider the randomness of sensing task arrival, the dynamic participation of participants and the complexity of task allocation. To this end, an online control mechanism is presented to maximize the profit of platform while guaranteeing system stability and providing personalized location privacy protection. By exploiting Lyapunov optimization theory, we transform the optimization problem into a queue stability problem, decomposing it into three subproblems further. Through rigorous theoretical analysis, we prove that our time-averaged profit is approximately optimal. We also carry out extensive simulations to verify the superiority of our proposed mechanism.

Blockchain-based femtocell networks aim to build decentralized frameworks which enable easy deployment and low power consumption, thus they have been seen promising technologies to make up the coverage of cellular networks in the next generation communication system. This article aims to employ power control to support quality-of-service provisioning, especially the guarantee for the transmission rate of a macrocell user (MUE) and the time delay of femtocell users (FUEs) in two-tier femtocell networks, where the MUE and FUEs share the same communication channel. We formulate the interactions among the macrocell base station and FUEs as a Stackelberg game to maximize the utilities of MUE and FUEs by obtaining the optimal power allocation and pricing strategy. Considering the uncertainty of channel gain which is expressed as a function of transmission distance, we propose a worst-case method to transform the uncertain optimization problem into a deterministic one. We then design two algorithms by considering the dynamics of FUEs, i.e., FUEs may join and leave femtocells. Numerical results verify the convergence and superior performance of our proposed algorithms.

Although the Cognitive Radio Ad‐Hoc Network (CRAHN) is an effective technology to fully utilize the spectrum resource, the appearance of selfish nodes seriously reduces the communication efficiency of CRAHN and generates unfair resource competition. In this paper, a new incentive strategy is proposed to tackle selfish nodes in CRAHN. In our CRAHN model, the Secondary‐User (SU) cooperates with the Primary‐User (PU) in a spectrum leasing mode. Since PU can select multiple SUs as relays but only leases a common authorized spectrum usage time to SUs, the SU has the selfish tendency to reduce its power in relay task, which seriously damage the partnership between PU and SUs. We propose an evaluation coefficient to evaluate the behavior of each SU, where the evaluation coefficient establishes the reward and punishment mechanism to suppress the selfish behavior of SU in relay task. Meanwhile, in order to solve resource allocation problem, a Stackelberg game between PU and SUs is formulated and the optimal solutions are determined in a distributed manner. Simulation results validate that the incentive strategy can effectively suppress the selfish behavior of SUs, in the meantime, the total communication throughput is increased.

The latest increasing popularity of high-speed railways (HSR) has stimulated growing demands for wireless Internet services in HSR networks, especially for video streaming. However, due to the high variability and unpredictability of wireless communications in HSR networks, it is still difficult for the existing solutions to provide high-quality video streaming services to HSR passengers. This paper addresses this crucial problem first by reporting on field experiments performed to investigate the characteristics of HSR networks. Then the paper formulates an intractable optimization problem for dynamic adaptive streaming over HTTP (DASH)-enabling service in HSR networks considering various factors, including packet loss, energy consumption, video service quality, etc. By leveraging Lyapunov optimization approaches, the formulated optimization problem is transformed into a queue stability problem which is of high scalability and generality. Moreover, in order to overcome the intractability of the initial optimization problem, the queue stability problem is further decomposed into three subproblems which can be easily solved individually. Finally, a novel joint stochastic DASH optimization (JSDO) mechanism consisting of three algorithms for the derived subproblems is proposed. Rigorous theoretical analyses and realistic dataset-based simulations demonstrate the effectiveness of the proposed JSDO mechanism.

The recent proliferation of public hotspots has given rise to Wi-Fi advertising where venue owners promote their business by pushing advertisers' advertisements on their hotspots. However, a small business usually has insufficient budget to make a purchase for a whole webpage. Therefore, in this paper, we propose GAWA, a Group-buying based Auction mechanism for Wi-Fi Advertising among a venue owner, group leaders and advertisers, which is composed of three phases. More specifically, in the first phase, we propose an algorithm to decide a group bid for each group leader and winning advertisers for each group. In the second phase, the venue owner assigns venues to group leaders by a novel winning group leader determination algorithm. In the third phase, the mechanism determines how much each winning group leader should charge each advertiser in the winning group. We prove that GAWA is computationally efficient, and possesses excellent economic properties such as individual rationality, budget balance, and truthfulness. We evaluate the proposed algorithms using large-scale simulations, and demonstrate the effectiveness and efficiency of our design when comparing with the state-of-the-art approaches.

Device-to-device (D2D) is a new paradigm in cellular networks that enhances network performance by introducing increased spectral efficiency and reduced communication delay. Efficient data dissemination is indispensable for supporting many D2D applications such as content distribution and location-aware advertisement. In this work, we investigate a new and interesting data dissemination problem where the receivers are not explicitly known and data must be disseminated to the receivers within a probabilistic delay budget. We propose to exploit data depositories, which can temporarily house data and deliver them to interested receivers upon requests. We formally formulate the delay-constrained profit maximization problem for data deposition in D2D networks and show its NP-hardness. Under the unique mobile opportunistic network setting, a practical solution must be distributed, localized, and online. To this end, we introduce three algorithms for Direct Online Selection of 1-Depository, Direct Online Selection of L-Depositories, and Mixed Online Selection of L-Depositories. To demonstrate and evaluate the system, we implement a prototype using Google Nexus handsets and conduct experiments for five weeks. We further carry out simulations based on real-world mobility traces for evaluation of large-scale networks and various network settings that are impractical to experiment.

It is a significant challenge for device-to-device (D2D) networks to deliver mobile videos among mobile users due to the highly nondeterministic and intermittent connectivity. In this paper, we propose to integrate the random mobility of users in mobile opportunistic D2D networks with crowdsourcing to push mobile video Ads. Incentives are key to the success of video Ads push as it heavily depends on how actively mobile users participate in it. To stimulate users to perform mobile video Ads push tasks, we model the interaction between depositories and the Ad provider as a reverse auction. More specifically, we try to maximize the utility of Ad provider by selecting a subset of depositories before a specified deadline. We first propose an online auction (OA) algorithm, which runs efficiently in polynomial time, guarantees individual rationality, profitability. However, it does not guarantee truthfulness and thus limits its practicality. We then introduce two truthful OA algorithms, i.e., TOA and TOA-MM. We carry out trace-driven simulation to verify the three OA algorithms. The simulation results corroborate that the proposed algorithms have superior performance and efficiently stimulate mobile users to make contribution to video Ads push.

A femtocell network which is supported by a macrocell base station and some femtocell base stations provides more reliable transmission, higher wireless capacity, and broader coverage. However, it may face eavesdropping risk, which provides an eavesdropper with a chance to overhear a macrocell user's confidential information. In this paper, we study a secrecy transmission problem for a downlink two-tier femtocell network with imperfect channel state information (CSI), where an eavesdropper wiretaps the legitimate macrocell user. More specifically, we aim to maximize the secrecy rate by jointly optimizing the power allocation and quality-of-service (QoS) requirement in terms of outage probability. We consider two types of CSI, i.e., instantaneous and statistic CSI, respectively, which can robustly guarantee the QoS of users in a complex communication environment. For the instantaneous CSI communication environment, where there exist estimated errors between instantaneous channel gains and their estimated values, we propose a novel conversion method to extract the approximate closed-form expressions of outage probability constraints. For the statistic CSI communication environment, where channel gains obey Rayleigh fading, we design a new method to obtain deterministic expressions by considering the expectations and variances of instantaneous channel gains. Then, the uncertainty and non-convexity of objective function are solved with the aid of variable substitution and Taylor expansion. Moreover, two iterative algorithms are proposed to derive the optimal transmission powers. Finally, we evaluate the proposed algorithms using large scale simulations, and present extensive evaluation results to demonstrate the effectiveness of our proposed algorithms.

This work focuses on the incorporation of incentive stimulations into data dissemination in autonomous mobile social networks with selfish nodes. The key challenge of enabling incentives is to effectively track the value of a message under such a unique network setting with intermittent connectivity and multiple interest data types. We propose two data dissemination models: the data pulling model where mobile users pull data from data providers, and the data pushing model where data providers generate personalized data and push them to the intended users. For data pulling, we present effective mechanisms to estimate the expected credit reward of a message that helps intermediate nodes to evaluate the potential reward of it. Nodal message communication is formulated as a two-person cooperative game, whose solution is found by a heuristic approach which achieves Pareto optimality. Under the data pushing model, “virtual checks” are introduced to eliminate the needs of accurate knowledge about whom and how many credits data providers should pay. The check buying process is formulated as an online auction model to further accelerate the circulation of credits. Extensive simulations carried out based on real-world traces show the proposed schemes achieve better performance than fully cooperative scheme, but significantly reduce communication cost.

Sleep monitoring system helps to diagnose various health problems. Traditional solutions for sleep monitoring are usually invasive or limited to medical facilities. Radio Frequency (RF) based methods require specialized devices or dedicated wireless sensors. Recently, Wi-Fi based methods without any wearable or dedicated devices obtain more attention, however, they all assume that all the users are in a relatively quiet environment without moving targets. In this paper, we develop a system called Mo-Sleep, which adopts off-the-shelf Wi-Fi devices to continuously collect fine-grained wireless Channel State Information (CSI) in a room. We introduce a motion detection module in our system to identify whether the CSI information has been interfered by a moving target. We then use Principal Component Analysis (PCA) to obtain accurate breath signal. Our prototypic system demonstrates that the proposed scheme can not only remove interfered CSI, but also obtain real time breath rate every five seconds.

In this work, we propose a time-sensitive incentive-aware mechanism for mobile video offloading by using the idea of crowdsourcing, where video packet holder cooperates with mobile users to deliver video packets to destination. The objective is to maximize video provider and mobile relay users' payoffs. We formulate the interaction among video packet provider and mobile relay users as a two-person cooperative game, where the video packets are treated as commodities. We apply the Nash bargain solution to obtain the optimal cooperation decision and payment. We carry out extensive simulation based on the real-world traces to validate the superiority of our proposed scheme.

Device-to-device (D2D) is a new paradigm that enhances network performance by offering a wide variety of advantages over traditional cellular networks, e.g., efficient spectral usage and extended network coverage. Efficient data dissemination is indispensable for supporting many D2D applications such as content distribution and location-aware advertisement. In this work, we study the problem of multi-copy data dissemination with probabilistic delay constraint in mobile opportunistic D2D networks. We first formally formulate the problem and introduce a centralized heuristic algorithm which aims to discover a graph for multicasting, in order to meet delay constraint and achieve low communication cost. While the centralized solution can be adapted to a distributed implementation, it is inefficient in a mobile opportunistic D2D network, since it intends to apply a deterministic transmission strategy in a nondeterministic network by delivering all data packets via a predetermined route. Based on such observation, we develop a distributed online algorithm based on the optimal stopping strategy that makes an efficient decision on every transmission opportunity. Extensive simulations under real-world traces and random walk mobility model are carried out to learn the performance trend of the proposed schemes under various network settings.

This work addresses the problem of how to enable efficient data query in a Mobile Ad-hoc SOcial Network (MASON), formed by mobile users who share similar interests and connect with one another by exploiting Bluetooth and/or WiFi connections. The data query in MASONs faces several unique challenges including opportunistic link connectivity, autonomous computing and storage, and unknown or inaccurate data providers. Our goal is to determine an optimal transmission strategy that supports the desired query rate within a delay budget and at the same time minimizes the total communication cost. To this end, we propose a centralized optimization model that offers useful theoretic insights and develop a distributed data query protocol for practical applications. To demonstrate the feasibility and efficiency of the proposed scheme and to gain useful empirical insights, we carry out a testbed experiment by using 25 off-the-shelf Dell Streak tablets for a period of 15 days. Moreover, extensive simulations are carried out to learn the performance trend under various network settings, which are not practical to build and evaluate in laboratories.