The rapid proliferation of emotion-aware web services has necessitated the analysis of multimodal user interactions. However, this introduces new vulnerabilities where adversaries exploit emotional signals to circumvent fraud detection systems. Despite its improved utility, the robustness of multimodal fraud detection against emotion-driven adversarial manipulation remains significantly underexplored. Existing paradigms often treat emotional cues as static features, overlooking the adversary's capability to strategically modulate multimodal signals (e.g., facial micro-expressions, vocal intonation, and textual styles) to mimic genuine behavior. Furthermore, prevalent evaluations are typically confined to unimodal perturbations and fail to account for context-consistent, cross-modal attacks, thereby compromising system reliability in real-world deployments. To bridge this gap, we propose Context-Emotion Adversarial Training (CEAT), a robust framework designed to fortify multimodal fraud detection against emotion-based attacks. CEAT leverages a Transformer-based architecture to synergistically model emotional features (e.g., visual dynamics and acoustic prosody) alongside semantic context derived from text, yielding a unified representation. Crucially, CEAT introduces a context-aware perturbation mechanism that injects noise into the emotional latent space during training. This process preserves semantic consistency while encouraging the learning of emotion-invariant and discriminative representations. Additionally, a contrastive learning objective is integrated to maximize the distributional divergence between genuine and adversarial samples within the latent manifold. Extensive experiments on multimodal benchmarks demonstrate that CEAT significantly outperforms state-of-the-art baselines, exhibiting superior robustness under simulated emotion-driven attack scenarios.

Modern cloud applications, from interative web services to mobile and WoT workloads, generate highly dynamic multi-tenant network demands. Guaranteeing priority-aware bandwidth remains challenging: legacy shapers like Linux Traffic Control Hierarchical Token Bucket are static and unscalable, while cloud-native solutions such as Cilium offer only coarse-grained rate limiting. We present BeeQoS, a cloud-native QoS system that delivers low-latency, adaptive, and scalable multi-priority bandwidth guarantees. BeeQoS consists of an eBPF-powered data plane for high-performance, fine-grained per-packet shaping, a demand-aware control plane that senses real-time flow requirements and adaptively reallocates bandwidth, and seamless Kubernetes integration for expressive policy specification and cluster-wide scalable deployment. Evaluation shows that BeeQoS scales to 1K+ flows with stable performance, boosts high-/medium-priority throughput by 14.6%/36.4%, cuts median latency by 72.4%, reduces deployment overhead, and improves video QoE by 27.3% over state-of-practice baselines.

While existing far memory and disaggregated memory solutions provide a foundation for addressing limitations of single-node memory capacity and inefficient resource allocation in data centers, they predominantly focus on host memory, overlooking the critical demands of GPU-centric workloads. A key bottleneck in scaling GPU memory is the lack of connectivity and interoperability between GPUs, which is exacerbated by their heterogeneity. To bridge this gap, this paper proposes UHM, a unified data transferring and memory pooling scheme for heterogeneous GPU memories. UHM establishes the communication channels between heterogeneous GPU/host memories and leverages double data buffers for pipelined and reliable transfer. Furthermore, UHM unifies both local and remote memories to build a memory pool. The pooling scheme effectively integrates local and remote resources, performs efficient caching management in local memory, and optimizes memory block management for remote memory resources. Evaluation on a heterogeneous GPU cluster demonstrates that UHM significantly reduces the data transfer latency (up to 87.2%), improves the cache hit ratio, reduces runtime memory allocation latency (up to 94.7%), while enhancing the overall memory utilization (24.7%).

In privacy-preserving mobile network transmission scenarios with heterogeneous client data, personalized federated learning methods that decouple feature extractors and classifiers have demonstrated notable advantages in enhancing learning capability. However, many existing approaches primarily focus on feature space consistency and classification personalization during local training, often neglecting the local adaptability of the extractor and the global generalization of the classifier. This oversight results in insufficient coordination and weak coupling between the components, ultimately degrading the overall model performance. To address this challenge, we propose FedeCouple, a federated learning method that balances global generalization and local adaptability at a fine-grained level. Our approach jointly learns global and local feature representations while employing dynamic knowledge distillation to enhance the generalization of personalized classifiers. We further introduce anchors to refine the feature space; their strict locality and non-transmission inherently preserve privacy and reduce communication overhead. Furthermore, we provide a theoretical analysis proving that FedeCouple converges for nonconvex objectives, with iterates approaching a stationary point as the number of communication rounds increases. Extensive experiments conducted on five image-classification datasets demonstrate that FedeCouple consistently outperforms nine baseline methods in effectiveness, stability, scalability, and security. Notably, in experiments evaluating effectiveness, FedeCouple surpasses the best baseline by a significant margin of 4.3%.

In the paradigm of decentralized learning, a group of agents collaborate to learn a global model using a distributed dataset without a central server; nevertheless, it is severely challenged by the heterogeneity of the data distribution across the agents. For example, the data may be distributed non-independently and identically, and even be noised or poisoned. To address these data challenges, we propose ROSS, a novel robust decentralized stochastic learning algorithm based on Shapley values, in this paper. Specifically, in each round, each agent leverages its local gradient and the cross-gradients from its neighbors (i.e., the derivative of its local loss function and the derivatives of its neighbors’ local loss functions evaluated at its local model) to update its local model in a momentum-like manner, while we innovate in weighting the derivatives according to their contributions measured by Shapley values. We perform solid theoretical analysis to reveal the linear convergence speedup of our ROSS algorithm. We also verify the efficacy of our algorithm through extensive experiments on public datasets. Our results demonstrate that, in face of the above variety of data challenges, our ROSS algorithm have oblivious advantages over existing state-of-the-art proposals in terms of both convergence and prediction accuracy.

Mobile Edge Computing (MEC) utilizes edge servers to offload the computational burden from cloud infrastructure. By providing low-latency and high-bandwidth services, MEC enables mobile users and IoT devices to efficiently offload and execute computational tasks at the network edge. However, optimizing communication and computational resources in a multi-user, multi-server MEC environment remains a significant challenge. In this paper, we propose TSAJS, an efficient multi-server joint task scheduling scheme designed to enhance the effectiveness of MEC offloading. We model the task offloading and resource allocation problem as a Mixed-Integer Nonlinear Programming (MINLP) problem, aiming to maximize user offloading gain by minimizing task completion time and energy consumption. A heuristic algorithm for offloading is introduced by combining threshold-triggering and simulated annealing to effectively avoid local optima and converge toward the global optimum. Meanwhile, the optimal solution for resource allocation is derived using the Karush-Kuhn-Tucker (KKT) conditions. Experimental results demonstrate that TSAJS delivers near-optimal performance, outperforming traditional methods in terms of user offloading effectiveness. Its efficiency enables solution finding within polynomial time, while also adapting to the preferences of users and service providers.

As network scale and complexity continue to increase, managing network configurations has become an increasingly challenging task. Existing configuration tools often depend on low-level, abstract intermediate representations, which require users to have substantial technical expertise. This reliance not only increases the learning curve but also heightens the risk of configuration errors. Recent advances in Large Language Models (LLMs) have demonstrated strong potential for automating tasks across various domains. However, their applications to network configuration generation remain limited due to several challenges, including hallucination, restricted context length, and insufficient adaptability to domain-specific requirements. To address these issues, we propose ConfAgent, an advanced network configuration generation system powered by a multi-model intelligent agent. ConfAgent comprises four key components: a conflict detector, an information extractor, a routing algorithm coder, and a formal synthesizer. These components collaborate to accurately interpret complex configuration intents, detect potential conflicts, and generate robust code and network configurations through intuitive natural language interactions. Extensive experiments conducted on the NetConfEval benchmark demonstrate that ConfAgent consistently outperforms existing state-of-the-art methods by margins ranging from 36 % to 100 %, particularly excelling in configuration tasks for large-scale network topologies.

Network traffic classification plays a crucial role in a wide range of applications, e.g., Quality of Service (QoS) enhancement, resource management, and network security. However, the widespread adoption of encryption protocols (e.g., SSL/TLS) and the emergence of anonymous communication systems (e.g., Tor) have introduced significant challenges due to the presence of complex and varied network noise. Although considerable effort has been made to improve the robustness of the traffic classification, the performances of existing state-of-theart methods cannot be guaranteed in the presence of a mixture of noises, and are not stable in different application scenarios. In this paper, we innovate in proposing a network traffic classification method based on MoA (Mixture of Agents), namely MOTA. By leveraging a light-weight MoA architecture, MOTA efficiently fine-tunes mainstream Large Language Models (LLMs) to adapt to different application scenarios of traffic classification, and fully exploits the collaboration of the LLMs to ensure the robustness against mixed noises. Our extensive experiments show that, the classification accuracy is$\geq 99 {\%}$across multiple public datasets injected with mixed noises, significantly outperforming existing SOTA methods. Moreover, despite incorporating multiple LLMs, MOTA maintains millisecond-level inference latency on a server equipped with four NVIDIA GeForce RTX 4090 GPUs, owing to its lightweight design.

In recent years, with increasing user demands for convenience, privacy, and personalized experiences, gait recognition has been widely studied across various domains, such as indoor intrusion detection and smart homes. Although computer vision solutions are extensively researched for their visual intuitiveness, Wi-Fi sensing is emerging as a new research focus due to its ability to preserve privacy. However, previous studies have primarily relied on abstract features with limited interpretability or required multiple Wi-Fi links. To address these issues, we propose Wi-GR, which utilizes a Wi-Fi link to extract robust and highly interpretable gait features for user recognition. First, we construct a multi-path gait signal model to establish a clear relationship between Channel State Information (CSI) and gait motion. Then, we design a gait signal separation and enhancement method to mitigate the effects of external non-target reflections and internal multi-part reflections, which significantly impact the extraction and interpretability of gait features. Finally, fine-grained gait features that visualize gait patterns are generated using MUSIC-based and GAN-based multi-part velocity profile generation algorithms, tailored for single-person and multi-person scenarios, respectively. Numerous experiments have demonstrated that Wi-GR achieves single-person recognition accuracies of 95.3%, 94.0%, and 93.2% for 30 persons in the meeting room, corridor, and lobby, respectively, and an average accuracy of 88.3% for two-person recognition.

Federated learning (FL) is an effective mobile edge computing framework that enables multiple participants to collaboratively train intelligent models, without requiring large amounts of data transmission while protecting privacy. However, FL encounters challenges due to non-independent and identically distributed (non-IID) data from different participants. The existing methods, whether focusing on local training or global aggregation, often suffer from insufficient unilateral optimization. Achieving effective local-global collaborative optimization, particularly in the absence of additional reference models or datasets, is both crucial and challenging. To address this, we propose a novel approach:Dual-AggregatedFederated learning based on a tripleSiamese network (FedSiam-DA). This method enhances the FL algorithm on both client and server sides. On the client side, we establish a triple Siamese network incorporating a stop-gradient scheme, which leverages a contrastive learning strategy to control the update directions of local models. On the server side, we introduce a dual aggregation mechanism with dynamic weights for local updates, improving the global model’s ability to assimilate personalized knowledge from local models. Extensive experiments on multiple benchmark datasets demonstrate that FedSiam-DA significantly improves model performance under non-IID data conditions compared to existing methods.

Neural network models have become integral to Internet of Things (IoT) systems, with applications spanning from industrial automation to critical infrastructure management. Despite their prevalence, the deployment of these models within IoT systems introduces distinctive security vulnerabilities. In particular, adversaries may execute model poisoning attacks, which aim to alter the decision-making processes of embedded models, leading to erroneous outcomes. Existing model poisoning attacks necessitate access to extensive auxiliary datasets, such as the training dataset itself or one with same distribution. These requirements often render such attacks impractical in IoT contexts, given the constrained storage and computational resources of IoT devices. This paper proposes the first model poisoning attack against interpreters without auxiliary datasets to manipulate the model’s behavior. We evaluate the attack on three real-world datasets, and results indicate that this attack can successfully coerce the targeted interpreters to produce outcomes aligned with an adversary’s intentions, while maintaining nearly indistinguishable performance from the original model, thereby ensuring its stealthiness. Furthermore, beyond directly affected interpreters, our experiments reveal that four additional interpreters coupled to the poisoned model are indirectly influenced, underscoring the attack’s transferability.

Internet of Things (IoT) devices are frequently deployed in highly dynamic environments and need to continuously learn new classes from data streams. Incremental Learning (IL) has gained popularity in IoT as it enables devices to learn new classes efficiently without retraining model entirely. IL involves fine-tuning the model using two sources of data: a small amount of representative samples from the original training dataset and samples from the new classes. However, both data sources are vulnerable to Membership Inference Attack (MIA). Fortunately, the existing MIAs result in poor performance against IL, because they ignore features such as the similarity between old and new models at the old classification layer. This paper presents the first MIA against IL, capable of determining not only whether a sample was used for training/fine-tuning but also distinguishing whether it belongs to the representative dataset or the new classes (unique in IL). Extensive experiments validate the effectiveness of our attack across four real-world datasets. Our attack achieves an average attack success rate of 74.03% in the white-box setting (model structure and parameters are known) and 70.08% in the black-box setting. Importantly, our attack is not sensitive to the IL hyper-parameters (e.g., distillation temperature), confirming its accurate, robust, and practical.

With the convenient design and prolific functionalities, wireless earbuds are fast penetrating in our daily life and taking over the place of traditional wired earphones. The sensing capabilities of wireless earbuds have attracted great interests of researchers on exploring them as a new interface for human-computer interactions. However, due to its extremely compact size, the interaction on the body of the earbuds is limited and not convenient. In this paper, we proposeUi-Ear, a new on-face gesture recognition system to enrich interaction maneuvers for wireless earbuds.Ui-Earexploits the sensing capability of Inertial Measurement Units (IMUs) to extend the interaction to the skin of the face near ears. The accelerometer and gyroscope in IMUs perceive dynamic vibration signals induced by on-face touching and moving, which brings rich maneuverability. Since IMUs are provided on most of the budget and high-end wireless earbuds, we believe thatUi-Earhas great potential to be adopted pervasively. To demonstrate the feasibility of the system, we define seven different on-face gestures and design an end-to-end learning approach based on Convolutional Neural Networks (CNNs) for classifying different gestures. To further improve the generalization capability of the system, adversarial learning mechanism is incorporated in the offline training process to suppress the user-specific features while enhancing gesture-related features. We recruit 20 participants and collect a realworld datasets in a common office environment to evaluate the recognition accuracy. The extensive evaluations show that the average recognition accuracy ofUi-Earis over 95% and 82.3% in the user-dependent and user-independent tasks, respectively. Moreover, we also show that the pre-trained model (learned from user-independent task) can be fine-tuned with only few training samples of the target user to achieve relatively high recognition accuracy (up to 95%). At last, we implement the personalization and recognition components ofUi-Earon an off-the-shelf Android smartphone to evaluate its system overhead. The results demonstrateUi-Earcan achieve real-time response while only brings trivial energy consumption on smartphones.

Submodular optimization has been identified as a powerful tool for many data mining applications, where a representative subset of moderate size needs to be extracted from a large-scale dataset. In scenarios where data points possess sensitive attributes such as age, gender, or race, it becomes imperative to integrate fairness measures into submodular optimization to mitigate bias and discrimination. In this paper, we study the fundamental problem of fair submodular maximization subject to a knapsack constraint and propose the first streaming algorithm for it with provable performance guarantees for both monotone and non-monotone submodular functions. As a byproduct, we also propose a streaming algorithm for submodular maximization subject to a partition matroid and a knapsack constraint, significantly improving the performance bounds achieved by previous work. We conduct extensive experiments on real-world applications such as movie recommendation, image summarization, and maximum coverage in social networks. The experimental results strongly demonstrate the superiority of our proposed algorithms in terms of both fairness and utility.

The surge in popularity of wireless headphones, particularly wireless earbuds, as smart wearables, has been notable in recent years. These devices, empowered by artificial intelligence (AI), are broadening their utility in areas such as speech recognition, augmented reality, pose recognition, and health care monitoring, thereby enriching user experiences through novel interactive interfaces driven by embedded sensors. However, the widespread adoption of wireless earbuds has spurred concerns regarding security and privacy, necessitating robust bespoke security measures. Despite the miniaturization of mobile chips enabling the integration of sophisticated algorithms into smart wearables, the research and industrial communities have yet to accord adequate attention to earbud security. This paper focuses on empowering wireless earbuds to authenticate their legitimate users, tackling the challenges associated with conventional authentication methods. Instead of relying on input interface authentication methods like PIN or lock patterns, this research delves into leveraging Inertial Measurement Unit (IMU) data collected during interactions with devices to extract novel biometric features, presenting an alternative approach that nonetheless confronts challenges related to signal capture and interference. Consequently, we propose and design BudsAuth, an implicit user authentication framework that harnesses built-in IMU sensors in smart earbuds to capture vibration signals induced by on-face touching interactions with the earbuds. These vibrations are utilized to deliver continuous and implicit user authentication with high precision and compatibility across various earbud models. Extensive evaluation demonstrates BudsAuth’s capability to achieve an Equal Error Rate (EER) of 0.0003, representing an approximate 99.97% accuracy with seven consecutive samples of interactive gestures for implicit authentication.

The development of Internet of Things (IoT) networks causes an increasing demand for high-quality video streaming, which results in the burden of traditional mobile cloud computing (MCC) networks, such as high energy consumption and dissatisfaction with the user’s Quality of Experience (QoE). To address this issue, mobile-edge computing (MEC) networks have recently been widely employed for high-quality video streaming scenarios under time-varying wireless channels. However, in MEC networks, limited resources deteriorate the video transmission rate and the quality of videos. Therefore, in this article, we investigate a MEC network with cloud-edge integration for adaptive bitrate (ABR) video streaming, where the edge server (ES) performs cloud-edge selection to decide whether the requested video should be directly obtained from the cloud server (CS) or through a transcoding process. We first design edge caching and transcoding processes to enhance resource utilization and reduce computational consumption through bitrate adaptation strategy and cloud-edge selection decision. Moreover, we utilize the energy efficiency by jointly considering the user’s QoE and energy consumption to measure the network’s performance, and then formulate the optimization objective to maximize energy efficiency. In further, we employ a deep deterministic policy gradient (DDPG)-based scheme to solve the optimization problem by applying video quality adaptation technique, allocating computational resources and transmit power. Finally, simulation results demonstrate that the proposed scheme accomplishes superior energy efficiency compared to the competing schemes at least 41.1%.

With the fast development of the 5G wireless communications, the Internet of Things (IoT) becomes a hot research topic. Unmanned aerial vehicles (UAVs), due to the high mobility and low labor cost, have a big potential to be applied in the future IoT communication networks, e.g., data collection in remote areas. In this paper, we take a UAV as a monitor and an IoT device as an agent, and study how to utilize UAVs to establish network coverage and enhance the overall performance. Given ground agents and aerial monitors, each agent needs to be supervised by one monitor that can at most take charge of certain amount, while such assignment should guarantee the required service quality. This is much different from the conventional assumption that each monitor owns exactly a fixed number of agents without considering sensing quality under limited transmit power. Suppose that a monitor supervises an agent with a cost as the negative value of transmission rate in Rician fading, we then maximize the sum of transmission induced by every monitor-agent connection constrained with workload capacity for each monitor. To achieve the above goals, we first present a fast algorithm to report the least-cost assignment plan. Then, we seek for the minimum number of monitors to maintain the required service quality. Last, we discuss the assignment problem in the scenario of dynamic agents and dynamic monitors. We also give a set of strategies on how to initialize assignment, optimize monitor locations and manage power consumption. Extensive experimental results on both simulated datasets and real-life traffic data demonstrate our effectiveness and high performance.

How to incentivize strategic workers using limited budget is a very fundamental problem for crowdsensing systems; nevertheless, since the sensing abilities of the workers may not always be known as prior knowledge due to the diversities of their sensor devices and behaviors, it is difficult to properly select and pay the unknown workers. Although the uncertainties of the workers can be addressed by the standardCombinatorial Multi-Armed Bandit(CMAB) framework in existing proposals through a trade-off between exploration and exploitation, we may not have sufficient budget to enable the trade-off among the individual workers, especially when the number of the workers is huge while the budget is limited. Moreover, the standard CMAB usually assumes the workers always stay in the system, whereas the workers may join in or depart from the system over time, such that what we have learnt for an individual worker cannot be applied after the worker leaves. To address the above challenging issues, in this paper, we first propose an off-lineContext-Aware CMAB-based Incentive(CACI) mechanism. We innovate in leveraging the exploration-exploitation trade-off in an elaborately partitioned context space instead of the individual workers, to effectively incentivize the massive unknown workers with a very limited budget. We also extend the above basic idea to the on-line setting where unknown workers may join in or depart from the systems dynamically, and propose an on-line version of the CACI mechanism. Specifically, by the exploitation-exploration trade-off in the context space, we learn to estimate the sensing ability of any unknown worker (even it never appeared in the system before) according to its context information. We perform rigorous theoretical analysis to reveal the upper bounds on the regrets of our CACI mechanisms and to prove their truthfulness and individual rationality, respectively. Extensive experiments on both synthetic and real datasets are also conducted to verify the efficacy of our mechanisms.

Recent years have witnessed the fast penetration of Virtual Reality (VR) and Augmented Reality (AR) systems into our daily life, the security and privacy issues of the VR/AR applications have been attracting considerable attention. Most VR/AR systems adopt head-mounted devices (i.e., smart headsets) to interact with users and the devices usually store the users’ private data. Hence, authentication schemes are desired for the head-mounted devices. Traditional knowledge-based authentication schemes for general personal devices have been proved vulnerable to shoulder-surfing attacks, especially considering the headsets may block the sight of the users. Although the robustness of the knowledge-based authentication can be improved by designing complicated secret codes in virtual space, this approach induces a compromise of usability. Another choice is to leverage the users’ biometrics; however, it either relies on highly advanced equipments which may not always be available in commercial headsets or introduce heavy cognitive load to users. In this paper, we propose a vibration-based authentication scheme, VibHead, for smart headsets. Since the propagation of vibration signals through human heads presents unique patterns for different individuals, VibHead employs a CNN-based model to classify registered legitimate users based the features extracted from the vibration signals. We also design a two-step authentication scheme where the above user classifiers are utilized to distinguish the legitimate user from illegitimate ones. We implement VibHead on a Microsoft HoloLens equipped with a linear motor and an IMU sensor which are commonly used in off-the-shelf personal smart devices. According to the results of our extensive experiments, with short vibration signals (≤ 1s ), VibHead has an outstanding authentication accuracy; both FAR and FRR are around 5%.

The last decade has witnessed a tremendous growth of service computing, while efficient service recommendation methods are desired to recommend high-quality services to users. It is well known that collaborative filtering is one of the most popular methods for service recommendation based on QoS, and many existing proposals focus on improving recommendation accuracy, i.e., recommending high-quality redundant services. Nevertheless, users may have different requirements on QoS, and hence diversified recommendation has been attracting increasing attention in recent years to fulfill users’ diverse demands and to explore potential services. Unfortunately, the recommendation performances relies on a large volume of data (e.g., QoS data), whereas the data may be distributed across multiple platforms. Therefore, to enable data sharing across the different platforms for diversified service recommendation, we propose aPrivacy-preserving Diversified Service Recommendation(PDSR) method. Specifically, we innovate in leveraging the Locality-Sensitive Hashing (LSH) mechanism such that privacy-preserved data sharing across different platforms is enabled to construct a service similarity graph. Based on the similarity graph, we propose a novel accuracy-diversity metric and design a 2-approximation algorithm to select$K$services to recommend by maximizing the accuracy-diversity measure. Extensive experiments on real datasets are conducted to verify the efficacy of our PDSR method.

Deep Learning (DL) has been widely adopted in almost all domains, from threat recognition to medical diagnosis. Albeit its supreme model accuracy, DL imposes a heavy burden on devices as it incurs overwhelming system overhead to execute DL models, especially on Internet-of-Things (IoT) and edge devices. Collaborative inference is a promising approach to supporting DL models, by which the data owner (the victim) runs the first layers of the model on her local device and then a cloud provider (the adversary) runs the remaining layers of the model. Compared to offloading the entire model to the cloud, the collaborative inference approach is more data privacy-preserving as the owner’s model input is not exposed to outsiders. However, we show in this paper that the adversary can restore the victim’s model input by exploiting the output of the victim’s local model. Our attack is dubbed Ginver 1: Generative model inversion attacks against collaborative inference. Once trained, Ginver can infer the victim’s unseen model inputs without remaking the inversion attack model and thus has the generative capability. We extensively evaluate Ginver under different settings (e.g., white-box and black-box of the victim’s local model) and applications (e.g., CIFAR10 and FaceScrub datasets). The experimental results show that Ginver recovers high-quality images from the victims.

In this article, we study a distributed privacy-preserving learning problem in social networks with general topology. The agents can communicate with each other over the network, which may result in privacy disclosure, since the trustworthiness of the agents cannot be guaranteed. Given a set of options which yield unknown stochastic rewards, each agent is required to learn the best one, aiming at maximizing the resulting expected average cumulative reward. To serve the above goal, we propose a four-staged distributed algorithm which efficiently exploits the collaboration among the agents while preserving the local privacy for each of them. In particular, our algorithm proceeds iteratively, and in every round, each agent i) randomly perturbs its adoption for the privacy-preserving purpose, ii) disseminates the perturbed adoption over the social network in a nearly uniform manner through random walking, iii) selects an option by referring to the perturbed suggestions received from its peers, and iv) decides whether or not to adopt the selected option as preference according to its latest reward feedback. Through solid theoretical analysis, we quantify the trade-off among the number of agents (or communication overhead), privacy preserving and learning utility. We also perform extensive simulations to verify the efficacy of our proposed social learning algorithm.

We consider a$K$-armed bandit problem in general graphs where agents are arbitrarily connected and each of them has limited memorizing capabilities and communication bandwidth. The goal is to let each of the agents eventually learn the best arm. Although recent studies show the power of collaboration among the agents in improving the efficacy of learning, it is assumed in these studies that the communication graph should be complete or well-structured, whereas such an assumption is not always valid in practice. Furthermore, limited memorization and communication bandwidth also restrict the collaborations of the agents, since the agents memorize and communicate very few experiences. Additionally, an agent may be corrupted to share falsified experiences to its peers, while the resource limit in terms of memorization and communication may considerably restrict the reliability of the learning process. To address the above issues, we propose a three-staged collaborative learning algorithm. In each step, the agents share their latest experiences with each other through light-weight random walks in a general communication graph, and then make decisions on which arms to pull according to the recommendations received from their peers. The agents finally update their adoptions (i.e., preferences to the arms) based on the reward obtained by pulling the arms. Our theoretical analysis shows that, when there are a sufficient number of agents participating in the collaborative learning process, all the agents eventually learn the best arm with high probability, even with limited memorizing capabilities and light-weight communications. We also reveal in our theoretical analysis the upper bound on the number of corrupted agents our algorithm can tolerate. The efficacy of our proposed three-staged collaborative learning algorithm is finally verified by extensive experiments on both synthetic and real datasets.

This paper studies the privacy-preserving adaptive ranking problem for multi-armed-bandit crowdsourcing, where according to the crowdsourced data, the arms are required to be ranked with a tunable granularity by the untrustworthy third-party platform. Any online worker can provide its data by arm pulls but requires its privacy preserved, which will increase the ranking cost greatly. To improve the quality of the ranking service, we propose a Privacy- Preserving Adaptive Ranking algorithm called PPAR, which can solve the problem with a high probability while differential privacy can be ensured. The total cost of the proposed algorithm is ${\mathcal{O}}(K\ln K)$, which is near optimal compared with the trivial lower bound Ω(K), where K is the number of arms. Our proposed algorithm can also be used to solve the well-studied fully ranking problem and the best arm identification problem, by proper setting the granularity parameter. For the fully ranking problem, PPAR attains the same order of computation complexity with the best-known results without privacy preservation. The efficacy of our algorithm is also verified by extensive experiments on public datasets.

This article studies the core maintenance problem in dynamic graphs. The core number is a fundamental index reflecting the cohesiveness of a graph, which is widely used in large-scale graph analytics. The core maintenance problem requires updating the core numbers of vertices after a set of edges and vertices are inserted into or deleted from the graph. Previous works focus on the scenario of single-edge updates and process the edges one by one when multiple edges are inserted/deleted. We initiate the studies of processing multiple edges concurrently to improve the efficiency of core maintenance. Specifically, we discover a structure of inserted/deleted edges,superior edge set, which can be processed together to greatly reduce unnecessarily repeated visits of vertices in the procedure of sequential edge processing. Based on the structure of the superior edge set, efficient algorithms are then devised for incremental and decremental core maintenance, respectively. Compared with single-edge processing algorithms, our algorithms show a significant speedup in the processing time. Furthermore, our algorithms admit parallel implementations, which can further improve the update efficiency. We also conduct extensive experiments on different types of real-world, temporal, and synthetic data sets, and the results illustrate that the proposed algorithms exhibit good efficiency, stability, and scalability.

Crowdsensing is an emerging paradigm of ubiquitous sensing, through which a crowd of workers are recruited to perform sensing tasks collaboratively. Although it has stimulated many applications, an open fundamental problem is how to select among a massive number of workers to perform a given sensing task under a limited budget. Nevertheless, due to the proliferation of smart devices equipped with various sensors, it is very difficult to profile the workers in terms of sensing ability. Although the uncertainties of the workers can be addressed by conventional Combinatorial Multi-Armed Bandit (CMAB) framework through a trade-off between exploration and exploitation, we do not have sufficient allowance to directly explore and exploit the workers under the limited budget. Furthermore, since the sensor devices usually have quite limited resources, the workers may have bounded capabilities to perform the sensing task only few times, which further restricts our opportunities to learn the uncertainty. To address the above issues, we propose a Context-Aware Worker Selection (CAWS) algorithm in this paper. By leveraging the correlation between the context information of the workers and their sensing abilities, CAWS aims at maximizing the expected cumulative sensing revenue efficiently with both budget constraint and capacity constraints respected, even when the number of the uncertain workers is massive. The efficacy of CAWS can be verified by rigorous theoretical analysis and extensive experiments.

This article studies privacy-preserving collaborative learning in decentralized Internet-of-Things (IoT) networks, where the agents exchange information constantly to improve the learnability, and meanwhile make the privacy of agents protected during communications. However, the harsh constraints in IoT make executing collaborative learning much more difficult than well-connected systems composed by servers with strong computation power, due to the weak capacity of devices, limited bandwidth for exchanging information, the asynchronous communication environment, and the necessity of privacy preserving. We show that even if with the harsh constraints in IoT, it still can devise efficient privacy-preserving collaborative learning algorithms, by proposing the first known decentralized collaborative learning algorithm for the fundamental multiarmed bandits problem under the framework of local differential privacy. Rigorous analysis shows that the proposed learning algorithm can make every agent learn the best arm with a high probability and keep the privacy preserved meanwhile. Extensive experiments illustrate that our learning algorithm performs well in real settings.

Dynamicity is one of the most challenging, yet, key aspects of wireless networks. It can come in many guises, such as churn (node insertion/deletion) and node mobility. Although the study of dynamic networks has been popular in distributed computing domain, previous works considered only partial factors causing dynamicity. In this work, we propose a dynamic model that is comprehensive to include crucial dynamic factors on nodes and links. Our model defines dynamicity in terms of localized topological changes in the vicinity of each node, rather than a global view of the whole network. Obviously, a localized dynamic model suits distributed algorithm studies better than a global one. The proposed dynamic model makes use of the more realistic SINR model to describe wireless interference, instead of the oversimplified graph-based models adopted by most existing research. Under the proposed dynamic model, we develop an efficient distributed algorithm accomplishing local broadcast services in the abstract MAC layer that was first presented by Kuhnet al.[24]. Our solution paves the way for many new fast algorithms to solve high-level problems in dynamic networks, such as consensus, single-message broadcast, and multiple-message broadcast. Extensive simulation studies indicate that our algorithm exhibits good performance in realistic environments with dynamic network behaviors.

Age of Information (AoI) is an emerging concept to model information freshness from the perspective of destinations of information deliveries. Serving as a metric to characterize the data delivery timeliness, the peak age indicates the maximum value of the AoI prior to a data packet reception. In this paper, we present a distributed scheduling algorithm for peak age optimization in a wireless network where a number of sensor nodes attempt to deliver their sensed data to a data collector over a wireless channel. In particular, each sensor node accesses the channel for data delivery independently according to an adaptively tuned transmission probability. The beauty of our algorithm lies in that, even with neither centralized infrastructure nor coordinations among the sensor nodes, our algorithm asymptotically approximates the optimal solution by only a constant factor. We perform solid theoretical analysis and extensive simulations to verify the efficacy of our algorithm. To the best of our knowledge, it is the first fully distributed scheduling algorithm for AoI optimization in wireless networks.

Nodes in wireless sensor networks (WSNs) are usually deployed in an unattended even hostile environment. What is worse, these nodes are equipped with limited battery, storage, computation, and communication resources. Therefore, it is challenging to ensure the security of a WSN without decreasing its network performance. Data aggregation (DA) combined with a security mechanism can provide a good scheme for solving the aforementioned problems. This article presents a comprehensive review of secure DA (SDA) in WSNs, including its security goals together with the existing problems. The traditional network topologies as well as new emerging ones are discussed and compared in order to indicate the application scenes and security levels of different topologies. Meanwhile, the contrastive analyses of security strategies are presented which divides SDA protocols into five categories according to different security mechanisms, security goals, and network topologies. Besides, the discussion points out some open issues which may be the valuable topics of SDA in the future.

Emotion detection in online social networks (OSNs) can benefit kinds of applications, such as personalized advertisement services, recommendation systems, etc. Conventionally, emotion analysis mainly focuses on the sentence level polarity prediction or single emotion label classification, however, ignoring the fact that emotions might coexist from users' perspective. To this end, in this work, we address the multiple emotions detection in OSNs from user-level view, and formulate this problem as a multilabel learning problem. First, we discover emotion labels correlations, social correlations, and temporal correlations from an annotated Twitter data set. Second, based on the above observations, we adopt a factor graph-based emotion recognition model to incorporate emotion labels correlations, social correlations, and temporal correlations into a general framework, and detect the multiple emotions based on the multilabel learning approach. Performance evaluation demonstrates that the factor graph-based emotion detection model can outperform the existing baselines.

We study the fundamental problem of fault-tolerant distributed sampling towards uniform probabilistic distribution in dynamic multi-hop wireless networks. Whereas uniform sampling has been extensively studied without concerning fault tolerance, only quite few proposals investigate how the uniform sampling algorithm tolerate Byzantine faults on dynamic networks with very special topologies, e.g., regular graphs with constant node degree. Therefore, designing fault-tolerant uniform sampling algorithms for more general graphs is still an open problem. To this end, we propose a fast and highly fault-tolerate randomized algorithm, such that nearly-uniform sampling is achieved in O(log2n) rounds, while up to O(√n/(polylog(n)·Δ)) Byzantine nodes can be tolerated, where Δ is the maximum degree of the network. Moreover, the proposed algorithm is also communication efficient in the sense that only O(log n) bits need to be exchanged on each link in every round. To show the power of distributed uniform sampling, we apply the proposed algorithm in designing polylogarithmic time distributed algorithms for two typical fundamental issues, i.e., to achieve agreement or data aggregation in Byzantine dynamic networks.

This paper investigates distributed Dominating Set (DS) and Connected Dominating Set (CDS) construction in dynamic wireless networks under the SINR interference model. Specifically, we present a new model for dynamic networks that admits both churns (due to node arrivals/departures) and node mobility. Under this dynamic model, we propose efficient algorithms to construct a DS and a CDS with constant approximation ratios w.r.t. the corresponding minimum ones in O(log n) time with a high probability guarantee. To the best of our knowledge, these algorithms are the first known ones for DS and CDS construction in dynamic networks assuming the SINR interference model. We believe our dynamic network model can greatly facilitate distributed algorithm studies in mobile and dynamic wireless networks.

Most existing proposals for indoor localization are “unnatural,” as they rely on sensing abilities not available to human beings. While such a mismatch causes complications in human-computer interactions and thus potentially reduces the usability and friendliness of a localization service, it is partially entailed by the need for low-cost/effort sensing with resource-limited mobile devices. Fortunately, recent developments in smart glasses (e.g., Google Glasses) signal a trend toward realistic visual sensing and hence make the sensing ability of mobile devices more compatible to that of human users. Leveraging such front-end developments, we propose VisioMap as a natural indoor localization system that intentionally mimics the human skills in visual localization. VisioMap uses very sparse photograph samples to reconstruct 3-D indoor scenes; this is facilitated by the facts that photographs are taken at the eye-level with high stability and regularity, and that the reconstruction is lightweight as it exploits geometric features rather than image pixels. Localization is in turn performed by matching the geometric features extracted online to the reconstructed 3-D scene, making VisioMap: 1) natural to users as they can see the matched 3-D scene and 2) dispensed with the need for dense fingerprints/POIs toward accurate localization.

Whereas a lot of efforts have been put on energy conservation in wireless sensor networks (WSNs), the limited lifetime of these systems still hampers their practical deployments. This situation is further exacerbated indoors, as conventional energy harvesting (e.g., solar) may not always work. To enable long-lived indoor sensing, we report in this article a self-sustaining sensing system that draws energy from indoor environments, adapts its duty-cycle to the harvested energy, and pays back the environment by enhancing the awareness of the indoor microclimate through an “energy-free” sensing. First of all, given the pervasive operation of heating, ventilation, and air conditioning (HVAC) systems indoors, our system harvests energy from airflow introduced by the HVAC systems to power each sensor node. Secondly, as the harvested power is tiny, an extremely low but synchronous duty-cycle has to be applied whereas the system gets no energy surplus to support existing synchronization schemes. So, we design two complementary synchronization schemes that cost virtually no energy. Finally, we exploit the feature of our harvester to sense the airflow speed in an energy-free manner. To our knowledge, this is the first indoor wireless sensing system that encapsulates energy harvesting, network operating, and sensing all together.

Recent large-scale deployments of wireless sensor networks have posed a high demand on network throughput, forcing all (discrete) orthogonal ZigBee channels to be exploited to enhance transmission parallelism. However, the interference from widely deployed WiFi networks has severely jeopardized the usability of these discrete ZigBee channels, while the existing CSMA-based ZigBee MAC is too conservative to utilize each channel temporally. In this paper, we propose ART (Adaptive fRequency-Temporal co-existing) as a framework consisting of two components: FAVOR (FrequencyAllocation for Versatile Occupancy of spectRum) and P-CSMA (Probabilistic CSMA), to improve the co-existence between ZigBee and WiFi in both frequency and temporal perspectives. On one hand, FAVOR allocates continuous (center) frequencies to nodes/links in a near-optimal manner, by innovatively converting the problem into a spatial tessellation problem in a unified frequency-spatial space. This allows ART to fully exploit the “frequency white space” left out by WiFi. On the other hand, ART employs P-CSMA to opportunistically tune the use of CSMA for leveraging the “temporal white space” of WiFi interference, according to real-time assessment of transmission quality. We implement ART in MicaZ platforms, and our extensive experiments strongly demonstrate the efficacy of ART in enhancing both throughput and transmission quality.

Serving as a key step for applications of image processing, superpixel generation has been attracting increasing attention. RGB-D images are used pervasively in scenes reconstruction and representation, benefiting from their contained depth data. In this paper, we present a novel framework for generating superpixels focus on RGB-D images of indoor scenes, based on a weighted geodesic driven metric that combines both color and geometric information. In particular, taking into account the unique structures of indoor scenarios, we first denoise the given RGB-D image, and construct the corresponding triangular mesh. A new weighted geodesic driven metric is defined by introducing a weight function constrained with normal vectors and colors. Under this metric, an energy function is defined to measure our over-segmentation of the triangular mesh, by optimizing which, we can acquire an optimal over-segmentation of the triangular mesh with object boundaries respected, such that vertices in each sub-region have similar geometric structures and color intensities. Re-mapping the over-segmentation of the triangular mesh to the RGB-D image results in desired superpixels. We perform extensive experiments on a large-scale database of RGB-D images to verify the efficacy of our algorithm. The results show that our algorithm has considerable advantages over the existing state-of-the-art methods.

Although the problem of k-area coverage has been intensively investigated for dense wireless sensor networks (WSNs), how to arrive at a k-coverage sensor deployment that optimizes certain objectives in relatively sparse WSNs still faces both theoretical and practical difficulties. Moreover, only a handful of centralized algorithms have been proposed to elevate 2-D area coverage to 3-D surface coverage. In this paper, we present a practical algorithm, i.e., the Autonomous dePlOyment for Load baLancing k-surface cOverage (APOLLO), to move sensor nodes toward k-surface coverage, aiming at minimizing the maximum sensing range required by the nodes. APOLLO enables purely autonomous node deployment as it only entails localized computations. We prove the termination of the algorithm and the (local) optimality of the output. We also show that our optimization objective is closely related to other frequently considered objectives for 2-D area coverage. Therefore, our practical algorithm design also contributes to the theoretical understanding of the 2-D k-area coverage problem. Finally, we use extensive simulation results to both confirm our theoretical claims and demonstrate the efficacy of APOLLO.

Whereas a few physical layer techniques have been proposed to locate a signal source indoors, they all deem multipath a "curse" and hence take great efforts to cope with it. Consequently, each sensor only obtains the information about the direct path; this necessitates a networked sensing system (hence higher system complexity and deployment cost) with at least three sensors to actually locate a source.

Many applications of wireless sensor networks involve monitoring a time-variant event (e.g., radiation pollution in the air). In such applications, fast boundary detection is a crucial function, as it allows us to track the event variation in a timely fashion. However, the problem becomes very challenging as it demands a highly efficient algorithm to cope with the dynamics introduced by the evolving event. Moreover, as many physical events occupy volumes rather than surfaces (e.g., pollution again), the algorithm has to work for 3-D cases. Finally, as boundaries of a 3-D network can be complicated 2-manifolds, many network functionalities (e.g., routing) may fail in the face of such boundaries. To this end, we propose Localized Boundary Detection and Parametrization (LBDP) to tackle these challenges. The first component of LBDP is UNiform Fast On-Line boundary Detection (UNFOLD). It applies an inversion to node coordinates such that a “notched” surface is “unfolded” into a convex one, which in turn reduces boundary detection to a localized convexity test. We prove the correctness and efficiency of UNFOLD; we also use simulations and implementations to evaluate its performance, which demonstrates that UNFOLD is two orders of magnitude more time- and energy-efficient than the most up-to-date proposal. Another component of LBDP is Localized Boundary Sphericalization (LBS). Through purely localized operations, LBS maps an arbitrary genus-0 boundary to a unit sphere, which in turn supports functionalities such as distinguishing interboundaries from external ones and distributed coordinations on a boundary. We implement LBS in TOSSIM and use simulations to show its effectiveness.

While the increasing scales of the recent WSN deployments keep pushing a higher demand on the network throughput, the 16 orthogonal channels of the ZigBee radios are intensively explored to improve the parallelism of the transmissions. However, the interferences generated by other ISM band wireless devices (e.g., WiFi) have severely limited the usable channels for WSNs. Such a situation raises a need for a spectrum utilizing method more efficient than the conventional multi-channel access. To this end, we propose to shift the paradigm from discrete channel allocation to continuous frequency allocation in this paper. Motivated by our experiments showing the flexible and efficient use of spectrum through continuously tuning channel center frequencies with respect to link distances, we present FAVOR (Frequency Allocation for Versatile Occupancy of spectRum) to allocate proper center frequencies in a continuous spectrum (hence potentially overlapped channels, rather than discrete orthogonal channels) to nodes or links. To find an optimal frequency allocation, FAVOR creatively combines location and frequency into one space and thus transforms the frequency allocation problem into a spatial tessellation problem. This allows FAVOR to innovatively extend a spatial tessellation technique for the purpose of frequency allocation. We implement FAVOR in MicaZ platforms, and our extensive experiments with different network settings strongly demonstrate the superiority of FAVOR over existing approaches.

For indoor Wireless Sensor Networks (WSNs), as the conventional energy harvesting (e.g., solar) ceases to work in an indoor environment, the limited lifetime is still a threaten for practical deployment. We report in this demo a self-sustaining indoor sensing system. First of all, given the pervasive operation of heating, ventilation and air conditioning (HVAC) systems indoors, our system harvests energy from airflow introduced by the HVAC systems to power each sensor node. Secondly, as the harvested power is tiny (only of hundreds of μW) such that the exiting sensor products cannot be afforded due to their high energy consumption, we exploit the feature of our harvester to sense the airflow speed in an energy-free manner, which can pay back the environment by enhancing the awareness of the indoor microclimate. We also present two complementary algorithms to synchronize the duty-cycles of the sensor nodes to adapt to the energy harvesting. To our knowledge, this is the first indoor wireless sensing system that encapsulates energy harvesting, network operating, and sensing all together.

Whereas a lot of efforts have been put on energy conservation in wireless sensor networks, the limited lifetime of these systems still hampers their practical deployments. This situation is further exacerbated indoors, as conventional energy harvesting (e.g., solar) ceases to work. To enable long-lived indoor sensing, we report in this paper a self-sustaining sensing system that draws energy from indoor environments, adapts its duty-cycle to the harvested energy, and pays back the environment by enhancing the awareness of the indoor microclimate through an "energy-free" sensing.

Although the problem of k-area coverage has been intensively investigated for dense wireless sensor networks (WSNs), how to arrive at a k-coverage sensor deployment that optimizes certain objectives in relatively sparse WSNs still faces both theoretical and practical difficulties. In this paper, we present a practical algorithm LAACAD (Load balancing k-Area Coverage through Autonomous Deployment) to move sensor nodes toward k-area coverage, aiming at minimizing the maximum sensing range required by the nodes. LAACAD enables purely autonomous node deployment as it only entails localized computations. We prove the convergence of the algorithm, as well as the (local) optimality of the output. We also show that our optimization objective is closely related to other frequently considered objectives. Therefore, our practical algorithm design also contributes to the theoretical understanding of the k-area coverage problem. Finally, we use extensive simulation results both to confirm our theoretical claims and to demonstrate the efficacy of LAACAD.

With the development of ever-expanding wireless sensor networks (WSNs) that are meant to connect physical worlds with human societies, gathering sensory data at a single point is becoming less and less practical. Unfortunately, the alternative in-network data management schemes may fail to operate in the face of communication voids (or holes) in WSNs (especially 3D WSNs). In response to this challenge, we propose harmonic quorum systems (HQSs) as a lightweight data management system for 2D/3D WSNs. HQSs innovate in exploiting a few scalar fields (constructed using pure localized algorithms) to guide data accesses. This liberates HQSs from depending on any routing mechanisms or location services, hence making HQSs efficient and robust against anomalies in WSN topologies. We implement HQSs in TinyOS, and we perform intensive simulations using TOSSIM to validate the performance of HQSs.

This paper proposes novel mechanisms to access sensory data in a distributed fashion in 3D wireless sensor networks. As these networks have their nodes deployed in 3D volumes, we first propose a volume parametrization algorithm to transform irregular volumes into a regular one; it also allows us to extend network protocols from 2D to 3D (which would otherwise be highly non-trivial). Based on this transformation, we propose a new quorum system, 3DQS, to handle distributed data access. The tunability of 3DQS enables it to adapt to different application requirements. We demonstrate the efficacy and efficiency of 3DQS through both analysis and simulations.

A wireless sensor network becomes dynamic if it is monitoring a time-variant event (e.g., expansion of oil spill in ocean). In such applications, on-line boundary detection is a crucial function, as it allows us to track the event variation in a timely fashion. However, the problem becomes very challenging as it demands a highly efficient algorithm to cope with the dynamics introduced by the evolving event. Moreover, as many physical events occupy volumes rather than surfaces (e.g., oil spill again), the algorithm has to work for 3D cases. To this end, we propose UNiform Fast On-Line boundary Detection (UNFOLD) to tackle the challenge. The essence of UNFOLD is to inverse node coordinates such that a "notched" surface is "unfolded" into a convex one, which in turn reduces boundary detection to simple convexity test. UNFOLD is uniform as every node behaves the same (performing coordinate inversion and convexity test), and it is super fast as both computation and communication involve only one-hop neighbors. We prove the correctness and efficiency of UNFOLD; we also use simulations and implementations to evaluate its performance, which demonstrates that UNFOLD is 100 times more time and energy efficient than the most up-to-date proposal.

A new method for constructing triangular patches to pass the C1interpolation conditions (boundary curves and cross-boundary slopes), on the boundary of triangles is presented. The triangular patch is constructed by a basic triangular operator and an error triangular operator. The basic operator is a polynomial of degree six, which approximates the interpolation conditions with a higher approximation precision, while the error operator is constructed by the side-vertex method, which passes the C1error boundary conditions. The C1error boundary conditions are formed by the C1interpolation conditions minus the boundary curves and cross-boundary slopes taken from the basic operator. The basic operator and the error operator are put together to form the triangular patch. Comparison results of the new method with other two methods are included.